905
THE IMMUNOGLOBULIN KAPPA LIGHT CHAIN REPERTOIRE EXPRESSED IN THE SYNOVIUM OF A PATIENT WITH RHEUMATOID ARTHRITIS S O 0 KON LEE, S. LOUIS BRIDGES, JR., WILLIAM J. KOOPMAN, and HARRY W. SCHROEDER, JR. Objective. To analyze the nature of the B cell response in the synovial tissue of a patient with rheumatoid arthritis (RA). Specifically, we sought to determine if the pattern of immunoglobulin expression was consistent with polyclonal stimulation of B cells or an antigen-driven response. Methods. We generated an unrestricted complementary DNA (cDNA) library from the diseased synovium of a rheumatoid factor (RFkpositive patient with an 18-year history of RA. A random sample of K light chain recombinants was identified, and sequence analysis was performed. The variable domains were comPresented in part at the 54th Annual Meeting of the American College of Rheumatology, Seattle, WA, October 1990, and the Annual Meeting of the American Society for Clinical Investigation, Seattle, WA, May 1991. From the Divisions of Clinical Immunology and Rheumatology, and Developmental and Clinical Immunology, the Departments of Medicine and Microbiology, and the Comprehensive Cancer Center, University of Alabama at Birmingham. Dr. Koopman’s work was supported by a VA Merit award and NIH grants AR-03555, AI-18958, AI-18745, AR-20614, and DK-40117. Dr. Schroeder’s work was supported by NIH grants AR-03555, AI-23694, and AI-30879. Dr. Bridges is a T e m Gotthelf Lupus Research Scholar. Dr. Schroeder is an RJR Nabisco Research Scholar in Immunology. So0 Kon Lee, MD, PhD: Research Associate, Division of Clinical Immunology and Rheumatology; S . Louis Bridges, Jr., MD: Assistant Professor of Medicine, Division of Clinical Immunology and Rheumatology; William J. Koopman, MD: Howard L. Holley Professor of Medicine and Microbiology, Division of Clinical Immunology and Rheumatology; Harry W. Schroeder, Jr., MD, PhD: Assistant Professor of Medicine and Microbiology, Division of Developmental and Clinical Immunology, Comprehensive Cancer Center. Address reprint requests to Harry W. Schroeder, Jr., MD, PhD, Division of Developmental and Clinical Immunology, TI 362 UAB Station, University of Alabama at Birmingham, Birmingham, AL 35294. Submitted for publication October 17, 1991; accepted in revised form March 31, 1992. Arthritis and Rheumatism, Vol. 35, No. 8 (August 1992)
pared with an extensive database of germline and cDNA K sequences. Results. We found a light chain repertoire enriched for K transcripts containing 2 VK gene segments (Humkv325 and Humkv328) that are frequently associated with paraproteins expressing RF activity. Kappa variable domains from synovium contained numerous somatic mutations which resulted in frequent replacement of amino acids that encode the classic antigenbinding site. Unexpectedly, many of these K transcripts contained non-germline-encoded nucleotides (N reK The combination of gions) at the site of V ~ J joining. N-region addition and variation in the sites of VK-JK splicing generated unusually long complementaritydetermining region 3 regions and charged amino acids near the VK-JK splice site. Conclusion. The pattern of somatic mutations found in this patient sample supports the hypothesis that these synovium-derived plasma cells are the product of immunoglobulin receptor-dependent (i.e., antigendriven) selection. The extent of N-region addition raised the additional possibility that these antibodies derive from an unusual set of B lymphocytes that have escaped normal regulation.
Immunoglobulins are heterodimeric proteins composed of 2 heavy and 2 light chains (1,2). Each chain can be divided into a variable domain, which defines the antigen specificity of the molecule, and a constant domain, which encodes effector function. Variable domains contain 3 regions of sequence hypervariability (termed complementarity-determining regions [CDRs]) that are separated from each other by 4 regions of relatively constant sequence (termed framework regions [FRs]).
906
In humans, a K light chain variable domain is assembled by splicing together I of more than 85 variable (V) gene segments (50 potentially functional) to 1 of 5 joining (J) gene segments (1,3). The V Kgene segment encodes CDRs 1 and 2, whereas CDR 3 is the product of V-J joining. In the mature protein, the heavy and light chain CDRs are juxtaposed to form the antigen-binding site (2). Associated with exposure to antigen is a process of variable domain-specific hypermutation which can introduce non-germline-encoded base pairs into both heavy and light chains (1). This combination of genetic and somatic mechanisms allows the generation of a highly diverse repertoire of antigen-binding sites from a limited number of germline gene segments. Rheumatoid arthritis (RA) is an inflammatory disease involving the synovial membrane of the joint ( 4 3 . Plasma cell infiltration of the synovium is a characteristic feature of the disease. Indeed, local immunoglobulin production may reach levels equivalent to those in the spleen (6,7). This local immunoglobulin production is thought to participate in the pathogenesis of tissue injury in RA, although the exact mechanism of tissue damage is unclear (5). Rheumatoid factors (RF), antibodies which bind the Fc portion of IgG, are the most common autoantibodies in RA, comprising 7-25% of synovial immunoglobulin (IgM, IgG, and IgA) in patients who are seropositive for RF (7). Because elevated levels of RF are associated with a more aggressive form of RA (8,9), many studies of immunoglobulin production in RA have focused on analysis of RF. Major insights into the genetic composition of RF have been gleaned from the study of IgMK paraproteins with RF activity (10). Analysis of the light chains of these paraproteins has been particularly revealing. IgMK RF are enriched for the members of 1 of the 4 V K gene segment families, VKIII. In particular, 2 monoclonal antibodies, 6B6.6 (11) and 17.109 (12), that recognize 2 distinct VKIII-associated cross-reactive idiotypes (CRI) present on approximately two-thirds of IgMK RF paraproteins, have been generated (10). These CRI are encoded by the germline gene segments Humkv328 (6B6.6) (12-14) and Humkv325 (17.109) (15,16), although the precise epitopes are unknown. However, the contribution of the 6B6.6 and 17.109 CRI to the peripheral blood or synovial fluid R F repertoire is small (-1%) (1 1). Indeed, a predominant CRI either among RF or among total synovial antibodies expressed in RA has yet to be detected (17,18). Thus, the composition, the germline derivation, and,
LEE ET AL therefore, the nature of the stimuli leading to synovial immunoglobulin production, both RF and non-RF, remain unknown. To further define the nature of the immunoglobulin expressed in unmanipulated RA synovium, we used molecular cloning strategies to examine the K light chain repertoire expressed in the diseased synovium of a patient with longstanding, seropositive RA. We found a mutated set of light chain variable domains that contained unusual VK-JKjunctions and that were enriched for the presence of IgMK paraprotein RFassociated VKgene segments (10). These findings have important implications for our understanding of the forces that lead to the expansion of the antibody repertoire expressed in diseased synovium of patients with RA.
MATERIALS AND METHODS Patient. The synovium that was studied was obtained in August 1989, from a patient undergoing total knee arthroplasty at the University of Alabama at Birmingham Hospital. The patient, a 62-year-old white woman, had an 18-year history of seropositive RA that met the American College of Rheumatology (formerly, the American Rheumatism Association) criteria for classic disease (19). Laboratory evaluation revealed a serum IgM-RF level of 51 d m l . Twentyfour grams of synovial tissue was obtained at surgery. Cell isolation. Synovial tissue was processed immediately after surgery, as previously described (20). Briefly, the tissue was dissected free of cartilaginous material, minced into l-2-mm3 fragments, and washed in chilled minimal essential medium (MEM). Tissue fragments were placed into MEM containing Dispase (Boehringer Mannheim, Indianapolis, IN), and incubated at 37°C with constant agitation for 30 minutes. Medium containing dissociated synovial cells was collected by aspiration and placed into RPMI 1640 with fetal calf serum. The extraction procedure was repeated until sequential cell counts indicated insignificant cell yields. Synovial cells were washed, and cell viability was determined by trypan blue exclusion. The cell yield was -5 x 106/gm of tissue. Quantitation of synovial immunoglobulin and RF production. Total IgG, IgM, and IgA in culture supernatant were quantitated by solid-phase radioimmunoassay (RIA) (20). Measurement of culture supernatant IgM and IgA-RF was performed as previously described, using the quantitative solid-phase RIA (21). The presence of RF-associated CRI in supernatant and sera was determined by a direct-binding enzyme-linked immunosorbent assay, as previously described (22). Generation and analysis of 2 unrestricted complernentary DNA (cDNA) libraries. Using the guanidinium isothiocyanate technique (23), 257 pg of total RNA was isolated from 1 x lo8 dissociated synovial cells. By means of oligo d(T) cellulose affinity chromatography, 24 pg of poly(A)+ RNA was purified and then stored in 2-pg aliquots. A A g t l O
K
LIGHT CHAIN REPERTOIRE IN RA SYNOVIUM
cDNA library of 6 x lo6 recombinants was prepared from 4 pg of the poly(A)+ RNA, as previously described (24). Briefly, double-stranded cDNA was generated by the method of Gubler and Hoffman (25). The cDNA was then Eco RI methylated (New England Biolabs, Boston, MA), ligated to Eco RI linkers (Pharmacia, Piscataway, NJ), cut with Eco RI, and size selected over a Sepharose 2B column (Pharmacia). An aliquot of the size-selected cDNA was ligated into the Eco RI restriction site of AgtlO (26). In vitro packaged phage was plated at low density on C600 hfr bacteria and the resultant plaques screened with a modified 5‘ CK antisense oligonucleotide (5’-CCAGAATTCAACTGCTCATCAGATGGCGGGAAG-3’), using standard techniques (27). CK+ recombinants present in an aliquot of 120,000 recombinants derived from this first library were identified and numbered sequentially by plate. VK-containing inserts were subcloned from purified Agt 10 recombinants by polymerase chain reaction (PCR) amplification (28). We synthesized 2 amplification oligomers from AgtlO sequences that flank the Eco RI cloning site (“forward” S’GGTTTTCCCAGTCACGACGTT-3‘ and “reverse” 5’-ACAGGAAACAGCTATGACCA-3’).Each Agt 10 recombinant underwent parallel amplifications with the same antisense 5’ CKoligonucleotide used for screening and 1 of these 2 AgtlO primers. PCR conditions were as follows: 25 cycles of annealing at 37°C for 2 minutes, and extension at 72°C for 4 minutes, with final extension at 72°C for 10 minutes. In each case, 1 of the 2 amplifications gave rise to a CK-containing product, as identified by hybridization analysis with an internal antisense CKoligomer (5‘-TGAAGACAGATGGTGCAGCCA-3’). These amplified inserts were cloned into pUC19, and double-strand sequence was obtained by the dideoxynucleotide termination method (29). In one case, clones derived from the same original recombinant were independently amplified to determine the extent, if any, of sequence variation introduced by the PCR technique. No sequence variations were seen. A second cDNA library was generated from another 2-pg aliquot of the same synovial sample, and then 240,000 recombinant clones from this second library were screened with an unmodified CKoligonucleotide probe (5’-TGAAGACAGATGGTGCAGCCA-3’)complementary to the amino terminal sequence of CK. These CK+ clones were subsequently screened with oligonucleotides specific for the FR t of Hurnkv328 (5‘-GAAATAGTGATGACGCAGTCTCCAGCCA-3’),and for the FR 1 of Humkv325 (5’-GAAATTGTG’ITGACGCAGTCTCCAGGCA-3’). Although the theoretical melting temperature of these oligomers in 1 x SSC (0.15M NaCl and 0.015M sodium citrate) was 76”C, we used our sequenced synovial cDNAs as controls and determined the optimal melting temperature under our reaction conditions to be 42°C and 57”C, respectively (data not shown).
RESULTS Production of immunoglobulin, RF, and major CRI by dissociated synovial cells. Synovial cell cultures secreted predominantly IgG (IgG 197 ng/ml, IgA 23
907
ng/ml, IgM 12 ng/ml). Synovial IgM-RF and IgA-RF represented 16-21% of the total IgM and IgA production. As is typical for patients with RA, the 6B6.6 and 17.109 CRI were detected on
THE IMMUNOGLOBULIN KAPPA LIGHT CHAIN REPERTOIRE EXPRESSED IN THE SYNOVIUM OF A PATIENT WITH RHEUMATOID ARTHRITIS S O 0 KON LEE, S. LOUIS BRIDGES, JR., WILLIAM J. KOOPMAN, and HARRY W. SCHROEDER, JR. Objective. To analyze the nature of the B cell response in the synovial tissue of a patient with rheumatoid arthritis (RA). Specifically, we sought to determine if the pattern of immunoglobulin expression was consistent with polyclonal stimulation of B cells or an antigen-driven response. Methods. We generated an unrestricted complementary DNA (cDNA) library from the diseased synovium of a rheumatoid factor (RFkpositive patient with an 18-year history of RA. A random sample of K light chain recombinants was identified, and sequence analysis was performed. The variable domains were comPresented in part at the 54th Annual Meeting of the American College of Rheumatology, Seattle, WA, October 1990, and the Annual Meeting of the American Society for Clinical Investigation, Seattle, WA, May 1991. From the Divisions of Clinical Immunology and Rheumatology, and Developmental and Clinical Immunology, the Departments of Medicine and Microbiology, and the Comprehensive Cancer Center, University of Alabama at Birmingham. Dr. Koopman’s work was supported by a VA Merit award and NIH grants AR-03555, AI-18958, AI-18745, AR-20614, and DK-40117. Dr. Schroeder’s work was supported by NIH grants AR-03555, AI-23694, and AI-30879. Dr. Bridges is a T e m Gotthelf Lupus Research Scholar. Dr. Schroeder is an RJR Nabisco Research Scholar in Immunology. So0 Kon Lee, MD, PhD: Research Associate, Division of Clinical Immunology and Rheumatology; S . Louis Bridges, Jr., MD: Assistant Professor of Medicine, Division of Clinical Immunology and Rheumatology; William J. Koopman, MD: Howard L. Holley Professor of Medicine and Microbiology, Division of Clinical Immunology and Rheumatology; Harry W. Schroeder, Jr., MD, PhD: Assistant Professor of Medicine and Microbiology, Division of Developmental and Clinical Immunology, Comprehensive Cancer Center. Address reprint requests to Harry W. Schroeder, Jr., MD, PhD, Division of Developmental and Clinical Immunology, TI 362 UAB Station, University of Alabama at Birmingham, Birmingham, AL 35294. Submitted for publication October 17, 1991; accepted in revised form March 31, 1992. Arthritis and Rheumatism, Vol. 35, No. 8 (August 1992)
pared with an extensive database of germline and cDNA K sequences. Results. We found a light chain repertoire enriched for K transcripts containing 2 VK gene segments (Humkv325 and Humkv328) that are frequently associated with paraproteins expressing RF activity. Kappa variable domains from synovium contained numerous somatic mutations which resulted in frequent replacement of amino acids that encode the classic antigenbinding site. Unexpectedly, many of these K transcripts contained non-germline-encoded nucleotides (N reK The combination of gions) at the site of V ~ J joining. N-region addition and variation in the sites of VK-JK splicing generated unusually long complementaritydetermining region 3 regions and charged amino acids near the VK-JK splice site. Conclusion. The pattern of somatic mutations found in this patient sample supports the hypothesis that these synovium-derived plasma cells are the product of immunoglobulin receptor-dependent (i.e., antigendriven) selection. The extent of N-region addition raised the additional possibility that these antibodies derive from an unusual set of B lymphocytes that have escaped normal regulation.
Immunoglobulins are heterodimeric proteins composed of 2 heavy and 2 light chains (1,2). Each chain can be divided into a variable domain, which defines the antigen specificity of the molecule, and a constant domain, which encodes effector function. Variable domains contain 3 regions of sequence hypervariability (termed complementarity-determining regions [CDRs]) that are separated from each other by 4 regions of relatively constant sequence (termed framework regions [FRs]).
906
In humans, a K light chain variable domain is assembled by splicing together I of more than 85 variable (V) gene segments (50 potentially functional) to 1 of 5 joining (J) gene segments (1,3). The V Kgene segment encodes CDRs 1 and 2, whereas CDR 3 is the product of V-J joining. In the mature protein, the heavy and light chain CDRs are juxtaposed to form the antigen-binding site (2). Associated with exposure to antigen is a process of variable domain-specific hypermutation which can introduce non-germline-encoded base pairs into both heavy and light chains (1). This combination of genetic and somatic mechanisms allows the generation of a highly diverse repertoire of antigen-binding sites from a limited number of germline gene segments. Rheumatoid arthritis (RA) is an inflammatory disease involving the synovial membrane of the joint ( 4 3 . Plasma cell infiltration of the synovium is a characteristic feature of the disease. Indeed, local immunoglobulin production may reach levels equivalent to those in the spleen (6,7). This local immunoglobulin production is thought to participate in the pathogenesis of tissue injury in RA, although the exact mechanism of tissue damage is unclear (5). Rheumatoid factors (RF), antibodies which bind the Fc portion of IgG, are the most common autoantibodies in RA, comprising 7-25% of synovial immunoglobulin (IgM, IgG, and IgA) in patients who are seropositive for RF (7). Because elevated levels of RF are associated with a more aggressive form of RA (8,9), many studies of immunoglobulin production in RA have focused on analysis of RF. Major insights into the genetic composition of RF have been gleaned from the study of IgMK paraproteins with RF activity (10). Analysis of the light chains of these paraproteins has been particularly revealing. IgMK RF are enriched for the members of 1 of the 4 V K gene segment families, VKIII. In particular, 2 monoclonal antibodies, 6B6.6 (11) and 17.109 (12), that recognize 2 distinct VKIII-associated cross-reactive idiotypes (CRI) present on approximately two-thirds of IgMK RF paraproteins, have been generated (10). These CRI are encoded by the germline gene segments Humkv328 (6B6.6) (12-14) and Humkv325 (17.109) (15,16), although the precise epitopes are unknown. However, the contribution of the 6B6.6 and 17.109 CRI to the peripheral blood or synovial fluid R F repertoire is small (-1%) (1 1). Indeed, a predominant CRI either among RF or among total synovial antibodies expressed in RA has yet to be detected (17,18). Thus, the composition, the germline derivation, and,
LEE ET AL therefore, the nature of the stimuli leading to synovial immunoglobulin production, both RF and non-RF, remain unknown. To further define the nature of the immunoglobulin expressed in unmanipulated RA synovium, we used molecular cloning strategies to examine the K light chain repertoire expressed in the diseased synovium of a patient with longstanding, seropositive RA. We found a mutated set of light chain variable domains that contained unusual VK-JKjunctions and that were enriched for the presence of IgMK paraprotein RFassociated VKgene segments (10). These findings have important implications for our understanding of the forces that lead to the expansion of the antibody repertoire expressed in diseased synovium of patients with RA.
MATERIALS AND METHODS Patient. The synovium that was studied was obtained in August 1989, from a patient undergoing total knee arthroplasty at the University of Alabama at Birmingham Hospital. The patient, a 62-year-old white woman, had an 18-year history of seropositive RA that met the American College of Rheumatology (formerly, the American Rheumatism Association) criteria for classic disease (19). Laboratory evaluation revealed a serum IgM-RF level of 51 d m l . Twentyfour grams of synovial tissue was obtained at surgery. Cell isolation. Synovial tissue was processed immediately after surgery, as previously described (20). Briefly, the tissue was dissected free of cartilaginous material, minced into l-2-mm3 fragments, and washed in chilled minimal essential medium (MEM). Tissue fragments were placed into MEM containing Dispase (Boehringer Mannheim, Indianapolis, IN), and incubated at 37°C with constant agitation for 30 minutes. Medium containing dissociated synovial cells was collected by aspiration and placed into RPMI 1640 with fetal calf serum. The extraction procedure was repeated until sequential cell counts indicated insignificant cell yields. Synovial cells were washed, and cell viability was determined by trypan blue exclusion. The cell yield was -5 x 106/gm of tissue. Quantitation of synovial immunoglobulin and RF production. Total IgG, IgM, and IgA in culture supernatant were quantitated by solid-phase radioimmunoassay (RIA) (20). Measurement of culture supernatant IgM and IgA-RF was performed as previously described, using the quantitative solid-phase RIA (21). The presence of RF-associated CRI in supernatant and sera was determined by a direct-binding enzyme-linked immunosorbent assay, as previously described (22). Generation and analysis of 2 unrestricted complernentary DNA (cDNA) libraries. Using the guanidinium isothiocyanate technique (23), 257 pg of total RNA was isolated from 1 x lo8 dissociated synovial cells. By means of oligo d(T) cellulose affinity chromatography, 24 pg of poly(A)+ RNA was purified and then stored in 2-pg aliquots. A A g t l O
K
LIGHT CHAIN REPERTOIRE IN RA SYNOVIUM
cDNA library of 6 x lo6 recombinants was prepared from 4 pg of the poly(A)+ RNA, as previously described (24). Briefly, double-stranded cDNA was generated by the method of Gubler and Hoffman (25). The cDNA was then Eco RI methylated (New England Biolabs, Boston, MA), ligated to Eco RI linkers (Pharmacia, Piscataway, NJ), cut with Eco RI, and size selected over a Sepharose 2B column (Pharmacia). An aliquot of the size-selected cDNA was ligated into the Eco RI restriction site of AgtlO (26). In vitro packaged phage was plated at low density on C600 hfr bacteria and the resultant plaques screened with a modified 5‘ CK antisense oligonucleotide (5’-CCAGAATTCAACTGCTCATCAGATGGCGGGAAG-3’), using standard techniques (27). CK+ recombinants present in an aliquot of 120,000 recombinants derived from this first library were identified and numbered sequentially by plate. VK-containing inserts were subcloned from purified Agt 10 recombinants by polymerase chain reaction (PCR) amplification (28). We synthesized 2 amplification oligomers from AgtlO sequences that flank the Eco RI cloning site (“forward” S’GGTTTTCCCAGTCACGACGTT-3‘ and “reverse” 5’-ACAGGAAACAGCTATGACCA-3’).Each Agt 10 recombinant underwent parallel amplifications with the same antisense 5’ CKoligonucleotide used for screening and 1 of these 2 AgtlO primers. PCR conditions were as follows: 25 cycles of annealing at 37°C for 2 minutes, and extension at 72°C for 4 minutes, with final extension at 72°C for 10 minutes. In each case, 1 of the 2 amplifications gave rise to a CK-containing product, as identified by hybridization analysis with an internal antisense CKoligomer (5‘-TGAAGACAGATGGTGCAGCCA-3’). These amplified inserts were cloned into pUC19, and double-strand sequence was obtained by the dideoxynucleotide termination method (29). In one case, clones derived from the same original recombinant were independently amplified to determine the extent, if any, of sequence variation introduced by the PCR technique. No sequence variations were seen. A second cDNA library was generated from another 2-pg aliquot of the same synovial sample, and then 240,000 recombinant clones from this second library were screened with an unmodified CKoligonucleotide probe (5’-TGAAGACAGATGGTGCAGCCA-3’)complementary to the amino terminal sequence of CK. These CK+ clones were subsequently screened with oligonucleotides specific for the FR t of Hurnkv328 (5‘-GAAATAGTGATGACGCAGTCTCCAGCCA-3’),and for the FR 1 of Humkv325 (5’-GAAATTGTG’ITGACGCAGTCTCCAGGCA-3’). Although the theoretical melting temperature of these oligomers in 1 x SSC (0.15M NaCl and 0.015M sodium citrate) was 76”C, we used our sequenced synovial cDNAs as controls and determined the optimal melting temperature under our reaction conditions to be 42°C and 57”C, respectively (data not shown).
RESULTS Production of immunoglobulin, RF, and major CRI by dissociated synovial cells. Synovial cell cultures secreted predominantly IgG (IgG 197 ng/ml, IgA 23
907
ng/ml, IgM 12 ng/ml). Synovial IgM-RF and IgA-RF represented 16-21% of the total IgM and IgA production. As is typical for patients with RA, the 6B6.6 and 17.109 CRI were detected on
