s222 D I SCU S SI ON: M ech a ni s ms of I m m un e R eg u 1 at i o n Hahn: Dr. Gershon, would you clarify whether you meant that Ly-l cells and Ly-23 cells can function in different situations either as helpers or as suppressors, or whether it is the Ly-123 precursor that can function either way in each situation? Cershon: There is no substantial evidence to indicate that the Ly-1 cell can do anything other than what it is tions of Fc receptors distinguish macrophages which have either suppressor or helper effects when mixed with appropriate T cells. Warner: Would the NTA antibody in NZB mice poten- tially interfere with any of the Ly typing that you performed? preprogrammed to do, that is, to help B cells or to induce a DTH response. This Ly-l cell seems to have differentiated to a state where it has lost all its previous options, and its only decision is whether to turn on or not; after turning on with Con A, it still functions the same way. The same is true for Ly-2 positive cells, although, of course the function of the Ly-2 cell is quite different. However, there is a large population of cells that are not known to be committed to a single function, and it is in this population that all kinds of strange things happen. The Ly-123 cells can act as amplifier cells, but also seem to be able to act as a cell population that can differentiate into either helper or suppressor cells. The work ahead of us is to try to fractionate the Ly-123 population by some mechanism in order to determine whether the different effects they produce are stable cells, or Ly-123 just on a maturation path. Gershon: If the NTA antibody were noncomplement- Williams: Since suppressor T cells are supposed to have Fc receptors for IgG and helper T-cell receptors for the Fc part of IgM, have you had an opportunity to look at these receptors in the subpopulations defined with respect to Ly typing? Hahn: I would like to make the point that one can Cershon: We haven’t looked at the Fc receptor distribu- tion among the cells. Bob Stout and the Herzenbergs in collaboration with Harvey Cantor showed that the Fc positivity cuts across all lines and a cell with a single function cannot be isolated by means of Fc receptors. No one I know has looked at subpopulations of Fc receptors on T cells in mice. There is also a problem in that the amount of Fc receptors on T cells will vary with their stage of differentiation. So this marker does not seem to have the advantage of the stability that the Ly markers have. But, we are looking at subpopulations of Fc receptors on macrophages and Erwin Diener’s group is doing the same thing. There is a definite suggestion that subpopulaArthritis and Rheumatism, Vol. 21, No. 5 Supplement (June 1978) fixing and did combine with theta antigen, that could be why we find theta negative cells. If the NTA is not gamma-1 and it is complement fixing, its presence would have been evident in the normal serum controls. Moreover, since there is an excess of Ly-l cells (over theta-positive cells), it is clear the NTA isn’t blocking in the Ly-l studies. Pernis: I was under the impression that Bob Stout could identify suppressor activity in the Fc receptor gammapositive T cells. Gershon: I agree that he has shown that, but there are helper cells that are Fc positive as well. So this particular marker seems to cut across functional cell lines and has not, so far, been helpful in distinguishing cells with defined functions. induce antigen-specific suppressor T cells in the NZBI W hybrid. Dr. Lewis Parker and I administered singlestranded DNA on poly-D-lysine carrier 3 times a week to male or female B/W hybrid mice beginning at birth. I n this study we showed that disease onset can be significantly delayed, with less severe nephritis and reduced amounts of circulating antibodies to DNA. The next question was whether we could transfer this phenomenon with cells, and the answer was that we could. This state of tolerization can be transferred by placing tolerized spleen cells into lethally irradiated syngeneic B/W recipients. Tolerance to DNA was assayed by measuring plaque-forming cells to DNA in the spleens of recipients repopulated with tolerized or normal spleen, thymus, or bone marrow cells. Recipients of tolerized spleen, thymus (2:l to normal bone marrow), or bone marrow cells were tolerant to sDNA. In contrast, the plaque-forming response to S223 DISCUSSION immunization with SRBC was the same in all three groups. This same approach prolongs life in older mice, but we have been unable to transfer tolerance with cells from the adult. Warner: I have some questions for both Dr.Steinberg and Dr. Waldmann: I would plead for a more open view of the suppressor system. Although it may be essential that a suppressor T cell is involved, this may not be the cell that is producing the final suppressive factor. I wonder if we might keep open the possibility that there is a defect, for example, in the macrophage lineage. Could one envision the possibility that the suppressor T cell, when activated, stimulates the secretion of a factor from the macrophage? And, of course, there are other data that suggest macrophage defects in these mice. Since the word macrophage is probably just as all-embracing as the word lymphocyte, we must recognize that there could be many subpopulations of macrophages. In relation to this, could prostaglandin be a factor? Have you cultured your cells in the presence of indomethacin? Waldmann: Rich and Pierce showed that T cells are absolutely required for the generation of SIRS. Dr. Krakauer has confirmed these observations. If you remove T cells from spleen cells of mice with antitheta antisera or by affinity columns, or if you use nude mice, SIRS cannot be generated by the B cells and macrophages that remain. It is true that macrophages are required for SIRS generation as well, according to Pierce and coworkers. In addition, I’m sure you are aware that the immediate target of SIRS, as viewed by Dr. Pierce, is the macrophage. We have not determined whether the macrophages in our system produce prostaglandins of the E series, as has been reported for the inhibitory macrophages in Hodgkin’s disease by Goodwin and coworkers. Gershon: I want to comment on Dr. Warner’s point. In the collaborative work of Pierce and in that of Harvey Cantor, it was shown that the Ly-2 cell makes the SIRS factor. The evidence that it is a T-cell product is fairly strong. Whether that is the suppressor factor in this case is not known. I would also like to make a comment to Dr. Waldmann: we haven’t looked for SIRS production in old NZB mice, but since they are loaded with Ly-2 cells, it is quite likely that they can make SIRS.The production of SIRS is a very complicated phenomenon because in the presence of Con A the Ly-1 cells make a helper factor that can inhibit or counteract the presence of a suppressor factor. The old NZB mice, by our analysis, have excess of both types of cellshelpers and suppressors-and they’ve lost modulators. You might be able t o alter the culture conditions or the dose of Con A that you are using and find that you can produce SIRS from these animals. Waldmann: I think that it is obvious t o everyone dealing with Con A that this lectin stimulates the activation of both helper and suppressor cells and their products. This makes analysis of the effect of Con A dependent on the study protocol. For example, if low doses of Con A are used or if it is added late in the culture period, stimulation rather than inhibition is observed. If it is added at the onset, suppressor effects dominate. Did you try any other regimens of Con A? Gershon: No, we tried identical ones. Steinberg: Dr. Gershon’s supposition is correct. If the supernates are fractionated, one can get both helper and suppressor factors with regard to certain functions, for example, nonspecific proliferation. Green: I would like to briefly present some studies done along with Drs. Sakane and Steinberg on suppressor cell measurement in patients with SLE. The experimental design was similar to that reported in the Journal of Experimental Medicine (143: 1 100, 1976) by Shou, Schwartz, and Good in which a two-stage culture of lymphocytes from the same individual was employed. Pure T and B cells were prepared and the T cells were stimulated with Con A for 3 days. The same individuals were bled again 3 days later, and again pure T and B cells were prepared. The Con A stimulated cells from the first culture were added t o cells obtained from the second bleeding. Inhibition of mitogen or MLR responsiveness in the second culture was used as the index of suppressor activity. In the normal individuals we have done thus far, we observed a very significant depression from 63% to 71% of the mitogen response of the second culture when purified T cells, but not B cells, from the first culture were added. This occurred with PHA, Con A, or pokeweed mitogen as the stimulant. To determine which subset of T cells was responsible for this suppressor function, T cells were separated on BSA gradients according to their buoyant density prior t o exposure to Con A. The results of these studies revealed that the T cells of higher density gave the greatest degree of suppressor activity. Of interest was the fact that the cells incorporating the most ’H- S224 thymidine were present in the fraction of cells of lowest density. These studies were published in detail in the Journal of Immunology (1 19:1169, 1977). So far we have measured suppressor cells of 5 patients with systemic lupus. In all of these patients some degree of defect was noted, at least to some of the mitogens. For example. when Con A was used in the second culture as a stimulator, there was a marked difference between normal and SLE cells. The SLE cells from the first culture did not inhibit the Con A response, whereas the normal suppressor T cells could inhibit this response. The SLE cells when used as responder cells in the second culture were inhibited by normal stimulated T cells, but when the normal cells were used as the responder in secondary culture, the SLE T cells from primary cultures did not inhibit this response. Thus, the SLE cell can respond to a suppressor T cell from a normal individual but by itself cannot generate suppressor cell activity. suggesting that these suppressor T cells are actually missing to some degree in lupus patients. Kunkel: Dr. Steinberg, do you have any evidence that anti-thymocyte serum was effective at body temperatures in the animal? This is primarily a cold-reactive antibody. Have you investigated whether it is possible that other things in the serum of these animals might be producing your effect? Steinberg: Good questions. With regard to the cold reactivity, there is no question that the in vitro phenomenon that was used to describe the activity of the antibody is cold-dependent. That, of course, doesn’t bear one way or another on its potential in vivo binding to cells and alteration of those cells, with reference to either trafficking or functions. Perhaps not all of the possible controls have been done. We don’t find any good reason to believe that what we are looking at is anything other than NTA. We can isolate the IgM fraction of the serum to get this effectof course, the IgM fraction contains other things including immune complexes-but these adsorb with thymocytes, thereby removing the activity. We have done a number of other controls, such as adsorption with liver powder. There is a very good chance that the active factor is NTA, but I don’t think it has been rigorously proved. Kunkel: One other aspect of that: have you absorbed with B cells? DISCUSS ION Steinberg: I’ve adsorbed just lately with nude cells and some other B cells. These studies are going on now and I don’t yet have the results. Barnett: We’ve heard of different responses in nude mice to so-called T-dependent and T-indpendent antigens. From the data with SIRS and low-molecular weight suppressors, what would be the prediction of response at different ages with so-called T-dependent and independent imm unogens? Steinberg: Obviously, there are a lot of abnormalities in New Zealand mice. As they get older, they have poorer primary responses to T-dependent antigens. They make reasonably good secondary responses to Tdependent antigens, and they tend to make good responses to T-independent antigens. SIRS has not been demonstrated to have in vivo effects in contrast to the low molecular weight suppressor. The latter causes enhancement rather than suppression of the response to disease in old NZ mice. Studies in nude mice indicate that the timing of the administration of CONS is important-given with antigen, it suppresses, whereas given several (74) days before antigen, it enhances. Waldmann: SIRS would act to inhibit both T-dependent and T-independent humoral immune responses without affecting cell-mediated immune responses. This contrasts with Con A activated T cells that suppress both humoral and T-cell proliferative responses. Hahn: I n regard to your experiments with NTA accelerating nephritis in B/W mice, can you elute NTA or antibody to it from the glomeruli? Steinberg: No, we don’t know anything about that yet. Williams: With respect to Dr. Waldmann’s model, we know that there are many components in the supernatant of the Con A stimulated cells. Have you had occasion to look at delayed-type hypersensitivity, skin testing, or any other parameter in the treated animals? Waldmann: We have not looked at even primary circulating antibody responses, much less delayed-type immune responses. Con A activated T cells act on multiple functions, including both cellular and humoral functions. However, Parker and coworkers had DISCUSSION looked a t cellular versus humoral responses and concluded that the humoral product-SIRS-acts only on antibody or immunoglobulin-type responses, not cellular ones. Williams: The hypogammaglobulinemia you produced was striking. And since SIRS are known to be prone to low pH, digested by pronase, and have a relatively well-defined molecular weight, have you looked at this in adult acquired agammaglobulinemia patients to see if they produce a lot of these SIRS? Also, wouldn’t it be possible finally to devise a radioimmunoassay for the material so that one could titrate patients, if this is what they need? Waldmann: We haven’t purified SIRS, but we are using crude supernatants of Con A stimulated cells that might contain a number of active factors in addition to the protein described by Rich and Pierce. Ziff: Dr. Waldmann, might the amount that you give0.5 p1 twice a week or so-be too small an amount to act on a whole animal? Is there any possibility that it is recruiting other cells? Waldmann: We gave 0.05 ml of Con A supernatant 5 times a week to a 20 gram mouse. We have no information as to whether it would recruit other cells. NZB/W animals given the Con A supernatants do not have suppressor cells in the spleens before or after in vitro treatment with Con A. I would not anticipate that the Con A supernatants would activate other T cells to become nonspecific suppressors. Gershon: I would like to comment that it was Pernis’ pioneering work with the delta-antibody that suggested to us how to find this Lyb 3 antigen. We had been playing around with it for 6 months, and not until we applied the techniques pioneered by Dr. Pernis were we able to get the augmentation and find the antibody activity. Warner: Dr. Paul, would you elaborate a little more on the mu/delta story? You make the interpretation that there is more m u rather than less delta. Paul: The reason w e reached the conclusion that the abnormal mu/delta ratio on CBA/N B cells probably results from increased m u rather than diminished delta is that fluorescence-activated cell sorter analysis indicates that CBA/N B lymphocytes have, on the S225 average, more m u than do normal B lymphocytes. Indeed, we also reached the conclusion from these studies that i n normal neonates there was an enormous heterogeneity in the amount of immunoglobulin per cell and that, on the average, there was probably somewhat more IgM per B cell than was found on adult B cells. The questions have been asked: 1 ) what relationship exists between the Lyb 3 marker that Huber et al. described and the Lyb 5 marker that I discussed; and 2) if injected into a recipient, does anti-Lyb 5 increase responsiveness to antigens as anti-Lyb 3 has been reported to do? The answer to the first question is that we have not yet established that Lyb 3 and Lyb 5 are determinants on the same molecule. It seems likely that this may be so, but these experiments have not yet been performed. With regard to the second question, preliminary experiments have not shown that anti-Lyb 5 increases antibody responses. However, sera with anti-Lyb 5 activity block in vitro anti-TNP antibody responses to TNP-Ficoll, suggesting that this differentiation antigen is not only a phenotypic marker but is also a membrane structure important in regulating B cell activation. Gershon: In that regard, there is one other important distinction between the two antisera: your serum only reacts with the cells of some mouse strains. The Lyb 3 antigen, as far as we can tell, reacts with all strains of animals; it is not reacting with an allotypic determinant, whereas the Lyb 5 apparently is. That doesn’t mean the two sera are not reacting with the same molecule, but it does mean that they are reacting with a different antigenic determinant. Paul: I would certainly agree. Most of you may not be aware of how anti-Lyb 3 is produced. It is a serum produced by the immunization of CBA/N X BALB/ C F, male mice with BALB/C cells; thus it is an F, anti-parent antiserum. The F, recognized the Lyb 3 antigen presumably because it lacks that particular antigen, but there is no reason why it should make antibodies to polymorphic antigenic determinants rather than to antigenic determinants common to mice of all strains. Thus, it is not at all surprising that the antibody is not allotypic. On the other hand, Lyb 5 is clearly an alloantigen, and backcross analysis indicates that it is controlled by a single locus. S226 Friou: I am quite interested in your K-cell data because, if I recall correctly, the mouse cell that lyses cultured cell lines is a nonphagocytic cell, as is the human cell that has that type of activity, Since we found a reduction in that component of the K-cell activity in active SLE, would you speculate on what kind of presumably secondary activity might be involved in yielding reduced activity in your animals? Paul: We really don’t have a good explanation for that result. It was, I must say, unanticipated. It’s still quite possible that it does represent a secondary effect. Of course, in lupus, circulating immune complexes are an obvious possible cause for reduction in ADCC activity through the saturation of the Fc receptor on the K cell. We have no evidence that circulating complexes exist in CBA/Ns, but we have not looked for such complexes. Another possibility is that there is a central defect in K-cell function; perhaps the K cell is in the lineage of B cells, although not formally identifiable as such a cell. Kincade: I want to make a few comments about some recent functional studies that we’ve done relevant to the B cell differentiation studies described by Dr. Paul. The assay is an in vitro cloning technique where individual B cells are dispersed in a semisolid gel matrix and stimulated with mitogens such that they proliferate and form macroscopic colonies. Neither helper nor suppressor T cells affect the assay. Macrophages can profoundly influence cloning proficieny, both enhancing it and suppressing it, but we can standardize for that to get a linear assay. The central finding was that CBA/N mice do not have any clonable B cells a t any stage of development. As expected, the F, male mice that carry the mutant X chromosome also do not give B-cell colonies. In the females of this cross, with one mutant and one normal X chromosome, the number of colonies is about half what it is in the age-matched wild type mouse. Thus it appears that the kind of cell that gives colonies correlates perfectly with the kind of cell that CBA/N mice lack. One finding that favors the concept that in normal mice this is a distinct lineage rather than a stage in differentiation is that fetal liver and spleen early in gestation contain B cells that form colonies. Talal: Dr. Roubinian and I have studied the regulatory effect of sex hormones and the thymus by investigating the sequential development of IgM and IgG anti- DISCUSSION bodies to two nucleic acid antigens-double-stranded DNA and synthetic single-stranded RNA (polyadenylic acid). Both of these antibody responses occur spontaneously in B/W hybrid mice and also occur in lupus patients. Very early in life, B/W mice have very low levels of antibodies. Starting at 4 months of age, females develop a progressive glomerulonephritis associated with an increase in these antibodies, particularly in the IgG response. This occurs later in males, again showing the influence of sex in the animal model for SLE. The response to the synthetic single-stranded RNA polyadenylic acid, is different. In very young B/ W mice there is no IgG antibody at all. The response is exclusively IgM. Polyadenylic acid is not present in the immune complexes found in the kidneys. One wonders what the immunogen might be for this response. We know that sequences of polyadenylic acid are associated with messenger and viral RNAs. The effect of thymectomy performed at 2-3 days of life is different in the male compared to female B/W mice. Mortality in the males is augmented by thymectomy, whereas in the females, mortality is decreased by thymectomy. The enhanced mortality of males induced by thymectomy is reflected in earlier appearance of anti-DNA antibodies. However, the thymectomized males who are dying earlier failed to develop IgG antibodies to polyadenylic acid. The IgG anti-poly A response is essentially aborted by the removal of the thymus. Thus, these experiments suggest modulating influences both for the thymus and for sex hormones. The effect of the thymus is different on these two spontaneous antibody responses to DNA and poly A. Our results are consistent with the removal of a suppressor influence for the DNA response, causing an earlier appearance of antibodies to DNA. They also suggest the removal of a helper influence for the poly A response that would, without thymectomy, appear later in life to permit the development of IgG antibodies to poly A. Davis: I want to make one comment about nucleic acids. We have been interested in free DNA in human sera and human plasma. One frequently finds free DNA after surgery, injury, and burns, associated with certain drugs, and after pulmonary embolization-sometimes large amounts of free DNA. We consider this phenomenon something that is probably quite repetitive and may occur in all of us from time to time. DISCUSS ION Here, then, is another example of our tolerance to a ubiquitous material. I don’t know whether there is any evidence that free DNA has led to disease in any of these patient groups, except for those with SLE. Steinman: A large number of syndromes are associated with transitory free DNA in the circulation. We also found that there is at least one fairly common clinical condition associated with repeated exposure of the individual to circulating DNA over long periods of time, and that is chronic hemodialysis. We found that release of circulating DNA following dialysis occurs uniformly. There has also been a report of the occurrence of ANA in these same patients. We wouldn’t know whether these patients were at risk for glomerulonephritis, but they clearly don’t get anything else that looks much like lupus. So we can certainly tolerate DNA alone for long periods of time. Kunkel: Before we discuss this point, I think it is extremely important to stress that you have to know whether it is ssDNA or relatively native DNA. Relatively native DNA doesn’t immunize experimental animals. These animals have been examined repeatedly, yet nobody has found any expression of disease. But when the molecule is denatured, the bases are very antigenic. So the fact that these animals tolerate DNA without getting any manifestations of disease is not at all surprising from the experimental animal point of view. But if you have denatured DNA around, that is another story. That needs to be specified. Steinman: We haven’t looked at that specifically, but it seems likely from what we know about the DNA molecule that any piece of double-stranded DNA that has been subjected t o physical shear, as it would in the circulation, is likely to have gaps along it and perhaps single-stranded ends, so that there would be at least some exposure to the bases. Ziff: Loss of suppressor T cells has been suggested as a reason for the development of a spectrum of autoantibodies of which those to DNA are, of course, the best known. I think it is appropriate at this time at least to consider any possible connection between depression of supressor T cells and viruses. This brings up the question of the thymus and the virus: What is the role of the virus in affecting T-cell function through the thymus? S227 Talal: It seems to me that the argument becomes circular. On the one hand, we have heard from several people in the last few days-Paul Black said it earlier-that activated cells, particularly B lymphocytes, are a very fertile soil for virus production. This argument says that first you get an activated lymphocyte, and then that activated lymphocyte produces virus. On the other hand, we have also heard-as Dick Gershon said-that the immune system behaves as if it is stimulated by chronic infection. This argument suggests that the first event is expression of the virus. Dr. Warner suggested that virus may be selectively expressed on certain subpopulations, as for example suppressor T cells. Thus, there is the circular argument: you keep going around and around and maybe, as often happens in this kind of situation, both arguments are correct. The problem is: Is there any experiment that would allow us to determine whether lymphocyte activation or virus expression comes first? I was disappointed to learn a number of things at this meeting that would appear to argue against the idea of a suppressor cell deficiency. Dr. Dixon and David Katz recently examined a number of suppressor functions in NZB/W mice and found them to be normal. Ralph Williams has said that he couldn’t find T-cell binding of DNA and would have expected to find it, if there were suppressor cells for DNA. I’m going away from this meeting with a feeling that the now almost conventional idea of suppressor cell deficiency needs to be reexamined. Warner: Just two comments on this point which I think are worth stressing. The question of the existence of suppressor cells and their absolute number in NZB mice is something that can now be analyzed. Dick Gershon gave us some preliminary data on Ly typing of NZB, which if anything showed an increased number of Ly-23 cells, and he made the point that in the absence or reduction of Thy- 1, this is compatible with differentiation toward terminal stage. However, it is now clear there are at least two different Ly-23 positive cells, namely the suppressor cell and the cytotoxic cell. They can be distinguished on the basis of the I-J gene product that is expressed on the cell surface of only suppressor cells. Hence, NZB mice should by typed for the number of I-J positive cells. There may well turn out to be further complexities, but a t this stage, this does seem to be a specific marker for suppressor cells; so it should be looked at. S228 I would like Dr. Kunkel to comment on a problem: The idea of a regulatory defect is attractive, but we have to differentiate between a generalized suppressor defect that affects the entire regulatory network which is antigen-independent, or an antigen specific defect. The problem with the idea of a generalized regulatory defect mediated through a hormone, through NTA antibody eliminating the suppressor function or other agent, is that we don’t see the perpetuation of any antibody response that the mice naturally encounter. There have been demonstrations in the literature of hyperresponsiveness in NZB, but 1 think it is fairly clear now that it is not a completely generalized phenomenon. I wonder how this is reconciled. Must we have a separate set of suppressors that deal with autoantibody producing cells, or is it going to be an antigenspecific suppressor cell that we are dealing with? Kunkel: To some degree we are overly conscious of the autoantibodies because we do not usually measure other types, so the fact that antibodies are not there, and then suddenly they are there, may be the most visible index we have of a lack of suppression. Perhaps if we could study these patients’ responses to immunization in more detail, we would see the effect on other antigens. I think the T-cell defect is a very real one that we have to cope with. There are so many bits of evidence that came out a t this meeting and that are coming out in the literature, particularly recently in the human clinical studies, it is a logical extension to suppose that it is a “generalized” lack of suppression that is involved. Lipsky: I would like to mention one piece of work presented in Miami by Dr. Tom Chused which might be of some interest. He demonstrated in the NZB mice that by taking single-cell suspensions of spleen cells and letting them synthesize immunoglobulin in vitro during short-term culture for 4 hours, these mice make increased amounts of immunoglobulin compared to normal strains. And this can be demonstrated on the first day of life. This finding suggests at least that these animals are born with some sort of regulatory defect, again raising the question of whether some of the suppressor abnormalities that develop later in life might not, in fact, be an interesting but secondary pheonomenon. In this same context regarding nonspecific Tcell help in patients with lupus, we have studied a DISCUSSION variety of patients with active disease and looked at their peripheral blood lymphocyte response to pokeweed mitogen as a polycolonal inducer of immunoglobulin secreting cells. We find that in all of the lupus patients whom we have studied so far, this response is highly abnormal, and they do not respond to pokeweed mitogen. This defect, at least, can be ascribed to an absence of an appropriate T-helper cell in the peripheral blood. Clearly, we cannot make any comment about what might be going on with regard to the internal lymphoid organs of these patients. But, at least with regard t o their peripheral blood responsiveness, we have clear-cut evidence here that there is no nonspecific T-cell helper. In fact, we cannot induce a Thelper cell with a nonspecific activator such as pokeweed mitogen. Ziff: Don’t you think that, in spite of the fact that the specific suppressor may be important in the control of specific antibody disease, it is more likely that we are dealing with broad effects? Because as we mentioned before, we get a group of autoantibodies emerging at the same time. Krakauer and coworkers have in fact demonstrated that SIRS can knock the whole business out. So broad regulation seems to be a more important aspect of lupus erythematosus than the specific control that may be involved in individual autoantibody responses. The fact that a patient develops a lupus-like syndrome during an attack of infectious mononucleosis is evidence for a broad effect. Tan: I just want to make a comment that perhaps man may be a little more complicated than the mouse. I think that in the mouse, and to a degree in man, we are pretty sure there is a T-cell defect that may account for the syndrome of systemic lupus. Immunogenetics may play a role. But the mouse is not exposed to certain drugs as man is, for example, to hydralazine or to procainamide. And the mouse is certainly not exposed to many things that we are exposed to in food preservatives which are highly reactive chemicals that may be very reactive with cell membranes and intracellular components. I think we are all convinced that lupus is a multifactorial disease. Another feature is that certain environmental influences may also predispose or help to induce the disease-the studies that we have shown with ultraviolet light, for example. Ultraviolet light can induce the thymidine dimers in DNA which are DISCUSSION highly antigenic, but this is not the whole story. UItraviolet light can also induce many changes between nucleic acids and proteins such as cross-linkings that may be very antigenic and that may in part be responsible for the induction of some of the autoantibodies to nuclear antigens. There is also the controversial finding concerning the deficiency of repair enzymes which are responsible for repair of damaged or altered nuclear macromoleclues such as thymine dimers. There are so many factors that can help to produce the disease that we should not limit our S229 thinking to just T o r B cells or the genetics but expand it to cover these other areas. Talal: There is no question that we have to visualize the pathogenetic mechanism of lupus as a very dynamic system, not static at all. I t doesn’t necessarily start in one place or in a single event. It’s like Yin and Yang in which several events may be intertwined. This leads to a degree of circularity in which cause and effect become blurred. Thus the question for all of us to ponder is: how can we diminish the tendency to be caught in this circularity and move toward finding primary mechanisms?