Chapter XI DNA Antibodies Chairman: MAXIMESELIGMANN, M.D. DR. SELIGMANN:Four or 5 years ago when we presented our work about DNA antibodies in 1upus,1 many people thought we should have quotation marks on both “DNA” and “antibodies.”As far as quotation marks on DNA are concerned, I think everybody now agrees that the reaction is with DNA itself and there is no need to discuss it; it is precisely because we are dealing with a relatively pure substance that so much interest in the subject was raised in different laboratories. As for the antibody nature of these factors in lupus, I think (just as in the case of the rheumatoid factor) that this meeting presents a good opportunity for anyone who has strong evidence against the antibody hypothesis to give it during the discussion. This is also a reason why I think it useful to report briefly the different findings which we think confirm that anti-DNA antibodies in lupus are true antibodies. First, they are detectable by quite different immunologic technics such as precipitation in liquid or gel medium,2v3 complement fixation4-0 and passive cutaneous anaphylaxis reactions.1° Second, the precipitin curves that were obtained if the pH was above 7 were quite analogous to those of an immunologic reaction and we could recover in the precipitate all the DNA in the antibody excess zone. Third, and we thought this the most important, it was possible to extract by action of deoxyribonuclease the proteins (or at least a great part of the proteins) linked to DNA in the washed precipitate, and it was easy to demonstrate that this protein was immunochemically a 7s y-globulin.ll This finding .~ DNA antiwas exactly confirmed by Deicher, Holman, and K ~ n k e l These bodies, in all cases of lupus sera we studied, appeared to be immunologically 7 s 7-globulins and behaved as such on DEAE columns. Now there are two disconcerting facts which well could be against the antibody hypothesis: ( 1 ) The fact that apparently these antibodies lacked specificity, in that we had similar positive reactions with human DNA, phage DNA, bacterial DNA, and so on. In Ouchterlony plates we saw a continuous line of precipitation with these different kinds of DNA.” ( 2 ) The fact that it was extremely difficult, if not impossible, to obtain DNA antibodies in immunized animals. We have some new results on these two subjects which I am sure will be important topics for us to discuss. The apparent lack of specificity for different kinds of DNA led us 2 or 3 years ago to try to look for the reactive sites on the DNA molecules.12I would Iike to present some of these results even though they are rather old because some of them show discrepancies with some of Dr. Levine’s results and it may be stimulating for discussion. 542 543 DNA ANTIBODIES g globulines cpg) - Serum No 1635 +a DNAnatif 0--0 DNA dWnt6grb par ultra Sons \ 0 1 5 10 20' '\ -4 DNAparmldeserum()LQ 1 50 Fig. 1.-Precipitin curve of DNA degraded by ultrasonic waves with fragments. Although we did not use as sensitive technics, we saw, just as he did, that considerably denatured and degraded DNA still reacted with the lupus antibodies. But when we studied DNA degraded by ultrasonic waves with fragments which should be approximately 300,000 molecular weight, we had a positive complement fixation reaction and the precipitin curve (fig. 1) was very similar to that obtained with native DNA except that, as you could expect, in the antigen excess zone it went down more quickly. This kind of curve was found with four lupus sera. However, Dr. Levine only found inhibition with sonically treated DNA. Now, you can notice on this curve the very high ratio of protein to DNA in weight and this means a considerably high molecular ratio, although what we used here is what we thought to be native calf thymus DNA. It may have contacted some denatured DNA but in any case it is not easy to understand this tremendous ratio. Another discrepancy is that when we destroyed DNA by DNase, the end products obtained with high amounts of DNase not only did not give us any positive precipitin or complement fixation reactions but were unable to inhibit the reaction even when 500 times more than the amount of native DNA required for absorbing all precipitating antibodies was added. What we saw also in this study is that when we used acidified DNA at pH 1.7 without heating, we had no decrease of the immunologic reactivity. When we used DNA which was first acidified-and then heated for increasing times, we had no more reactivity and at the end we were unable to inhibit the reactions with an amount 80 times more than the amount of native DNA required for neutralizing the precipitating antibodies ( table 1). Therefore, we thought that the purine bases were involved in the reactive sites. I may add that RNA did not inhibit the reactions with DNA. Perhaps some of these discrepancies mentioned are due to the fact that from one serum to another and from one patient to another we are dealing with different kinds of DNA antibodies. 544 IMMUNOLOGIC ASPECTS OF RHEUMATOID ARTHRITIS AND SLE Table 1.-Reactions with Acidified and Heated DNA Reactions with Lupus Serum Heated PH at for Precipita- 70Purine tion by GI04 Basis H0,K N Liberated - Ring-test +++ ++ + Neutralization Complement of precipitating fixation antibodies; +++ ++ + lxn=+ 97 2xn = 90 12 3xn = 86 18 4h 8h (+) 6xn=+ 72 40 (+I 20xn = t + 34 02 24h 80xn = 24h 24 98 80xn = 24h 3 100 *n = minimum amount of native DNA necessary for neutralization of the precipitating 1.7 1.7 1.7 1.7 2.1 1.7 1.7 37C. 37C. 37C. 37C. 37C. 30C. 37C. Ih 3h + + + antibodies. Another important point in Dr. Levine’s studies is that in some way we are able to see that the reaction with the different kinds of DNA is also different quantitatively from one DNA to another. As far as the production of anti-DNA antibodies in immunized animals is concerned, all our attempts in rabbits were unsuccessful regardless of the method employed. I think that until the work of Dr. Levine with T4 phage DNA, most of the positive results in the literature were at least open to discussion. This work, however, gives the first conclusive evidence that you could obtain DNA antibodies despite the fact that the antibodies were a particular kind of DNA antibody specific to glucosylated DNA.13 Now, the fact that we are able to induce formation of anti-DNA antibodies in animals gives stronger evidence that you may have true antibodies to DNA in lupus but does not mean at all that these anti-DNA antibodies are pathogenic and noxious in the patients. As Dr. Dameshek noted, I do not agree with the term “connective tissue disease,” but neither do I agree with the term “autoimmune disorder.” I think that all we can say is that lupus is a disease with autoantibodies and, among others in some cases, with anti-DNA antibodies and nothing more. But the studies on experimentally produced anti-DNA antibodies may be very important and interesting for future research about the eventual pathogenic role of the anti-DNA antibodies in lupus. First of all, I think that with these kinds of antibodies (even if they do not come from lupus patients) we may perhaps have a definitive answer to this very important question: Is an antinuclear antibody able to penetrate a living cell? I think we have, as far as lupus antinuclear factors are concerned, much evidence against it, but this question is still open to discussion and has not yet been definitely solved. I think also that the soluble complexes eventually formed in vivo in lupus patients (complexes between different antinuclear factors, and perhaps especially anti-DNA and the homologous antigens) may be pathogenic and lead to formation of vascular lesions similar to those of Dr. Dixon’s model. I think that it would be very important to detect the antigens in these lesions where we find y-globulins and complement; among these antigens, it would 545 DNA ANTIBODIES Table S.-Positive Reactions with DNA (per cent) Precipitation Complement Fixation Bentonite Flocculation 75 75 87 46 48 76 10 0 12 0 28 10 47 48 70 Systemic Lupus Erythematosus (217 samples from 93 patients) Before treatment Under treatment but no real improvement Under treatment and great improvement Apparent clinical remission Per cent patients giving, at laast once, a positive reaction Other Dkemes Rheumatoid arthritis (73 patients) Scleroderma, polyarteritis nodosa, dermatomyositis, discoid lupus, (60.patients) Miscellaneous (300 patients) 1.3 1.3 5 0 0 0 0 0 _____ 0 be perhaps relatively easy to detect DNA and the experimentally produced DNA antibodies could be useful for this purpose. I would like to know, during the discussion, if anyone has some evidence for the presence of DNAanti-DNA soluble complexes in the blood of lupus patients. We have had some evidence for it, but we could not draw from our experiments definitive conclusions. Aside from these very interesting theoretic problems, there are a few practical questions we could well ask: Are DNA antibodies useful for the diagnosis of lupus? I would first like to show our results (table 2 ) . I think it is extremely important in lupus to give results in connection with the stage of the disease. This is especially true for DNA antibodies. The spectrum of antibodies present in lupus sera differs considerably if you have untreated patients or patients already in treatment or patients in remission. That is why we give our results by number of samples and according to the stage of the disease. You can note that 75 per cent of the patients with definite lupus, prior to any treatment, gave precipitin and complement fixation positive reactions; when they are treated and improve (but not in remission), this incidence falls to 10 or 12 per cent; in remission the reaction was never positive. Now, in the other diseases we had altogether one single serum giving a weak positive reaction with these technics: this was a rheumatoid arthritis patient who had L.E. phenomenon but not its clinical nor pathologic features. Since last year, and after the. paper of B o z i c e v i ~ h ,we ~ ~ have worked with the bentonite flocculation test. We had much difficulty the first months, but now we have reproducible results. This test does indeed increase the sensitivity of the reaction and we have still some positive sera during remission. The problem was to see if with this increase in sensitivity we had more specificity for lupus. To the present we found only 5 per cent positive 54.6 IMMUNOLOGIC ASPECTS OF RHEUMATOID ARTHRITIS AND SLE results in cases of rkeumatoid arthritis (patients were selected on the basis of also having positive L.E. phenomenon). I am quite aware that we have a much higher percentage of positive reactions with DNA than most people in this country have reported. At the beginning I selected nontreated cases in order to study anti-DNA antibodies, but now this contains a wider sampling and I can only present my data even if they are not completely in accordance with yours. Our tests were made at a pH of 7.5 or 7.8. SPEAKER:Were these rheumatoid arthritis patients examined for L.E. cells and for antinuclear factors? DR. SELIGMANN: Yes, they were examined for L.E. cells. The 5 per cent are all patients with positive L.E. phenomenon and they all have antinuclear factors in fluorescence; the incidence of antinuclear factors in fluorescence is of course much higher than these 5 per cent. The one patient giving a positive reaction in complement fixation is a true rheumatoid arthritis patient with L.E. phenomenon. DR. LEVINE:While investigating" the immunochemical nature of T2 bacteriophage, it was found that rabbits immunized with alkali ruptured bacteriophage produced antibodies to DNA. The anti-DNA in these sera reacted more effectively with denatured (single strand) than native (double strand) DNA.13J6These bacteriophage DNA antibodies were specific for the DNA of T2, T4, and T6 bacteriophage,16 and did not react with DNA preparations from about 40 other sources. It was found subsequently that the antibodies to phage DNA were directed in part to the glucosyl moieties17J8(either a, f3 or diglucosyl residues of the hydroxymethylcytosine unique to these DNAs ) . It appeared that in the native, double strand, helical form, antigenic determinants of DNA were less available for reaction with their homologous antibodies. As a result of collapse of the double strand helix, these antigenic determinants were unmasked and reacted more effectively with anti-DNA. If the antigenic determinants are masked in the helicaI form of DNA, tlte single strand state should be more reactive in other DNA immune systems. Dr. David Stollar looked at the other DNA immune systems known at that time; the DNA reaction with anti-DNA present in SLE sera4-6,9J2 just discussed by Dr. Seligmann. The data shown in figure 2 describes the complement ( C ) fixation of six SLE sera with native and thermally denatured T4 phage DNA. It can be seen that denatured DNA reacted more effectively than native DNA.l9 In general most of the SLE sera gave these results, however, some of the SLE sera fixed C' with native DNA (although more fixation was observed with denatured D N A ) , and one SLE serum reacted as well with native as denatured DNA, It must be remembered that this C' fixation technic20 used diluted antiserum (1:500with serum C.C. in figure 2 ) . More concentrated SLE serum does give C' fixation with native DNA. "This is publication No. 207 from the Graduate Department of Biochemistry, Brandeis University, Waltham, Mass. Aided in part by grants from the National Institutes of Health (E-1940)and the American Cancer Society (E-222). 547 DNA ANTIBODIES Serum Hu. Serum N.A. Fig. 2.-C' fixation by LE sera and varying quantities of native and thermally denatured DNA. 0- - 0, native DNA; 0 - - 0, thermally denatured DNA.2f The anti-DNA in SLE sera reacted with about 40 DNA preparations from various sources,z1 confirming the earlier ~tudies.~J' Quantitative differences were observed, however, both in the extent of C fixation by a given serum with denatured DNA of different origin and the amount of C' fixation by diflerent SLE sera with native and denatured DNA. Again, with a single SLE serum, denatured was a more effective antigen than native DNA.21 The variation found in SLE sera was ascribed to differences in specificity of antibodies and the variation from DNA to DNA with respect to reactivity of native and denatured DNA was probably due to variable extents of single strand areas in our DNA preparations. Dr. Stollar next attempted to identify the antigenic determinants of the DNA, DNA digested by pancreatic deoxyribonuclease did not fix C with the anti-DNA in SLE sera. The digests, however, did inhibit the homologous DNA-anti-DNA reaction although the amount of enzymatic digest required for inhibition varied from serum to serum. With two SLE sera, pyrimidine oligonucleotides (prepared by treatment of DNA with diphenylamine and formic acid) and the enzymatic digest were equally as effective for inhibition of the DNA-anti-DNA reaction. It appeared that the antigenic determinants 548 IMMUNOLOGIC ASPECTS OF RHEUMATOID ARTHRITIS AND SLE mpMOLES INHIBITOR Fig. 3.-C' fixation inhibition of serum M.M. and denatured BaciUus natto DNA by increments of pyrimidine oligonucleotides.22 in DNA which reacted with the antibodies in these two SLE sera were pyrimidines. Pyrimidine oligonucleotides of varying chain lengths were isolated by DEAE chromatography and tested for inhibition of an L.E. antibody. The effectiveness of inhibition increases with increasing size of oligonucleotide, up to the pentanucleotide.22The oligonucleotides from each chromatographic peak were separated by paper chromatography according to cytosine and thymine content. Inhibition by the separated pyrimidine tetranucleotides is shown in figure 3. The effectiveness of inhibition reflects the thymine content of the oligonucleotide. When polythymidylic acids of varying chain lengths were tested for inhibition it was found that the pentathymidylic acid was only slightly more effective than the tetrathymidylate.22 Most of the SLE sera we have studied are inhibited by enzymatic digests much better than pyrimidine oligonucleotides,23suggesting that purine oligonucleotides or purine-pyrimidine oligonucleotides are antigenic determinants. The earlier studies of Dr. Seligmann12 had also suggested the antigenic role of the DNA bases. We are presently attempting to isolate and identify purine oligonucleotides from apyrimidinic acid and oligonucleotides from enzymatic digests. The anti-DNA in one SLE serum (G.B.), which we thought might be directed to a small antigenic determinant because its C' fixation response resembled quantitatively that obtained by Butler et al.24with antibodies to puronyl-conjugated bovine serum albumin, was investigated. These antibodies in the serum (G.B.) were characterized as anti-DNA by their lability to pancreatic deoxyribonuclease and their greater reactivity with denatured DNA ANTIBODIES 549 Fig. 4 . 4 4 C' fixation by serum G. B. (1/200) and Bacdlus nutto DNA kept for 10 min. at O", 51", 65", 73",82' ( a ) ,91" ( A ) , and 100" ( O ) ,and then quickly chilled. (b)Immunologic thermal denaturation profile of Bacil2us natto DNA as measured by serum G. B. DNA.23It can be seen in figure 4 that the C' fixation can be used to follow the transition of BaciEZus mtto DNA from the native double strand to single strand state. This serum was analyzed by inhibition with deoxyadenylic, deoxyguanylic, thymidylic and deoxycytidylic acids. The quantity of nucleotides giving 50 per cent inhibition was found to be 3Q, 230, 400, and 1000 mpmoles respectively, suggesting that adenylic acid was part of the antigenic determinant reacting with the anti-DNA in G.B. Further studies showed that the adenine moiety of deoxyadenylic acid was the active inhibitor. The data in table 3 show the inhibition of serum G.B.and denatured calf thymus DNA by purine derivatives, histidine and chloroquin. The more effective inhibition by some of the purine analogues and especially theobromine and theophylline raises the possibility that the DNA-anti-DNA in G.B.serum is really a cross-reaction. For example, if G.B.had produced antibodies to theophylline of theobromine, the C' fixation with denatured DNA could be due to the structural similarity of adenine and these purine analogs. Chloroquin is also an effective inhibitor of the DNA-anti-DNA in G.B.serum (50 per cent inhibition with 50 mpmoles of chloroquin). Unlike the nucleotides and nucleosides, chloroquin also inhibits the DNA-anti-bacteriophage DNA reaction. Dr. Stollar has demonstrated that chloroquin inhibits the DNA-anti-DNA reaction by binding not to the anti-DNA but to the DNA.25The binding was measured by equilibrium dialysis, fluorescence, and spectrophotometric technics and by inhibition of bacterial DNA transformation. Holman has shown that chloroquin inhibits the L.E. cell phenomenon.2* h. SELIGMANN:Could you tell us how many different reactive groups you were able to find for such a number of sera? Second, I wonder if it wouldn't be useful to perform with strong sera 550 IMMUNOLOGIC ASPECTS OF RHEUMATOID ARTHRITIS AND SLE Table 3.-C% Inhibition of Serum G.B. (1/200) and Denatured Calf Thymus DNA b y Purine Derivatives, Histidine, and Chloroquin Inhibitor Theobromine Theophylline 6-Methylaminopurine 6-Chloropurine 6-Mercaptopurine 2-Aminopurine Isoguanine Purine Adenine Xanthine L-histidine Chloroquin mpmoles Inhibitor Required for 60% Inhibition 6 10 15 15 40 55 60 65 75 225 1000 50 precipitin curves both with native DNA and with Sins-Heimer +x-174 single strand DNA. We were only able to see that it gives strong reactions but we did not have enough of the material to perform quantitative studies. Third, have you studied reactions between the DNA antibodies of a particular patient and his own DNA extracted from his cells? This might be quite interesting because I wonder if this is the homologous DNA.* DR. LEVINE:With the L.E. system we have never performed quantitative precipitin experiments. We have looked at the DNA from +x-174 bacteriophage. This DNA is in the single strand state as isolated from the purified phage. This DNA gave identical fixation with native and thermally treated DNAz1when tested with an SLE serum such as in figure 2. DR. TOMASI: Have you tried any synthetic derivatives of either DNA or RNA? DR. LEVINE: Yes, we have. The anti-DNA in serum G.B. is inhibited by polyadenylic acid. It was no more effective than monomeric adenylic acid in terms of the number of purine residues required for a given degree of inhibition. With the anti-DNA in LE sera directed to pentathymidylic acid, poly AT was not inhibitory. Apparently the inhibitor must have iininterrupted thymidine tracts. We have not performed inhibition studies with Kharana’s synthetic poly-T‘s.. What do you mean by single strand? I know that your meltDR. TOMASI: ing temperatures look like Doty’s melting temperatures. Is this coiled in a OHijmans and Klein (unpublished results) have recently had the opportunity to answer this question. Liver and spleen were obtained within 2 hours after death of a patient suffering from SLE whose serum contained antibodies precipitating with DNA. DNA was prepared from these organs by the detergent method of Zamenhof. Calf thymus DNA was prepared by a similar method. Figure 5 shows the precipitin curves (at pH 8.5) obtained when the serum of this patient reacted with autologous and heterologous DNA, before and after ultrasonic breakdown. The shape of the different curves in the region of antigen excess, quite different from those of figure 1, remains unexplained. In any case, there are no indications from this experiment that autologous DNA has a greater affinity for the antibody than heterologous DNA. 551 DNA ANTIBODIES .* = CALF L - THYMUS = CALF THYMUS ULTRASON. o = HUMAN SPLEEN (AUTOLOGOUS) A A L = HUMAN SPLEEN ULTRASON. 0 ; 10 15 D N A PHOSPHATE 20 Fig. 5.-Precipitation of LE serum B with DNA. single strand helix? Or do you think this is a random specific distribution in terms of tertiary structure? DR. LEVINE: That is a good question. The temperature profile of DNA obtained immunologically is identical to that obtained by Doty et aL2' using physical measurements. There is a 2-3" displacement to lower temperature when compared to the physical data obtained at ambient temperatures. The nature of the DNA along this temperature profile is thought to be ( 1 ) a double strand helix; ( 2 ) a state best described as double strand with single strand areas or bubbles; and ( 3 ) completely separated strands. The completely separated strands on quick chilling can coil over on itself, but all of the antigenic determinants are not masked in this state. Formaldehyde prevents the intramolecular bonding and increases the serologic response of the denatured DNA slightly.28 The DNA containing bubbles may be metastable. If you allow the separated DNA strands to incubate at conditions which favor recombination to double strands (renaturation), you lose C' fixing capacity. The similarity of denaturation and renaturation of DNA measured immunologically and physically are two of several criteria we use to identify the antibody as anti-DNA. DR. SELIGMANN:When you say 'lose immunologic activity," it is not a complete loss? DR.LEVME:Under the conditions of our C fixation assay, loss or gain of serologic activity is demonstrated by comparison of C' fixed by the completely separated strands ( a calibration curve) and the C' fixed by the DNA preparations subjected to various experimental conditions. The amount of denaturation is determined by the ratio or thermally denatured (loo" for 10 minutes ) to the experimentally treated sample required to give equivalent fixation in the region of antibody excess. 552 IMMUNOLOGIC ASPECTS OF RHEUMATOID ARTHRITIS AND SLE DR. HOLMAN:When you cool rapidly you get separate strands-do they remain separated? DR. LEVINE:Yes. We routinely separate the strands by boiling DNA at about 5 pg. per ml. at 0.15M salt concentratian and pH 7.5. The solution is rapidly chilled in an ice bath. The strands remain separated by this procedure. With DNA preparations of high guanine and cytosine content one may have to increase the temperature or decrease the salt concentration or both to get separate strands rather than double strand DNA with “bubbles.” If you use tetrathymidylic acid and it reacts with the antiDR. HEIMER: body like a hapten, as in this case, do hapten antibody reactions fix complement? I do not think a hapten antibody reaction in which a three-dimensional complex is not formed will fix complement. DR. LEVINE:In the precipitin analyses of Dr. Seligmann, if you use calf thymus DNA, and assume it is 6 million molecular weight, you would have a mole ratio in extreme antibody excess of about 50,000 antibodies per DNA molecule. Is this not out of proportion? DR. SELIGMANN: We have made the same calculation and have found in the extreme antibody excess zone a molar ratio of 3,000 to 5,000 with the different sera which have been studied and at the equivalence zone a molar ratio cf 1,200 to 2,500. I suppose you mean 5,000 to 1. Nevertheless, we were astcnished by this high ratio in these precipitin curves performed at a pH above 8. I remember that in Dr. Deicher’s quantitative precipitin analyses showing the recovery of DNA in the precipitin, the mole ratios were also very high. DR. DAMESHEK: May I intrude on a hematologic note in which DNA hypersensitivity appears to be the central feature. In a patient with many bouts of a peculiar hemorrhagic eruption, Levin and PinkuP demonstrated a hypersensitivity to DNA. More recently we studied a similar patient30 with the help of Dr. Levine. In our patient there was an extraordinary skin lesion which eventually resulted in disability, She would break out with enormous hemorrhagic welts; these were finally traced to an unusual DNA hypersensitivity. The lesions could be reproduced very quickly with tiny amounts of DNA of various types; there was no reaction to RNA, and DNase acting upon DNA stopped the sensitivity reaction. There were no humoral antibodies. Apparently, these were cell bound antibodies and the reactions occurred only on the arms and legs. We transplanted skin from the arms or legs to the back and there was no reaction there; but, on the other hand, skin from the back transplanted to the forearms did result in a reaction. The patient was treated with chloroquin with a remarkable response and she is now well. DR. BUTLER:The studiesZ4 I would like to discuss were carried out in the laboratories of Dr. Sam M. Beiser and Dr. Stuart W. Tanenbaum in the Department of Microbiology at Columbia University. Like other investigators, Dr. Beiser has been interested for the past decade in making antibodies to DNA experimentally and to date he also has been unsuccessful. 553 DNA ANTIBODIES Before I arrived at Columbia, Dr. Beiser along with Drs. Erlanger and Tanenbaum had tried to elicit antibodies to certain purines and pyrimidines but without much success in terms of specificity. During the past year Dr. Sasson Cohen and Dr. Aaron Bendich at the Sloan-Kettering Institute synthesized a purine which has proved quite useful in our work. The compound which they used was 6-trichloromethyl purine. This compound is very readily synthesized from &methyl purine by a straightforward chemical reaction. In studying this compound in glycine HCl buffers, Dr. Cohen noted that it reacted very readily with amino groups under slightly alkaline conditions and at room temperature as shown below: c1 I c1 -c 4 1 OH wi C-I#-R When R is an amino acid, the product in the above reaction is a purinoyl amino acid. In other words there is now a carboxyl group attached to the 6-carbon of the purine ring and linked by a peptide-like bond to the amino group of the amino acid. It became apparent that this compound might serve our purposes quite nicely in terms of conjugating it to a protein carrier. And it turns out that in bovine serum albumin (BSA), there are about 59 available epsilon-amino residues of lysine to which it might be possible to conjugate the 6-trichloromethyl purine. Accordingly, we added this compound to BSA under slightly alkaline conditions at room temperature and, without going into the chemistry of it, we were able to get on about 25 residues of our purinoyl compound per molecule of BSA. The conjugate with human serum albumin (HSA) was also prepared in a similar manner. The reason we were not satisfied with using a compound which will be linked to protein in the 6-position is because in DNA the purine is linked to the sugar by the &position. However, we felt, since we had a conjugate with which we could work, that if we could elicit antibodies to this hapten, this would be encouraging for further studies whereby we might investigate compounds that are linked at the 9-position. The purinoyl-BSA conjugate in Freund's adjuvant was injected into rabbits in three weekly injections. All animals injected developed large amounts of antibody which precipitated both with purinoyl-BSA and purinoyl-HSA. Surprisingly little of this antibody precipitated either with BSA or HSA alone which made us believe that the antibody being elicited had great specificity 554 IMMUNOLOGIC ASPECTS OF RHEUMATOID ARTHHITIS AND SLE for the purinoyl residues. This was confirmed by appropriate hapten inhibition studies in which amino acids were quite effective as inhibitors of this precipitin reaction. We then studied these antisera with DNA from various sources, both native and denatured. In precipitin reactions we have been totally unsuccessful in detecting any precipitation whatever with any of the DNA preparations studied to date. However, with Dr. Levine’s complement fixation technic, on addition of thermally denatured pneumococcic DNA or denatured DNA from a number of bacterial sources to anti-purinoyl sera, we observed complement fixation. Furthermore upon the addition of unheated single strand +x-174 DNA we got complement fixation. For example, in a typical antiserum we observed the fixation of complement by a 1:400 serum dilution with denatured pneumococcic DNA. There was essentially no reaction with native pneumococcic DNA at a 1:400 dilution of this serum. With this particular antiserum we did get slight amounts of fixation with native preparations of pneumococcic DNA at a dilution of 1:lOO. It required 4 pg. of native DNA to get about 70 per cent fixation of complement, whereas, with the 1:100 dilution of this particular antiserum, we got similar fixation of complement with 0.03 pg. of the denatured DNA. To date we have detected no reaction with DNA either by precipitation or by complement fixation reactions. It may be that the molecular weight of the RNA is too small to evince fixation in this system. We have tried various methods of “denaturation” of the RNA according to some protocols suggested by Dr. Bendich, but as yet we have detected no complement fixation whatever. I might also mention that this complement fixation reaction is inhibited quite specifically by our purinoyl amino acids. We found essentially 100 per cent inhibition of a typical complement fixation reaction by less than 1/100 of a pmole of purinoyl epsilon-amino caproic acid which is structurally the closest derivative to purinoyl epsilon-amino lysine that we have available at the current time. We get less striking inhibition by adenine and other purines and very poor inhibition by pyrimidines. L.E. preparations have not yet been performed on these antisera. We have actually been interested in these antisera for purposes other than the study of lupus erythematosus. Globulin fractions from these antisera have already been shown to inhibit DNA-dependent bacterial transformation in a pneumococci. Drs. Galis and Beiser have carried out this study. Drs. Bendich, Rosenkrantz, Borenfreund and others are studying the possible effects of these antibodies on various viral and tumor DNA preparations. We have also instituted some immunohistochemical studies. Obviously, the extension of this type of work is to conjugate better haptens to proteins, compounds that have more resemblance to the structures in DNA. This is not going to be easy; however, it is a highly promising area for further investigation. DR. SELIGMANN: You said that you had no complement fixation with RNA. Did you look for inhibition? DNA AN'IIBODIES 553 DR. BUTLER:We have not completed these studies but we have not detected any inhibition so far. DR. SELIGMANN:Did you refer to studies with fluorescent antibodies or nuclei? And would you comment on the pyrimidine conjugates with which you are now studying? DR. BUTLER: Dr. Pappas has been conducting some preliminary studies and has observed some fluorescent antinuclear staining but he is not yet prepared to comment on its specificity and significance. Dr. Tanenbaum has prepared another conjugate, a pyrimidine-protein conjugate. Antisera to this conjugate also react with denatured DNA. (These antisera are currently being studied by Drs. Tanenbaum and Beiser.) DR. TOMASI: How do you know that your site of attachment is a protein? Have you actually evidence it is attached to an amino group? Could it not attach to a serine hydroxyl group? DR. BUTLER: Dr. Cohen does not feel that this particular reaction occurs. He does feel that there is some possibility that it may react with other groups but he thinks that the reaction with amino groups is by far the most important. Also, the fact that the purinoyl epsilon-amino caproic acid is the most effective inhibitor in both the precipitation and complement fixation systems would indicate that this is the predominant means of attachment. Spectrophotometric data also point to purinoyl-amino residues in the conjugate. DR.BENACERRAF: Do you know how many groups you had on your antigens that were so antigenic? DR. BUTLER: Both by estimation from spectrophotometric measurements and by actual alkaline digestion of the conjugate, we have determined that both the BSA and HSA conjugates contain about 25 purinoyl residues per protein molecule. We do Folin reactions which I have to convert each time and it strikes me that we have got probably 800 pg, of antibody nitrogen per ml. and most of anti-purinoyl; less than 5 per cent in most antisera is anti-protein. It is quite unusual in that respect. With some of the other haptens studied in our laboratory we had not observed this high a percentage of antibody directed against our hapten. DR. HEIMJTR: I think the answer to Dr. Tomasi's question is that the conjugation occurs in alkali and in esters. An ester would be formed in alkaline wnditions and esters are not stable in alkali. That is probably the reason why amides could be stable under those conditions but esters could not. I am curious why you did not get a lysine derivative. DR. LOSPALLUTO: DR. BUTLER: Six-trichloromethyl purine reacts with both amino groups of lysine. We have tried to block the alpha-amino group in an attempt t o prepare the epsilon-amino derivative. We have not yet been successful, but expect to synthesize this compound in due time. The available hapten which most closely resembles purinoyl epsilon-amino lysine is purinoyl epsilonamino caproic acid. This derivative can be readily synthesized and is an extremely effective inhibitor of anti-purinoyl sera. W e have tried copper chelate but this did not work. It should be remem- 556 IMMUNOLOGIC ASPECTS OF RHEUMATOID ARTHRITIS AND SLE bered that copper has some peculiar interactions with purines, so that failure to obtain the epsilon-purinoyl derivative from the copper chelate is not too surprising. DR. SELIGMANN:Does anyone have new data or ideas about the eventual pathogenicity of antinuclear antibodies in lupus patients? DR. KAPLAN: I can report a negative experiment. Just as you might- localize rheumatoid factor in the tissues by using a labeled reactant, we have labeled isolated beef thymus-nucleoprotein-in various ratios with fluorescein-hoping that this product might be useful for the localization of at least antinucleoproteinfactors in the tissues. It turns out that at all ratios of labeling fluorescein to nucleoprotein, the nucleoprotein loses its reactivity completely with L.E. sera. So, whatever the reason, the resulting labeled material is no longer reactive with L.E. sera. REFERENCES 1. Seligmann, M.: La sp&&cit6 des rkactions sbrologiques du lupus Brythbmateux disskmink et les anticorps antiacide dhoxyribonucleique. Immunopathology, 1st Internat. Symposium, Seelisberg, 1959. Basle, Benno Schwabe, p. 402. 2. -: Mise en kvidence dans le sQum de malades atteints de lupus krythbmatuex disskmink d'une substance d& terminant une &action de prbcipitation avec l'acide d8soxyribonuclkique. Compt. rend. Acad. sc. 245:243, 1957. 3. Deicher, H. R. G., Holman, H. R., and Kunkel, H. G.: The precipitin reaction between DNA and a serum factor in systemic lupus erythematosus. J. Exper. Med. 109:97, 1959. 4. Robbins, W.C., Holman, R. H., Deicher, H. R. G., and Kunkel, H. G.: Complement fixation with cell nuclei and DNA in lupus erythematosus. Proc. Soc. Exper. Biol. & Med. 96:575, 1957. 5. Ceppellini, R., Polli, E., and Celada, F.: A DNA-reacting factor in serum of a patient with lupus erythematosus diffusus. Proc. SOC. Exper. Biol. & Med. 96:572, 1957. 6. Seligmann, M., and Milgrom, F.: Mise en kvidence, par la fixation du complhment, de la rhction entre acide dksoxyribonucl6ique et Ie skrum de malades atteints de lupus brythbmateux disskmink. Compt. rend. Acad. sc. 245:1472, 1957. 7. Pearson, C. M., Craddock, C. G., and Simmons, N. S.: Complement fixation reactions with DNA and leukocyte material in systemic lupus erythematosus. J. Lab. & Clin. Med. 52:580, 1958. 8. Asherson, G. L.: Antibodies against nuclear and cytoplasmic cell constituents in systemic lupus erythematosus and other diseases. Brit. J. Exper. Path. 40:209, 1959. 9. Hijmans, W.,and Schuit, H. R. E.: Studies on the LE cell phenomenon 111. Complement fixation with nuclear substances and the LE factor. Vox. Sang. 4:376, 1959. 10. Deicher, H. R. G., Holman, H. R., Kunkel, H. G., and Ovary, Z.: Passive cutaneous anaphylaxis reactions with a systemic lupus erythematosus serum factor and isolated DNA. J. Immunol. 84: 106, 1960. 11. Seligmann, M.: Etudes immunologiques sur le lupus krythbmateux disskmink, Rev. franc. Etudes clin. biol. 3:558, 1958. 12. Barbu, E.,Seligmann, M., and Joly, M.: Rhactions entre des acides dksoxyribonuclkiques diversement dhaturbs ou degrades et les anticorps anti-acide d6oxyribonuclbique du sbnun de malades atteints de lupus 6rythkmateux disskmink. Ann. Inst. Pasteur 99: 695, 1960. 13. Murakami, W. T.,Van Vunakis, H., Grossman, L., and Levine, L.: Immunochemical studies of bacteriophage deoxyribonucleic acid. 11. Characterization of the active antigen. Virology 14:190, 1961. 557 DNA ANTIBODIES 14. Bozicevich, J., Nasou, J. P., and Kayhoe, D. E.: Deoxyribonucleic acid (DNA) bentonite flocculation test for lupus erythematosus. Proc. SOC. Exper. Biol. & Med. 103:636, 1960. 15. Levine, L., Murakami, W. T., Van Vunakis, H., and Grossman, L.: Specific antibodies to thermally denatured deoxyribonucleic acid of phage T4. Proc. Nat. Acad. Sci. 46:1038, 1960. 16. Murakami, W. T., Van Vunakis, H., Lehrer, H. I., and Levine, L.: Immunochemical studies on bacteriophage deoxyribonucleic acid. III. Specificity of the antibodies. J. Immunol. 89: 116, 1962. 17. Townsend, E., Murakami, W. T., and Van Vunakis, H.: The antigenic determinant groups of bacteriophage DNA. Fed. Proc. 20:438, 1961. 18. Levine, L.: Determinants of specificity of proteins, nucleic acids and polypeptides. Fed. Proc. 21:711, 1962. 19. Stollar, D., and Levine, L.: Antibodies to denatured DNA in a lupus erythematosus serum. J. Immunol. 87: 477, 1961. 20, Wasserman, E., and Levine, L.: Quantitative micro-complement fixation and its use in the study of antigenic structure by specific antigen-antibody inhibition. ]. Immunol. 87:290, 1961. 21. Stollar, D., Levine, L., and Marmur, J.: Antibodies to denatured DNA in lupus erythematosus serum. II. Characterization of antibodies in several sera. Biochim. et biophys. acta 81:7, 1962. 22. -, -, Lehrer, H. I., and Van Vunakis, H.: The antigenic determinants of denatured DNA reactive with lupus erythematosus serum. Proc. Nat. Acad. Sci. 48:874, 1962. and -.: Antibodies to denatured DNA in lupus erythematosus serum, IV. Evidence for purine determinants in DNA. Arch. Biochem. 101:417, 1963. Butler, V., Beiser, S., Erlanger, B., Tanenbaum, s., Cohen, s., and Bendich, A. : Purine-specific antibodies which react with DNA. Proc. Nat. Acad. Sci. 48:1957, 1962. Stollar, D., and Levine, L.: Antibodies to denatured DNA in lupus erythematosus serum. V. Mechanism of DNAanti-DNA inhibition by chloroquin. Arch. Biochem. 101:335, 1963. Holman, H., and Deicher, H. R. G.: The reaction of the LE cell factor with deoxyribonucleoproteins of the cell nucleus. J. Clin. Invest. 38:2059, 1959. Doty, P., Marmur, J., Eigner, J., and Schildkraut, C.: Strand separation and specific recombination in DNA, physical chemical studies. Proc. Nat. Acad. Sci. 46:461, 1960. Stollar, D., and Grossman, L.: The reaction of formaldehyde with denatured DNA: Spectrophotometric, immunologic and enzymic studies. J. Molecular Biol. 4:31, 1982. Levin, M. B., and Pinkus, H.: Autosensitivity to deoxyribonucleic acid (DNA): report of case with inflammatory skin lesions controlled by chloroquine. New England J. Med. 264:533, 1961. Schwartz, R., Lewis, F. B., and Dameshek, W.: Hemorrhagic cutaneous anaphylaxis due to autosensitization to deoxyribonucleic acid. New England J. Med. 267:1105, 1962. 23. -, 24. 26. 26. 27. 28. 29. 30.