DISCUSSION: Methodology of Virus Recovery Pincus: On the basis of informal discussion, several people have suggested a review of a few basic concepts regarding type C viruses. The complex polymorphism of endogenous type C viruses in many species has become increasingly apparent in recent years. Many criteria have been used in preliminary classifications, including host range (e.g., ecotropic, xenotropic), source of isolation (e.g., exogenous tumor, endogenous tissue), structural polypeptides (e.g., Class I, 11, 111). However, a comprehensive classification relating structure and function is not yet available, and a relatively detailed characterization must be provided of each particular virus isolate. It is also necessary to enumerate the technique used to detect the presence of a particular type C virus. Endogenous type C viruses have been detected in at least three very different forms: 1) Copies of viral genomes are found in normal tissues of all mouse strains and many other species, using nucleic acid hybridization. These genetic sequences are usually not expressed, either as viral proteins or infectious virus. 2) Gene products of the “viral” genome are frequently expressed as viral structural proteins, in the absence of complete infectious virus. A well-characterized example involves the GIx antigen, which is a variant of the viral envelope glycoprotein gp70. Most murine strains show low levels in the gp70 glycoprotein, as well as the major internal p30 protein, in normal tissues. 3) The least common form of type C virus expression involves actual infectious virus which can be isolated from animals or tissue culture. Infectious type C virus appears after chemical and physical stimuli (e.g., carcinogens, x-ray) in many mouse strains; endogenous viruses are also isolated from sporadic aging mice, with different incidences for various strains. However, infectious virus is found from birth in only a few strains, e.g., A K R mice, which express ecotropic virus and develop lymphoma, and NZB mice, which express infectious xenotropic virus and develop autoimmune disease. New Zealand mice share many features found in patients with systemic lupus erythematosus (SLE), and spontaneous expression of infectious xenotropic type C virus in these mice has provided a rationale to search for a type C virus in SLE. However, as we have heard here from Drs. Schwartz and Dixon. virus expression and autoimmune disease are not necessarily Arthritis and Rheumatism, Vol. 21, No. 5 Supplement (June 1978) correlated, and the possible role of virus expression in autoimmunity remains enigmatic. Furthermore, there is no conclusive evidence at present for the existence of a unique human type C virus; all type C viruses isolated from human tissue are identical in protein structure to known primate viruses. Finally, the finding of altered immunologic responsiveness in NZB mice and SLE patients, including elevated antibody titers to most common human viruses found in sera of SLE patients, raises the possibility that virologic phenomena may reflect rather than induce altered immune function. On the other hand, most of the immunologic abnormalities seen in SLE can be found with clinical viral infection, supporting the “viral hypothesis” in SLE, although viruses other than type C viruses must remain under consideration. Schwartz: May I ask a question about a point of nomenclature that has to do with the emphasis placed on the distinction between exogenous and endogenous viruses? For example, the mouse ecotropic virus is put in the category of exogenous virus. But in the mouse, there are integrated viral genomes that segregate as mendelian traits. These genes code for ecotropic viruses. Thus, there may be a point of confusion between true exogenous viruses like the feline leukemia virus or the bovine leukemia virus and the mouse ecotropic virus. By contrast with the former, the latter are truly integrated and segregate as an endogenous virus, like the xenotropic virus. Todaro: I think we both agree that the mouse ecotropic viruses are endogenous viruses. For virologists it is a very important distinction; it makes a great deal of difference in deciding whether one is dealing with an agent coming in from the outside, the way a virus does, or one is dealing with an endogenous agent that becomes activated. The mouse ecotropic viruses would be endogenous agents even though they may move from cell to cell within the animal, and perhaps recombine with each other. Feline leukemia and bovine leukemia are more complicated. From the epidemiology it is clear that the disease is a consequence of infection from the outside. Indeed, you could also say the same thing about certain kinds of murine viruses-that they too could be transmitted in a horizontal manner. s53 DISCUSSION Paul: Dr. Aaronson, would you tell us more about the mapping of genes for p30? Have any p30s been shown to exist in polymorphic forms? Aaronson: Genetic mapping in the murine system has been done by the use of temperature sensitive mutants and defective sarcoma viruses that turn out to be recombinants between a portion of helper virus and some other, presumably transforming sequences. Dr. Mariano Barbacid of our laboratory has defined the gag gene order that I described. Paul: What about the place within the genome of the mouse where the p30 gene exists? Aaronson: In the mouse, those types of mapping studies have not yet been finished. It is known that the structural information for one type C virus in the AKR mouse is in linkage group 1. Also, there is evidence with other endogenous mouse type C viruses that independently segregating inducibility loci contain viral structural information. The other point to make is that these viruses appear to be present within the genome as complete viruses; certainly this seems to be so in the mouse system. In other words, all of the information for a given virus can be localized to a discrete locus within the mouse genome. Steinberg: Since the AKR virus may be an unusual virus, can any of the other viruses be mapped? We've heard that there are 10 to 20 copies per genome. Would these be linked to a specific region of chromosomal material or are they randomly inserted in the genome? Todaro: That is not entirely clear yet. I mentioned one set of experiments in the cat where we have done crosses t o genome-negative cats. First of all, the F, has half the copy number and then the backcrosses seem to segregate all or none of the copies. This would suggest that the viral genes are not randomly distributed throughout the genome. But this is a case where there is a potentially infectious virus. It is worth pointing out that in some of the other cases like the primates, even though we can detect hybridizing sequences, we d o not know if in fact the whole genome is there. It may be defective-missing a piece necessary for replication. Alternatively, the genome could be scattered around the cellular genome such that the possibility of making a whole virus would be extremely small. Benacerrai: You have listed a variety of type C viruses but you have not said which ones you favor. I wonder if you would care to comment more precisely as to which is more important. Can you also tell us what approaches could be used to determine the evolutionary value of type C viruses for the host? Todaro: I d o not think this is an either/or situation: The genes can be helpful to the host in certain situations and harmful in others. Genes that confer disease resistance would have a strong selective advantage to the recipient. Certainly the resistance to infection is an important one, and I would expect resistance in the cases where the virus has been transmitted to a recipient host-for example, the baboon virus into a cat host. The baboon virus, which grows in many species, will not grow in domestic cat cells but will grow in leopard cells, so we are now in the process of growing hybrids and backcross cell cultures to see if there are a limited number of integration sites or if there is chronic expression, for example, of gp70, which blocks the receptors. In chickens such expression serves to block exogenous infection. But in the case where it is transmitted to a new species and acquired by a new species, the selective value would be that of conferring resistance to infection. Benacerraf: Is it a domesticated pathogen? Todaro: Yes. When the animals are first exposed to the virus, the prediction is that it might be quite lethal and that survivors are selected from among those able to successfully integrate the virus. For example, the virus from the Asian mouse is quite lethal to gibbons by producing leukemias and lymphosarcomas. If left unchecked, the mouse virus might integrate into some of the gibbons and would then offer resistance to superinfection, perhaps by occupying a limited number of integration sites. That mechanism obviously could not explain the whole process. Another part of it could be the gp70 differentiation story. There is not just one function, and I'm not sure whether one should consider the function of the whole virus or each of the individual components. The reverse transcriptase may also have a normal function; having that gene might be useful for a cell whether or not it has or expresses a whole virus. Tan: I wonder if Dr. Todaro or other virologists would tell us a little more about the proteins that are transcribed by the gag genes. I was interested in his saying that p12 is an RNA binding protein. Is this a basic type of protein or is it an acidic type of protein? In trying to draw a comparison with the basic proteins DISCUSSION and acidic proteins of mammalian cell nucleii, is there anything more known about the p12, p15, and the p30 types of these proteins? There are a number of autoantibodies in SLE that are specific toward not only basic but also acidic proteins, as Dr. Kunkel pointed out this morning. Todaro: The p12 of mice leukemia viruses is an acidic phosphoprotein; the p10 is a very basic protein that will bind nonspecifically to a variety of cellular DNAs. Ziff: Is there an immune response to gp70 in the NZB? On the one hand, we have Dr. Dixon’s data on the failure to find serum antibody and on the other Dr. Mellor’s description of staining for viral constituents in the glomerulus. Pincus: The controversy stems in part from the different methods used to detect viral antibody. There is evidence for naturally occurring antibody in sera of many mouse strains to whole intact and disrupted type C virus preparations. However, most murine sera that bind to whole virus preparations do not bind isolated viral proteins in radioimmunoassay. The explanation for this apparent discrepancy is not known. The whole virus assay may have greater sensitivity and it is possible that antigenic determinants that are bound with low avidity to natural antibody might be altered during protein purification. On the other hand, a nonviral protein adsorbed onto the virus and reactive with normal serum may be detected using radioimmunoprecipitation of whole virus. Ziff: But others might say that the immunoglobulin binding to virus is nonspecific. Pincus: There may be a specific explanation for reactivity with nonviral proteins. Rabbit antisera prepared to fetal bovine serum (independent of virus) show significant binding to whole virus preparations. It is possible that a complex of viral protein with a fetal bovine serum component may be involved in this type of radioimmunoprecipitation reactivity. As noted before, reactivity is generally not seen with isolated specific viral proteins. Warner: Dr. Nowinski showed that by using the naturally occurring mouse antibody in a precipitation assay, and analyzing the precipitated material by polyacrylamide gel electrophoresis, the viral p15 and p30 components were present. Pincus: This is correct. If an immunoprecipitate of normal serum with whole radiolabeled virus is analyzed on polyacrylamide gels, these proteins are demonstrated. However, I reiterate that one cannot demonstrate immunoprecipitation by normal sera of the isolated p30 or p15E proteins purified prior to immunoprecipitation. It’s a question of the order in which analysis of viral proteins is performed. This problem pertains directly to the question of a unique human type C virus. Normal human immunoglobulins are similar to normal murine immunoglobulins, in that both show the property of precipitating whole mammalian type C virus preparations. The human reactions fulfill many criteria for specificity, including lack of reactivity with uninfected cells or with avian type C viruses, which are immunologically noncrossreactive with the mammalian viruses. The situation is complex, and not fully understood at present. Schwartz: In the radioimmunoassay you described, Dr. Aaronson, I assume you take the bulk tissue and grind it up and make an extract and then test that preparation for antigen? Aaronson: Yes, you grind it up and sonicate the extract. You can perform immunoassays on the extract, but you can’t stop there because of the problem of nonspecific reactivity. Further purification procedures are required. Schwartz: It is certainly conceivable that within that mass of tissue there will be differential expression of a putative antigen on a minority of the cells. If, for example, only 3% of the cells in the tissue express that viral associated antigen, would you be able to detect this in so few cells in the tissue? Aaronson: The question pertains to how you do the assays and what you do to ensure that you aren’t losing something. I think that in any kind of system of evaluation you have to develop appropriate positive controls. In many endogenous virus systems, the approaches which we frequently described have worked in isolation of antigens that can be shown to crossreact with the p30 protein of some known type C virus. Black: There is evidence that mice develop immune responses both to their ecotropic and to their xenotropic viruses, with respect to both humoral and cellmediated immunity. In other words, there’s no apparent tolerance to their own endogenous virus. It may be difficult to distinguish from the immune response whether the virus is endogenous, where the genetic information is vertically transmitted and integrated into the DNA, or exogenous viruses, and where the trans- DISCUSSION mission is horizontal. This question relates to the two viruses that were supposedly isolated from human leukemia. One is a simian sarcoma virus from the woolly monkey and the other is the baboon endogenous virus. Humans do not have information in their DNA to code for such a simian sarcoma virus; so if we do apply this criterion, it implies horizontal transmission, like any classical infectious agent. Therefore one should develop antibodies to an acquired virus as one does with any infectious agent. The baboon infectious agent, as George Todaro outlined, is an endogenous xenotropic virus in the DNA of the baboon. The proviral DNA is present in a complete copy in the baboon. Thus, it’s fair to ask, do humans have an endogenous virus? So far, the only evidence that we do is that we have in our DNA about 10% to 25% of the sequences of the baboon endogenous virus and the immune response data that I have discussed. Schwartz: But there is a problem here, and I think it’s an important one. First of all we don’t know what the hypothetical agent of lupus might be. The tenor of the whole morning session was that a type C virus would be the best candidate. Of course, there’s no proof of this. Therefore, the kind of positive control you mentioned may be an antibody, rather than an antigen. In an “unknown” system a viral antigen could be expressed in a chronically infected individual on only a minority of cells. In fact, Dr.Viola referred to that this morning when he talked about subacute sclerosing panencephalitis and indicated that only certain parts of an infected brain may contain the virus, whereas other parts of the brain could be completely normal. My point is that maybe the approach of taking the whole spleen, for example, and making an extract of it and then looking for an antigen is not going to be fruitful because you will lose 1%, 5%, or 10% of cells that have the antigens you are seeking. Dixon: Knowing the sensitivity of your test, one should be able to calculate the threshold-approximately what proportion of cells would have to be positive, or how many specific molecules per 1Og cells would be needed for reliable detection. Schwartz: If you are talking about a low fraction of antigen-positive cells, it seems to me that detection of antibody to the putative antigen would be a far more sensitive technique. Phillips: I want to make a comment regarding the specificity of the measles antibody reaction This is old s55 data now that I don’t think we should have to reiterate. We showed 4 years ago that you can adsorb reactivity specifically with measles antigen and not with control culture antigen; the antibodies are very specific and they are elevated. Aaronson: I think that lupus provides a very important test system from the perspective of a virologist. Here we can try to determine whether there may be aberrant reactivity against endogenous viral proteins, whether p30, p10, pl5, or gp70. It’s a tedious screening process but it seems worth it to try to find evidence of immune reactivity against type C viral proteins. Patients with this disease may make immune responses against their endogenous viruses. So those kinds of studies are both worthwhile and in progress. Dixon: You were very clear about the failure to find antibodies against type C viruses in humans with a variety of diseases. How about the antigens? With your best assays have you found any of these viral antigens that you can correlate with lupus? Aaronson: As of now, we have looked at five lupus kidneys and the results have been negative. In terms of looking at affected tissues of a number of disease states, we’ve looked at over 100 or so-which is quite a number when you have to d o the kind of purification procedures that we believe are necessary. I want to stress that I believe that endogenous viruses will be found in humans just as they have been found in so many other species. But at the moment the evidence is not convincing. Phillips: Let me speak briefly about the point on the baboon endogenous viruses because the experiments that we’ve done to isolate the virus were specifically aimed at this possibility, and failed to reveal infective virus. The particles are there microscopically, whether they are type C virus or not. But we don’t get baboon virus out. So I think that the question is moot. One can almost say with today’s knowledge that the baboon virus is not there. Ziff: Dr.Black, why did you use placental material in your study of immunity t o type C virus in humans? Black: We looked in pregnancy for immune reactivity to a possible endogenous virus because 1 ) viruses are seen in placentas; 2) the baboon virus itself was isolated from a baboon placenta; 3) Gross and Feldman have found several type C viruses in rat placentas. When the viruses are seen in the placenta, they are seen at the maternal-fetal junction. Of course, this is S56 where one might expect to have an immunologic reaction-on the maternal or host side against the paternal antigens of the fetal graft in the placenta. Thus the expression of endogenous type C viruses might result from a host versus graft reaction; we have demonstrated that viruses can become activated from such reactions. Winchester: Dr. Black, in the cell-mediated immunity that you see at this point in pregnancy-with the virus-infected fibroblast system vigorously controlled for considerations of histocompatibility-have you, for instance, been able to use the fibroblasts of maternal origin to show that there is absolutely no possibility of the cellular response resulting from a weak DISCUSSION transplantation antigen enhanced by viral infection? Furthermore, for considerations raised by the Doherty-Zinkernagel experiments, there should be some histocompatibility between target cell and effector. Black: No, we haven’t done that. We attempted to utilize a system where effector and target cells share a histocompatibility antigen, but we haven’t used the mother’s own fibroblasts. However, the reaction is virus-specific because with the uninfected fibroblast we get no reaction and the same effector lymphocytes give no reactivity with similar cells infected with another type C R N A virus, the simian sarcoma virus. Also, adsorption of effector lymphocytes with disrupted virus absorbs the reactivity.