OPENING DISCUSSION -FREDERICK STOHLMAN, JR., M. D.

The data presented at this meeting seemed to be leading toward the inevitable conclusion that leukemia in man is associated with a reverse transcriptase which may represent the "foot prints" of an RNA tumor virus. From this it may be suggested that human leukemia is in fact of viral origin. The chemical and immunologic properties of transcriptase which have been described in acute myelocytic leukemia, chronic myelocytic leukemia, and the lymphocytic leukemias are similar. The end results, i. e., the leukemia, however differ substantially in their clinical manifestations. One might suggest therefore various subspecies of transcriptase or virus with similar general physical and biochemical properties but each with significant differences that permit the development of varying types of leukemia. Alternatively the determinant as to the type of leukemia may rest within the cell which the virus infects, or perhaps the site ofinfection within the pluripotent cell. To date the studies on transcriptase have required fantastic amounts of tissues which have not permitted adequate controls from normal tissues or the sequential evaluation of the leukemia as the patient enters remission and relapse. Perhaps the best control is the phytohemagglutinin stimulated lymphocyte but even this has been criticized by some. Further, the relative insensitivity of the techniques currently available for evaluation of transcriptase have precluded detailed study of the cells in question. The purification of the enzyme promises that specific antibodies may be developed which can be applied with immunofluorescent techniques to evaluation of the single cell. This together with differential separation ofbone marrow cells should permit the evaluation of the importance and distribution of transcriptase in stem cells. Development of such technology in my view is critical to gaining meaningful insight into the role of reverse transcriptase and viruses in human leukemia. In reflecting on the role of reverse transcriptase there are several considerations which come to mind. Leukemia may be viewed as a disease in which there is abnormal information presented to the cell, presumably at the pluripotent stem cell level, which results in abnormal growth patterns. Although there are several different types of myeloid leukemia, e. g. acute myelocytic, progranulocytic, etc., the number is circumscribed. Does this reflect a high specificity of the viral effect at a few sites on the DNA molecule or does it reflect that most of the viral directed effects are lethal and hence a leukemic clone does not develop, there being a random chance that the transcriptase affects the human stem cell in such away that leukemia eventuates. I gather from the discussion at this meeting that it is yet to be established whether the viral information is passed vertically and awaits only derepression or whether a more complicated hypothesis must be invoked. It is clear, however, that whatever theories are advanced they must take into account the leukemogenic effects of irradiation, alkylating agents and other leukemogenic drugs such as chloramphenicol. In the case of irradiation there is a suggestion of a dose-effect relationship at high dose levels. Is this due to somatic mutation, derepression of an oncogene or a more complicated method of initiating viral infection ? Clearly information on reverse transcriptase in patients with acute myelocytic leukemia thought to be secondary to irradiation or alkylating agents is of importance. The lag phase between the primary insult and the development of clinical leukemia and the events which transpire at a molecular level need to be explored. Investigation of these relationships, however, is dependent on the development of microtechniques for the evaluation of reverse transcriptase and RNA tumor virus. Without it we are left with gross correlations. The characteristic of the cell in myeloid leukemia should be considered. These differ in several respects from that of the normal cells. The generation time of the cell in acute myelocytic leukemia may be longer or the same as that of the normal cell; as a result the fractional turnover rate is not increased but, due to the greatly expanded pool size resulting from the failure of differentiation or loss of a "death" function, the total growth of leukemic cells is greater than seen in the normal myeloblast compartment. The leukemic myeloblast or its progenitor may migrate from the marrow into the peripheral blood and proliferate in extramedullary sites, where normal myelopoiesis is not seen. This suggest changes in membrane properties and the interaction between microenvironment and the leukemic cell. Additionally, there is some data to suggest that the leukemic myeloblast may have specific antigens; these are being explored immunologically and may permit improved therapeutic strategies. A third consideration is that the leukemic cell differentiates partially, i. e. from stem cell to myeloblast and even to progranulocyte but further differentiation usually is not observed. Using the soft agar technique it has been claimed that leukemic clones may differentiate normally under the direction of colony stimulating factor. As I mentioned in my formal presentation, however, the evidence on this point from a morphologic and functional standpoint is inconclusive. Further, it seems to me that to view, acute leukemia solely as a failure of the normal interaction between a granulopoietic regulator and the myeloblast is an over simplifaction. Failure of differentiation and as a result the lack of a "death function" of course leads to an ever increasing number of proliferative cells but the capacity for growth in extramedullarysites is necessary for the complete evolution of the disease. I t is possible of course that if we were able to switch the balance of differentiation from leukemic to normal clones, leukemia could be controlled but in my view this most likely will require manipulation of differentiation of the pluripotent stem cell not the committed myeloid cell or myeloblast.

Discussion following Dr. Stohlman's Talk

Dr. Ostertag: Dr. Stohlman, you are talking about three compartments of the hemopoietic system: the compartment of the pluripotent stem cell, the compartment of committed stem cells and the differentiated compartment of megakaryocytes of myeloid and erythroid elemen ts. You further say that in your view leukemia originates back in the pluripotent stem cell compartment. As supportive evidence you cite the chromosomal change as observed in CML as Philadelphia chromosome, which can be found in cells of all three compartments. Now you do get a specific chromosomal change in many patien ts all suffering from chronic myelocytic leukemia; why can't you get the same kind of chromosomal translocation at the same time in different compartments of the hemopoietic system of the same patient. A viral etiology could easily account for that.

Dr. Stohlman: If we assume for the moment a viral etiology, and I would suggest we are really talking about a virus, then this virus theoretically might affect DNA in a number of ways. The numbers of different types of leukemia are quite restricted, which leads me to think that most effects of viruses are lethal. Whether the immunologic surveillance system identifies and destroys those transformed cells which are recognized as abnormal, perhaps due to membrane changes, or the transformation results in an intrinsically abnormal cell which dies, perhaps after a few divisions, I don't know. The reason a person gets chronic myeloic leukemia, acute myelocytic leukemia or acute lymphocytic leukemia may be either due to the species of virus that transforms the molecule or the same species may hit DNA at random, most of the lesions being lethal, only a few of them being compatible with proliferation; the Philadelphia chromosome is one manifestation and is associated with CML. It would be most improbable for all patients to have three different cell types affected simultaneously.

Dr. Ostertag: I would like to disagree. Let's say a virus is going into the cells affecting the same place in the DNA and breaking the same chromosome in all cells. You would get the same change in all different compartments of the hemopoietic tissue.

Dr. Gallo: That is a perfectly acceptable alternative possibility, and most particularly since a common chromosomal alteration appears limited to CML. Dr. Stohlman: You also have CML without the Philadelphia chromosome.

Dr. Gallo: Where there are marker chromosomes in AML or ALL you usually do not see them in all three cell lines.

Dr. Stohlman: But you do see the same abnormalities in all three cell lines.

Dr. Hardesty: If leukemia is a stem cell disease, what do you think is the cytological basis of the disease: do the stem cells replicate too rapidly or do their daugh ters stick around too long? You further mentioned a colony stimulating factor that is involved. in the regulation of myelopoiesis. If one assumes that the normal myeloblast and the leukemic one has a factor, the normal cells should be dominant over the diseased cells. Has somebody tried to make a fusion between leukemic and normal myeloblasts? If the hypothesis is true the normal cells should be able to cure a cancer cell.

Dr. Gallo: To answer your second question first: I don't know if it has been tried yet in vitro. However, many people with abnormal bolle marrows were transfused with Ilormal cells. In a few illstances apparellt trallsformation of the donor normal cells ill the recipiell t leukemia patiell t developed. Whether this was due to cell fusioll remaills doubtful but this would be all argumellt agaillst your hypothesis (i. e. dominallce by a normal cell) .To your first question I would say, I am not collvinced that leukemias are "stem cell" diseases. If the origin (first cell affected) is always, indeed, the stem cell, thell I would say, proliferating stem cells give rise to too many myeloblasts in places they normally should not be. Normal granulocytes go all over the body, they often for instance localize in areas of infection. In short, it might be said that they "metastasize" just like cancer cells, but they know how to terminate, to die. They do not have the potential for continued replication.

Dr. Stohlman: Normal myeloblasts, progranulocytes and myelocytes stay where they are supposed to, namely the bone marrow but abnormal leukemic myeloblasts are seen elsewhere. The abnormal myeloblast grows in extramedullary sites, divides and replicates itself, or in my view it is more likely that the leukemic stem cell is responsible and it and its progeny ( the myeloblast etc. ) grow in extramedullary sites. The normal fate of agranulocyte is to be released from the bone marrow and to go into the peripheral blood, to go from the peripheral blood to sites of infection. This cell does not have the capacity to divide. The polymorphonuclear granulocyte is an end stage cell just as the red cell, the normal fate of which, after extruding its nucleus, is to enter the peripherial blood, circulate, provide oxygen, and then die in anywhere from 30days to 10 months depending upon the species. But the myeloblasts and myelocytes do not normally enter the peripheral blood. There is the reason to suggest that this is due to membrane characteristics, namely stickiness and a lack of deformability which prevent early myeloid elements from entering the peripheral blood. Part of maturation is to develop deformability, and lose the "stickiness". In order to release myeloblasts into the peripheral blood there must be an abnormal bone marrow architecture or an abnormal membrane either of which would alter the interaction of the cells with the environment. When you inject myeloblasts into the peripheral blood or stem cells into the peripheral blood, they hone and grow only in the bone marrow or in the spleen.

Dr. Gallo: Let's discuss tuberculosis or other like diseases where immature normal cells including myeloblasts are released into peripheral blood.

Dr. Stohlman: What Dr. Gallo is referring to, is a leukomoid reaction which one occasionally sees in tuberculosis and usually what this means is that immature cells are circulating into the peripheral blood. Myeloblasts are not seen in these situations and moreover the cells in leukomoid reaction don't proliferate.

Dr .Hardesty: What happens when a patien t goes into remission ? I t seems that this is a pertinent question with respect to whether it is a specific block or whether it is somethingwayback.

Dr. Stohlman: I'll give you my thought on that. Normally, you have, what I've been referring to as a pluripotent stem cell and for purposes of conversation it's very easy to consider this a homogenous or uniform compartment. However, it probably is not, one might suggest that a given pluripotent stem cell has a finite limit to the extent of this replication and once exhausted another more immature stem cell is triggered to begin differentiation and you have what is called "clonal" succession. Now, if you treat a leukemic patient with chemotherapeutic agents and induce a remission what may be happening is that a normal stem cell line has differentiated out, remission will last as long as one continues to have normal stem cells differentiate. Relapse comes when either the succeeding stem cell is leukemic or perhaps virus has transformed another normal cell, the presumption being that if you have the capacity for remission then you have normal stem cells present along with leukemic cells.

Dr .Hardesty: Yau haven't answered the question really. What happens in this respect with Friend virus induced mouse leukemia? In my experience, the spleens are just literally enlarged. The animal is overwhelmed with cells. If there is a remission in this kind of situation do these cells go on and differentiate or do they die ?

Dr. Stohlman: In human acute myelocytic leukemia you kill the cells. If you give a patient with acute myelocytic leukemia Daunomycin it wipes out almost all of his cells. You then go through an aplastic phase and then if there is regeneration from a normal pluripoten t stem cell, recovery is seen. The Friend virus which produces an erythroleukemia in mice does not have a counterpart in human tissue and hence I am not certain that discussion of this is germane to the present problem.

Dr. Hofschneider: I would like to know the difference between a chronic and an acute leukemia in cellular terms or whatever terms you like. Do you have cells in chronic leukemia that do not grow as fast as in the acute form ?

Dr. Stohlman: Chronic myelocytic leukemia is a disease of myeloid cells in the peripheral blood and bone marrow. Immature cells (myelocytes, progranulocytes and myeloblasts) which normally are not present in the peripheral blood are present. It may be associated with anemia, polycythemia, thrombocytosis or thrombocytopenia. Classically there is a decrease in leukocyte alkaline phosphatases. Acute myelocytic leukemia, of the classical variety in young people, is associated with a large increase in the number of myeloblasts in the bone marrow and myeloblasts are present in the peripheral blood without evidence of differentiation such as is seen in the chronic myelocytic leukemia (i. e. in chronic leukemia differentiation is seen). In AML there may be growth of the leukemic cells outside the bone marrow. It could be the lungs, the heart, the brain, the skin and so on. Patients with chronic myelocytic leukemia, at some point develop a blast cell crisis which mimics acute myeloblastic leukemia.

Dr. Hunt: If human leukemias are caused by viruses, do you say the virus infects and is present in the genome of the pluripotent stem cells, and again if that is true, why does it only affect one cell line? What is your explanation for that? Have you ever tried looking in these leukemic patien ts in any other cell types for the presence of virus particles or viral genomes in the DNA ? And what is the answer?

Dr. Gallo: The answer is no, we have not looked at other cell types or tissues, only the affected cells, but it is a very good point. Perhaps I can speculate. I think it's possible that some people are infected with a leukemia virus from without which carries information essential to the induction. This has not been ruled out. However, I don't think infectivity from without is going to be the common mechanism for cancer causation in man. There may be people who disagree with that. By "infectivity from without" I mean active viral infection within in a relatively short period of time relative to the onset of the disease .

Dr. Stohlman: If I may comment on this point. The myeloblast of course predominates in this disease and one can obtain enough cells to examine. But to determine whether there is virus or transcriptase in other cells, erythroid, megakaryocytes etc. will require more refined techniques as there are not enough cells to permit study with current methods.

Dr. Gallo: You're right. We need abetter microtechnique. The antibody idea I think is now very feasible. We have available antibodies to primate Type C-virus reverse transcriptase. They cross with the human enzyme, at least some that we've recently isolated. Therefore, we can develop, hopefully a possibility of looking at small quantities of cells by immunochemical methods. This will be a major thrust of our future work in this area.

Discussion following Dr. Gallo's Talk

Dr. Hunt: Dr. Callo has been reviewing the properties of reverse transcriptase and also informed us about recent experiments to produce antibodies against it. Before we come to the second point, can I ask about one property of this enzyme, its capacity to make double-stranded DNA. What is known about it? Does it in fact make double-stranded DNA ?

Dr. Gallo: In most of our experiments the endogenous reaction has been done in the presence of actinomycin D. We, therefore, have not adequately looked at this. However, Peter Duesberg has.

Dr. Duesberg: Yes, they do, but you know we do not know much about this synthesis of DNA other than short pieces of DNA are made perhaps 5 or 6 Sin size given a template of 3 or 4 million daltons at least. In the primary reaction it makes complementary DNA, and then the secondary reaction leads to double-stranded DNA also of small size. Whether this is the whole story, that is the complete transcription from RNA that ultimately leads to a complement of double-stranded DNA, as it presumably happens in the cell, is completely open.

Dr. Gallagher: I would like to go back to your (Dr. Callo) hypothesis that you drew with regard to a hot spot in the DNA because if that is true you might predict anew form of virus and this could get out and infect other cells and so forth. You might be able to test this. Spiegelman's lab could perhaps tell us if there is an increase in the differences between normal and leukemic DNA over a period of time in a leukemic patient, perhaps during the course of CML or something like that. Have they checked it?

Dr. Hehlmann: No, we don't have sequential data in one patient through the course of the disease except one case of ALL, in which we detected viral related RNA during the acute phase of the disease that was not detectable after remission. (Spiegelman and his group have completed these results. There are no leukemic DNA sequences in leucocytes during remission. Moreover, leukemic DNA sequences have been found in the leukocytes of only the leukemic member of identical twins; Proc. Nat Acad. Sci., 70,269-2632).

Dr. Kufe: At first, according to this hypothesis you said that you think that the malignant state requires the addition of exogenous information and then you went on to evolve the hot spot hypothesis and proposed it was a mutation or addition via recombination of that hot spot. Now is that saying that there was oncogenic potential in that sequence that just had to be altered. Is this just a variation of the oncogene hypothesis only that it requires abase change or something like that?

Dr. Gallo: I believe Dr. Kufe wants to know if the proposed hot spot is either just a sequence that is specially receiving a carcinogenic "boost" thus beingjust a variant of the oncogene theory, somatic mutation, or added new information ? The proposal demands new information. There was only oncogenic potential by virtue of its unusual susceptability to change. This is clearly distinct from the oncogene hypothesis. However, regarding the nature of the change, I don't think it is useful to attempt to distinguish between the alternatives since as yet the data available, including the important paper that your lab published in this respect, might be explained by amplification, i. e. a difference in some nucleotide sequences between normal and leukemic cells, sequence X after transformation becomes X 50. Your experiment may not differentiate between those two possiblities. Moreover, it is of course, not yet proven that those "extra" sequences are pertinent to leukemogenesis, although I would like to assume with you that they are.

Dr. Kufe: I have to answer that according to the sensitivity of these assays, it would be impossible to have X originally to be amplified. That is, X had to be introduced from the outside because we would have picked up X on the hybridization assay.

Dr. Gallo: Are there viral (type-C RNA tumor vires) genes in some normal cells? Everyone by now must believe that there are some virogenes in at least some normal cells. I would like to know where they came from -or which came first -are these virogenes in fact really cell genes which the virus utilizes? Duesberg should speculate on this.

Dr. Duesberg: That's too much for me. That's like all theories on the origin of life: Where do whales come from, where does God come from, where does a "clean chicken " come from ? But I would like to return a question to you, may be somewhat related to that. I think we can at least divide those viruses which cause cancer and those which are sub-virus like things which may be a consequence of cancer. I think that those which are causing cancer may be like the men and the other more or less like the boys. So I think that shouldn't be confused too much. I think these sub-viral particles or endogenous viruses or incomplete endogenous viruses or enzymes might in fact well be a consequence of cancer rather than its cause. But I think there is little doubt that Rous SY or AMY can be the cause of cancer .

Dr. Gallo: I kind of agree with that, at least they cause chicken cancer. I think the information for carcinogenesis may be packaged into only very special type-C RNA tumor viruses. But I wouldn't even make those viruses that you call boys any less important because boys can become men. Moreover, we have now demonstrated that the reverse transcriptase in human leukemic cells and the viral related nucleic acid is related not to endogenous non-oncogenic type-C viruses, but specifically to typeC viruses which in fact are oncogenic such as the woolly monkey simian sarcoma virus.

Dr. Duesberg: That is absolutely right. I could have called them girls but I gave you boys. Maybe I could ask one more question. When you talk about leukemia or certain types of myeloblasts that you clinically find are these all genetically or antigenically homogenous ? In a given type of luekemia, is there always a unique population of cells ? Or could there be heterogeneity, could it be a random thing, just a random messing up of differentiation ? Or could it be that it all results from a single cell and leukemic cells are all identical? I think the identIty of leukemic cells would be more compatible with a genetic or viral theory whereas "'radom" could be regulation or who knows what?

Dr. Stohlman: There are a restricted number on types of leukemia and frequently one sees a monotonous type of cells morphologically. I don't know that anyone has analy sed the genetic information from these to say it is identical from cell to cell.

Dr. Duesberg: In these chromosome linked diseases like the Philadelphia chromosome, do you see the change only in the leukemic cells or also in other cells ?

Dr. Stohlman: The erythroid (red) cells and the megakaryocytes (platelet precursors ) all have the same chromosome.

Dr. Hehlmann: I now refer to DR. Gallo's very interesting data on antibodies prepared against reverse transcriptases of primate vjIuses which cross react well with your hurnall leukemic enzyme and offer a new immunologic approach. You have just said you have not prepared an antibody against your human leukemic reverse tran. scriptase. You had that enzyrfie fairly purified in the past. What are the difficulties in producing all antibody?

Dr. Gallo: We have been giving it to Bob Nowinski, and he is inoculating rats with the pure enzymes. So far we are losing a lot of enzymes. i.e. no antibody to date. I am becoming very discouraged unless we come in with about a 1 000 grams of leukemic cells so that we can get a lot of this enzyme and are then able to hand him one or two milligrams of enzyme. We tried Only twice, we failed, and we didn't relish the idea of losing more of this enzyme.

Dr. Stohlman: I would like to ask Bob one question which shows my immunological incompetence, is it necessary to really clean up and purify this enzyme. Don't some people suggest that you have a better chance of forming good antibodies if things are dirtied up a bit and thten you absorb the stera later ?

Dr. Gallo: I don't know that I am more comptetent than you are in immunology but I will answer as best I can. In the first trials when the antibodies to viral reverse transcriptast; were prepared, the reverse transcriptase wasn't purified enough and it is true, success was achieved easier in those laboratorities which didn't purity as much. It is also apparently true that when you purify more, you reduce antigenic potency. However, if you finally succeed with the purest preparation you are obviously in a much better position. In the long run the results as well as the antigen are cleaner .

Dr. Hofschneider: I have to apologize if I don't ask about reverse transcriptase and such things. I would like to come back to the colony stimulating factor. I have just met, maybe one or two weeks ago, some cell biologists and have asked them for the factor and they told me it is better to forget about it as a specific agent. Apparently, here in the audience are many people who believe in this factor. I would like to have some more information. Is it known what is the chemical nature of the factor, has it been enriched and to what extent, and has it been applied to animals and what was the effect?

Dr. Stohlman: It's a glycoprotein and various molecular species have been reported from 15.000 to 60.000. It has been given to normal animals. There is a problem in giving it to normal animals in that it is difficult in the experiments done thus far so separate effects of the "release factor", the release of granulocytes from the storage compartment, from true proliferation. The studies to date just don't separate them. I can't answer the question of its physiologic role. when human marrow is cultured with CSF after 12-14 days you get a significant number of eosinophilic colonies, maybe 30, 40, 50 percent, and in the normal human being you certainly don't see this degree of eosinophilic myelopoiesis. So I would raise the question if maybe CSF is a triggering substance, there being other regulators. Most of the evidence suggesting a physiologic role for CSF is inferential. I'm sure it does have one but for various reasons I don't think we have worked it out.

Dr. Torelli: Since I have been under provocation by Dr .Gallo to give my views I would like to say something about the nature of the leukemic cell. I think it is quite evident that we are talking about the etiology of leukemia and we're talking about the virus which is probably being brought into the leukemic cells but we still have to deal with the question, what are the leukemia cells. Because it is quite clear that we are trying to get rid of this question simply by saying: well, these cells do not mature, these cells are blocked in maturation. I think that we should be carefully comparing normal immature cells with leukemic cells. It's quite true that for a long time, in attempting to compare leukemic cells with normal cells, studies were hampered by the fact that many were comparing cells which were proliferating (leukemic) and cells which were not proliferating (normal). I think results at this stage of our studies were useless. We are really faced now with one main question. What is the key difference in the leukemic cell and the appropriate normal cell control. I think that point has to do with the introduction of a viral genome into the cell. This introduction should not bring the cells to limited progression. I think we should look for major changes which are brought into the cell by the introduction of the virogene.

Dr. Gallo: There is one point which I don't think came up at any time in the meeting. I am referring to some cases of bone marrow transplatation. There were two reported cases of recipients that were leukemic who received bone marrow from normal donors and apparently when relapses occured the normal donor cells had transformed in the recipient patient. Now there's a lot of discussion ofhow valid those observations were; how clear were the results which were based on cytogenetics. If, however, these results are valid, there is obviously a very important lead which directs us to almost only one conclusion, that a transforming agent remains in at least some leukemic patients. We can talk about cell-cell fusion, but I doubt whether this would occur under these conditions in viva. Even in vitro under the best conditions, it is difficult. Normal donor cells then apparently can transform in recipient leukemia people, and the most likely interpretation is that the "transforming factor" is still present after destruction ofleukemic cells.

Remarks Concerning the Discussion

It was quite impossible to include in this book the whole discussion, which lasted for more than 12 hours, in its entirety. We had to, unfortunately, leave out the greater part, and limit the discussion to the summary reviews of Dr. Fred Stohlman, a clinical hematologist, and Dr. Robert Gallo, a medical molecular biologist. Even from the discussion following Dr. Stohlman's and Dr. Gallo's reviews we were forced to cut a great many interesting critical remarks, and were only able to include 30 to 40 per cent. For the critical selection we thank Drs. R. Hehlmann and T. Hunt. Because of space limitation many of the authors have included a considerable part of the discussion in their articles and many questions, arising from the different investigational trends, have been summarized in the introduction.

Rolf Neth