Ann. rheum. Dis. (1974), 33, 258 Ann Rheum Dis: first published as 10.1136/ard.33.3.258 on 1 May 1974. Downloaded from IgG anti-IgG in rheumatoid arthritis and certain other conditions

JOHN L. ABRUZZO AND RALPH HEIMER From the Departments ofMedicine and Biochemistry, Jefferson Medical College, Thomas Jefferson University, Philadelphia, U.S.A.

Rheumatoid factor (Rose, Ragan, Pearce, and latex test for the detection of IgG anti-IgG antibodies Lipman, 1948), known to be a frequent finding in the (Heimer and Abruzzo, 1972) which, in the conditions sera of patients with rheumatoid arthritis, was employed by us, is specific for the detection of IgG thought to be an unusual in that it occurred anti-IgG. in the 19S and 22S forms (Franklin, Holman, In this study we demonstrate the of Muller-Eberhard, and Kunkel, 1957) without show- this latex test and report the occurrence of IgG anti- ing evidence for 7S analogues. The activity of this IgG antibodies in rheumatoid arthritis, osteoarthritis, antibody appeared to be directed against gamma- hepatitis, and primary biliary cirrhosis. globulin. It was not until the early 1960s that 7S antibodies, Material and methods presumably of the IgG class, with similar activity were also noted. These were first described (Kunkel, PREPARATION OF IMMUNOADSORBENT Muller-Eberhard, Fudenberg, and Tomasi, 1961) as 1 g. human IgG (Fraction II, Squibb Laboratories) was circulating intermediary 9-17S) complexes composed dissolved in 20 ml. phosphate buffer (pH 7-2, 0.1 M) and copyright. of 7S gammaglobulin monomers, occurring in a few heated to 63°C. for 10 min. After cooling, 4 ml. of 2-5 per selected cases of severe rheumatoid arthritis. Addi- cent. aqueous glutaraldehyde was added drop by drop with tional evidence of their existence was later presented stirring, following the general procedure of Avrameas and Ternynck (1969). The mixture was allowed to stand for by Chodirker and Tomasi (1963) and Heimer and 2 hrs at room temperature, and the resultant gel was Levin (1966), but it was not until 1967 that Torrigiani homogenized in a blender in the presence of ammonium and Roitt (1967), with a combination of immuno- formate buffer (pH 3 0, 0-1 M). The homogenized material adsorption and radial with specific was distributed in 40-ml, plastic test tubes, each containing antisera, reported the wide occurrence of IgG anti- approximately 125 mg. immunoadsorbent. The latter was http://ard.bmj.com/ IgG antibodies in rheumatoid sera. There is evidence suspended in phosphate buffer (pH 7-0, 0 1 M) and stored (Winchester, Agnello, and Kunkel, 1970) to suggest at 4°C, until used. The immunoadsorbent could be used that IgG anti-IgG antibodies may have an important repeatedly, but care had to be exercised in removing all role in the pathogenesis of the chronic synovial previously adsorbed material by washing with pH 3 0 inflammation which characterizes rheumatoid arth- buffers. ritis. SERUM IMMUNOADSORPTION Further reports (Torrigiani, Roitt, Lloyd, and Sera were used either when freshly obtained or upon on October 2, 2021 by guest. Protected Corbett, 1970; Torrigiani, Ansell, Chown, and Roitt, storage in the frozen state. 1 ml. serum was exposed to 1969; Adachi, Atsumi, Saito, Nakamura, and 0-05 M mercaptoethanol for 1 hr at room temperature Horuichi, 1969; Panush, Bianco, and Schur, 1971; and then diluted with phosphate buffered saline to 30 ml. Bianco, Panush, Stillman, and Schur, 1971; Tapanes, The solution was added to 125 mg. immunoadsorbent and Rawson, and Hollander, 1972) have appeared of the shaken at room temperature for 1 hr. The tubes were then presence ofserum IgG anti-IgG antibodies in rheuma- centrifuged and the supernatant discarded. The immuno- adsorbent was washed usually four times with 30 ml. toid arthritis and related arthritides, utilizing tech- phosphate-buffered saline. The criterion for the final wash niques of immunoadsorption for purification fol- was an absorbance of less than 0 005 at 280 mg in a 1 cm. lowed by radial immunodiffusion for identification light path. IgG anti-IgG was then eluted in 10 ml. ammo- and quantitation of these antibodies. nium formate buffer (pH 3 0, 0-1 M). A single such extrac- Previously we have described a modified procedure tion usually accounted for approximately 80 per cent. of of immunoadsorption and a simple tube-dilution all recoverable IgG. In order to avoid further dilution of

Accepted for publication September 9, 1973. Re rints requests should be addressed to: John L. Abruzzo, M.D., Thomas Jefferson University, 1025 Walnut Street, Philadelphia, Pa. 19107 USA. This work was supported in part by the Eastern Pennsylvania Chapter of the Arthritis Foundation, American Cancer Society Grant IC37B, The National Institutes of Health Grant AM 14274, and the Arthritis Clinical Research Center grant from the Arthritis Foundation. IgG anti-IgG antibodies in rheumatoid arthritis and certain other conditions 259 Ann Rheum Dis: first published as 10.1136/ard.33.3.258 on 1 May 1974. Downloaded from eluate, additional extractions were not carried out. The Table Reproducibility ofresults eluate was brought to pH 8'4 with the addition of con- centrated NH4OH. Test no. Number oftubes agglutinated* LATEX 1 4.5 Materials eluted from the immunoadsorbent were neutral- 2 4.5 ized and serially diluted in 0 5 ml. phosphate buffer 3 4.5 (pH 8-4, 0-1 M). The latex reagent was prepared with 4 4.5 No. 615 Lytron particles (0-2 ,u diameter, Monsanto 5 4.5 Chemical Co.), suspended in phosphate buffer (pH 8-4, 6 4-5 0-1 M), and adjusted in a Hitachi-Perkin Elmer spectro- 7 4.5 8 5'0 photometer to an absorbance of 0'80 at 400 m,r in a 1 cm. 9 6-0 light path. After addition of 0-5 ml. reagent to each tube, 10 5'0 the mixtures were incubated at 56°C. for 30 min. and read 11 5-0 after standing overnight at room temperature. The agglutination patterns were stable for at least 3 days. ZXi = 52'5 Mean = 4-77 tubes POPULATION Standard deviation = ±0'47 tubes Serum samples were obtained from 38 persons with definite or classical rheumatoid arthritis (Ropes, Bennett, Cobb, * 'Tubes agglutinated' represents the number of tubes containing serially diluted eluate showing agglutination. If the last tube was rated Jacox, and Jessar, 1959). Thirty had serum rheumatoid 'trace agglutination', it was recorded as 0O5, hence the 4 5 recorded in factor titres of 1:80 or greater. Eight were negative for the Table. As the first tube tested represented a dilution of 1: 10, four serum rheumatoid factor or had titres of less than 1: 80 as tubes is equivalent to 1: 80, 4-5 tubes to 1: 120, and five tubes to 1:160. measured by the latex test of Singer and Plotz (1956). The patients were selected only on the basis ofadequate clinical group. The difference in titres between these two data and availability of serum. groups is highly significant (P (Student's test) less Serum samples were also obtained from thirty persons with osteoarthritis and from twelve healthy young adults than 0'001). The geometric mean titre in the healthy randomly selected from among hospital employees. There young adult group was 1: 15. The range of individual copyright. were nineteen sera from patients with hepatitis, nine of titres within each group was such that the rheumatoid whom were positive for Australia , and five and the healthy young adult control groups had patients with primary biliary cirrhosis. The diagnosis of few overlapping values. The osteoarthritis group, on hepatitis and primary biliary cirrhosis was based on typical the other hand, contained a wide range of titres that clinical and laboratory features combined with liver overlapped with the other two groups. biopsy histology consistent with those entities. Our results appear to be consistent with those ob- The mean age of the rheumatoid arthritis group that of was rheumatoid factor positive was 60 years. The mean tained by other methods of purification IgG anti- age ofthe rheumatoid factor negative subgroup, the osteo- IgG followed by quantitation using radial immuno- http://ard.bmj.com/ arthritis group, and the healthy young adult group was diffusion (Torrigiani and Roitt, 1967; Torrigiani and 51, 65, and 25 years respectively. The mean age of the others, 1970; Panush and others, 1971; Tapanes and hepatitis group was 34 years and that ofthe primary biliary others, 1972). cirrhosis group 47 years. The geometric mean titre of the IgG anti-IgG anti- body in the hepatitis group was 1:27 and was evi- dently independent of the simultaneous presence or Results absence of Australia antigen. The geometric mean In order to determine the reproducibility of our titre as well as the range was similar to that observed on October 2, 2021 by guest. Protected method of testing IgG anti-IgG antibody, eluates in the osteoarthritis group and significantly greater recovered after immunoadsorption of a serum of a than that of the healthy young adults whose age patient with rheumatoid arthritis were tested on eleven range is similar to that of the hepatitis group. separate occasions by two technicians independently No evidence of significant titres of IgG anti-IgG in two separate laboratories. Serial dilution was was found in primary biliary cirrhosis. The highest started with an aliquot equivalent to a 1:10 dilution titre in this group was 1:20, well within the range ofthe original serum. A high degree ofreproducibility measured in the healthy young adults. The number of was achieved (Table). The mean titre was 4'77 tubes patients with primary biliary cirrhosis, however, is agglutinated with a standard deviation of +0'47. small and does not allow us to draw a definite con- Having established the reproducibility of the clusion. agglutination test, we then analysed a panel of sera obtained from our clinic population. As the Figure Discussion (overleaf) shows, the highest titres of IgG anti-IgG antibody were found within the rheumatoid group. IgG anti-IgG antibodies can be found in relatively The geometric mean titre in the rheumatoid group low concentrations in the serum of many patients was 1:79 compared with 1:27 in the osteoarthritis with rheumatoid arthritis (Torrigiani and Roitt, 260 Annals of the Rheumatic Diseases Ann Rheum Dis: first published as 10.1136/ard.33.3.258 on 1 May 1974. Downloaded from

-a

0320 - 0 4, --t160. I0 *4 0000, 00 .0. 0O *- 80- -000------00000 5-A 0@ 0...... 000 S-- 0 c 40" 000 000I * 0@ - @- A- 0O* * 0 : 20 I 0@ 0 00 a *000 @0.-- 0 X 10 * 0000* 000 0 0

-J 0* RHEUMATOID ARTHRITIS HEALTHY HEPATITIS PRIMARY RF(+) RF(-I SEARHII ADULTS AU (+) AU (-I BILIARY CIRRHOSIS * titre refers to oriqinal serum

FIGURE Titre of IgG anti-IgG antibodies in eluates of serum of patients representing a variety of clinical groups RF= rheumatoidfactor; A U = Australia antigen Horizontal broken lines represent median titres

1967; Torrigiani and others, 1970; Panush and others, immunodiffusion and latex agglutination at the same 1971; Tapanes and others, 1972). These results are time. Our test has the disadvantage in regard to quan- confirmed by our study, which employed a modified titation that is inherent in all serial dilution proce- immunoadsorbent for isolation and latex agglutina- dures, but our results indicate that this disadvantage

tion for quantitation of IgG. Our test system is based appears to be of little consequence. copyright. on the observation (Heimer and Abruzzo, 1972) that Our finding of IgG anti-IgG antibodies in patients suspensions of certain preparations of latex particles with chronic active viral hepatitis and their relative are agglutinated when exposed to even very small absence in cases of primary biliary cirrhosis is of amounts of IgG anti-IgG antibody in the absence of interest. It is not surprising that IgG anti-IgG anti- inhibiting substances. The presence of other IgG bodies exist in diseases other than rheumatoid appears to inhibit the agglutination of the test par- arthritis. IgM anti-IgG antibodies have been sub- ticles by IgG anti-IgG antibody, hence the isolation of stantiated in patients with viral hepatitis (Risemberg,

IgG anti-IgG antibody by immunoadsorption is a DeGomez, and Rife, 1969) and anti-IgG antibody http://ard.bmj.com/ necessary procedure. Our test appears to have certain production is probably the result of stimulation by advantages over previously published methods for denatured IgG or IgG present in an immune complex detection of IgG anti-IgG antibody. The immuno- (Milgrom and Witebsky, 1960; Abruzzo and adsorbent composed of heated and glutaraldehyde Christian, 1961; Aho and Wager, 1961; Williams and cross-linked IgG is more efficient than cross-linked Kunkel, 1963). Hence it is possible that Australia IgG which has not been heated (Heimer and Abruzzo, antigen and presumably additional operative 1972). Pre-treatment of the serum with 0-05 M in viral hepatitis when complexed with antibody lead mercaptoethanol and 30-fold dilution of the sample to the stimulation of IgG anti-IgG antibodies. If the on October 2, 2021 by guest. Protected before exposure to the immunoadsorbent renders essentially negative response observed in biliary eluates free of detectable IgM and Clq (Heimer and cirrhosis can be confirmed, it would suggest that there Abruzzo, 1972). While serum albumin was occasion- may be no stimulus for anti-IgG production in that ally observed to be a trace contaminant, even highly disorder active at the time of our testing. It would be concentrated eluates failed to react with three different ofinterest in this regard to examine patients with early anti-IgM preparations. Our latex test is sensitive in active disease in order to examine that question more that serum concentrations ofIgG anti-IgG antibodies thoroughly. as low as 5 ,ug./ml. give a titre of 1:10 (Heimer and It is not surprising that a relatively large percentage Abruzzo, 1972). Furthermore, the test results are of patients with osteoarthritis were found to have quite reproducible and as no final concentration step increased titres of IgG anti-IgG antibody in their is required as is for radial immunodiffusion, some serum. Persons with osteoarthritis are generally older labour is saved and potential errors usually associated and titres of IgG anti-IgG antibody activity may with concentration procedures are avoided. The cost reflect a cumulative, more varied, and prolonged of a latex agglutination test is negligible when com- immunological host experience with a multiplicity of pared with that of radial immunodiffusion, and since antigens. Additi6nal evidence for this hypothesis is the test is so very sensitive, one can elect to do radial the known fact that IgM anti-IgG antibodies (rheu- IgG anti-IgG antibodies in rheumatoid arthritis and certain other conditions 261 Ann Rheum Dis: first published as 10.1136/ard.33.3.258 on 1 May 1974. Downloaded from matoid factor) occur in older persons (Heimer, Levin, for serum IgG anti-IgG antibody. The highest titres and Rudd, 1963) in significantly greater quantities were found in the rheumatoid arthritis group; the than in healthy young adults. geometric mean titre being 1: 79. The geometric mean titres in the osteoarthritis and hepatitis groups Summary were both 1: 27 and in healthy adult controls 1: 15. The possible significance of serum IgG anti-IgG A modified technique of immunoadsorption and a antibodies in these various groups is discussed. new test were the simple latex used for detection of We are grateful to Dr. Hyman Menduke for providing the serum IgG anti-IgG antibodies. As used by us, the statistical analyses, to Drs. Gordon Benson and Jay latex test is both specific and reproducible. 38 subJects Ziegenfuss for supplying certain serum specimens and with rheumatoid arthritis, thirty with osteoarthritis, clinical information, to Miss Audrey Rouse and Mrs. nineteen with hepatitis, and five with primary biliary Ophelia Buella for technical assistance, and to Miss Jane cirrhosis, and twelve healthy adults were examined Wherrity for secretarial and typing services.

References ABRUZZO, J. L., AND CHRISTIAN, C. L. (1961) J. exp. Med., 114, 791 (The induction of a rheumatoid factor-like substance in rabbits) ADACHI, M., ATSUMI, T., SAITo, N., NAKAMURA, M., AND HoRiucm, Y. (1969) Int. Arch. Allerg., 35, 77 (Detection of IgA and IgG rheumatoid factors by antiglobulin augmentation technique) AHO, K., AND WAGER, 0. (1961) Ann. Med. exp. Biol. Fenn., 39, 79 (Production of 'anti-antibodies' in rabbits. Appearance in rabbit serum of 'anti-bodies' reacting with autogenous and isogenous antibody, following auto-stimulation with protein antigens) AVRAMEAS, S., AND TERNYNCK, T. (1969) Immunochemistry, 6, 53 (The cross-linking of proteins with glutaraldehyde and its use for the preparation of immunoadsorbents) BIANCO, N. E., PANUSH, R. S., STILLMAN, J. S., AND SCHUR. P. H. (1971) Arthr. and Rheum., 14, 685 (Immunologic studies ofjuvenile rheumatoid arthritis) CHODIRKER, W. B., AND TOMASI, T. B., JR. (1963) J. clin. Invest., 42, 876 (Low-molecular-weight rheumatoid factor) copyright. FRANKLIN, E. C., HOLMAN, H. R., MULLER-EBERHARD, H. J., AND KUNKEL, H. G. (1957) J. exp. Med., 105, 425 (An unusual protein component of high molecular weight in the serum of certain patients with rheumatoid arthritis) HEIMER, R., AND ABRUZZO, J. L. (1972) Immunochemistry, 9, 921 (A latex test for the detection of human IgG aggregates and IgG anti-IgG antibody) - AND LEVIN, F. M. (1966) Ibid., 3, 1 (On the distribution of rheumatoid factors among the immunoglobulins) ,AND RUDD, E. (1963) Amer. J. Med., 35, 175 (Globulins resembling rheumatoid factor in serum of the aged) KUNKEL, H. G., MULLER-EBERHARD, H. J., FUDENBERG, H. H., AND ToMASI, T. B., JR. (1961) J. clin. Invest., 40, http://ard.bmj.com/ 117 (Gamma globulin complexes in rheumatoid arthritis and certain other conditions) MILGROM, F., AND WITEBSKY, E. (1960) J. Amer. med. Ass., 174, 56 (Studies on the rheumatoid and related serum factors. 1. Autoimmunization of rabbits with gamma globulin) PANUSH, R. S., BIANCO, N. E., AND SCHUR, P. H. (1971) Arthr. and Rheum., 14, 737 (Serum and synovial fluid IgG, IgA and IgM antigammaglobulins in rheumatoid arthritis) RISEMBERG, A., DE GOMEZ, A. G., AND RIFE, U. (1969) Ann. rheum. Dis., 28, 428 (Agglutinating factor in serum of patients with viral hepatitis) RoPES, M. W., BENNETr, G. A., COBB, S., JAcox, R., AND JESSAR, R. A. (1959) Ibid., 18, 49 (1958 revision of on October 2, 2021 by guest. Protected diagnostic criteria for rheumatoid arthritis) ROSE, H. M., RAGAN, C., PEARCE, E., AND LIPMAN, M. 0. (1948) Proc. Soc. exp. Biol. (N. Y.), 68, 1 (Differential agglutination of normal and sensitized sheep erythrocytes by sera of patients with rheumatoid arthritis) SINGER, J. M., AND PLOTZ, C. M. (1956) Amer. J. Med., 21, 888 (The latex fixation text 1. Application to the serologic diagnosis of rheumatoid arthritis) TAPANES, F. J., RAWSON, A. J., AND HOLLANDER, J. L. (1972) Arthr. and Rheum., 15,153 (Serum anti-immunoglobulins in psoriatic arthritis as compared with rheumatoid arthritis) TORRIGIANI, G., ANSELL, B. M., CHOWN, E. E. A., AND RoITr, I. M. (1969) Ann. rheum. Dis., 28, 424 (Raised IgG antiglobulin factors in Still's disease) AND Ro-rr, I. M. (1967) Ibid., 26, 334 (Antiglobulin factors in sera from patients with rheumatoid arthritis and normal subjects) ,LLOYD, K. N., AND CORBETT, M. (1970) Lancet, 1, 14 (Elevated IgG antiglobulins in patients with sero- negative rheumatoid arthritis) WILLIAMS, R. C., JR., AND KUNKEL, H. G. (1963) Proc. Soc. exp. Biol. (N. Y.), 112, 554 (Antibodies to rabbit y-globulin after immunizing with various preparations of autologous y-globulin) WINCHESTER, R. J., AGNELLO, V., AND KUNKEL, H. G. (1970) Clin. exp. Immunol., 6,689 (Gamma globulin complexes in synovial fluids of patients with rheumatoid arthritis: partial characterization and relationship to lowered complement levels)