The workshop took place in the Natcher Auditorium, National Institutes of Health, Bethesda, Maryland, at 8:30 a.m., Andrew Lewis, Chair, presiding.

PRESENT:

TOM KELLY, Co-Chair
ARTHUR LEVINE, Co-Chair
BRIAN MAHY, Ph.D., Co-Chair
ROBIN WEISS, Co-Chair
ROBERT BRIGHT, Ph.D., Speaker
MICHELLE CARBONE, M.D., Ph.D., Speaker
JAMES COOK, Speaker
JAMES DeCAPRIO, M.D., Speaker
PRISCILLA FURTH, Speaker
HARVEY OZER, Speaker
HARVEY PASS, Speaker
JAMES PIPAS, Speaker
MARY KATHLEEN RUNDELL, Speaker
SATVIR S. TEVETHIA, Ph.D., Speaker
JANET S. BUTEL, Ph.D., Panelist
KRISTINA DOERRIES, Panelist
RICHARD FRISQUE, Ph.D., Panelist
ROBERT GARCEA, M.D., Panelist
FRANK O'NEILL, Ph.D., Panelist
KEERTI V. SHAH, Panelist
HOWARD STRICKLER, Panelist

AGENDA

Panel-Audience Discussion
Moderator: ARTHUR LEVINE, NICHD
Topic: SV40 As A Putative Human Commensal
Panel Members: Janet Butel, Kristina Dorries, James Goedert, Richard Frisque, Robert Garcea, Anthony Morris, Frank O'Neill, Keerti Shah

Session 4 Part 1: Mechanisms of SV40 Oncogenesis 1
Chair: BRIAN MAHY, CDC

SV40 Oncogenicity in Hamsters,
MICHELE CARBONE

SV40-Rodent Tumor Models as Paradigms of Human Disease: Transgenic Mouse Models,
PRISCILLA FURTH

SV40 Transformation of Rodent Cells in Vitro,
KATHLEEN RUNDELL

SV40 Transformation of Human Cells in Vitro,
HARVEY OZER

Experimental Tumor Induction by SV40 Transformed Cells,
JAMES COOK

Session 4 Part 2: Mechanisms of SV40 Oncogenesis 2
Chair: TOM KELLY, Johns Hopkins

SV40 DNA replication and Transformation Requires the DnaJ Chaperone Domain of Large T Antigen,
JAMES DeCAPRIO

SV40 Large T Antigen Functions as a DnaJ Molecular Chaperone: Implications for Tumorigenesis,
JAMES PIPAS

SV40-IGF-1r Mechanisms: Studies in an SV40 Induced Hamster Mesothelioma Model,
HARVEY PASS

A Strategy for Assessing CTL Responses to SV40 T Antigen in Humans,
SATVIR TEVETHIA

Immunotherapy of SV40-Induced Tumors in Mice: A Model for Vaccine Development,
ROBERT BRIGHT

Panel-Audience Discussion 3:
Moderator: ROBIN WEISS, Chester Beatty Laboratories
Topic: SV40 As An Oncogenic Virus and Possible Human Pathogen:
Panel Members: Antonio Giordano, Priscilla Furth, James DeCaprio, James Pipas, Harvey Ozer, Howard Strickler, Antonio Procopio, Robert Garcea, George Klein

Concluding Remarks

PROCEEDINGS

(8:38 a.m.)

DOCTOR LEVINE: Good morning. I'm Art Levine. I'm going to moderate this morning's panel session. We seem to have a somewhat smaller audience, possibly having shed the lawyers and the reporters, but we are down to science.

For those of you who weren't here yesterday, I would like to take just a moment and summarize what I believe to have been the major conclusions from yesterday's meeting.

First, SV40, at least in the form of its DNA, is or is not present in human tumors, and is or is not present in normal human tissues. And, we heard compelling data on both sides of that question, all from good labs, and I think that the question will only be resolved by an appropriately blinded study.

The reason for the disagreement probably lies with the complexity of the PCR assay used to detect the SV40 footprint in modern times, and the sensitivity, and specificity and nuances of that assay.

But, even if the SV40 footprint is in human cells, there was no evidence from the strong epidemiologic studies presented by Doctors Strickland and Olin that any apparent harm occurred as a consequence of the apparently massive exposure to SV40 in the early era of poliovirus vaccines. And, in fact, one might hypothesize that if SV40 were truly harmful for human beings, and it had been harmful right along as an endemic agent, then surely the rates of some cancers should have increased during the massive exposure to SV40 in the poliovirus vaccine. And, the fact that there was no increase in any rate that we could see, granted the epidemiologic studies aren't perfect, given that there was no increase in the signal rate of any individual tumor assayed, then one might comfortably say that SV40, in fact, is not a human pathogen.

However, that leaves open the question of whether it's a human Commensal. Is this an agent that we live with, and that we've always lived with, independent of the poliovirus vaccine exposure.

May I have the first slide, please?

Okay. Well, just to rehearse some of the data on whether SV40 is an infectious agent for humans, there are a couple of studies I would like to call your attention to, and I think most of these are referenced in the bibliography that was handed out yesterday morning.

Doctor Morris, who is here on our panel, instilled SV40 by the nasopharyngeal route in 1960, because a vaccine that had been constructed at that time devoted to the respiratory-cincitial virus was known to contain SV40. So, a series of volunteers was inoculated by the nasopharyngeal route, and they received a very high dose of SV40, about 10⁴ tissue culture infectious doses. And, indeed, about 70 percent of those volunteers developed neutralizing antibody at a moderate titer, and, in fact, about 30 percent of those volunteers had SV40 recoverable from their throats about one or two weeks after the nasopharyngeal installation. So, we certainly know from this study that SV40 could infect humans.

Secondly, as you heard abundantly yesterday, the oral live poliovirus vaccines in the early days contained high titer SV40, and yet, they induced no, or at least not detectable by the assays of that day, no neutralizing serum antibody, but many of these people did have positive cultures, up to about 30 percent of the recipients had positive stool cultures, for up to five weeks after they swallowed the poliovirus vaccine. So, it appears that by the oral route this virus, SV40, can colonize the human gut, but it doesn't cause a systemic infection.

However, the injected or killed poliovirus vaccine, which still contained low titer SV40, since, as Doctor Hilleman pointed out, SV40 is more resistant to formalin than is the poliovirus vaccine, probably because of its double-stranded covalent close configuration, injected killed poliovirus vaccine induced high to moderate antibody titers against neutralizing antibodies against SV40 in about 20 to 25 percent of those people who were knowingly inoculated, and those antibodies were still present after three to ten years, depending upon the studies, and the same data were true of the adenovirus vaccines.

Next slide, please.

So, that brings us to the question of whether SV40 is endemic in human populations. Well, several studies shown on this slide have demonstrated that close contact with rhesus monkeys or people working with monkey tissues in the laboratory have neutralizing antibodies and variable titer to SV40 in as many as 50 to 60 percent of those who have such contact. And, those studies were not only from this country, but the Soviet Union.

Incidentally, one thing that was not mentioned yesterday is that there was a massive exposure to SV40 in the Soviet Union, in the early poliovirus vaccine era. To the best of my knowledge, those people were never followed up, or at least not followed up for a long period of time, but I'm told that access to their records still exist somewhere in whatever that part of the Soviet is currently called.

Another study done by Brown, et. al., in 1975, looked at isolated populations. These included the Papua New Guineans and the Yorubas, Alaskan Eskimos, Brazilian Indians and so forth, who had no exposure to any vaccine and no exposure to monkeys, and Brown, et. al., found that in people in these populations who were seronegative for the BK virus, they, nonetheless, had low titer neutralizing antibody to SV40 in about five percent of those people, but in those who were BK positive in about 35 percent, suggesting cross reactivity, at least in that assay, which was the plaque reduction assay, between BK and SV40, but it was in the direction of BK spilling over onto SV40, rather than the other way around.

Individuals who had been bled before 1954, in other words before the poliovirus vaccine or who were born after 1963, in other words after the SV40 contaminated polio vaccine had been taken from the market, in those two populations neutralizing antibody to SV40 existed in about four percent, four to five percent, mostly in low titer.

Individuals receiving the polio vaccine during the period of SV40 contamination, 1955 to 1962, had antibody to SV40, serum antibody to SV40, in about 20 percent, perhaps, 25 percent, and Doctor Shah will say more about that. These are his data. So, there's no question that the incidence of infection with SV40 went up as a consequence of the contaminate poliovirus vaccine, but then it went down again, apparently, to what it had been before the poliovirus vaccines.

Next slide, please.

Now, finally, various groups have been tested for neutralizing antibody to SV40 capsid antigens by the ELISA assay, the enzyme linked immunosorbant assay, which is quite sensitive and reasonably specific, and here there are some interesting data. One paper by Zimmerman and Geissler from 1983, and another by Geissler, et. al., in 1985, these are fairly obscure papers, but important.

These workers, first of all, looked at 51 medical students at the University of Wisconsin in Madison who had been bled in November of 1952, considerably before the poliovirus vaccine era, and they found that 12 percent of this population, in fact, were positive for neutralizing antibodies to SV40.

Then, they looked at an entirely different group of people in Germany, who were born between 1959 and 1961, and found that 24 percent of them had positive sera. This is consistent with Doctor Shah's data, and those born after 1962, when contaminated vaccine was no longer on the market, went back down to about 13 percent. So, you can see that the rates of infection were the same before the contaminated vaccine and after the contaminated vaccine, these are higher than Doctor Shah's rates, but it is a more sensitive assay.

And, finally, as with the plaque reduction assay, workers in the laboratory, with high-intensity SV40 exposure, had about a 55 percent carriage rate of SV40 neutralizing antibodies, and, again, that was consistent with the other data that I showed you by the plaque reduction assay.

Cancer patients were looked at also, they had quite a variable incidence of neutralizing antibodies, between ten and 30 percent.

Now, there's additional data about SV40 isolated from human tissues before the polio vaccine era. Two patients with progressive multi focal leukoencephalopathy, a disease we heard about yesterday known to be associated with the JC virus, and born in 1915 and 1933, had SV40 isolated from their brains by techniques that are still reliable. And, one patient wit metastatic melanoma born in 1894, so there was no question of exposure to the vaccine, had SV40 isolated both from tumor tissue and from pleural exudate. This is a very good study by Soriano, et. al., reported in Nature in 1974, and this was the virus itself, in addition to expressivity of T-antigen and neutralizing antibodies.

May I have the next slide, please?

That brings us to the next question by way of background, which is the following, is SV40 neutralizing activity in human sera explained purely by cross reactivity to the capsid antigens of BK and JC. Well, we know that JC, BK and SV40 all share capsid antigens, but in every assay that I was able to examine homologous reactivity was at least 100 fold greater than heterologous activity, and that was true by plaque reduction, fluorescent antibodies or immunoelectron microscopy assays, and these assays are known to be less specific than hemagglutination inhibition, CF immunodiffusion or immunoelectrophoresis, and there are abundant references in support of this notion that there is cross reactivity, but the homologous reactivity is very much greater than the heterologous activity.

I would call your attention to one paper in particular, that of Penney, et. al., who did an immunoelectron microscopy and was able to directly visualize the interaction between the antibody and the capsid.

Now, as I mentioned, Brown, et. al., in 1975, studied isolated populations, and found that 35 percent of those who had BK positive sera were also positive for SV40 in low titer, but five percent of the BK negative sera was also SV40 positive in moderate titer by plaque reduction, again suggesting some cross reactivity, mostly in the direction of BK and possibly JC spilling over onto SV40, but not the other way around.

And, I also mentioned the study of Brown and Morris, in which they instilled the respiratory cincitial virus vaccine known to be contaminated with SV40 by the nasopharyngeal route, and then they went back and looked at that same sera from that same group a couple of years later and found antibody to the BK virus by hemagglutination inhibition, which is a sensitive assay.

And, I would simply call your attention to the fact that the T-antigens of all these viruses are more highly related than are the capsid antigens.

So, I think with that background at hand, we are now poised to hear at the beginning of this panel some formal comments from one or two people on the panel who have asked to speak, starting with Doctor Shah.

After that, we'll move on to the questions that we have formulated for the panel to address.

DOCTOR SHAH: I tried to discuss this question about the SV40 neutralizing antibodies yesterday in the review portion of my talk, but I did not have enough time to talk about it. So, could I have the first slide, please?

This is a picture that I showed yesterday. It's in a temple where the monkeys live in close ecological contact with the human populations.

In 1965-66, one of my major interests was to see if SV40 was transmitted to humans in this situation, and the stimulus for that was the SV40 exposure in the United States. And, as we discussed yesterday, in 1966-67 the follow-up period after the exposure of the human population was only three or four years, and it may take a long time for a virus-induced cancer to occur, so the rationale was that if SV40 is being transmitted to humans in India, and that is probably occurring for centuries, so if you studied the cancers in India you may get some sense of whether or not SV40 is involved in any human cancer.

May I have the next slide, please?

We did this study, and this was a summary of the antibody data from quite a large number of studies. This first line are the rhesus that were infected intra nasally, subcutaneously or orally, and after four to six weeks and 29 to 31 weeks these were the antibody titers. These tests were done, not by plaque neutralization, but by neutralization in test tubes. So, you always got these relatively high titers after experimental infection.

These are naturally infected rhesus monkeys which were bled in India, and, again, the titers are in the same range.

These are the individuals who received the SV40 subcutaneously. They received sometimes live SV40, but at the same time they received a great deal of inactivated SV40, so they were exposed to the antigen, inactivated antigen, in addition to the live virus, and they may have been exposed to the antigen alone a number of times without having the live virus.

And, these are, again, they are not too far away from what are found in the monkeys, although the titers were somewhat lower, but in all instances they are sort of in the high range.

And, these are the same people after three years, these are not my data, these are taken from literature, I think these are probably from Doctor Gerber's paper.

And, these are Doctor Morris' data, and he's here, internasal inoculation of SV40, and, again, the titers tend to get lower as you see here.

These are the oral polio vaccine, other people's data, no antibodies, and this is what we found in India, and only about five percent had low levels of antibodies, but look, these titers are extremely low, and quite different from what you see here.

We thought for some time, can I have the next slide, please, we thought for some time that the infection in man may not be very efficient, so they are having a low level replication and low level of antibodies, at that time we were studying the animals, the serum specimens from North India, and this is where the rhesus monkey is distributed and where there is contact between the rhesus monkey and man, and we thought for some time that these are probably SV40 antibodies.

We subsequently worked in South India, and where there is another macaque species, the bonnet macaque, or the macaque irradiata, which is free of SV40. There are many macaque species, of the 19 macaque species there are several which are not infected in nature. So, when we studied the human sera here, their pattern of antibodies was identical to what we had seen in North India.

Can I have the next slide, please?

So, what we concluded then, that the low levels of antibodies were detected on a small proportion of human sera, we could not ascribe them to SV40 infection because prevalence in North India was similar to that in South India, and subsequent studies were, I mean especially the study by Brown, Sih and Babacek, which Doctor Levine referred to in the isolated populations, and that study suggested that the antibodies might be the cross reaction might -- the low level of SV40 antibodies may be due to cross reactivity with BKV antibodies.

Now, if I remember correctly, they did not have at that time access to JCV antigens, so there was a portion they could not explain, I think Doctor Levine said that about five percent of the BKV negative sera also had antibodies, but then the JCV was not looked for. So, it may be that this might be cross reactivity with the non-viruses.

Actually, we had proposed several years before JCV and BKV were identified that there are, perhaps, cross-reacting human viruses which are responsible for these antibodies. So, most of the data, most of the serological data that we had, could be, perhaps, explained as a result of cross reactivity, but there were some isolated instances which we could not explain, one which Doctor Levine referred to. One of the two cases of PML that was reported from Hopkins as due to -- well, they identified SV40 in the brain had extremely high titers of antibodies, high just like the sera from experimentally infected monkeys, and in serum specimens that Doctor Roh gave us many years ago, which were his collections for some other reason, we found something like four or five human sera, and this is all documented so I won't go into detail about that, of people who were bled, for example, in 1952, and in some instances there was more than one serum from the same person so we could check both samples, there were titers that seemed to be too high to be this cross reactive antibodies, but we never really completely solved that question.

I would like to show on our transparency if I may -- I think this is the last slide, yes, last night we were discussing in a group about what this virus might be which might be circulating in the human communities, I don't, myself, conceive that, but supposing it is there, what would be its property, what would be its characteristics? And, Doctor Butel and I came up with these characteristics and, perhaps, most people would agree with them or might, perhaps, challenge some of it, one, that the virus that has been found in the mesotheliomas, the osteosarcomas, and in ependymomas is SV40 itself. It is not BKV or JCV, and it may be -- it is, for all intents and purposes, the monkey virus, the SV40, so we are really look for SV40, and it is not a question of BKV or JCV being misclassified as SV40.

The second question was that originally was introduced, now it is circulating independently of the polio vaccine contamination, because all the ependymoma patients, most of the osteosarcoma patients, and, perhaps, some of the mesothelioma patients, were not exposed to the polio which was contaminated.

So, they must have occurred, the virus seems to be circulating independently of the initial exposure, initial non-exposure, and you think of two possibilities. One, that is has always been there, even before the polio episode, and the polio vaccine may have simply increased the amount, or that the polio vaccine introduced it into the human population and then it is now circulating independently.

As I think I said before, this is not easy to conceive, because these viruses are very highly species specific, they don't really cross boundaries, their genetic make-up is not like that of the RNA viruses. But anyway, the data suggests that it is independent of polio vaccine contamination, it is circulating.

And, the third characteristic we thought one might ascribe to is that it is capable of being transmitted very early in life, and this is the data from the ependymoma cases, where the children were two, three, four years old when they had cancers, so that the infection must have occurred at birth or very soon after birth, perhaps, in utero even. And so, this then suggests that it may be either transplacental transmission or in some other way.

Now, transplacental transmission would only occur, I think, if the mother was having a primary SV40 infection with a large amount of virus, a large amount of virus in the blood viremia. So, one would think that the mothers would have very high antibody response if they were capable of transmitting the virus to the fetus, and yesterday there was a suggestion from the floor as to what happened to the parents, did the mother show evidence of infection?

And, one of the ways one could follow in order to clarify all these issues is to study the individual cases of these tumors from an epidemiologic point of view, in the families, in the surroundings, and see if you can find any evidence for this naturally separating SV40, and preferably isolate the virus itself, relying not so much on presence of antibodies. And, it was an attempt to do that sort of thing that we had looked at these 165 urines from transplantions, not transplantations, but HIV infected patients, to look for this independently circulating virus, but the contacts of the patients who have these tumors, I think they would be a very rich source for investigating this. And, these are very rare tumors, but I think you can get to them if the resources are mobilized.

Before I came here, I spoke to Doctor Grossman at Hopkins, who is the head of a tumor bank, a brain tumor bank, and a consortium of ten institutions where they look at brain cancers, and there are three such consortia in the United States. Now, he had, in his freezer, six ependymoma cases, one choroid plexus papilloma, he said these are rare tumors but you will find them within the ten institutions, there will be one, or two, or three patients currently undergoing treatment. If you want to reach these people, look at the families, look at the patients for evidence of these viruses, I think it would be a big thing. And, I think such a study can only be done by CDC or NIH, it would be beyond the capacity of individual workers to do such a study.

Thank you.

DOCTOR LEVINE: Are there any other members of the panel who would like to make a statement before we continue?

Okay, then if not, may have the next slide, please, or my next slide?

While we are waiting for the slide to go on, I would like to invite the members of the audience to take part in this discussion. I think it should be uninhibited and informal, and I thought yesterday we heard some superb comments and data from the floor, and my expectation is that this morning will be equally robust.

Okay. So, here we pose some questions, and I would like the panel to begin to address them. We have talked about all of these questions already to some extent, but now we will focus more, and, again, I'd invite the audience to address these questions as well.

What are the current data that suggest that SV40 is an endemic agent in the human population? How can SV40, which supposedly replicates poorly in human cells, spread as an infectious agent in humans? What additional data are needed to determine whether SV40 is endemic or not? And, what are plausible sources of human exposure to SV40, for example, vaccines, primates, humans themselves, or unrecognized recombinants.

So, I need a member of the panel to lead off this discussion. Doctor Butel?

DOCTOR BUTEL: I'll just comment that we have been doing some sera epidemiology assays using a plaque reduction test, looking for the presence of SV40 neutralizing antibody in various groups of people.

In general, the numbers that we're finding agree with Doctor Shah's published data. If the group of people are of an age where they probably received a contaminated polio vaccine, we are finding neutralizing antibodies in, roughly, 20 percent of those people.

In older people, that most probably did not get a vaccine, we are also finding antibody more in the neighborhood of maybe ten percent of those people.

In younger people born after 1963, we are finding neutralizing antibody anywhere from two percent to ten percent, depending on the group that we are looking at.

The titers are not too high, more like the natural infection data that Keerti just showed. They range from a neutralizing titer of one to 20, we've had some that are one to 320, and I guess I'm not persuaded that all of this can be explained by cross reactivity with BK or JC antibody.

We've tried to grapple with this, and I would like to hear suggestions as to how to absolutely prove this, but we looked at the BK antibody titer in a group of people that were positive for SV40 neutralizing antibody, and compared that with a group that were negative for SV40 neutralizing antibody, to see if the one group had very high BK titers and the other group were low, and that is not what we found. We found a range of BK titers in both groups, and so the SV40 neutralizing antibody did not correspond to having a high BK antibody titer.

So, we would like some suggestions as to how to maybe absolutely prove that this is SV40 neutralizing antibody that we are finding.

DOCTOR LEVINE: Doctor Frisque?

DOCTOR FRISQUE: I don't have an answer for you, but I think I have a related -- closely related question, which is I think the data which were nicely reviewed, years to corroborate this by Doctor Shah and Doctor Levine, certainly provide tantalizing evidence that SV40 exists in the human population, perhaps, was enhanced by the poliovirus experience.

My problem is, is that the data that have been summarized and published certainly don't need meet my standards for current rigorous sero epidemiology. As many of you know, I come particularly from the HIV perspective, and the two-tailed issues of sensitivity and specificity just can't be addressed, I don't think, in the algorithm of applying a single assay, particularly, with a virus that has a relatively complex lifestyle, that involves both replication of structural antigens and then subsequent latent antigens that may or may not be picked up by a single assay.

And, I think to move this field ahead, it seems to me as though, you know, a high sensitivity assay, followed by corroboration with a high specificity approach, is really what's needed, and that's certainly the state of the art in HIV and Hepatitis C virus and the specifics depend on the virus and, you know, how it manifests in the population.

But, for example, if you take the Geissler paper, which I think is among the most tantalizing of them, it's the one that applied an ELISA technique of a high sensitivity, I don't see any corroboration that that proves that that's, in fact, SV40, and ELISA's, of course, are, you know, renowned to be prone to various kinds of cross reactivities that are uncharacterized.

So, I would like suggestions as to how, you know, what would be needed to sort of move this ahead in the two-tailed front of a screening assay and then a corroboration assay.

DOCTOR BUTEL: Well, let me respond to something you said. HIV serology is very different, I think, from what SV40 serology would be, because HIV has many more antigens, and in the case of SV40 we know there's the neutralizing antibody is directed against VP1, and there aren't other antigens that I think would be important in neutralization.

And, I would like someone to explain what assay is better than a neutralization assay for identifying a specific antibody, where you have a plaque assay for the virus. I think it's more specific than CF or HI, or even an ELISA.

DOCTOR LEVINE: We have some comments from the audience. Doctor Pass?

DOCTOR PASS: If I can show a couple of slides that I just gave to the gentleman in the back. With the help of Paula Rizzo, when she was in my lab, because we were doing some hamster assays, we also set up an ELISA, not only to measure antibodies, serum antibodies to SV40 in the hamsters, but also hamsters, as an indirect ELISA. And, essentially, the question is, what do you compare your numbers to?

So, we, essentially, have two baselines that we compared to in these patients. I felt that it was reasonable, this was before the blood and urine paper, that it would be reasonable to compare titers to cord sera, so we got cord sera and used that as a baseline, but also from patients who come to the NIH we took their blood and got sera and called the baseline the mean of those patient sera plus three standard errors. And, if the levels were then above that baseline, we would call it positive. And, the results are summarized here, I have the graphic results in the next slide, but this slide shows that if you compared a cord sera, both mesothelioma patients and normal volunteers, will have levels above cord sera, and that's not significant. But, if you then look at the levels above the baseline, which is comparing to general population I guess, you find a statistically significant increase in levels of antibodies to T-antigen, at least, SV40 T-antigen, in the sera in the mesothelioma patients.

Next slide.

DOCTOR LEVINE: Could I just ask a question? Did you use JC and BK antigens as controls?

DOCTOR PASS: No. No. Again, the purpose of this was to see whether the mesothelioma population as a whole had some difference from a normal population with regard to neutralizing antibodies of T-antigen, and the data is graphically depicted here with the normal volunteers at the bottom and at the top.

Very frankly, when we started to do this, I didn't know how much -- I didn't think there was much in the literature that really guided us, so we sort of just started from scratch and did it because we had the sera.

DOCTOR LEVINE: Thank you.

DOCTOR GOEDERT: Could I make a comment? I think as a first cut, I think that, you know, sort of diversifying the approach to detecting antibodies in this particular problem is probably an extremely valuable undertaking. You clearly need some kind of a confirmation or sorting out of the different virus reactivities that could be SV40 and could be other epitopes.

DOCTOR URNOVITZ: May I add to that comment? I agree with the last speaker that -- I'm Howard Urnovitz with Chronic Illness Research Foundation, I was a manufacturer of one of the HIV test kits, and I think that as we went forward and looked into other bifluids in blood, we found a completely different distribution pattern of antibodies, and what we found out is the ELISA is a heck of a great screening test. The confirmation is even better to tell us which is the difference between false positives.

We didn't go out of our way, because then we started to recognize our false positives were antibodies to human endogenous retro viruses, which was not HIV-1. And, the bottom line is, is the further we look the more we realize that all we could really get out of an antibody test is what most of the intended statements say on the manufacturer's test, is that it suggests an exposure to the virus. I mean, that's about as far as you can go with the data. So, I would respectfully submit that we may want to think that the antibody tests have given us great direction in which way to go to research, but I think I'm just cautioning the group that you may be over-interpreting the antibody data, and I think you need to move to more of a molecular biology approach. I just don't see how you can go any further with the antibody assays and conclude that it was an endemic.

DOCTOR LEVINE: Anybody else from the panel? Doctor Dorries?

DOCTOR DORRIES: We did ELISAs and HI tests on BK virus and JC virus, and I only can say that our homologous activity was clearly much higher than the heterologous activity. In HI tests, it was much better than in ELISAs, and I think then that indirect ELISAs might be even better.

DOCTOR LEVINE: From the floor?

DOCTOR OZER: Harvey Ozer from New Jersey, a standard way of trying to discriminate cross reactions is to do competition assays. And, in fact, if there is antibody to SV40 and you are suspicious of BK, why not try to block the antibody to SV40 with BK virus.

In fact, Keerti and I, many, many, many years ago, used an assay to verify the neutralization, which was, essentially, an immunoprecipitation of purified virus of SV40, which was radio labeled, and that was susceptible to competition assays.

So, I think there are assays out there that people can do. They are not convenient for screening, and so I think the issue is, if you identify antibodies that you think are interesting, from patients that you think are interesting, to follow them up with research protocols.

DOCTOR LEVINE: I'd just point out that Geissler actually did do a blocking test, but he probably did it with the wrong antigen, he used polyoma and a lambda protein, but not BK or JC.

DOCTOR OZER: And, I would just conclude with the comment that we -- since it neutralizes SV40 virus, the one that Janet was talking about, it must be on the BK virus virion, and, therefore, so it's very clear what the competitor should be.

DOCTOR LEVINE: Other comments from the panel? Doctor Oxman?

DOCTOR OXMAN: I was going to say the same thing, that really if you cross absorb the five percent of the sera that are neutralizing in the plaque assay with both native and disrupted BK and JC, you'll either find you are absorbing those cross reacting antibodies or you've got some residual that's SV40 specific.

DOCTOR LEVINE: From the back?

AUDIENCE: I'd like to make some comments about some of the things Keerti Shah said. This business of cross reactivity is really an interesting problem, and I think we shouldn't forget other papova viruses, like LPV and that sort of thing, so maybe this is a time to rethink some of our strategies and think again about viruses like LPV and how they fit into this whole equation.

The second thing, with respect to virus neutralization studies, I'd like to remind everybody about the Ashkenazi-Melnick experiment, I think it was '62, but they infected monkeys with SV40 and found out, number one, even with infected monkeys, using the techniques they had back then, it was hard to show and recover virus from some of the infected monkeys, and the antibody levels they saw in these monkeys in some cases didn't go over a one to ten dilution based on virus neutralization tests.

So, the point is, depending on who is doing the experiments, you know, it's not really that different from what you see in monkeys and humans, in some cases. Also, if you use different species of monkeys, you might get somewhat different results.

How do you compare, though, the results you get with a certain monkey and what you might see in humans? That's a very important question.

DOCTOR LEVINE: Thank you.

Anybody from the panel wish to comment? Okay. Yes?

DOCTOR MINOR: Can I go back to something that Doctor Butel said about her sero studies?

DOCTOR LEVINE: Sure.

DOCTOR MINOR: If I understood it right, my name is Philip Minor, I'm from NIBSC in the United Kingdom, as I understood it, the people who are of an age to have been exposed to the -- vaccine were 20 percent seropositive, or thereabout. Were those highest titers or lowish titers? I mean, my difficulty is that they were exposed at least 30 years ago probably, and this is an awful long time for an antibody to persist, if it's, you know, just straightforward antibodies you saw. Can you say a bit more about the titers and how long it was since the exposure and so on?

DOCTOR BUTEL: The titers, in general, were low, one to 20, one to 40. The exposures, presumably, were at the time when the contaminated polio vaccine was given, although, there's always the possibility that if SV40 is circulating in the human population they might have been exposed again.

Many of these sera were collected, oh, ten, 15 years ago, they've been in storage.

DOCTOR LEVINE: I think in Joe Fraumeni's study he looked at antibodies shortly after the contaminated vaccine and followed that group initially for four years, and then I think there was a subsequent follow-up at ten years. And, if I remember his data correctly, he found high titers initially, and they fell to what he called a moderate level thereafter.

But, Doctor Morris, you were of that era, can you comment?

DOCTOR MORRIS: Yes. The titers that occurred in about one third of the volunteers were relatively low, one to ten, one to 20.

DOCTOR LEVINE: Thank you.

Doctor Shah?

DOCTOR SHAH: Yes. I think in the antibodies that occurred as a result of SV40 contaminated polio vaccines were a good bit higher than what we saw in the human population, and I believe they were followed for at least 13 years, without any marked reduction of titer.

And, there was a controversy at the time that the antibodies are persisting, so there must be live virus there, on the other hand, whether the mammary cells can persist. I think that controversy was not resolved, but the antibody titers did not decrease, at least not markedly, for at least 13 years.

DOCTOR LEVINE: But, we should go back and look at Joe's data, because he did, I think, have a careful record, but I couldn't find it in the publications of what the titers actually were on given patients.

DOCTOR SHAH: Even to children Doctor Fraumeni followed?

DOCTOR LEVINE: At least out to the first three or four years, but, Jim, maybe do you know anything more about it?

DOCTOR SHAH: I think they were children who got oral polio vaccine, I think, so they were probably antibody free.

DOCTOR LEVINE: Well, he didn't follow the ones that were antibody free, because they didn't have antibodies, so it must have been the group that had the Salk vaccine.

Other comments or questions? Doctor Strickler?

DOCTOR STRICKLER: Yes, thank you. I think the big problem that we have in trying to validate any sero assays is that we don't necessarily have good exposure data on any particular individuals. We don't know exactly who is infected. We, in our one, tried to see if the virus could be detected in urines, as a manner of figuring out who is infected and non-infected, we were at least so far unsuccessful.

Doctor Butel took the first step in looking at comparison groups, in which she had an understanding of the probability of exposure by looking at different birth cohorts, but that's exactly the type of thing that we need to do as we move forward with these assays. We have to be very clear about what our exposure groups are as we validate this, lab workers who were exposed to monkeys likely to be contaminated is one group, the birth cohorts that we've defined, and, obviously, the patients in whom we think that we are detecting SV40 DNA, are individuals we would like to test, but we need to compare them also to appropriate groups.

For example, it would be interesting to know if Doctor Butel was able to detect antibodies with her assay in any patients with the tumors that have been found to contain SV40, but I think it's also important to look in other cancer patients, because as people become immune compromised there can be other reasons for antibodies to different viruses to increase. So, I think we need to be very careful about our comparison groups, and we need to look towards individuals who we think we understand what their exposures were, as we try to validate these assays, and not just make it based on competition assays and so on. We need to try to validate them based on the amount of exposure data that we have.

DOCTOR LEVINE: Thank you.

DOCTOR DORRIES: I would like to comment to the ELISA titer -- to titers in the natural monkey infection. We recently did some ELISAs on persistently infected, naturally infected monkeys, and the ELISAs were fluorescent based, and we got titers in the range of one to 5,000, and one to 10,000.

So, in reviewing the new methods, one might get another possibility to check out.

DOCTOR GOEDERT: T-antigen or V antigen?

DOCTOR DORRIES: V-antigen, we used purified SV40 particles.

DOCTOR LEVINE: Doctor Carbone?

DOCTOR CARBONE: I would like to ask the panel, the members of the panel a question, that I'm not sure I understand very well, and it's the following. The oral polio vaccine were used, one of the main reasons that they were used was because they are excreted, and so you vaccinated one kid and then you vaccinated an entire family.

Now, if I understand correctly, SV40 contaminated some of those vaccines, and SV40 was also excreted. So, I would assume, and I'm not sure I'm correct here, but would assume that if the other members of the family got infected and vaccinated with the polio, they also maybe got SV40.

Now, I hear that there is a lot of concern whether you are checking the antibodies or the tumor induction in kids or in adults, but this may be, in fact, not completely correct, because if the kids were mostly the ones who received the vaccines, but then if the thing works the way that the polio worked, then everybody got it.

Am I wrong?

DOCTOR LEVINE: Doctor Shah, do you want to comment?

DOCTOR SHAH: I did not exactly get the question.

DOCTOR LEVINE: I think Doctor Carbone wants -- is raising the issue of whether SD40 was spread by the oral fecal route from infants who were immunized to their family members.

DOCTOR SHAH: Right. In the U.S., only a few thousand infants received oral polio vaccine that was potentially contaminated, because the licensed oral polio vaccines were free of SV40.

I think it is absolutely right that if there's a polio excretion in a family, every member will become infected with polio virus. I think this was done by Doctor Melnick and many other people long, long ago, and you could imagine that the similar thing could occur with the SV40, which would be in the vaccine.

The studies that were done on excretion of SV40, in people who received contaminated oral polio vaccine, and they were able to do it because the stools were saved for polio studies and they went back and looked at the same stools, there were about three or four studies, maybe one or two found SV40 excretion, which was intermittent, low level for about five or six weeks, but some other studies were negative, they did not find SV40 persisting in the stools. Whether it could have infected family members, of course, I don't know.

DOCTOR MORRIS: I'd like to make a comment about the individuals. You said there were very few individuals for which the exposure is known. Well, I think there were 31 volunteers participated in the studies that we carried out in the early 1960s. Those sera were stored, and they were used most recently in 1966. So, for those persons, and I still -- I believe that those sera are still stored here at NIH, if you want an individual with a known exposure, with virus recovery from throats, with antibody rises, you have subjects that are available, that is, the materials recovered from these subjects are available, and they should be stored still at NIH for use in further studies.

DOCTOR LEVINE: I should point out that we have another bank of sera that may be relevant here. In the late 1950s, there was a national study, at that time sponsored by the Neurology Institute, of cerebral palsy, and to carry out that study sera were collected from 45,000 pregnant women, and those sera still exist. They date from about 1958 through about 1965, and the histories were well documented, so we know which mothers received the vaccine and which did not. The cord sera and infant sera from 20,000 progeny of those mothers exist in the bank as well, and so that's going to be a valuable resource, once we decide what assays might be applied, so that we don't squander the sera, which brings up the question, actually goes back to what we might --

DOCTOR BUTEL: Could I say something before you change the subject?

DOCTOR LEVINE: Sure.

DOCTOR BUTEL: To sort of address Michele's question, some of the sera, but it doesn't exactly answer the question, some of the sera that we've had access to were from daughters and mothers. The daughters were of the age to have been exposed to the contaminated vaccine, the mothers were older and would not have been.

And, we looked at some of the matched sets of the mothers' and daughters' sera to see if the daughter was positive for SV40 neutralizing antibody, was there a better chance then that the mother was going to be positive, most probably because the virus had spread in the family from the vaccine.

The results were that there didn't seem to be any correlation between whether the daughter was positive and the mothers were positive.

DOCTOR LEVINE: Yes?

AUDIENCE: I'm very encouraged this morning to hear that the government is interested in looking at the parents of the children who have SV40 associated --

DOCTOR LEVINE: Go ahead, I think the microphone is on. You were on.

AUDIENCE: -- to see if they also are carrying SV40.

I was exposed to the potentially contaminated vaccines in both Minnesota and Colorado, when I was a child. I think that parents or mothers in my generation would be very interested to have you do a problem study and find out how many of us are carrying it, and also to find out how many of our children are carrying it.

I think it's very important for the independent researchers who came here and who have done these studies, that have pointed out this problem, be funded and be part of anything official that is done to find out if there are SV40 associated cancers that are causing health problems.

I think that the government is in charge of promoting vaccination, and it took many, many years for you to admit that the oral polio vaccine can cause polio in some children and in close contacts, and the public would not have confidence, frankly, if these studies that you are talking about are only conducted by the government and do not include independent researchers in some kind of oversight.

DOCTOR LEVINE: Thank you for your comment, but I would like to point out that I don't -- it's not my sense that the government has been conspiratorial or guilty of obscurantism in dealing with the issue of whether SV40 -- let me just -- hear me out for a second -- whether SV40 is in these vaccines, because as Doctor Helleman pointed out yesterday, the discovery of a new -- what was then a new virus, and elucidation of its biology, and ridding that the uncurrent vaccines of SV40 was all accomplished with a two-year period, which is quite remarkable.

Secondly, I would like to point out that the data that we heard at this meeting has been entirely generated, or at least primarily generated, by non-government scientists, and, in fact, there's no reason to believe that that won't be sustained.

But, finally, I do need to remind you of the data that Doctors Olin and Strickler showed yesterday, which is very powerful epidemiologic data suggesting that no harm came from the vaccines that were knowingly contaminated with SV40 with respect to cancer.

Could I have the next speaker, please? Doctor Lednicky?

AUDIENCE: I'd like to reply to what you said.

DOCTOR LEVINE: I will let you in just a moment.

DOCTOR LEDNICKY: Perhaps, the panel might address this. Now, as a virologist, I have a question about doing antibody tests when we suspect an association with tumors, because if it's not a Lytic disease can we really expect that you would have high titers of neutralizing antibody against virus?

So, perhaps, as was talked about earlier, using ELISA tests directed against T-antigen might be the better way to go, because I'm not sure you can really learn anything about screening perspective -- patients with perspective SV40 tumors for antibody.

The second thing is, doing neutralization tests, a lot of labs can't talk to each other because the antibody levels that are detected aren't always exactly the same, and that goes back to basic virology. One reason for this are the reagents. If you have a lot of defective interfering particles in your virus prep, for example, this can affect your result. So, I would like to suggest that we standardize that test. Doctor Butel has people doing literally hundreds of these tests a week, and it might be a good idea to standardize some of these tests by having virus come from one source.

Thank you.

DOCTOR LEVINE: Thank you.

Would the panel like to respond?

DOCTOR GOEDERT: This is, I guess, a related question. I was wondering, I was going to ask if others beside Doctor Pass and our group have tested sera from mesothelioma patients or others whom we believe would have a high likelihood of having virus detected, at least by PCR. And, you know, we did not detect neutralizing antibodies, and I think the reason may, in fact, be that they don't have anything but latent antigens expressed if they are infected, and I think there's a reasonable chance that they are, and so I'm just wondering if others have experience with populations of people who have some validation that they actually are infected with the virus.

DOCTOR LEVINE: Does anybody have data?

Let me finish with the panel first, Doctor Frisque?

DOCTOR FRISQUE: Okay. I just wanted to make a point about the reagents and discussion of reagents. One factor we haven't really considered, I believe, is the source of the antigen used in these assays, and, clearly, there are antigenic variants in these viruses. With BK virus, for example, BK wild type versus the second strain called BKVAS, which is 95 percent sequence identity, but those viruses if you make antibody to BK wild type you will not see a cross reaction with BKVAS. On the other hand, if you make antibodies against BKVAS, you will see a good cross reactivity in the other direction with BK wild type.

So, I think there is consideration in the antigen sources that we use in these tests, that that's an important factor as well.

DOCTOR LEVINE: Thank you.

Now, let's go back to the panel. Yes?

DOCTOR KYLE: Yes, sir, Walter Kyle, I'd like to comment briefly on Doctor Carbone's question. There are a couple of things.

First, I think everybody should realize, as Doctor Ratner attempted to point out yesterday but was interrupted, the inactivated polio vaccines of the '50s weren't actually inactivated. They contained live polio viruses, they caused many cases of polio.

In addition, very significantly, when you mix formalin with protein you get plastic. You had plastic encapsulated polio viruses that presented themselves after 30 days of inoculation, and I assume the SV40 was live encapsulated also in a lot of those early injections. I spoke about this at the National Academy of Sciences in 1992.

I've had the opportunity to gather and review a lot of the manufacturing records from that time, of the manufacturers that were making these vaccines.

The other comment I have is to Doctor Shah, I keep hearing this reference to after the vaccine was licensed in 1963, there was no SV40 in it. You should all know, I know, I have the records, most, if not all of the licensing lots of the oral Sabin vaccine were SV40 contaminated.

Now, if they removed it afterwards, I would hope so, the fact is that the government agencies in charge for the safety of the vaccine have been found negligent in its licensing and release. It's the only vaccine or product where the government has been sued successfully, and that's -- that case was litigated for over 20 years, I was a part of it for a while.

The documents we've gathered from it may be very helpful to some of you here, and I'd be happy to share them with the independent scientists.

Thank you.

DOCTOR LEVINE: Well, I don't -- if I may take the Moderator's prerogative -- I don't think that there's any question that there was live SV40 in the polio vaccines. That's not been an issue.

Yesterday, we heard representatives from the manufacturers describe their technology in detail, assuring us that to the best of their ability, and with all current technologies, they can't detect SV40 now, but I don't know whether any of those people are still here in the audience. Is anybody here from the -- yes, would you like to comment? May we have a response to your question?

AUDIENCE: I understand what's going on now wasn't what happened in the past, and I know -- I have the records of what Doctor Kirschstein has seen in the polio vaccine over the years, and her comments and her meetings with the manufacturers.

For the manufacturer to get up here and assert that there's no viable microbial agent in their vaccines belies the evidence that the NIH and the Bureau of Biologics has as to what was in there over the years, into the '70s, into the '80s, after the '80s they started cleaning the monkeys up because a big problem arose, it was AIDS.

But, the retro viruses, they actually discovered in the vaccines, the oral vaccines, in the mid-'70s, permitted their release, and didn't recall them from the market in the '80s when they knew.

DOCTOR LEVINE: Right. I think we are getting into a territory which is remote from what we want to discuss, but if we can have a brief response to that comment I would appreciate it.

AUDIENCE: I mean, I think leaving aside the retro virus question, I'm not a manufacturer. I'm the CBER equivalent, if you like, in the United Kingdom.

In 1962, there were international requirements produced, which were the consensus of national requirements, okay, which say that SV40 should be excluded.

In 1965, this was made a statutory requirement, if you like, for WHO things. The requirements were updated around 1972, I think, and this remained in there, in 1989 it's still in there. So, from 1965 international requirements, which are the consensus of national requirements, require that SV40 shall be absent.

At NIBSC, which is an independent testing laboratory, we have records from at least 1966 to say that we were testing every batch of vaccine that came in for SV40 and finding them negative, so we were actually checking this out.

At least as far as we are concerned, from the documented point of view, from 1966 the vaccines were free.

The data that Dave Sangar presented yesterday on the PCR aspects of the batches that we are looking at at the moment start around 1996, because these are the recent batches that we can do something about, we've got back as far as 1980, and they were all negative. The proposal, really, is to go back even further and show that they are negative yet.

In my view, the methods which the manufacturers used to try and detect this kind of contamination were really the best available at the time, and I think that the modern methods which we are now applying prove that they actually were adequate to remove it.

DOCTOR LEVINE: Thank you.

Panelists, anybody?

Sure, we can, but let me just finish up with the questions from the audience. Doctor Procopio?

DOCTOR PROCOPIO: Yes, I just want to go back to the antibodies against T-antigen. We have screened the '60s sera from a group in Italy from different clinical groups, and some of them were from mesothelioma patients. We have seen by Western Blotting that most of them were -- that they had been reacting with antibodies against SV40 antigen specific antigen.

However, we have not seen correlation with the mesothelioma or other tumors, so it seems to be that also T-antigen specific antibody, you know, you use the T-antigen as a source of specificity may over estimate the infection, potential infection.

I would like a comment from the panel about whether it is possible to select specific peptides of T-antigen that may differentiate BK, JC or SV40 T-antigen.

DOCTOR LEVINE: Panel?

DOCTOR FRISQUE: I think it's important to point out, I guess most people understand this, but the polyclonal sera against T-antigens are highly cross reactive, so that I think it's going to be very difficult to get a specific test for T-antigen. I think that's the wrong approach to try to take.

Whether you take a peptide or whatever, you may find small areas, but the homology is so high it would be very difficult to find epitopes, which probably do exist to some extent, to make a test, though, that was specific for one of these T-antigens versus another. I think it's the wrong approach.

DOCTOR LEVINE: Could you leave some light on the panel, but put the questions back on the board, so we can remember exactly what it is we are addressing this morning?

DOCTOR GARCEA: I just want to make a comment about the antipeptide antibodies. I mean, first of all, I'd like to say that I think that the serology problem is a big one, the technical problems are immense. The capsid cross reactivities between these viruses are immense. I don't think T-antigen is the way to go, there's even more cross reactivity.

But, for a particular experiment that we've actually done, we've made antipeptide antibodies against the BC and DE hyper variable loops on SV40 in an attempt to get capsid-specific antibodies, and these antibodies generated by antipeptide antibodies cross react with polyoma, among other viruses.

So, I think that there's a terrible problem here, it certainly does deserve a tremendous amount more work, but I think that, for example, even in the ELISA assays that were judged to be excellent, I think that by my criteria they are not excellent, and they are not even very good for screening.

DOCTOR LEVINE: Doctor Martin?

DOCTOR MARTIN: I have, one is a comment, and the second is a question, a legitimate scientific question.

The first comment is, I think we are forgetting some of the historical perspective. I was in medical school between '56 and '60, and on the wards in '57 and seeing the last few -- this is in Massachusetts General, in iron lungs, and went out immediately and got inoculated with the Salk vaccine. I think it was criminal not to have given the Salk vaccine in those days, so that's the comment.

The question is, in view of Andy Lewis' comment, I wonder to what extent the adenovirus vaccines and/or isolates, random isolates of adenovirus have been looked for by PCR for SV40 sequences, because while it's true that there might be a cross reactivity with the T-antigen between BK, JC and SV40, it's also true that if SV40 is being introduced via adenovirus recombinants, naturally occurring, I'm not saying this is -- that you are never going to see them.

And, if, like the LEV -- do I have the nomenclature right, it's LEV and HEV --

DOCTOR LEVINE: L-E-Y, H-E-Y.

DOCTOR MARTIN: L-E-Y, thank you, it's been so long since I've been in SV40, if those things are around you are going to -- the mode of infection may be via adenovirus, but the thing that's doing the dirty work could still be SV40.

DOCTOR LEVINE: Does anyone have data on the SV40 sequences by PCR in the hybrids? No, I don't think such data exists.

I'd like to continue with questions that are focused on the topic at hand, which is the sensitivity and specificity of serologic assays for SV40 antigens.

Yes?

AUDIENCE: I think it's unfair and inappropriate for you to characterize our call for participation and funding of independent researchers in these studies that the governments are suggesting.

Conspiracy was a word that you used, not me. I made it very clear yesterday that we understand that you didn't know at the time that you released those vaccines in the '50s, you didn't have the tests.

However, it's very important to also emphasize that when you did know that these vaccines carried monkey viruses, that fact was not communicated to the public. And, all the epidemiological data in the world, including that that was presented by Doctor Strickler and Doctor Olin yesterday, is not going to negate the fact that there are now SV40 associated cancers that are occurring in adults and children.

And, again, I think the public is interested in a full examination with the participation of independent researchers.

DOCTOR LEVINE: Well, let me respond in a couple of ways. First of all, I certainly wasn't suggesting that you were accusing the government of being conspiratorial.

Nonetheless, one reason for this workshop is to try to, in fact, do just what you suggest, to try to, by putting our heads and our collective data together, to arrive upon what is scientific truth.

There had, however, been, before this meeting, a sense in the lay press of a conspiratorial quality in the government's actions. Once again, you needn't -- we really don't want to have a debate about this, because I'm not accusing you of having accused me of being conspiratorial.

The fact remains that this is a very grey area. You heard yesterday that we are not sure, even in the best of hands, whether the SV40 footprint is or is not in tumors. Most importantly, if it is there, we don't know whether it's effect or cause. And, finally, you've heard this morning that there is much work to be done before we can, in fact, hit upon assays that tell us correctly whether this virus is or is not endemic in the human population.

Doctor Weiss?

DOCTOR WEISS: Robin Weiss from London.

To get back to the questions on the screen, I'd like to echo what Harvey Ozer suggested, that it might be worth investing some technology in competitive assays. And, I don't mean just competitive absorption with antigen, but by taking reference sera, it could be monkey sera, it could be rabbit sera, that have a known high titer, titrating them out, labeling them, going in with your human test sera to see whether you compete out the labeled antiserum.

This works superbly well with HIV, and, you know, I take in what Janet said, that SV40 is not HIV, but it enabled us to establish that slim disease in Africa actually was HIV infection, when a previous report got about 50 percent false positives. This kind of competitive assay yielded no false positives.

It's enabled clinical virologists to distinguish in rapid masquerading assays between Herpes Simplex Type II and Herpes Simplex Type I, which is of the kind of order that we might need to distinguished between BK, JC and SV40.

So, I think there are tricks of the trade that are reasonably well known to clinical virologists, but which most of us molecular virologists are not quite so good at, that could be applied fruitfully, or at least would be worth investigating, to try and see if we could get up a mass screening serological assay. Then those should be followed by confirmatory tests by Western Blotting other kinds of assays, assays for other SV40 proteins, and, of course, by PCR.

But, I think it's worth doing, and I don't think it would be vastly expensive, and I could suggest one or two names of clever people who might be able to help.

DOCTOR LEVINE: I think that would be a very good approach. If and when we get to PCR, we will, of course, have to agree on how to do it, and what it means, so are there responses from the panel, further comments from the panel on this issue?

Doctor Goedert?

DOCTOR GOEDERT: I would just endorse the idea of a competitive ELISA, because in addition to the examples that Robin cited, I think it worked extremely well for HTLV at a time when it was, you know, rare, and it is rare, and it has some other analogies, and I think it's the kind of thing that you ultimately do need some additional confirmatory assays, but it's the kind of thing that can be set up and executed pretty easily.

DOCTOR LEVINE: I had one more question that we haven't addressed on the board, the last question, how good are assays for SV40 specific antibody in immunocomprised patients? There's no particular reason to think that they are not good, since we are finding them there, and certainly, in patients with HIV infection we find reasonable levels of antibody to at least some antigens.

But, I wonder if anybody on the panel would like to address that question.

Anybody in the audience? Yes, Doctor Weiss?

DOCTOR WEISS: Well, in the majority of HIV infected patients, until they reach very late stage AIDS, there's actually a hyperplasia of B lymphocytes, so antibodies to all sorts of things you've had in the past go up. So, we've got to distinguish that HIV patients are selectively immunocompromised, and part of the HIV syndrome is an elevation of antibodies. So, I think if there has been past experience of SV40 infection, it's quite likely to be detectable.

DOCTOR LEVINE: Thank you.

DOCTOR GOEDERT: Yes, I would have to agree as well, with patients that have PML, almost invariably they have -- they are immunocompromised severely, and they still retain high levels of antibody to JC that stay fairly level. So, those antibody levels don't go down with this state.

DOCTOR LEVINE: Yes?

DOCTOR CHEN: Yes, this is Bob Chen from CDC. My comment is more of a second comment, just to correct the record in terms of when the vaccine associated paralytic polio cases after oral polio vaccine were more or less established and reported to the public, right after the oral polio vaccines were used in the early '60s there was a Surgeon General report that came out, and I believe it was 1962 or 1963, so I think in terms of some kind of cover up I don't think that really was the case at all.

Then, the second issue is in terms of epidemiology studies, I think, from Doctor Olin's report yesterday, it sounds like in Sweden they are definitely well defined cohorts of exposure, et cetera, and I think once the serology assays are made more sensitive and specific, I think it will be very nice to go into settings like that, to kind of figure out what the exact prevalence of SV40 is, and, similarly, in the U.S. kind of every ten years the NHANES does a serosurvey, and, presumably, from those we could construct certain cohorts of seroprevalence also.

Thank you.

DOCTOR LEVINE: Right, and we have the child development study bank that I talked about earlier, with 45,000 sera. Just out of curiosity, does anyone know of any other serum bank that antedates the polio virus vaccine or that followed the vaccine progressively thereafter?

AUDIENCE: I just wanted to comment. There's another cohort that you might want to look at, and this is people that have worked with SV40 for a lot of years, many who have kept antibody data. And, I know when I was in Paul's lab, Paul Berg at Stanford years back, that data were accumulated over years, and that some individuals had surprisingly high titers, so that might be an interesting source of information to look at.

I think Chuck Cole had one to 3,000 antibody titer, for example.

DOCTOR LEVINE: Right. Geissler has data, too, from the high-intensity SV40 exposure in the labs, and, again, just to remind you, the data were that 50 to 60 percent of people so exposed had titers, and they were at high levels by the plaque reduction assay.

Yes?

DOCTOR KLEIN: This may be quite a far-fetched question, George Klein, Stockholm, but, perhaps, it's not irrelevant with regard to the question of the SV40 tumor association in humans.

In the days when Bob Shubner discovered the SV40 T-antigen by complementary -- there was then an expression called the tumor of the hamster, and the titers against the large T-antigen, these were also by immunofluorescence, were highest and as an SV40 induced tumor was growing in the hamsters. And, actually, it could be used horizontally to follow the tumor as it were.

Now, with the uncertainty that now exists, as I understood it yesterday, with regard to the association between SV40 and the tumor rates detected by PCR, whether it's in all cells and all that, is there any evidence that titers are high and increase in patients with SV40 carrying tumors? And, could that be followed horizontally?

DOCTOR LEVINE: That's an interesting question. Does anybody have data on the evolution of the antibody titer in people who have been bled pre-tumor, early tumor, late tumor and so forth, anybody doing that? Doctor Carbone, are you doing that, wherever you are?

DOCTOR GARCEA: I think we discussed yesterday that there's a difficulty in a lot of these samples are archival, and for the IRB studies you can't go back and get blood samples.

We currently have before the major pediatric oncology groups, the Children's Cancer Study Group and the Inner Group Sarcoma Study Group, two prospective collections, where we look at all ependymoma, choroid plexus and osteosarcoma patients that will occur in the United States in the pediatric population. And, coincident with this, we'll gather the sera, and also test the tumors for the virus. But, in the retrospective archival studies, we cannot do those kind of studies, but, hopefully, we'll have the data, but I think the purpose of this discussion is actually, once you collect those sera how are you going to make sense of the titers.

DOCTOR LEVINE: Yes. I would also point out that, getting back to the question of cause versus effect, that if SV40 is endemic in the human population, one could hypothesize that it finds a comfortable home in a tumor cell, as a function, perhaps, of the replicative machinery of the tumor cell. And, particularly, if its replication is episomal, as opposed to integrative, it is entirely possible that, in fact, you'd begin replicating more SV40, as well as JC and BK, as a consequence of the genome simply being in the tumor cell.

So, that, in itself, is not going to separate cause from effect.

Yes?

DOCTOR TEVETHIA: Tevethia. I just wanted to send a word of caution about detection of T-antigen, and no two groups, as what I have read, are using the same monoclonal antibodies to T-antigen for the detection.

And so, a large number of these monoclonal antibodies do react, number one, with JC, and that has to be distinguished. Number two, the number of monoclonal antibodies to SV40 T reactive with seroproteins, and they must be avoided to make sure people don't -- maybe the first thing they should know about it, second, they shouldn't use it. Third thing, they should use the monoclonal antibody as precisely mapped, you know, to about eight or nine immunoacids, and everybody should use the same monoclonal antibody that are high titers, and precisely well defined, and third thing, an important test, in my opinion, is, for example, recently our lab, well, a while back, and in Levine's lab together, and now Dick Frisque, have generated monoclonal antibodies to JC that don't react to SV40 T. So, any sample that you have reacts to SV40 but not the JC antibody, to make sure definitely you are detecting SV40 T.

DOCTOR LEVINE: Thank you.

We'll take a question in the back.

DOCTOR IMPERIALE: Mike Imperiale, University of Michigan. One suggestion is to use RNA aptomers to look for cross reactivity, because those aptomers can be exquisitely sensitive and may be able to actually help out in distinguishing whether the antibody is reactive against one or the other of the viruses.

DOCTOR LEVINE: Okay.

One more question, but devoted specifically to the issue of the sensitivity and specificity of serologic assays for SV40. Do you have such a question?

AUDIENCE: No. I want to ask if anybody on the panel will discuss the famous leukemia cases from Niles, Illinois, eight leukemia cases associated with one school, and it seems that nobody is discussing that.

DOCTOR LEVINE: I'll be happy to discuss it.

AUDIENCE: Thank you.

DOCTOR LEVINE: Although I'm not an epidemiologist, but there have been many instances in the history of cancer epidemiology in which clusters of tumors have been found, the Niles leukemia epidemic, the Albany Hodgkin's epidemic and so forth.

And, when the borders of these regions have been carefully drawn and redrawn in a variety of ways, the cause and effect relationships become obscure.

But, I really would prefer Doctor Goedert to comment on this.

AUDIENCE: Each --

DOCTOR LEVINE: Could I have Doctor Goedert respond also, and then I'll give you another chance.

AUDIENCE: -- each of these leukemia cases had three or more polio shots in the '50s.

DOCTOR LEVINE: Right, so did a lot of other people who didn't get leukemia.

Doctor Goedert?

AUDIENCE: This is a famous cluster.

DOCTOR GOEDERT: Yes, I'll be happy to talk with you at length during the coffee break. Basically, cancer clusters do occur by chance alone, some of them by some kind of common exposure mechanism. It's usually an extremely difficult and inefficient way to try and sort out causation, to pursue the clusters of cancer cases, and the example that you cite, I think Doctor Levine has certainly the right answer, is that virtually every child had three polio shots, and so I think that it would not be unexpected for seven children in Niles, either randomly selected, or selected because they showed up at the hospital with a serious disease, are unlikely to be different.

DOCTOR LEVINE: Okay.

May I have the next slide, please?

Let's go on, so that we try to finish our questions before the coffee break.

Here we asked a question, should we look for SV40 specific antibody in the mothers of infants or young children with tumors that have been associated with SV40 DNA sequences. That question we have already addressed, I think everybody is in agreement that that would be an interesting study to do. Is there any other comment from the panel or the audience on that question?

Doctor Morris?

DOCTOR MORRIS: Yes. I'd like to make a proposal that might partially answer the last questions there, about the persistence of SV40. The experiment that you described earlier that we carried out the mid-1960s, these experiments were carried out under almost ideal conditions. The prisoner volunteers were housed in the clinical center here on this campus, very careful records were made of their experience while here, but I'd like to propose that someone be assigned the task of trying to find these people now, those who have died in the interim, what they died of, those who are still living, the health status. This could be carried out by a graduate student given a grant to do such work, but I think it's important that these patients might be followed, after all these years, more than 30 years, to find out whether or not there is a virus persistent, or whether or not there is not.

Irrespective of this test, of this examination, the results would be of interest.

DOCTOR LEVINE: I think that's an excellent suggestion. How many were there altogether in the original group?

DOCTOR MORRIS: There were 31 prisoner volunteers, and there were 11 controls, so some of these people must still be alive, but those who have died, the cause of death would be of interest.

DOCTOR LEVINE: Right.

The second question on the slide, does SV40 persist in humans? Which organs does it persist in? And, is viral gene required for persistence was suggested by Doctor O'Neill, Doctor Morris has started on that question.

Doctor O'Neill, would you like to dilate on the question?

DOCTOR O'NEILL: Yes.

DOCTOR LEVINE: I hesitate to use the word amplify, so I used dilate instead.

DOCTOR O'NEILL: I think the observation that Doctor Shah made that the antibody titers to SV40 persist at a level, a steady level for several years, perhaps, 30 years, I think that suggests that the virus is persisting in those individuals.

And, it would be interesting to determine which organs that virus is residing in, and how it persists, and we might try to compare that with persistence of BK and JC viruses.

DOCTOR LEVINE: Does anybody have data on the issue of SV40 in human kidneys, selectively? No.

Doctor Lednicky?

DOCTOR LEDNICKY: Addressing the second question, we shouldn't overlook the monkey model. No one has done this recently. And, I suggest that we do that, we look at some of these monkeys and using our current methodology try to culture the virus from different organs, as well as to do PCR.

And, in fact, one population of monkeys that might be interesting to look at are the ones from southern India that Doctor Shah looked at, because it would be interesting to find out if the virus was persisting in slightly different cells or organs in those particular monkeys.

DOCTOR LEVINE: Thank you.

Doctor Strickler?

DOCTOR STRICKLER: In response to Doctor Morris, I would like to say that we had the same idea to follow up on those individuals exposed during that volunteer study, and we've not been able to find the records yet. So, if any help can be brought to be able to identify these people and so we could follow up on them, we'd be very interested.

I'd like to also mention that we are also endeavoring to follow up on the patients who doctor Fraumeni identified early on in his study in Cleveland working with Doctor Mortimer there, where they were following more than 1,070 children who were exposed in the first days of their life to contaminated oral polio vaccines.

DOCTOR LEVINE: Thank you.

DOCTOR MORRIS: Sir, what is your name?

DOCTOR STRICKLER: Strickler.

DOCTOR LEVINE: Doctor Lewis?

DOCTOR LEWIS: I wanted to follow up on the question that Doctor Lednicky raised about the bonnet macaques in southern India, and I was going to ask Doctor Shah if there's any overlap in the range between the bonnet macaques in southern India and the rhesus macaque in northern India, because if there is, the question is why two species which are equally susceptible to the virus, one is carrying it endemically as an infection in a fairly significant portion of the population, and the other is completely negative. And, the question comes as to whether that's sort of a model for what could be going on in humans.

And, the analogy that I would raise is that it's my understanding that we really don't know how JC virus is spreading among us either. And, I'll ask Doctor Frisque to comment on that, but if we have JC viruses that is present in the population in a very large percentage of us, what, 80, 90 percent by the time we are 20 years old, and we have SV40 which is similar, both these viruses share properties in the sense that they really don't, in tissue culture, don't infect human cells very well. So, the question is whether there's some analogy here between human species and simian species in this regard.

DOCTOR SHAH: I think -- should I answer?

DOCTOR LEVINE: Yes.

DOCTOR SHAH: I think it is quite true that some of the macaque species are free of SV40, at least as judged by serological studies of 50, 60 monkeys, that is all that this is based on, but, for example, the Swedish vaccine was made in Javanese monkeys, they were free of SV40.

The study on the South Indian monkeys was based on looking at 60 or 70 specimens, and they were free of SV40 antibodies.

Even in the rhesus macaque, the infection can be lost, and there was one documentation of it, because a group of rhesus monkeys were moved to Cayo Santiago, which is off Puerto Rico, for a natural history study by some ecologists, primotologists, in 1938, and no new monkeys were added.

And, in serological studies we were able to show that one of the -- some of the monkeys that were born on the island of Puerto Rico had antibodies, so the infection was brought to the island, but all the animals which were under a certain age were completely free of antibodies, so as if the infection was brought, it persisted for a while, and then it was lost.

If you look at human populations, I think you can see similar things, for example, measles can be lost if the number of people in a population -- it will not survive, they cannot sustain it, and I think even in Doctor Brown's study and some of these isolated populations, I think they had some instances where the antibody prevalence was quite low.

So, with the current density of human populations, we can maintain these viruses, but in small populations I think they could be lost.

DOCTOR LEVINE: We do have access, of course, to the regional primate centers in this country, which would be an excellent resource for looking at the natural history of SV40 in various macaque and other old world/new world species. So, that's another resource that's available.

DOCTOR BUTEL: Can I?

DOCTOR LEVINE: Yes, Doctor Butel.

DOCTOR BUTEL: I'd like to ask Doctor Dorries or Doctor Frisque to just bring us up to date on the most current thinking about the transmission of JC and BK.

DOCTOR DORRIES: I think at least in part urinary excretion and transmission is correlated very closely, and several Japanese studies from families who really very nicely show that it is transmitted in the families, and the genetic footprints really show that some of the families got the virus by the parents and the virus types were, similar virus types, were transmitted to the children. And, in other families, different virus types came in the family after the kids came to school.

So, I think that the transmission, at least in part, is going within the families, and as we have heard yesterday, probably respiratory tract infection also might be involved in JC virus infection, as it is shown for some years in BK virus infection.

So, both sides of infection might be responsible for the transmission.

DOCTOR LEVINE: Doctor Frisque?

DOCTOR FRISQUE: I would have given the same answer, I believe. The only thing I would add is, again, to remind you that this is infection that occurs in children primarily, and an excretion occurs in adults at a high level, perhaps, over 40 percent of us will excrete JC in our adult life at various times, and so that, it is probably through the urine that this virus leads, as I said yesterday, probably as archetype, and that's what enters our body usually at a young age.

DOCTOR LEVINE: Yes, Doctor Dorries.

DOCTOR DORRIES: I have another comment to the Puerto Rico monkeys. They were negative for SV40 up to the last year that we got monkeys from Puerto Rico, to getting in Germany.

And, shortly after the monkeys seroconverted to SV40 positivity, and we have seropositive monkeys in Göetting, so I think transmission also was going on by urinary excretion in these monkeys, because they are caged in different cages.

DOCTOR LEVINE: Okay.

I'd like to get to the last slide, the last two questions, because we've already touched on them and we're -- our time is growing short.

The questions are as follows: Could recombinant viruses, which include SV40 sequences, affect our interpretation of the endemicity of SV40 in humans? And, could SV40 strains isolated from human tumors differ from the archetypical SV40, and if so, could this affect our interpretation of the endemicity of SV40 in the human population? These are both questions that we've already touched on, but, perhaps, Doctor Frisque, you can continue with either the first or the second question.

DOCTOR FRISQUE: Yes. I might take your second question first, and that is, in terms of the archetypical type of SV40, that is found certainly in animals, and I would say that if you look at the sequence, small amount of sequencing that we've done, that the sequence is essential identical that we've looked at with wild type forms, forms that have been rearranged.

So, in terms of changing antigenicity, I think most of the changes that occur in archetype versus rearranged forms of these viruses occur within the transcripts control region, not within coding ranges where you might see antigenic changes.

That's not to say that antigenic changes, antigenic variants do not occur, but if you are comparing archetype with rearranged forms, I don't think that's the point.

In terms of the other question, whether there's recombination with human viruses, again, we talked about that possibility yesterday, and I think it's possible. It certainly could complicate things, depending on how much rearrangement occurred and how many sequences were rearranged. Small parts of SV40 put into JC might make JC a bit more active, maybe more tumorigenic, but it might be difficult also to define those SV40 sequences in those JCs unless you do a lot of sequencing.

DOCTOR LEVINE: Right.

Other comments from the panel? Doctor O'Neill?

DOCTOR O'NEILL: Is there any indication that there are individual cells that contain both JC and BK virus?

DOCTOR LEVINE: Anybody have data? Doctor Dorries?

DOCTOR DORRIES: Actually, I would expect that similar cells might be co-infected, but I think nobody has really shown it up to now.

DOCTOR FRISQUE: I certainly haven't seen it.

DOCTOR LEVINE: Doctor Carbone?

DOCTOR CARBONE: Just a comment on the last question. From the data that we heard yesterday, it's obvious that today fortunately all the vaccine that we have are SV40 free.

However, from what is written there exactly, and from the data of Doctor Butel, it is also clear that there are some differences with all the strains, that differ from the 776 that we dealt in the past, that is the virus that, the Lyse quickly CV1 cells.

And so, it's at least conceivable to think that a virus that is only 172 base pair, or something like that, may grow less efficiently in CV1 cell, and that if that happens, and we are just testing the CV1 cells for lysis, if one is particularly unlucky could miss it.

So, what I was just suggesting is that, given the fact that -- is the PCR, whether one could just simply add what you think of that, simply add to those tests and just look at lysis of disease in molecular tests that is very sensitive, to exclude the viruses that are not SV40, as we call this before today, but are similar to it, may eventually one day be there.

DOCTOR LEVINE: You mean PCR with sequencing the product?

DOCTOR O'NEILL: Yes, any PCR amplification of that would completely rule out that something that is close to SV40, but is not the SV40 776 that we have been talking about that is a virus that will grow and lyse the cell, and so it will be obvious that it's there, could be missed.

DOCTOR LEVINE: Right.

Doctor Dorries?

DOCTOR DORRIES: I have a question. Has anybody really sequenced another wild type virus completely?

DOCTOR BUTEL: Yes. Judy Tevethia had sequenced VA 4554, and we've sequenced the Baylor strain of SV40, Renee Stewart, in the lab, has done that, and we've sequenced the SVCPC isolate, and SVPML-1 and SVMEN.

DOCTOR LEVINE: You've sequenced the entire genome?

DOCTOR DORRIES: And, they are all conserved?

DOCTOR BUTEL: Yes.

DOCTOR LEVINE: Doctor Lednicky?

DOCTOR LEDNICKY: Something else to consider related to question one. We should also think about DI particles. Now, recall that when Krieg cloned what he called SV -- what we call SVMEN, there was another virus that was cloned, and in particular it was a truncated version that had two ecolar one cites, it wasn't a complete virus. I think it was something like 3.5 kb. So, the point is, it could also be that in some of these infections that we clear the wild type virus but some DIs linger, and, you know, we might be detecting DI particles in some of these cases.

DOCTOR LEVINE: That's an interesting point.

Are there any other comments from the panel or from the audience? Doctor O'Neill?

DOCTOR O'NEILL: Well, one of the problems with the defectives is that if you clear the infectious virus, the defectives probably won't hang around very long, because they need the wild type helper, and once that helper is gone, a particle that had a genome that's only 3.5 kb is not likely to be infectious on its own, so it probably will be lost.

DOCTOR LEVINE: Doctor Dorries?

DOCTOR DORRIES: We -- in JC work, we never have seen really defective particles, in terms of rearrangement, major rearrangements, whether the TCR archetypes might be defective in a certain sort of way we don't know yet.

DOCTOR LEVINE: Doctor Oxman?

DOCTOR OXMAN: Art, can we have a word from Andy Lewis on what we know and how much follow-up of E46 with respect to the first question, the military who received E46 adeno SV40 hybrid vaccine?

DOCTOR LEVINE: Doctor Lewis?

DOCTOR LEWIS: Mike, I'm not aware of any data that was done on an analysis of military recruits. I don't think it's been done.

DOCTOR LEVINE: Did Steve Baum have any data on that?

DOCTOR LEWIS: No.

DOCTOR LEVINE: No? Okay.

Any other questions or comments? Well, if not, I think our time is virtually up. I did want to make one summary comment, though, before everybody gets up, because I think it's important to try to put the things that we hear at this meeting in perspective.

I did want to point out that I think on the basis of our discussion it's fair to say that despite the question of specificity of the earlier assays that have been used to determine whether or not neutralizing antibody to SV40 is present in the population, the fact is that virtually all studies have shown that while there is some cross reactivity, the homologous reactivity is at least 100 fold greater than the heterologous reactivity. Therefore, putting all the data together that I've heard, I think it's safe to say that SV40 is endemic, at least at a very low level, in the human population, and that shouldn't surprise us since we know that the infection is transmissible at some level from monkeys to people, and we also know that lab workers who have been exposed to high levels of SV40 for long periods of time have high titers of antibodies, at least some of which appear to be specific.

Nonetheless, there's little question that we need to improve and to standardize our assays. A suggestion of using, of course, a blocking antigen is a very salient suggestion. Whether or not we go to molecular techniques to determine endemicity will depend upon our agreement that we are using the right assay with the right conditions, the right standardization, the right sensitivity, and the right specificity. It's not an easy question in molecular biology.

And finally, once we've done all that, we must beg the question of transmission, and there the concept of looking at mothers and infants, families of people known to be infected, workers with monkeys and so forth, all become germane.

So, on that note, I thank you. Enjoy the coffee break.

(Whereupon, at 10:28 a.m., a recess until 11:06 a.m.)

DOCTOR MAHY: It's time to start this session, so please take your seats. My name is Brian Mahy, I'm Director of the Division of Foreign Rickettsial Diseases at the Centers for Disease Control, and we're moving on to a subject now which, if you like, is a little less debatable, the question of SV40 and its oncogenicity.

The papers we are going to have this morning are going to be primarily concerned with oncogenicity in rodents and cells in vitro.

We do have the opportunity, if people speak for less than the appropriate time, to have some questions, and I'd quite like that. But, the first speaker has just ten minutes, Michele Carbone, from Loyola Medical Center in Maywood, Illinois. Are you here, Doctor Carbone? God.

Return to Agenda

DOCTOR CARBONE: Thank you.

In this talk, I was asked to review the data about SV40 oncogenicity in hamsters.

Can I have the first slide? Great, okay.

So, I'm going to review the data about SV40 oncogenicity in hamsters. In 1961, Doctor Eddy reported that tumors were induced in hamsters by monkey kidney cell -- and a year later, in 1962, she identified the substance present in these kidney cell -- responsible for the oncogenicity of SV40. These were subcutaneous tumors that from an histologic point of view would be called sarcomas.

But, at the same time, Doctor Kirschstein reported that if you injected the virus in the brain of hamsters, they would develop ependymomas, and ependymomas also develop in mastomys, which I understand is a kind of rat, when they were injected with SV40 into the brain.

Some years later, Doctor Diamandopoulous started what would happen if you injected the virus systemically, and he injected SV40 to the femoral vein. And, when he did that, with his great surprise, he found that only specific cell type would develop, and they are indicated here. One hundred and fifty animals were injected, 125 of them developed tumors, the tumors, as listed below, obviously, more than one tumor developed in some animals, and you have the tumors that developed were abdominal and mediastinal lymphomas, also sarcomas, and one single lymphocytic lymphoma.

Now, the oncogenicity of SV40 is related to the large T-antigen. The role of the small T-antigen, if any, in the oncogenic process was unknown. And, Doctor Lewis and Doctor Martin studied this problem and addressed it. And, they published that if you injected hamsters with small T -- they published a study in 1979 in PNAS -- that if you inject hamsters with SV40 multi mutants these mutants are still able to induce tumors. However, they found it prolonged the latency.

A few years later, Kathleen Dixon reported that if you injected these multi mutant subcu, in addition to these tumors with the prolonged latency that were reported before by Doctor Lewis, some animals developed abdominal metastases. Now, these metastases never developed in animals injected with SV40 wild type.

But, at that time, actually, two years later, I began a post doc in the lab of Doctor Lewis, and we studied those tumors, and to our surprise we found out that those so-called metastases were not metastases, those were primary lymphomas that would occur in these animals.

So, for some reason, when you inject the se multimutants subcutaneously a portion of animals would develop primary abdominal lymphomas, it's an ugly terminology, but these are macrophage lymphomas, or to use a human term, true histiocytic lymphomas.

We were particularly intrigued by this finding, why that would happen, and why that would happen only with multimutants, and so we decided to inject a variety of multimutants systemically through the heart into hamsters to see what was the oncogenicity of these multimutants when different organs were exposed to the virus, and, of course, we had the control group animals that was injected with wild type SV40, and in the control group of animals we found the most unexpected results, but let's go in order.

These were the tumors that developed in animals injected with multimutants, these are lymphomas, and all of them were -- actually, not all of them, most of them were true histiocytic lymphomas, few of them were bilymphomas. So, for some reason when you inject the virus systemically, that's basically the only tumor that you see when you use this multimutant. There must be a different reason, and I don't have time to go through all of them.

Rarely, a multimutant injected animal will develop an osteosarcoma that is shown here, also sarcomas did develop in the control group of animals that was injected iwth SV40 wild type.

The most surprise thing was that in the control group of animals injected iwth wild type, we saw these tumors, and these tumors are mesotheliomas and you can see in A and in C the epithelial -- the basic pattern that is kind of characteristic of this malignancy. Now, we were very surprised by that, because, as I mentioned yesterday, mesotheliomas have never been associated with anything else than asbestos exposure, at least in mammals.

In D, you can see a cell line that -- cell culture derived from one of those tumor, and this is one of the tests that we did to characterize these tumors in these cells. On the last day of the original tumor, -- of the cells derived from it, indicating that they are representative of the original tumor. They really look like a -- and there are some of the characteristic electron microscope in these mesotheliomas, including those branching microvilli that some of you can appreciate.

Obviously, you would like to see that the virus is, in fact, integrated into these tumors, and that is responsible for these tumors, and that is, in fact, the case. In panel A, you have the line alternate, tumor cell line, tumor cell line, each line is derived from the tumor, and they are cut with a single cut of Bam HI or EcoRI. It's obvious a number of things. First of all, that the pattern of integration in the tumor and in the cell line is quite different, and so probably rearrange them and -- at least in cell culture, maybe in the tumor, too.

The second thing that is obvious is that there is -- in A but there is also about 5.1, 5.2 kb band, which could represent episomal DNA. So, in B, we got to that DNA with a non-cutter, and when we use a non-cutter for SV40, only high molecular weight is there, indicating that the -- tumor at least, all or most of it, is SV40 integrated and not episomal, and finally in panel C Hind III showing the characteristic pattern of the early region, that is, the region that calls for large T and small T, and that you would expect to find there if the tumor if the virus is playing at all in tumor -- in oncogenesis.

The lines that are white in panel C, the panel cut with Hind III, are from the heart of these animals, indicating that at least by Southern Blot you cannot detect SV40 in the organs of these animals that do not contain tumor, but only the tumor.

These cells express large T-antigen, 90 kilodalton, and if they derived from the wild type also there's multi-antigen, 17 kilodalton, 19 kilodalton. They were strongly positive by immunoperoxidase experiments, and these experiments were done by Doctor Harvey Pass at NCI a few years later than those I showed before. When you take this, these are hamster mesothelioma, SV40 induced mesotheliomas and cell lines derived from these tumors, when you take these cells and you inject them into hamsters they are highly oncogenic, they are oncogenic up to 10² cells, when you inject 10² cells they will develop tumors in about eight weeks, when you inject 10⁵, 10⁶ cells they will develop -- the animal will develop tumor in about four weeks.

And, it doesn't matter where you inject it, you can inject subcu and you get the same thing, the tumors grow very quickly.

So, these are the conclusions from the hamster studies, at least these are our studies. Wild type SV40 injected intracardially, 60 percent of animals develop mesotheliomas, 40 percent true histiocytic lymphomas, five percent osteosarcomas, five percent sarcomas, when we inject the wild type SV40 into the pleural space, 100 percent of animals develop pleural mesotheliomas in three to five months. When wild type SV40 is injected intra peritoneally, mesotheliomas and true histiocytic lymphomas can both develop. When wild type SV40 was injected intracranially by other investigators, ependymomas and choroid plexus tumors developed, and if you delete this multi-antigen and you inject the virus systemically only true histiocytic lymphomas develop, true histiocytic lymphoma will also develop if you inject the virus subcu in a minority of animals, in addition to the local sarcomas.

And so, these are the conclusions of this study. In hamsters SV40 preferential induce mesotheliomas, osteosarcomas, sarcomas, specific types of lymphomas and ependymomas. These same tumor types have been shown to contain SV40-like sequences in humans. Deletions of this multi-antigen out of the oncogenicity of SV40.

Thanks.

(Applause.)

DOCTOR MAHY: Thank you very much, indeed.

There's time for one quick question, if anybody has one. We are short of time, but Doctor Shah?

DOCTOR SHAH: In the earlier studies in the '60s, the lots of SV40 innoculated hamsters, did they also detect mesotheliomas, or this is a new finding?

DOCTOR CARBONE: There is one single report of a cell line that was derived from mesothelioma. I am very sorry I can't remember the first author of that paper, however, it went this way. There was the cell line, it's called TU something, and I called him, and he explained to me that at that time, in the '60s or so, they were injected 100s of hamsters between the scapula, and that then they would send these tumors to pathology, and that one of these tumors came back saying that this was a mesothelioma, and that everybody was intrigued with that and he'd send the cell line around to many places.

And, his own opinion was that since these were small animals, he thought that probably the technician who did the injection in that case went deep enough to reach the pleura and that's why the mesothelioma came out.

So, there is one only report and the cell line is called TU-800.

DOCTOR SHAH: Thank you very much.

DOCTOR MAHY: Okay.

We're going to move on now to a paper by Priscilla Furth, who is from the new Institute of Human Virology in Baltimore, Maryland, on SV40 rodent tumor models as paradigms of human disease transfusing mouse models.

Return to Agenda

DOCTOR FURTH: Thank you, and Doctor Lewis invited me here to discuss some of the transgenic mouse data using T-antigen.

First slide. And, as a way of introduction, of course, the vast majority of transgenic mouse studies do not look at viral infection, rather, they focus on a specific viral protein, and I'm going to tell you this will be the SV40 large T-antigen. In many of the constructs, the small T-antigen is there as well, but it may or may not be expressed.

So, these studies differ considerably than from what you've heard before, because we are looking at the action only of the transforming region from the SV40 virus.

Most of the studies have focused on viral oncogenesis at the process of viral oncogenesis, and they have looked in a number of different tissues. I would venture to say that the large T-antigen coding sequence may be the most frequently injected DNA sequence in transgenic animals. It's been expressed in a wide variety of tissues, and it's also been used as a tool, basically, to look at some new technologies as well. So, there's a number of studies out there, and what I will do is present data from my own laboratory which illustrates some of the common themes that have been seen.

I would mention that there is a transgenic mouse model which does use the promoter from SV40, and it does target to the choroid plexus, and Terry van Dyke's laboratory has done a lot of work on that particular model.

But, if you look around and do a MEDLINE search you can come up with models, liver, pancreases, Doug Hanahan has done a lot of work there, intestine, lung, kidney, the lens of the eye, bone and cartilage.

In my own laboratory, we focus on the mammary gland and the salivary gland.

The large T-antigen is interesting for those of us who are interested in viral oncogenesis, because, of course, it binds to and inactivates two very important tumor suppressor genes, the retinoblastoma protein and P53.

So, now I would like to discuss our mammary model, and similar to all of the other animal models, if you express SV40 large T-antigen in mammary epithelial cells you will, in fact, produce tumors, and this is a transgenic mouse which has a mammary tumor. Mice, of course, have ten mammary glands, five on each side. These tumors come up in these mice after about three to four pregnancies, or about four to five months of age. So, there's a considerable latency there.

What's interesting to us about this particular model is that we can look at early steps which follow expression of large T-antigen. The promoter, which we use to drive T-antigen expression to the mammary epithelial cells, is under a strict developmental control. We use the whey acidic protein promoter. What's important for you to remember is that this promoter turns on around day 13 of pregnancy in these cells, and what we've done is then look at the glands at day 18 pregnancy. This would be five days after initial T-antigen expression, to see what, in fact, has happened to the cells.

On the top panel, this is alveolus in the mammary gland here, and this is another alveolus that's here. These alveoli are embedded within a fat pad in the mammary gland. This is T-antigen immunohistochemistry, and so you can see the typical staining of the nuclei in these cells. In this particular transgenic line, T-antigen expression is virtually homogeneous in these cells, in other words, all the cells express T-antigen.

One of the interesting findings that we saw was that expression of T-antigen in these cells actually induced programmed cell death or apoptosis, and you can recognize the cells which are undergoing programmed cell death because they appear brown here in this in situ stain, which can identify cells undergoing this process. So, again, we have the alveolar structures here, and you can see one to two cells in each of these alveolus is actually undergoing programmed cell death.

This was somewhat interesting to us, because, of course, the P53 and retinoblastoma protein are bound by T-antigen, P53 has been implicated in a number of processes involved in programmed cell death. In these animals, P53 is, in fact, bound up and inactive, yet we see induction of apoptosis, and the studies that we followed up with, in fact, demonstrated that this is P53 independent apoptosis.

So, we see that very early after expression of this oncoprotein, the cells are fighting back and trying to eliminate probably those cells which may, in fact, have DNA damage.

One thing I should mention is that when we express T-antigen in these cells, virtually, all the cells are polyploid, so if we do a fogan stain, which can recognize DNA and stains do DNA, all these cells will light up with multiple copies of their cellular DNA, so we've introduced into these cells an abnormal cell cycle, and it may well be that these cells are undergoing programmed cell death because they are recognized as being defective.

What happens during the process of oncogenesis, however, is that the cells develop a resistance to P53 independent apoptosis, and this we can see in the mammary gland by looking at the involution. Involution is the process which follows lactation. It is the time when all the mammary epithelial cells die, through the process of programmed cell death, and what you can see in this normal gland, which is now 13 days after lactation has ceased, you can still see all of the residual ductal structures, but you do not see the epithelial cells here.

In contrast, this is a T-antigen animal, which has undergone three pregnancies, three lactations, and what you'll see now is that this gland no longer regresses, and this is evidence that the gland has developed resistance to P53 independent apoptotis, so this would be one of the themes that you see during the process of T-antigen oncogenesis.

We've also looked at these cells at the day 18 pregnancy for other evidence of abnormalities. This is an H&E stain, again, the alveoli. This is the fat pad within here. You can, in fact, pick up those cells which are undergoing apoptosis, they stain darkly within these alveoli.

What I want you to note on this particular slide is that the cells do not appear dramatically transformed. They are single layer cells, and they are resting upon a basement membrane. Nevertheless, the cells have considerable functional abnormalities. One of the things that happens in the mammary epithelial cells is they are no longer able to process and secrete milk proteins, and this is immunohistochemistry in the T-antigen animals here, and controls, we look for milk proteins, one of them the whey acidic protein, and this is an antibody that recognizes total milk. You can see that the T-antigen animals do not secrete any of these milk proteins, and this can be demonstrated again on a Western Blot, three, four and five are transgenic samples, one and two are wild type animals.

Initially, we thought this was because the cells was dedifferentiated, T-antigen has been "associated" with the dedifferentiated phenotype, but we found, in fact, that these cells were largely differentiated from mammary epithelial cells.

In this Northern Blot, we are looking at the expression of milk protein genes in mammary epithelial cells. These can be used as a measure of differentiation, and what we saw, that this is control animals here, beta casein, wap, alpha lactalbumin, we can see that in these T-antigen animals we saw RNA expression of all these genes, this means that the deficit that we are seeing in these animals is very specific and related to protein synthesis. And, here's your T-antigen expression.

While these genes are expressed normally, there are examples of other genes which have their transcription patterns interrupted. One of the interesting ones is something called WDNM 1. It's interesting because this gene was originally identified in mammary cell lines as a gene which is down regulated in metastasis. Metastasis, of course, is a very late phenomenon in oncogenesis, but we see down regulation of this gene only five days after T-antigen expression, so it probably indicates that this gene is not, in fact, a metastasis factor, but something related to early changes within the cell.

And now, to look at some of the later steps in oncogenesis, I'm going to turn to a different transgenic model, and this model we used a binary system, and the reason that we have chosen to do this is we would like to temporally control the expression of T-antigen, in other words, we want to use an animal, but we want to be able to turn T-antigen expression on at a specific time, and then later we want to be able to turn it off at a specific time. This enables us to perform experiments in which we can look at the time dependency of viral oncogenesis.

The system that we use is a tetracycline responsive gene expression system. It consists of two genes here, and what we do is drive T-antigen expression off a promoter, which I've termed here the tet-op promoter, this is a minimal promoter, it contains 50 base pairs around the CMV tata box, and it is linked to seven tet-operator sequences from the E. Coli repressor transposon.

Those DNA sites which are linked the CMV minimal site, contain no eukaryotic transcription factor binding sites. What that means is, this promoter is, essentially, silent in a eukaryotic cell. So, it's off.

We turn this promoter on by expressing a specific transactivator, which can bind to the promoter and increase gene transcription. This transactivator is contained on a second transgene here. It consists of the protein domains which can recognize and bind to these DNA domains, and also comes from the repressor, from the E. coli transposon. This is linked to an activator domain from herpes simplex virus.

And, what essentially that does, this hybrid transactivator is a eukaryotic transcriptional activator and has turned what was a repressor protein in a bacterial system into an activator.

We control the binding of this transactivator by the administration and withdrawal of tetracycline. The DNA sequences from the repressor contain a binding site for tetracycline. When tetracycline binds to this particular protein, it changes its confirmation, and it can no longer recognize and bind to DNA. However, in the absence of tetracycline it can bind to this promoter, and you see gene transcription.

So, this is a system then that we can specifically turn gene expression on and off at time points.

The cells that we used the system to look in are the striated ductal cells of the submandibular salivary gland, and we chose to do the experiment in those cells for a very good reason. We had, in this particular transgenic model, very homogeneous expression of T-antigen in those cells, and this would allow us to read out, if we turned off T-antigen expression in a large population, the phenomenon that we wished to look at.

Striated ductal cells are characterized by these pink striations in here. They are absorptive cells. These are their nuclei, rather round here, and placed eccentrenically.

This slide, we sometimes will put a reporter gene here besides T-antigen. This is a nuclear localized betagalactacydase gene, and what it demonstrates is we target betagal expression to the striated ductal cells, similar to what we did with T-antigen, so these are these cells with T-antigen expressed in them for approximately four months, and what you can see now is the cells have, in fact, become transformed, the nuclei are eccentric, and you can see a hyperplasia is developing off here.

This is a Western Blot demonstrating expression of T-antigen in the presence of the transactivator here, and you can see we see excellent levels of expression. This is actually the earlier model, you looked at the mammary gland model, you can see we express actually significantly greater amounts of T-antigen protein using the binary system here.

This is a single transgenic animal. It contains only this construct, and you can see that there's no T-antigen expression in that animal.

If we look over time in these animals, the first time point that we've been able to examine is two weeks of age, and at two weeks of age you can find foci within the submandibular salivary gland which are hyperplastic, but if you look in a larger field you'll see that these are rather limited, and there are many cells which do not exhibit any hyperplastic changes.

If you follow that mice up to about four months of age, however, you'll see that the vast majority of the striated ductal cells are now hyperplastic.

We chose this time point then to turn off gene expression and ask whether or not the hyperplasia that we were observing was dependent on continued T-antigen expression. In other words, if we turned off T-antigen expression, would the hyperplasia reverse or would it be maintained.

So, to turn it off then, we placed these animals on tetracycline, and what we saw, in fact, was a very dramatic reversal of the hyperplastic changes.

Coming across here, this is the nuclear localized lacZ that you saw before, but you are seeing it at a lower power. It outlines the structure of the striated ducts.

This is a non-transgenic animal, you can see that there's pink staining within here. Those are all striated ductal cells. This is a single transgenic animal, it does not express T-antigen, and it looks indistinguishable from wild type. This is a double transgenic animal, in which T-antigen has been expressed for four months, we see dramatic hyperplasia here. We placed this animal on tetracycline for three weeks, and what you see is extensive reversal of this hyperplastic phenomena, and this would suggest then the hyperplasias at four months of age are, in fact, dependent on continued T-antigen expression.

One of the interesting controls about this is that you can see that the density of the striated ductal structures is actually increased over this animal, and that let's you know that, in fact, hyperplasia was there.

If we look at a higher power, we can see an animal that is not on tetracycline, so T-antigen expression, hyperplasia, we see the immunohistochemistry for T-antigen. This animal has been on tetracycline for three weeks. The histology of the cells is reversed, and there is no T-antigen on immunohistochemistry.

We then performed the same experiment at seven months of age and asked the same question. And, in contrast to the results at four months, by seven months of age we find that reversal of the phenotype is very limited, and this is illustrated here. This is a seven month sample. You see hyperplasia in the absence of tetracycline. We placed that animal on tetracycline for three weeks and the hyperplastic changes remain.

This would suggest then that this hyperplasia no longer needs T-antigen expression, and in a brief look at mechanism we'll return to our four month sample, one of the things that you do find as T-antigen expression is correlated with the appearance of polyploidian cells.

And, this is a fogan stain here, the DNA which is stained in multiple copies, appears dark pink here, these are cells which have T-antigen in them, and these cells are polyploid.

When we turn off T-antigen expression at four months, there are a few foci which remain transformed. They do not express T-antigen, however, they maintain their polyploidy state, and, therefore, we would suggest that there may be genes, cellular genes, which have become mutated in this process, which are now sufficient to keep the cell in a transformed state.

Thanks.

And, I'd like to acknowledge Ming Lan Lee, is a Post-Doc in my lab that worked on both of the projects, and Dagmar Avol, a Post-Doc in -- lab that did the second project.

(Applause.)

DOCTOR MAHY: Thank you very much, Doctor Furth. That was a very interesting presentation. We have time for a couple of questions.

AUDIENCE: Have you looked for P53 localization at any time during this? For example, is P53 localized to the T-antigen expressing cells early on, and does it stay in the nucleus, for example?

DOCTOR FURST: Yes. Most of the P53 studies I've done are on the mammary gland, and when you express T-antigen P53 is localized to the nucleus in those cells.

In mammary epithelial cells, there's actually very low levels of P53, although, it's somewhat strain dependent. So, it's difficult to pick up, but it's there, it's in the nucleus.

AUDIENCE: And, that reverses upon tetracycline?

DOCTOR FURST: I did those studies in the mammary gland, not in the salivary gland. Probably P53 plays a different role in the salivary gland, and it may well -- that will be something interesting that we could look at.

If you take an oncogeny, and you express it equally in mammary tissue and salivary tissue, you breed that animal into a P53 null. You will find that you don't change the mammary tumor incidence, but you do change the incidence of salivary tumors. Bottom line, P53 may play a more prominent role in salivary gland tumorigenesis than it does in mammary gland tumorigenesis.

DOCTOR MAHY: Question in the back?

AUDIENCE: Yes. A very nice study. The question is on the mice that you did show that the reduction hyperplasia, when you didn't have complete removal of hyperplasia, did you look at the promoter to see if it was mutated or recombined, or was it just the distribution in tetracycline? Can you explain why you didn't get complete reversal?

DOCTOR FURTH: In the four month, why that one foci remains there?

AUDIENCE: Yes.

DOCTOR FURTH: I suspect it's because those -- you put T-antigen in, you cycle these cells, and you start to accumulate DNA damage. A lot of it may be silent. Some it eventually statistically will hit on a critical gene, maybe rats or something, mutates it.

So, my working hypothesis is that particular clone of cells contains a mutation in another oncogeny, and we need to demonstrate that.

DOCTOR MAHY: Last question, Doctor Ozer?

DOCTOR OZER: Harvey Ozer. The question I have is, I'm familiar with the tet system, and we've used it, and some people do find a trace of leakiness in some of the systems.

So, I was just going to ask you, how close to zero is the T-antigen gene in these cells, and have you done RT-PCR or something?

DOCTOR FURTH: I have not done RT-PCR. There may be a threshold. All I can say is that we are below the threshold of detection by immunohistochemistry, because we're not so much interested in RT-PCR as we are interested in protein.

There are studies, of course, that have looked at levels of T-antigen and oncogenesis, and there may be some correlation there. So, I know that we are down below the level that we can detect on a protein level.

One of the other provisos with the tet system is that, yes, in tissue culture cells I've got my own data where it can be leaky. We have, reassuringly, found that in transgenic animals it has acted fairly tight, but, again, I can't tell you if we went in with RT-PCR if we might pick up a few transcripts. But, I'm looking at a particular threshold.

DOCTOR MAHY: Thank you very much, indeed. This is a very nice study.

Now, we're going to move on to transformation of cells in vitro, and Kathy Rundell from Northwestern Medical School is going to talk to us about transformation of rodent cells.

Return to Agenda

DOCTOR RUNDELL: I'm going to spend about the first third of my talk talking generally about transformation systems that have been used to study SV40, and I've drawn very largely, of course, on laboratories like those of Jim Pipas, and Chuck Cole and Judy Tevethia, who have studied so many of the mutants, especially in the SV40 large T-antigen.

I want to start with just describing a couple of transformation systems, in which the large T-antigen is sufficient. It's very clear that this is the major transforming protein of the virus, and there are many assays where this is the only protein you need, and a couple are listed here, in focus formation assays, in particular, and in many agar colony formations in certain mouse lines.

The important large T-domains are three in particular. First, you've already heard about the domain that binds to the RB protein and its family members. This is clearly an important region of the protein, and that's been well documented in transgenic mice.

There's also a very important region of large T that maps to the C terminus, and this is the region that binds the P53 protein.

In the transgenic studies that have been done, there is a relationship between the amount of apoptosis that goes on and the sequences being present from this region of the protein, and it raises the interesting possibility that part of the effects of P53 binding are just to suppress the apoptopic responses of the cells, and you just heard some reference to this in the last talk, although, we also, of course, have non-P53 related apoptotic mechanisms.

Another interesting possibility was just found recently in reports that large T-antigen may interact with a cellular protein, P300, through its binding of P53, and this raises the possibility that some of the transformation effects of large T could be mediated through effects on this P300 protein.

A third domain that I'll come back to in a few minutes, that's very important in transformation, maps to the immunoterminus, is a region that's been heavily studied by Jim Pipas. There are at least two regions that seem to be involved, sequences mapping to a region in residue of 17 to 27, another that maps to 42 to 47, and I'll come back to these in a few minutes with respect to small T that I'd like to talk about, and you'll be hearing more about this region this afternoon, because this is now identified as a DNA J region of large and small T-antigen.

Although, only large T-antigen is required in some transformation systems, there are several that also require a small T for efficient results. The ones that were initially identified had to do with the ability of cells to grow in semi-solid media, or agar assays, or anchorage independent growth assays. The initial reports were in rat F111 cells, and I just mentioned these on the last slide, that you don't need small T for these cells to be transformed to produce foci.

However, Noëlle Bouck originally showed that when cells were infected with viruses that lacked small T-antigen, and then plated in semi-solid media, their ability to form colonies was greatly impaired.

We've subsequently showed that the ability of small T to bind to cellular protein phosphatase, pp2A, is critical for this ability of anchorage -- inducing anchorage independent growth.

The Livingston Laboratory extensively used mouse 3T3 micro colony formation, and this is also a semi-solid growth agar assay, and showed a dependence on small T for that transformation.

And then, I'd also like to note, because I thought this is a very interesting study historically, work done by Setlow & Martin in hamster -- Chinese hamster lung cells, where they had the same finding that was made in rat F111 cells, that small T was required for efficient growth when cells were plated directly in agar, but as shown in this cartoon slide, if they took the cells and plated them directly into agar they would get a very low transforming inefficiency, but they did the interesting subsequent experiment where they took the same infected cells, passed them in culture once or twice, and then found that the requirement for small T was greatly alleviated, and this raised the possibility that what small T was primarily providing in these assays was growth impetus. And, I think that is very much similar to the kinds of studies that Michele Carbone just described this morning, in the kinds of tumors that appear in animals, with and without small T, and that the tissues that are non-growing tissues are much more likely to require small T for tumor formation.

So, to summarize the results of several laboratories up to this point, it appears that efficient transformation by SV40 is always intimately linked to its ability to induce cell growth.

I wanted to remind you at this point of the experiments by Hscott & Defendi nearly 20 years ago, where they showed that infection of mouse embryo cells with SV40, it was possible to induce several rounds of growth in these tissues or these cells, and that if you took -- did the same experiments with viruses that lacked small T that cells would undergo a single round of division, but then proceed no further, again, underscoring a requirement for small T in prolonged growth.

This kind of situation may be what's going on in a third kind of assay that's dependent on small T, and that's when small T is required for focus formation. In the assays where small T is required to form foci, and two are listed here, the assays that are used tend to be monolayer overgrowth assays, and so the distinction that I make here is that you are asking the cells, not only to behave as transformed cells and to appear morphologically transformed, but they also have to overgrow contact inhibition and density arrest.

Other kinds of assays have been done frequently, where cells are plated at rather low density, and observed for morphologic appearance of transformed cells, and frequently those cells are never forming full monolayers, and it's a different kind of assay, and it rarely depends on small T.

The other cell type that shows this dependence on small T for monolayer overgrowth is human diploid fibroblasts, and I'm going to -- we've been spending a lot of time looking at cell cycle parameters in these cells that I'm going to show you some data on in a few minutes.

So, my laboratory has been interested in defining for some time what is it that either large T or small T, or both proteins, can contribute to transformation systems that are dependent on small T. And, this cartoon just diagrams some of the important regions of small T that we've focused on, and these include, roughly, there is a unique C terminal half to small T-antigen that's diagramed here. It's a very cysteine rich region. I've already referred to sequences that regulate the ability of small T to bind and inhibit protein phosphatase 2A. These regions have been required in every assay we've ever looked at, if it's dependent on small T these sequences are required.

This region here I won't talk much about. It's involved in zinc binding and stabilization of the protein.

But, I'd like to point out particularly this region here. Now, this half of the protein is shared with large T, and there is a very heavily conserved hexapeptide HPDKGG that we became interested in simply because of its high conservation, and you are going to hear more about this this afternoon with large T antigen, because this is the DNA J homology region, part of it.

These sequences in SV40 small T-antigen are linked to its ability to transactivate, and this was originally described by Mary Lakin, that small T could transactivate various promoters when co-transvected into cells.

We've been studying this region, not only for that transactivating activity, but more recently we've been excited by the ability of these sequences to transactivate the cyclin A promoter in a transient assay, and even more importantly to increase transcription of the endogenous cyclin A gene when small T is introduced into cells.

It turned out that this same region was necessary for small T or for the virus, for SV40, to transform human diploid fibroblasts in the monolayer overgrowth assays.

There is also a dependence on a region down here in the sequences that map in amino acids 17 to 27, and that region is required in human diploid fibroblast transformation as well.

Surprisingly, in the assays that we've performed, either large T or small T has been able to contribute either this region or this region, and so they seem to be complementing one another for whatever this is providing in transformation of cells.

Recently, we have turned to FACS analyses to begin to study a little more carefully the systems that require small T and large T for efficient transformation, and to do this, and to separate the functions of the two viral proteins, we have used adenoviral vectors in which inserted into the E1A region of these defective adenoviruses is a transcription unit with the cytomegalovirus promoter driving either the expression of small T or the expression of large T.

And, when you introduce these into cells, this is now, for those of you who don't look at FACS analyses, cells are stained with preputium iodide, which intercalates into the DNA, you get a very nice 2N peak or normal diploid peak of DNA in most cell cultures. Those cells that are either in the G2 phase of the cycle or that are possibly even naturally polyploid or tetraploid are shown at a double fluorescent area.

When you infect CV1 cells, and that's what is shown on this slide, when you infect CV1 cells with this adenovirus that expresses small T, you can see a significant increase in the number of G2 cells, or 4N DNA content cells. And, in CV1 cells, which is just a permissive monkey cell line, either small T or large T can drive these cells into the cell cycle. You don't need both proteins.

In contrast, the other systems that are related to those that I just explained to you, that are small T dependent transformation assays, will require both large T and small T for efficient cycling.

The first system I'd like to tell you about is, again, with the F111 cells, the rat cells, that if we suspend these cells on agar coated plates, so they are no longer able to spread out and adhere, you see very little evidence of cell cycle progression in these cells. This is the 2N region or G0G1, and what's very obvious is this sub G1 population of cells that are reminiscent of apoptotic cells.

We haven't formally proven yet that this is apoptosis that's going on, but others have shown in other cell lines that if cells are normally adherent, and you try to get them to survive in non-adherent states, that this is no unusual for apoptosis to occur.

This is a rather preliminary experiment, but it was quite intriguing, that if we pre-infected, and this was two days now before the experiment was started we pre-infected with SV40 in the absence of serum, we seem to have reduced considerably the amount of this sub-gene 1 or apoptotic peak, and we are now pursuing this direction and trying to ask whether infections with SV40, with or without small T-antigen, may delay or reduce the amount of apoptosis that occurs when cells are no longer allowed to adhere to substrates.

We know that this cannot completely suppress, because of experiments like this one shown here. In this experiment, we have infected the F111 cells with either the adenovirus that expresses large T or with two viruses, one expressing large T and one small T, and there are two points I'd like to make here. One is that you have a tremendous amount of cell disintegration. That's all this material here, and that's whether or not the viral proteins are present, that occurs.

But, more interestingly, it's a little hard to see because the scale is so high here, the G2 populations, there are clearly -- to the extent that these cells are able to cycle when they are placed in non-adherent conditions, that cycling is dependent on both the large T and small T antigens being present.

I haven't shown you the other controls here, but there's considerably less G2 when the cells are infected with large T alone.

The second system that I'd like to show you briefly is the human diploid fibroblast cells, and I remind you that in these assays the system that requires small T for transformation is overgrowth of monolayers or focus formation above a confluent monolayer.

And so, the first question we needed to ask is what is the arrest like in the cells that are maintained at confluence as opposed to serum deprive or sub-confluent cells, and that's what's shown here. In this experiment, we took either sub-confluent cells or confluent cells, and then deprived them of serum for two days before infecting them with the various adenoviruses, or, I'm sorry, before restimulating them with serum to just follow their normal cell cycle progression.

In the top panel, you can see that sub-confluent cells rather rapidly and efficiently reenter a cycle, and the majority of cells here have already gone either into S or into G2, and, ultimately, come back to a nice G1 population.

In contrast to that, the cells that are confluent, when they are restimulated with serum, the rate of movement through the cycle is much delayed, so that we are only getting really significant G2 peaks 30 to 34 hours out, and the efficiency is very low. A large number of cells never move into this cycle at all.

So, the important thing here, I think, is that the block that cells face when they are confluent is more than just a serum deprivation or the kind of block that you see when you are depriving cells of growth factors.

And then, to look at the effects of the various viral proteins on these cells, the important point from this experiment is, when we take these same kinds of cells that are held at confluence, infect them with the various adenoviruses that either express small T or large T, that you need both large T and small T for efficient movement into the cell cycle, the pronounced G2 peaks here.

Neither small T or large T alone is very efficient at doing this, and this is related, as I told you earlier, to levels of cyclin A that these viruses are able to induce in these cells. Now, this is a Western, not a Northern or RNA analysis, but you can see very clearly that when cells are infected with the adenovirus expressing small T you definitely get some induction of cyclin A, you get some, as well, with large T expressing virus, but for really efficient expression of cyclin A you need both viruses to be present.

And then a final point that I would like to make that's clear in this analysis is, the very striking -- we repeatedly find this very striking accumulation of 4N cells in cell cycle analyses. Now, you've heard this referred to just in the last talk very nicely, but also yesterday when Mike Imperiale talked about BK virus, that there does seem -- and, this was originally reported, or not originally, but it was reported recently by John Layman's lab, looking at large T-antigen in monkey kidney cells, that there does seem to be a possibility for these viruses to induce tetra ploidy and certainly aberrant DNA profiles in cells. We aren't yet positive what's going on here, but we are currently trying to sort out whether this a cycling 4N population or whether it's an arrested G2 population, but we see it pretty reproducibly in these kinds of experiments.

So, in summary then, in vitro transformation systems have been extremely important in understanding SV40 transformation in tumorigenesis. Over the years, these studies, and those of other DNA viruses, have highlighted important cellular molecules, such as P53 and the RB family members, proteins that play key roles in controlling and regulating cell cycle progression.

It's important to remember, I think, that the SV40 small T-antigen contributes to these activities as well, and may be of particular important under particular conditions of growth arrests, such as those imposed by density or by anchorage independence. Thank you.

(Applause.)

DOCTOR MAHY: Thank you very much, Doctor Rundell.

There's time for just one quick question, if anybody has one. Yes.

DOCTOR FRISQUE: Dick Frisque. Is there any evidence that the 17KT has any influence on the transforming behavior?

DOCTOR RUNDELL: Can you say it again?

DOCTOR FRISQUE: Is there any evidence that 17KT has --

DOCTOR RUNDELL: Oh, 17K.

DOCTOR FRISQUE: -- has any influence?

DOCTOR RUNDELL: We looked -- we tried to look at that, and we couldn't find any evidence for much of a role for 17KT in transformation. I think in the original reports of Depper it showed that it was weakly transforming, it has a very weak transforming activity. I didn't find anything really to add to that.

DOCTOR MAHY: Thank you very much.

Okay, we'll move on now to Doctor Harvey Ozer, from the New Jersey Medical School, who is going to talk to us about SV40 transformation of human cells in vitro.

Return to Agenda

DOCTOR OZER: As you undoubtedly believe, SV40 does transform human cells. SV40 also, as you know, undergoes a permissive infection, and so -- termed semi-permissive -- and so, there is actually a complex virus cell interaction.

We were interested in trying to sort out the two parameters, namely, virus replication and transformation, so we chose to investigation the interaction of SV40, we cloned sequences which have a six base pair deletion at the begal 1 site of SV40, so they are origin defective. They cannot replicate their DNA, and, consequently, they cannot -- neither can they express late viral genes efficiently.

Now, in fact, such a construct transformed cells better than wild type virus does or wild type viral DNA. If I could have the first slide, and I will try it out on the right.

This is a standard focus assay, similar to what Doctor Rundell was just describing, and these are normal cells that are confluent. When we transvect them with an origin defective construct that I mentioned, you see a number of large, well describe foci.

This is the same experiment using the same construct, which contains an SV40 origin and otherwise normal sequences, and you can see the number of foci is still apparent but much reduced in number.

This is if you use free viral DNA obtained from the -- supernatant, so it's not an artifact of the plasmid sequences.

This phenomenon is even more pronounced if you look for single cell cloning assays, such as growth in agar. So, we picked foci such as these and analyzed them further for their transform phenotype and growth patterns. And, I would point out that typically there's one to two integrated copies, there's no free viral DNA as you would expect, and the viral sequences tend to be quite stable in their integration site.

If I can have the slides on both sides now.

This summarizes the features. This is a growth pattern of normal cells. Normal human cells, as you may know, have a limited life span, and this is reflected here. You can passage them repeatedly and then they lose their proliferative capacity, and thus, become a model of cellular senescence. Introducing SV40 can result in the transformed phenotype as such that I just showed you, and, in fact, that alteration in the transformed phenotype growth in agar, growth in monolayer, focus formation, is dependent on small T as well as large T.

However, there's an additional phenotype which we can score for easily in human cells, and that's the extension of life span beyond that which is typically observed, i.e., the overcoming of senescence. And, that phenomenon, extension of life span beyond senescence, is independent of small T.

Can I have the next slide on the right?

In fact, if that period of extended life span, and that period of normal life span both show T phenotypes, if you shift them up at a time when they still are within their period of normal life span, they lose the transformed phenotype, and this is shown with a temperature sensitive T-antigen, TSA 58 transformed cells. At 35 degrees they grow to a high saturation density, if you shift them up to 39 degrees soon after isolation, in fact, they plateau, and you can passage them a few times and they show a low saturation density.

However, as you passage the cells progressively at 35 degrees, they lose the ability to then grow when shifted up to 39 degrees, and this is because they are now in that extended life span period and that is dependent on T-antigen.

If I can have the next slide on the right? No, keep the one on the left.

These cells subsequently die, and they die in a phenomenon called crisis, which has a large component of apoptosis in it, and may be due in part to the reappearance of senescence as well. This has not been extensively studied, other than in a phenomenological way.

In addition, one can get rare cells which come out of some populations in crisis, which now grow continuously, and it's typical that tumor cells grow continuously, but normal human diploid cells do not spontaneously become immortal. So, this is, again, a T-dependent phenomenon, in the sense that T promotes the frequency of immortalization from virtually zero to a detectable number, but I would emphasize it's a low number.

What we decided to do was to look for what is going on to analyze this a bit further, and the most obvious thing to do was to, again, look at the T-dependence, and that's done very simply by showing the next slide on the right. If you take such immortal cells, again, generated with a TSA 58 mutant of SV40 which is origin defective, and this is a growth curve at 35 degrees, when you shift them to 39 degrees, the open circle, they cease to grow and, in fact, die. So, immortal SV40 transformed cells, similarly, still require T function.

If I could have the next slide on the right and the left. I'm sorry, take it off on the right.

We decided to look at what T functions were there, and as you would expect based on the information you already know, there are TRB complexes in these cells at 35 degrees, and when you shift up to 39 degrees the TRB complexes dissociate. So, we do not -- so they are temperature dependent for proliferation, and also for TRB complex formation.

T binding to RB has also been shown to be important because when you take a mutant of SV40 T-antigen, not a temperature sensitive, but the well known mutant K1, it does not induce DNA synthesis in senescence cells.

There is another system of SV40 immortal cells generated by something other than a temperature sensitive mutant, there's one where it is dependent on dexamethasone. It's a T-antigen which is driven by the MMTVLTR, which was developed by Woodie Wright and Jerry Schay. And, in their system as well, K1 does not stimulate DNA synthesis.

However, TRB complexes, though necessary, are not sufficient for continued proliferation, as we've shown in some public studies.

If I could have the next slide on the left.

There are also T P53 complexes, as you would expect. Again, a mutant in SV40 which does not bind P53 was shown by Lynn & Simmons not to extend the life span of human fibroblasts, so again, binding of P53 seems to be an important thing, and again, T P53 complexes are necessary but not sufficient for extended -- for permanent proliferation.

And, if I can have the slide on the right. We can show this most simply that there are T P53 complexes that precipitate extracts from cells prepared at 35 degrees, you can see T P53 complexes by Western Blot, however, if you prepare extracts from cells at 39 degrees you do not -- there is T-antigen present at both temperatures, and this is a non-TS cell line, which does not show temperature dependent complexes, and, in fact, you can restore the growth properties, and the P53, and RB binding by reintroducing a wild type T-antigen which is shown here, so the properties are due to T-dependent processes which are being expressed or not expressed in these cell lines.

So, if I could have the next slide on the left, so we would conclude that T binding to RB and P53 are critical, but is the T binding to P53 working in inactivating it, and that seems to be the case, because, in fact, some people, not ourselves, have looked for mutations in SV40 transformed immortal cells and they have not found mutations in P53. This has not been thorough, it's not been done to exhaustion, but there are a couple of studies.

I think one has to keep in mind the fact that there may be P53 effects which are not being totally controlled, and that is the fact that O'Neill has recently, who is here and he can talk about that, has recently shown that there's an excess of P53 in immortal human fibroblasts than that that can be explained by binding to T, and also these experiments do not deal with the issue of transient changes in P53, nor whether all the P53 functions are affected, particularly, the P53 repression function.

So, if I can have both slides off and then the slide on the left, we come to the conclusion that there are both T-dependent factors, but there must be also T-independent factors, because T function is no different in the immortal cell and the pre-immortal cells, as far as we can tell. So, immortalization must require something in addition to T.

We've looked at a variety of phenotypes. What I'd like to do is take the last few minutes to talk about those which indicate that there's a growth suppressor in chromosome 6 which is important to the immortalization and, therefore, the transformation process.

Can I have the next slide here?

It is known that immortal cells and human cells may be immortal for a limited number of reasons. Pereira Smith and her husband, Jim Smith, Livvy Pereira-Smith and her husband, Jim Smith, did systematic crosses, cell fusion crosses among human tumors, and they defined four different complementation groups. In support of the complementation groups, it was found that, for example, complementation group -- well, two things, immortalization is recessive to limited life span, and so you could fuse to immortal cells and now have suppression of the immortal phenotype by cross correction, classical genetic complementation.

To support the idea that the B group was a specific group, they found that chromosome 4 would suppress group B cells, including HeLa cells, but not other groups. Similarly, there was a group C which could be suppressed by one, and there's a group D that could be suppressed by chromosome 7.

There have been no -- they did not do similar studies, and other labs have not done similar studies with SV40 human fibroblasts. We, in fact, have done such studies, and I'll show you that they, in fact, are suppressed by chromosome 6.

But, in fact, multiple SV40 transformed cells, both the -- minus constructs that we have described, and ones from other labs that they've studied, fall into this same A complementation group.

I point out that if our's -- if the SV40 transformed human cells are binding RB and P53 and suppressing their function, similarly, HPV, which is present in HeLa, and adenoviruses present in 293 are doing the same thing, so we are not talking about RB and P53 in this phenomenon, although, they are obviously important.

Not everyone agrees with the complementation groups, so that's why the other is in there, and not all SV40 transformants fall into the A group.

I'd like to now show you the evidence that, in fact, there is an immortalization gene or a growth suppressor, which is involved in immortalization by chromosome 6.

So, if I could have the next slide here, and the slide on here. We've done three types of studies. One is to look for a chromosome rearrangement in immortal cells, as compared to matched pre-immortal cells. We've generated multiple sets, and in one set which was generated with a temperature sensitive mutant grown at the permissive temperature, it had a quite good karyotype, even though it was transformed, in fact, it's a normal diploid karyotype except for the fact that it had only one copy of 16.

This cell line gave immortals at a high frequency, and so even when there was a mixed population of normal cells we could identify immortal cells within it, and they, similarly, had a single copy of 16 and a certain number of rearrangements.

We isolated a clonal isolate from that, and it had, again, the same rearrangements as its precursor, but had a few others.

We also isolated immortals independently from the same transformant, and it also had a single copy of 16 and a few rearrangements.

But, if you look closely, there's only one thing that they all share in common, they all have lost sequences, all the immortals have lost sequences, a long arm of chromosome 6.

And so, if I can have the next slide on the right, this generates the following model, that in the normal cells there are two copies of chromosome 6, while in the transformed, but not immortal, one of the copies undergoes a mutation of chromosome 6. I'm now going to call that gene SEN 6 for convenience. When we lose the -- and it's indicated here by the Xeroxing artifact as a mutant gene, when we lose the normal copy we now have no normal SEN 6 and only the mutant copy and the cell is now immortal. Similarly, we could lose the whole chromosome 6.

Now, this makes a very clear prediction. If we were to put back a wild type chromosome 6, we should now go back on the curve and, in fact, therefore, suppress growth.

And so, in the next slide on the right, we, in fact, introduced chromosome 6 using the microcell mediated chromosome transfer technique, we isolated a limited number of colonies and two different immortals from our cell line, independent immortals in our cell line, all the colonies isolated failed to grow progressively.

When we put 6 into another tumor cell, transformed cell line with non-SV40 it grows well, and chromomes 2, 8 and 9 changes do not suppress growth. So, there seems to be a growth suppressor on chromosome 6.

Now, if this was true, it shouldn't be true just of our cell lines, so, in fact, we got, if I could have the next slide on the right, we got SV40 immortal cell lines from a number of other laboratories, these are our's, these were generated at Los Alamos by Paul Kramer, and this was generated by Jerry Schay in a third cell line. They are all human diploid fibroblasts.

What I've done here, if I can have the slide on the left, is to analyze these different cell lines using a combination of molecular genetic techniques, dinucleotide polymorphisms and CA repeat, RFLP analysis of fluorescent in situ hybridization.

Let me give you a few examples. AR5, the cell line that I showed you in the previous slide, when scored for these polymorphic markers shows loss of heterozygosity, namely, an open circle.

Halnea, which is a relative, again, loss of heterozygosity but retains this marker.

An example here, you lose only these two distal markers, retain the others.

If you total it up, 12 of the 17 have lost sequences in chromosome six, five of the 17 show no losses, and, therefore, are only closed circles within the markers that we have checked for.

If I could have the next slide on the right. If you now tabulate this data on a figure, solid line is loss of sequences, you can see that a lot of these cell lines have lost large fragments, but some have lost less, the closed part, and, in fact, if you define a minimal region of common loss it comes down to 6Q26 to 27.

We've now analyzed that further, and we've come down to a one megabase region, which is selectively deleted in SV40 immortal cells, and we are now attempting to clone the sequence.

So, could I have the slide -- second slide on the right, skip this, and the one on the left, I would say that SV40 transformation should really be thought of transformation and immortalization. And, it represents both T dependence and SV40 dependent direct effects, which would involve RB and P53, and, perhaps, others, but also bear in mind that there are SV40 independent effects which are not dependent on continuous function of T-antigen, and one of those would be the growth suppressor that we've defined in chromosome 6 and, presumably, possibly other growth suppressors which may exist.

Another area that one must bear in mind is the fact that immortal cells must deal with the telomere problem. Normal cells have -- do not express telomeres typically, and under those conditions the ends of their chromosomes, containing telemeric sequences, progressively shorten. Introduction of T-antigen does not correct this defect.

An immortal cell, in order to stay immortal, must, in fact, stabilize its telomores, and so there must be a mechanism in that as well, it's not purely T dependent. Similarly, overcoming crisis and, perhaps, other things involve changes in susceptibility to apoptosis, as was already mentioned two talks before, and I would also point out that there are mRNA expression differences that we found in screening cDNA libraries.

So, I would say that tumor formation by SV40, in fact, is a quite complex phenomenon.

Thank you for your attention.

(Applause.)

DOCTOR MAHY: Thank you very much, Doctor Ozer.

A couple of quick questions. Yes, please.

AUDIENCE: Harvey, does 6Q26 show a loss of heterozygosity in any human tumors?

DOCTOR OZER: Yes, actually, 6 is rearranged often in human tumors, but there are three human tumors which show rearrangements in this region. One is Berkitt lymphomas, another is ovarian cancer, and a third is mammary tumors, mammary cancer of certain types.

It's not clear that it's the exact same region, but it's in the neighborhood.

DOCTOR BUTEL: Harvey, how are you distinguishing between immortalization and transformation in your system?

DOCTOR OZER: Okay. I would define transformation as changes in the phenotype, such as focus formation growth in agar, which appear soon after introduction of T-antigen, and which can be seen as persistent growth for a limited period of time. A limited period of time can be six months, as many generations we are talking about, it can be as many as 80 or 90 generations.

However, progression of the same phenotypes, but, most important, the growth phenotype, after 100 population doublings is what I would define as immortal, the immortal phenotype, added onto the transformed phenotype.

DOCTOR MAHY: Doctor Weiss?

DOCTOR WEISS: Robin Weiss. Harvey, just to follow up on the first question, has this LLH in chromosome 6 been looked for in, say, mesothelioma for the brain tumors that we've been discussing be more associated possibly with SV40, is it worth looking for?

DOCTOR OZER: So far as I know, there's been no systematic study of it in mesotheliomas or osteosarcomas. We are looking in cell lines derived from mesotheliomas in collaboration with Harvey Pass, that there isn't sufficient data to comment on.

DOCTOR MAHY: Thank you very much.

Okay. Now, we have the final talk today from James Cook, the final talk of this morning's session at least, on experimental tumor induction in SV40 transformed cells, from the National Jewish Center in Denver.

Return to Agenda

DOCTOR COOK: Thanks. I'd like to thank Andy Lewis for organizing such a really interesting meeting and for his collegiality over the years.

What I'm going to talk about takes off a bit from what Harvey and Kathy have talked about, and that is to assume that cells now are transformed and try to understand what it is about these cells that determines their ability to form tumors in animals.

And, I'll start by giving a little overview of the prototype model, that is, the SV40 transformed hamster cell, and then talk about some studies that Lauren Sumperac and I have been doing together on SV40 transformed rat cells. And, I'll try to convince you that rat cells may be somewhat more like human cells than human cells are like hamster cells.

May I have the first slide, please?

The first thing I'd like to talk about is SV40 transformed hamster cells and how they behave. As Michele Carbone told you this morning, he actually reviewed quite well the discovery of SV40 tumor induction in hamsters by newborn inoculation. The next step was that SV40 transformed hamster cells were shown to be quite efficient, relatively efficient in inducing tumors in hamsters, with an efficiency that requires maybe tens of thousands of cells.

These cells were useful for a couple of other reasons. I think, at least to me, two important reasons, one is that comparisons of SV40 transformed hamster cells that are quite tumorigenic in the immunocompetent normal animal, which is a contrast to cells from other species, allowed the ability to dispel a couple of myths. One is that hamsters are somehow immunologically incompetent to deal with SV40 transformation. There are no measures of anything done in hamsters that have suggested that there's anything wrong with them immunologically, other than the fact that they have a fairly restricted number of MHC molecule diversities.

The other myth is that SV40 transformed hamster cells just aren't immunogenic, and that's why they sneak by and make tumors in hamsters. It turns out when studies have been done to actually quantitate this, and these were studies done when I was a post-doctoral fellow in Andy's lab, that SV40 transformed hamster cells are actually quite immunogenic in tumor protection assays.

And then the last thing is that there is something distinct about SV40 transformed hamster cells, when it comes to their ability to survive injuries by the punitive components of the early parts of the cellular immune response.

And, on the next slide I just show you a very simple experiment, where SV40 transformed hamster and mouse cells were overlaid on monolayers of activated macro phages, and this was done while I was a post-doc in John Hibb's laboratory. This first row of cells was overlaid with SV40 transformed mouse cells called TCMK, and the second row of cells was overlaid with SV40 transformed hamster cells, a cell line called SV40 HE1.

In the first two columns, the wells contained confluent monolayers of macro phages that were highly activated by previous exposure of the donors to BCG, and then the question was asked, what would happen to the cells when they were allowed to co-cultivate with these macro phages or by themselves in the right-sided wells. And, as you can see, the SV40 transformed hamster cells survived this encounter with activated macro phages, and actually eventually overgrew them. This is taken several days after the assay was done, whereas, the SV40 transformed mouse cells were largely destroyed.

The other interesting thing that's hard to see is that normal fibroblasts that came from the peritoneal exudates used to prepare these macro phages that were slightly contaminated with fibroblasts because of the inflammatory response, actually grow quite well in these monolayers and suggest that SV40 transformed hamster cells are a lot more like normal cells than they are like SV40 transformed mouse cells, that is, they are resistant to this destructive effect.

So, we have interpreted this to say that SV40 transformed cells are inherently resistant to many of the types of injuries they might encounter in a normal immune response in vitro, and this is not true of cells from other species that are transformed by SV40, including human cells.

Next slide, please.

So, what I think this might mean about other cells, now stepping aside from hamster cells and looking at cells from other species, is that there may be a stepwise process required to get to the point that SV40 transformation can result in tumorigenicity, experimental or otherwise. SV40 transforms hamster cells from many species, as you've already heard. It's a fairly long list. Tumor production by SV40 transformed hamster cells -- I'm sorry, by SV40 transformed cells in species other than hamster, almost always requires immunosuppression, so the cells have to be put into nude mice, or animals that in the old days were thymectomized or irradiated, to get tumor formation at any frequency.

But, the other interesting thing is that once these cells are established, once these tumors form, and you reestablish these cells in tissue culture, they quite often acquire a quite different tumor-inducing capacity that is much more like SV40 transformed hamster cells. So, it appears that this initial tumor formation under the cover of immunosuppression allows these cells to now become tumorigenic in immunocompetent animals, whereas before they never were.

Next slide, please.

So, a way to think of this in sort of a graphic way is that maybe transformation of hamster cells by SV40 results in the direct acquisition of primary tumorigenicity in the immunocompetent host, whereas transformation by cells from other species may go through some step that Harvey and others have described, that is, immortalization or transformation, or whatever you will, but that is the ability to grow in vitro, but the inability to efficiently grow in the in vivo environment unless immunosuppression is put n the middle, and that some secondary event or events, probably a series of events, occurs during this immortalization process that's required for these cells to now acquire a high-level tumorigenicity in the immunocompetent host.

Next slide, please.

So, now I'd like to describe the model that Lauren and I are working on, and these are all preliminary observations. For many years, Lauren has been studying the sequences of SV40 that are minimally sufficient to immortalize rat cells. All of these studies were done with the entire SV40 sequence, encoding both large and small T-antigens, and what we'll talk about is the conversion of cells from normal cells to what I'll call operationally immortalized cells, that is, cells that can grow in vitro, and then what may occur as a stepwise progression of cells from that immortalized state onto a fully tumorigenic state, again, analogous to what SV40 transformation of hamster cells can do in the first place, and then consider the hypothesis that this may require from some acquisition of other cellular mutations that are required on top of what SV40 can do to create this phenotype.

Next slide, please.

This is the model. Lauren immortalized primary Fischer rat embryo fibroblasts two different ways, with two different gene-containing constructs. One is an SV40 plasma that was simply transvected into these cells. The cells were selected for growth in soft agar, and then a cell line called 10-1 was created.

The second cell line created some months later was done by infection with a retro virus vector containing SV40, and these cells were selected for growth in geneticin, because the construct contained the neogene, but weren't selected for growth in soft agar.

Next slide, please.

Then, what we did was to characterize these cells for their tumor forming capacity, realizing that it probably would be weak. These cells were inoculated at high dose, 10⁷ cells, and in some experiments L titrations were done to determine minimal sufficient numbers of cells to cause tumors subcutaneously into either adult nude mice, and by adult I mean about two months old, or into Fischer weanling rats, and these rats were usually six to eight weeks old.

Next slide, please.

And, these are the data. When the 10-1 cell line was inoculated into nude mice, this is four different experiments with about three animals each, about 75 percent of the animals cumulatively developed subcutaneous tumors, but it required almost two and a half months for this to occur. These was usually no tumor nodule apparent, and around two and a half months two millimeter or so nodules could be seen under the fairly transparent skin of the nude mouse. And, in these experiments, the weanling Fischer rats were also inoculated simultaneously and got no tumors.

The second cell line gave exactly the same phenotype, and, that is, tumors occurred fairly efficiently at this high cell inoculum, but required a fairly long period of time to begin to appear, and no tumors were observed in two experiments in which weanling rats were challenged.

The next slide, please.

So then, what we wanted to do was to ask whether this stepwise process could be reproduced in a rat model, and we reestablished these cells in tissue culture from the tumor, and then called these nude mouse passage one of either of the two cell lines, and then characterized their tumor-inducing capacity after they'd been established. This is based on models that Andy and I did in an adeno 2 transformed system, and others have done with SV40 transformed human and rodent cells, to show that some increased tumorigenicity can be acquired after a period of passage in vivo.

Next slide, please.

So, this shows the same data table set as was shown before, except in this case we are comparing the nude mouse passage cells from the 10-1 cell line. The parental cell data is shown here, these are the control experiments for the experiments done before. So again, it took over two months for tumor formation that occurred relatively efficiently, no tumors formed in the Fischer rats. When the nude mouse passage number one was tested, these two experiments, tumors formed in both cases, but they have formed quite rapidly. So, we had two to five millimeter tumors within ten days after the inoculation. In the interval period, there was nothing detectable after the fluid was reabsorbed.

The striking thing was that these cells also formed tumors quite rapidly in Fischer rats, in fact, more rapidly than we could have imagined based on SV40 transformed cell tumor induction in hamsters, and these were just masses now, they weren't histopathologically looked at at three days, but this is when the mass started, and from that point progressed.

The same thing was true with the second nude mouse passage, which was simply this tumor reestablished in culture and put back into nude mouse. Again, the tumor masses appeared quite rapidly, and these cells induced tumors in all weanling rats that were challenged. So, they had acquired an ability to make tumors now in the immunocompetent host, that they never had had before.

The next slide, please.

The other thing that was done was to look at the efficiency of tumor formation by these cells, now that they could actually form tumors in nude mice with some rate that we could measure. We could quantitate this by titrating the numbers of cells required for an endpoint of tumor formation, and what we looked at was the endpoint required for 50 percent tumor formation in nude mice.

As you can see, there was quite a striking difference between the parental cell, the SV40 immortalized Fischer rat fibroblast, and the second nude mouse passage, requiring only about 30 cells to form tumors at a 50 percent endpoint, in contrast to hundreds of thousands of cells of the parental cell, and again, the latency period is factored in here as well.

Next slide, please.

So, what I want to do now is to break from this and say that we don't understand the mechanism by which these cells have acquired this increased rate of tumor formation and efficiency of tumor formation, but since it's so analogous to what goes on in human systems we think that this may be a model in which we could study both ends of the spectrum without really knowing what went on in the middle, to ask what the key changes are that these cells might have to experience before they are able to form tumors in immunocompetent animals, like the weanling Fischer rats.

Now, it's like a lot of things, we do what we can do, and we've been studying how E1A affects, the E1A gene of adenovirus, how it affects the susceptibility of rodent cells to killer cell injury. And so, since we were able to do that in the laboratory and had some experimental models, we wanted to ask a fairly easy question, and, that is, what happens to these SV40 transformed cells after their passage through nude mice and acquire this increased tumor-inducing capacity, when they are put into contact in a very experimental in vitro model with two kinds of killer cell circumstances.

Cytotoxic T-lymphocytes or natural killer cells have at their use two different ways to kill their targets. One way is called degranulation dependent killing, and it's calcium dependent. It requires exocytosis of the killer cell granules that are transmitted to the target cells and lead to DNA degradation and other kinds of injury, and this is called porphyrin granzyme killing.

The other mechanism does not require degranulation, can occur in the absence of calcium, but does require activation of the cells to express on their surface Fas-ligand, that then interacts with its cognate antigen on the cell surface Fas-antigen, and through this triggering kill cells in a TNF receptor-like mechanism, since Fas-antigen is in the TNF super family.

So, it's possible to measure these two things independently, again, in fairly contrived in vitro models, but what we wanted to ask was, did these cells have any differences in their susceptibility to these two kinds of killer cell injury in these in vitro assays that might correlate with their acquisition of tumor-inducing capacity in an immunocompetent rat.

Next slide, please.

What we found was that the cells that had been passaged either one time or two times in nude mouse, had lost a good bit of susceptibility to degranulation-dependent killing in this assay.

Now, these are preliminary data and a number of other things have to be looked at in terms of cell surface expression of a variety of ligands, but it appears that one of the phenotypes that has occurred, at least, on both ends of the spectrum, is the loss of this ability to be killed by porphyrin granzyme that the killer cells might use. And, this is true, as I say, for both passages after the first passage.

The next slide, please.

Looking at Fas-ligand dependent killing, when we use activated cells now expressing Fas-ligand in the absence of calcium, so they couldn't degranulate and use that other mechanism, what we found, that only those cells that have been passaged twice through nude mice had lost susceptibility to this for of killing, so it appears that there's a dissociation here in this kind of injury. It's not clear what this means. This has been reproduced, actually both of these observations have been reproduced with the RW-9 cells, so it looks like it happens with both of these clones. Again, we don't know the mechanism by which this occurs. We do know that Fas-antigen is expressed equally on the target cells before and after passage, so it's not simply the fact that they've lost the ligand that the receptor needs to interact with.

The next slide, please.

So, what we think this says is that, not understanding the mechanisms by which it happens, that SV40 can immortalize cells through a first step, and that other things are required, perhaps, in a stepwise mechanism, to allow cells to acquire the tumorigenicity that SV40 transformation of hamster cells can create automatically.

What these secondary events may be that occur along the way are unclear, but it is clear from reviewing the literature that a number of things can do this, that is, increase the tumorigenicity of SV40 immortalized cells. There are things as simple as serial tissue culture passage that's associated with accumulation of mutations in the cells, irradiation of the cells, chemical exposure to mutagens, other kinds of injury that can lead to probably serial genetic mutations that occur with the permission of SV40 that has created the immortalized cell to let this occur.

Now, with the talk by Doctor Furth, I suppose it's possible that SV40 could set up this circumstance, and then either become minimally expressed or, perhaps, even non-expressed, and these other mutations could take over and create this circumstance. Our evidence, at least in terms of the susceptibility of cells in the E1A model is that the early genes must continue to be expressed for this susceptibility to continue. We know in these cells that they still express SV40 T at high levels. They still express wild type P53 at high levels, so we don't think that any other trivial explanations explain this difference in susceptibility of cells.

Next slide, please.

So, in summary, we think that, as everybody else has said and I think knows, is that immortalization is the first step that has to happen. These cells must become immortalized before all these other events have the chance to happen.

Perhaps, for cells, for all other types of cells, or most other types of specie cells other than hamster cells, something else must happen that might be a mutational event, that allows SV40 transformed cells to acquire tumorigenicity that we can measure experimentally in vivo, and that one class of these other types of events, at least the phenotype that is shown, is the loss of susceptibility to a variety of injuries that these cells may encounter when they are confronted by the host cellular immune response.

Clearly, there are other things that may also occur, in terms of growth factor receptors, other kinds of cell surface changes, structural protein changes that may occur in cells during this transition, and the question eventually will be, what's the minimal sufficient thing that must occur for a cell to become highly tumorigenic. The possible relevance for this meeting is that if rat cells are like human cells, in the sense that the transformation occurs but doesn't usually convey to those cells high- level susceptibility, what are these other changes that must occur in human cells to allow them to now form tumors, as they may do in the context of mesotheliomas and osteosarcomas, if those turn out to be truly SV40 T-antigen induced neoplasms.

Thank you.

(Applause.)

DOCTOR MAHY: Thank you very much, indeed, Doctor Cook.

A question?

AUDIENCE: We have done the same experiment in the mouse system, where we have transformed the primary mouse embryo fibroblasts, -- include the nude mouse.

DOCTOR COOK: I don't think your microphone is working, I can't hear you very well.

AUDIENCE: Okay. We have done the same experiment in the mouse system. You take the mouse embryo fibroblast, transform them with full SV40, and then they are transplanted right away in nude mice.

No matter how many times you pass through nude mice, they are still non-transplantable in the immunocompetent mice. So, I think there must be something special about the rat system that does not apply to the mouse system.

And, my question is, that are your cells still expressing -- molecules to the same extent as the ones that went into the nude mice before?

DOCTOR COOK: We've only looked at one of the two clones, but the level of class one expression is comparable on the nude mouse passage cells with the 10-1s.

In the mouse model, as you talk about, others have found that, it's not clear, but passage of cells like MKSATU-5, which is a tumor-derived cell line, but didn't induce tumors very efficiently, could create a cell line like MKSAASC that acquire greatly induced tumor-inducing capacity.

That's not directly analogous to passage in nude mice, but there does appear to be, in some mouse cells, the ability to acquire increased tumorigenicity, for example, simply from passage for 50 times in tissue culture, and I know that probably isn't the case with your cells, but it looks like it may not be quite that mouse cells can't do it at all, and only rat cells can, but I'm sure there must be some explanation for the difference.

DOCTOR MAHY: Okay, thank you very much, indeed.

I thank all the speakers of this session. We are going to meet again at 1:30 after lunch in the cafeteria.

(Whereupon, the meeting was recessed at 12:42 p.m., to reconvene again at 1:30 p.m., this same day.)

AFTERNOON SESSION

(1:45 p.m.)

DOCTOR KELLY: Good afternoon. I'd like to welcome you all to the last scientific session before we settle all outstanding issues at the end of the day.

I think we've got a very interesting session. We are going to continue with mechanisms of SV40 oncogenesis, and the session contains some new information about SV40 T-antigen, as well as other material.

In the interest of generating some discussion, I've administratively cut every speaker's time by ten percent, and that's non-negotiable. So, sorry, Jim.

So, with no further ado, let's begin the first talk by Jim DeCaprio, from Dana-Farber, and he's going to be talking about SV40 DNA replication and transformation, requiring the DnaJ chaperone domain of large T-antigen.

Return to Agenda

DOCTOR DeCAPRIO: Thank you very much, Doctor Kelly.

First, I'd like to just acknowledge my post-docs and grad students that did the work, as well as the collaboration with Doctor Tom Roberts and his grad student, Kathryn Roberts. Most of the work you'll see has been performed by Hilde Stubdal, a grad student in my lab, as well as Juan Zalvide, a post-doc.

SV40 large T-antigen alone is sufficient to transform a variety of normal cell types, and this protein has been the focus of many, many investigators in the mechanisms for transformation.

Doctor Fanning and others have pointed out that there are several domains within large T-antigen that are required for this transforming process, including C terminal domain, combined to the P53 tumor suppressor, as well as this yellow block sequence coded by the residues leucine something, cysteine something, and glutamate, the LXCXE domain combined to the retinoblastoma tumor suppressor gene, as well as two other members of the RB family, and a third and terminal domain that's required for transformation that has not been previously well characterized, but is clearly involved in transformation.

T-antigen is thought to transform cells by binding to tumor suppressor genes, such as P53 and RB, and inactivate their tumor suppression functions, and thereby allowing growth of cells in various transformed phenotypes and various transformation assays.

Well, it wasn't clear whether the 107 and 130 genes, which look very similar to RB, whether they, in fact, were also targets of the LXCXE domain, and, that is, did T-antigen have to bind to these two proteins, and did it have to inactivate their growth suppression properties, in order to transform cells, or was RB the only relevant member of this family that T-antigen had to inactivate.

So, what we did was, we asked the question whether a large T-antigen could, or a mutant of T-antigen that was mutated in the LXCXE binding domain that could no longer bind to RB or 107 and 130, could that transform cells that had no RB. So, we prepared mouse embryo fibroblasts from cells that had the RB gene knocked out, RB minus cells, ad we asked whether wild type T could transform those cells, and whether the LXCXE mutant could transform those cells.

And, in a series of experiments, Juan Zalvide showed, in fact, that wild type T could transform wild type cells. It could also transform RB minus cells. These are colonies grown in soft agar. Wild type T could transform RB minus, but the mutant T-antigen in a LXCXE domain that could not bind to RB 107 or 130 could not transform either wild type cells or could not transform cells that had no RB.

And, this experiment suggested that 107 and 130 were or could be targets of T-antigen, and that you needed that domain of T-antigen for transformation.

Well, then we started looking at the interaction of P130 with large T-antigen, as well as 107, and this is a Western Blot for the P130 RB2 protein that we've heard about yesterday from Doctor Giordano, and here you can see the panel over here in lane five, here is the P130 expression and you can see, actually, a series of bands extending over several kilodalton, and in the cells that have a mutant T-antigen that can now bind to this P130 you see there are several phosphorylated bands, but in a cell line that contains wild type T-antigen, there actually is only a fast migrating form of P130, you don't see the upper phosphorylated forms, and with treated phosphatates you can see actually in a mutant T line P130 collapses to this fast migrating form. This actually is P107 down here, we won't discuss that right now.

This effect of T-antigen reducing the levels of phosphorylated P130 we have also observed in monkey cells here, CV1P cells showing phosphorylated P130 compared to COS cells, which contain large T-antigen and show actually no phosphorylated P130 and actually very little P130, and this effect of T-antigen on none of the levels of phosphorylated RB, phosphorylated P130, but also the total amount of P130 is reduced by T-antigen. We'll come back to that topic.

This slide shows that wild type T can reduce the levels of phosphorylated P130, and it suggests, in fact, that you need the LXCXE binding domain to get this effect.

Well, we did a series of experiments then looking to see if the LXCXE domain was required and if it was sufficient for this effect, and what we found was that you needed both the LXCXE domain, as well as the N-terminus domain, so here is just a schematic showing that we had a wild type T combined with P130 and reduced its phosphorylated state, a mutant T that could not bind to it could not affect the phosphorylation.

Here is a T-antigen that was truncated at the N-terminus originally made by Ellen Fanning, and we adapted it as a cDNA, this combined to P130 but it leaves the phosphorylated P130 alone. There is phosphorylated P130 in cells transformed with this T-antigen.

And, here's a T-antigen that truncates off the C-terminus. This combined with P130 and reduced its levels of phosphorylation.

That observation caused us to look at this N-terminus, and we looked at it by performing a search against the gene database, and the bottom line here is actually part of the N-terminus of SV40 large T-antigen compared to polyoma large T-antigen, and what we noticed is, in fact, there were several residues in the N-terminus that are conserved with a class of proteins known as DnaJ. DnaJ is found in E. coli here on the top line. There is more than 20 varieties of DnaJ in the -- genome, and here are two human DnaJs that had been cloned, although the exact function of these two are not known. In particular, the HPDK or HPD residues are absolutely conserved among all DnaJ homologs, as well as all popova viral T-antigens from all the species that we heard about, and this is actually conserved with both the large, small T-antigens as well as middle T-antigen of polyoma, and there are several other residues, in particular, this Leucine at position 17 in T-antigen.

But, what is a DnaJ? Here is a cartoon borrowed from Hartl, published in Nature this past year. DnaJ typically contains a 70 residue structure known as the J domain. This is the region that's shared with T-antigen, that is usually at the N-terminus but can be at different positions in different homologs, and a variable C-terminal domain that is thought to be involved in protein folded and chaperone functions.

The J domain interacts with DnaK, also known as Hsp 70, another heat shock protein. Heat shock protein 70, Hsp 70, contains ATPase, and it is thought that the interaction of DnaJ with Hsp 70 activates the ATPase of Hsp 70.

Here's a cartoon of the general schematic of what DnaJ does when it interacts with Hsp 70, so here is a DnaJ homolog. Here is the J domain sort of in this loop, with the HPDK forming a loop, the C-terminal domain which is highly variable interacted with some protein substrate here, representing an unfolded state, could be needs to be translocated, could be denatured, a variety of different things. The J domain attracts Hsp 70 loaded with ATP, and the J domain promotes the hydrolysis of ATP to ADP, and through some magical process now will refold up substrate or translocate a substrate, a variety of different activities dependent on a specific DnaJ.

What we would argue now in this is that T-antigen is a DnaJ homolog, here is its J domain, and here, for example, is the LXCXE domain binding to P130.

Well, here's a Western Blot actually looking at some of those specific residues that were conserved with the DnaJ, so here is P130 in the first lane, co-expressed with wild type T, showing a reduction in the levels of phosphorylated P130, and here H42Q in lane four, and D44N in a highly conserved HPD domain. These T-antigens combine with P130 but do not affect its phosphorylation state.

To test this hypothesis more directly, we constructed chimeric proteins. Here a substitute in the N-terminus with T-antigen with the J domains from two human DnaJ genes that had been cloned previously, HSJ1 and DnaJ2. We fused them with the C-terminal part of T-antigen, and also made point mutations in the J domain to test with specificity the J activity.

And, this Western Blot here of P130 co-expressing with T, you can see in lane two P130 with wild type T-antigen or with the DnaJ2 chimeras in lane four or in lane six the HSJ1 T chimeras, showing again the reduction of levels of phosphorylated P130, and compare that to the point mutations in lanes five and seven, the point mutations of the two chimeric proteins do not affect the levels of phosphorylated P130.

So, T-antigen is having an effect on P130, and it seems to be mediated through its J domain. One of the studies -- it looked like that there was always less P130 around when wild type T was present, and we tested this by performing a pulse chase experiment, and in the dotted line is actually the half life of P130 alone, about a five hour half life. However, in the presence of T-antigen a half life is less than an hour. In blue is a T-antigen co-expression with the T-antigen that can't bind to it, and in green, actually, is the J domain mutation of T-antigen that combined to P130 and actually may lead to a bit of stabilization of that P130.

So, we think that T-antigen not only promotes the loss of phosphorylated P130, but actually specific degradation of P130, through its activity as a DnaJ homolog.

Well, we wanted to test whether this J domain was required for transformation, did it participate in the transforming activities of T-antigen, and was the N-terminal transforming domain, did that represent a J domain. So, we established mouse embryo fibroblasts stably that expressed T-antigen in lane one, a mutant T-antigen that couldn't bind to any of these proteins, K1, two different J domains, H42Q and D44, and a chimeric T-antigen in lane five and a mutant chimeric in lane six, and you can see, again, in stable lines, and this is the endogenous P130, that, again, the same effect in a phosphorylated P130 as we saw in the transient assays is present in these stable lines here. However, the T-antigens that either have a mutant LXCXE or a mutant N-terminal transforming domain leave P130 alone.

You can see in a P107 panel that, in fact, the phosphorylation of P107 is also affected in a manner similar to 130, and the mutant T-antigens do not affect the phosphorylation status of P107.

In contrast, we don't think that RB phosphorylation is affected by the presence of T-antigen, that RB phosphorylation is not affected in any way, and that RB goes about its normal business of cell cycle dependent phosphorylation.

This bottom panel is just showing the expression of the various T-antigens, and all of these T-antigens have a very stable half life of more than 18 hours, very similar to the wild type T-antigen.

We did a series of different types of transformation assays. In this assay here, we took wild type mouse embryo fibroblasts that were established by expression of these various T-antigens, and we plated them at low cell density, fed them every two days with complete media containing ten percent serum and asked how dense could they -- how many cells could grow on a particular plate.

And here, wild type T, shown in red, and also a chimeric T-antigen, continues to grow to very, very high cell densities here at ten days, where the LXCXE mutant in blue, or the various J domain mutants in green, reach a certain level of confluency and then do not continue to grow. They will become density arrested at a certain density in this particular assay.

To examine whether 107 and 130 were specifically targeted by the J domain, we established cell lines that were made from mouse embryo fibroblasts that were genetically deleted for both 107 and 130. We obtained these from Tyler Jacks and Nick Dyson, who constructed these various chimeric -- various homozygous deleted mice. And, what we found here is that wild type T-antigen, that chimeric T-antigen grew very well, and the three different J domain chimeras grew as well as wild type T antigen in this assay. In contrast, the LXCXE mutant that could not bind RB, could not bind to 107, it could not bind to 130, here becomes density arrested and remains contact inhibited in this assay.

What this result suggests is that T-antigen is eliminating the growth suppression functions of P107 and P130, and you need both the LXCXE and the J domain to do that. In the absence of an in-tact J domain, if you've lost the genes for 107 and 130, then you don't need that J domain, and now T-antigen can transform these cells in this particular assay.

We also performed this experiment in 107, only knock out the 130, in the RB only knock out, and the J domains were more likely LXCXE mutants in all those assays. So, it looked like the J domains needed to target both 107 and 130, in order to give you complete transformation.

So, what we would propose then is that this N-terminal transformed domain is a J domain, that the function of the N-terminus to transform actually requires an interaction with HSC 70. In work that we haven't shown that Kathy Campbell performed, actually, is that the J domain actually is interactive with HSC 70. This observation was originally demonstrated by Doctor Butel several years ago, showing that T-antigen could bind to HSC 70 in a specific manner, and Kathy Campbell's extended that by showing specifically that the J domain or the chimerics can bind to HSC 70 but not these point mutations in specific J domain activities, and that the N-terminus cooperates with the LXCXE in order to reduce the growth suppressing properties of 107 and 130.

In addition, Kathy Campbell, with Tom Roberts, looked at the replication activities of SV40 large T-antigen in a plasma replication assay that contained the SV40 origin. And, what she was, she compared the ability of wild type T-antigen to replicate this plasmid, and shown here are about 6600 counts, and the point mutation in this J domain actually reduced in about five-fold the levels of replication activity.

Here is testing, actually, the chimeric T-antigen DnaJ2 or HSJ1, showing both of those can replicate this plasma DNA more efficiently than the mutant T-antigen, and the point mutations in the J domain and the chimerics are also defective or greatly reduced now several fold compared to the chimeric proteins here.

The J domain of T-antigen is probably not necessarily required for in vitro replication with work done by Jim Pipas several years ago, but at least in this sort of co-transvection assay it appears to cooperate and contribute to the replication activities of T-antigen here.

I'd like to stop and thank you very much, the organizers, for the talk.

(Applause.)

DOCTOR KELLY: We have some time for some questions. Mike?

AUDIENCE: Could small T, wild type small T complement large T J mutants?

DOCTOR DeCAPRIO: So, small T does have this J domain, and shares it with the large T-antigen, but we have not looked at whether small T in trans could complement this transformation activity.

We have started to do that type of work, but we don't know. All those were done in cDNA, so we've eliminated the small T problem, but now we're going to have to go back and see if it works in trans.

We tried to look at the effect on P130 phosphorylation in trans, and it's possible that two different T-antigens, with two different activities, can work with that, but it's not the same as asking about a small T-antigen.

DOCTOR KELLY: Mike?

AUDIENCE: Jim, do you know if these J domain mutants oligomerise normally?

DOCTOR DeCAPRIO: We do know that they can form multimers with various T-antigens, but I don't know how efficient it is.

So, for example, the K1 mutant, which runs a lot faster than the other mutants, I could distinguish that from any J domain mutants, and that can oligomerise with K1. So, it can, I don't know how efficient it is.

Of course, you know, Ellen Fanning showed, actually, her half T, or 82 to the end, it could oligomerise, but it was inefficient. I don't know how efficient this is.

DOCTOR KELLY: Can I ask, so what do you think the role of the J domain is in reducing phosphorylation of P107, do you think it's prying off kinase or something like that?

DOCTOR DeCAPRIO: I think that what happens is that when T-antigen binds to P130, that -- let me step back -- when P130 gets -- P130 normally gets phosphorylated in a cell cycle dependent manner, and that if T-antigen is stuck to it, having a J domain there, when that phosphorylation occurs that actually targets it for degradation.

DOCTOR KELLY: Okay.

DOCTOR DeCAPRIO: So, I think that some of the differences between P130, P107 and RB can be explained then by that. We know that RB phosphorylation is not affected by T, and I think the reason why is that only unphosphoralated RB binds to T, so when RB gets phosphorylated it falls off and has protected itself from this J domain activity of T.

P107 is not phosphorylated during G1, it only becomes phosphorylated in S phase, so only the S phase fraction of 107 is affected by T. However, P130 is always phosphorylated, there's more phosphates maybe in S phase than in G1, but it is heavily phosphorylated in G1, and whenever T-antigen binds P130 is phosphorylated, and I think that's what triggers this degradation process.

So, we are trying to make the mutants now in P130 to address that.

DOCTOR KELLY: Phosphorylation site immune. DOCTOR DeCAPRIO: Phosphorylation site immune, to see if we can eliminate -- make it stable now and see if it's dominant over T-antigen transformation.

None of these are required for replication. The K1 can replicate, so that doesn't seem to be -- that J domain activity is probably a different J domain activity in replication.

DOCTOR KELLY: Ellen?

DOCTOR FANNING: Just a quick question. The immunoacid 17 seems to be conserved in these J sequences as well. Have you done anything to check whether that one behaves the same way, because at least one mutant with a mutation at 17 is unstable.

DOCTOR DeCAPRIO: Yes. L17K, we actually made that point mutation. That has a half life of 18 hours in our transient pulse chase, and that is defective in affecting P130 phosphorylation. We did not test actually that in transformation, but on the transient assay, looking at P130, it was defective. So, all the ones I showed you were all stable.

This J domain probably forms a very highly ordered structure, and it has several helices, and affecting the helix structure of it I think really promotes its own degradation.

DOCTOR KELLY: Okay.

If there are no other questions, we'll move on, and the next talk will be given by Jim Pipas, from the University of Pittsburgh, and he'll give us some complementing data on the DnaJ domain of large T-antigen.

Return to Agenda

DOCTOR PIPAS: Thanks. This slide is here to remind me to thank the organizers for convening such a nice meeting, focusing on the biology of SV40, and it's also to remind me to thank the pioneers in this field for providing us with such a powerful system for probing site or functions such as tumorigenicity, and it also reminds me to tell you that we can look at this virus in two different contexts, one illustrated by over here, where you see plaques. This is a productive infection, plaques of SV40 on a monolayer of BSC monkey cells, or in this context, where we are looking at transformed foci of multilayered cells on a monolayer of REF 52 cells. And, it's this latter thing I'm going to focus on for the rest of this -- latter transformation phenomenon I want to focus on for the rest of the talk.

And, here's the issues I think the field is trying to get at. We want to know what viral functions contribute to SV40 induced tumorigenesis. We want to know what the site or targets of each of those activities are, what T-antigen does to the target, and finally, how all that adds up to the transformed phenotype.

We've come some way towards answering this question as a field, and here's two take-home lessons. First of all, SV40 large T-antigen, as we've heard, is complex. It contains multiple activities required for transformation, at least three, and I emphasize at least, and to further complicate the issue which activity or combination of activities is required to transform depends on the cell type.

Now, Jim and Ellen Fanning earlier reviewed some of the known targets of T-antigen, so we know that from work from a number of groups that P53 is a target relevant to transformation and binds to a bipartite region out here near the C-terminus. That was mapped by Tim Kierstad and Judy Tevethia. And, we know that the RB family, as Jim DeCaprio just mentioned, is an important target for transformation. And then, what Jim focused on and what I'm going to focus the rest of my talk on is this amino terminal region, which also seems to play a critical role in transformation, and we showed some time ago, both in cell culture and transgenic mice that this region encodes an activity of unknown cellular targets, so we called it X.

Now, some time ago Keith Peden and I characterized a whole series of aminoacid substitution mutants that extend through this region and found that mutations in this region could affect multiple viral functions. So, you can get mutants in this region that affect virion assembly, but are fairly normal for viral replication or transcriptional control, you can get mutants that are defective for replication, but still transform normally, or you can get mutants that affect all three of those functions, assembly, replication and transformation. So, mutations in this region can have a very pliotrophic effect.

Some time ago, Ashok Srinivasan in my lab, showed that a fragment, consisting of the first 121 aminoacids of T-antigen insufficient to transform a number of cell types. What I'm going to focus on today is a new mutant which was constructed that is a little more user friendly than that one, it's called TN136, and it makes an amino terminal fragment of T-antigen that consists of the first 136 amino acids, so it goes from the amino terminus right through the nuclear localization signal, and then truncates.

All right. Let me point out one other thing here. If we look at this region, there's really three genetic elements, and Jim DeCaprio just touched on them, that lie within this minimal transforming region. The first is the RB binding motif. The second is this conserved HPDKGG sequence, which seems to occur at the loop of a J domain, and then there's this region down here that shares homology with the adenovirus E1A and -- virus E7 proteins. That's the immunoacid 17 to 27 region we've heard referred to.

So, what I'm going to show you is the

results of transformation assays, where we mutate each of these elements, both in the context of a full length T-antigen and in the context of this truncated TN136 protein.

All right. So, on the top, what we are looking at here is the numbers of foci induced -- transformed foci induced by a plasmid expressing wild type or mutant T antigens on two different cell lines, C3HT10T½ or REF 52, wild type T-antigen transformed both with an equal efficiency. This first mutant, 1135, deletes the 17 to 27 motif at the extreme immunoprecipitation terminus, and it fails to transform either cell type, in the context of a full length T-antigen.

These are two mutants within the loop of the putative J domain, D44 to N, and this one carries multiple substitutions. Most of these make stable proteins, but here we can see some subtle differences arising in the immunogenic analysis. D44N still transforms both of these cell types, although with a reduced efficiency, but this mutant is absolutely defective for transforming both cell types.

Here are the RB binding motif mutants, and these mutants are -- this region is not required to transform the 10T½ line. We still get some foci, although it's reduced efficiency, but it's absolutely required for the transformation of this REF 52 line.

Now, where this gets much more interesting is when we look at the effects of these mutations in the context of the N136 protein, so what we know is, if we just express the first 136 amino acids we retain the ability to transform the 10T½ line, and now we want to know what the contribution of each of these sequence motifs within the immunoprecipitation terminus, how each of these motifs contributes to the transformation of this line.

And, what we see is, if we mutate either the 17 to 27 immunoacid motif, the J domain loop, or the CR2, mutations in any three of these elements leads to a loss of the ability to transform.

Now, we've done quite a bit of -- just to anticipate a question -- we've done some experiments where we attempted to see a small T-antigen could complement this defect in the 10T½ cell system and in the REF 52 system, and it does not.

All right. So, Jim outlined some evidence that the immunoprecipitation terminal part, the large T, small T region of large T-antigen is a J domain. And, we first noticed this as a result of some work by Hasa Georgopolis and Walt Kelly, that pointed out a homology between the immunoprecipitation terminus of T-antigen and J domains, and he's done an experiment in E. coli where he's taken the J domain region of both SV40, BK virus and JC, and shows in a chimeric molecule that it functions in E. coli as a J domain.

What we wanted to do is look at the biochemistry of this region and see if had the biochemical attributes of a J domain, and so we purified these mutant proteins, diluted two by four, and performed the test that I'm going to tell you about in a minute.

Now, before I go on to this test, I want to remind you that the J domain homology region is also present in small T-antigen, and so we also were curious as to whether small T-antigen might carry any of the activities of a DnaJ molecular chaperone.

All right. This is just a sample gel showing the purity of the proteins we are using, so this is immunofinity purified wild type, large T-antigen, PTN 136, and this is a small T-antigen that was purified by Kathy Rundell, and a mutant of small T which mutates two of the residues in the loop of the J domain. And so, we used these in several in vitro assays to test for J domain-like biochemistry.

Now, as Jim alluded to, Jim DeCaprio just alluded to, one of the attributes of a DnaJ molecular chaperone is its ability to stimulate the ATPase activity of its cognate DNAK partner. So, the first thing we tested, and this was work done by Jai Vartikar in my lab, was to test the ability of T-antigen of the PTN 136 transforming protein to stimulate the ATPase activity of HSC 70, one of the mammalian DNK homologs.

And so, here we see the intrinsic ATPase activity of PTN 136. The T-antigen has no ATPase activity. This is the intrinsic ATPase activity of HSC 70, and this is a commercially purchased prep, and then we see when we add in 136, mix in 136 with the HSC 70 we get an acceleration of HSC 70 ATPase activity, reminiscent of a J domain activity.

Down here in C, Jai has just shown that this is a stoichiometric effect, so as we increase the molar ratio of N 136 to HSC 70 we increase the acceleration of the -- we increase the rate of the ATPase reaction.

And, over here in B, we tested whether or not wild type full length T-antigen could also affect the ATPase activity of HSC 70. This assay is complicated, because T-antigen itself carries an intrinsic ATPase activity, so what you are looking here is the ATPase activity of HSC 70 alone, of T-antigen alone, or of the mixture of the two.

Now, for a number of reasons we were not satisfied with the HSC 70 that we were able to get, and so we started a collaboration with Jeff Brodsky, who is also at Pittsburgh, looking at the ability of T-antigen to stimulate the ATPase activities of some yeast DNAK proteins.

Now, the system we chose was a yeast DNAK homolog called SSA1P, and its cognate DnaJ partner is called YdJ1p, and in this graph you are looking at the intrinsic ATPase activity of the yeast DNAK, SSA1P, the J protein, yeast J protein has no ATPase activity, mix them and you get a simulation about four to five fold in the ATPase activity of the DNAK protein.

Here is the results when we mixed the purified PTN 136, the T-antigen fragment, with SSA1P, and here we see about an eight-fold, ten-fold stimulation of the ATPase activity. If we pretreat the T-antigen fragment with a thermal lysin or any other protease, we destroy its ability to stimulate DNAK activity, so this is not the result of some small peptide stimulation.

In this graph, we are showing the ability -- we note the monoclonal antibody to T-antigen, PAb 419, whose epitope resides within that immunoprecipitation acid 17 to 27 region, can partially block the stimulation of SSA1P ATPase by T-antigen. So, again, SSA1P ATPase alone and 136 alone, mix the two get the stimulation. The antibodies have no intrinsic ATPase activity. If we do a triple mix of the T-antigen, SSA1P and the monoclonal antibody we see about a 2.4 fold reduction in the ATPase activity of SSA1P. This is an irrelevant antibody that does not reduce the -- that binds T-antigen but does not reduce the ATPase -- the ability to accelerate ATPase activity.

As I mentioned, we also wanted to look at small T-antigen to see if it might carry a J domain-like activity, and so this is the material purified by Kathy Rundell. Again, here is the ATPase activity of SSA1P. Small T alone has no ATPase activity, but if you mix the two you get about a four-fold -- four to five-fold stimulation of SSA1P APTase, about the same as YdJ gives you.

This is the results when we look at small T that carries a double amino acid substitution within the loop of the J domain, and we see that we lose the ability to stimulate SSA1P ATPase.

And finally over here, we looked at the ability of full length T-antigen to stimulate the ATPase activity of the yeast homolog, and in this case to get around the problem that T-antigen carries its own ATPase we used a mutant T-antigen that carries an immunoprecipitation acid substitution out in the ATPase domain, and inactivates this ATPase activity.

So, here we see ATPase activity of SSA1P alone, the mutant full length T-antigen alone, or if we mix the two we get the stimulation of SSA1P ATPase.

Now, stimulation of ATPase by DNAK, ATPase by a DnaJ protein is just one attribute of DnaJ function. The other is that DnaJ stimulates the release of denatured peptides from a DNAK protein, and so that's what we are testing in this assay.

I forgot to mention that this work and the other yeast work was done by Ami McClellan, a graduate student in Jeff Brodsky's lab.

So, what this is, is we take an irrelevant protein, this is carboxy methyl lactalbumin, we denature it severely with urea, iodate it, and then what you are looking at is the way that denatured iodinated protein runs on a native gel. So, it runs the blur down here.

If you mix it with DNAK, a DNAK protein, in this case SSA1P, you see a band shift and this represents the denatured peptide bound to the DNAK protein.

If you then add DnaJ protein to this, in this case we are adding the YdJ1p, the yeast cognate partner, then the bound peptide is released from the DNAK protein, as we see here.

In these three bands, we show that the -- or, in these three lanes we show that PTN 136, or two mutant full length T-antigens that are -- for ATPase, also stimulate release of the bound peptide from SSA1P.

So, based on this data, we conclude that both large T-antigen and small T-antigen carry a DnaJ-like activity, and that in the case of large T-antigen it seems to have most of the attributes or many of the attributes of a DnaJ molecular chaperone.

Now, this slide, I'm getting close to the end, allows me to speculate rapidly about what this means. All right. So, what I've told you is, that the minimal transforming region in the immunoprecipitation terminus of T-antigen consists really of two elements, the J domain, which is common to large and small T, and the motif that's responsible for binding, the retinoblastoma family of proteins.

Now, this is not speculation yet, but I'll get to it in a second. In a cell what we know happens is, is that there's a -- complex between the retinoblastoma protein E2F1, or an E2F family member, and a DP family member, and T-antigen stimulates the release of the E2F DP heterodimer from that complex and you are left with a TRB complex.

In small T-antigen, we also have this J domain next to a motif, and in this case, as Kathy Rundell pointed out, this motif is involved in affecting the activity of the phosphatase PP2A. And, in that case the reaction is listed over here, small T comes along and displaces the B regulatory subunit of the enzyme, and you are left with an AC small T complex and release B.

So, one feature that's in common between these two proteins is that they both are involved in stimulating the rearrangement of multi protein complexes. And so, one hypothesis that we're testing right now is that T-antigen does not act by displacing these transcription factors by differential affinity, but rather, there is an energy requirement, that is, T-antigen participates as a co-chaperone, and that chaperone activity, perhaps, with HSC 70, is required to lead to this release. And, we're testing that hypothesis right now.

Now, one of the things that strikes us about T-antigen is that it has a plethora of activities and seems to bind almost every protein you can name, and this has bothered a lot of us in the field. Why can this protein bind so many other proteins? And here, this is not a complete list, just a quick list, this shows you some of the multi protein complexes that T-antigen has to act on in a number of biological processes. So, in transformation, we're well familiar with these proteins that it acts upon. T-antigen binds a number of transcription factors, and, perhaps, the pre-transcription initiation complex is one of the mothers of all the protein complexes. In DNA replication, we know that T-antigen binds DNA prolimerase alpha, and Ellen Fanning has shown that one of the contacts for the major subunit of prolimerase alpha is through this J domain, and a number of other proteins in the pre-initiation complex.

So, again, one of the common themes in all these processes is that T-antigen stimulates the rearrangement of multi-protein complexes, and so, for our general speculation we think that maybe the J domain is acting as a crow bar or a lever, and in cooperation with energy provided by a DNAK homolog is stimulating the release of these proteins. And so, if that's true, we'd say that the T-antigen action on HSC 70 or cooperation with HSC 70 represents a new target of T-antigens to be relevant for a number of these biological phenomenon.

And so, for the last slide, I just want to thank the collaborators. In my lab, it was Ashok Srinivasan, Jai Vartikar, Alice Castellino, who really started purifying a lot of these T-antigens, and Paul Cantalup and Ian Marks, who have continued the biochemical characterizations, Jeff Brodsky and Ami McClellan, who have done the SSA1P and work in yeast, and finally, Kathy Rundell from Northwestern, who collaborated with us on the small T-antigen work, and, as always, Keith Peden, we continually dig back into his bag of mutants and come up with interesting findings.

So, thank you.

(Applause.)

DOCTOR KELLY: Questions?

AUDIENCE: In the ATPase experiments, did you -- there are some reports that you needed a third protein, a kip protein or something, for DNAK, DnaJ increased ATPase activity, and was that present or is this a crude system?

DOCTOR PIPAS: Yes. I think the question was, in some systems for DnaJ/K interactions you need a third protein, and I think you are probably referring to grup e, and that's required in the -- that's present in the E. coli and DnaJK system, but no grup e homolog, as far as I know, has been found in mammalian cells. So, it's not clear that it's required or some other type of protein is required. But, apparently, you don't need it for any of these in vitro reactions.

DOCTOR KELLY: So, if I got it correctly, small T doesn't complement mutations in the J domain, so I guess that would suggest that the substrate and the chaperone sort of have to be in sis or something, or have to be on the same molecule, is that the idea?

DOCTOR PIPAS: Very good, that's something I didn't have time to talk about. We've done a series of complementation tests to look at the ability of mutations in different regions of the molecule to complement, and without going through all the nitty-gritty, it appears that our interpretation right now is the J domain must be in sis with CR2 and must be in sis with the C-terminus to carry out transformation.

But, that may not be the case all the time. That's the case in our systems. Kathy Rundell clearly has a case that she just published where small T can complement in some cell type systems some of these same type of defects.

DOCTOR KELLY: Any other questions?

Thank you very much.

All right. Our next speaker is Harvey Pass from Wayne State, and he's going to return to the SV40 induced hamster mesothelioma model.

Return to Agenda

DOCTOR PASS: Thank you, Doctor Kelly.

I'm going to present actually two portions of this talk. One is going to be concerned with IGF 1 receptor, and one will be concerned with T-antigen experiments. All these collaborations were done with Michele Carbone's model, which we had in our lab at the NIH when I was here, and these collaborations were done with also Ron Kennedy, providing us T-antigen for the second portion. All the experiments were done in my lab, done by Jessica Donnington, who was a fellow in my lab at that point.

Could I have the first slide, please?

The question is why such a topic should be on this forum, and I think that what it's leading into is the potential for treating these tumors that we've talked about, and either with peptides or with some sort of molecular gene therapy if you can identify certain targets.

And, my interest in IGF1 receptor, next slide, came from the fact that IGF1 seemed to be involved with many, many activities of the cells, mainly for regulation of growth, and that if you had a knock out mice that you did not have IGF in it, of course, this animal did not grow.

Also, there are many, many cell cycle phenomena that cooperate IGF with PDGF to promote cell cycle activities and proliferation.

Next slide.

The receptor is interesting because the receptor also is very important, not only for cell cycle activities, but also has been shown by Renata Baserga to be involved in apoptotic activities now. So, there was some data that there were certain tumors in which the IGF1 and the IGF receptor autocrine growth was important.

And, there's certain tumors that were interesting, in that they were brain tumors that seemed to be involved in IGF, but certainly mesothelioma was not mentioned here when we started our studies.

But, there was some interesting data from Trojan & Ilan that showed that if you took an antisense to the ligand, to IGF1, in a murine model, they completely lost tumorigenicity, the clones that were the antisense clones, and that there may be some sort of immunologic response that causes this, as well as a recall phenomenon.

About that time, a paper came out from Sloan Kettering, in which they looked at mesotheliomas and found that IGF1, IGF2, as well as IGF1 receptor, were expressed in normal mesothelium as well as in mesothelioma cell lines, and this was fairly relevant considering the antisense work that had been done by Trojan.

So, we asked the question, is there an IGF1 mechanism that is operative in mesothelioma, at least in the murine models that we could use, and if you then were able to block IGF1 receptor, as opposed to ligand, could you then decrease tumorigenicity in this system?

The model that we used then was the one that was already in the lab, and that was the H9A SV40 induced hamster mesothelioma model that Michele has alluded to with this slide, that again shows that if you take the cell culture and place it into the animals' bellies you will get a picture that looks exactly like human mesothelioma.

But, before we could undertake any of these experiments, certainly we had to show that there was an IGF1 mechanism involved here, and the reason of looking for this model instead of going to other models was that there was intriguing data from Renata Baserga's laboratory, who was my collaborator, that in order to get cell transformation by SV40 T-antigen you had to have a in-tact IGF1 receptor. And, in fact, when you had cell transformation by SV40 T-antigen the levels of IGF1 as well as receptor go up remarkably.

And then finally, he had shown in the laboratory that if you took an antisense oligonucleotide to the IGF1 receptor that you were able to decrease the growth stimulatory activity of T-antigen.

So, it seemed like a logical step that there was SV40 T-antigen IGF1 receptor relevancies here, and thought that this model then would be one that would be good.

But, in order to show that we had an IGF1 mechanism here, we actually did RTPCR looking to see whether we had IGF1 or 2, and then the respective receptors. And, what you see here is the H9A, which is the hamster mesothelioma model, compared to a spontaneous and asbestos-induced mesothelioma model in rats that Cheryl Walker had, and what you see is that like the rats the H9A mesothelioma model, indeed, expresses IGF1 and expresses IGF1 receptor, as well as IGF2 receptor. So, at least we knew that ligand was there, and we knew that the receptor was there, and then using a standard radioimmunoassay we took the cells, the H9A cells, plated them out in serum free media and then measured the development of IGF1 in the media and found that, indeed, either on the basis of N per CC of supernatin or on the number of cells that there was a progressive increase in the amount of IGF1 in the media. So, at least we knew that the cells were making IGF1.

Here is the construct that Renata Baserga sent us. This construct, essentially, is under the control of the heat shock promoter, and is a 309 base pair construct in the sense of the antisense position for the IGF1 receptor, that you could clone out then using neomycin resistance. We used standard techniques of calcium precipitation, using this to get the sense or the antisense vectors into the H9A cells, and then cloned them out in G41A, and we're hoping that then we could see some temperature dependence of these clones.

Well, what you would expect from cell cycle analysis is that, indeed, if you were able to get the antisense expression vector in, you should decrease IGF1 receptors and then you should hold those cells in G1 and have decreased numbers of cells in S phase.

And, to verify that, we actually did the facts analysis of these clones. The B9 clone is the antisense clone, the A3 clone is the sense clone, and what you see is that at 34 degrees, which is when the vector is minimally active, you had about the same number of cells in S phase between wild type H9A, sense and antisense, as well as in G1, but then when you turned up the heat to 39 degrees there was no difference between the sense and the wild type clones, however, a reduction in the number of cells in S phase and an increase in the number of cells in G1 in the antisense clones.

This actually translated into functional data, in that, here is the number of cells plated down after three days at 34 degrees, with the sense clone and the antisense clone not too much different at 34 degrees, but then when you turn up the heat for those three days the sense clones will proliferate nicely, while the antisense clones drop dead.

Well, in vivo, we thought then that we could take these clones and implant them. This was a subcutaneous model, as well as a IP model, to see if we were able to then decrease tumorigenicity. We concentrated on the subcutaneous model here. We would expect that in the animal that the antisense clones would give less tumors than the sense clones, and we also had wild type controls.

Here's the results of those experiments, which were done twice with more than 30 animals in each of these groups, and what we found was that the antisense clones formed less tumors than the sense and the wild type. When you analyze the tumors that actually developed or escaped this protection, this loss of tumorigenicity in the antisense group, what we did was, we actually extracted the DNA and then looked to see if the vector was still there by looking at a certain region, axons one and two, and what we found was that here are the plasmids without anything done to them, and here is the sense tumors, they are all there, but this is the wild type, of course, that wouldn't have the vector in it, the ones that escaped therapy that were antisense somehow had lost the vectors. So, essentially, they lost that tumorigenicity lack and made tumors.

Does this have any human relevance? Well, it does have human relevance, because if you look at our cell lines that we developed from the patients that I operated on, and stain them for IGF1 receptor, comparing mop C to now the IGF1 receptor antibody, they all have IGF1 receptor, and also if you put them in culture this is how they'll grow in serum free media, if you don't have the IGFs, this is one of my cultures, Gates, if you add IGF1 and IGF2 there's a greater cellular proliferation. So, I think there's going to be a role in the future for targeting the IGF1 receptor, and that may be directly related, possibly, to SV40 interactions in these tumors.

Well, finally, because we're talking a little bit about possible therapies, I'd like to concentrate a little bit on our preliminary work, in which we have attempted to use this model in a collaborative effort with Mike Shear and Ron Kennedy, to see if we can do protection assays in vivo. Essentially, what we started out with was purified T-antigen from Ron Kennedy from a bacula virus system, in which we took very small doses of T-antigen and gave them to these hamsters, in a short vaccination schedule, only days one and days 14, and then we implanted tumor taking some sera to look at the development of antibodies to T-antigen, and then looked for tumor growth.

And, this was our original -- this was our original ELISA. The ELISA, this is the serum from an animal that just has the tumor. Of course, this animal is a T-antigen expressing tumor, so naturally that's the positive control, and then if you look at the animals that received T-antigen this one with Freunds you see that there is an increased level of antibodies to T-antigen, but not great, and, again, these were low dosages of T-antigen that were given.

Here is the T-antigen group. This is the development of tumors, animals who did not develop tumors would be up here. The T-antigen group sort of gives you an idea that you may be able to protect against a subsequent challenge, but what we did was, we modified this by going to a higher dose of T-antigen and a longer vaccination schedule, 20 micrograms and ten and ten, then challenging them IP with H9A, but also bleeding them to look at their development of antibodies to T-antigen in this long vaccination.

And, here we are again with the positive controls. These are individual animals now, positive controls show antibodies of T-antigen, but the long vaccination certainly enhanced our ability to detect antibodies to T-antigen in these animal sera. Here are the controls, of course, showing no antibodies, and in that particular experiment the animals seen here were the animals that were vaccinated, this is survival, versus the controls, saline versus serum albumin.

So, again, this is a protection sort of assay, but, again, gives credence to the next series of talks that will be given by Doctor Tevethia and by Rob Bright, and really are our initial efforts to see if we can try and treat this based on T-antigen as an immunogen.

Thank you very, very much for having me present this data.

(Applause.)

DOCTOR KELLY: I think we have time for one or two questions, if there are any.

Did you actually look at the level of expression of IGF1 receptor in the cells that had the antisense, to see what level of ablation you got?

DOCTOR PASS: Well, actually, we did scratch analysis in vitro. What we found by the scratch analysis was that there was a 44 percent decrease in the number of IGF1 receptors when you turned up the heat in the antisense compared to the sense. And, it's interesting that the levels of IGF1 receptor in these tumors, including the wild type, was very, very high compared to what you see in the literature.

DOCTOR KAST: Hi, Martin Kast, Loyola University. Could you speculate how you --

DOCTOR KELLY: Could you talk a little bit louder, Martin?

DOCTOR KAST: Could you speculate how antibodies against large T have a protective role?

DOCTOR PASS: Antibodies against large T have a protective role?

DOCTOR KAST: Well, that's what you show in your model.

DOCTOR PASS: What I show in my model is that -- oh, you are talking about the mechanism.

DOCTOR KAST: The mechanism, right.

DOCTOR PASS: Absolutely. I'm going to leave that to Rob Bright, who works for me, but I think Rob can go into that a lot better than I can.

DOCTOR KAST: Harry, I was asking you the same question, anyway, there is some reports then that already there is adenovirus around that has been produced that has antisense Baserga, the same kind of antisense against IGF1 receptor.

Now, what is happening, is that this adenovirus, however, may not be able to suppress the receptor expression, since the adenovirus just induced over expression of the receptor by itself. How you can think that this gene therapy strategies can be done?

DOCTOR PASS: Well, I'm not so sure -- you've got to pick your vector carefully, and I've had discussions with Renata about that, because adenovirus will increase your expression of IGF1 receptor to begin with, so, obviously, that's not the vector. But, other types of vectors may be useful.

My own inclination is that I'm not even sure whether this mechanism of delivering IGF1 receptor is going to be the good one, because now Baserga has developed a soluble receptor, which is a dominant negative, which may be even more interesting to look at, and we plan to do that.

DOCTOR BRIGHT: Robert Bright from Wayne State.

I just -- not a question, but just want to add a point regarding Doctor Kast's question about mechanisms, in this particular model in specific.

I think the rationale behind monitoring T-antigen antisera level was just as an indication of an ongoing immune response, but not to imply that there's a humoral active response going on here, and the problem being that to follow a cellular response would require sacrificing the animals prior to tumor challenge.

But, Doctor Pass does have data that demonstrates that lymphocytes from these animals do proliferate, indicating some sort of cellular response when incubated in vitro with T-antigen or the inactivated tumor cells.

DOCTOR PASS: It's a very difficult model to work with, to tell you the truth, because these animals, of course, are not syngeneic, so I don't have data that talks about the non-cellular aspect of it.

DOCTOR KELLY: Okay. Thank you very much.

Okay. We'll move on to Satvir Tevethia, from Penn State College of Medicine, who is going to talk about CTL responses to SV40 T-antigen.

Return to Agenda

DOCTOR TEVETHIA: Well, I want to thank the organizers, too, for inviting me.

One of the nicest things in working in SV40 for about 25 years, Janet, is that you get to show old slides and find out if they are still relevant.

So, if I could show you, this is a slide that I made in 1974 for a meeting, and it shows the following, that interaction of SV40 with a non-permissive host, then the first thing that happens is that the T-antigen synthesized in the cell surface, transplantation antigen appears. And, what we said at that time, that the normal lymphocyte gets sensitized, to become sensitized lymphocyte. And, remember, at that time we don't know anything about CD8 positive lymphocytes and so on and so forth.

Then, we see how the transformation continues here, and then cell proliferation leading to the tumor formation.

Now, all of these events since then have been really worked out and we know exactly how the immune system works, and I can tell you, at least in the mouse system, this is strictly a T-cell mediated phenomenon.

One of the things that we proposed at the time, that let's see what happens if we were to consider the SV40 infection with the permissive host, could be a JC in the humans or SV40 in monkeys, similar events occur, that lymphocytes get developed. In the meantime, the cell dies, because of the fact they are losing the virus.

Now, the virus then, at the same time the neutralizing antibody will develop, and the neutralizing antibody will block further infection in any -- if you want to speculate in 1974, that there were going to be tumor cells developing, they'll be taken care of by the cytotoxic T-cells at that point.

The more interesting thing is, in light of this meeting, we know everything there is to know about this system. We know nothing about this system. At least I don't recall a single experiment that has been done to prove or disprove this system. This is especially relevant in light of the number of papers that were presented yesterday on JC viruses, that every one of the speakers has said that JC virus is activated under immunosuppression. Therefore, one can conclude that the immunological control must be keeping HC viruses latent.

What are those controls? And, I don't recall anybody has tried to look at those controls in this manner. So, I think we need to pay attention to this route in the permissive host.

So, there are two points I wanted to make from this slide, number one, that if one isolates a virus, which turns out to be SV40 for human tumors, you can verify this by sequencing, but there are other ways of verifying the -- SV40 and it's exactly what we have done and I'm going to describe to you. We received three viruses from Doctor Janet Butel when she isolated from the choroid plexus tumors in humors and two other human viruses.

And, the second part I want to tell you about is some highly preliminary experiments, give some evidence along the line that I'm talking about.

So, we all know that SV40 T-antigen is an extremely immunogenic protein. T-antigen can immortalize cells, everybody heard that, purify T-antigen immunizes mice against tumor transplantation. With Bob Pigeon in 1980, or early on, we showed that as low as .25 micrograms can immunize the animals. DNA encoding T-antigen immunizes mice against tumor transplantation. The T-antigen in transformed cells then induces a generation of cytoxic T-cells that ca be restricted by different -- molecules, and purified T-antigen in 1980 with Bob Pigeon showed it can induce a generation of cytoxic T-cells.

And, CTL now participates in tumor rejection. This is shown, not just by our work, but Martin Kast and his group in Holland has done that a number of years ago.

Okay. This is just to show you a very old experiment, some new data mixed in, if you immunize mice with SV40 virus or irradiated tumor cells, or with DNA, and challenge them with tumor cells, you can see they are all protected. This goes to show you there are a number of ways of immunizing the mice that will induce protection.

Now, this is another model we've been working with, this is the transgenic mouse model that Arty Levine worked with, with a picture of the mouse that was shown yesterday, so there is a tumor appearance in the choroid plexus, and you can see the entire ventricle, all ventricles are filled with the tumor. And, all these mice died by 104 days.

Now, when we infuse into these mice the immune T-cells, T-cells that are immune to SV40 T-antigen, and this is what happens. This is, you can see that the tumor, or hardly any tumor, it looks like almost like a normal choroid plexus, and these mice lived rather than 105 days, anywhere from 150 to 250 days. So, one can see it is possible to do some immunotherapy using the immune lymphocytes.

Now, this is just to remind me that I'd first like to discuss that virus that Doctor Janet Butel sent us, then how did we characterize then the -- SV40 rather than BK and JC. That goes back to the fact that, just to show you that it is possible to generate the cytoxic T-cells. When you go to use the cytoxic T-cell clones as a -- just like you'd use for Southern Blots, to really document the presence of the epitopes that are characteristics of SV40, and one can immunize the -- mice with the T-antigen expressing cells, remove the spleen cells, you stimulate them in vitro, and then you can establish cytoxic T-cell clones, and we have done that, and before I go and describe to you those clones I just want to briefly refresh your memory about how the T-antigen is processed or -- protein. When a virus infects the cell, proteins are synthesized, these are broken down into peptide, the peptides are shunted through the TAP transporters, and then where they assemble with the peptides assembled are recognized by the class of molecules, where the class of molecules assemble, and then ultimately are transported through the endoplastic reticulum to go out to the cell surface, where it is seen by the T-cell receptor of the cytoxic T-cells.

And, this is where the peptide is presented by the MSV molecule, is essentially a -- peptide, it fits in the groove formed by the alpha one and alpha two helices of the class one molecule.

So, in SV40 T-antigen, restricted by H2B class one molecule, there are only four CTL epitopes. One, 206 to 215, 223 to 231, 404 to 411, 489 to 497, and these are the respective CTL clones, K1, K11 and so on and so forth.

Now, if one then takes a cell transformed by JC and SV40 and compared the reactivity of these CTL clones, especially for SV40, this is what one finds. Clone epitope one is recognized by two different clones, Y-1 and YK-11, and you can see Y-1 clone recognized SV40 epitope and JC, but not BK, and within epitope 1, K-11 recognized only SV40, not JC and not BK. Epitope 2-3, where the sequence is identical between BK, JC and SV40, recognizes all three transformed cells. Epitope four recognizes SV40 and JC, but not BK, and epitope five does not recognize either JC or BK.

Now, you can tell very easily, by using three of these different clones, or three of these different clones, you can have a cell line which is expressing the either BK, JC or SV40, and you can actually do a mismatch or other matching to see whether it is SV40 or a JC virus in those transformed cells.

So, this is just to give you an example of what is the basis of this discrimination the CTL clones see in terms of the epitopes that are present in BK, JC and SV40. This is the epitope one, it's a ten amino acids long epitope over here, the sequence of SV40, just remember the JC virus, the first, this epitope at residue 212 alanine to cysteine, there's a substitution over here, and you can see that Y-1 CTL clone recognized this, that means it knows the cysteine, can recognize it, but not K-11. So, K-11 -- dictated by the change from alanine to cysteine, BK you have two changes from alanine to cysteine, not recognized by either Y-1 or K-11. In a number of ways I think it is more specific than doing the PCR analysis, and it's not subjected to any criticism with regard to the contamination and temperature changes and so on and so forth.

So, and the other point I wanted to make, this is the function alignment that we have done for the epitope one over here and two, but I want to mention this, the residue 212 is an alanine, and remember I showed you the change from alanine to cysteine. Now, this residue then interacts with the T-cell receptor, where the two residues here are the anchor residues, they interact with the MHC molecule. So, the T-cell receptor interacting residues, and no wonder they change from alanine to cysteine now, abrogates the reactivity with the T-cell clones.

Now, we have good slides, not we, Stan --with our collaboration, have good slides of this particular peptide with the MHC molecule and analyzing the 3D structure for this.

So, now coming back to this, and you can see that going over this diagram again, these are the three viruses you see from Jan, PML, MEM, CPC, when the viruses came Judy Tevathia was very kind enough to open them only in one particular hood and directly went with the primary mouse embryo fibroblast, they transformed as well as any of the viruses, and you can see -- recognize all were CTL clones, they are all T-antigen SV40. That just goes to show us the power of the CTL clones, you know, can lead to some definitive conclusions about the origin of the viruses.

Now, I want to move on to the second part of the talk, and then I'll try to be brief, Mr. Chairman. The strategy for assessing the CTL responses to T-antigen in humans, and the first thing was determining the HLA, HLA-A2, and the reason we chose HLA-A2 is because the HLA-A2 binding motifs were available through the hard work of Jurkat Raminses and others, and a lot of work done by Martin Kast and his group later on with the HPV.

And so, you first determine the HLA binding sequences in T-antigen, based on the motifs, so you do the computer analysis, you synthesize corresponding peptides, and then what we did, we then took the sequence of these peptides and expressed them in the vaccinia virus behind a refined E glycoprotein E19 promoter, so then all the peptides that are made will directly be shunted into the endoplastic reticulum.

And then, we determined the binding peptides, binding capacity of these peptides with the HLA molecules, and then we immunized the HLA-A2 transgenic mice, and I'll explain that in a minute, with synthetic peptide and assayed the CTL activity against cells that are expressing HLA-A2, plus the peptide.

Now, this is the computer analysis of both the SV40, BK and JC for the peptide that might be possibly a candidate for presentation with HLA-A2 molecule, and you can see that the peptide over here, 197-205, is common between SV40, BK and JC and so on and so forth. And, for controls we used the peptide 82-90, that had been shown by Martin Kast's group to be an immunogenic peptide.

And so, these peptides were synthesized, the same sequence I showed you before. I won't belabor the point. And, this is the protocol. What we did, we used a transgenic mouse, which is expressing HLA-A2 molecule, and what this is about, that the alpha one, alpha two domains is contributed by the HLA-A2, because these are the ones that present the peptides. The alpha three domain and transmembrane region and cytoplasmic tail comes from the H2KLB molecule, these mice were provided by Doctor Linda Sherman in Scripps, and both at Scripps and Martin Kast's group, who is here, have shown previously that these mice respond to making cytoxic T-cell which are restricted by the human HLA-A2, and since I've been told that 40 percent of the population have HLA-A2, we thought this might be a good candidate for this.

So, we immunized them with 100 micrograms of the peptide, and 140 micrograms of HBV core T helper peptide, in -- Freunds -- we then take the spleens out, we culture them with syngeneic LPS blasts from the same animals with the micro molar of A2 binding peptide, and we acid in the chromium release assays.

Now, what I'm going to show you are highly preliminary experiments and they need to be confirmed several times before we can say this is correct.

So, this is a chromium release assay, now this is a control peptide, HPV-16 82-90, you can see, if you have your target in a cell line that now has this peptide on the cell surface, and you can see that these lymphocytes now lyse the cell effectively 67 percent and 44 percent, and they don't see the irrelevant T-antigen peptide to any appreciable degree.

Here, the T-antigen peptide, it now induces a very good response, 68 to 52 over here, and it does not touch the targets which are coated with the HBV peptide. This peptide also gives us a response, but not in this particular experiment. The rest of them proved to be non-immunogenic.

Now, this slide shows that the cell lines, which are HLA-A2 positive, when you infect them with vaccinia virus expressing the peptide are capable of processing this particular peptide are of the vaccinia virus context, and then shunted, and then peptide associating with the class one molecule being presented, and, you know, then it's recognized by cytoxic T-cells.

What is interesting is that, in the interest of time, that you can see that the CTL to peptide 140 to 150, they recognize both the peptide -- cells as well as the vaccinia infected cells, but unfortunately they don't see the peptide process directly from the full length T-antigen which we have expressed in the vaccinia virus. I think we know what the problems are and we are trying to solve that.

Okay. And, this is to show you that this is our -- not for long, last week we marched with the Geisinger Medical Center, we don't know what the name will be.

Thank you for the opportunity.

(Applause.)

DOCTOR KELLY: Thank you very much. Questions for Tev?

AUDIENCE: Satvir, do you have any idea if your two epitopes are being transported by TAP?

DOCTOR TEVETHIA: No.

AUDIENCE: You mean they are not transported or you don't know?

DOCTOR TEVETHIA: The question was, whether any of these two peptides are transported by TAP. The reason we don't know that, because the vaccinia construct that we made was made behind an adenovirus in 19 glycoprotein ES sequence, and that directly deposited the peptides into the endoplastic reticulum. No, we don't know that.

AUDIENCE: Doctor Tevethia, do you have any evidence that lymphocytes from HLA-A2 positive humans will lyse targets that foster that peptide?

DOCTOR TEVETHIA: Not humans, transgenic mice. Okay, the question was?

AUDIENCE: The question was, do lymphocytes from HLA-A2 positive humans recognize the epitope in the context of HLA-A2?

DOCTOR TEVETHIA: We are only up to the level of the transgenic A2 transgenic mice. We haven't -- actually, I wanted to mention that, we have the -- you know, when I got my grant renewal this experiment was proposed, but we are not set up or funded for doing the human work. I think we want to leave that to other people.

DOCTOR KELLY: Okay.

We'll move on to our last talk by Robert Bright, who is going to be talking about immunotherapy of SV40 induced tumors in mice, a model for vaccine development.

Return to Agenda

DOCTOR BRIGHT: I think it's obvious from the title that my presentation falls under the category of other materials.

I'd like to take this opportunity, though, to thank Doctor Lewis and the organizers of the meeting for the invitation and the opportunity to participate in this meeting, and present some of the experimental data that we've produced over the last few years, examining the possibility of treating SV40-associated or induced tumors by immunotherapy.

Just to review briefly, and the way I put this talk together, I left out a lot of the data in the interest of time, and also knowing that by and large the audience wouldn't be immunologically oriented. So, I tried to orient this talk so that I could skip over the nomenclature and the immunobiological mechanisms of immune responses and it would still be clear that the point we want to make is that T-antigen represents a target that you can treat lethal tumors with, at least in this model system.

And, just to review the first two points, as Doctor Tevethia pointed out, that animals are protected with inactivated syngeneic transformed cells, or by injection with virus, and as he demonstrated, too, you can extract T-antigen from those transformed cells and immunize animals and protect them from challenge.

And, in so doing, demonstrated that SV40 T-antigen represents tumor specific, a viral encoded tumor specific antigen on these transformed cells. And, our question was, would it be possible to generate synthetic and recombinant vaccines that are, not only cell free, but virus free, and see the same effects in animal models.

And, just to set up the animal model briefly, all this data is done primarily in the BALB/C system, with a BALB/C syngeneic SV40 transformed cell, specifically, the MKSA cell line. MKSA was generated a number of years ago, as was mentioned earlier, it's tumorigenic in immunocompetent animals, adult animals. It expresses SV40 large T-antigen in the nucleus, the cytoplasm, and as Doctor Butel and Jarvis have shown not too long ago, to a small extent on the surface of the transformed cells as a sort of integrated protein, and specifically the amino and carboxy terminuses of T-antigen are accessible to the environment.

Now, if you take these MKSA cells and inoculate them interparatoneally into BALB/C mice, they generate a tumor that closely approximates abdominal mesothelioma in humans, and just to show you the tumor titration of the model we set up, what we wanted was a read out endpoint that was controlled so that we could just ask questions about vaccinations, strategies, dosages and so forth, and we chose a lethal model over a measurement of tumor nodules, so that the endpoint was very definite and subjectivity in measuring the tumors would be eliminated.

The other thing is, we chose a protection model over a treatment model, because, historically, protection models are much easier to demonstrate immunologic responses, and, again, the point being just to identify synthetic recombinant vaccines in the system, and to later approach treatment strategies in model systems. So, this is a protection model with a lethality endpoint.

And, if you look at the titration here, what we did was, we decided to establish an LD50, a lethal dose of 50 percent of the animals inoculated, and we started with 10 million cells and worked down to 1,000. And, if you go to the middle of the screen, C57 black 6 mice, which are MHC mismatched in this scenario, obviously, the tumors don't grow in any of the animals no matter how many live cells you inoculate sub cu, IP or otherwise.

Now, if you inoculate BALB/C mice, however, anything above 100,000 cells is 100 percent lethal in 100 percent of the animals repeatedly in our laboratory, and, more specifically, at the LD50 point of 50,000 cells, repeatedly we get about 50 percent of the animals that succumb within 30 days, the other 50 percent can linger on for another month or so, but most of them eventually die, the same scenario we saw with the same cell dosages in the F1 cross of C57 black 6 and BALB/C.

So, what we chose was a two to five times an LD50 dose to use as our standard inoculant for protection strategies, and you can see in that range 100 percent of the animals die, and I want to point out that they succumb to the tumors within three weeks, roughly, around 21 to 24 days all the animals in the unprotected groups will die.

So, the question we had was, could we generate recombinant SV40 large T-antigen and use it in a vaccine scenario as an alum precipitate or absorbed in alum, and treat animals against a lethal tumor challenge with syngeneic SV40 transformed cells. And, to do that we generated SV40 large T-antigen in a vaculovirus system, and the T-antigen was characterized as being pretty much identical to wild T-antigen or T-antigen isolated from the transformed cells, both in post-translation modification and in biochemical function.

And, just to show you, the T-antigen we used is immunoaffinity purified, and this is silver strain gel, STS PAGE gel, with molecular white markers in lanes A and D, and VSA as a reference in lane C, and this is SV40 large T-antigen generated in vaculovirus and immunoaffinity purity in our lab.

And so, we took this protein preparation as an alum precipitate and inoculated animals IP, and asked the question, would this generate an immune response sufficient to protect against a lethal tumor challenge. And, this slide is representative of number of tumor challenges we've done, both in BALB/C in CV6 F1, and as I stated earlier, we haven't done any challenge experiments in the black 6 mice, because in our hands or to our knowledge the black six lines that are available are not tumorigenic in vivo, so we couldn't ask those questions. But, this is representative of a CV6 F1 or a BALB/C system.

And, typically, what we do is immunize with alum alone, an irrelevant protein, in this case it's represented by IgG, but we also had Hepatitis B surface antigen as a vaculovirus recombinant protein and alum as well, which would fit into this row.

Recombinant SV40 T-antigen and alum, and then live inactivated mKSA -- not live, but inactivated mKSA cells as a positive control, and as you can see that all animals immunized with either alum alone or irrelevant protein and alum, succumb to the tumors right around the three week point. And, if they were immunized prior with T-antigen or the inactivated cells, they survive, and, in fact, this number would go out to 180 days or however long we chose to keep them and pay for their upkeep, without any evidence of disease.

And, I want to point out, too, that routinely in our experiments of the data I will show, the immunization schedules are 20 micrograms of protein in alum, followed two weeks later by ten micrograms, two weeks later by ten micrograms, usually up to four vaccinations. This is an over kill, I just want to point out, and I won't show the data on the dosage, but we had evidence that a single inoculation with this recombinant protein at 100 nanograms in alum is enough to protect against two LE50s. But, again, for the focus of this talk, all the studies, for standardization reasons, will be done on that four immunization scheme.

Okay. And, as pointed out by Doctor Tevethia, and also by others who study tumor immunity in humans and in murine systems, Tevethia's group and Doctor Kast's group in HPV, CTL has been demonstrated to be very important in rejection of tumors in vivo, and the first thing we wanted to do was analyze the cellular response. I'm not going to show you data on cell proliferation, but we need to look at that too, but what I will show you is a representative table at a very high E:T ratio of splenocytes from animals that were immunized or immunized, challenged and survived, looking at both cellular means of immunity non-specific in K-type lyses, or CTL activity. And, normal is naive animals that were either immunized with alum alone or not immunized splenocytes, and then these were immunized with recombinant T-antigen.

And, as you can see, the NK activity is very similar to the normal, and not very high, so it probably doesn't play a significant role here, and again, we could not demonstrate CTL activity in this system.

I don't show you the data on the black 6 mice, but we were able to demonstrate induction of CTL in black 6 mice with this recombinant protein against the black 6 SV40 positive tumor.

And, just to reiterate that this isn't unique to our lab, if you look at the literature closely, a number of other labs have seen the same phenomenon, and this is just a brief summary from Barbara Knowles group in '79, where I believe the coin was termed non-responder with respect to CTL activity in BALB/C mice, and what they did was to take BALB/C mice which are H-2d restricted and inoculate them with inactivated BALB/C SV40 transformed cells, and then challenge them, or look for CTL activity against the black 6 MHC mismatched target as a control and against the inoculating tumor and couldn't demonstrate CTL activity compared to the black 6 scenario where you could demonstrate potent CTL activity that was MHC restricted.

And, interestingly, in the CV6 F1 crosses, all the CTL activity seemed to be restricted by the H-2b element shared by the black 6 parent, in that they only lysed the black 6 target, and we've seen similar results in our laboratory looking at the protein inoculation.

So, our question was, if it isn't CTL or cellular immunity, and I don't want to tell you that we don't know, or that absolutely it's not involved, we don't know for sure, but what we do know is that they make a very high titer T-antigen specific antibody response, these animals, when they are immunized with this protein, and this is just to show you that all three strains immunize either two, three or four times, as I described earlier, and in looking at endpoint titers as a reciprocal dilution of serum as a mean of ten animals, you can see that they develop an endpoint titer after formulation that's very high, 400,000 in many cases, and that all the animals respond to T-antigen with an antibody response very similarly.

So, our question was, what role are these antibodies playing in this immune response in the light that we cannot demonstrate a cellular response.

And, this is a summary slide of a lot of data that was done, to ask the questions about antibody mediated mechanisms for rejection of tumors in the system, and just to summarize, this is data from five animals from a group of ten, to show you that we looked at antibody mediated complemented the dependent cytolytic activity against the tumor, we looked at ADCC, we looked at NK and C tail as I said earlier, and we also looked at isotope of the antibodies in the serum, compared to endpoint titer.

And, what we did was, we took the serum and heat inactivated it first, to eliminate the endogenous complement as a control. And, I also want to say that all the effector cells used in the ADCC assays were not able to lyse the tumors without the presence of specific antibody, nor was complement from guinea pigs without -- or in any case, alone or with any of the antibodies.

As you can see, starting on the left there was no CDC activity to speak of, and that wasn't surprising because when we looked at the IgM in these serum, after multiple immunizations, there was very little IgM, which one would expect when you are driving immune response toward an IgG type response with multiple immunizations.

But, when you look at ADCC from five of the animals, and we saw it in every animal we looked at, you see ADCC activity that's very significant, but I want to point out, it doesn't necessarily correlate with endpoint titer, which leads us to believe that it's the quality of antibody in the sera, as opposed to the quantity that's mediating the destruction.

Now, there's some data we looked at using monoclonal antibodies that were mapped for specificity epitopes in T-antigen to ask these questions that had specific subtypes, such as IgG-2, which are known to mediate ADCC through the FC receptor gamma 2 on effector cells, and we saw very similar patterns of ADCC with T-antigen specific monoclonals of the right subtype.

So, we were confident that it's very likely that antibody mechanisms were involved here, and I also want to remind you that the T-antigen -- certain domains of T-antigen are accessible on the surface to a small degree of SV40 transformed cells. It just so happens the areas are in the immuno and carboxy terminus, so what we wanted to ask is if we could take synthetic peptides that might represent B cell epitopes from SV40 large T-antigen and repeat this data to try to identify what areas of T-antigen were eliciting antibodies that were capable of destroying the tumors.

The way we did that was to employ a computer algorithm that looked at secondary structure, beta turn, et cetera, and came up with a series of synthetic peptides that we thought would best approximate secondary structure in the antigen, being apart from the antigen, the whole antigen, and the six that I'm going to show you today are here, and interestingly, two of them were in the immuno terminal end, which should be accessible on the cell surface, and the other four in the carboxy terminal end.

And, just the sequences, we modified them with CG so that we could couple them to KLH for immunization, and replaced any cysteines with serenes to forego any disfulphide bond formation.

And then, these are the antibodies we used, or the peptides we used to immunize animals, and we immunized animals from all three strains. Initially, we looked at antipeptide serum responses against free peptide by ELISA, to get around the KLH reactivity, and demonstrated that all of these peptides were immunogenic in BALB/C, most of them in C57 black 6, and most of them in the CV6 F1 cross, and, in particular, probably the most immunogenic, with respect to an antipeptide antibody response were the ones in the immuno terminal end.

Now, when we looked at -- this was interesting, but it wasn't relevant with respect to whether or not they'd recognize T-antigen. So, if you look at T-antigen recognition by ELISA, one can see that only the peptides, at least in BALB/C mouse, that represent the carboxy terminus of T-antigen were capable of recognizing T-antigen by ELISA.

And, by Western Blot, these are Western Blots of BALB/C, black 6 and the cross F1s, with preimmune or irrelevant protein immunization, T-antigen immunization, in the six peptides here again you can see by Western Blot that the serum from peptides representing the carboxy terminus were able to recognize T-antigen by Western Blot.

And, I'm not going to show you the data, but we also have flow cytometry data demonstrating that these peptide-induced sera also recognize the service of SV40 transformed cells as well, as through anti-T-antigen serum.

So, if you do the challenge experiments, of course, the alum is the negative control, 5077 is an HIV peptide synthesized in a similar manner, KLH coupled and immunized as a control, and then the rest of the peptides, you see that only those peptides that were recognizing -- induced responses that recognized T-antigen by those other assays were capable of protecting animals. In repeated experiments, only about 50 percent of the animals would come up, and in a sort of side experiment added on here we wondered if we combined these two peptides ad mixed we might increase the survival rate, and we didn't, it still was around 40 to 50 percent.

Now, we were concerned that we didn't have any data to demonstrate CTL activity, and in light of what is known in tumor immunology we decided we would take another approach, and that is introduce the antigen as an intracellular source by inoculating with the gene, and the studies that prompted this were some of the studies in influenza virus and HIV showing that you can inoculate animals with naked plasma DNA, generate humoral cellular immune responses, so we did this in a vaccine protocol similar to the protein. And, we started following it by antisera, generation of anti-T antibodies, and surprisingly, we couldn't generate -- we couldn't demonstrate a T-antigen antibody response when they were inoculated with the plasma. So, we weren't sure it was working when compared to animals inoculated with the protein.

So, we went ahead and did the challenge experiment, and as you can see the animals were protected. And, just briefly, the O represents the recombinant T-antigen and alum, control collectively represents alum, saline alone, controlled peptide, controlled T-antigen, or controlled plasmid, which is PSV2 neo, which is this plasmid less the T-antigen gene. And, if you IP inoculate with the gene, you didn't get protection, but if you IM inoculated the red squares you got protection that was very close to that with the protein.

And, we demonstrated then in these animals after one in vitro restimulation splenocytes from animals receiving the DNA that we do get a CTL response, and probably the precursor frequency is low, which is the reason the responses really aren't that high, but if you do increase the E:T ratio you can see that the CTL responses do increase in BALB/C mice, and they are MHC restricted, in that they don't recognize black 6 targets which are here. And, T-antigen again immunized mice didn't generate a CTL response.

So, just to sort of summarize, we came up with a dichotomy of mechanisms in this one model of what might be happening to protect animals in vaccination strategy, and it seemed that if we used soluble protein or peptides we had a preference of a humoral immune response, although we are not convinced the CTL aren't playing some kind of role.

And, in the DNA immunized mice, we couldn't demonstrate antibody responses, but we could demonstrate CTL activity, so certainly there is a cellular response involved in the DNA vaccination or the gene therapy.

So, the conclusions are that T-antigen based vaccines that are synthetic do induce T-antigen specific immune responses that are protected from tumor challenge, and there's an evidence in this model for a role of cellular and humoral immunity, which makes the model unique for asking questions about specific humoral and cellular immune responses that might be of interest in any given human scenario.

And, it also left us with some questions, and, that is, what role do the CTL play in the soluble protein vaccination, because certainly other models have demonstrated, and will the CTL peptide epitopes generate protective tumor immunity. And also, I mentioned before about treating tumors, and we're working on models to do that now. And then, of course, clinical relevance, which is the relevance for this meeting, and, that is with reference to malignant pleural mesotheliomas in humans, and the questions are, do MPM patients possess CTL with specificity for T-antigen in the peripheral blood, and if they do, do these CTL recognize SV40 T-antigen positive tumors in an MAC restricted fashion. And, if you could do that, then potentially you could identify those epitopes that they are recognizing, and use those to generate either ex vivo lymphocytes to use in ad optive therapy transfers in patients, or as recombinant vaccines, as I showed in the animals, to actively vaccinate patients, sort of as an adjuvant to surgical resection.

Some other questions that I had put up here that I won't talk about but just mention is the concerns of tumor suppression or immune suppression, as was pointed out earlier, as well as those tumors that are T-antigen negative.

And, I just want to mention a few people that were involved in this. All of this data, in fact, was generated in the laboratory of Doctor Ronald Kennedy in Texas by Michael Shearer and myself, and Michael Shearer and Doctor Kennedy have since moved to Oklahoma. Doctor Lanford and Doctor Beames were involved in various aspects of the murine tumor model, and more recently Doctor Pass at Wayne State and the Karmanos Cancer Institute in looking at human immune response to the T-antigen mesothelioma patients.

Thank you.

(Applause.)

DOCTOR KELLY: Can we have the lights up?

Any questions for Doctor Bright?

AUDIENCE: I have some questions and some remarks. What you should do in your animal models is avoid alum at all costs.

DOCTOR KELLY: Can you speak into the microphone a little better? I don't think you can be heard.

AUDIENCE: What you should in your animal models is avoid alum at all cost. In general, it cuts your CTL responses in half, and in some models you don't even get CTL response with alum-based vaccines.

So, if you would take your protein in IFA, I would bet you would get your CTLs.

DOCTOR BRIGHT: We haven't tried that. The impetus for alum was the hope, at the time we did these models, there was obviously not a definite -- there wasn't any clinical relevance at all. And, it was one of these scenarios where you have a model that relevance may have shown its head later, was that if something did show up that we would already have data on an FDA approved adjuvant. That was the reason for the alum, certainly, and that's something that we like to look at, because I'm not convinced entirely that there aren't CTLs that are induced in that system.

AUDIENCE: Yes, there are FDA approved or adjuvants like montenile, which is a mineral oil, which is now also allowed in the clinic, and that induces good CTL responses also in mice, so that's a good choice.

And, a question is, have you looked to an CD4 mediated killing?

DOCTOR BRIGHT: Aside from proliferation data on splenocytes, not looking at specific subtypes, T-cell subtypes, not specifically, other than we have looked at briefly T-helper profiles as cytokine production in this thing. We looked at aisle 4, aisle 5, interferon gamma in aisle 2 production in these lymphocytes in vitro that were immunized, and it points toward a T-helper 1 type response, even in the protein immunized animals, which is another reason I'm not convinced that there aren't CTL there somewhere.

AUDIENCE: Okay.

DOCTOR KELLY: Okay. I think we should adjourn at this point, and I thank all the speakers for a really excellent session.

We'll take our coffee break now and maybe come back in in about 15 minutes.

(Whereupon, at 3:32 p.m., a recess until 3:54 p.m.)

Return to Agenda

DOCTOR WEISS: All right. Can you hear me? Is this on? No. That's better, I think. Talk into it, ah, bend down. Okay.

Well, I commend those of you in the audience for staying until the last session, and that we have a full quorum up here on the panel. There's one change from as advertised, George Klein had to leave to catch a plane, so he's not with us. Michele Carbone felt he hadn't had enough exposure at this meeting, and had more to say, so he's substituting for George. We hope you'll bring us the same depth and breadth of experience in viral oncology that Doctor Klein might have brought to the panel.

We have a few specific questions to ask. I'm not sure they will be answered, and then to wrap up, really, and see what we've learned over the last two days, and what we need to do next.

I don't think that -- I'm not confident that we'll be able to satisfy any of you individually or as organizations, because I think we've seen, really, that there are still far more questions than answers.

Before we get into the discussion as a genuine panel, one or two panel members wish to present some data, and I think, Michele, you are one of them, and you wish to present something that you haven't already.

DOCTOR CARBONE: After what you said about the exposure, now you are making me very scared to go up there.

DOCTOR WEISS: Well, have you loaded some slides, or are you going to just talk?

DOCTOR CARBONE: Yes, I have.

DOCTOR WEISS: Okay. If there's slides from Doctor Carbone, those are the ones that will come up first, so come and speak.

Now, what is the point you are going to address?

DOCTOR CARBONE: I'll pass this first slide, since you have seen it too many times.

Since this session was about mechanisms, I was going to present some new data that we have recently developed in the lab, data that are not published, but data that we have recently developed, and that I could fit well into this discussion.

We found this, as has been discussed here, that T-antigen or something like a T-antigen protein that is present in mesotheliomas. So, whatever it is, if it's T-antigen or SV40 T-antigen, or something similar to it that we are unable to distinguish, and anyway, what we know about T-antigen, and you heard that a lot today, is that it mediates transformation through many mechanisms, some of which are summarized in these slides, and they are its ability to bind several tumor suppressor genes, its ability to induce IGF-1, that it's apparently very important in the process of transformation of T-antigen, and then its ability to induce random chromosomal alteration. These are some of the activities that are related to its transforming properties.

Now, if you assume that there is a T-antigen in a tumor cell, that you may think that some of those things could happen, and so what we wanted to do, after we convinced ourselves that there was a T-antigen like protein, whatever it is, into these tumors, was to see whether something happened into these tumor cells.

And, we decided to start to check about the possible relationship between T-antigen and P53 in mesotheliomas, and the reasons I hope will be good in the following slide.

Briefly, P53 -- wild type P53, as everybody probably knows, is a tumor suppressor gene. The half life of wild type P53 is about 20 minutes, and it is undetected usually by immunohistochemistry, and cells containing wild type P53 are sensitive to -- chemotherapy because one of the functions of P53 is that its able to recognize mutations in a cell, and so will prevent the cell from dividing and actually the cells will go into apoptosis and die.

So, mutated P53 in study is considered an oncogene, has a prolonged half life, and so becomes detectable by immunohistochemistry, and in many labs of pathology the detection by immunohistochemistry is used as a quick and quite sensitive assay for P53 mutation.

Mutated P53 is associated with various aggressive tumor phenotypes, and cells containing mutated P53 received -- chemotherapy, and the idea is that now the cells basically will not stop division because P53 will not block a cell that has a mutation, and so the cells will go on and eventually going on may accumulate other mutation and become even more aggressive.

T-antigen binds and inactivates wild type P53. The complex T-antigen P53 is relatively stable, and so sufficient amount of P53 accumulates in the cell that can be detectable by immunohistochemistry, and that T-antigen does not bind mutated P53, so another mechanism by which you can see P53 in a cell if it is bound with T-antigen.

About P53 human mesothelioma, what is published is that P53 mutations are rare in human mesothelioma cell lines. Human mesothelioma cell lines contain high levels of wild type P53, detectable by immunohistochemistry, and mesotheliomas are completely resistant to therapy.

If you follow what I said before, evidently there is something that doesn't fit into this picture, because Y mesothelioma, one of the most aggressive human cancers, contains high levels of wild type P53, and -- high levels of wild type P53 enable the oncogenic phenotype and induce sensitivity to therapy. So, obviously, since the T-antigen like proteins appear to be present into some mesotheliomas, one possibility was that in some of these cases T-antigen could maybe be bound to P53, and then be responsible for death.

So, we studied the relationship of T-antigen and P53 in mesotheliomas, and these are the results that we found. They are summarized here. We can assume that P53 mutations are very rare in mesotheliomas, at least compared to other tumors. By SSCP analysis, we found that P53 wild type in 27 of the 31 -- studies, then 31 of the 52 mesotheliomas expressed P53 by immunohistochemistry, and 32 of the 52 mesotheliomas expressed T-antigen by immunohistochemistry.

Now about that, I don't have a slide that says that, but I would like to say that while there is a stronger statistical correlation between the co-expression of the two, there were samples in which we detected T-antigen without detecting P53 and vice versa.

And then, RNA in situ hybridization experiment that demonstrated in some specimens co-expression of T-antigen and P53, and I'll show that to Doctor Shah, with whom we discussed it yesterday. This is one example of the P53 mutation that we detected in one of those tumors.

This, obviously, is the best or one of the best pictures that we had, so I wouldn't want that the take-home message is that that's what you will always find in a mesothelioma. However, in this particular mesothelioma, or in some of them, we have a control on the left, a T-antigen staining and a P53 staining. Those are not sequential sections, so that is not co-expression of T-antigen and P53 in the same cells, but it's the same tumor, and the cells stained for T-antigen and for P53 monoclonal antibodies.

If you make an immunoprecipitation with antibodies against T-antigen and P53, on the right you have a control that is the hamster mesothelioma cell line that we derived from one of the tumors that I showed you before, that you precipitate a T-antigen like protein with a nine kilodalton molecular weight, a 17 to 19 like protein that is in the range of small T-antigen, and a protein range at 53.

This is an RNA in situ hybridization showing in this particular case that the same cells that have a message for T-antigen have a message for P53, and they are indicated by the arrow. You can also see on this slide that there are cells that -- actually, you can see on the right that all the cells in this particular case have a message for P53, only a portion of the cells have a message for T-antigen, not all of them.

Finally, Waf-1, or P21, or chip 1, as you prefer to call it, is in use by wild type P53, so you expect to find expression of Waf-1, if you have wild type P53 that is biologically active. And, we were unable to detect Waf-1 expression, except then in one of these specimens, and for this we just used immunohistochemistry and we used two reagents, one that was an antibody that was provided to us by Doctor Apella, and the other that is an antibody that is commercially available.

And, the conclusion that I could make out of this is that, in some of these mesotheliomas there seems to be a protein that we cannot distinguish from T-antigen, at least with the reagents that we have, that is detectable by immunohistochemistry and that seems to be associated with P53 because we can precipitate the two together.

Given the role of P53 in controlling cell division, it is a possibility that if the like sequences are present in tumor cells they may be able to interfere with the function of P53.

And, that's it. Thanks.

(Applause.)

DOCTOR WEISS: Thank you, Michele. Come and sit down.

So, there we have some evidence and some beautiful immunohistochemistry that would indicate that there's P53 sustained expression, possibly wild type, and T-antigen expression in one mesothelioma.

If SV40 is causally related to these human mesotheliomas, then we should expect the majority of them to have a picture, as Michele Carbone has shown us. In that case, they should be positive by Southern Blotting for the T-antigen gene, and should be very readily detectable.

So, I think it's important to know in these analyses, and I understand they may be preliminary, really how many of these tumors give this kind of picture, because if this is a one of, it may be a case history, and what this workshop is really about is whether there's an epidemiological case for an association of SV40 with these tumors.

Would you like to comment on that?

DOCTOR CARBONE: I'm not sure exactly what's the question, because while I was working here you had started to talk, and I was not listening to you, sorry about that.

DOCTOR WEISS: Is this a one of case --

DOCTOR CARBONE: Excuse me?

DOCTOR WEISS: -- the example you've shown, is that one tumor or is this a general case with mesothelioma?

DOCTOR CARBONE: No, it's obviously not one tumor. What we found was that out of 52 mesotheliomas that we found, we found that 31 or 32, I don't remember exactly what it was, 31 or 32 you could detect by immunohistochemistry wild type P53, I mean, P53, and then you could also detect the T-antigen.

The percentage of positive cells will vary in different tumors, and there is a considerable variation between different tumors, for which I do not have an explanation, and that is true both for P53 and for T-antigen.

The thing is more complicated by the fact that in mesotheliomas it's very difficult to distinguish between malignant cells and non-malignant cells, because the biphasic appearance of mesothelioma, that means that some cells have an epithelial appearance and some have a fibroblastic appearance, is very difficult to understand where you are dealing with reactive stromal tissue that is normal tissue, or whether those are, in fact, malignant cells that represent tumor cells.

So, given this limitation, there is a number -- a different population of cells, some of which will stain positive, and some of which will stain negative.

I don't know if I answered it, what was the question.

DOCTOR WEISS: So, we know that 31 or 32 tumors out of 52 have at least one or two cells that express these antigens, and that you showed us the best example.

DOCTOR CARBONE: How many of these?

DOCTOR WEISS: We still have no idea --

DOCTOR CARBONE: Yes.

DOCTOR WEISS: -- what proportion of cells express these antigens out of those 31 tumors. But, as you've told us, it's actually quite difficult, at least for your pathologists, to tell the difference between a tumor cell and a stromal cell.

DOCTOR CARBONE: If we take the epithelial component that is the easiest one to identify the malignant component, usually you have a proportion of cells that can vary between 30 to 50 percent that stains positive.

AUDIENCE: I think -- let me just say that there was a specific grading system for this immunohistochemical stuff that was done, and Michele didn't look --

DOCTOR WEISS: Can you speak closer to the microphone?

AUDIENCE: -- yes, I sure can. Michele didn't look at the slides. I mean, these were blinded, first of all.

Second of all, the grading system was less than 25 percent, 25 to 50, and greater than 50 percent positive, as well as a strength of grading from zero to four. And, we didn't even call a specimen positive unless it was greater than two, and we didn't call specimens positive unless greater than 25 percent of the cells were stained.

So, it's not like one or two cells in a specimen, as you intimated, it's a specific grading system that's standardized for immunohistochemical techniques that are, you know, qualified all over the world.

DOCTOR FRIED: Could I ask, did you look for --

DOCTOR WEISS: This is Mike Fried speaking, for the record. Carry on.

DOCTOR FRIED: -- does BK cross react? I mean, could there be BK T-antigen you are detecting, or JC? I mean, those things are there also, are they not?

DOCTOR CARBONE: Yes. I do not know, I cannot exclude the possibility that that would be BK or JC T-antigen, because from what I understand there is cross reactivity among different monoclonal antibodies.

DOCTOR FRIED: But, I mean, from what you've done, how many of these tumors contained BK or JC? I mean, some people --

DOCTOR CARBONE: I never looked into this tumor for BK sequences.

DOCTOR FRIED: And, how many of the tumors could you immunoprecipitate? Did you do then -- of the immunoprecipitate, did you do a Western with anti-P53 to actually demonstrate that that band was P53?

DOCTOR CARBONE: Yes. We used that, from five frozen mesotheliomas we were -- we used only five frozen mesothelioma tissues, and those five frozen mesothelioma tissues that were positive for SV40 like DNA, we were able to precipitate this protein.

DOCTOR WEISS: Yes?

AUDIENCE: I'm just wondering, Michele, out of the 52 samples, how many were positive for both elevated P53 and the presence of T-antigen?

DOCTOR CARBONE: The answer to that is not -- I don't remember the exact number, I remember another thing, that is that the statistical correlation between the two was very strong, and was 0.0001 or something very close to that, but exactly the number of that I don't remember.

DOCTOR WEISS: Okay.

DOCTOR CARBONE: Can I say another thing, I want to add to that that I didn't show there, and that is something that at present we do not understand. We have seen, in tissue culture, in cells that are derived from these tumors, that you can sometimes stain these cells for T-antigen, I mean, with a monoclonal antibody against T-antigen, these cells would strongly stain against a monoclonal anti-T-antigen, and that while you pass these cells in culture, the number of the cells that is positive keep decreasing. And, I, at the moment, do not have any explanation for that, and if anybody has an explanation for that I would love to hear it.

DOCTOR WEISS: You did mention, Michele, that 27 out of 31 analyzed had wild type P53, so that suggests that four of those 31 did not have wild type P53. Perhaps, those stained immunohistochemically, because they have mutant P53 but no SV40. Have you really put this together and worked it out?

DOCTOR CARBONE: Yes. The problem there is that in my experience it's very difficult when you use an antibody against P53 to really be -- there are antibodies that are against wild type P53 and antibodies that are against mutated P53, but at least in my experience it very often happens that with an antibody against wild type P53 you get both or vice versa.

In any case, in this particular case we used the antibodies of oncogene signs, and the one that you saw before was one that is called AP6, that would be able to recognize both, the wild type and the mutated P53.

So, we would not be able to discern on an immunohistochemical staining whether that was wild type or that was mutated P53, so the immunohistochemic stain will assume that that is mutated P53, if it detects it, or is it P53 that is there at higher levels for whatever other reason.

Does that make sense?

DOCTOR WEISS: Yes, but I may have misunderstood you, I thought you'd sequenced 31 of these tumors and found that 27 were wild type.

DOCTOR CARBONE: No, we analyzed them by SSCP analysis for -- five and nine, and there is where you find 95 percent of mutation of P53 usually in human tumors. And, of those, 27 were wild type, and the others instead contained some detectable mutation. DOCTOR WEISS: Okay. I think there are other members of the panel who wish to present some data briefly.

Antonio, would you like to go next, Antonio Procopio.

DOCTOR PROCOPIO: Okay.

Please, the first slides, both, there are two sets. Okay.

I would add to the discussion of the panel just three main questions with what we have in mind. The first is whether or not the evidence of the possibility of amplifying SV40 like sequence in mesothelioma could be of importance for diagnosis. The second is why we have any doubt about prognosis, the possibility that these data can have a prognostic value, and the third, if they could affect in a certain way the therapy, the therapeutic strategy of mesothelioma.

These questions have be raised to me in several meetings, since in Italy there is a great concern about mesothelioma, since we have more than 1,000 people dying of mesothelioma each year.

And, to my experience in the first paper with Michele, with Harvey Pass, and also from experience with other samples in Italy, we have never been able to find the positivity of these samples for peritumor and normal tissues, or from -- the next slide, please -- or, from cancer of other sorts invading the pleura.

On the left side, we have screened and blinded a number of different specimens come all over -- and we did not find any positivity on cancers that were coming from not mesothelial origin.

Instead, the lower line we have found positivity in mesothelioma samples, but this positivity was changing from 20 to 60 percent, depending from the institution which was sending us this material.

And, we found also some positivities in pleural involvement by tuberculosis. We did not know why tuberculosis should make positivity for SV40-like sequence, so we went back to the samples and we tried to sort out what kind of -- where positivity was assumed to be and what theoretically the material where the sequence could be amplified.

And, this is a slide that shows the tuberculoma that was heavily calcified, and on the top of it you can see that there is an activation of mesothelial cells that we call the mesotheliosis.

When we cut out this piece of material, we amplified that this was positive, the rest of the material was always unamplifiable.

So, we think that this could be the first evidence of non-neoplastic mesothelial associated, in fact, involvement in SV40-like sequence expression.

Next slide, please.

The second question was, can we use these, why do we have any possibility to sort out whether this evidence are prognostic. We have 86 mesotheliomas, and we have not seen any difference statistically significant in SV40 positive and SV40 negatives surviving in these tumors.

One possible explanation is shown here on the right side of the panel, we have tested lac activity against the mesothelioma cells in the presence of the pleural exudate. There is a large body of literature and findings that these, of course, will suppress any cytotoxic activity due to the heavy production of TGF beta 1 or the production of sulubri fibronectin.

In the other slide instead, we show Tunicin expression into mesothelioma, Tunicin expression is in the right part of the slide, and this is the part that is neoplastic. Tenacina is able to strongly suppress T-cell receptor mediated activation of T-cells, so we have to take in mind that we have a stronger immunosuppressive environment in which mesothelioma may develop.

Next slides, please.

Okay. We have tried to see whether we could link SV40 expression with any morphological phenotypical evidence of difference within the tumors, and up to now we have been not able to see any difference in the production of extracellular matrix, or integrating expression, or proliferation, or differentiation in the chemotactic and apoptotic ability of mesothelioma cells and the mesothelioma cell lines and mesothelioma primary cultures.

But, we were able, as Michele told before, to immunoprecipitate from a number of mesothelioma tumors T-antigen by using P53 antibodies. This is an experimental co-immunoprecipitation. However, we have not been able to do the opposite. We tried several times, but we were unable to immunoprecipitate P53 antigen from the tumor by using an anti-Tag antibody.

Next slides, please.

Okay. What we have done next was to try to use a more direct approach to the problem, and by use of adenovirus that have been raised in our lab in collaboration with a group here at NIH, we have infected the mesothelioma cell lines with adenovirus that have a bearing cDNA for wild type P53. And, of course, we were able to see a dose dependent increase of P53 expression within the cells, and also buf 1 expression, and also BCL2 became apparent when the lytic unit of virus were increased.

Next slides, please.

When you, in fact, with the related gene by gene therapy with P53, some cells you may expect cell death by apoptosis most of the time, terminal differentiation of loss of malignancy.

The slide on the right shows that we are able, with this virus construct, to inhibit most -- the whole -- all four cell lines, mesothelioma cell lines that we have been able to use.

Next, please.

This is following which is showing the effect on the right of an adenovirus with respect of the control.

Next.

And, this is, instead, the effect on soft agar tissue culture, in which there is no growth in the infected cells on the right, and with respect to the uninfected on the left, or the control virus infected cells.

Next.

And finally, these are the results of in vivo studies. We have done several different studies by injecting mesothelioma cells, human mesothelioma cells into nude mice after or before infection with adenovirus, and what I can say is that the results are pretty impressive. We can block proliferation of adenovirus, of mesothelioma cells in vivo, and the more impressive results are in the group that's been injected intrapleurally, where the survival is also -- has become double or have multiplied three times.

(Applause.)

DOCTOR WEISS: Thank you, Antonio.

Any questions or comments on Antonio Procopio's presentation?

AUDIENCE: I realize the numbers are still very small, but is there any correlation between gender of patients, age of patients, anything that would indicate there's a different distribution of the mesotheliomas that are T-antigen positive and those which are T-antigen negative? Anything?

DOCTOR PROCOPIO: We have only analyzed 86 patients, so we have a very tiny group of data, and we have only data up to now for an Italian statistic, with expression with paraffin embedded tissues. Okay. So, we cannot see anything about T-antigen, because there is no reagent working now.

But, I think that we should have a large number of data before going to address carefully this evidence, because there is a wide variability of approach, clinical approach to these patients, and it is very difficult to standardize everything.

DOCTOR WEISS: Jim Goedert?

DOCTOR GOEDERT: Slightly off the topic, but I was wondering if Doctor Procopio or Doctor Carbone, or others who have examined mesothelioma tissue, have looked for other infectious agents in the mesothelioma tissue, and I'm thinking in particular of Human Herpes Virus 8, which appears to have a particular tropism for the pleural and peritoneal tissue.

DOCTOR PROCOPIO: I have only the data, we have also checked for JC and BK, all the data I've shown, the groups, the samples that I've shown. In that sample, sir, there was -- in the samples I've shown there was not JC or BK infection, though some other samples could be -- this virus could be amplified. So, they don't fit.

DOCTOR GOEDERT: But, not for HHV-8.

DOCTOR PROCOPIO: No.

AUDIENCE: In the nude mouse experiment, was the presence of P53 in the adenovirus necessary, because you are putting in a human -- a virus that will only grow and kill human cells, so, I mean, if you just put wild type adeno would it have done the same thing?

DOCTOR PROCOPIO: The cell lines had the wild type P53, yes. And, we have used a number of cell lines for these studies.

AUDIENCE: Yes, but if you just put adenovirus in.

DOCTOR PROCOPIO: Oh, no, okay, we had the null adenovirus virus, or we had the adenovirus with betagal insert. Okay. The adenovirus with betagal insert were a bit toxic. Okay. The null was not, and there was no increase of P53 expression, and there was not any of this evidence seen.

AUDIENCE: You were showing that the tumor was destroyed in the nude mouse, right?

DOCTOR PROCOPIO: Yes.

AUDIENCE: So, I'm just saying that since you are putting a virus that kills -- for human cells into a mouse, does it have to have P53 to kill the human cells when it's in the mouse? I mean, if you just put regular wild type adeno, you might have also killed the cells.

DOCTOR PROCOPIO: Regular adenovirus were not working, this is -- I don't know if I answered it carefully. The control virus was not effective as the adenovirus with P53.

AUDIENCE: It seems that one of the questions that keeps coming up is whether SV40 can co-localize in sites of inflammation, and, perhaps, selectively infect and be expressed in cells that are inflamed. There was data from France presented yesterday that said that 16 percent of reactive pleura had SV40 sequences. These data that tuberculous pleuritis might be sites where you find SV40 sequences, Michele's data that suggests that 60 percent of animals that were inoculated by the intra cardiac route, despite George Diamandopoulous' data that intravenous injection didn't cause mesotheliomas, makes me wonder whether it would be pretty easy to hit the pleura when you are trying to hit the heart in a small animal. In fact, I'm sure it is.

So, it seems like an experiment that could be done would be to do George Diamandopoulous' experiment to shoot SV40 intravenously into a hamster and then cause some kind of reactive pleuritis and see if you end up with SV40 in the pleura. You know, does virus that makes it into the intravenous system somehow tend to like to go to places at sites of inflammation, and could that be a cause for finding sequences, without necessarily invoking SV40-induced transformation as a cause of tumor?

DOCTOR CARBONE: Jim, I would like to answer. I asked myself when we found the mesotheliomas how it was that Diamandopoulous didn't find the mesotheliomas. He's a pathologist, and certainly he would have recognized the mesothelioma. And, I have my theory of that, and my theory of that is that when you do an intra cardiac injection you have to go through the pericardium or through the pleura, if you make a mistake, for sure you go through the pericardium. And, so that, the virus went into the pleura directly from the needle. Of course, I have no proof of that, I'm just saying what I think. But, what I think is that I have the syringe, I put the syringe inside the tube where I have the virus, now I stick the syringe into the heart of the animal, and now I'm getting a mesothelioma.

Now, if that is true, that means that the amount of virus that is needed to induce a mesothelioma is much, much lower than the amount of virus that is needed to induce a sarcoma, for example, and I hope that I will be able one day to do the experiment and see whether this is true.

But, that would be my explanation for that.

AUDIENCE: Well, it might be less or it might be more. You don't know how much virus that you shot intra cardiac ended up at an end site where the transformation occurred.

So, if you, like the Bernice Eddy studies, where no matter where you injected animals with transformed cells they always got subcutaneous tumors, whether it was intraperitoneal or intra anything else, they ended up with little subcutaneous nodules at the site of inoculation. It suggests -- I realize those are transformed cells, but it suggests that you might not take very much at the site.

But, I'm getting at the point of SV40 trying to find a site that it can grow, that it can replicate, or at least express some genome, that you might detect subsequently.

And, these interesting ideas about the pleuritis being sites where you find SV40 sequences just seemed like they would suggest an obvious experiment to do.

AUDIENCE: If I can follow on to that, I think, excellent comment, is, one of the things to note, that in the original preps of the monkey viruses there were cocksacky and other adenoviruses in there as well as polio.

And, to add to that, if I could -- since we're going to try to do your experiment, because I think it was a good suggestion, is create inflammation by adding a mouse cocksacky like cardiovirus with the SV40 as a way of doing co-infections, see if it does change the distribution. I thought that was a good comment, and I think that simultaneous infections are important, because there were about 26 monkey viruses in those preps. We are only focusing on SV40 here, but there were the other viruses, too.

DOCTOR WEISS: That will give the lawyers a lot of work to do, I think.

Harvey Ozer?

DOCTOR OZER: To have a brief comment, some of the things I was going to say were actually preempted by the data that were presented by Doctor Carbone, so it was really the posture that one should be able to do those experiments and find those results.

But, the other things I would like to remind people, of what Doctor O'Neill said in our own data and those of others, that SV40 does replicate in human cells, and it replicates its DNA quite well, even if there's a low level of virus production.

And so, one might, in fact, expect that there would be a negative selection against wild type virus under conditions of a longstanding infection, such as the induction of a tumor, if that's what it's postulating.

And, I wonder, I would not be surprised, therefore, to find mutant viruses, rather than wild type viruses, if that, in fact, were the mechanism.

The other thing is that that may be relevant to your comment of finding your tissue culture cell lines losing viral sequences, because if it were a cytopathic virus, one might clearly expect that sort of result.

And so, I would remind people that we are again talking about a virus infection in this system.

DOCTOR WEISS: Okay. Thank you, Harvey.

Who else wants to present some data? Jim Pipas? Have you got the slides for Doctor Pipas to project? I can't see any response from the projection booth, so, perhaps, the answer is no.

Have you anything to say, Jim, without slides?

DOCTOR PIPAS: Why sure.

No, we've just heard a lot of discussion during this meeting about recombinant viruses, and so, wanted to remind you of some work we did some years ago looking at the ability of SV40 to form viable recombinants with a close relative, SA12, which is a baboon virus.

Now, some time ago we showed that SA12, by sequence analysis, is closely related to the human virus BK, and we did a number of studies of two types. One is to try to mark or rescue SV40 mutants, either in T-antigen, or in the late region, with fragments from the SA12 genome, and this yielded a number of recombinants, viable recombinants. These were selected for viability on monkey cells that contained hybrid T-antigens and hybrid late regions.

The other method we used was to make chimeric dimers between full length SV40 and SA12 genomes. These molecules are too large to be encapsidated into SV40 virions and monkey cells, but if we infected with these chimeric dimers that they would resolve by recombination to yield an array of recombinants, showing that you can replace both the early regions and the late -- you could reciprocally transfer the early regions and the late regions of these viruses and make a number of viable hybrids that were within the T-antigen gene.

So, at least with these closely related viruses, the efficiency, I think, is hard to measure, and it may be low, but it's certainly possible to generate an array of viable recombinants with either large substitutions or small substitutions of the different viral sequences.

DOCTOR WEISS: Are you suggesting that this happens in the natural history of infection?

DOCTOR PIPAS: It's hard to say, because the efficiencies that we were measuring were very low, so it's hard to tell whether in a natural setting you would get these recombinants.

All I can say, at least between these pairs, and I suspect between BK and SV40, it's possible to generate viable recombinants, and an array of them in different parts of the genome.

So, in theory, yes.

DOCTOR WEISS: Thank you.

It seems to me that we've heard very ample evidence that SV40 is a potentially oncogenic virus. We know how oncogenic it is in hamsters, provided at a very high dose, and the dosage of infection that might have occurred, say, through polio vaccine is much less.

On the other hand, one might need a high dose in hamsters because it's a non-permissive system and non-replicative. And, we don't really know what degree of amplification we would get in human infection.

But, it still seems to me unclear whether there's a causal relationship between SV40 infection and human tumors. We heard on the first day that there are really polarized views as to whether SV40 is genuinely present. Some groups just don't find it, others do, and we've seen the evidence presented. And, we've heard the epidemiological analysis that there's no significantly increased relative risk for the types of tumors that are thought to be related to SV40 in hamsters or in humans, when one looks at -- does a cohort analysis in terms of both cohorts, both in terms of the analysis in Sweden and the analysis in the States.

I'd like to ask Howard Strickler whether he could expand on that a little. We heard earlier this afternoon that T-antigen can elicit very strong CTL responses, and both the talks on cell mediated immunity were sort of conceptually directed towards thinking of immunotherapy or immunological protection.

But, let's turn that around to an epidemiological question. If, like other human tumor viruses, the expression of a viral antigen in the tumor is antigenic, then in immunosuppressed populations you would expect a higher instance of these tumors.

So, I'd like to ask Howard, or any other epidemiological colleagues here, whether in AIDS patients, such as the analysis of the MACS cohort, whether in the registry of tumors in transplant patients there is any increase in the types of brain tumor, mesothelioma, osteosarcoma that we've been discussing over the last two days.

DOCTOR STRICKLER: I think the perfect person to respond to the issue of HIV to development of these cancers just stood up, which is Jim Goedert, but I'll take the opportunity to say that I think that the issue of causality, and how to properly assess that, is something that we need to think about more soberly, whether the detection of the virus in some tumors, and also normal tissues, takes us very -- how far that takes us down the line, and think about the questions which we need to address in order to look at the issue of causality.

And, I'm not going to bother here listing the different elements of that in terms of their general principles, but I think that I'd like to remind all of us, we've yet to even show that the virus is specifically in the tumors by in situ hybridization or other methods like that, show that the virus is in all cancers. So, in terms of the strength of the association, I think that that's yet to be fully worked out.

And, I think that we need to understand the issues of the transmission in the general population. So, I will leave it that I think that we need to be very careful in what we say regarding the issue of causality, and I'll leave it at that at this point.

DOCTOR WEISS: Thank you, Howard. I quite agree, the principles of causality are complicated. We haven't quite got as far as satisfying cross postulates yet. Recently, in reference to HHV-8 and Kaposi's sarcoma, I tried to elucidate some modern postulates, and my first one was that every virus needs its Duîsburg, that we need the challenge to the data as well.

DOCTOR STRICKLER: I would hate for that to become --

DOCTOR WEISS: No, don't take that personally.

DOCTOR STRICKLER: -- I'm really just trying to raise an issue of caution, and I think that it's important at this point, and I just have been sitting here waiting for the issue of causation to come up and someone to address it, and part of me wanted to address it in a formal way and go point by point, strength of association, and on down the list, the specificity and so on, but there's some very compelling data at this meeting, people are finding it in tumors, and I do not wish to be the Duîsburg. This is an important topic, but I think a lot of people are talking about cause in this, and I would like to see caution in the tone.

DOCTOR WEISS: Thank you. I think we'll come back to causality.

Jim is going to answer the specific question, whether there's an increase in these types of tumors in immunosuppressed people.

DOCTOR GOEDERT: The analysis is still preliminary, hopefully, to be done shortly, in terms of other malignancies besides non-Hodgkin's lymphomas, Kaposi's sarcoma, that are increased risk among persons with AIDS.

In the pediatric population, which you'd expect would be at most vulnerable risk, it looks like the AIDS-associated malignancies are entirely confined to Kaposi's and to two that look like they are associated strongly with Epstein-Barre virus, which is non-Hodgkin's lymphomas and leiomyosarcomas or benign leiomyomas.

In the adult population, there's a little bit more of a diverse spectrum. The risk of these things among persons with AIDS are far, far less than Kaposi's and non-Hodgkin's lymphomas, but they do, in some respects, look like they are virus-associated tumors, and, in particular, I'm thinking of anal cancer, which is clearly excessive on persons with AIDS, and it looks like it's not only due to homosexual practices, and Hodgkin's disease, which is looking more and more like, at least in part, an Epstein-Barre virus malignancy.

There's a couple others that are lower on the spectrum that are hard to explain. We have not seen ependymomas, either in adults or, as I say, in the children. We also have not seen, but haven't looked real closely for, mesotheliomas. There have been excess risks of other things that I suppose you could make a stretch, and I'm thinking in particular of adenocarcinoma of the lung and leukemias, particularly, lymphocytic leukemias, appear to be excessive.

And then the last one, which I think as I listened to the last almost two days of meetings, raises some concern on my part, is there does appear to be an increased risk of malignant gliomas of the brain among persons with AIDS, and I think, as I say, the magnitude of that is not very large, but it is statistically significant, and I think it's one thing to look at.

DOCTOR WEISS: Thank you.

Next question.

AUDIENCE: One thing, you know, going back to the basics, there's something that we really should remind ourselves, and that is, none of these papova viruses in nature, it's not a primary function of any one of them to cause tumors. So, in considering SV40 as a possible human pathogen we must consider that in people who are immunocomprised it could be causing other types of infection, especially lytic infections, perhaps, kidney disease or something, and those may actually be much more prevalent than seeing an association with tumors.

DOCTOR WEISS: Thank you.

We're confining this workshop to malignancies in human SV40 infection, if we want to get out tonight, but it's a point very well made.

Please.

AUDIENCE: Yes. Just a little comment for better understanding of the causality of the data, of the different data, of detection of SV40, the sequences in malignant and non-malignant tumor, I would like to suggest we organize maybe multicentric studies with anti-laboratory controls.

DOCTOR WEISS: I'm sorry, I didn't hear you clearly. Can you make a concise statement speaking right into the microphone.

AUDIENCE: Maybe it's my English, yes, for a better understanding of the causality, and of the different data and the difference studied during this meeting, I just would like to suggest we organize multicentric studies with anti-laboratory controls for the detection of SV40 DNA sequences.

DOCTOR WEISS: This is back to the question of exchanging materials for detection, which I think we agreed yesterday would be a worthwhile thing, and, perhaps, is one of the summary statements that could come out. Thank you.

DOCTOR PASS: Harvey Pass. I think that the question of immunosuppressed populations is an excellent one. I'm not so sure that I would have thought of the AIDS population as the first one to look at, but I do think there is a population that the epidemiologists could look at, and that are very relevant to the mesothelioma population.

And, I'm sure they've thought of this, but there are still plenty of people who have the diagnosis of asbestosis, and there are still plenty of people who are being followed, just as Doctor Levine had said for the children with big serum banks, that had asbestosis, that have serum samples.

And, my only question is that I think we shouldn't forget that population, because we are talking about a tumor that has a causal relationship, which is asbestos, and the question is, what is the SV40 or the virus doing at all to this. And, I would not forget that population with regard to collecting samples, or looking at when we decide upon how we are going to screen to see whether they have an increased exposure, see what the effect is in a prospective fashion and see if they develop mesotheliomas, and see then with the data whether they had a higher incidence of having antibodies to T-antigen, or however you want to look, but I think that that's an immunosuppressed population that we need to concentrate on.

DOCTOR WEISS: Thank you.

Well, the epidemiologists, at least in Europe, tell us we are just at the beginning of the asbestos-related mesothelioma epidemic, so it does look sadly as if there's going to be a lot more clinical material to study in the future. To try and relate that to possible SV40 infection would be useful. We should remember that all cancers are multifactorial, and in some cases there may be more than one route to getting the same end result.

I'd like to bring up the old chestnut, that SV40 is certainly oncogenic in terms of transforming cells, and the very elegant works that's being done with transgenic mice, and forming tumors in hamsters, but we could say the same, perhaps, quantitatively different, about other viruses for which we have a much greater confidence and a much longer knowledge, that they are natural ongoing human infections, adenoviruses and BK and JC.

So, there seems to be a paradox between the experimental oncogenicity of all these viruses, and the easel, shall we say, difficulty with which we can show that they are oncogenic or associated with tumors in their natural host.

We have Maurice Green sitting here in the audience, who spent many, many years searching for evidence of adenovirus genes on genomes in human tumors, and after an exhaustive search, admittedly in the pre-PCR days but with Southern Blotting, we couldn't really find any evidence for it. And yet, E1A and E1B are just as good at sequestering retinoblastoma protein, or P53, and transforming cells in culture, and inducing tumors in hamsters, as SV40.

Now, does this mean that all the experimental evidence of oncogenicity we've heard is irrelevant to the discussion? I'm trying to provoke some response here from the panel. Who would like to answer that?

DOCTOR OZER: Well, okay, I'll wander into it. No, I think there's a major conceptual problem that we've dealt with in thinking about the DNA viruses, as to whether they are going to be transforming in the same organism in which we have a permissive infection, and there is data from mouse with polyoma that maybe we should allude to, or Mike should, because Tom Benjamin is not here, but the point is that we see even in tissue culture that, in fact, cells that are producing virus, or, perhaps, producing very large amounts of DNA, and I would emphasize that some of these cells can produce very large amounts of viral DNA, that that may not be compatible with going on in further steps towards carcinogenesis, and that, therefore, the experiment has not been done with recombinant viruses or the defective viruses that would be the better candidates for that sort of a situation.

DOCTOR WEISS: I would have thought the rate at which replicating viruses generates defective particles, that can nevertheless be encapsidated and spread to other cells, that in natural in vivo infection there should be a sort of throw off of particles that have the transforming genes and not the replicative genes, and that those would potentially lead to tumors.

DOCTOR OZER: Well, we have a very potent immune response to at least SV40, as was pointed out, that would also be targeted at the acutely infected cells. And so, maybe it's -- and, if we believe that immortalization is a key part of carcinogenesis, and there are a lot of problems in that in the human system as contrasted to the rodent system, we may just have enough mechanisms of clearing the cells before they've accumulated enough other mutations.

DOCTOR WEISS: Mike Fried?

DOCTOR FRIED: Just to say, the defectives of polyoma and SV40, which some of them are worked on, were mainly origins of replication which were duplicated, because that is the selective advantage. If you have ten origins, you are going to replicate and you are going take over, and hardly any of them were found with the transforming genes, which were in tact. So, that was the regulatory region containing the origins, and that's the way the DNA viruses, in a replicating system, are competing, and those are the things that went out.

DOCTOR WEISS: So, does that mean we could argue, actually, to the opposite, and that we would not expect DNA viruses that are in permissive hosts, where they cause natural infections because their replication is lytic, that we would not expect them to be oncogenic in their natural host?

Well, in that case, is SV40, is the human population a natural host for SV40 infection? We'll come back to yesterday's discussion.

I'd like to, perhaps, take that argument a little bit further, and, perhaps, provoke Keerti Shah's response. Is it possible that the natural reservoir of SV40 is the human population, and that rhesus monkeys, are they sensitive indicator species, and we have the opposite of a zoonosis, we have a sort of anthroponosis. Because, as far as I can tell, there have really been minimal studies of genuinely wild rhesus monkeys that have not been in close contact with man, like the monkeys in the Nepalese temple.

Keerti has also told us that other closely related species of the same genus, macaque, like the bonnet macaque in southern India, the cynomolgus monkey, macaque fasicularis in southeast Asia, are generally free of SV40. We are also told, these popova viruses co-evolve with their hosts. There's something a little bit fishy here, so I'm maybe the only one in the room, but could if I propose, if only to see it knocked down, that maybe the human population is the natural host for SV40, and that monkeys have courted off us.

DOCTOR SHAH: Actually, Andy Lewis asked me the same question a few days ago, whether it is going in the reverse direction. I think the evidence does not support any such idea. If you look just for antibodies, I mean, looking at infection in any way you wish, the antibody prevalency with age, antibody prevalence in close population, finding the virus in the kidney, finding in urine, there is a huge amount of data on SV40 in the monkeys, and there is relatively no data, I think there is not a single instance shown which says that SV40 is circulating in the normal population. And, we did not get it here in this last two days.

And, I think if you look at these viruses, each virus is exquisitely tuned to its host, and different species, animals of different genera, do not share a single polyoma virus. I think it's -- that I don't think would hold.

And, I'm not so sure, but people like Harvey and Tom might know how the SV40 irregulatory region might be right for the rhesus cells, as opposed to BKV and JCV, which may have elements which make it grow in humans instead, this might be available, perhaps.

DOCTOR WEISS: Even so, it is captive or commensal monkeys that very clearly pass it on from monkey to monkey, but that's similar to, say, SIV, which is never found in wild macaques, but is very readily transmitted from one monkey to another in captive encaques, whereas, the natural host, African species of monkeys, and it transfers across with contact in captivity and then causes disease in the unnatural host.

DOCTOR SHAH: so, is there a part that SIV comes from another reservoir altogether, not simian reservoir?

DOCTOR WEISS: Simian, but not macaques.

DOCTOR SHAH: Oh, I see, okay.

DOCTOR WEISS: But, it was first discovered as AIDS in macaques, that they are unnaturally infected in captivity. SV40 might be the same.

DOCTOR SHAH: It is not impossible that there is another macaque species, but I really don't think the evidence will suggest -- for example, in the instances where people were exposed to SV40, those of Doctor Morris, Tony Morris' study this morning, and oral polio vaccine, there was very little multiplication, there was practically no antibody response with the oral virus, with the respiratory infection there was, again, an extremely low level multiplication and a very low level antibody response, but decreased suddenly. I think -- I mean, anything is possible.

DOCTOR WEISS: Well, at least my idea wasn't totally crazy, if Andy Lewis was suggesting the same a few days ago.

John Lednicky.

DOCTOR LEDNICKY: I think Frank O'Neill's experiments address adequately the question with regard to if SV40 can grow in human cells, you know, why might you see tumors or vice versa in monkey cells, and the reason is because not all the cells are permissive for SV40.

And, turning that question around, you know, this is the reason monkeys don't liquefy if they have SV40, right, because not every monkey cell is permissive to SV40.

DOCTOR WEISS: Thanks.

I'd like to come back to the question of the polio virus. We heard this morning, Doctor Levine summarized it, and Jan Butel gave some data, that with the serology, if we accept the specificity of the serology on face value, that, in fact, if you look at it by the age cohort of the potentially exposed population the serology positivity approximately doubled, went from somewhere around ten percent or less to 20 percent, and now it's come down again.

I wondered if there are any confidence limits on these percentages, whether they are statistically significant. Is Jan still here? Could you come to the microphone? It seems to be that we are close to accepting that SV40 may be a human infection, and may have existed before polio, and may be continuing after the polio vaccines were cleaned up, so how significant is the increase in the polio exposed population?

DOCTOR BUTEL: We haven't analyzed the data that I can give any confidence limits for you. These studies are ongoing.

DOCTOR WEISS: Well, I hope you'll be able on a future occasion then, because it's very important, particularly, given the public concern and media interest in this.

DOCTOR BUTEL: But, the numbers are very similar to what Keerti published in the 1970s.

DOCTOR WEISS: Yes.

DOCTOR BUTEL: And, Geissler also.

DOCTOR WEISS: Yes, we shouldn't forget Irwin Geissler, who did these pioneering studies when he was quite on the wrong side of the Berlin Wall, and when we complain about lack of facilities and resources, just consider how he managed and produced those data, and then we should consider ourselves very lucky, us on this panel.

DOCTOR GARCEA: Does that mean it's going to get worse?

AUDIENCE: Doctor Weiss, I'd like to take a whack at you. Since I'm from Berkeley, I'm assume the role of Duisburgian. Since you brought up adequately that this really is a multifactorial disease, should we really be applying Coates postulate to it? I mean, I look at the SV40 data, I heard this morning that it's probably endemic now. I'm a great disbeliever in antibody data, so I can't conclude whether it was before the vaccine or not, and I don't care to.

But, to me, if you have the presence of SV40 in your blood cells, which I think some of the investigators showed yesterday, to me that means it puts you at risk for the possibility of developing a mesothelioma or the other tumors, but I just don't see a causal relationship. I think it's just a component in the whole multifactorial process.

The reason why I asked you is whether we should state a Coates postulate or not is, it has everything to do on how we design epidemiologic studies, are you asking all the right questions? What else have you been exposed to, what other viruses do you have?

I think Doctor Furth gave an absolutely excellent presentation in Science that was superb, that says, to my mind, that T-antigen is more like a cigarette lighter that gets the whole hyperplasia going, and since she didn't resolve all the hyperplasia, enough DNA damage was done that you might find tumors that have no SV40 in it, yet it was involved.

So, I'm worried about applying Coates postulate, which works fabulously for acute diseases, whether it really is the right way to go for the chronic illnesses and approach it more from a multifactorial position.

DOCTOR WEISS: Well, thank you. Yes, we visited this with Multiple Sclerosis, with Rheumatoid Arthritis, with cancer many times before, and it is very difficult.

AUDIENCE: Yes. I'd like to address the issue of viruses in the blood. Is Doctor Dorries here? I attended a meeting in Cambridge about five years ago, in which she reported that about 60 or 70 percent are actively carrying BK and JC viral genomes in our peripheral bloods, and there has also been work on adenoviruses as well. So, there are any number of viral genotypes associates with human peripheral blood cells, and I think that to think that there's some etiologic association between those sequences in the blood and the tumors that we happen to get at some point in time is a fair stretch of the imagination at this point in time.

DOCTOR WEISS: That question was also raised, if SV40 is involved it might be a hit and run type of infection, and we may not be too surprised not to see it in the final tumors, while it's in a relatively small number of mesothelioma cell lines, for instance. I guess that's true of EBV that's not found in most cell lines, the few cells lines that have been developed, nasopharyngeal carcinoma, it's also reported with Kaposi's sarcoma that the established cell lines do not have HHV-8, there's a dispute whether those cell lines actually are Kaposi's sarcoma.

So, with the herpes viruses, there do seem to possibly be cases where there's strong epidemiological evidence, what's lacking here with the polyoma viruses is strong epidemiological evidence for association of the tumor for the lack of genomes in the tumor. Here it's almost the opposite way around.

Would anyone like to comment on that?

DOCTOR DORRIES: I just would like to comment on the viruses in lymphocytes, BK virus and JC virus both are found in lymphocytes, although we, in the moment, don't know which cell type really is affected.

But, what we really don't know is whether they are replicating in the lymphocytes in the same way as they are replicating in persistent infected organs like the kidney, or like the central oligodendrocyte or astrocytes in the kidney.

We only have found DNA. There is some antigen found, but we really can't say what is going on in the normal persistent infections that might be activated.

In the other organs, the virus is very effectively replicating. The DNA always is, or mostly is, complete. We don't find any defectives and so on, so I really don't know what's going on in the lymphocytes. There it must be a little bit different than in the other organs.

DOCTOR WEISS: And, if tonsilar tissue is the first port of call, there might at least be some transient infection in lymphocytes, I guess.

Jim Goedert?

DOCTOR GOEDERT: I'll take a partial stab at your hit and run question. It strikes me as though, under some circumstances, a hit and run mechanism is certainly plausible, but it also strikes me as though it almost has to be the exception to the rule.

And, you mentioned a number of virus cancer associations, and the one that came to my mind was Hepatitis B with hepatocellular carcinoma, where the vast majority of those cases, leaving aside Hepatitis C virus, but the vast majority attributed to Hepatitis B virus have chronic active Hepatitis B replication and chronic Hepatitis B surface antigenemia, and it's only the exceptional case that appears to have had defective integration of the Hepatitis B X gene, upstream from the -- or rather, upstream from the necessary oncogenies, in which case the, you know, very, very detailed investigation can postulate a sort of partial hit and run or exclusion hypothesis, but it really seems to me as though it needs to be the exception to the rule, and we are so far away from that at this point that, you know, let's see if we can establish what's common.

DOCTOR WEISS: Thank you, that's a point well made.

Well, I think we have to look a little to the future. So, we've had a lot of debate in the last two days. We've heard interesting data. We've heard controversial data, some groups finding one thing, others another, or not finding the same thing. I'm not sure we're going to leave much the wiser in coming to firm conclusions about the prevalence of SV40 in humans, the provenance of that SV40, how it got into us, and its relationship to malignancy.

So, what do we need to do to extend this and firm up the evidence that has been presented or to refute it? One clear conclusion from the discussions of detection is to prepare some blinded samples, not necessarily oblige everyone to use the same techniques of detection, but to see whether a broader set of laboratories than has been reported so far, whether there's some consensus on detection.

It's not clear to me how this is being organized or which organizations are proposing to do it. The NCI was going to put something together. Howard, would you like to say anything about that, and, perhaps, Andy, for the FDA?

DOCTOR OZER: We are trying to put together a selection of test specimens which would allow multiple laboratories to test the same specimens. We were focusing on ependymomas, since those seem to be among the tumors with the strongest evidence suggesting that SV40 DNA could be detected in the tumors, as well as some negative controls.

We've been speaking in private with some of the groups working here and presenting here, to begin an initial investigation looking at inter and intra laboratory agreement, and those discussions also began to focus on some of the things that your data pointed to, which was which exact assays should be used to optimize everything. But, we've not been able -- we've not solidified any specific program, but we hope to move ahead to be putting together a series of test specimens. So, we'd be interested in hearing from people interested in testing specimens that could be looked at in common.

DOCTOR WEISS: Thank you, and I would have thought it was worth extending it from ependymomas to mesothelioma, and, perhaps, osteosarcoma, too, and to specifically incorporating some samples that Bob Garcea's and Michele Carbone's groups have already found to be positive, but suitably blinded, so that it's not just assembled by the skeptical school, so to speak. Would you be agreeable to that?

DOCTOR OZER: We'd be agreeable to that. I suppose the reason, up until now we were not looking to take the samples other people had been already finding positive, is to the degree to which the issue over contamination was a concern, if a specimen is already contaminated, sending it to additional laboratories, to also find that same contaminated specimen positive isn't necessarily all that enlightening, and we were hoping instead to start with specimens in which we would be able to prepare them in a way that would allow the laboratories to be testing specimens as similar as possible. So, we wanted to start in specimens in which we knew we were in tumor, that they validated the tumor, and then repeated sections from the same tumor specimen, and then verifying at the end of all the sections that we would make that we were still in tumor at the end of preparing the specimens. And, so that, each of the laboratories would have sufficient material to work with, and that the comparability of those materials would be as similar as possible.

The addition of additional tumor types is obviously a useful thing, but we just wanted to get going with something where we could have clear positives, clear negatives, just to measure the replicability in laboratory's hands.

DOCTOR WEISS: And, should we be extending laboratory-based epidemiological studies to serologic assays?

DOCTOR OZER: Obviously, that would be very useful, and we do have serum specimens that if a laboratory were able to show that they had an assay that they felt was worth testing, we have ample serum samples on osteosarcoma patients, patients with mesotheliomas, that could be tested against.

One of the things I wrote myself down as a note during this meeting was that it would be very useful to also put together a panel of serum specimens of the type that I was suggesting we needed to do to examine whether the assays are picking up what we think that they should be picking up, for example, serum from laboratory workers highly exposed to SV40 would be a very useful panel of specimens, I think, for people to be testing. The cancers that I mentioned, other cancer patients' serum as a comparison group and normal populations, including individuals specifically from the birth cohorts in which we expected them to be exposed to the contaminated vaccines. And so, nothing has begun in terms of putting together those sort of panels, but I would be happy to see such a thing coordinated through us.

DOCTOR WEISS: Would there be general agreement to the extension of such studies from the members of this workshop? Any dissent?

DOCTOR LEDNICKY: I think we would all like to do it, but a very important question is for people like me, who have no money --

DOCTOR WEISS: I was going to come on to that.

DOCTOR LEDNICKY: -- yes, our lab costs per sample, because of the number of primers, because we resynthesize primers, we use brand new reagents every time, you know, they are astronomical.

DOCTOR WEISS: Research costs money. That wasn't the reason you are standing near the microphone, Jim, let's hear from you first and then come back to that.

AUDIENCE: I was going to add a caution about the serum studies for antibodies. It seems to me as though you must have a gold standard to begin with, and in my mind what you'd want is serum samples from one or more people who have shown to have the virus. That means you need to have a link probably between at least your PCR result and your serum collection.

At the moment, I don't believe that we have that in hand, and so I think that the people who have been doing the work and demonstrating the virus, perhaps, do have some reagent quality sera from people who have SV40 infection as detected by at least PCR and serum samples as well, that could be added into the panel for the serum antibody studies, and that would not carry with it some problems about, you know, carryover contamination, that sort of thing.

AUDIENCE: Jim, are you talking about positive controls?

AUDIENCE: Yes.

DOCTOR WEISS: Yes, I think that's very valuable. If I think of the two papers published last year on the vast question of human foamy virus, sorry I keep coming back to retro viruses, where some 10,000 samples were studied and the general conclusion now is that foamy virus is not human, the positive controls were immensely useful, both serologically and for PCR, that they are just a small number of people, monkey handlers, who are genuinely infected with Simian foamy viruses, and they provided the gold standard against which we could measure the well population and find, after 25 years of claiming that this was a natural widespread human infection, that it's probably not there at all. So, we need a gold standard.

AUDIENCE: Yes. I take issue with the serology also, that, you know, if all you have is a hammer and everything is a nail, you are going to fall into a problem.

We saw data today, and it's been published for a long time, you gave 30 odd volunteers SV40, and they nicely placed it in their feces for five weeks, and yet, they made no antibody response. So, I'm not sure, what can you conclude when you know that not everybody makes an antibody response to SV40, and if you have a tumor that doesn't have the antibody do we conclude that this person wasn't exposed to SV40? I think this is a dinosaur technology that shouldn't even be addressed.

You need to upgrade it to the genetic analysis, it is expensive, you've got to find the money, but I don't think the antibody results will give you the kinds of answers to the questions you are asking.

AUDIENCE: I would disagree. I think there certainly are limitations on the serology, but it can be made more specific and more selective. I think that using a competitor, and specifically using JC and BK as blocking agents in a serologic assay, will give it a high degree of specificity.

And, I certainly don't think --

DOCTOR WEISS: Please, use the microphone.

AUDIENCE: -- I simply said that I think that the serological assays can be made useful by using a competitive blocker, and specifically JC and BK. So, I don't think we should completely discard serologic assays, because in terms of how one goes about doing a mass epidemiologic screening to seek endonicity in various populations, I think PCR is going to be very impractical, plus which I would remind you that we haven't yet agreed upon a standardized set of conditions for PCR or how it would be interpreted.

The importance, I think, of making these distinctions was well illustrated by a question that you raised, Robin, that I don't think has been adequately answered, which is this. In the tumors, if there is, in fact, sufficient T-antigen to bind P53 and/or RB and render those proteins dysfunctional, why then is there so little SV40 DNA in the tumors.

And, one obvious answer could be that, in fact, the T-antigen that's being measured is JC or BK T-antigen, as you pointed out. So, I think that it's absolutely critical to resolve the issue of which virus we are talking about, and I think that the first step ought to be to, in fact, ensure that we have, not only sensitive, but specific serologic assays.

DOCTOR WEISS: Well, although there was some difference in opinion as to what was worthwhile, I think there was consensus that we have to develop more specific and more sensitive technologies before we can address it on a very widespread scale.

AUDIENCE: This is a quick point, which is that if we are going to look at serum by any mechanism or device, we need to be sure that we know what we are doing, because many of these serum banks have vanishingly small amounts of the serum, and they are invaluable resources, particularly, the old ones, so we shouldn't be plunging ahead to look at serum until we are positive that we are doing the right thing. Otherwise, we will quickly squander a valuable resource.

DOCTOR WEISS: Good point.

Mike Oxman?

DOCTOR OXMAN: Robin, I think we'd all agree that there have been lots of introductions of real SV40 into the human population, but I, for one, am not at all persuaded that it's endemic. It would be very easy, I think, to persuade me and lots of other people if five percent of BK negative, approximately, five percent of BK negative people currently in the population have low levels of alleged SV40 neutralizing antibody, if you picked 500 pregnant women who had that serologic picture and looked in their urine, if SV40 is endemic in the population, it's almost certainly to be shed in the urine by those pregnant women. And, if it isn't, I think it would be extraordinarily unlikely that it's endemic.

DOCTOR WEISS: Well, we did hear about HIV positive patients research in urine, and that, surprisingly, was negative.

Keerti?

DOCTOR SHAH: I think I must be one of the few people who has looked for BKV, JCV and SV40 antibodies in human sera.

If you give me 1,000 human sera, and if we did tests with all three viruses, regardless of what test we did, there will be, perhaps, more than 850 sera which will give completely unequivocal results.

So, it is only a matter of resolving some things that are not very clear, and I liked very much a suggestion which Doctor Oxman and Doctor Ozer had made yesterday, that you could get a panel of sera, and Doctor Olin from Sweden even said that he had all the sera set up, where you could do a sort of a screening test with all these, and then those that are difficult to interpret there are millions of different ways in which you can clear it up by making the proteins in vitro, or labeling the virus. There are, I mean, many, many ways in which this can be done, so I don't think that's a major question.

The second point is, what Doctor Oxman mentioned, is that if you did find SV40 antibodies, supposing we looked at women or men, either immunosuppressed transplant patients, pregnant women, HIV positive women or men, and then if we found SV40 antibodies we would look for the virus. And, I think unless we find this virus which is thought to be circulating, everything we will get will be with many reservations.

And, I think, Doctor Weiss, you said a number of times that we agree that SV40 is endemic, I don't think there is agreement that SV40 is endemic, but what is clear, that the virus that is being found in all these tumors is SV40.

So, if there is a human virus circulating, then it will be SV40, rather than -- it is not a BKV or JCV that we are picking up. So, I think one should make a major effort to look for the virus, and I really like Doctor Pass' suggestion, that all these asbestos-exposed people, there are probably many thousands of them, and they are going to develop mesothelioma, so if a prospective study was done now, and we look for SV40 serologically, perhaps, in the urine, I think that might be extremely valuable in sorting out the causal relationship, how frequently it is present, and things of that sort.

DOCTOR WEISS: Thank you very much.

The last question of the tentative ones that Andy Lewis put down that might be raised is how can the resources necessary for these analyses to evaluate SV40 infection in humans be identified and made available.

Well, one way is public pressure. We have a workshop here with television cameras in the back. We can be pretty sure that that will divert funds away from, perhaps, other equally worthy avenues of medical research, and equally important ones where the television cameras aren't present, that it is one way of making sure that resources come to this.

But, I wonder if the representatives of the NIH and the FDA, who have sponsored this meeting, have suggestions of realizing the not excessive resources, the necessary resources to develop further analytical tools and then to apply them.

I'm asking those who can provide the resources, not the advocates of patient groups.

DOCTOR LEWIS: I think I should say that that's certainly under consideration. There's been no resolution to that as yet, and --

DOCTOR WEISS: Can you speak into the microphone, Andy?

DOCTOR LEWIS: Now better?

DOCTOR WEISS: Yes.

DOCTOR LEWIS: Okay, sorry.

Those problems are under consideration, but as far as I'm aware there's really no systematic way to do it yet. We are still thinking about it.

The question is, what was going to come out of the meeting, and how we might be able to respond to that, and I think that's all we can say now.

DOCTOR WEISS: Well, we had Ruth Kirschstein open the meeting, the Deputy Director of NIH, so, perhaps, that message can be carried back.

AUDIENCE: I would like to make a statement on behalf of the public, because, after all, public health officials are responsible to the public for the choices they make on their behalf.

DOCTOR WEISS: Yes. Are you sure you represent the public, not your own organization?

AUDIENCE: Excuse me?

DOCTOR WEISS: Are you sure you represent the public at large, rather than a small sector of the public?

AUDIENCE: We represent parents who want to vaccinate their children. We represent health care professionals, who are very concerned about vaccine safety, and we've done this for 15 years, and we've sat on government advisory committees, and we have worked within the infrastructure for the public health.

DOCTOR WEISS: Okay, go ahead and make your statement briefly, please, as we have to close in a moment.

AUDIENCE: I think the proof of whether the government is going to seriously address the issue of whether Simian viruses have played a role in human disease will depend upon what you do after this conference. It will depend upon whether or not you are willing to involve the independent researchers that came forth with the data.

In any scientific investigation that you make, whether or not you are willing to commit immediately funds to conducting this research, and whether you are willing to be forthright with the public about what you are doing and what you find out. DOCTOR WEISS: Thank you for reiterating your views.

Frank O'Neill?

DOCTOR O'NEILL: I'd like to make a comment on the so-called contamination problem. This has been touched on earlier.

If there is SV40 in some human tumors, it seems to me that some of these copies of the DNA should be integrated. And, I think it would be very useful to identify the cellular flanking sequences and show that they are human.

And, in each case, if it works like it has worked in human transformed cells and animal transformed cells, the integration site should be unique. So, in each case that the viral DNA is integrated, there should be human DNA, and that human DNA that flanks the viral DNA should be different and unique in each case, and I think that would be one step forward in solving the contamination problem.

DOCTOR WEISS: Good point, if it is integrated then we can show it's integrated in human DNA.

Michele, did you want to respond to that?

DOCTOR CARBONE: More than responding to them, I want to just make a small comment, if you allowed me.

We always have been very careful when we talk about the sequences, calling them SV40-like sequences. And, Doctor Butel then has been able to isolate one SV40 virus and to sequence it.

But, I think that it's very easy to slip into the confusion and making everything all. The fact that one of these sequences corresponded to wild type SV40, or -- SV40, I think that still the possibility that Doctor Shah brought before, and I'm happy to agree with Doctor Shah 100 percent, that these sequences may not always correspond to SV40, but may correspond to something that today we call SV40 because that's the best that we can do when we put the sequence in the computer.

But, they may be all the things that we discussed here, such as hybrid virus, a virus that has been endemic in the population. It seems that suddenly all these things have been discovered, and then all the discussion that I hear is, well, it has to be SV40, so it's coming from the vaccine, it's not coming from the vaccine, is endemic, we have to do a test for SV40, I think it is also important to see whether it's another virus that is similar to SV40 or maybe is both things. Sometimes it is SV40, something it's something similar.

Thanks.

DOCTOR WEISS: Thank you.

Well, I think we should draw this panel to a close. One more point from the floor there.

DOCTOR ROSS: I'm Malcolm Ross, from the U.S. Geological Survey. I'm a mineralogist, it may seem a little strange that I'm at this audience, but I've been involved with minerals and mineral dust in health for quite a while.

And, we've been trying to get intelligent public health initiatives in asbestos, various asbestos dust, and we particularly have been involved with the EPA for years in trying to bring sense about this asbestos abatement in schools and public buildings.

And, the United States has spent probably $100 billion removing asbestos. As a public health initiative, it's certainly not cost effective, if you think of $100 billion, and not health effective either, because of the very poor abatement.

I would just bring this up, to indicate the economic importance of the issues we are addressing here today and yesterday, particularly, with regard to mesothelioma asbestos disease.

Now, the abatement initiative has gone to Europe, and the fear of asbestos is promoting, oh, let's rip it out of all the buildings, as a public health initiative, and I think we really need to guide the public policy in this area to make proper initiatives that are not so costly.

I gave a lecture at the University of Paris last October, explaining how we mismanaged our abatement, and in Paris they are going ahead, they are starting with a $300 million initiative at the university there in the Latin Quarter, and I can see this fear of asbestos taking over Europe like it took over here. And, I fear that, you know, with, say, viruses that conceivably the public could fear even having vaccine, so I just wanted to point out the economic importance of good public health policy and good science being used to direct that policy.

Thank you.

DOCTOR WEISS: Thank you. I think that's very important, and I think it's very important for the media to get over a sense of the complexity of science and the uncertainty of science, and not uncertainty to create fear, but it is difficult in a very complex issue like we've been discussing in the last two days, to obtain clear evidence and clear weighting of different factors in these types of malignancy.

I'd like to thank all the panel this afternoon. I'd like to thank Andy Lewis and his colleagues for convening this workshop, which I think has been open, has had -- many of us are quite independent of government funding, there was only one rare lapse of judgment from Andy in asking a retro virologist to chair the last panel, though, sometimes a comparative view helps.

I think we leave having had a good discussion, not knowing where funds are going to come from for future studies, and knowing that there's a great deal more work to be done before we can either lay this to rest or establish a clear causality.

That's all I have to say now. Are there any announcements from Andy?

As he comes up here, let's all give him and all the people who have helped in this workshop some applause.

(Applause.)

Return to Agenda

DOCTOR LEWIS: Thank you very much, Robin, and folks in the audience.

It sort of falls on me to summarize the events of the last two days. I'm going to be very brief, because the time is late, and we have to have a couple other things to do before we leave the building.

The first observation that I think I should make is that confidence in using PCR assays, under even the most carefully controlled conditions, to amplify DNA sequences that occur at low levels in reaction mixtures, will require, it looks like, the develop of specific primer conditions for each specific set of primers, work with defined and standardized reagents.

It also seems that using PCR assays to amplify DNA sequences occurring at low frequencies in samples from nature, that is in tumors, normal tissues, DNA from archive specimens, also are going to require some additional work which will include sequence analysis of the amplified products.

The questions about the administrative mechanisms which we've heard a good bit here in the last few minutes of the panel, to support the development and the continued evaluation of standardized PCR reactions and serological reactions need to be resolved.

I am going to suggest that the sponsors of the workshop will need to develop such mechanisms, and at least from my understanding these things will be very seriously considered in the future.

Based on the serological data, in the isolation of SV40 from humans over the years, culminating with a report by John Lednicky in '95, SV40 in some form certainly seems to be out there, and has probably been in the population most likely preceding the large-scale exposure to SV40 in the polio vaccine back in the '50s and early '60s.

The question of interfamily transmission of SV40 in the prospective study of the families of children who developed choroid plexus tumors and ependymomas in the future, certainly seems to be an important direction to go.

These types of studies, in fact, could be greatly enhanced by the development of improved serological assays for detecting SV40 antibody.

Even though a number of laboratories have detected SV40 sequences in human tumors, the detection of SV40 specific sequences in normal and non-neoplastic tissues and body fluids by the same techniques, and the low frequency of SV40 sequences in tumor cell, it seems to me precludes at this time forming any conclusions as to SV40 cause and effect relationships in neoplastic development in humans.

I do believe this concept is supported by the epidemiological studies of multiple tumors in the United States, as well as in Sweden, and from this data it seems that the relative risk of developing any of these neoplasms has not increased for those who were exposed to SV40 in the polio vaccine.

If the SV40 DNA sequences continue to be detected in humans and in other types of environmental materials, a possible role of recombinant virus in the dissemination of SV40 will have to be kept in mind.

A number of interesting cell culture systems in animal models have been presented today. These models suggest that mechanisms of SV40 oncogenicity are being methodically unraveled.

The cys gene system described by Harvey Ozer provides an opportunity to evaluate in human tumors mechanisms that could be SV40-related events in the transformation of human cells.

In closing, I certainly would like to thank the session chairmen, Doctor Kirschstein, Breiman, Snyder, Mahy and Kelly for their assistance in managing the meeting. I think the panel moderators, Mike Fried, Art Levine and Robin Weiss, really require special thanks, because they put in a tremendous amount of work in organizing and managing the panel audience discussions. This is really science as it should be as I see it.

I'd especially like to thank each of the speakers for the enthusiasm with which they expressed when they were asked to participate in holding the workshop and the efforts they put in to review a huge amount of data that's accumulated over the past 40 years, and to present new data related to the issues that the workshop was designed to address.

I think the audience also needs to be recognized for its participation and contributions, for the questions and the discussion that they raised, and for the agenda items that they addressed.

In this regard, I think the comments and the support of the members of the parents group who monitor safety issues also needs to be mentioned.

I'd like to thank the members of the media who have been here throughout, for their cooperation in allowing the business of the workshop to proceed unimpeded.

Finally, I think it's important to recognize the contributions made by the co-organizers, Howard Strickler, Jim Goedert at NCI, and Bill Egan in the Office of Vaccine Research and Review.

In this regard, I think special recognition goes to Doctor Carolyn Hardegree, the Director of the Office of Vaccine Research and Review, for her interest and advice on numerous occasions as we developed the format for the workshop.

To the speakers, you need to know that the proceedings of the meeting will be published in the Developments of Biological Standards Series, instructions to the office for preparing these papers will be sent out to you in the next week or so. I'll ask that you try to get your papers back to me the first of June. Once the papers are -- we collect the papers and we get them in the hands of the publisher, it will take about five or six months to the publication of the final volume.

Doctor Florian Horaud, of the Pasteur Institute, needs to be recognized and thanked for his help in arranging the publication of the proceedings of the workshop.

I certainly hope that we can continue to participate in these discussions at the DNA Virus meeting coming up in Cambridge. I asked Mike Fried to think about how we might do this, and he's assured me that he'll give it pretty serious consideration.

The last thing on the agenda is that I want to remind you that there will be a press availability meeting in Conference Room B immediately following our exit from here, in addition to the FDA representatives, we are asking the following people to help us with this discussion, Doctor Levine, Doctor Dixie Snyder, Doctor Howard Strickler, Doctor Mike Fried and Doctor Robin Weiss, and Doctor Jim Goedert.

Thank you very much. It's been a great event for me. It's been fun to do, and it's been fun to meet and discuss all these issues with you.

Thank you very much.

(Applause.)

(Whereupon, the meeting was concluded at 5:50 p.m.)

Return to Agenda

Updated September 24, 1999

CBER Home Page | CBER A-Z Index | CBER Search | Contact CBER
FDA Home Page | Search FDA Site | FDA A-Z Index | Contact FDA | Privacy | Accessibility | HHS Home Page

FDA / Center for Biologics Evaluation and Research