Serologic Specificity of Antibodies to Ribonucleic Acid in Normal and Rheumatoid Arthritis Sera

Home , Cytidine monophosphate, Thymidine monophosphate, Thymine

This dissertation has been microfilmed exactly as received 69-4926

LARKIN, Gary Freeman, 1941- SEROLOGIC SPECIFICITY OF ANTIBODIES TO RIBONUCLEIC ACID IN NORMAL AND RHEUMATOID ARTHRITIS SERA.

The Ohio State University, Ph.D., 1968 Health Sciences, immunology

University Microfilms, Inc., Ann Arbor, Michigan SEROLOGIC SPECIFICITY OF ANTIBODIES TO RIBONUCLEIC ACID IN NORMAL AND RHEUMATOID ARTHRITIS SERA

DISSERTATION Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

By Gary Freeman Larkin, B»S<,, MoS„

The Ohio State University 1968

Approved by

Adviser Microbiology ACKNOWLEDGMENTS

The writer wishes to express his sincere appreciation and thanks to Dr. Matthew C. Dodd whose sound advice and encouragement were invaluable during the course of graduate studies at this Uni v e rsity . The writer is also deeply grateful to Dr. Nancy J. Bigley for her enthusiasm, interest, and most valuable criticism during the present investigation. Appreciation is also extended to James C, Darnel’ for his sug gestions and interest in this study. VITA

July 16, 19;+1 Bom — Massillon, Ohio

1963• B.S., The Ohio State U niversity, Columbus, Ohio 196^ 1966.... Teaching A ssistant, Department, of Microbiology, The Ohio State U niversity, Columbus, Ohio

1966...c.e.o. M.S., The Ohio State U niversity, Columbus, Ohio 1966-196?.... Graduate Research Assistant, Department of Micro biology, The Ohio State U niversity, Columbus, Ohio 1967-1968.... Graduate Research Associate, Department of Micro biology, The Ohio State U niversity, Columbus, Ohio

FIELDS OF STUDY

Major Field: Microbiology Studies in Immunology. Professor Matthew C, Dodd, Profes sor Nancy J. Bigley, Professor Frank V/, Chorpenning Studies in Pathogenic Microbiology. Professor Melvin S. Rheins Studies in Bacterial Physiology. Professor Chester I. Randles , Studies in Food Microbiology. Professor Harry H. Weiser

Studies in Virology. Professor David A. Wolff

i i i CONTENTS

Page

INTRODUCTION...... ___ ...... 1

REVIEW OF THE LITERATURE...... 3

MATERIALS AND METHODS...... « . « . . . 2 3

RESULTS OF EXPERIMENTATION...... 37

DISCUdSION...... 6oe..o.ffoe..o.....0oe*effeooee.ee«eoo..o. 72

SUMMARY . . . .o. o. . .9 0 . oe...... «.o.o. . 87

LITERATURE CITED...... o o o ......

xv TABLES

Table Page 1. Percentage of complement fix atio n in h ibition of anti-RNA a c tiv ity in normal human serum by- purine and pyrimidine bases...... 39 2. Percentage of complement fix atio n in h ib itio n of anti-RNA a c tiv ity in normal human serum by purine and pyrimidine ribonucleosides...... 39 3. Percentage of complement fix atio n inh ib itio n of anti-RNA a c tiv ity in normal human serum by purine and pyrimidine ribonucleotides»...... 40 4. Percentage of complement fix atio n in h ibition of anti-RNA a c tiv ity in normal human serum by adenine-containing compounds...... • . . . •». 41 5. Percentage of complement fix atio n inh ib itio n of • anti-RNA a c tiv ity in normal human serum by guanine-containing compounds...... 41 6. Percentage of complement fix atio n in h ib itio n of anti-RNA a c tiv ity in normal human serum by cytosine-containing compounds...... 42 7. Percentage of complement fix atio n inh ib itio n of anti-RNA a c tiv ity in normal human serum by thymine-containing compounds...... 43 8. Percentage of complement fix atio n inh ib itio n of anti-RNA a c tiv ity in normal human serum by uracil-containing compounds ...... 43 9. Results of whole serum electrophoresis of rheumatoid arthritis sera and normal sera...... 46 10. Conglutination reactivity of rheumatoid arthritis sera and normal sera ...... 48

v Table Page

11. Passive hemagglutination reactivity of rheumatoid arthritis sera with nucleic acids and their derivatives...... p . 31 15. Results of passive hemagglutination absorptions with various antigens upon rheumatoid az tin x tis seia...... oooo...... 5^*

13. Results of passive hemagglutination absorptions with base pairs upon rheumatoid a r th r itis s e ia ...... 63

v i INTRODUCTION

Complement fixation techniques as well as precipitation reactions have demonstrated that antibodies exist in serum with specificities directed toward cellular as well as subcellular portions of tissues and organs. Furthermore, different specificities can be shown in comparisons between normal and malignant cells and subcellular components. With the discovery of, and subsequent investigations into, the area of nucleic acid antigenicity, work in this laboratory and others has been carried out in order to determine the exact im munological. specificity of these-antigenic substances. The work of Friou, Finch and Detre (30) and the subsequent identification of the lupus erythematosis factor as a gamma globulin with deoxyribonucleoprotein specificity was a major achievement in the increased body of knowledge regarding the immunologically associated autoimmune disease processes. Previous work in this laboratory has shown that ribonucleic acid (RNA) linked to bentonite particles can detect antibodies to RNA in sera of both diseased and normal individuals. Furthermore, there are differences in the specificity of the antibody response, to this antigen. It is the purpose of 'Hie present study to further elucidate these differences and to study in greater depth the iimn.unocheiin.cal relationship between the antigen and its specific antibody.

1 By utilizing the quantitative micro-comploment fixation and micro-complement fixation inhibition techniques of Wasserman and Levin© (92), normal serum anti-RNA antibody sp e c ific ity was studied. Since rheumatoid arthritis sera were anti-complementary, attempts were made to identify these factors. Passive hemagglutination titrations and absorptions were performed upon rheumatoid arthritis sera using nucleic acids, nucleotides, nucleosides and bases as anti gens. Base pair absorptions were also carried out. Direct comparisons between the bentonite flocculation technique, the passive hemagglu tin a tio n technique and the complement fix atio n technique were made. By using various nucleotides, nucleosides, and purine and pyrimidine bases, the immunologic specificity of such antibodies was also com pared. The nature of the antigenic determinant was investigated. REVIEW OF THE LITERATURE

Previous reports have indicated a wide range of immunolog- . ically active substances present in tissue and organ extracts. Duran-Reynals (27) reported that sera from chickens and other fowls flocculated crude salin e extracts of huiuan tissu e s. Using complement fixation, Kidd and Friedewald (45) demonstrated the reactivity of normal rabbit serum with saline extracts of rabbit kidney,.liver, lung, brain, spleen, and heart tissue. Further investigations have shorn th a t animals immunised with wholo brain tissu e react with alco- . holic extracts of brain (.96), both of homologous and heterologous origin. Thyroid tissue appears to contain both organ and species specific factors. Isoantibodies may be produced in the rabbit causing changes in the organ of the immunised animal (69, 97). Sim ilarly, many patients with chronic thyroiditis possess antibody reactivity toward the microsomal fraction of the extracted gland (84). Furthermore, lens (37), uveal (22), and pancreatic tissue (98) can serve as auto antigens as well as iso- and hotero-genetic antigens. Gajdusek (34) used complement fixing antibody techniques to demonstrate autoanti bodies to human tissue antigens in the sera of patients known to have acute or chronic diseases as well as for the detection of autoantibody to normal human tissue (33)» In addition to normal tissue and organ antigens, there are many specific antigens which are produced as the

3 kr result of experimentally induced tumors (55) and naturally occurring tumors (85t 99) • Wiederman (95) has shown that antibodies with specifieity directed toward the cytoplasmic constituents of the liver, namely the mitochondria, lysosomos, microsomos and the soluble fractio n appear in sera of patients with lupus erythematosis and infectious hepatitis. Asherson (1) demonstrated furthermore complement fixation reactions with the cytoplasmic fraction of cells as well as with nuclear compo nents in the sera of patients with systemic lupus erythematosis, lupoid hepatitis, syphilis, and macroglobulinemia. Deicher, eh al. (2h) demonstrated that sera of patients with SLE and other diseases when mixed with subeellular fractions (human liv e r, kidney or calf liv e r, and r a t liv e r microsomes and mitochondria) bound complement. Treatment of these preparations with DNase, RNase, or trypsin did not change the titer of the complement fixation reac tions indicating that nucleic acids were probably not the cause of antibody production. As a logical consequence of these studies, Macks.y and Gajdusek (52) stated that cell proteins, particulate compo nents, polysaccharides, and nucleoproteins could become antigenic and act as autoantigens if they were released from the cell and came into contact with cells of the retieulo-endothelial system. When, in 19h8, Hargraves, Richmond and Morton (hO) discovered the LE serum factor which is responsible for LE cell formation and when Haserick, et al. (hi) determined it to be a gamma globulin, interest was turned toward discovering the specificity of the LE fac tor in serum. Friou, Finch, and Detre (30) showed that this serum facto r combined with iso lated n u c le i, and Friou (31) proved th a t the antibody would react with deoxyribonucleoprotein of both natural and synthetic origins. Further investigation of lupus erythematosis serum has re sulted in a great Increase of information regarding the antigenic properties of nucleoproteins (?1, 72, 90). Antibodies in the serum of a p a tien t with lupus erythematosis fixed complement and precipi tated the DNA from normal human leukocytes, lymphatic leukemia cells, and myeloid leukemia cells (20). Absorption of this sera with RNA had no effect upon reactivity. Seligmn (?8) reported the presence of DNA antibody in some LE sera which were detected by liquid or gel precipi tation, passive cutaneous anaphylaxis, or by complement fixation. Five per cent of rheumatoid arthritis sera were also positive for anti- DNA (79)o Robbins,- e t a l . (61) showed positive complement fixation with cell nuclei and D.upus erythematosis sera, and Holman (^3) has shown an affinity between the LE factor and cell nucloi and nucleoprotein, concluding that the LE factor could act as an autoantibody.

Barnett (7) has similarly reported the presence of an antinuelear fac tor in rheumatoid arthritis. Musehel, et al. (5*0 have demonstrated that the anti-DNA reactivity in lupus erythematosis is not species specific and postulated either a bacterial or viral nucleoprotein as the antigenic stimulus. The exact role of antibodies in autoimmune diseases is highly disputed. Seligman (80) reported that studies of LE sera before thor- apy was initiated showed an average in excess of eighty per cent positive for antinuelear antibodies. After therapy was completed, 6 the number of positive sera correlated well with the degree of re covery. In cases of complete clinical recovery, the incidence of positive reactivity is well below the ten per cent level. Levi and Poverenny (50), using passive hemagglutination and hemagglutination inhibition reactions characterized the specific anti bodies of SLE patients against both denatured and native DNA. They hypothesized that antibodies are formed first at the outset of the disease and are directed toward denatured DNA. Later, antibodies are formed against native DNA, depending upon the severity of the autoim mune process. These authors believe that the presence of DNA and DNA antibodies alternated in the serum.

Tan, Sehur, Carr and Kunlcel (89) studied the possible mecha nisms of renal 3.esion production in systemic lupus erythematosis. Using gel diffusion, they demonstrated not only DNA antibodies, but also native DNA free in the serum. In addition, it was shown using serial serum samples that there was an inverse correlation between the presence of native DNA in the serum and the presence of DNA antibodies. If sera with free DNA were reacted with sera containing anti-DM, complement fix atio n and p re c ip itin reactions were po sitiv e. This sug gests a possible mechanism of in vivo antigen-antibody complexes form ing and causing tissu e damage. Antibodies with specificities directed towards ribosomes and their subunits have also been discovered in LE sera. Sturgill and Carpenter (87) and Sturgill and Preble (88) reported that fifty-one per cent of forty-one patients with SLE had antibodies to ribosomes by using an indirect immunofluorescence spot test. Precipitation in agar 7 would occur if the titer was greater than56 2 as indicated by the spot test. The antigen was sensitive to both RNase and trypsin and mi grated toward the cathode in an electric field, Schur, Moroz, and Kunkel (74) described antibodies to a ribo- sornl antigen in 11 of 85 SLE patients by direct precipitation and in 35 of the 85 patients studied by using the fluorescent spot test. Those patients having precipitating antiribosomal antibodies all had renal implications. Of the remaining 74 SLE patients not demonstrat ing the precipitation reaction, renal complications were present in only half. These authors postulated that ribosomes may gain access to the circulation where they combine with circulating antibodies and be come deposited in the glomeruli of the kidneys. Treatment of the ribo somes with trypsin or RNase destroyed the antigenic reactivity, im plying that the antibody response is directed toward the intact ribosomal p a rtic le with i t s complex nucleic acid and protein in teractio n s. Barnett, Bienenstock, and Block (11) have demonstrated anti- nuclear factors of immunoglobulin origin in the synovia of patients with rheumatoid arthritis. Sera contained anti-nuclear factor reac tivity in most instances, but in some patients, only the synovia demonstrated re a c tiv ity . Hemolytic complement levels as w ell as the reactivity of C'2 were greatly reduced. It was proposed that the reaction of the antinuelear factors in the synovia with nuclear anti gens may contribute to the pathogenesis of the disease. Recently, Friou (32) has discussed the diagnostic significance of anti-DNA and anti-deoxyribonucleoprotein antibodies in systemic disorders. Anti-DNA when present in significantly high titers is of 8 great significance for clinical diagnosis whereas anti-deoxyribonucleoprotein is never completely diagnostic alone. If both anti bodies are present at high titers, indications of a systemic involve ment are most strongly favored. Beck, et al, (12) described anti-nuclear precipitating auto antibodies to a wide variety of organs and tissues in Sjorgren's Syn drome which is characterized, by keratoconjunctivitis sicca, salivary gland involvement, and rheumatoid arthritis. In addition to tho antinuelear and tissue antibodies, the rheumatoid factor is present as well as thyroglobin autoantibodies. The abovo mentioned investigations have shown the relationship between the presence of an autoimmune phenomenon and the resulting autoantibodies to subcellular organelles including ribosomes and nuclei. Although Lacom’, Harel and Harel (^7) have successfully pro duced antisera to mammalian ribosomes, recent work by Dodd, Bigley and co-workers ( 25, 26) have shown the development of an autoimmune proc ess in experimental animals. The antiserum from these animals immu nized >rith rat liver ribosomes contained antibodies to ribosomes, yeast soluble RNA, and ribosomal protein. Ribonucleic acid in adju vant produced practically all the same abnormal aspects of the ribo some immunized animals. In another study, Bigley’, ..jot al, (15) havo shown th a t rabbits injected with r a t or homologous liv e r ribosomes in Freund’s adjuvant developed hemolytic anemia, leukopenia and multi- organ pathology. Normal rabbits also demonstrated the same autoim mune syndrome when passively immunized with anti-ribosorcal sera from an immunized rab b it. Furthermore, i t was shown th a t nuclei, soluble 9 RNA, and. ribosomal protein in Freund’s adjuvant would cause production of specific antibodies with definite nucleotide and nucleoside speci ficity as determined by passive hemagglutination and hemagglutination inhibition tests. Anti-ribosome antibodies were completely inhibited by 500 micrograms of guanosine, cytidine, and uridine, and partially inhibited by adenosine. Antibody activity directed toward soluble ENA was inhibited by a ll four nucleotides and nucleosides present in ENA. The anti-nuclear antibodies were partially inhibited by guanosine mono phosphate and guanosine and completely inhibited by thymine monophos phate, cytosine monophosphate, adenosine and cytidine. Because of the autoimmune nature of the experimentally induced disease, other workers have investigated sera of patients with rheuma toid arthritis for the presence of anti-nucleic acid antibodies. Rheumatoid a r th r itis is a disease suspected of an autoimmune origin. There is evidence that the immune response relates to the cause of the disease (21). • There are definite clinical and patholog ical analogies between serum sickness and the connective tissue syn dromes, which include rheumatoid arthritis, since the host responds to foreign serum proteins in a way similar to the response in rheumatoid arthritis, rheumatic fever and systemic lupus erythematosis. Using serologic techniques, studies have shown th a t autoimmune processes are active in rheumatoid, arthritis since antinuelear factors, positive LE cell formation and the presence of the rheumatoid factor are all found in a majority of cases.

The sedimentation constant of tho rheumatoid factor is 19S, but the f a c to r. is combined with the low molecular weight gamma 10 globulin in sera which contains very large amounts of the globulin„ forming a 22S complex. In addition to the 19S factor, there is also recent evidence for a7S rheumatoid factor which reacts with other 7S gamma globulins. Prolonged immunization of experimental animals with killed Escherichia coll, Bacillus subtllls, and Salmonella typhimurium causes production of substances which mimic rheumatoid factors. The sero lo g ical sp e c ific ity is detectable in the form of an immune complex or in an aggregated form. It appears that prolonged immunization forms an in vivo immune complex, and the eventual breakdown of these com plexes is the stimulation necessary for the production of autoanti bodies. This experimental work indicates that perhaps there is a con stant antigenic stimulation which is responsible for the disease syn drome . There are two other pieces of clinical information which can be cited as possible iiumunologically oriented implications of rheumatoid arthritis. First, a common feature of rheumatoid arthritis is a hypergammaglobulinemia and secondly, the synovial membranes and joint structures which are affected contain plasmacytes and nodular arrange ments of lymphocytes which resemble the primary structural feature of lymphoid tissues, or the lymphoid follicle. It has been found by Hollander (4-2) that cytoplasmic inclusions of gamma globulins exist in the polymorphonuclear leukocytes in the synovial fluid. A possibility exists that these gamma globulins are being actively phagocytized by the polymorphonuclear leukocytes causing the release of leukocytic 11 lysosomal enzymes which may, in turn, cause further destruction of the synovial tissues (2). S im ilarly, complement lev els have been studied in the sera of patients having diseases of suspected immunological origin23). ( Dermatomyositis, anaphylactoid purpura, scleroderma, polyarteritis nodosa and rheumatoid arthritis are associated with normal or elevated complement t i t e r s while SLE with renal involvement shows a decreased complement le v e l. However, Cooper and Fogel (23) point out th a t not a l l immune complexes fix complement and th a t normal or elevated com plement levels do not completely rule out the possibility of an auto immune etiology for the disease procoss. In a similar manner, de creased complement levels are not always indicative of in vivo com plement fixation by immune complexes and a relationship to immuno logical processes. Recently, Bienenstock and Block (14) studied the levels of iimiunoconglutinin in the sera of patients with rheumatic and autoim mune diseases. Imimraoconglutrurin tit e r s of 16 or greater were found in 51 per cent of patients with rheumatic diseases and 71 in per cent of patients with SLE, myasthenia gravis, Hashimoto's thyroiditis, and Sjorgren's syndrome. Since immunoeonglutinin can combine with the bound Cf3 component of complement and prevent fu rth er fix atio n of complement components, these authors postulate th a t immune con glut in -jn could lim it the circu latio n of p o ten tially harmful activated comple ment components. However, another im plication arises since immuno- conglutinin is a gamma M antibody and could bind complement by i ts e l f . They postulate that this combination could provide more C!4,2a sites 1 2 causing the activation of more C*3 molecules, and the further activa tio n of complement would occur. This could cause fu rth er damage in an autoimmune sta te . Scheetz (81) utilized the bentonite flocculation technique of Bozieevich (16) to investigate the occurrence and immunologic specific ity of antibodies to DNA. He demonstrated that purine bases did not inhibit DNA-bentonite flocculating reactivities in lupus erythema tosis or normal sera, but did havo the capacity for inhibiting DNA reactivities present in rheumatoid arthritis sera. The pyrimidine bases inhibited only a small percentage of antibodies in all three serum types while the deoxy-counterparts, the deojypurines and deoxy- pyrimidines, were shown to inhibit the DNA flocculation in rheumatoid sera in a significantly higher percentage than in the sera of normal individuals. Similar work (48) was carried out using the bentonite floc culation method with RNA as the antigen. A heat stable factor cap able of flocculating RNA-bentonite was found in the sera of some in dividuals with rheumatoid arthritis, systemic lupus erythematosis, and miscellaneous collagen diseases. A heat Labile factor was noted in the sera of all normal individuals which was absent in the collagen disease sera. The disease sera were shown to havo unique serologic specificities by nucleotide, nucleoside and base inhibition studies while normal unheated sera displayed a much broader range of speci f ic ity . Feldbush (29) extended the above mentioned work by investi gating the immunochemical relatio nship between c a rrie rs, red blood 13 cells in the passive hemagglutination reaction, and bentonite in the bentonite flocculation technique. There were two different antibodies detected by the hemagglutination reaction. One was characterized as heat labile, predominately 19S and B ^, which, was inhibited by DNA, The othex- was determined to be heat stable, 7S and 19S globulins which were inhibited completely only by RNA or its nucleotides, Thi.s work correlates well with the investigations carried out by Barnett and co-workers (6, 8, 9) in which all three major classes of immunoglobu lins were found to contain anti-nuclear reactivity in rheumatoid arth ritis sera and lupus erythematosis sera. Feldbush also showed the occurrence of a reversible antigen--antibody complex when either RNA or the low molecular weight hapten inhibitors were used, as antigens. It is interesting to note that a difference exists in the de tection of antibodies to PINA in normal individuals using passive hemag glutination (29) and bentonite flocculation (AS), Only 29 per cent of normals were weakly positive by passive hemagglutination reactions while 100 per cent were positive by the bentonite flocculation test.

Recently, Poverenny and co-workers (66,67 ), using the passive hemag glutination reaction found 60 per cent of all normals to react with DNA. Three groups of antibodies were demonstrated. The first wore those reacting with formalin denatured DNA. The next group demon strated activity with heat denatured DNA, and the last was a group with a broad range of serological activity including heat denatured, formalin treated, and native DNA. It is this last type of antibody which can inhibit the activity of DNA as a primer for RNA polymerase reactivity and can therefore influence DNA translating ability. 14

Rosenkrantz, et al. (70) have demonstrated that antisera directed against purine and pyrimidine containing albumin complexes can pens- trate and become fixed in the cells of fertilized, eggs of Arbacia punctulata. The arrestment of development was dependent upon the dilution of antibody and its specificity. The interaction of antibody is postulated to be with the single-stranded DNA in the cell. P anijol (57) has recently shown th a t there are anti-RNA a n ti bodies in the sera of germ-free animals. These antibodies will precip itate only synthetic poly I and poly U. Furthermore, the precipitin reaction is specifically inhibited by various RNAs and with either pod.y A or poly C. There is no cross-inhibition with DNA from any source. The authors propose that these antibodies are a result of either auto immunization or are a part of a normal clearing mechanism involving genetically controlled, normally synthesized globulins. Thivolet4 Monier, Ruel and Richard (91) discovered anti nuelear antibodies of autoimmune nature in Swiss mice who were thymec- tomiaed at birth. These antibodies appeared at 11 and 15 weeks after thymectomy. Since the thymectomy should suppross a l l immunological responses, a role of normally synthesized antibody as an active physiological clearing mechanism must be considered. Antinuelear antibodies wore also found in normal human sub

je c ts (19) by using complement fix atio n and inmmofluorescent tech niques. Ninety-seven per cent of all normal blood donors showed positiv e complement fix atio n with antigens of calf thymus origin. Antibodies with a weak reactivity for calf thymus nuclei were also 15 demonstrable with normal human serum by indirect immunofluorescence using monospecific anti-gamma G, gamma A, and gamma M. In a similar manner, normal rabbit sera has also been shown to contain demonstrable anti-nuclear reactivity (10). Since it was known that subcellular organelles were antigenic, investigative efforts wore turned toward nucleic acids and their pos sible role as antigens. An early report (9^0 stated that nucleic acids possessed no antigenic activity. Phillips, ot al. (60) demon strated precipitating antibody in the sera of rabbits immunised, with a DNA rich preparation of Brucella abortus. The substance had a strong

Fuelgen reaction which showed decreasing absorption Z60 at and 277 millimicrons when treated, with DNase. Further work by Plescia and co-workers (61) showed that the reactive complex was sensitive 'bo per iodate treatment also, implicating the presence of other antigenic determinants in addition to the nucleic acid moiety. The sensitivity of antigenic preparations to DNase in other studios (b8) has further demonstrated the antigenicity of DNA complexes. Levine (51) found a cross-reactivity with DNAs from tho T-even coliphages T2 to T6, but not to Escherichia coll or calf thymus DNA, indicating that some specificity may lie in the 5-hydroxymethyl cyto sine common to these phages. Murakami and co-workers (52) have extended this work with their experimental approach in which thermally dena tured phage DNA was used to uncover the antigenic determinants of the molecule. Antibodies were directed in part to the glueosylated 5-hydroxymsthyl cytosine of the phage DNA. Townsend, et al. (92) showed hapten inhibitions of DNA reactivity present in the sera of 16 rabbits immunized with bacteriophage DNA with diglycosyl-dHMP, cello- biose, gentiobiose, amygdaleno, alpha glycosyl-dHMP and maltose. Several approaches have been used by Beiser and Erlanger (13) for the production of DNA specific antibodies. Hypsrimniunization with Gram negative bacteria, immunization with purines or pyrimidines coupled to proteins or polypeptides, or immunization with denatured DNA or polynucleotides electrostatically coupled to methylated bovino serum albumin all caused production of nucleic acid specific anti bodies. Further work by these authors (18) has shown that bovine serum albumin conjugates containing the 6~purinoyl and 6-purinoyl-B-alanyl moieties caused production of antibodies which gave equivocal results when examined for purine or pyrimidine specificity. The effectiveness of various compounds as hapten inhibitors was investigated. It was found that those inhibitors which possessed an intact ring structure were far more effective as inhibitors than those compounds containing a ring structure which was open. Erlanger and Beiser (28) further investigated antibodies specific for ribonucleosides and ribonucleotides. They demonstrated th a t an ti-cy tid in e, anti-guanine, anti-AMP and anti-UMP react with thermally denatured DNA but not with native DNA. It was noted that AMP was a better inhibitor than either adenosine or adenine. Also, cytidine appeared to be better than either cytosine or cytidylic acid. It was concluded that the antibodies were specific for the entire mole cule, not just the purine or pyrimidine bases and that serologic differences do reside in the sugar-phosphate backbone of the nucleic acid structure, also. 17 Butler, Tanenbaum, Stewart, and Beiser (17) have demonstrated that purin~6~oyl protein conjugates form antibody with purine speci f ic ity which faeces complement in the presence of denatured DNA, Only purines and purine derivatives specifically inhibited the reaction. The 50 por cent endpoint in the complement fix atio n technique was never reached with the pure base only. Base derivatives were able to inhibit at the desired endpoint, however, Sela, et al. (83) have shown that antibodies with specificity toward uridine may bs obtained in rabbits by injection of synthetic molecules in which uridine--carboxylic acid is bound through an amide bond to the amino term inal groups of poly-DL-a.lan.yl side chains of a multichain synthetic polypeptide. The antigenic specificity of this compound is due to the nucleoside unit as a whole since uracil and ribose alone did not inhibit the reaction while uridine and uridylic acid did. Thymidine also caused inhibition which indicates that neither the methyl group at position 5 nor the hydroxyl group at position 2 of the riboso moiety is important in determining the combining site. There was no inhibition by purine nucleosides, implicating the im portant role of the pyrimidine ring in specificity. Further work by

Sela (82) has shown th a t the uridine-poly A complex possessed anti-DNA re a c tiv ity . Calf thymus DNA did not p recip itate the antbody, while heat denatured DNA did. There was an enhancement of precipitation when the DNA was heat denatured in the presence of formalin. Recently, Seaman, et al. (75) coupled methylated bovine serum albumin and T4 phage DNA. All of the antisera contained antibodies directed toward the alpha-glucosylated hydroxymethyl cytidilic acid of 18 the phage DNA. All the antibodies showed specificity toward T-oven phage DNA except in cases in which the phage underwent photooxi dation. These antibodies reacted with heterologous DNAs lacking the glucosylated. moiety. It was also noted that the removal of phos phate from the cytidilic acid decreased its effectiveness as a hapten in h ib ito r. Halloran and Parker (38) prepared micleotide-protein conju gates using various carbodiimides as the coupling agents. Mononuc leotides j, oligonucleotides, and DNA were coupled directly to both bovine gamma globulin and human serum albumin through the formation of a nitrogen-phosphate bond with protein epsilon amino groups or by the formation of a phospliodiester bond with protein seryl and threonyl residues. Experimental evidence indicates the formation of N-P bonds as the principle means of linlcaga. These methods of conjugation occur without any chemical pretreatment which might destroy the antigenicity of the nucleotide portion of the conjugate. Furthermore, these authors (39) have reported that the antibody response to a tetranucleotide of thymidylate may not be to the entire tetranucleotide. There also were no obvious differences in antibody responses to two different types of DNA, differing greatly in their base contents and ratios. This appears to be in direct apposition to the work of Stollar, Levine, Lehrer and Vasa Vvmakis (86) in which they postulate a tetra- or penta-nucleotide sequence as making up the antigenic determinant which determines the combining site of the antibody.

Similarly, Plescia and Braun 65 ( ) described nucleic acid specific antibodies produced in response to oligomers of thymidylate ranging 19 from two to six residues in length or even by a ribonucleotide made up of three residues.

Pleseia (61, 6^) has shown that soluble or transfer KM can function as a hapten when conjugated to methylated bovine serum al bumin. Similarly, synthetic co-polymers of deoxyadcnylate-thymidylate when complexed with methylated "bovine serum albumin were haptenic in rabbits. These antibodies cross-reacted with heat-denatured calf thymus DM, and they inhibited the transform ation reaction of B acillus subtills DIM. Furthermore, ribonucleotide homopolymers (63) have been conjugated to methylated bovine serum albumin. Antibodies were pro duced with specificity for tho ribonucleotides which made up the poly mers whereas antibodies made in response to the soluble RNA conjugates showed a wider range of reactivity to all the common nucleotides. Using methylated bovino serum albumin ribopolymer complexes as immunogens, Seaman, Van Vunakis, and Levine76) (produced antibodies directed toward polyadenylie, polyinosinic, and polyeytidilic acids. Polyinosinic acid cross-reacted with antibody to polyadenylie acid by gel diffusion and complement fix atio n , whereas tho reciprocal cross reaction between polyadenylie acid and anti-polyinosinie antibody was not observed. Only the polyinosinic acid—anti-polyinosinic acid system was inhibited to any significant extent by mononucleosides. This implies that the antigenic determinants are larger than mono nucleosides. Recently, Karol and Tanonbaum(1 1 ) studied some immunogens which cause stimulation of antibodies cross-reacting with RNA. These included 2,^~quina2olinedione-6-aao-albumin, pseudouridine, and 20 dihydrouridine conjugates of albumin. A uridine specific antiserum and the globulin from a dihydrouridine antiserua can be shown to react with ENA by complement f 3ratio n . A ll of the above conjugates cross- rea ct in gel with c alf thymus DKA. Barbu and Dandeu (5) and Barbu and Panijel (3» *0 have de scribed the presence of antibodies reacting with both ribosomal pro tein constituents and with the ribonucleic acid from bacterial ribo somes. Panijel, Barbu, and Quash (56) found a normal factor in the sera of humans and non-immunized rabb its which reacted with. ENA. That tolerance was not observed to ribosomal ENA is shown by the observation that non-immunised adult rabbits exhibited a delayed type hypersensi tivity towards small doses of ribosomes of diverse origin including autologous ribosomes. New-born rabbits inoc.iil.ated during the first seven days after birth with bacterial ribosomes as adults exhibited the cutaneous delayed hypersensitivity reaction only after larger dosages of ribosomes than were required by the non-iimaunizod group. These investigators maintain that tolerance cannot be established to RNA, but it can be provoked to the protein moiety of the ribosomes. P an ijel, Cayeux, Sacquet, and Charlier (57) in discussing anti-RNA facto rs in germ-free animals place the re a c tiv ity in the classes of NGI and NGII antibodies. NG refers to anti-irucleic acid antibodies. Tho NGII globulins are the result of autoirmiunisation and reflect an intrinsic quality of normally synth.esi.zed globulins. The NGI globulins are produced in response to immunization by ribosomes and are thought to bolong to the 73 globulin class. This globulin typ9 has been investigated by Sandor and co-workors (73) in horse 21 serum. It has been described to be a 7S gamma globulin which is immunologically identical to a 19S gamma-1(beta 2)--M~globulin. This factor is present at levels lower than 0.1 per cent in normal horso serum but makes up 1-2 per cent of the globulins in horses immunized with ribosomes. The extent of organization of polyribonucleotides and their involvement in the immunochemical behavior in reacting with NGI a n ti bodies have been studied in great detail by Panijel, Souliel, and Cayeux (58). In studying homopolymers of one or several chains of poly I, poly U, poly C, and poly A, these investigators v.sed the pre cipitation of NGI antibodies of horses immunized with ribosomes of bacterial origin to determine the optimal conditions for such precipi tation reactions. When the anti-RNA antibodies are prepared in non- eleetrolyte solutions, all four polyribonucleotides precipitated in an identical fashion, showing that the bases do not determine the pre cipitation reaction either quantitatively or qualitatively. When the ionic strength is increased, the polyribonucleotides undergo a reor ganization which changes their ii>jmunochemical behavior. This reorgan ization results in an unmasking of antigenic sitos. For instance, it was determined with poly U that at increased, ionic concentration, there was a molecular conformation in which the molecule folded over upon it self. As the ionic strength was increased, the intermolecular confor mation became more and more folded with a subsequent unmasking of antigenic sites, causing the precipitation of more globulins. Further studies with poly A and poly C implicate an intramolecular reorgani zation with tho same unmasking of antigenic sites. In the case of 22 poly I, increasing ionic concentration showed a variable number of fragments which increased the order of organization and allowed a max imum amount of antibody to be p recip itated . To summarize th is work, the quantity of antibodies which a certain amount of polyribonucleo tides can precipitate is a function of the number of antigenic sites accessible to the antibodies along the chain. The number of antigenic sites available is directly related to the physical configuration of the molecule with larger amounts of antibody being precipitated as a result of the unmasking of antigenic sites. The authors expanded th e ir work to complexes of homopolymers such as complexes of poly A and poly U (59). It was determined that the formation of complexes causes a masking of antigenic sites and reduces the amount of antibody which is precip itated . However, when an additional strand of poly U is added, there is no increase in the masking of antigenic sites. It was hypothesized that tho additional strand of poly U is inserted between tho helical structure formed by the complex of poly.-A and poly U. This work show's th a t the estab lish ment of a helically organized structure determines the degree of masking of the antigenic sites which are located along the polyribose- phosphate backbone. This organisation produces a steric hindrance which reduces the amount of anti-RNA antibodies which can bo precipi tated . Further work regarding the characteristics of anti-ribosome antibodies is reviewed by Plescia and Braun65 ). ( Using polyinosinic acid as a test substance, it was found that human sera from individ uals with pathological conditions and other mammalian sera contained 23 factors which precipitated poly I. They were present at very low le v e ls, while the sera of ribosome immunized animals contained sig nificantly higher levels. Furthermore, the natural occurrence of basic proteins leads to the reaction of anti-ribosomal serum with DNA, polycytidilate, heparin and dextran sulfate in addition to RNA. Pre cipitation was carried out at 0.14 M NaCl and 0.3 M NaCl with only RNA being precipitated at both concentrations. It was found that there were two fractions of anti-ribosome serum. One of these fractions was the more basic and was precipitated preferentially by RNA and the polyribonucleotides. Pleseia and Braun (65) further describe the problems associated with the studies of anti-nucleic acid antibodies. The prime difficulty lies with the heterogeneity of antibodies because of nucleotide se quences which vary in length and chemical make-up. In addition, structural alterations in the nucleic acid antigenic complex may well influence immunologic reactivity. Furthermore, it has not been defi nitely established that the nucleotides must be in a direct sequence or that reactive nucleotides must be on tho same chain structure. Lastly, antibodies directed toward nucleic acids are also thought to have soma specificity toward the sugar-phosphate backbone in which case nucleotide sequences and base composition would play a reduced role in immunological considerations. The purpose of this dissertation is to further elucidate the antibody specificity of antibodies to RNA in both normal sera and in the sera of individuals with clinically diagnosed rheumatoid arth r i t i s and to compare these sp e c ific itie s . Comparisons w ill be made with reference to tho three techniques employed: (1) bentonite floe culation, (2) q uantitative complement fix atio n , and (3) passive hemagglutination. MATERIALS AND METHODS

Rheumatoid A rthritis Sera

All rheumatoid arthritis sera used in this investigation were obtained through tho courtesy of Dr. Vol K. Phillips of the Columbus Medical Center. Twenty-five pooled sera from serial bleedings of individual patients wore used, for all serological studies in this in vestigation o Normal sera were obtained from donors in tho Department.

Buffers and Solutions

Phosphate Buffer — Phosphate buffer (Na^HPO^ — KHgPO^ sys tem) , 0.15 M, pH 7.3 was used for a ll hemagglutination and hemaggluti nation absorption reactions. Modifi ed Barbital Buffer — Modified barbital buffer was pre pared and used fo r a l l complement fix atio n reactio ns. To prepare th is buffer, 2.8750 grams of 5»5~diethylharbituric acid wore dissolved in 250 ml of distilled water near tho boiling point. The remaining in gredients wore then added: 1.8750 grams sodium 5»5~diethylbarbiturats, *1-2.500 grams sodium chloride, 0.1103 grams CaCl °2H„0, and 0.5083 grams MgCl^' 6lig0. The solution was cooled and diluted to 1000 ml with distilled water. To use the buffer, one part of the stock solution prepared above was diluted with four parts of distilled water. The final pH of the buffer was adjusted, to 7.3.

25 26

Barbital Buffer for Electrophoresis — Tho high resolution barbital buffer for electrophoresis was made as follows: 6,0 grams Tris (hydr oxyme thylaniinoethane) , 2*5 grams 5»5~diothylba.rbitu.rie acid, and 10.4 grams sodium 5 ,5-diethylbarbiturate were dissolved in 2 liters of distilled water. Tho pH was adjusted to 8.6 with sodium hydroxide. Tho buffer was stored at refrigerator temperature and was used cold during electrophoresis, Tris—NaCl Buffer for Sephadex Gel Filtration — One-tenth M Tris HC1, pH 8.0 in 1 M NaCl was used for the swelling of Sephadox Gels and for the actual separation of serum components in the gel f i l tration techniques. A 1:10,000 dilution of merthiolata was used to prevent contamination of the gels and serum. Aquaclde Solution —* A 5 per cent solution of Aquaeide (Calbiochem) was used in order to concentrate globulins obtained from gel filtration and precipitation procedures. The material was dis solved either in distilled water or phosphate buffer, pH 7.3, de pending upon the serological technique which was to be employed. Normal Rabbit Serum —- Phosphate Buffer — Rabbit Serum was heat inactivated at 56°C for30 minutes and then absorbed three times with washed, packed human group 0, Rh+ erythrocytes volume for vol ume. One ml of this inactivated and absorbed normal rabbit serum was added to 99 ml of 0.15 M phosphate buffer, pH 7.3, making a 1 par cent solution to bo used in all passive hemagglutination techniques. Saturated Ammonium Sul-fate Solution — Seventy-five grams of (NH^gSO/j, were added to 100 ml distilled water at room temperature. 2?

This saturated solution was used for tho precipitation of gamma globulin fractions of rheumatoid and normal sera.

Serological Procedures

Tho q uan titative micro- complement fix atio n mothod of Wasserman

and Lovino 92 ( ) was used with some modifications. Spectrophotometric Standardization of Sheep Erythrocytes — Sheep erythrocytes were washed with modified barbital buffer until the supernatant was free of hemolysis. A suspension of these washed and packed erythrocytes which upon lysis had an optical density reading of

O.7O at 413 mu in a Bausch and Lomb Spectronic 20 spectrophotometer was made in modified barbital buffer. To lyse the red blood colls, 0.4 ral of the blood coll suspension was added to 2.4 ml distilled water and then thoroughly mixed. The suspension of erythrocytes was adjusted with washed and packed cells until tho lysing procedure produced an optical density reading of0.70 a t 413 mu. Titration of anti-Shsep Erythrocyte Hemolysin — Commercially prepared sheep erythrocyte hemolysin (Grand Island Biologicals, Grand. Island, New York) was pooled and subsequently was titrated against tho

standardized sheep erythrocyte suspension. Tho highest dilution of hemolysin which caused complete hemolysis in an excess of complement was used fo r a l l the complement fix atio n reactions. I t was stored a t refrigerator temperature.

Preparation of Sensitised Sheep Erythrocytes — The standard

ized suspension of sheep erythrocytes was incubated with tho proper 28 dilution of hemolysin for one hour at 37°C before the cells were added to the test system proper. Titration of Guinea Pig Serum Complement — Guinea pig com plement (Grand Island B iologicals, Grand Island, New York) was recon stituted, pooled, and titrated as follows. A 1:200 dilution of pooled complement was made in tho modified b a rb ita l b uffer. A series of tubes was se t up in to which various amounts of the 1:200 complement dilution were added, ranging from 0.05 ml to 0.80 ml. The volume was then brought to a total of 2,4 ml by the addition of barbital buffer. These tubes were then incubated for 18 hours at 4°C before the ad d itio n of the sensitized erythrocyte suspension since complement loses approximately 30 per cent of its activity in dilute solutions under those conditions. After this incubation period, 0.4 ml of the sensi tized erythrocytes was added and the tubes were incubated at 37°0 for 30 minutes. At this time, the tubes were centrifuged in an Adams Serofuge for one minute. The supernatant fluid was then transferred to spectrophotomet er tubes and analyzed at 413 mu. The dilution of 1. complement used in the te s t procedure was th a t azaound which gave90 ' per cent homolysis of the standardized erythrocyte suspension. Titration of the Antigen — Serum dilutions ranging from 1:40 through 1:40,960 wore made in modified barbital buffer. The antigen used was soluble RNA of bovine liver origin (Nutritional Biochemicals, Cleveland, Ohio). The concentrations used ranged from 40 ug to 100 ug per ml. To each of the serum dilutions, the various antigen concen trations were added. Complement was added and the tubes were incu bated at 4°C for 18 hours. At the end of this time period, sensitized 29 sheep erythrocytes were added, and the hemolysis was read at 413 mu to determine the degree of complement fix atio n . Protocol for the Test System — To each of the tubes were added the following reagents. Each test system and control were set up in triplicate. Test System: Antibody — 0.4 ml of each dilution Antigen — 0.4 ml of each concentration Complement — 0.4 ml of d ilu tio n giving 90 per cent hemolysis of standardized RBC suspension Buffer — 1,2 ml

Antigen Control: Antigen — 0.4 ml Complement —■ 0.4 ml Buffer — 1.6 ml Antibody Control: Antibody — 0.4 ml Complement — 0.4 ml Buffer — 1.6 ml

100 per cent Hemolysis Contro l: . Complement — 0.8 ml. Buffer 1.6 ml

Complement Control : Complement — 0.4 ml Buffer ~~ 2.0 ml Cell Control: Buffer — 2.4 ml These tubes were then incubated for 18 hours at 4°C to insure maximum complement fix a tio n . At th is time, 0.4 ml of tho sensitized sheep erythrocytes were added and the entire mixture m s further incubated for 30 minutes at 3?°C. They were then analyzed spectrophotometrically at 413 mu. Since® tho standardized erythrocyte suspension has an optical density of 0 , 70 , the complement control should have an o p tical density of 0.63 because the 90 par cent hemolysis endpoint is used in 30 th is procedure. The percentage of complement fix atio n was determined hy observing the difference in optical density readings of the test system and the antibody control. The change in optical density di vided by the optical density of the antibody control when multiplied by 100 indicated the percentage of complement fix atio n . In h ib itio n Studies with the Q uantitative Micro-Complement Fixation Technique — For inhibition studies, nucleotides, nucleosides, ribose, ribose phosphate, DM and RNA were made up in modified barbital buffer to a concentration of 100 ug per ml. From these stock solu tions, concentrations of 4, 8, 12, and 16 ug per ml were prepared in modified barbital buffer. The pH of each solution prepared in this manner ranged between 7.2 and 7.4. The sera to be teste d were d i luted in these inhibitor dilutions to the desired concentration of antibody and were allowed to incubate for 24 hours at 4-°C„ At this time, 0,4 ml of the antigen concentration and 0.4 ml of complement were added to the system, and the entire mixture was incubated for an ad ditional 18 hours at 4°C. Sensitised sheep erythrocytes were added and further incubation at 37°0 for 30 minutes was carried out. The tubes were then centrifuged, and the supernatant fluid transferred to spectrophotometer tubes. Analysis was carried out at 413 max. To determine the percentage of complement fix atio n in h ib itio n , the optical density of the tost system is divided by the optical density of the antibody control and multiplied by 100. Passive Hemagglutination Technique — The passive hemagglutin ation technique as described by Bigley, et al (15) was carried out in which antigens were covalently coupled to human group 0, Rht red blood 31 cells with bis-diazotized benzidine vising both precipitated gamma globulins and serum fractions from gel filtration of rheumatoid arth ritis sera. Bis-diazotized benzidine was prepared using the pro

cedure of Gordon, et al36 (). To 0.23 grams of benzidine base in ^5 ml of 0.2 N HC1, 0.175 gram of KaNO^ in 5.0 ml of distilled water was added at 0°C. The reaction proceeded for 30 minutes with stirring. The material was divaried into 0.7 ml aliquots and was quickly frozen in a dry ice—acetone bath at -78°C. It was stored at ~70°C until use when it was used in a 1:16 dilution in phosphate buffer, pH 7.3. Coupling of Antigens to Erythrocytes — To couple the antigens to the erythrocytes, the following'procedure was visc-d. Freshly ob tained human group 0, Rh+ erythrocytes were washed with 0.15 M phos phate buffer until there were no traces of hemolysis in the superna

tant fluid. To 9.75 m3, of phosphato buffer was added 0.25 ml of washed, packed erythrocytes, followed by thorough mixing. The antigens to be coupled to the erythrocytes were added at a final concentration .of 1000 ug per ml. Then 0.5 ml of the 1:16 dilution of bis-diazotized benzidine was added to the erythrocyte mixture. This was fol3.owed by incubation at room temperature for 10 minutes with frequent mixing. At the end of this time, the sensitized cells were centrifuged and washed once with a 1 per cent solution of normal rabbit serum--phosphate buffer. The colls were centrifuged and resuspended in the rabbit serum—phosphate buffer solution. Titrations — The pooled gamma" globulin fractions as well as the fractions obtained from Sephadex gel filtration were spectrophoto- metrica3.1y standardised to 1500 ug per ml using a Beckman DU 32 spectrophotometer and an ultraviolet light source* Doubling dilutions were made in 0.5 ml amounts in normal rab b it serum—phosphate buffer. For the titrations, to each of those dilutions 0.1 ml of the coupled antigen preparation was added. The mixture was then incubated at 3?°C for 30 Minutes at which time the tubes were centrifuged for 30 seconds in an Adams Serofuge. The tubes wore gently tapped to break up the pellet of cells and were observed for both, macroscopic and micro scopic agglutination. The degree of macroscopic hemagglutination was determined with maximum hemagglutination being designated as 4+ while microscopic agglutination was determined on a plus or minus basis, Hemagglutlnln Absorption — After titrations were performed, serum samp3.es wero absorbed with each of the antigens. The coupled c e lls were centrifuged and added on a volume fo r volmu© ra tio with the antiserum. They wore then incubated fo r 60 minutes a t 37°C. Hie ab sorbed sera wore then tested for reactivity with the absorbing antigen by microscopic examination. Absorptions were continued until there was no reactivity demonstrable for the absorbant. Tho absorbs! sera were then tested against tho remaining antigens in the samo manner as described above. In the cases of ribose and ribose phosphate, tho sera were inhibited with 20,000 ug per ml of these substances before testing sine© these two compounds cannot b© coup3_ed to erythrocytes with bis- diazotiaed benzidine. Base Pair Absorptions — Passive absorptions were performed using two bases coupled to human group 0, I?h+ erythrocytes, Mixtures of the base pairs A-C, A~T, A-U, G-C, G~T, and G-U were prepared to a final concentration of 1000 ug per m3., lia3Lf of which was a purino base 33 and the other half a pyrimidine base. Absorptions were made until there was no reactivity with the absorbants. The absorbed sera were then tested with the other antigenic substances and were examined for both macroscopic and microscopic hemagglutination Precipitation of Gamma Globulins with Ammonium Sulfate ~ Kheumatoid a rth r itis and normal serum pools were used, Y7ith constant stirring, ammonium sulfate solution, saturated at room temperature, was added dropwise to effect one-third saturation. At the completion of the ammonium sulfate addition, the pH of the solution was adjusted to 7.8 with 1 N NaOH. The suspension was then stirred for two hours after which it was centrifuged at room temperature for30 minutes. The precipitate was dissolved in phosphate buffer, pH 7.3, to its orig inal volume. Two additional precipitations were performed as above in order to purify the gamma globulin fraction. After the last pre cipitation, the final volume of phosphate buffer used to dissolve the precipitate was one-half the original volume. Dialysis of this glob ulin solution against distilled water was carried out at ^°C for three days with frequent changes of water. The globulins were then dialyzed against phosphate buffer, pH 7.3, for ?J-V hours. They were then con centrated by using a 5 per cent solution of Aquacide, Standardization of the globulins to 1500 ug was then performed using the phosphate buffer as the diluent. Sephadex Gel F iltra tion — Sephadex G-100 and G-200 were used to fractionate undiluted sera. The gels were allowed to swell for at least 72 hours in an excess of 0.1 M Tris HC1, pH 8.0, in 1 M NaC3., Two 45 centimeter by 2.5 centimeter columns were used. Packing of the y - v columns was done separately and then were connected to each other. To pack the solumns, a small amount of the Tris—NaCl buff ex' was added to each of the solumns „ The slurry was then added to almost completely f i l l each column. Air bubbles were removed by s tirrin g with a glass rod. The outlet was then opened and the buff ex- was allowed to leave the column. When the buffer lev el was about one inch above the packed gel, more slurry was added and followed by stirring in order to remove a ir bubbles as woll as to avoid the formation of an in te r face between the packed gol and the new slurry. When tho layer of packed gel reached a level of 10 centimeters from tho top of the column, a sample applicator was added to the G-100 column. At the top of the G-200 column, an upward flow adaptor was placed. The upward flow adaptor was connected to a Uvicord Scanner (LKB Productur, Stock holm, Sweden). The G-100 column was connected to the G-200 column so that the G-100 acted as a descending filtration gel and tho G-200 acted as an ascending filtor. A buffer reservoir of 0,1 H Tris—NaCl was connected to the G-100 column. To apply the sample, buffer was allowed to leave the system until it was at the level of the gel be low tho sample app licator. Throo ml of undiluted sera were placed in the sample applicator and were allowed, to enter tho gal. The buff ex* reservoir was then reconnected and the column allowed to fill to tho top. Tho flow rate was maintained at 20 to 25 ml per hour. Tho frac tions obtained were dialyacd against frequent changes of distilled water for two days at ^°C and then were restored to their original vol umes in either modified barbital buffer or phosphate buffer, pH 7-3» after concentration using a 5 per cent solution of Aquacide. The 35 samples were checked electrophoretically for purity. Only the peak of the 7S fraction was collected to avoid contamination by other serum components, Electrophoresis Procedures — Whole rheumatoid arthritis sera and normal sera as well as their fractions were subjected to electro phoresis vising the Gelman Model 513-70 Electrophoresis Chamber and cellulose polyacetate electrophoresis strips (Gelman Instrument Com pany, Ann Arbor, Michigan), Tho technique used was that described in the Gelman Manual #70176-B (35)• Scanning of the stained and cleared strips was performed on a Beckman Analytrol equipped with a Gelman Scanatron, Conglutination Technlque A 0.5 per cent suspension of sheep red blood cells was sensitized with whole bovine serum which was heated at 56°C for 30 minutes. A 1:5 dilution of horso serum served as the source of complement. Rheumatoid sera as w ell as normal sera were heat inactivated at56°C fo r 30 minutes and absorbedk times with washed, packed sheep erythrocytes, after which serial doubling di lutions were made in saline. Each tube contained the following mate rials: 0.5 ml of a 1:5 dilution of horse serum; 0.5 ml sensitized sheep erythrocytes; and 0.5 ml of test serum dilution in saline. The tubes were incubated at 37° C for 30 minutes, centrifuged inAdams an Serofugs for ^5 seconds and were read for hemagglutination with maxi mum hemagglutination designated as *M-. Positive controls consisting of bovine serum as a source of conglutinin, and negative controls con sisting of cell suspensions prepared with heat inactivated horse serum 36 wore included in each ’experiment. The titer was expressed as the last tube to show positive agglutination on resuspension of the pellet. RESULTS OF EXPERIMENTATION

The purpose of this investigation was to study and delineate further and more completely the specificity of anti-nucleic acid antibodies in the sera of individuals with clinically diagnosed rheumatoid arthritis as well as in tho sera of normal adult individ uals. Twenty-five pooled sera wore used in all sorological studies. These pools were made up from serially obtained bleedings of 25 indi viduals whose sera were shown previously to contain antibodies to RNA by the bentonite flocculation method. Nineteen of the serum pools were from patients who showed positive latex flocculation reactions for the rheumatoid factor at each bleeding. Six pooled samples wore included from patients with positive reactions with RNA-bsntonite but with negative latex flocculation reactivities. In addition to these pathologic sera, _s.er& from five normal, healthy adults were also studied for their anti-nucleic acid specificities. Tho quantitative micro-complement fixation technique of Was- serraan and Levine (92) was employed to determine the percentage d if ferences in the inhibitory capacity of various nucleotides, nucleo sides, and bases as well as DNA, RNA, sodium ribonueleate, sodium deoxyribonueleate, ribose and ribose phosphate. This technique em ploys the 90 per cent endpoint of ly sis as the d ilu tio n of complement to be used.

37 38

Dilutions of normal serum in modified barbital buffer were made ranging.from 1:40 to 1:40,960, The concentrations of RNA used ranged from 100 ug par ini to 40 ug per ml. Each serum d ilu tio n was tested with each of tho antigen concentrations in order to determine the dilution of serum and concentration of antigen which would produce optimal complement fix atio n . A concentration of 4-0 ug RNA per ml and a 1:5120 d ilu tio n of normal human serum brought about maximum comple ment fix atio n . The actu al concentration of RNA used was 16 ug per 0.4 ml since th a t volume was the one used in the actual te s t pro cedure. In order to study the specificity of the reaction, the sera were diluted, to the proper dilution by using inhibitor—modified bar bital buffer solution as the diluent. Four concentrations of each in hibitor wore prepared—4, 8, 12, and 16 ug per ml. The results of these inhibition studies appear in the following Tables. Table 1 shows the relative effectiveness of tho purine and pyrimidine bases as in h ib ito rs of complement fix atio n . I t is to be noted that guanine is most effective at a concentration of 12 ug per ml. with the remaining concentrations being less e ffic ie n t. However, the other purine base adenine is most effective a t i t s maximal con centration of 16 ug per ml. Similarly, the pyrimidine bases uracil and cytosine appear most effective at 12 ug per ml concentration while thymine is most effective at a concentration of 16 ug per ml. On an overall view, however, cytosine and thymine are not as good as in hibiting agents as is uracil on a percentage basis. It should be noted. th a t the pyrimidine bases inhibi.t complement fix atio n to a much higher degree than do the purine bases.

TABLE 1 .—Percentage of complement fix atio n in h ib itio n of anti-RM activity in normal human serum by pur line and pyrimidine bases

In h ib ito r Concentration of Inhibitor (ug) 4 8 12 16

Purine Bases Guanine 15 24 33 14 Adenine 12 16 27 42 Pyrimidine Bases U racil 36 68 78 63 Cytosine 33 50 70 61 Thymine 4 16 18 58

Table 2 shows the inhibitory effects of several, ribonucleoside upon normal serum. Ui’idinc was the most inhibitory substance tested with peak activity7- at 12 ug per ml concentration. Furthermore, cytidine also had its greatest effect at the same concentration while guanosine displayed peak reactivity at the 8 ug per ml level. As was previously shown with the bases, the pyrimidine ribonucleosides were much more effective as inhibitors than were the purine ribonucleosides

TABLE 2 .—Percentage of complement fix atio n inh ib itio n of anti-RM activity in normal human serum by pur ino and pyrimidine ribonucleosides

In h ib ito r Concentration of Inhibitor (ug) 4 8 12 16

Uridine 58 81 91 85 Cytidine 27 30 42 20 Guanosine 13 36 14 14 40

Ribonucleotide inhibitions of normal sera are shown in Table 3. Adenylic acid and cytidylic acid appear to be equally inhibitory at 12 ug, but cytidylic acid is much more effective at the lower eon- centrations whereas adenylic acid is more effective at the higher con centrations. Guanylic acid is only a moderate inhibitor at all con centrations while uridylic acid is the most effective overall inhibi tor. Thymidylic acid appears to be only a moderate inhibitor and is only effective at the highest concentration used. As lias been shown previously, a pyrimidine base compound is a much better inhibitor than the purine counterpart.

TABLE 3 .—Percentage of complement fix atio n inh ib itio n of anti-RNA activity in normal human serum by purine ■ and pyrimidine ribonucleotides

In h ib ito r Concentration of Inhibitor (ug) 4- 8 12 16

Purine Ribonucleotides Adenylic acid 24- 36 80 76 Guanylic acid 31 4-2 4-5 4-7 Pyrimidine Ribonucleotides Uridylic acid 72 88 93 89 Cytidylic acid 74- 81 82 4-3 Thymidylic acid 11 33 40 73

In addition to comparing bases, ribonucleosides, and ribonuc leotides, comparisons were made among compounds containing the same base. Table 4 shows a summary of normal serum inhibitions with com pounds containing adenine. These data demonstrate that the greatest percentage of inhibition is obtained with adenylic acid at a concen tration of 12 ug per ml. Similarly, adenosine triphosphate shows a 41 a psak inh ib itio n a t 12 ug. Adenine displays maxamua complement fix - ation inhibition at 16 ug. Furthermore, adenosine triphosphate ap pears to be a better inhibitor at lower concentrations than either adenylic acid or adenine.

TABLE 4 .—-Percentage of complement fix atio n inh ib itio n of anti-RNA activity in normal human serum by adenine-containing compounds