Simian Virus 40 (SV40:)
A Possible Human Polyomavirus Workshop

The Workshop took place in the Natcher Auditorium, National Institutes of Health, Bethesda, Maryland, at 8:30 a.m., Kathryn C. Zoon, Director, CBER, presiding.

PRESENT:

KATHRYN C. ZOON, M.D. DIRECTOR, CBER
ROB BRIEMAN CO-CHAIR
MIKE FRIED CO-CHAIR
RUTH KIRSCHSTEIN CO-CHAIR
DIXIE SNIDER CO-CHAIR
BONNIE D. BROCK, V.M.D. SPEAKER
JANET BUTEL, Ph.D. SPEAKER
MICHELE CARBONE, M.D., Ph.D. SPEAKER
KRISTINA DOERRIES SPEAKER
ELLEN FANNING SPEAKER
RICHARD FRISQUE, Ph.D. SPEAKER
ROBERT L. GARCEA, M.D. SPEAKER
ALLEN GIBBS SPEAKER
MAURICE R. HILLEMAN, Ph.D. SPEAKER
MICHAEL J. IMPERIALE SPEAKER
KAMEL KHALILI SPEAKER
ANDREW LEWIS, M.D. SPEAKER
MARIA C. MONACO SPEAKER
LUCIANO MUTTI SPEAKER
FRANK O'NEILL, Ph.D. SPEAKER
PATRICK OLIN SPEAKER
DAVID SANGAR SPEAKER
KEERTI V. SHAH SPEAKER
HOWARD STRICKLER SPEAKER
MAURO TOGNON, Ph.D. SPEAKER
JIM C. WILLIAMS, Ph.D. SPEAKER
JOHN LEDNICKY, Ph.D. PANELIST

ALSO PRESENT:

DR. GALATEAU-SALLE
HARVEY PASS
ETHEL de VILLERS
ROBIN WEISS

CONTENTS

Introduction and Welcome by Dr. Zoon

SESSION 1 Presentations:

Dr. Fanning
Dr. Shah
Dr. Garcea
Dr. Butel
Dr. Carbone
Dr. Gibbs
Dr. Mutti
Dr. Giordano
Dr. Tognon
Dr. Shah

SESSION 2 Presentations

Dr. Dorries
Dr. Imperiale
Dr. Khalili
Dr. Frisque
Dr. Monaco

LUNCHEON RECESS Afternoon Session

Audience Participation

Presentation by Dr. Lednicky

Panel Discussion

SESSION 3 Presentations

Dr. Hilleman
Dr. O'Neill
Dr. Lewis
Dr. Brock
Dr. Williams
Dr. Sangar
Dr. Olin
Dr. Strickler

PROCEEDINGS

8:35 a.m.

DIRECTOR ZOON: On behalf of sponsor's agencies today, which include the National Institute of Child Health and Human Development at NIH, the Division of Cancer Epidemiology and Genetics at NCINIH, the National Center for Infectious Diseases and National Immunization Programs at CDC, the National Vaccine Program Office, and the Center for Biologics, Evaluation, and Research of FDA, I'd like to welcome you to this workshop on SV40.

We, the sponsors of this workshop are pleased that so many of the national and international scientific community have come to discuss this very important topic. The workshop was prompted by recent reports demonstrating the presence of SV40 viral sequences in tissue, including certain rare, human tumors, and the fact that SV40 was an unsuspected contaminant in early polio and adenovirus vaccines.

The potential connection between SV40 and these various tumors is confounded by the finding of SV40 sequences in individuals who are too young to have received the SV40-containing vaccines. As a result, we must ask whether SV40 was present in the human population and if so, whether SV40 was present in the human population prior to polio vaccines.

Moreover, the role of the human polyomaviruses such as BK and JC, which are closely related to SV40, need to be explored. Therefore, the purpose of this workshop is twofold.

The first is to consider the possibility that SV40 is an infectious agent that is endemic in the human population; and second, is to stimulate the effort required to determine if SV40 is a causative agent in human disease.

We look forward to discussions, both formal and informal, over the next two days, that will better define these scientific issues. Furthermore, we hope that these discussions will lead to important new research collaborations.

Finally, at the outset I would like to acknowledge the enormous effort by those individuals who organized this conference, including Drs. Strickler, Levine, Egan, and particularly, Dr. Lewis.

Before I turn the meeting over to the Chair of the first session, Dr. Ruth Kirschstein, there's several housekeeping issues I'd like to go through. Lunch will be available upstairs at the cafeteria. There will be coffee breaks served in the foyer down here. Buses will return participants to the Marriott Hotel.

Transcripts and videotapes of the workshop will be available and how to obtain them is presented in your registration package. In the context of that I would like to ask those of you who come to the microphone to please identify yourselves so we will have a record of it.

The registration package also contains a comprehensive bibliography of papers relevant to the topics being discussed at this meeting. If participants know of additional materials, please send them to Dr. Lewis and we will see to it that they will be disseminated to the registrants.

Finally, we welcome the media coverage of this meeting but we ask that in the spirit of facilitating the scientific discussions that are so vital to the success of this meeting, that the media refrain from questioning the participants during the meeting.

There will be a press-availability session at the end of the meeting in conference room B, and representatives from each of the sponsoring agencies will be available. Other workshop participants are also welcome.

With that, I'd like to now turn the meeting over to Dr. Ruth Kirschstein.

CHAIRMAN KIRSCHSTEIN: Good morning. I want to add my welcome on behalf of NIH as a whole. We're pleased to be helping to sponsor this conference.

This conference is a little bit of a nostalgia trip for me since I started working in this field many, many, many years ago. And indeed, I was scheduled to Chair a different session than this one. For reasons of scheduling, I had to change that, and when I looked at the program, it is perhaps even more appropriate that I chair this session since some of the talks relate to work that I did over 30 years ago.

With that, I'd like to call on our first speaker, Ellen Fanning, who is in the Department of Molecular Biology at Vanderbilt University, who will give us a brief review of SV40 biology, and an overview of the organization and expression of the SV40 genome.

Dr. Fanning.

Return to Table of Contents

DR. FANNING: Thank you very much. When I got a call from Andy Lewis asking me to review the biology at the beginning of this meeting -- of SV40 -- in 20 minutes, I felt a little bit like Sarge in the Beetle Bailey cartoon. He said, you've been working on SV40 for 20 years; you know a lot about it. I says yes, but I don't know all about it, and 20 minutes? Andy persisted, however, so it's of course a little bit daunting.

SV40 has been studied from the beginning of its first reports in 1960 by Sweet and Hilleman, by some of the brightest minds in science. This virus provided a model system to study cell transformation and growth control, eukaryotic gene expression and DNA replication, chromatin structure.

It's led to the discovery of enhancers and promoters -- many of the factors that bind to them -- to the discovery of RNA processing and eukaryotes to signals for nuclear protein transport and to the first reports of the tumor suppressor, p53. SV40 has taught us a lot about eukaryotic cell biology.

So what I'm going to try to do in the next 20 minutes is to start at the beginning, to mention a few of the milestones along the path in learning about this virus, and then over the course of the next two days, my colleagues will provide you with some data on the questions that are at hand and to be discussed in this meeting.

As was already mentioned, SV40 was a byproduct of the early polio vaccines. The virus was discovered in polio vaccines that had been produced in monkey cells in culture. The virus was named after the cytopathic effects that it produced in infected monkey cells which became highly vacuolated, hence the name. The SV40 standing for simian virus, or vacuolating virus number 40.

The virus particle is relatively simple. It's small, it's made up of 72 capsomeres that contain three different viral proteins. It's just slightly larger than ribosomal subunits. Inside the virus there's a mini-chromosome of 5,243 base pairs, which is complexed with cellular nucleosones made up of cellular histones.

The mini-chromosome as shown up here, it's a very compact structure. When it's prepared for electronmicroscopy sometimes it opens up like this so that you can see the typical beads on a string. These over here are intermediates in viral packaging.

SV40 normally infects permissive cells -- that was how it was discovered -- such as this little -- according to this little scheme which is typical of the way that it infects CV1 monkey cells in culture. The virus particle enters the cell, progresses into the nucleus. The viral mini-chromosome is set free in the nucleus, it's transcribed by cellular transcription factors and enzymes to produce the so-called early messenger RNAs.

One of the products, protein products, that's generated from this early messenger RNA is called T-antigen. This is a short abbreviation for tumor antigen. This protein was discovered by virtue of the fact that very early on, it was established that SV40 could cause tumors in rodents and that these rodents produced antibodies against a new antigen that was not found in the virus particle. It was called tumor antigen.

This protein, T-antigen, as it's now known, is found in infected cells as well as in these tumor cells. This protein migrates back into the nucleus where it carries out the next stage of the productive infection, initiating viral DNA replication and stimulating late transcription.

The late messenger RNAs encode the capsid proteins which again migrate back into the nucleus, assemble around the mini-chromosomes to produce the new virus particles.

Not all cells are permissive for SV40. Most cells in fact, are semi-permissive like this species here, or non-permissive like the rodent shown down here in the grass. The infection in non-permissive cells is depicted here in a slide borrowed from Arnie Levine's book, Viruses. It stops after the transcription of the early viral genes.

The virus particle comes into the cell, the early messenger RNAs are made, the T-antigen is synthesized, it migrates into the nucleus, and normally the infection stops there. However, infrequently, the viral genome can become integrated in the host DNA in such a way that the T-antigen continues to be expressed on the other early genes as well. This can lead to cell transformation.

All right. So we have seen that somehow the virus infection does not progress beyond the very early phase. Viral DNA replication does not take place and late gene products are not synthesized.

The year 1978 was a milestone year in SV40 research. One of the important advances that was made in this year was that the first reports of the sequence of the SV40 genome appeared by Walter Feur's lab, and independently from Sherman Weisman's lab.

You can notice several things about the SV40 genome here. There are two sets of genes. The early genes are here; the late genes are over here; one transcribed on the Watson strand; the other transcribed on the Crick strand.

You can also note that the genes are overlapping. That is, the early genes contain several protein products, there's a large T-antigen, the small T-antigen -- and not depicted on this slide is another early viral gene product called the 17KT antigen which was recently identified.

The late genes are also overlapping. The three capsid proteins: VP1, VP2, and VP3. In between these two sets of genes is a control region which is depicted in more detail on this slide down at the bottom, here.

Here we have the early messenger RNA being transcribed this way, the late messenger being transcribed this way, the alternative splice products of these two sets of transcripts lead to the different viral gene products.

If we look at these control sequences in more detail we see two prominent binding sites here for SV40 T-antigen. Now, as the concentration of T-antigen increases in the cell, T-antigen binds to these sites specifically. It down-regulates the early transcription through this binding event, and initiates viral DNA replication through this second binding event.

The sequences that control transcription are located here. There's a so-called tatabox promoter, and you'll hear more about these copies of the 72 base repeats which are called enhancers. All of these are required for transcription.

Now, as viral DNA replication takes place and T-antigen changes the activity of the cellular transcription factors, the late genes are turned by indirect mechanisms.

So if you're starting to get the idea that T-antigen is a rather complex protein, you've got the right idea. T-antigen is perhaps the most multi-functional protein that's ever been discovered. Notice all of the arms here. T-antigen is something that we still don't know everything about. Notice the mysterious spatial expression here.

We do know quite a bit about it, however. The strategies that the T-antigen follows in directing the viral infection in permissive cells is depicted here.

First of all, the T-antigen must prepare the cell to support the viral infection. It does this by kicking a quiescent, differentiated, resting cell back into the cell cycle and forcing it into the S-phase. It needs this in order to replicate its own viral DNA which is dependent on cellular replication enzymes.

Having done that, T-antigen then initiates replication of the viral DNA, recruits the cellular proteins to replicate it, and through the indirect mechanisms that I mentioned earlier, leads to stimulation of late transcription -- viral transcription.

T-antigen somehow then senses that it should not initiate DNA replication anymore, but allow the viral genomes to be packaged into new virus particles. In other words, the T-antigen function changes with time after infection.

All right. So now, how does T-antigen do all of this? Generations of graduate students and post-docs have mutagenized the T-antigen gene, and this is the simplified version of some of what they've found.

The protein encodes 708 amino acids; it's sensitive to proteases at the sites marked by the asterisks, which tells you that the protein is folded up in different folding domains. These domains tend to correlate with functional properties of the T-antigen molecule. For example, this domain is responsible for specific binding to those two sequences in the viral control region.

T-antigen has a number of other intrinsic, biochemical activities. It binds to ATP and hydrolyzes ATP in a DNA-dependant manner using sequences located in the carboxyl terminus of the protein. It uses both of these domains to encode a DNA helocase that was first recognized in 1986 in Rolf Knipper's lab.

It needs all of these activities in order to replicate viral DNA. Now in addition to these intrinsic biochemical activities, T-antigen also interacts with a large variety of cellular proteins. Some of the proteins that it interacts with are depicted here.

The DNA preliminaries alpha primase, interacts with T-antigen at two independent sites diagramed here. T-antigen interacts with nuclear location protein transport machinery located at this region here, designated NLS. T-antigen also interacts with tumor suppressor proteins such as Rb and the other members of the pocket protein family.

T-antigen also interacts with p53, and actually there are two independent regions that are involved in binding p53: one here and one here, as shown in Judy Tevethia's lab. There's also a sequence down here which was not very well understood, which helps SV40 determine what type of monkey cells it can replicate in.

Now, T-antigen's phosphoprotein, the sites have been mapped to serienes and threonines and two clusters in the amino terminus and the carboxyl terminus.

All right, so having said all that, let's try and look at how T-antigen carries out this strategy. I'm going to discuss first -- because we know most about it -- how T-antigen directs the replication of viral DNA, very briefly; and then say a few words about how T-antigen prepares the cell to support the viral infection.

All right. More than ten years ago Tom Kelly's lab developed a system which would replicate SV40 DNA in a test tube. This system was dependent only on one viral protein of course, the SV40 T-antigen, as well as ten cellular proteins that have been defined and studied in some detail in Bruce Stillman's, Jerry Hurwitz's, and Tom Kelly's lab.

T-antigen's functions in this system are threefold, basically. First of all, T-antigen binds to the viral origin of replication and assembles there as a multimer. Having done that, it proceeds to unwind the two strands of the parental DNA so that they're available to be replicated by the cellular proteins.

You can see here that a mutant T-antigen -- this is the wild type up here -- this mutant T-antigen is stuck at the origin and cannot proceed further to these bidirectional unwinding of the parental DNA.

The third function that T-antigen carries out in the viral replication phase of the infection is that it interacts with key cellular proteins involved in replication -- in getting replication started -- such as DNA preliminaries alpha primase.

Its DNA preliminaries alpha primase is the molecule which is responsible for determining the host specificity of viral replication at least, in a cell-free system, as first shown by Jerry Hurwitz's lab. And in fact, as it's turned out in studies that were carried out in my lab, it's only one of these subunits of DNA preliminaries alpha primase which is sufficient to determine whether or not SV40 DNA can be replicated by this preliminaries alpha primase.

What we did was to generate recombinant enzymes, human or mouse enzymes, or rehybrid enzymes which contain only one subunit from mouse or one subunit from human. And using this system what we were able to find was that only a single subunit, the large subunit of DNA preliminaries alpha primase must be from humans in order to allow SV40 replication. If that subunit's from mouse cells, SV40 DNA cannot replicate in the test tube.

Replication is also controlled by the phosphorylation state of T-antigen. If we look at the form of T-antigen that's most common in productively-infected cells or in transformed cells for that matter, it's a highly phosphorylated form of T-antigen; in particular at two key seriene and one threonine residues.

This form of T-antigen is not able to replicate viral DNA, although it represents the bulk of the protein in the infected cell. The form of T-antigen that's able to replicate SV40 DNA is an under-phosphorylated form which lacks phosphorylation at these two key seriene residues. This is a minor form in infected cells, but fortunately for biochemists like us who want to study it, it's the major form that's produced in recombinant baculovirus infected insect cells.

The unphosphorylated protein has made any cholase also inactive. So we have a lot of different things going on. We have a multi-functional protein whose activity is then being regulated more exactly by its phosphorylation state.

I'd like to turn then to the early stage of the infection when T-antigen is trying to prepare the cell to support the viral infection. If you look here you'll see a representation of a cell cycle in eukaryotic cells. Most cells that T-antigen would infect in an animal would be in a resting state.

And as soon as T-antigen concentration builds up, presumably in this highly phosphorylated form, it will have the effect of forcing the cell back into the cell cycle, forcing it through the early G-1 phase of the cell cycle and into the S-phase.

It does this by circumventing some of the signal transduction pathways that normally would control cell growth. It does this through its interactions with cellular growth control proteins such as the Rb tumor suppressor protein, the p53 tumor suppressor protein.

Also in 1978 it was first shown by Adolf Gressman that T-antigen was sufficient to stimulate cells to re-enter the cell cycle and progress into the S-phase. And this experiment was reproduced in my lab, shown here, either in secondary African Green Monkey kidney cells or in CV-1 cultured cells.

The cells were serum-starved and then treated with -- and were microinjected with SV40 DNA, or treated with serum and then the kinetics of re-entry into the S-phase were followed. You can see down here that T-antigen protein does this faster than SV40 DNA.

There are a number of functions of T-antigen besides binding to Rb and binding to p53 that are involved in this growth stimulation functions. Each of the mutations shown by these red bars will inactivate these growth stimulating functions but does not affect the ability of T-antigen to replicate viral DNA.

So it's possible to specifically knock out these growth stimulating functions, and this is by no means all of them. There's only four of them here. They're interacting independently with cellular growth control proteins.

To give you an idea of the importance of these interactions, in the biological activity of T-antigen in stimulating cell growth and eventually transforming the cells, bear in mind that not only SV40 T-antigen interacts with these cellular growth control proteins, but also other groups of viruses.

Adenoviruses and code early proteins that target the same cellular proteins, and the human papillomaviruses -- which we know are risk factors in human cancer -- are early proteins which target the same set of -- at least some of the same set -- of cellular growth control proteins.

So I'd like to stop there and hope that I've prepared you to fit the biology, together with some of the data that you're going to here over the course of the next two days on SV40, as a possible human pathogen. Thank you.

CHAIRMAN KIRSCHSTEIN: Thank you, Dr. Fanning. Indeed, you have kept the time beautifully and also prepared us for the next session of speakers.

Our next speaker will be Dr. Keerti Shah, who is in the Department of Molecular Microbiology and Immunology at Johns Hopkins, and will talk on SV40 as an infectious agent in simians and humans. Dr. Shah.

Return to Table of Contents

DR. SHAH: This morning, while registering for this meeting, I met a lot of people whom I had not seen for 10 to 20 years and I said, we are all old-timers. One of them told me that I was the oldest of the old-timers.

What I want to do is, I was going to review some of the background for the infection of human beings with simian virus 40, which was a contaminant mainly in the inactivated Salk polio vaccine. And what I was going to cover was the natural infection in Rhesus monkeys in India who are the donors of the kidneys in which the vaccines were made.

I'll describe a little bit of the experimental infections in these animals because that gives us some insight into how the virus might be transmitted, and then describe briefly the circumstances of human exposure to SV40 and the early studies of finding out if this virus was pathogenic for man or not.

And I had reviewed this topic in some detail in 1976 -- and this citation is shown on the slide -- and almost everything I'm going to say is contained in that review paper.

The three viruses which are very similar is: one is the simian virus 40 of the Macacs, and the two human viruses, BK viruses and JC virus, which are very similar biology to the simian virus 40. The BCV and JCV were identified in 1971; SV40 in 1960.

They're called polyomaviruses after the polyomavirus of mice, which was the first virus of this subfamily that was characterized. And there are many polyomaviruses scattered in a large number of species -- as shown there in rabbits and mice, in parakeet, in cattle -- so that it is widely distributed. And each polyomavirus is very well-adapted to the species in which it grows. So they're highly species-specific and no polyomaviruses is shared between two different species.

The primary infections with these viruses are almost completely harmless and after the virus enters the body there is probably some multiplication at the local site. Then there's a period of viremia on the viruses in the blood and it reaches its target organs by viremia.

The target organ in most instances is the kidney. The viruses in their primary infection produce viremia that reach the kidney. There is probably some virus excretion in urine, and after that the viruses remain latent in the kidneys indefinitely, perhaps for the lifetime of the particular infected species.

They are reactivated in times of immunological impairment and transplant patients, patients with AIDS, are the ones in whom these viruses are found very frequently. Viruses are found most frequently -- BK and the SV -- in the urines of these exposed individuals.

This is the distribution of the Rhesus monkey which provided the kidneys in which the collections were made. And it's a Macacus species which lives in north India. And we have done some work on the infection in these Rhesus monkeys.

In south India there's another Macacus species, a bonnet Macacus, which is not naturally infected with SV40. But the Rhesus Macacus, whose picture you see -- here there is a female with a baby -- this is an angry male, and he looks as if he might transmit something more than SV40. Actually in India, very often these monkeys will bite individuals who are nearby.

And many of these Rhesus live in ecological contact with human beings as shown here in this temple in Nepal, where they are there in large numbers. One of the questions which I looked at at the time, was whether the virus from the Rhesus monkey was transmitted to people naturally, and if this were to happen it would occur in a country like India. And the first NIH grant that I obtained was to study if simian virus 40 was responsible for any human cancers in India.

Only some of the Macacus species are infected with SV40, and in the Rhesus Macacus only about 20 percent of juvenile monkeys, and perhaps all of the adult monkeys, have antibodies to the virus. So it is not widely prevalent.

In some situations, as in a colony that was established in the island of Cayo Santiago off Puerto Rico, where the Rhesus monkeys were brought there in 1938, they came to the island with SV40 infections, but then it was eventually lost -- the SV40 infection was lost from this Rhesus colony. So this can occur.

Although the infection is not widely prevalent in young Rhesus, when they are brought together and caged together as they were in India prior to their transport to the United States, and then in the U.S. before they were used for vaccine production, the antibody prevalence reaches practically 100 percent, because there is a great deal of transmission from infected animals to non-infected animals.

If you infect the Rhesus experimentally, it is infected extremely efficiently whether the virus is given subcutaneously or -- orally this virus was introduced into the stomach of the Rhesus monkey, or by the intranasal route. In all instances they have a period of viremia, generally in the first week that this virus is in the blood.

The period of virulia, virus in the urine which is seen two to six weeks post-inoculation, in this particular experiment the virus was not recovered from rectal swabs and throats swabs. And then all the animals, no matter how they were infected, they developed very high titers of antibodies to -- neutralizing antibodies to the viruses, and they also developed antibodies to the T-antigen that Dr. Fanning described in the earlier presentation.

There are many other studies done at the time simply to see where the virus is. And in a number of studies in the African Green Monkeys it was shown that the virus is excreted in the urine, it is latent in the kidney, and there may be a lower-level shedding of the virus during infection. In African Green Monkey the virus was recovered sometimes from throat swabs and stools.

SV40 does not produce much or less in the Rhesus Macacus. It is extremely rare that it would produce any less in the Rhesus Macacus. No tumors, benign or malignant, have been ascribed to SV40, any tumors in the Rhesus Macacus.

But just as JC virus and BK virus will produce human disease in immunosuppressed people, so does simian virus 40 produce disease in Rhesus Macacus, especially when they're immunosuppressed. When monkeys that have the immunodeficiency virus in them, the simian virus 40 will produce an illness which resembles PML, progressive multifocal leukenepalopathy, which is a degenerative disease of the nervous system, demyelinating disease.

It also sometimes produces renal pathology, renal tubular necrosis, which is very similar to that produced rarely by BK virus. So with all of these viruses, most of the illnesses occur only in immunosuppressed populations.

Now, the factors that determine how much virus will be in the vaccines are listed here. First, the source of cells that are used. Now, many of the vaccines were produced in Rhesus cells, but some were produced in cynomolgus cells, which is another Macacus species.

The Cynomolgus Monkey is not naturally infected with SV40. So if the cells were of Rhesus origin, there's a greater chance of that being contaminated than if it was cynomolgus cells.

The type of culture of the cells are grown in a monolayer culture. They expressed simian virus 40 and replicated simian virus 40 very readily, whereas in some instances the vaccines were made in what are called the Maitland cultures, where the cells are not in monolayer form, but they are in the form of minced kidney tissues. This Maitland-type of culture did not support the replication of SV40 as well as the monolayer cells.

Then in many instances the kidneys pooled. The number of studies then would show that if the vaccine was made in a single -- all the cells derived from a single animal, it has a low chance of being contaminated with SV40, but if you pooled the kidneys, then any one infected kidney would contaminate all the rest, and then you have a much higher chance of getting contaminated vaccine.

One of the big, major factor was -- especially in terms of live SV40 and only matter of importance is live SV40, not inactivate SV40 -- depended upon whether the vaccines were live vaccines or inactivated vaccines. In the live vaccines such as the oral Sabin vaccine -- which are not inactivated -- the SV40 remained in high titer; whereas in the Salk vaccine the formalin that was used to inactivate the polio virus, also inactivates SV40 to a large extent.

So people would get, in the contaminated vaccines, either live SV40 along with a good bit of inactivated SV40, and they would get smaller amounts of SV40 in the Salk vaccines than they would get in the Sabin vaccine.

And very probably, although we are not completely sure about this from the data that we had available to us in 1976, only a proportion of the Salk vaccines had contamination with SV40 because a large proportion of the vaccines were made in the Maitland-type culture which do not replicate SV40 very well.

The most important exposure is the third that I've listed here: licensed inactivated polio virus vaccine. But in 1955 and 1961 the vaccine was contaminated -- some lots were contaminated. As we said before, being an inactivated vaccine it would have low amount of live SV40, but 98 million people had received the vaccine by 1961.

The live polio vaccine which would have large amounts of SV40 in the United States, only the experimental lots contained live SV40. By the time the live polio vaccine was licensed it was required to be free of SV40. So in the U.S. people were not -- not a large number of people were exposed to SV40 in the live vaccine.

The live RS vaccine which is the first line there, it's given to very few people and it's important simply to see what happens to SV40 when it is given intranasally. The second one, the inactivated adenovirus vaccine -- I think it's an error on my part here because those were live virus vaccines that were given to military recruits.

If you give the SV40 intranasally to individuals, the virus excretion occurs in throat excretions to some extent. There's a very low level antibody response. With oral vaccine there is almost no antibody response and the virus is recovered very infrequently, suggesting that the infection is very transient. There is no information with respect to the subcutaneous vaccine, how often it is -- if the virus is disseminated from people who are infected by subcutaneous vaccine.

And I'll pass this over just to show, the people who are most likely infected with SV40 vaccine, the year of birth, 1941, '61 -- people born between 1941 and '61 -- have a high probability of being infected. Those were born up to 1963 and later, those are very small probability of being infected.

The are a number of studies -- this last slide -- the number of studies done in the United States to see if the virus had bettered in a city of poor people -- and I have just listed them. I think we may have a chance to discuss many of them in the course of the two days.

But it is summarized at the bottom that while the studies did not reveal any ill effect of SV40, they did not have enough numbers, there was not sufficient period of follow up. The most susceptible would be infants who were infected in early life; not many of them could be followed. So while the data did not show any pathogenicity for SV40, there were -- all of the studies had their limitations.

Thank you.

CHAIRMAN KIRSCHSTEIN: Thank you, Dr. Shah. We will continue now with SV40-like sequences in choroid plexus tumors and ependymomas by Dr. Robert Garcea who's at Children's Hospital in Denver, Colorado. Dr. Garcea.

Return to Table of Contents

DR. GARCEA: Thank you. What I would like to present today is the data from a New England Journal paper in 1992. And I'd like to present it in the context of exactly how the experiments were carried out. And so it's an historical, sociological presentation rather than a -- although I will give you all the facts, I think you'll see -- I want to give you the flavor of how we proceeded with these experiments.

My laboratory at that time was located at the Dana Farber Cancer Institute, and our laboratory is primarily interested in the structural biology of papomaviruses and really not in translational work, although I am a Pediatric Oncologist. In 1989 John Bergsagel came to the laboratory to look for a post-doctoral fellowship and wanted to pursue whether there was any link between polyomaviruses and human, in particular, pediatric malignancies.

I didn't think any existed, but I was struck -- and this paper had always stuck in my mind -- this is from 1984, Brinster and Pometer, who described -- I think it was the second transgenic animal ever made with the SV40 large T-antigen driven under its own enhancer/promoter.

These animals developed very peculiar tumors, all of them in the coriplexus of the animals. And so I thought that this was quite a striking result and therefore I told John to go learn PCR and examine all the coriplexus tumors he could get his hands on at the Children's Hospital in Boston, the presence of polyoma-like sequences.

Now, coriplexus tumors are very, very rare tumors. They're only three percent of all of childhood malignant intercranial tumors, and we estimate that there's about 30 to 60 cases per year in the United States. Very interestingly though, all of these tumors are -- the majority of them occur within the first year of life, and they are in the differential diagnosis of hydrocephalus in utero.

A related tumor that we decided to study along with coriplexus tumors, because of the animal data of tumorigenicity of SV40, and because the ependymal cells are also a lining cell of the ventricular cavity of the brain, were ependymomas. Ependymomas are a little bit more common but again, still strikingly rare.

So John went to the pathology department at Children's Hospital, to the archival specimens. And he decided to -- what I thought in my wildest hopes was that maybe he would find BK or JC in these tumors and in my less-than-wildest hopes I knew that we would probably get PCR working in the laboratory.

John was to do a computer-assisted search of the polyomaviruses and align the sequences and find out where there were very close homologies between the viral genomes to make PCR primers. And not surprisingly, what he came up with was a region just after the splice site in large T-antigen, which is the Rb binding site. This site is very highly conserved among most of the polyomaviruses. And so John made primers in this region.

And the strategy that we initially employed was to -- these regions are identical for the forward and reverse primers, for both the BK and the JC viruses. I think we were trying to economize -- at that time it was expensive to buy primers, and so I think we were trying to be a little bit economical here.

And so these primers at the end would amplify either BK or JC, and the strategy therefore, was to amplify the specimen and then probe with an internal probe that was unique for either BK or JC.

And so what we did, John went to the archives, scraped slides off -- just in passing and I think we'll get back to this -- is that at least at Boston Children's, before 1976 most of the brain samples were fixed in Bouin's solution which has I think, picric acid in it, and we found that the DNA was highly fragmented. So most of our specimens came subsequent to 1977.

So what John did was, for most of these specimens which were on slides and from paraffin-embedded material, he precipitated the DNA. He then analyzed with globin oligos to see whether the DNA was intact, then he PCR'd with the BK and JC oligos, did a southern blot, and he probed with a specific oligos for either BK or JC.

And this is one of the original blots that John got. It's what we call low stringency. Actually, the whole PCR reaction was of low stringency and that cost us some consternation. And the other problem was that John, at that point being a physician, didn't quite understand the necessity for monitoring temperature in southern blots.

But what we found in some of the first samples was that there was a -- for example, we could get amplification in many samples of BK, for BK and even of JC, but many, many times these bands were hybridizing in both situations to both the BK and JC probe. And just for interest sake, this went on for quite a number of months, about three months, before I became a little frustrated and I told John to stop this and just sequence one of them so we could figure out what was going on.

And this was when the first matrices surprise happened. And that is, when John sequenced two of these bands he found that they were neither JC or BK but SV40. And there's a very characteristic nine base pair change in this region that he sequenced between those viruses.

And when we went back and looked at the primers, indeed the primers that we had -- since this region was so conserved, would have amplified SV40 under our conditions. And therefore, we went back and we used those primers with now a probe that was more specific for an internal region of SV40, and we used higher stringency conditions in the southern blot at 52 degrees -- and John had realized the importance of temperature by that time I think.

And most of our samples then, were amplified and blotted with this SV40 specific probe, although still -- and I think this is a topic for further discussion -- some had also BK or JC sequences in them, which we could not figure out why, but we were still now stuck with the fact that we had sequenced SV40 and that many of our samples were amplified and being identified with a probe for SV40.

John then went back and made new primers. The original primers were these right here after the splice site, PY forward and PY reverse with the probe in the middle.

John made now, another set of primers that amplified across the intervening sequence, SV forward-2 and SV reverse, and also another set of primers that amplified a very small, 107 I think, base pair region that lie just outside the previous amplification.

So when John -- we were concerned here that maybe there was some CDNA contamination in the laboratory. Contamination of course, is going to be a major issue in our discussion.

And so for example, he took some of the tumor DNA -- now this is, we have found that -- we were lucky in that our first set of primers amplified about 170 to 180 base pair fragment, because from the paraffin-embedded material it was very difficult to get long amplifications, and so when we got some fresh tumor specimens it was possible to do these longer amplifications.

And this for example, is an amplification with those long primers, with those primers that amplified across the intervening sequence, and it generated a band which could be specifically cut with an enzyme that would only cut SV40 and not BK or JC. So this is, besides sequencing, another way we determine that some of our samples were probably SV40.

At that time when we got this result, I wanted to verify the data with at least some immunohistochemistry. And so I called up Janet Butel and along with Milton Finegold, the pathologist at Texas Children's Hospital, we did immunohistochemical studies of these tumors -- and this is an example of immunohistochemistry with a large T-antigen of an ependymoma, and you can see that there's some very darkly-staining nuclei in this field, whereas the others don't stain.

And this for example, is a coriplexus tumor which is hard to see, but there are very darkly-staining nuclei here. And I think we'll get to the problems with the immunohistochemical cross-reactivity of different T-antigens, but this was the data that Janet and Milton obtained.

We went back then, with the set of primers that amplified that short region, and re-analyzed all the specimens that we had in hand. And this is the table of all the tumors we had at that time. And so we had 20 coriplexus tumors, ten of which amplified with those short, 107 base pair amplifying primers, ten of 11 ependymomas amplified.

We put neuroblastomas -- we analyzed those because transgenic animals with JC or BK also give rise to neuroblastomas and so we were very curious in that, also because it was a pediatric malignancy, but we didn't find any sequences in those tumors.

These are other controls. We studied normal brains, seven normal brains. We had a problem in the beginning because it was difficult to find brains that were not pickled for a month, and seven specimens were capable of globin amplification.

One was positive for SV40, which we thought was unusual. This was a 28-week, premature infant that died shortly after birth. The neuroblastomas I talked about, and we also examined normal blood. Fifty were studied at time and they were all negative. Subsequently we've studied several hundred and they've been all negative.

So at that time in this paper, we concluded that a segment of DNA corresponding to SV40 T-antigen was amplified from these tumors, and we concluded that perhaps SV40 or a related virus -- at that time we didn't know whether it was intact SV40 since we'd only amplified from one part of the viral genome, and maybe some hybrid virus was involved here.

And the tumors that we were finding this in were similar to those tumors that were induced in experimental animals by SV40. And so the questions that we had at that time were: is this a hybrid virus; is there full length copies of DNA in these tumors or are these episomal or integrated; where else are these sequences found? And of course, cause and effect is always a problem here, and I think that that's an issue for discussion.

There were two very interesting patients in our first study. One was a patient with Aicardi Syndrome, and this is a very rare syndrome that has agenesis of the corpus callosum in it. There's associated other abnormalities and rarely there's coriplexus tumors in these patients.

But perhaps more interesting, one of our coriplexus patients was a member of a Li-Fraumeni family, and this of course, most are aware is, the proband is a young patient usually with a sarcoma, and there's two 1st-degree relatives with cancer -- oftentimes breast cancer or osteosarcomas.

At that time, there were two kindreds identified by Judy Garber and Fred Li, with coriplexus tumors as part of the phenotype, and subsequently several other kindreds have been found having coriplexes tumors as part of this syndrome. About half of these individuals will have some mutations in p53, and the other half I think, are still up in the air.

Because of the occurrence of one Li-Fraumeni patient in our initial series, and because David Malkin and Steve Friend were across town at Mass General, I asked David for all of the Li-Fraumeni DNA specimens he could give me in a blinded fashion. And so we analyzed 163 Li-Fraumeni-related specimens -- and I'll talk about how they were related in a minute -- because we were interested in trying to follow up this one patient.

We'd also done some other controls. We looked at Wilms tumor because it's a kidney tumor and these virus seem to be trophic for kidneys. Subsequently, we looked at lung cancer because afterwards as you'll hear, Michele Carbone had found the virus in mesotheliomas and we were interested in other lung-related tumors.

So for the Li-Fraumeni specimens we analyzed 163 of these, and only 19 were positive. Now we were amplifying for the large fragment -- that fragment that amplifies across the intervening sequence. And I think what I was struck most about this analysis -- and so my first surprise was finding the SV40 sequence, and my second surprise was in decoding the specimens and finding that five were from osteosarcomas, four were from the blood from osteosarcoma patients, and one was from a lymphocyte cell line made from an osteosarcoma patient.

Now, all of these DNAs have been prepared by David Malkin and Steve Friend. And so there was a preponderance here of, half or over half of these in a very large series were related to osteosarcomas. And at that time I think I got my first call from Michele Carbone telling me about the mesothelioma data. And I told him about the osteosarcoma data and he said, oh yes, we've seen that too.

And so at that point, that's when Michele and I started to collaborate on the bone tumors which he's going to tell you about. And we also told immediately, Janet Butel, who's also got some data on the osteosarcomas. So I'm not going to give any more of that except as an introduction to their talks.

So this is a further breakdown of the Li-Fraumeni. There were three from unknown tumor types. I would say in the 163 samples there were many, many breast cancer specimens because this was part of the Li-Fraumeni syndrome, and none of those were positive.

There were somatic sources, and this was from blood or fibroblast cell lines that had been made from these patients. But again, a number were related to the osteosarcomas.

So that in short, is the introduction to Janet's and Michele's talks, I think. And I think Dr. Butel and Dr. Carbone will give more details on the subsequent analysis of these sequences in brain and bone tumors.

And I would just like to point out my collaborators at the Dana Farber when we were there, was John Bergsagel as a post-doctoral Fellow who really, from an M.D. coming into the lab, did a spectacular job. And I have to say that we didn't publish the New England Journal paper for at least a year of repeating all of these things over and over again from John, and he was a very meticulous person.

Wendy and Kristie Johnson worked up the Li-Fraumeni specimens in my laboratory. We had a very nice collaboration with the Baylor group, with Janet, Milton, and John, and at the University of Chicago, now at Loyola with Michele Carbone.

And thank you very much.

CHAIRMAN KIRSCHSTEIN: Thank you, Dr. Garcea. Before we go on to the next talk, if Dr. Mutti is in the audience, I understand he wants to talk to me, and if he'll come up here we'll try to settle the problem before he's announced to speak.

Our next speaker is Dr. Janet Butel from Baylor College of Medicine. She will talk to us on evidence for the presence of authentic SV40 in human brain and bone tumors. Dr. Butel.

Return to Table of Contents

DR. BUTEL: Thank you very much. I'm going to pick up the story now from where Bob Garcea left off. We thought about his finding of the SV40 DNA fragment in human brain tumors and two major questions emerged: did the DNA represent authentic SV40, or was it a new virus or some recombinant virus; and if in fact it really was SV40, were there human-specific variants that we could identify?

So in the next 15 minutes I'm going to present some evidence that suggests very strongly that authentic SV40 is present in at least a few human tumors. And since I'm going to go over several types of data rather quickly, I wanted to just summarize what the main points were going to be.

And that is, by PCR we've been able to detect SV40 fragments from four different regions of the viral genome, and sequencing showed that it's really SV40 DNA. The regulatory region structures are typical of a non-duplicated enhancer region that's found in some natural isolets.

We detected variability in the C-terminal T-antigen gene sequences. We isolated infectious SV40 from one human brain tumor, and as Bob mentioned briefly, we have some similar results with bone tumors.

So Bob sent us blinded samples of DNA from some of the brain tumors for a follow-up study. And this map shows the SV40 genome and it shows the positioning of different sets of primers that we use for PCR analysis. Here was the Rb proximal region that Bob described.

We looked at the regulatory region of SV40 down here at the end of the T-antigen gene, and then a region in VP1, the structural protein of the virus. Our idea was, that if separated regions from the viral genome could be detected, then most likely the virus that was present in the tumor was SV40.

Source controls for the regulatory region studies. We used a series of plasmas that John Lednicky had constructed previously, and these contained one, two, or three copies of the 72 base pair region; that's the enhancer region. Lab strains usually have a duplication in this region and some natural isolets from monkeys have had a single copy.

This is a gel showing the PCR results using these primers from the regulatory region. A number of these blinded brain tumor samples were positive, and the size of the fragment that was generated was the same size as the fragment containing a single 72 base pair region.

When John sequenced those PCR products, in fact the sequence was exactly SV40 from each of those, and in fact, they did contain non-duplicated enhancers.

This shows the results of PCR assays using the VP1-specific primers. The same tumors that had been positive for the regulatory region were also positive and generated a VP1 fragment. And the sequencing of those PCR fragments showed that it was an exact match with SV40.

We then examined the samples using the T-antigen primers. Now, from everything we know about SV40 we would predict that the full length T-antigen gene ought to be present if infectious virus were present or if T-antigen were involved in tumor development. And using the primers from the C-terminus of the T-antigen gene, the same tumors that had been positive for the other parts of the viral genome were positive and yielded products.

When the products were sequenced there were some nucleotide changes that were detected. This is compared to strain 776 here at the top. Each of these slides represent something slightly different about the sequence and of the five brain tumors that were sequenced, each one yielded a slightly different T-antigen sequence.

And some of those nucleotide changes would result in amino acid changes when you look at the predicted amino acid sequence for T-antigen. And that's what this slide shows. There were some substitutions, there were some deletions, and there were some insertions in what was found in the brain tumors.

We tried to isolate virus from the brain tumor samples and we did succeed in one case. The tumor DNAs were lipofected into monkey kidney cells: both TC7 and CV1. Only one of the samples, sample number 12, produced CPE. It took a pretty long time for the CPE to show up -- six weeks. The cells were extracted and John cloned the DNA and we called this virus isolet SVCPC.

The virus induces typical CPE vacuolization. We're starting to characterize the virus. Renee Steward in the lab has done growth curves comparing our Baylor wild type strain of virus with the human tumor isolate, SVCPC, in both monkey kidney cells and in human cell lines. This happens to be a renal tumor cell line that we got from the American Type Culture collection.

And the point is that both the wild type virus and the human isolet grow very well in both monkey and human cells. So it's clear that SV40 can grow well in human cells.

The next question is, were there changes in other parts of the T-antigen gene among different virus isolets? So Renee sequenced the entire early region, not only of SVCPC but of several other isolets including two old human isolets, SVMEN isolated by Dr. Kreig in 1984 from a meningioma, and SV40PML isolated in '72 by Drs. Weiner and Shah.

There were a few nucleotide changes that were found but no huge differences, and bear in mind we're comparing here both monkey and human isolets and viruses that were recovered over a span of about 35 years.

Now, when the nucleic acid sequence was converted to the amino acid sequence, it was really remarkable to discover how conserved the T-antigen protein is. There were no changes in the protein at all among any of these isolets until we got down to about the last 90 amino acids, and then we saw a cluster of changes. So on this basis we refer to this as the variable domain of T-antigen and this is the region where we're seeing some variation in the tumor associated sequences.

Renee has just about completed sequencing the late region of SVCPC. This slide summarizes the results from the coding regions. If there's a vertical line below this horizontal bar that means there's an amino acid change, and we didn't find any amino acid changes in the agnoprotein, in VP2 or VP3, and a single change in VP1. So it's clear that these human isolets are typical SV40.

Now, when I heard from Bob Garcea and Michele Carbone that they were finding SV40 DNA in osteosarcomas, we wondered first of all, could we confirm that on an independent set of samples, and then secondly, perhaps it would be possible to identify a distinct bone tumor associated virus variant.

So my pathologist colleague, Milton Finegold, obtained for us ten osteosarcoma samples from St. Jude's Hospital, and this slide just summarizes some of the patient information. I want to point out that all of these patients were born 1965 or later, after the use of the contaminated vaccine was discontinued.

The samples were provided to us blinded, they were extracted -- John Lednicky will discuss some of the important steps in processing these samples and analyzing these samples by PCR during the panel discussion. But we used the same four sets of primers to analyze the osteosarcomas that we had used in the brain tumor study.

This shows the PCR results using the primers for the regulatory region of SV40. Five of the ten osteosarcoma samples were positive and that they yielded a product. And when John sequenced those, each contained a single 72 base pair region -- that is, a non-duplicated enhancer -- and the sequence was typical SV40. We also used primers specific for JC and BK regulatory region and didn't pick up anything in these osteosarcomas.

This just summarizes the VP1 results from the osteos. The same five samples were positive and when the products were sequenced, again, there was an exact match with SV40, VP1.

The T-antigen gene was also present in the same five samples. We got positive signals with both sets of primers: those from the Rb proximal region and then those from the C-terminus of the T-antigen gene. I'm just showing the amino acid calculated sequence.

The take-home lesson is that again, that each of the osteosarcomas contained a slightly different T-antigen sequence when we looked at the variable domain of T-antigen. Three of the tumors yielded new sequences that we hadn't seen before, and in two cases, it was a known sequence.

So finally, I want to end by telling how we addressed the question of whether this variable domain at the end of the T-antigen gene is highly mutable -- meaning it changes rapidly over time -- or in fact, is it stable and the variation that we were observing would reflect stable strain differences.

So to do that we analyzed high and low passages of two different strains of SV40 for which we had a known history. And this study was possible because I had frozen away in my freezer, some low passage stocks that had been frozen down for more than 25 years. And I'm just going to show you the story with VA4554.

This virus was received at Baylor in 1967. It was passed a couple of times and a stock frozen down in 1971. We reconstructed the history. We know that Dr. Hilleman sent the virus to the Enders laboratory in the early-60s.

Peter Tegtmeyher used this virus, manipulated it in all of his genetic studies. Then in the mid-70s he sent the virus to Judy Tevethia's lab where she used it in her genetic studies. And then in the early-80s Judy cloned the virus and sequenced it.

So John Lednicky went to this very low passage stock of VA4554 that we had, he cloned out isolets from that stock, and then compared the sequence of these isolets with this sequence that Judy had determined on this lineage of virus that had a very different history.

John cloned out two types of viruses. One had an archetypal regulatory region, one had a duplicated enhancer from the low passage stocks. The sequence of each was exactly like the sequence that Judy had determined for her virus.

And then finally, when we examined the sequence at the C-terminus of the T-antigen gene, the sequence was exactly what Judy had determined for her virus. So our conclusion is that the sequence is stable and we think that these variations reflect viral strain differences.

So I've run out of time and in summary, this just reiterates what I had said earlier was going to be the take-home lesson. And these are the people who have been involved in the work.

CHAIRMAN KIRSCHSTEIN: Thank you very much, Dr. Butel. I want to repeat, if Dr. Mutti is here I need to speak to him.

The next presentation will be by Dr. Michele Carbone, who is at the Loyola Medical Center in Illinois. Evidence for SV40-like DNA sequences in human mesotheliomas and osteosarcomas. Dr. Carbone.

Return to Table of Contents

DR. CARBONE: Thanks. And first of all, I would like to thank Dr. Lewis and Dr. Levine for having invited me here. I did my post-doctoral training with Dr. Lewis and I was hired by Dr. Levine as a visiting associate in his lab, and he gave me the possibility to start my career, so I'm particularly grateful to the organizer of this meeting for what they did for me.

Today I am going to talk about the evidence that we have for the presence of SV40 sequences in human mesotheliomas and in human osteosarcomas. Tomorrow I will present our new data that suggests that not only these sequences are present in the tumors, but they may also in some instances, contribute to the transformed phenotype. And tomorrow we'll also talk about, as before, oncogenicity in hamsters.

In order to start my -- this is SV40 viruses, small DNA tumor viruses. In our studies in humans we are prompted by our findings in hamsters. What we found when we injected SV40 intracardially into animals, was that only particular tumor types developed, specifically, mesotheliomas, osteosarcomas, lymphomas, and on the -- sarcomas.

The reason that we injected the virus in the heart was to expose more cell types to the virus and to see whether every cell could be transformed by SV40 or only specific cells were transformed. Now, from our tissue cultures -- not ours, but from tissue culture experiments that other people did -- we know that SV40 will infect more or less, every cell. Even in humans it can easily infect karyotinocytes.

You can clearly see that the most common cancer, at least those that developed in humans, never developed in a hamster when you inject SV40. We have never seen a carcinoma. I was particularly struck by the fact that the mesotheliomas developed and so we repeated the experiment injecting as before, into the pleural space. And in that case, 100 percent of the animals came down in tumors in three to six months.

Other investigators had found that when you inject SV40 intracranially in animals, only ependymomas or ancyroid plexus tumors develop. So the conclusion of this experiment is obviously, SV40 is a virus that for some reason, even if it can enter different cell types, will transform only particular cell types. They must be more susceptible to transformation by this virus.

Again, I was particularly struck by this tumor. This is how it looks histologically. This is a mesothelioma. Mesotheliomas are a tumor with incidences increasing inordinately. Today we have 2,000 or 3,000 cases in the United States. Actually, I just heard from the -- that 4,000 are projected for '97.

And this is a high number if you consider that until 1950 or so most books of pathology denied the existence of mesothelioma between they were so rare in that many people so that they didn't exist at all. So it's like going from zero to 4,000.

The reason for that is the use of asbestos. At the beginning of the century asbestos was used largely in all the western world, and it's obvious that exposure to asbestos induced mesotheliomas after approximately 20 to 50 years.

However, 20 to 50 percent of mesotheliomas -- and that depends what study you look at -- are not associated with asbestos exposure. Now, that's a big number because obviously there is a large number of mesotheliomas that are increasing from 1960 for which we cannot account. So we ask it, could SV40 or a related virus be related to the development of mesotheliomas?

And this slide summarizes whether that time was known as SV40 human pathogen. SV40 can transform human cells in tissue culture. SV40 human transformed cells induce tumors when injected into human volunteers. Millions of people were injected with SV40 contaminated adeno and polio vaccine and after 1963 vaccine should be SV43.

While I was trying to convince my -- at that time I was supervising Dr. Levine's lab -- but I couldn't convince them to look at these SV40 tumors because obviously it was a very risky project. So while I was struggling to convince somebody to work on this with me, Dr. Garcea published in New England Journal of Medicine that 60 percent of human ependymomas contained SV40-like sequences.

It's a study that was later confirmed by other investigators, including a recent paper by Martini, et al, in cancer research a couple of months ago, and from the laboratory of Dr. Butel where Dr. Lednicky isolated infectious SV40 from one of these tumors indicating that at the least, in that case, the SV40 life sequence was infected as we thought.

So eventually I was able to convince Dr. Procopio that had done his post-doctoral training at the NIH and was a tenured professor in Italy, to come to the NIH and do this work together to see whether there was any SV40-like in these mesotheliomas.

We used the same technical approach that had been used by Dr. Bergsagel in his studies and the reason is, as Dr. Garcea indicated before, that the Rb binding domain should be there if the T-antigen is doing something, because the Rb pocket binding domain is that region of the antigen that binds Rb, p107, Rb2 or p130. So the cell are proteins that must be inactivated by the antigen in order to receive transformation, therefore, that would be the region that you most likely would expect to find.

We started a great collaboration with Dr. Harvey Pass, that at that time was the Chief of Thoracic Oncology here at the National Cancer Institute. Harvey has an incredible collection of mesotheliomas and he candidly gave access to us to his collection and also he worked very closely with us during these experiments.

And this is a summary what we found; that is, 29 of these 48 mesotheliomas that Harvey gave us tested positive with primers that amplified the Rb pocket binding domain of T-antigen. Sequence analysis confirmed that those sequences, the type that amplified with the specific proper SV40, were in fact SV40, and the arrow points to that unique region of SV40 that Dr. Garcea already talked about so I'm not going to repeat it.

Immunoperoxidase staining indicated that some mesotheliomas cells contained a nuclear antigen that strongly reacted with the monoclonal antibody against, as before, the T-antigen. So we were seeing either as before, the T-antigen or something very close to it.

And actually, if you look at that again you can see that only the tumor cells stain and in fact, that the reactive fibroblasts that are around the tumor cells do not stain. And immuno-precipitation studies from frozen tissue precipitate a 90 kilodalton protein with the monoclonal antibody against the antigen that reacted in western blot with the monoclonal antibody against the antigen.

And this was the conclusion of that paper, we found SV40-like DNA sequences in 29 of 48 mesotheliomas stasis and demonstrated the antigen expression in 11 of 14 specimens. The associated lung did not contain SV40 sequences although they contained asbestos.

We suggested that an SV40-like virus may act independently or as a co-carcinogen with asbestos. Moreover, the selective T-antigen expression by mesotheliomas and not the surrounding pulmonary barenchyma, may have diagnostic and therapeutic implications.

At that time I gave the code to Bob Garcea and from what we knew in animals, there was another possibility for finding SV40 among the many human cancers was obviously osteosarcoma. And the other question was, are there other tumors in humans that may contain SV40 sequences?

So we started a collaboration -- me, Bob, the laboratory of Dr. Pass, and the laboratory of Dr. Procopio in Italy -- and we studied 345 different human samples including 159 bone tumors and sarcomas for SV40-like sequences by PCR.

Only after all four laboratories completed the analysis was the code identifying the origin of the bone tumor specimen broken and the results compared. And these are the results that we published in Oncogene last summer.

Fifty-three of 159 bone tumors and sarcomas -- that is exactly one-third -- tested positive. Of the 186 known bone tumor sarcoma samples only one out of one neurofibromatosis type, one was positive. And the other 185 were all negative.

Ten samples, all from bone tumors, gave conflicting results, meaning that at least one reaction in one of the four laboratories was positive. These samples were considered negative.

And these are some of the results that we obtained. In this particular study we used a different set of primers that are the long primers that Bob referred before. And they are indicated up there -- I don't have a stick to indicate them -- but you can see a hoop at the top line. They are indicated by SV42 and SV-rev.

They amplify the same RV pocket binding domain that we used to amplify before that is indicated by the primers before SV-rev. But they also amplify the intraregion of the antigen. And the reason to include the intraregion was that while you want to find SV40 T-antigen there so that's why you put the Rb pocket binding domain there, it would have been nice to find some mutations, because there is always the question when you do PCR, people will ask, are you sure that there's no contamination?

Well, this is what we found. This is the 574 page paper that you expect to amplify with this parameters. And the last time in the right under the H letter that indicates hamster, is a hamster osteosarcoma. So you can see that in addition to the 574 base pair band there are other bands with different molecular weights.

Particularly prominent is the band around 300 or so base pair. When Dr. Rizzo, that is my research associate, sequenced those bands that were hybridizing with the probe specific for SV40, the sequence indicated that that was SV40 T-antigen, but that there were deletions within the intraregion. And we never found deletion outside there; we never found deletion in the Rb pocket binding domain.

For some samples we also tested the other region of SV40 genome to see whether they were there, and for example for those two samples indicated with the numbers 103 and 105 that were positive for SV40-like sequences, in panel B we are amplifying the capsid throat in PV1 and in panel C we're using two sets of parameters that amplify the carboxyl terminal domain of the antigen.

So for at least these samples a lot of the SV40 genome was there. We also sequenced the regulatory region of SV40 and we found 272 base pair.

This slide points to an intriguing coincide that comes out from these studies. And these are the tumors associated with SV40 in humans. Ependymomas are induced by SV40 hamsters and in humans contain SV40-like sequences. The same is true for choroid plexus tumors, the same is true for mesotheliomas, the same is true for bone tumors, the same is true for sarcomas.

The last one, true histiocytical lymphomas that are induced by SV40 hamsters, are so rare in man that many people doubt that they exist at all.

And this is the conclusion for, this talk at least. The significance of SV40 and SV40-like sequences in human tumors is presently unclear. Specifically, it is unknown from what source these sequences originate; whether these sequences contribute to tumor development. And again, I am going to present some data suggesting that in some cases they can, and Dr. Weiss has been very kind to offer me to present this data at the beginning of this round table tree tomorrow afternoon.

And whether these sequences can be used as targets for designing new immunotherapeutic and molecular approaches for the management of malignancies expressed in T-antigen. And while that can seem futuristic or too much optimistic, I think that in fact, it's a very exciting area.

This is one of the best immunoperoxidase that I have, of course, but this is a human mesothelioma and you can see the staining of these mesothelioma cells that clearly distinguish them from the surrounding non-malignant cell. Now obviously the presence of a unique antigen on a tumor cell gives you, at least in theory, the possibility to attack those cells, and we currently are working on this hypothesis.

Thank you very much.

CHAIRMAN KIRSCHSTEIN: Thank you, Dr. Carbone. Our next speaker is Dr. Allen Gibbs who will present SV40 DNA sequences in mesotheliomas. Dr. Gibbs is from Llandough Hospital the United Kingdom.

Return to Table of Contents

DR. GIBBS: First of all, thank you. I want to thank the sponsors for inviting me to present the work that has been conducted in Cardiff in this. I'd just like to point out that the pronunciation of the hospital is "thrandock" and not "thrandough", but I will excuse the American pronunciation.

I'd also like to acknowledge Dr. Bharat Jasani -- he was also at this meeting -- who really is responsible for the actual methodological approach of the study.

Why did we conduct this study? Well first of all, we had seen the paper by Dr. Carbone on the finding of SV40-like sequences in mesothelioma; that my particular hospital has had a long association with examining mesotheliomas and I have an archival store of several thousand mesotheliomas. And these go back to the days of Chris Wagner who worked in my hospital and in fact, was the person who put the association with asbestos and mesothelioma on the map in 1960.

And we were particular interested, really, in looking at our archival material because a considerable proportion of that has been well worked up, both diagnostically and from the point of view of exposure data, and that includes a very detailed mineral analysis on the lung tissues.

So we were interested in trying to exploit this material, looking at the possible role of SV40 and asbestos in causation of mesothelioma. So this was a, basically a pilot study to examine frozen and archival, surgical and post-mortem mesothelioma tissues for SV40 DNA-like sequences, and we were using the method of PCR.

I'm not going to talk about the Tab expression using immunohistochemistry now, but it may possibly come up in one of the later sessions.

The materials and methods that we started with are cases. There were nine pleural mesotheliomas. I also took nine adenocarcinomas that were metastasized to the pleura, and nine reactive pleura. These varied from reactive eosinophilic pleuritis to non-specific chronic inflammation of the pleura.

All the cases had formalin fixed and paraffin embedded material, but in addition I mentioned to, as we went along, obtained four mesotheliomas which we had frozen down. These were complementary to the four other surgicals -- three surgicals and one post-mortem.

There were three post-mortem cases and six surgical biopsy cases when it came to the mesotheliomas. The breakdown was eight males and one female. The age range varied from 38 to 73 with a mean of 58.8, and there was a mixture of histologies here. There were both the epithelial, biphasic and the sarcomatous types. All of these had a history of exposure to asbestos.

We basically used the same technology that Michele Carbone used in his paper, and we employed the Bergsagel primers, the direct known as the SV40, specific DNA sequences which was 107 base pair, and then one -- sorry, 105 -- and then the 172 base pair sequence that recognized the papomaviruses BK, JC, etc.

And for controls we had some SV40 transfected human thyroid cells, one positive control and one negative, and the positive contained one copy per cell. And employed the controls both on frozen material and on formalin fixed, paraffin embedded material.

This shows the gel electrophoresis on the -- here's the mark here which tells you the size of the protein that's found. This was one of the mesotheliomas which was negative. This was the positive thyroid transfected cells, and then these other four were the four positive mesothelioma cases.

This shows the controls of three positive controls; again the marker here. This was the frozen, positive control and these two were both fixed in different ways, and they were positive, too.

So to summarize the results of our PCR analysis on these cases, we found the SV40 specific segment detection in four out of the nine mesotheliomas: one out of three post-mortem cases, three out of six surgical biopsy cases. We didn't find any present in the reactive pleurae and we didn't find any in the adenocarcinomas.

Just to go into some more detail on the actual mesotheliomas showing the concordance, or discordance between the SV40 and the papomavirus sequences, that the SV40 here was consistently present, both in the frozen and the paraffin material for each case that was positive, but that we got some cases which were negative to the SV40 which had the other sequence.

And again, these show the same thing for the extra cases. These were all just the paraffin embedded surgical cases.

This is just to compare our study with the published literature on studies of mesothelioma that -- Michele Carbone has already gone into this; that there's a somewhat similar rate in the two studies. A recent study by Strickler in fact, failed to find any SV40-like sequences in 50 cases, and then there's another study by Cristaudo that found positivity in 72 percent; again on archival material.

So the conclusion to the study was a simple one. That we found SV40 DNA-like sequences in some British mesothelioma cases. Thank you.

CHAIRMAN KIRSCHSTEIN: Thank you, Dr. Gibbs. The next talk will be shared by two speakers. First, Dr. Luciano Mutti who will speak on SV40 on mesotheliomas. Dr. Mutti is at the Fondazione San Maugeri in Italy. And the second part of the talk will be presented by Dr. Antonio Giordano from Jefferson Medical College, who will talk on a retinoblastoma family in mesothelioma. Dr. Mutti.

Return to Table of Contents

DR. MUTTI: First of all, let me thank the organizer of this fine workshop for my invitation. My work has been -- I share my work with Dr. Giordano from Philadelphia. And very briefly, as you can see, SV40 has been able as stated, to induce some tumors on hamsters, and as far as mesothelioma is concerned, injection of SV40 in pleural space, able to induce in 100 percent, mesothelioma, and while injecting in peritoneal space, is able to induce it near 50 percent of the animal.

In humans, raw data showing that SV40-like sequences are detectable in mesothelioma cells in about 60 percent of cases, and the T-antigen is detectable in a good number of cases. And SV40 has been so considered as a powerful cofactor in addition to asbestos fiber, for example, in inducing this tumor.

We studied this population of ten patients with this type of size of neoplasperitoneoma and nine of pleural mesothelioma, and with this type of histology. Some of these patients had not been exposed to occupational asbestos fibers, but had been exposed to an environmental exposure of a -- they lived in an area wherein a plant working on asbestos items had been working for at least 20 years.

So pollution of at least free fibers for meter cubed has been, nowadays, even now, detectable in this area. And these patients had been an occupational exposure for this period.

And there is another of these patients that had been exposed both to occupational and environmental exposure, because they lived in this plant -- they worked in this plant and they lived in this area.

Very briefly, we found SV40-like sequences in three out of ten patients we studied. And as you can see, one patient had both occupational and environmental exposure and the other two, only occupational exposure.

So we can see that there are some implications of these conclusions because SV40 can be considered actually as a tumorigenic factor, but it's important for us to assess -- to state that SV40 could be considered as an inducer, as a tumor, as an antigen, for mesothelioma.

So far it's not been possible to induce an effective neuro-response in mesothelioma cells, but there are a lot of studies that can demonstrate that a self antigen, a tumor or self antigen of mesothelioma cells do exist.

And one of the most important strategies to increase the new response is to increase the self-antigen expression, and with perspective in treatment of the -- of mesothelioma cells or mesothelioma is, without any doubt, to find out tumors that are dangerous, such as the antigen against it's possible to induce an immune response and a T-mediated response.

Just very brief, so I'll leave my talk to Dr. Giordano because he carried out another study about these patients, studying especially every family protein.

CHAIRMAN KIRSCHSTEIN: Dr. Giordano.

Return to Table of Contents

DR. GIORDANO: Okay, so thank you, Dr. Mutti, actually for allowing me to present some common data that we carry out in collaboration. In the first light --

CHAIRMAN KIRSCHSTEIN: Before you start, I need to make an announcement. Dr. Chanock has an urgent message at the message center.

DR. GIORDANO: Okay. So very quickly, the retinoblastoma families is formed by three members, Rb 107, and the Rb2 p130. The Rb2 p130 is the family member cloned a few years ago in my laboratory -- we code Rb2 and then code for under 30 kilo often. The 03 family member, okay, they share a pocket region, so-called pocket region because the functional domain that is a point of the structure is tackled by different genus tumor viruses like the virus C1 as is before T-antigen. Is seven of papomavirus and do it as a function of domain because this is how this family or protein go through the cell cycle machinery.

So once we cloned Rb 2 we went on in the collaboration with Dr. Knudsen at Fox Chase Cancer Center. And with Dr. Knudsen we marked the Rb 2 on a chromosome 16, 12.2. And that start actually, our interest in the mesothelioma, because 16.2 by Dr. Testa's group at Fox Chase has been found also being the region found deleted in 16 out of 25 mesothelioma tested by Dr. Testa.

So we went on and carried on a series of experiments. And as a first characterization that we did, we found that Rb 2 is important to growth suppressor gene. When you applied to different tumor cell line the gene, basically you express in appropriate way, you achieve a growth suppressive property that is pretty dramatic in several tumor lines: like osteosarcoma, like glioblastoma, like nasopharyngeal cancer.

So by immunochemistry, okay, immunochemistry we went home and first test a -- cell line cancer tissue. And what do we found? We found that in line cancer actually, there is -- we divide in different histological groups. We found very interesting, retinoblastoma family in line cancer, especially the Rb 2 P 130, as in vast relationship, okay, with the aggressiveness of the disease. More aggressive is the tumor, undetectable is the Rb 2 gene.

So again, we went on and start the collaboration with Dr. Mutti and we start to test the mesothelioma lines in the tumors, actually, from tumors that also we obtain from different sources.

If you see here, we just use the primers that amplify the regional T-antigen that is required to tackle the retinoblastoma family. And we found the sequences of T-antigen in four of the patients of the nine patients that we tested.

But more interestingly, when we performed immunohistochemical analysis, when we perform immunohistochemical analysis on these mesothelioma, alike up in the line cancer, we find 11 of the Rb family do not change. So there are high, medium high level of the Rb family in all the mesothelioma tissue that we tested.

So T-antigen, sequence of T-antigen we found in these patients, and when we extract actually, the protein from this fresh tissue using antibody -- monoclonal antibody, this T-antigen -- we find the protein the right size, okay, in which T-antigen migrate. This one is a cos cell line that -- this place, SV40 T-antigen, and this one a cell line HLA 60. This does not see any DNA tumor virus. So there is no expression of T-antigen.

So Rb family highly expressed in all mesothelioma lines. So a mechanism that the work probably we could predict, I mean, just considering the data by several laboratories, that T-antigens and other genus tumor virus binds in a physical association with the original blastoma family, lead us to an experiment in which we took the source of T-antigen in these patients having mesothelioma, and we ask if there is any physical association with the original blastoma family.

As you can see here, okay, the Rb, the p107, and the pRb 2 physically interact; they interact in a physical pro-interaction with those three family members. So no wild conclusion from this of course, but clearly there is no blastoma family, okay, member in -- that in line cancer for instance, and don't see intermediate cancer in the screening of studies we did, correlates, especially the Rb 2, P 130, correlates with the aggressive disease in mesothelioma do not change.

And one of the mechanisms that probably we suggest is that known mechanism suggested by several laboratories that T-antigen by binding these original blastoma family does not allow them to perform the suppressive property. Thank you very much.

CHAIRMAN KIRSCHSTEIN: Thank you, DRs. Mutti and Giordano. Our next speaker is Dr. Mauro Tognon from the University of Ferrara, Italy, who will speak on SV40 and mesotheliomas.

Return to Table of Contents

DR. TOGNON: Thank you very much, the organizer, for inviting me to this interesting meeting. And congratulation also, to the Chairperson for the very nice pronunciation of my name.

Probably the people are already tired to listen to the PCR amplification of the Rb pocket domains of the T-antigen, but in our investigation we did exactly the same thing. And in these experimental control we have in the first lane, the BK, the second, JC, and the SV40. One other, 72 base pair were amplified by the primer that is being named P4 and P5.

And as you can see ethidium bromide staining, we have the amplification of this region, and because this region is common to the three different polyomaviruses, we set up an experiment with the internal liogprobe that recognized specifically the three different polyomavirus.

We used a very high stringency condition just to be sure that this time we did the experiment with these two primers that amplify the Rb pocket domain of the large T-antigen which is common to the three polyomaviruses. All the time we have clear results by hybridization. In this case, we used the recombinant plasmid DNA that contains the entire genome of the three viruses.

In these experiments we analyze different human brain tumors and in particular I would like to say about the experiment, because we always start with 500 nanograms of human genomic DNA, and all the time we extracted DNA in -- what can I say -- in old fashioned way, we tried to avoid all the time to centrifuge or to precipitate the DNA with ethanol.

And the reason is that in our experience, if you precipitate the DNA with the ethanol, or you extracted DNA with the commercial kit, you lose the signal after the hybridization with the internal olioprobe.

So probably this makes sometimes the difference, is a possibility, between the high level internal percentage that we found in our human brain tumor and we can see a little bit later, also in the normal brain tissue, of the positive sample that we found.

In this case the control was always present. This is the 172 base pair. This is the only possibility because it's the control to see the amplified bands in the ethidium bromide staining, but in the other samples that are under analysis, the amplified bands never appear. In this case we amplify the DNA for 35 cycles.

The results come out only with the internal olioprobe that is specific in this case for SV40. As you can see, there are in these typical experiments, some primary tumors that come from ependymoma, across the plastispapilloma, plasticytoma, and neoblastoma. The neoblastoma was negative and also the sponsoblastoma was negative.

We analyzed also the normal brain tissue from people that died by car accident and we extracted the DNA in the same fashion, and the bands that amplify the control is always present. You can see the hybridization band that comes out after three or four days of exposure, our radiography, but none of the six samples that were analyzed in the experiment were positive for the SV40 sequences.

Okay, this is another experiment that we set up to investigate the presence of SV40-like sequences in the peripheral blood cells. As you can see, after the amplification and hybridization two of these samples are positive and indeed, we have the results of peripheral blood cells that sometimes are positive for these sequences, even if the percentage of positive sample in peripheral cell of normal individual.

This is not blood that comes from the patients but we took 70 different blood samples from the blood bank of the general hospital at the University of Ferrara.

Because we found the sequences specific for SV40, both in the tumor sample and in the peripheral blood cells of normal individuals, we think that probably there are sometimes contamination of the blood of our biopsies that come from the neurosurgeon.

And so we investigate the same kind of tumors, starting this time from the tumor cell line. All these samples were from glioblastoma cell lines. We analyzed, if I remember correctly, 18 glioblastoma cell lines. Nine were set in the lab of the Department of Pathology at the University of Verona, and they never worked with SV40, and the other nine were purchased from European culture collection.

And surprisingly, when we found that the results of our data after the hybridization with the internal olioprobes, comes out that practically all the Italian tissue cultures cell lines from neoblastoma were positive and none were positive from the tissue cultures that we purchased in England.

This is another data that comes from the experiments we did in the same fashion with other tumor -- primary tumors from the -- primary brain tumors. And in this instance we isolated the RNA from the tumor cell line, the glioblastoma cell lines, and among four different glioblastoma cell lines, three were positive for the messenger RNA. Even in this case, the experiments were set up with the specific primers from the Rb pocket domain.

This RT-PCR has been recently conduced, not only with the glioblastoma cell line but also with the osteosarcoma cell line. It's a new experiment that we set up recently. And some new data is now out from the normal tissue from people that are not affected by any kind of disease. These are three samples that come from buffy coats.

These are three different cell lines that were obtained from the Institute of Medical Genetics. Also these people -- of the University of Ferrara -- also these people never work with SV40. And these are three samples that come from the sperm fluid in another University, University of Modena.

As you can see, these experiments is only representative of the sample that I analyzed. Some of them are positive, some other are negative, and at the end I'll show you the table that summarizes all this data.

When we try to specify if the Rb pocket domains that the large T-antigen sequences are specific of SV40, we clone in sequence at the beginning from a glioblastoma primary brain tumor, a glioblastoma, the SV40 sequences, and they came out -- it was practically the same sequences of the SV40 wild type.

The difference that was also pointed out during the talk of Dr. Bob Garcea, is that the nine base pair insert that are present in BK and JC are always absent in the SV40 sequences, and also there are some other differences among the other bases that we sequence. So by comparison and by checking in the databank the sequence, we know there is a real SV40 -- I'm sorry -- okay, got it.

These are different sequences that were obtained from other primary brain tumors. This is an ependymoma and the sequence is the wild type SV40. These two come from two different glioblastoma cell lines, and we found a mutation in this position. The sequence is A, C, G, T.

And this is another primary brain tumor, a neoblastoma, that has two different mutations. Both are transitions in the first base of the triplets and at the end when we checked the amino acid we found that there are a substitution of the amino acid in this position.

Okay, after checking the specificity of the sequences for SV40, we checked also in two of the blastoma cell lines, the presence of the large T-antigen. And in this case we used a specific monoclonal antibody that react specifically with the large T-antigen SV40, and at the same time we compared these data with a polyclonal antibody that recognize SV40 and the BK virus T-antigen.

As you can see the results are similar in terms of positive data, but if you check here -- this is the control of BK transformed cell that is called T-53 -- this light doesn't react with the monoclonal antibody, specifically SV40 and the polyclonal antibody reactor, and we can see in the nuclei the positive fluorescence.

On the top we have the cos cell that are transformed with the large T-antigen, and expressed the large T-antigen in the nucleus, and as you can see even here, the monoclonal antibody, specifically SV40 react, and the polyclonal antibody that recognize both the large T-antigen, react with the same cells.

In a previous area of experiments we did more or less the same experiment in searching for the homology, not only for the SV40 but also for the BK. Actually, this was a first series of experiments and also in this case we used the internal olioprobe specific for BK to discriminate along the other polyomavirus sequences that eventually are present.

In this case, the same sample that I showed you before, ependymomas and plastispapilloma, so now to be positive for the BK, the same for the three or four blood cells, and the same data are for the RT-PCR in the glioblastoma cell line.

When we sequence the BK positive for the large T-antigen Rb pocket domains, it turns out in 12 different samples, ten from the tumor samples and two for normal tissue, that all the sequences were the same compared to the BK virus Dunlop strain. We couldn't find any mutation in the 12 different sequences that we performed in 12 different samples.

This tells me, I have to show you the table that summarizes -- wait a second, I go back for a second. No, it's just I want to show you this transparency. Okay. This is the table that summarizes the data for the detection of the specific BK virus DNA, and detection of SV40 DNA. And as you can see, all the primary brain tumor that was positive for the BK DNA were also positive for SV40 DNA in terms of much more positiveness for the BK with respect to the SV40 DNA.

Similar data were obtained also for the cell lines, and these tell us that the positive doesn't depend on the blood -- contaminated the samples, and we have some new data compared to the previous publication directed to the primary bone tumors, and the data are very similar to those that we had before from Dr. Carbone, Dr. Garcea, and Dr. Butel.

Approximately the five percent of the osteosarcoma are positive. We analyzed also some huge tumors, and it turns out that 33 percent are positive. And once again we analyzed those of the cell line and the osteosarcoma are 43 positive for SV40, but all of them are positive for the BK. And interesting, these giant cell tumors, these rare tumors, 80 percent are positive for SV40. And the small osteosarcoma is a particular kind of osteosarcoma -- 36 percent are positive for SV40.

Also some other tumors came out to be positive, but they are less of an extent for SV40 and positive also in the 42 percent for BK. But interesting, the normal brain tissue is quite different in terms of original data compared to the results that we heard before by other speakers.

And we have for example, the normal brain tissue, only one out of 13 is positive for SV40; the bone, none are positive; but the peripheral blood cell in general, that we obtained from density gradient centrifugation, 23 percent are positive for SV40.

But if you take the B and TD 4 side that are cell lines, in this case transformed by the ABV, you see that the SV40 is approximately 11 out of 15 samples were positive; 73 percent is quite high. Even the TD4 side contains the SV40 and as we published already, the nine out of 20 sperm fluids were positive for SV40.

The new and the last -- when you work with the PCR of course, you think sometimes you contaminate your sample. So this is a new analysis what we did very recently by PCR in human tumor and tumor cell lines from DNA that was extracted ten years ago, and in the laboratory that never worked with any kind of virus.

And as you can see from the results, the percentage of positive samples for SV40 are more or less the same that was entered in the previous table, both for the primary tumor and for the cell line.

So in conclusion, it seems to me that there are some different regional distributions of the SV40 sequences. And compared, as I say, to this country or in England, we have much more positive samples from the normal tissue compared to other data.

I have also a couple of more results I would like to tell to the people, that are obtained recently in the Institute of Microbiology by Professor Barbanti-Brodano, that unfortunately today is not here because he got a recent operation. And we tried to rescue the viruses of the SV40 by transfecting with lipofecting or lipofit -- I mean DNA from brain tumors, brain tumor cell lines, and from preferable cell experiments.

The only two isolates that we rescued so far were from preferable cell sample from the dysplasia of a vulva sample that was even in this case, obtained from a collection of a large DNA sample that was at minus 80 since 1985.

Now we are characterizing it to isolate it. The first analysis -- I mean the first sequence analysis seems to indicate that to isolate it at the two SV40, and in particular, the difference that we see as a variant are in the origin of replication. This data are just preliminary so we have to control a little bit in detail, these results.

Thank you very much.

CHAIRMAN KIRSCHSTEIN: Thank you, Dr. Tognon. Our final presentation today is by Keerti Shah: A Search for SV40 in Human Mesotheliomas.

Return to Table of Contents

DR. SHAH: We were also very intrigued by the finding of Dr. Carbone about the presence of SV40 in mesothelioma tissue, and we attempted to reproduce that. And this has been written up and published in this journal in 1996. We were not able to detect simian virus 40 sequences in pleural mesothelioma.

We also looked at serum specimens from mesothelioma patients and osteosarcoma patients for antibodies to simian virus 40. And we were not able to detect any evidence of SV40 neutralizing antibodies, the results documented in this paper.

And more recently -- something which is not in this paper -- we tried to look for simian virus 40 in urine specimens collected from immuno-compromised and non-immuno-compromised patients. We looked for BK virus, JC virus, and also simian virus 40, and we were not able to detect any simian virus 40 sequences in these urines.

So this is the mesothelioma study. There were 50 pleura-mesothelioma samples collected at the Armed Forces Institute of Pathology over a number of years -- this was from 1987 to 1992 -- and obtained from many different hospitals and clinics in the United States.

The patient's diagnosis was between -- age year diagnosis was between 43 and 88, and median age was 68 years. These were archive samples so we tested the paraffin sections with proteinase K and we amplified them with two sets of primers for SV40, and another set of primers for beta globin to see if the DNA was suitable for PCR.

The two primers for SV40 we used were described by Dr. Garcea. They amplified 103 and 205 base pair regions of the T-antigen for SV40. And they are shown here in illustration.

This is what we found. These are three identical filters. The top one is with beta globin probe, and these are amplified with the beta globin primers. And most of the specimens we were able to amplify beta globin easily. There were some failures -- one here, one here -- so in 48 out of the 50 mesothelioma tissues we were able to amplify beta globin so that the DNAs were thought to be suitable for PCR studies.

These are the same identical filters here and here they are tested with an SV1 probe, a probe that was described by Dr. Carbone. These are all mesothelioma samples. None of them show up SV40 sequence. And we have done a titration of SV40 cos-1 cells.

Cos-1 cells have one copy of the SV40 genome per cell, and these are 300 cells, 30 cells, three cells, and actually less than one cell. So with this particular primer path we were able to amplify one copy of this SV40 genome just by this particular data with the second set of primers; again 300, 30, and three SV40 copies.

So that you -- we thought that our technique was quite sensitive but we were not able to detect any SV40 genomes. So we also looked at 105 serum specimens: 35 were derived from mesothelioma, 35 from sarcomas, and 35 controls, matched to the mesothelioma patients. Then 97 -- and these were tested in SV40 plaque neutralization assay in BSC-1 cells -- 97 were completely negative and five were partially protective.

They did not completely neutralize the simian virus 40, but reduced the number of plaques. And they were scattered in the three groups. There were three out of 34 mesothelioma, one out of 33 osteosarcoma, and one out of 35 controls.

Then we looked at the urines and we collected -- these are 165 urines provided to us by Dr. Strickler and Dr. Gater -- and they came from a study of homosexual men, some of whom were HIV positive -- seven urines which were HIV-positive men and 78 from HIV-negative men. The median age was 38/39; ethnicity, they were 91 percent white.

And this time we thought that if we do not have good positive controls for SV40 no one is going to believe our data. So we took SV40 cos cells and spiked normal urine with the cells, and these tubes we sent back to the NCI where they were processed with the other urines.

And then when we received these 165 urine specimens they also contained other tubes marked similarly, which could not be distinguished but which contained 17 specimens which came from the spiked urine which were the SV40 positive urines, and contained approximately 300 cos-1 cells or 300 copies of the SV40 genomes.

These are the SV40 data. There are -- 17 urine specimens that we got were positive for SV40 and there were -- each of the 17 of the positive controls, the urine specimens that were spiked with SV40. The other 165 urine specimens, none of them contained SV40, but the prevalence of BKV and JCV was quite high. About 50 percent of the urines contained either BKV or JCV and some contained both.

This is the data from the -- for BKV and JCV from the urine specimen. In the HIV negative urines, or urines from HIV negative people -- only 2.3 percent were positive with HIV negative, and it increased with SV40 positivity and the degree of immunosuppression, something that has been reported before.

The SV40 prevalence was very high in both HIV negative urines as well as HIV positive urines, and we did not see any much greater increase in prevalence from what was already a very high prevalence in the HIV negative individuals.

So we could not detect SV40 sequences in mesothelioma tissues although we did 50 tissues. From the results of previous speakers we should have picked up at least 20, 25 positive specimens. We did not find serological evidence of SV40. One would think that if these people were developing tumors because of an SV40 infection, one would expect some evidence of SV40 infection, and the antibodies to SV40 would be a very good sensitive measure. Plaque neutralization test is very specific; we did not find this.

Now, this is just a beginning study. At least in this group of patients, immuno-competent patients and immuno-compromised patients, we did not find any virus in the urine.

One of the puzzles in all of these studies is that all these patients in whom SV40 has been found, many of them have been born after the vaccines were cleared of SV40. So this suggests that SV40 must be circulating in human communities.

And to get some evidence for that, whether SV40 is circulating in human communities, we looked at these urines from groups with generally shared lots of polyomaviruses like BKV and JCV. And so we did not find evidence of that also. Thank you.

CHAIRMAN KIRSCHSTEIN: Thank you, Dr. Shah. This concludes the first session, a very interesting session. There will now be a coffee break until 11:15.

(Whereupon, the foregoing matter went off

the record at 11:03 a.m. and went back on

the record at 11:30 a.m.)

CHAIRMAN BRIEMAN: I'm Rob Brieman. I'm the Acting Director of the National Vaccine Program Office, and I'd like to welcome you back for the next session.

I would like to say that the NVPO is very happy to sponsor this meeting and I think it is providing the opportunity to look at very interesting data and to examine the implications of the information presented from a public health standpoint.

The first speaker for this session is Dr. Kristina Dorries from the University of Wurzburg in Germany.

And I'm sorry. Before we begin, let me just say that if anyone has additional data that they would like to present in a short presentation -- either positive or negative information -- in time for the next panel, or the panel discussion which will be this afternoon and led by Dr. Fried who's standing here --please see him either during this session or immediately following to arrange that presentation.

So again, our first speaker will be Dr. Dorries.

Return to Table of Contents

DR. DORRIES: Thank you very much for the invitation, to the organizers, and we will proceed now to the human polyomaviruses, BK and JC virus, and I will try to summarize some of the essential molecular, biological, and pathogenic features of both viruses.

A common infection of BK virus is early in childhood with zero conversion rates of more than 90 percent in the young. In contrast, JC virus is coming a little bit later up and there we have zero conversion rate of 70 to 90 percent in the adult population.

According to PCR analysis, a virus persists in nearly 100 percent of young adults and both viruses persist in kidney epithelium. They are transmitted, at least in part, by urinary excretion. Both viruses are discussed to be involved in tumor induction in men.

These have the viruses -- and both viruses in common with SV40, but additional DNA homologies and protein homologies are followed by similar morphology of virus particles and by the same replication strategies of the three primate polyomaviruses.

This is a genome of JCV and you'll see small T-antigen and large T-antigen coding in the early part of the genome, VP1, 2, and 3 in the late part, and there is an open reading frame for the adenoprotein.

These DNA sequences are very homologous between the two viruses, however the TCRs, the transcription control region is different.

This is the non-coding control region with the origin of DNA replication. That is conserved with T-antigen mining sites 1 and 2, in the whole family of the primate polyomaviruses and the transcription control region is shown from SV40. And we have here heterogeneity of single and double promoter elements that are present in different -- that might be present in different organs.

The conserved first in hearts or promoter element contains binding sites for cellular factors that are gliocell specific or associated to the basic activation -- cellular activation -- and are associated with signal transcription elements.

The cause of human polyomavirus is not quite clear. We don't know the entry site, but primary infection always is followed by live, long persistence of the viruses. We don't know whether the infection is latent or low copy infectioned, but we can find viral DNA, episomal DNA in the affected organs.

Under limited changes in the immune state, as induced by pregnancy or older age, we find a temporary activated infection with short episodes of virus production and messenger RNA viral proteins that can be detected. This situation is asymptomatic.

In contrast, in the clinical overt state of infection under severe immune deficiency as induced by lymphoproliferative diseases, immunosuppressive therapy, or AIDS, we find an unrestricted virus grows with efficient virus production and the lysis of viral target cells.

Polyomavirus associated human disease are only described in the central nervous system for JC virus. It's progressive multifocal leukoencephalopathy, and for BK virus there is strong association with the urogenital system. However, recently there is a systemic disease described that involved tubular lynenphorytis and disthicia pneumonitis and the subacute meningeal encephalitis for BK virus.

This is summarized here on this slide but I would like to concentrate on the cell type specificity. Cells of the connective tissue are clearly associated with BK virus infection at epithelial cells as fibrocytes of the three organs, and specifically in the meningeal encephalitis we find epidermal cells associated with BK virus infection and interestingly, astrocytes.

The astrocytes are in contrast to JC virus infection from the glio type of cells -- oligodendrocytes predominantly are infected.

This is only shortly in in situ hybridization of the epidermal layer on the ventricular system. This is -- the brown color is BK virus DNA specific hybridization, and this is a double immuno staining of GFAP astrocyte specific protein, and in the nuclei polyomavirus specific antigen is found.

This BK virus associated disease is a rare disease. It was described once in Germany, and we might now have a second case. However, this is completely different to JC virus associated PML as found now in steadily increasing numbers.

The characteristics of PML is disseminated focus cytolytic infection of oligodendrocytes in cortex and white matter of the brain, and this is followed by the loss of myelin sheath.

This is a cartoon of a typical PML lesion with the oligodendrocytes filled with virus particles and the rim of the lesion, and at the center you see reactive astrocytes and giant cells in mitosis. This is believed to be a semi-permissive JC virus infections and there are cases where virus particles are found in these cells.

Additionally, it might be associated with transforming infection. Limited infiltration of lymphoid cells is described recently in predominantly in AIDS associated cases.

Here you'll see in situ hybridization of typical PML lesion. You'll see the highly concentrated virus DNA by the black grains. It's radioactive hybridization. This is at rim of such a lesion with less concentrated DNA in freshly-infected cells.

In contrast, in the periphery we have only an attenuated infection. These are epithelial cells surrounding the renal tubules, and although this cell contains not only DNA but also a viral antigen and an electromicroscopy virus particles where detected, you see that this is persistent, infected in the activated state -- a more or less attenuated infection.

A molecular characterization of the virus DNA population in these different sites of infection in the brain revealed presence of heterogenous TCR, transcription and control elements, that are repeated at the GSP prototype from Wurzburg, and in the kidneys you find the single, so-called archetype elements.

In the time before AIDS, lymphoproliferative diseases were closely related. About 62 percent of all cases in '84 were the basic diseases for PML, but now in the AIDS era we discuss between five and ten percent of AIDS patients with neurological symptoms might eventually die by PML. And therefore, several laboratories are now interested in factors or situations that might lead to the induction of disease.

And I summarized several factors that are discussed in the moment that as, first, virus-dependent factors and host-dependent factors. In case of virus-dependent factors, conditional viral dissemination is discussed that probably primary -- accidental primary infection of the effector organs, the central nervous system under immunosuppression as leading to disease.

The second possibility is that the genetic heterogeneity of the promoter/enhancer elements is involved in pathogenic -- in the induction of the disease, and this might be associated with the neurotropic selection of highly active, cell specific transcription elements.

From the host-dependent factors, genetic predisposition can be assumed, but what is obvious is the control of virus host interaction by the immune system really plays an essential role in the activation processes.

The first possibility was that primary infection might only happen under immunosuppression. Under this condition, no involvement of the central nervous system in persistent infection should be expected and therefore we analyzed cellular DNA from non-PML patients. And altogether I think we analyzed 70 patients by PCR analysis. And here, from several regions of the genome these were hybridized with JC virus-specific probes and the products were sequenced.

What came out is summarized here. This is a T-antigen amplification product that hybridized with JC virus DNA. And interestingly, from the same -- hybridization from the same material revealed that not only JC virus is present, but also BK virus is present. And these probes are species-specific for both viruses.

From this data we can say now that probably JC virus is persisting in the central nervous system and reaches the central nervous system long time before the induction of disease.

And that is a requirement for the selection of neurotropic lysal-specific TCR variants that might be selected by cytolytic -- might be selected by activation processes in the persistent state.

And to analyze this we characterize the variants that were present in the different ethnification products by cloning PCR amplification products and sequencing them afterwards.

And altogether we found seven different TCR subtypes. From the TCR type 1, three subtypes in TCR type 2, two subtypes type 1. Repeating the tatabox elements type 2 leaves tatabox element out and repeats only conserved sequences that are following.

What was interesting is that we found single elements -- these are so-called archetypes -- we found double elements here and here. And here, those are other duplicated elements. And additionally, we found triplicate from TCR type 1 DNA.

This was a finding that was astonishing but what really was new is that W1 and W6 were the dominant subtypes that we found in persistently infected brain tissue. W1 is similar to the mat 1 isolet from medicine, the American isolet, and the TCR type 2 is similar to the GSB isolets. The predominant type altogether was med types -- American subtypes.

Then we thought that possibly the situation might -- that these rearrangements were -- came up under persistent infection and we studied then a kidney tissue. And interestingly we only could find W1, double enhancer promoter elements in kidney tissue of eight different patients. Only the single W4, single archetypes are found in only low percentages.

From these data we only can assume that the different rearranged elements are present in the human population and are not rearranged anew in each patient.

This brings us back to the third possibility that changes in the virus host interactions due to severe breakdown of the immunological surveillance are the real requirement for the induction of clinical, overt CNS disease.

However, we have not mentioned under these conditions recent results that were found in lymphocytes. First it was described that V cells in PML tissue were actively infected by JC virus and in addition, later on, peripheral blood leukocytes were analyzed for the presence of JC virus DNA.

And as you can see here again, it has T-antigen PCR hybridization -- with T-antigen specific probe revealed that almost all healthy individuals were carriers of JC virus DNA. To a certain extent we also detected BK virus DNA. However, the concentration is lower than the JC virus concentration, although the primer is similarly sensitive for both viruses.

This opens up now, a complete new field of possibilities, how the virus could interact with host-related mechanisms and will bring us further I think, in the next time.

To summarize, the conditions that are associated with the pathogenesis of PML under immunocompetence, we find an asymptomatic JC virus infection of the central nervous system, life-long persisting viable JC genomes with highly active TCR types are present in the central nervous systems, a growing virus load in advanced stages of life was detected, and also in individuals undergoing changes of the immune state as induced by systemic tumors carry an enhanced virus load.

A virus infection is probably controlled by tight immunological surveillance, and the severe changes of immunological virus/host interaction and probably comes to a 2-step situation. First, we find an activation of viral expression in peripheral organs, we find a growing virus load by the activation event and peripheral tissue and in the cells, and this really probably is followed by an enhanced seeding of virus to the central nervous system.

And a second all simultaneous step we find an unlimited activation of viral expression in the central nervous system, and then CNS tissue is destructed by cytolytic multiplication, and additional enhancement of the virus load by infected lymphocytes may happen also.

And then possible interaction with the viral transactivators should be discussed. In vitro experiments with cytomegalovirus and HIV early antigens have shown that the early antigens are interactive with JC virus, and it could be that under these conditions we find an enhancement of virus replication.

Thank you.

CHAIRMAN BRIEMAN: Thank you very much, Dr. Dorries for that elegant presentation that was, I might also note, precisely timed -- perfectly. I need to announce again that Dr. Chanock needs to check at the desk because he has an urgent message.

The next speaker will address the question, "is BK virus a co-factor in human cancer?", and it's Dr. Michael Imperiale of the University of Michigan.

Return to Table of Contents

DR. IMPERIALE: I'd like to thank the organizers for inviting me to talk. What I'd like to do during my talk is sort of move into the realm of molecular biology a little bit.

Okay, so you've already heard a little bit about BK virus but I just want to repeat a few salient points here. The first is that there's a subclinical, persistent infection in about 80 percent of the population. BK virus encodes a T-antigen which is about 75 homologous to SV40 T-antigen which you've heard a lot about this morning. However, that homology is much higher in the transforming domains that Ellen Fanning referred to in her talk.

It's been shown that over-expression of T-antigen or a co-expression of T-antigen with an activated ras gene can transform cells -- this is both in vitro and also in transgenic animals.

And finally, as you heard before, BK virus has been associated with various types of tumors and also with hemorrhagic cystitis in amino-compromised patients.

So about two years ago what we wanted to set out to do in my lab was to ask the question regarding these tumors in an indirect sense and that is, does BK virus have the tools that it would need to be involved in human cancer?

This is just a table from a paper that was published in Virology in 1995 from the group in Ferrara, where they showed that BK viruses associated with various urinary tract tumors -- and indeed, in some of these tumors they found that they could detect viral sequences by southern blotting rather than by PCR, which implicated that there might be at least a higher amount of viral DNA in those tumors.

Okay, so the first thing that we wanted to look at was the interaction of viral T-antigen with the Rb family of proteins. This is just to remind you that these proteins inhibit cell cycle progression. They do this by binding to family members of the E2F transcription factors and block E2F function.

The next slide just talks a little bit about E2F. It's a cellular transcription factor. It's actually a family of heterodimeric proteins, and the function of these transcription factors is to regulate genes that involved in S-phase progression.

And the next slide is just a model of what's going on here. So this is a normal cell. Up here at the top, Rb and p107, p130 are bound to E2F. When the cell receives a signal to divide the proteins become phosphorylated; this results in release of E2F which is now able to activate transcription of its target genes which then drive the cell into S-phase.

In the model that's been worked out for SV40 T-antigen is shown down here at the bottom, which is that T-antigen binds to the retinoblastoma protein and its other related family members. This has the same end effect as a mitogenic signal up here, which is that the E2F is released and the cell can now enter the cell cycle.

So we wanted to ask whether BK virus T-antigen was functioning in a similar manner. The first thing that we did was to just look and see whether T-antigen could complex with the Rb family members in cells -- and these are a series of immuno-precipitation assays where we immuno-precipitate with antibodies against T-antigen, run the proteins on a gel and blot and probe for the three different Rb family members. And you can see that we find complexes of T-antigen with Rb, p107, p130 in the cells that are expressing T-antigen.

Now, what I need to point out here is that, in order to see this we have to use 100 times as much protein from these BK virus T-antigen expressing cells, as from cos-1 cells which express SV40 T-antigen. And the reason for that is that there's a hundred-fold less T-antigen in those cells.

So this is a T-antigen blot. We have to load 100 times as much protein to see the T-antigen as we do here. So any effects that BK virus T-antigen is having here as compared to SV40 T-antigen, these effects are occurring in spite of the fact that there's a hundred-fold less protein there. And that's something I'd like you to keep in mind as I go through this.

The next slide on the right is just a whole cell lysate blotted with antibodies against these proteins. And if you look at the normal cells compared to the cells expressing BK virus T-antigen, you can see that there's less Rb, p107, p130. What's left is mostly in the hypo- or under-phosphorylated form. And I think Jim DeCaprio is going to talk more about this tomorrow, so I'm not going to address that issue.

But what I want to point out here is that what this is saying is that most of what's left in the cell here is in this form, which should be the growth inhibitory form, and also most of what's there is not bound to T-antigen because there's so little T-antigen present in the cell. So the question is, are we getting an induction of E2F activity?

So to look at this, what we did was to take a gene that expresses a CAT reporter, under the control of E2F DNA binding site, and transfect that into cells, and the results have shown -- the next slide on the right -- so these are just CAT activities of these different cell lines.

You can see when we transfect this reporter gena to normal kidney cells we don't get any activity. If we put it into cells expressing BK virus T-antigen we get an induction of activity. SV40 T-antigen we also get an induction.

Notice that the induction only differs by about 4- or 5-fold even though again, there's 100-fold difference in the amount of T-antigen in these cells. And these are just transient transfections where we've done the same sort of assay showing the induction by BK virus T-antigen compared to another protein that interacts in the system, which is the adenovirus C1A printing.

So in spite of the fact that there's very little T-antigen binding to these proteins that we can detect and that most of the protein that's there in the cell is in this form here which should be bound to E2F, it looks like we're getting induction of E2F activity.

So the question we wanted to ask is, what does the E2F look like in these cells? What we've done then is to look at E2F binding activity by doing DNA band shift assays. The left side here are just straight DNA band shift assays, making extracts either from normal monkey kidney cells, cells that are expressing BK virus T-antigen, or cells expressing SV40 T-antigen.

What I want to show you on this side here is that, this is the free E2F, so this is the theoretically, transcriptionally active form of E2F. And although this is a little bit overexposed so that you can see some of these other bands, there is an induction of the amount of free E2F in the cells expressing the viral T-antigens as compared to the normal cells.

And I should point out, this is not due to differences in the number of cells that are in different portions of the cell cycle.

Now, the right side of the -- I'm sorry, the left slide over here shows that even though T-antigen is present in these cells, there's a significant amount of E2F that's still bound to Rb. And that's shown here. You can see that there's a little bit of E2F Rb complex that's left here, and this side of the slide just shows how we identify the complex, as this complex goes away if we add antibodies against Rb; this complex goes away if we add antibodies against p107.

This is to show that there are Rb E2F complexes still present in the cell. What we've done here is to use antibodies against the Rb protein to immunoprecipitate and then release any E2F that was present, bound to that Rb, and put that into the band shift assay.

And you can see that there's still E2F activity that's associated with Rb, both in the BK virus T-antigen expressing cells and in the SV40. And again, this is about equivalent, despite the huge difference in the levels of T-antigen in these two different cell lines.

Finally, the last slide on the left. So the question is, if we're getting induction of E2F activity, are we getting induction of cell growth? So these are some growth assays of these cells. If we take these cells and grow them in ten percent serum you can see that all of these different cell lines grow well.

However, if we put these cells into medium-containing load serums -- so this would be an assay for serum independent growth which is a hallmark of transformed cells -- we can see here that two different parental monkey kidney lines do not grow well in low serum, whereas the cells expressing BK virus T-antigen or SV40 T-antigen both grow in low serum.

So what this tell us then, is that BK virus has the ability to induce cells to grow. So this would be one of the things that one might expect for something that might be involved in tumorigenesis.

The next thing that we wanted to look at then, is the interaction of T-antigen with p53, and this is just a little cartoon here to show you how p53 works. Normally, if a cell is getting ready to divide and that cell receives DNA damage, there's an induction of p53 levels, and that results in arrest of the cell before it enters S-phase. That allows the cell to either repair the damage or if the damage cannot be repaired then that cell will be induced to undergo programmed cell death or apoptosis.

Now, in the presence of SV40 T-antigen what happens is, T-antigen binds to p53, the cell receives DNA damage, there's no induction of p53, no G1 arrest, and the cell can go on to divide with damage, or if there's enough damage there will be mitotic failure and cell death.

So we wanted to ask what was the interaction between BK virus T-antigen and p53. The next slide on the right is just an immunoprecipitation similar to what I showed you before. For the Rb proteins, if we immunoprecipitate with antibodies against T-antigen, we bring down all of the p53, or over 95 percent of the p53 protein that's present in the cell.

So this says that even though there's very low levels of T-antigen present in the cell, it is able to bind most, if not all, of the p53 in those cells. So the question is then, is it having any effect on p53 function?

So what we did then, was to take normal cells or cells that are expressing T-antigen and irradiate them and look at the response to DNA damage. The next slide on the right is just an amino blot -- oops, on the left, sorry -- it's just amino blot where we've looked at either the induction of p53 or one of its important target genes, p21, upon irradiation.

So the left side of this slide shows the normal cells; increasing amounts of irradiation we get an induction of p53 levels; and we get a concomitant induction of p21 levels.

In the cells that are expressing BK virus T-antigen, first you can see that there are higher levels of p53 to start with. This is because T-antigen binding stabilizes p53. However, there's no further induction of p53 upon irradiation and there's also no induction of p21 upon irradiation.

The next two slides just show some cell cycle analysis of this cells, the normal cells on the left, the transformed cells on the right, and you can see that when these cells are irradiated, the normal cells, one gets mainly arrest here in the G1 phase of the cell cycle. If you look at the cells expressing SV40 T-antigen however, one does not get arrest in G1; instead, one mostly gets arrest in G2 M-phase of the cell cycle.

So what this says then is that BK virus T-antigen is interfering with p53 function and removing its ability to block cells in the G1 phase of the cell cycle upon DNA damage.

Okay, finally then, we did an experiment based on some old experiments that were published for SV40 many years ago, where people showed that SV40 infection can induce chromosomal damage. And so we wanted to ask whether BK virus can also induce the same sort of chromosomal damage.

So what we did then, was to take some primary human fibroblasts and infect them with either BK virus or also JC virus, and look at their chromosomes. And the next slide on the left shows the result.

This is a karyotype from normal cells here. These are the karyotypes from either BK virus-infected cells or JC-infected cells. And in both cases you can see that we're getting chromosome fragmentation, there are translocations. And moreover, in the presence of BK virus, we're also seeing this thickening of the chromosomes which is indicative of endoreduplication.

And this actually makes sense when one looks at the cell cycle analysis I just showed you; that the cells may be getting hung up in M-phase. So BK virus is able to induce DNA damage in these cells.

And what I want to point out here is that we did this at a rather high multiplicity of infection so that we could get enough karyotypes -- enough infected cells to actually see a karyotype. And we assume that BK virus may be inducing a lot of undetectable chromosomal damage, even at lower multiplicities of infection that could potentially be occurring during a persistent infection.

So the next slide on the left then, I'd like to just sum up by asking a question -- and let me state from the outset, I don't know what the answer to this question is -- but the question is, is BKV a co-factor in human cancer?

As I showed you, it can induce DNA damage, it inhibits the p53 response to this damage so if cells receive damage then they're unable to arrest in G1 before that damage can be repaired, and moreover, it's inducing cellular DNA synthesis so it's inducting E2F activity. And the question is, can this lead to increased mutation rates which then potentially could lead to carcinogenesis.

And so I think I'll stop there, and just on the next slide acknowledge the people who have done this work. In my lab, Kimya Harris, Joan Christianson, and Eugenia Chang. And our collaborator in the Department of Human Genetics, who's done the karyotype analysis, Tom Glover. Thank you.

CHAIRMAN BRIEMAN: Thank you, Dr. Imperiale. That was very impressive. The next speaker from Allegheny University for the Health Sciences, is Dr. Kamel Khalili who will speak about JC and BK sequences in human tumors.

Return to Table of Contents

DR. KHALILI: I guess what I'm going to do today is focus my talk on JC virus. Can I have first slide, please?

This slide just demonstrates that the JC virus is not another SV40 or humana SV40, and it has its own personality and character. Kristina Dorries very elegantly laid out some of those characteristics, and I'm going to bypass some of these slides which I have initially, to describe a JC viral control region or lytic cycle and then demonstrate that.

In the next slide -- this is a virus which infects greater than 92 percent of human populations and as you heard from Kristina's talk, perhaps 100 percent PCR. And it's a virus which has no animal reservoir, and perhaps more importantly, it's a virus that we can talk about. When you talk about an AIDS epidemic you cannot ignore JC virus.

It's a virus which used to be rare and the disease of the virus was rare but not any more. And perhaps I should mention that greater than ten percent of AIDS patients which no logical disorders, which by itself is greater than 70 percent, come down with PML.

So suggesting that the virus has a chance to become reactive replicates in cells. The cells of the virus is replicating oligodendrocytes.

This is oligodendrocyte which is responsible for formation of myelin sheaths around the axon. In fact, oligodendrocytes sends a processes that -- this is an electromicrograph of normal brain -- shows that the oligodendrocyte sends a process, circulating processes, and these processes form the layers, very compacted layers, around the axon, an important event for the insulting axon.

Now, cytolytic destruction of these oligodendrocytes upon reactivation of endogenous JC virus, all the JC virus which gets to the brain by B cell or any other vehicle, results in the demyelinating disease of PML.

As Kristina mentioned, the JC control region sits between the viral early and late genome, and it has a unique characteristic. It has a 298 base pair repeat sequence and old genome viral DNA replication. It's a bidirection promoter of single to SV40. In one hand it shoots for the early gene expression; on the other hand, the other sides controls late gene expression.

Now, if you really go through the sequence you do see that there is not much release sequence homology or similarity between the JC control region and SV40 control region. That's one characteristic of the virus counts.

We spent about eight years -- over the last eight years -- trying to understand what confers the tissue specificity to the viral gene expression. We knew that early gene expression is make or break for the virus. So if you reduce immunosystem, virus gets opportunity to replicate, but it does not replicate in every cell. It replicates in all oligodendrocytes, the cell which I showed previously.

So there is two barriers in the viral gene expression: first is immune system and the second is its cell-type specific barrier. The latter was easier to address. We have the similar cell lines and we went on through the standard biochemistry to understand that one of the sequences within the viral control region, which, upon interaction with cellular factor, turn on early gene transcription.

Again, Kristina demonstrates a slide summarizes all the transcription factors which could bind to a JC control region. And through those studies, we and others learn that perhaps interaction of the protein some were inducible, some were tissue specific factors.

So the control region, per se, is not -- it's essential but still does not put the virus through the lytic cycle. So it seems that the interplay between protein upon the interaction plays more important role in this event.

So this slide is about a 3-year old slide. It just shows some of the proteins that we and others purified by standard cold room chemistry and cloned the genes, and demonstrated that association of the protein, for example in this case, YB-1 with the JC control region, is important to induce viral gene expression couple-folds, but sufficient -- it wasn't sufficient to put the virus through the lytic cycle.

Then once other proteins were purified, we realized that the communication of this protein with each other perhaps plays a more important role.

Now, another feature of the virus which I'm going to little bit talk about that, is the TAR region. TAR is a sequence which first identified in HIV genome, and this region, located downstream from the HIV transcription initiation site, and it's responsive to tat protein, a protein from HIV which is important for the viral replication.

Now, a higher incidence of PML among AIDS patients suggests that perhaps in addition to immune system, there is some direct communication between HIV-1 and JC Virus in the brain parenchyma.

So we asked the question of whether or not HIV regulator protein included in tat could activate JC promoter in the choleal cells. The answer to that question was yes, and we demonstrated that this activation of the viral promoter, JC promoter for HIV tat was mediated through the sequence, TAR log sequence which is located in the leader of the late gene transcription -- very reminiscent to the structure described in HIV promoter.

In addition, it was demonstrated that tat could bind to another cellular protein; a protein which initially identified based on this ability to transcribe, stimulate JC promoter, and as a result of this interaction, augment expiration of the viral JC virus.

So you see that we learn much more by studying JC virus in terms of the mechanism of viral-viral interaction and the cell type of specific transcription.

Now, let me move on, on the next slide. So the question was JC virus, under proper conditions, proper environment, produce a T-antigen, and once the T-antigen is produced such as other papomavirus, induce viral DNA replication, late gene transcription, and certainly the destruction, the virus causes PML.

Now, what happens if you block replication event so you still produce T-antigen under certain circumstances, and what happened to that T-antigen? Does T-antigen, the absence of replication that's supposed to be crossed, could induce the tumor?

So the first clue for that came from the animal studies, hamster model, where the JC virus cannot efficiently reproduce. In the replication does not happen because of the primary unit of DNA prolimerase is species-specific, cannot turn on JC viral DNA replication in hamster.

Therefore, if you take a JC virus, inject in the brain of a newborn hamster, what happens after two months, you get tumors. Eighty percent of the injected animals come down with a tumor. This was demonstrated by a number of people, and also later on, Sid Hoff and others demonstrated that intercerebral inoculation of the JC virus in our monkey, the squirrel monkey, induces glioblastoma multiformity. A very devastating brain tumor which we still don't know about the pathogenesis of the tumor and also we don't basically have a cure for this tumor.

So this slide illustrates the gross section of the brain of the newborn hamster. This is a tumor for an open injection of JC virus and this is a normal area. Now, if you take the tumor results here and stain it for T-antigen, you basically get 100 percent staining of the T-antigen in almost every cell.

There's a series of slides which -- studies that we did, and I'm not going to go through that -- and we will learn that these cells do have their own characteristic in terms of the cell growth and other factors; factors which are important for control of the cell proliferation.

And even the extent of the T-antigen in a variety of the cells were different. We cloned these cells; we are in process of analyzing it -- how low, high, and medium dose of the T-antigen in the various cells could utilize the different pathways which leads to the formation of the tumor.

And at the same time, we are utilizing the hamster model as a measure for studying formation of the tumor. This is a very unique model. You can inject JC virus in the hamster brain with notion that 80 percent of the injected animals come down with the tumor. So what you could do is, you could monitor development of the tumor; things that you cannot do in any other animal species, and clearly we cannot do in human.

So how you can do it in vivo, we have taken this hamster model and in during MRI on the living animal, we are monitoring formation of the tumor from the time that injected to the time that becomes significantly big and close to kill the animal.

So in parallel we are devising histopathological studies to understand the interplay, because some of those regulatory factors that Mike talked about, that E2F for example, cyclins during the formation of the tumor, and at the same time we are trying to see that whether or not the ring value stage of the tumor, the partner of the T-antigen changes.

We know that the T-antigen binds to p53 and Rb. These experiments are done in cell line, but we can utilize this study, animal model, to do experiments in whole animal model.

Now, another model is the transgenic mice, that expression of the T-antigen induced tumor in the transgenic mice. And analysis of the tumor showed that a tumor is formed in the abdominal area. It was NSC-positive, normal crest origin tumor.

And when you do a histopathological analysis of this tumor formed upon expression of JC in trans-genic mice, you see that they're highly differentiated cells that -- tumor cells, basically infiltrates between the wall of the colon and the muscle cells here. These are all the tumor cells and these are all T-antigen-positive cells.

Now, what we are trying to do, we are trying to use these transgenic mice to study the importance of some of the cell cycle regulators in whole animal model information of the tumor. First we demonstrated that T-antigen was expressed in RT-PCR, in almost in every tissue of this experimental animal.

But when you look at the Weston blot you see that the detectable level of the T-antigen was obtained only in the tumor but not in any other tissue derived from this animal.

So in the next slide here, we looked at the interaction of T-antigen with p53 tumor suppressor, our favorite experiment as Mark demonstrated, by coming and precipitating with one antibody, probe the Western blot with another antibody. The take-home message is yes, p53 associates with the tumor and the T-antigen associates with p53 in the old 2-reciprocal experiment.

And the working model was is that if the JC virus T-antigen associated with p53 takes away p53 from the loop as a result on same target to p53, p21 WAF, a protein which could suppress function of cyclin's associated kinases and the complex which eventually could phosphorylate Rb and liberate E12, could be other function in this case.

And what we know is, each WAF could regulate number of other cell cycle genes and put the cells into the rapid proliferating stage, and at the same time, T-antigen could associate with Rb, another way that liberates E12.

Now, we really went on through every step of this and examined every step in this tumor. For example we asked, what happened to WAF gene in this tumor tissue? We realized that there is no WAF gene, basically WAF protein in the tumor cells. And then we asked, if you don't have WAF, what happens to all the cyclin and CDKs?

What I'm going to do in the next slide just showed two examples of that. This is cyclin E and this shows the level of cyclin E, and this shows that it's a kinase activity.

And the partner to cyclin E is CDK2 -- the level of cyclin is CDK2 and its kinase activity. So you see that it is highly active in these tumor cells. What happens if the cyclin E, CDK2 is very active? After the previous pathway -- in the next slide -- we examined the Rb phosphorylation and you see Rb, unphosphorylated Rb which is supposed to be the single band in the tumor, becomes two bands. The top band is phosphorylated.

If the Rb phosphorylated dose E12 liberates from the Rb -- next slide -- it does, and it all regulates itself and you see massive amounts of E12. And if E12 is functional or not -- yes it is, because it activates PCNA -- it's on the same target and expressed in the tumor.

So you see that we can utilize this model to really go through all those cascade of events that has been described in the cell cycle, and identify the regulatory factors which really is important for the pathogenesis of the JC-induced tumor.

Now, what I'm going to do is, we switched to the human samples, and we all learned that JC virus could be associated with a number of tumors in humans, we learned this morning, and these are studies done previously by many other people.

Recently, maybe two years ago, a 61-year-old, immunocompetent, HIV-negative -- this is very important clinically -- individual came to neurology service and he had a multiple grand mal seizure.

This individual, after MRI, showed that the hypointension area in the left frontal lobe of the brain -- after about a year, gallilinium staining shows a new enhancement which sits right in the center of the previous enhancement, suggesting that the tumor is formed here.

Now, if you see that the tumor mass was so big that it pushed the lower, the left anterior -- the formal -- this structure, and you see that this structure, which is ventricular, should be like this. But the tumor is pushing this toward the right lobe.

Now anyway, so we got the samples from these patient, and the RNA protein DNA was extracted and we did RT-PCR for the presence of T-antigen DNA which was there, and the RNA was utilized for the primary extension for expression of JC virus RNA, and the results is here. Basically says yes, early RNA was expressed.

And then when we look at the protein by Western blot, here is a T-antigen of the JC in the tumor specimen. And if you do immunohistochemistry you see that the T-antigen is present in 50 percent of the tumor cells.

This slide shows the staining of cells with Chi-67. It's a self-proliferating marker showing that the cells are highly proliferating and this is Luxor blue staining showing that the level of the myelination -- which shows they're heavily myelinated.

Now, if that T-antigen doesn't belong to a JC virus, we amplify the control region of this virus, and after sequencing they turn out to be member of a JC virus -- it's a mat-4 strain of the virus which has a characteristic of 98 base pair and 79 base pair sequence.

Now, if the virus is there, why does it cause PML? Why cannot replicate and destroy the oligodendrocyte? We thought initially that it might be a mutation, replication which does not respond to T-antigen. As a result, you create a situation like hamster; that you produce T-antigen in the brain but DNA replication does not take place. T-antigen triggers a cascade of events, and lead to the formation of the tumor in this patients.

Clearly the T-antigen was intact. That was a wrong assumption. Then we learned from SV40 there's another protein in the leader of the late gene -- adenoprotein.

We learned that if you introduce mutations in the AUG of the adenoprotein -- this is worked on by Tom Shank many years ago and also some of the mutants that Jim Pepper has created, the mutation which has a T-antigen deletion of the carboxy terminal of a SV40 T-antigen -- what happens is you do not get a viral protein, late protein, be transported to the nuclei and form the virions.

And when we look at the adenoprotein of the JC virus we found that there's a mutation -- three nucleotides right at the initiation site which put the adenoprotein on a frame. At present we don't know that's the cause, that's the important event that put the virus through the other pathway, the tumor pathway or not. So that's a question that we're asking.

So I'm going to stop here, but before ending I would like to -- well, Dr. Shah mentioned something very interesting this morning. He mentioned that this meeting reminds him about 10, 15, 20 years ago when the people, the polyomaviruses were gathering and talking about this system. And indeed it does. It resembles like at those meetings that we used to go to, Cold Spring Harbor and talking about polyomavirus, mostly on SV40.

But there's one person that's missing in this today, and that's a person who did a lot of work, a lot of contribution on SV40; things that we learned on enhancer regulation of SV40, and the T-antigen. This person is George Khoury who left us about ten years ago. In 1987 he died of lymphoma, and I think it's very appropriate that at this meeting we just remember him and his contribution.

Thank you.

CHAIRMAN BRIEMAN: Thank you very much, Dr. Khalili. Our next speaker is Dr. Richard Frisque of Penn State University, and the title of his presentation if, "Rearranged and chimeric primate polyomavirus genomes".

Return to Table of Contents

DR. FRISQUE: Well, I'd like to offer my thanks also, to the organizers for the opportunity to speak. The focus of my laboratory is on the pathogenic and oncogenic potential of JC virus. Now, much of our work has relied upon the biological and molecular comparisons with related viruses BK and SV40.

This slide shows the comparison of the JC and SV40 genomes, and these are very similar. They share 69 percent sequence identity. JC and BK share even higher: 75 percent sequence identity. The organizations as you can see, are essentially identical.

There are three regions as already have been pointed out. I'll just quickly do that again. The early region which includes this T-antigen that we've all been talking about. In the case of JC there are five different proteins produced, now we know, in the early region, based on alternative splicing mechanisms.

The late region includes the capsid proteins and then this control region -- or the regulatory region as I call it -- is also present; the third region. It includes the replication origin as well as the signals for transcription.

Now although many biological similarities occur between these three viruses, JC is distinct. It has a prolonged lytic cycle; it is very inefficient at transforming cells in culture; and its expression is restricted to only a few cell types.

Over ten years ago we asked the question, what makes JC biologically distinct? Which regions of the genome, which sequences are important? And we began making in our first approach, chimeric virus between JC, BK, and SV40.

Now, our prediction was that if we replaced some of the JC sequences with those of SV40 or BK, we might make a more active virus and identify some of those sequences that contribute to JC's unique biology.

Now, in terms of relevance this workshop, I think that if co-infections to occur in the human host with one or more of these viruses, that in fact, if a combination can occur between these three viruses then perhaps we might generate more viable, perhaps more active, recombinance.

Alternatively, or maybe in addition, we also may see complementation occurring, where one virus could complement the growth of the second co-infecting virus. So that's the relevance then, to this workshop.

This slide shows you the first of the chimeric genomes that we've produced. These are what I call regulatory region chimeras. What we do is, we take the coding region of one virus -- in this case, JC -- replace its regulatory sequences with those of BK or SV40. And similar things were done with the other two viruses.

Now, if you can see it with this color -- I'm not sure how it looks back there -- but what we first did with these chimeras was to look at their transforming behavior in rodent cells. And what we found was that when we had the same coding sequences linked to the various regulatory regions, in all cases SV40 was always more potent, JC was always by far, the weakest at transforming cells.

Similarly, if we took viruses that had the same regulatory region and different coding regions, that again SV40 was always the most potent, JC was always the weakest. And in fact, the coding sequences seemed to have more of an influence upon this restricted behavior of JC than even the regulatory sequences.

Now, we've also done some tumorigenicity studies. This has been done with transgenic mice, and as others have shown as well, the regulatory sequences of JC, SV40 types of chimeras, the regulatory sequences are influencing the location of where the tumor occurs, whereas the coding region -- again the T-antigen, primarily -- is involved in the tumor induction itself.

Well again, these same chimeras are shown in this slide, but in this case now, we're looking at lytic behavior in terms of DNA replication and the ability to produce viruses themselves, and also looking at host range effects.

And what we found that, somewhat to our surprise, that those constructs that had JC regulatory sequences in either SV40 or BK coding sequences, that in fact, these were viable viruses -- viruses produced. And in fact, they were more active than wild type JC itself.

In addition, when we took one of these chimeras, JC SV40, we found that its host range was expanded over that of wild type JC. In fact, these not only grew in human cells but they also grew in monkey cells as well.

Finally, the last surprise from looking at lytic behavior of these chimeras is shown over here on the right, when we had constructs that had JC coding sequences linked to BK or SV40 regulatory sequences, which are more potent, that in fact these were dead viruses.

So if I could have the next slide, I would like to then look at that last point in more detail. Here what we've done is make another kind of chimera. In this case we've made chimeric replication origins. We've made chimeric regulatory regions to see if we could find out why the JC coding sequences in the particular T-antigen was unable to interact productively with the origin of SV40 and BK.

Now what we've done is, we've put the regulatory sequences of the various viruses onto a plasmid that does not contain any other sequence information for viruses, for the viral proteins. So just the regulatory sequences are on this plasmid.

And as we've shown already with the JCT protein, when these constructs were put into cell lines expressing JCT protein, both BK and SV40 were unable to -- their origins not replicate in the presence of JC T-antigen.

What we found from the chimeras was, when we had a JC on the early side of the replication and origin, core origin, that SV40 sequences on the late side, that in fact again, JC's T-antigen was not able to productively interact with those sequences. In other words, sequences on the late side of the core origin were responsible for JC's T-protein's ability to distinguish between the two origins.

And in fact, we've gone through the site- specific mutational analysis now, where there's three nucleotide differences within that small core region within the AT-rich region, that allows JC to distinguish between the two origins.

On the other hand, all the replication of origins, chimeric or wild type, were able to interact with the SV40 T-antigen as shown on the right.

On this slide I'm showing you a new set of chimeras. In this case we're producing chimeras that had T-antigens that were exchanged for sequences in three locations: either at the amino terminus, in the central region of the T-protein, or the carboxy terminus.

And in this slide I'm showing you a regulatory region -- and in this case, SV40's regulatory region, although we've also made chimeras with the JC's regulatory region as well. I'll just show you the data with the SV40 origin and transcriptional control signals.

And what we've found is, when we look at transformation in rodent cells, that as expected, JC's T-antigen if it was wild type, transformed very poorly. By replacing the amino terminus with SV40 sequences, we did not see much of an increase in transformed behavior. However, as we started to replace carboxy and central regions of the T-antigen with SV40 sequences in place of the JC, transforming behavior went up until we used the wild type SV40 T-antigen which is very potent.

Now in addition, in studies done in collaboration with Dr. Frank O'Neill, we did some immortalization studies with these chimeras, in human cells. And what we found was that most of these chimeras could not immortalize human cells. SV40 does immortalize human cells, but ten percent of transformed human cells become immortal.

What we were surprised to find was that constructs that had JC or BK at the carboxy terminus, these were also able to immortalize human cells and in fact, it's a much higher level about 50 percent of the lines that were looked at.

Well, in following with the way I've been going here, we've also taken these same kind of T-antigens and looked at their ability to stimulate DNA replication. Again, those are the same constructs as I showed you on the last slide using the SV40 replication, origin, and transcription signals.

And what we found was that the constructs that failed to interact with the SV40 origin for replication -- for DNA replication -- were those constructs that had JC within the central region of the T-antigen. So those, in human cells, were unable to replicate, thus identifying the part of T-antigen which was able to distinguish between the two replication origins: JC or SV40.

We also found that two of the constructs, those that had SV40 sequences at the carboxy end, again were able to replicate in monkey cells as well as human cells. So expanded host range involved as expected, the carboxy terminus of T-antigen.

Well, I just wanted to summarize the chimeric data at this point and again try to show you some relevance to this workshop. If in fact, co-infections do occur, we have the question of whether or not recombination can occur between these viruses. And if so, if a recombinant does occur, what would we predict, based on some of this data and some others, is the phenotypes would be relative to wild type SV40.

And I would like to suggest as possibilities, that the recombinant would probably show reduced transforming behavior, but perhaps higher or lower mortalizing potential. We'd expect probably, DNA replication activity would be reduced as well as virus yields, we'd expect that the host range would perhaps be more restrictive than SV40, and we would expect that the recombinant itself might actually be more active than wild type JC.

Alternatively, we can look at co-infections as a possibility of complementation occurring. Again, based on the data that I've shown you, we would predict that JC probably would not enhance SV40 growth if they co-infected the same cell. On the other hand, we might predict that SV40 would enhance JC growth.

I'd like to switch gears at this point and tell you about some experiments that we've been doing looking for JC's presence in human tissues, again as you've all heard, by PCR analysis and then followed by sequence analysis.

We have not found any evidence yet for recombination occurring between these three viruses. What we do see, considerable rearrangement occurring within the promoter/enhancer elements for transcription. And that's what I want to tell you about next.

What I'm showing you here is JC, and we're looking at the regulatory region, in particular, the transcriptional control region of the promoter/enhancer signals. In humans there are two types of JCs that can be found: this archetype that you've heard about, and what I call the rearranged form of JC.

Now archetype is shown at the top in the schematic, and what's unique about the archetype genome is that it has a single copy of the promoter/

enhancer region; there are no large tandem duplications. On the other hand, I show you four examples of rearranged forms of JC's transcription control region. These all came from different PML patient's brains.

And what you can see is that, unlike the archetype, there are differences in rearrange between each other as well as with archetype. And they involve deletion events usually involving the 66 base per block of information in archetype shown here, and sometimes its 23 base pair block of information is missing as well.

Following these deletion events, then duplication and sometimes even triplications occur, such that these sequences here are duplicated relative to archetype. What we believe is happening is that archetype, which is found in the kidney and in urine of healthy people, circulates in the population. We become infected by that.

Within our body we believe that these rearrangements may occur involving deletion events and duplication events. And these rearranged forms have been primarily found in brain, sometimes in PML kidney, and in lymphocytes.

In our first PCR experiments that we conducted about seven years ago, we first started by looking at PML tissues. Obviously, we were expecting to find JC in those tissues. And what I show you here are the tissues of brain and kidney from five different PML patients.

Over here, the brain specimens. Lane 7s in each case are the positive controls; lane 1 are the negative controls. These are the five brain specimens; these are the five kidney specimens; and as expected high levels of JC were found in these tissues.

When we went ahead and sequenced these isolets -- these PCR products, what we found was that these in fact, were rearranged forms of JC. As well in the kidney, we found primarily rearranged forms as well, although we have found archetype in some PML kidney.

More surprising at the time when we published this was that in fact, JC was also present in normal brain tissue. In other words, we've had 18 different specimens shown on this slide in which people had died of things unrelated to neurological complications. And in fact, these are the positive controls; negative controls are shown in the first part of each panel. And these are the positive samples that we've seen when we were doing PCR analysis and then blotting the PCR products with the JC-specific probe.

So JC we found, was present in greater than 50 percent of normal brain specimens that we looked at. When we sequenced these, as I said -- I believe I said -- these are rearranged forms. When we looked at normal kidneys, on the other hand, normal kidney was always archetype JC.

Well, at the same time when these initial studies were being done we did have access to five tumor specimens. These were five different patients with medulloblastoma. So we also looked for the presence of JC in those tumor samples and in fact, we found high levels of JC in each one of those five.

This here in lane 8 represents a normal brain -- one of our best normal brains in terms of the amount of JC that was present. You can see the difference in the amounts that were present. Again, this is a positive control and a negative control.

Now, more recently our PCR analysis has been conducted on a single PML patient. What makes this patient interesting we believe, it is unusual in terms of the way PML occurs. This occurred in a 5-year-old child. PML usually occurs later in life.

This child had severe combined immune deficiency. And so in addition to being young, we believe also that the PML arose following a primary infection rather than in most cases of PML where it occurs following the reactivation of an earlier persistent infection.

In addition, most cases of PML you can find JC in the brain, in the lymphocytes, and in kidney. When my grad student, Jason Newman, began looking at this patient, we had specimens from eight different tissues. And he found JC present in each one of these: brain, coeliac plexus, spleen, kidney, lymph node, liver, cardiac muscle, and lung. So JC was present in all of these, and this is just a southern blot to confirm the identity as JC.

Now, Jason was using primers that would lie outside of the regulatory region of JC in a highly conserved region for all JC isolets looked at so far. And so the PCR primers he was using would allow to amplify either archetype or rearranged forms of JC if they were present.

What Jason found -- this is again, the comparison, we're all going to compare with this archetype strain called CY. CY is a strain that was isolated by Dr. Yogo in Japan a number of years ago and I'm using that as a reference.

What Jason found was that in all the kidney clones that he looked at in sequence, they were identical to CY except for a single change at a hotspot of rearrangement for archetype at position 217. He also found, in cardiac muscle, archetype-like JC. That is, there are small deletions or intermediate size deletion without duplication. He also found in lung, another archetype-like strain.

This slide again -- with comparison to CY shown at the top -- this is the other tissues that he looked at. And what was he found was that there were multiple rearrangements occurring within these JC isolets. In fact, this is again unusual relative to most PML patients that are looked at.

Usually when you look at a rearranged form in a PML patient there's only one or maybe two types of rearranged forms. Here, we found multiple kinds of rearranged JC, and this is actually only a subset of what he found.

Now you also notice that in some cases the same clone was found in multiple tissues. In other cases you found that in the same tissue -- brain, brain, brain -- you found multiple clones. So there's a lot of rearrangement going on here.

Now as I said, the primers that Jason was using were laying right outside the regulatory region so they would amplify whatever was predominantly there. We wanted to see if archetype might be present in tissues other than kidney, since that really hadn't been looked at too closely.

This slide was constructed following JC's work with archetype-specific primers. In other words, he had two pairs of primers that he used, whereas one member of each pair lay within the 66 base pair sequence which would be archetype sequence that was present. Rearranged forms essentially in this patient, were always losing this region. So it would be specific for the presence of archetype JC.

And what Jason found, was in the brain and the lymph node that he could find the archetype JC or archetype-like variance within brain and lymph node, and this is the first time this has been shown.

So if I could have the last slide -- just some conclusions from this work with the pediatric patient that I've just been talking about. We believe that the immune status at the time of exposure -- this child has severe combined immune deficiency -- probably contributed to this widespread distribution of JC that we see in her body, and to the extent of rearrangement of this transcriptional control region.

The data I believe, shows us that rearrangements can occur early, perhaps after primary infection. This is in contrast to the past where we've been thinking that it usually occurred following the reactivation of a persistent infection. So at least sometimes, rearrangement occurs quite soon.

We believe lymphocytes are probably involved; that whether they're involved directly in rearrangement process we're not sure yet -- we may hear some more information about that next; and we believe that they certainly are involved in seeding the virus to secondary sites of infection.

As already mentioned, we do not know where the primary site of infection is, but lymphocytes may take it, wherever that is, to these other tissues that I've been showing you.

Finally, we would suggest that because we see different predominance of archetype and rearranged forms in different tissues, that in fact, this might indicate that the replication potential of these two forms differs in the differing tissues.

Now, given everything I've shown you so far, as I said, we have not seen anything to resemble recombination between JC, BK, and SV40 from these kinds of studies. I should add quickly, that we really haven't really looked that hard until recently. So in terms of whether or not rearrangement occur, that question is open.

Well, and partly for what I've shown, I think rearrangements certainly do occur; they occur in JC as well as in BK and SV40, leading to these archetype to rearrange the types of transitions that we see, so that this a highly plastic region of the genome that might be interesting to look at further.

Thank you very much.

CHAIRMAN BRIEMAN: Well, thank you, Dr. Frisque. You covered a lot of material within the time limit and I think that was very nicely done.

The next speaker is Dr. Maria Chiard Monaco of NIH, and she's been given a very limited time period to address the question of JCV infection of peripheral lymphoid tissue and the implications for viral latency.

Return to Table of Contents

DR. MONACO: Good afternoon. First I want to thank Dr. Lewis to give me the possibility to present my data today. As we heard from the previous speakers, JC is a DNA virus of the papomavirus family and as Dr. Khalili mentioned before, more than 80 percent of the human population has been infected by JC virus and few conversion occurs during childhood.

JC is thought as a neurotropic virus but as we can see from these slides, JC can also infect cells -- non-neural cells as well. Back in 1988, Dr. Sid Hoff and Dr. Major found that JC virus DNA associated to capitalize a lymphocyte from two PML patients in the bone marrow and spleens. And we expounded a host range of JC virus, focused our attention particularly on cells of lymphoid tissue.

In this slide we have a nested PCR amplification from two AIDS patients: patient 1 is not PML, and patient 2 was at the onset with PML. And this line marker, there's normal B cells contains DNA from a B cell isolated from peripheral blood mononuclear cells of JC virus-negative individuals.

Patient 1 and patient 2 -- in the lane of patient 1 and patient 2 we have DNA extracted from unfractionated peripheral mononuclear cells, or B and T subpopulation isolated by cell sorting technique.

And we found no evidence of JC virus DNA in the T cells from the PML patients and in the unfractionated peripheral mononuclear cells, or B and T cells from the non-PML patients. So this data shows us that PML can infect the peripheral lymphocytes in vivo, in particular, this population.

So next question we asked was, whether progenitor cells could be susceptible to JC viral infection. And to answer this question we looked at two progenitor cell lines, KG1 and KG1-A. These lines are derived from a patient who presented with leukemia but eventually developed acute myelogenous leukemia. But these two lines, CD34 antigen, that is a marker from stem cells.

As we can see in these slides we have in A, T-antigen-positive cells, and in B, T-antigen-positive cells detected with two different monoclonal antibodies, and in C we have some hybridization with the KG1-A positive cells using a biotinylated JC virus probe.

And the interesting feature of these cell lines is that when they are treated with formalizers they can differentiate -- KG1 can differentiate into a macrophage -- a mature macrophage. Instead, KG1-A cells are not affected by this treatment. So for this reason, we infected both KG1 and KG1-A untreated, and we saw susceptibility of these cells by JC virus.

When we infected KG1 treated, PML-treated cells, differentiating cells with microphage-like characteristics, these cells were no longer susceptible to JC virus infection. So this data clearly demonstrated that cells with monocytic lineage are not susceptible to JC virus infection.

And we confirmed the susceptibility of some cells, of precursor cells, with the primary CD 34, human parameter CD34 cells as we can see in these slides.

Previously, we described some of the interaction between tonsillar thermo cells, in particular B lymphocytes. Thermo cells are an important constituent of lymphoid organs, so we want to evaluate their involvement in the polyomavirus infection.

In this slide we have a representative field of tonsillar thermo cells infected by JC virus. The percentage of thermo cells infected by positive for JC virus T-antigen and viral DNA, was only 2.5 times lower than the percentage of human fetal iligos cells that are recognized as the most susceptible cells to JC viral infection.

So if JC virus occurs by respiratory route, tonsillar thermo cells, because they have relatively high susceptibility to JV viral infection, and for the natural interaction with the progenitor cells and lymphocytes are ideally positioning to be a site for initial infection and possibility to disseminate virus.

We are testing this working hypothesis by examining tonsillar tissue from children and other donors for the presence of JC virus DNA, and so far we've found 20 percent of these tonsils positive for JC virus DNA. We don't know yet in which cell type the virus, the DNA was sequestered, but this data are additional evidence that tonsils or other lymphoid organs could be a reservoir for the virus, or also initial site for primary infection.

This work has been done in Gene Major's lab and I want to thank him for his work. I also want to thank Blanche Korfman, Peter Jensen, Dala Galanti, Cathy Connor, for their help. Thank you very much.

CHAIRMAN BRIEMAN: Thank you, Dr. Monaco. Now, before going off to lunch in what, for those of you who haven't been here before, I think you'll find to be a very atypically excellent cafeteria for, you know, government cafeterias, let me remind you that we will return here at 1:40 for a panel-audience discussion that will be moderated by Dr. Mike Fried. And the topic is, "Issues related to the detection of SV40 DNA in human tissues". And again, if you have additional data, either positive or negative, regarding detection that you'd like to have presented, please contact him. Thank you.

(Whereupon, a luncheon recess was taken at 1:00 p.m.)

Updated February 12, 2001

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