Vox Sang 1991;61:196204
0 1991 S. Karger AG, Basel 0042-900719Y0613-0196 $2.7510
Polyreactivity of Human Monoclonal Antibodies: Human Anti-Rh Monoclonal Antibodies of IgM Isotype Are Fkequently Polyreactive A . Blanchera, E Roubineta, E Oksmana, I: Ternynckb, H . Broly', J. Chevaleyred, G. Vezond,J . D U C O S ~ 'Laboratoire Central d'Immunologie, C.H.U. Purpan, Toulouse; bInstitut Pasteur, Unit6 d'Immunocytochimie, Paris; 'C.R.T.S., Lille: dC.R.T.S., Bordeaux, France
Abstract. The specific aim of this study was to characterize human anti-Rh monoclonal antibodies cross-reacting with self-antigens. We studied supernatants from man-mouse hybridomas and from lymphoblastoid cell lines. Man-mouse hybridomas were established by fusion of peripheral blood lymphocytes from healthy individuals recently immunized against Rh alloantigens, with mouse myeloma (or man-mouse heteromyeloma) cell lines. Lymphoblastoid cell lines were produced by Epstein-Barr virus induction of lymphocytes from identical sources. Of the 55 monoclonal alloantibodies studied, 11also reacted with intracellular self-antigens as demonstrated by immunofluorescence assay on cryostat sections of human tissues. This cross-reactivity was mainly a property of monoclonal alloantibodies belonging to the IgM isotype (among the 11cross-reacting mAbs 10 were IgM). The cross-reactivities of these monoclonal antibodies were ascertained by absorption of alloreacting antibodies with red blood cells. Similar results were obtained on a panel of purified cellular antigens by ELISA. The results confirm that during an immune response against a foreign antigen (alloantigen), B cells that produce polyreactive antibodies are not excluded from the pool of responding cells. Therefore, polyreactive autoantibodies present in sera from healthy individuals may be the result of an immune response against foreign antigens.
Introduction A variety of studies have documented the presence of autoreactive B cell clones in healthy humans [l-31, mice [2, 41 and rats [5]. Hybridomas derived from preimmune healthy mice were shown to secrete monoclonal autoantibodies [6, 71. Most of these natural monoclonal antibodies (mAbs) are polyspecific [&lo] ; they react with more than two self- or nonself-antigens. In mice, all of these polyreactive natural autoantibodies were found belong to the IgM isotype [4], even though certain antibodies of IgG isotype are polyreactive [ll]. The presence of natural autoantibodies in hyperimmunized mice has also been demonstrated [12]. It has been hypothesized that cells carrying polyspecific natural antibodies as receptors are capable of clonal expansion after antigenic stimulation and that cells producing highly specific antibodies for epitopes of a given antigen are derived from them by somatic mutations. Thus polyreactive B cells are permanently renewed by medullary lymphopoiesis and could serve as precursors for highly specific B cells recruited during peripheral immunization [13].
Thus, it should be theoretically possible to isolate clones representative of an intermediate stage of the immune response which would produce antibodies specific for the immunizing antigen and simultaneously conserve part of their primordial auto- and polyreactivity. The aim of the present study was to verify whether, among alloreactive human monoclonal antibodies, some conserve their polyreactivity and yet cross-react with self-antigens. To achieve this, we chose the immune response against the Rh antigens because of their dependence on T-cell regulation rather than the usual T-independent responses to carbohydrate antigens [14]. Also, the frequency of naturally occurring antibodies to such protein antigens is extremely low [15, 161. We herein describe a study of 55 anti-Rh human mAbs. Their cross-reactivity with a panel of cellular antigens was studied by indirect immunofluorescence assay on cryostat tissue sections and by ELISA. We found that monoclonal anti-Rh antibodies of IgM isotype frequently cross-react with self-intracellular antigens. On the contrary, anti-Rh mAbs of IgG isotype exceptionally exhibit such cross-reactivity. The polyreactivity of
Polyreactivity of Human Monoclonal Anti-D Antibodies
anti-Rh of IgM isotype was compared with that of human IgM unreactive with Rh antigens. The data demonstrate that B cells producing polyreactive antibodies are not excluded from the responding pool in immune response against foreign antigens.
Material and Methods Human Monoclonal Anti-Rh Antibodies Anti-Rh mAbs (tablesl, 2) were produced in four laboratories: Toulouse (this study), Lille (this study), Bordeaux (this study), Bristol [17] from 25 donors. We studied 51 mAbs with specificity for Rh D and 4 mAbs with specificity for Rh E. Among the 55 mAbs studied, 39 were of IgG isotype and the remaining 13 of IgM isotype. The mAbs were derived either from heterohybridoma cell lines or from Epstein-Barr virus (EBV)-transformed lymphoblastoid cell lines (see tables 1and 2). Cell Fusion Peripheral blood mononuclear cells (PBMC) from individuals immunized against Rh antigen were isolated on MSL (Eurobio) and fused as previously described [18] with cells of murine myeloma cell line X 63 Ag 8.653 [I91or with cell of heteromyeloma cell lines derived from X 63 Ag 8.653. Transformation of B Lymphocytes by EBV PBMC were separated on MSL (Eurobio), incubated with EBV (1 ml culture supernatant from filtered mycoplasma-free B95-8 cell line per lo7 PBMC) at 37°C for 1 h and washed. Aliquots were plated at 0.5 . lo6cellslml in 100-pI wells in culture medium RPMI 1640 +lo% fetal calf serum (FCS) and O.Spg/ml ciclosporin A (Sandoz). The cultures were thereafter maintained by weekly renewal of half of the medium. Cloning Cells lines were cloned by limiting dilution. Cells were plated with or without feeder cells (mouse thymocytes). Cultures were fed once a week with RPMI 1640 +20% FCS or lymphoblastoid cell lines for with ISCOVE's medium +20% FCS for hybridomas. Gm Allotype Determination Sensitized red cells with each mAb were resuspended at 3% in PBS. Equal volume (50pl) of the red-cell suspensions and Gm reagents (human anti-sera kindly provided by J.M. Dugoujon, CNRS 8291, Toulouse) were incubated in round-bottomed wells of microtiter plates for 2 h at room temperature with gentle alternative stirring (maximum slope 30", frequency 2 min). The plates were placed on a 30" slope allowing the reading of agglutination. Serology The supernatants were tested against a panel of D-positive and D-negative red cells using the conventional saline, albumin displacement, two-stage enzyme (bromelin and papain) or indirect antiglobulin techniques. In all cases, 3% cell suspensions in glass tubes at a 1:l ratio of supernatantkells were used. Titrations of anti-D mAbs in microtiter plates (round-bottomed wells) were performed by adding 50 pI of 3% 0 RIRlpapain-treated red cells to 50 PI of doubling dilutions of supernatants in PBS.
197
After incubation at room temperature for 60 min the plates were read macroscopically. The titer was taken to be the reciprocal of the final dilution at which agglutination was observed. Anti-E mAbs were studied on a panel of E-positive and E-negative red cells and were titrated on 0 R1R2(CDelcDE) papain-treated red cells. Screening Human Immunoglobulins in the Culture Supernatants by Immunoenzyme Assay The plates were coated overnight (4°C) with rabbit antibodies directed against human x and h chains (DAKO) diluted U5,OOO in carbonate buffer (0.1M ; pH 9.6). Culture supernatants diluted U5 in phosphate-buffered saline (PBS) containing 3% bovine serum albumin (BSA) and 0.1% fieen-20, were distributed into the wells (100pY well), as well as serial dilutions of a reference human serum (Behring). After a 3-h incubation at 37", the plates were extensively washed. For each well, 100 pl of alkaline phosphatase (AP)-conjugated rabbit antihuman Ig antibodies were added (Behring), either anti-p chain-AP diluted U100 or anti-y chain diluted U160 in PBS-Tween). After a 90-min incubation, plates were again washed extensively. The reaction was revealed by the addition of 200 p1 of a substrate solution (p-nitrophenyl phosphate (PNP): 1mg/ml in 1M diethanolamine buffer, pH9.8, supplemented with 0.5 mM MgCI2). The optical density (OD) was read at 405 nm using a Titertek multiskan (Flow laboratories). Human Ig dosages in supernatants were confirmed by immunonephelometry (Beckman). The limit of detection for human IgG and IgM was 4 pg/ml. Indirect Immunofluorescence Assay (IFA) The search for antitissue autoantibodies was carried out using thin sections (5 pm) of frozen tissues. Fresh human tissues were obtained at biopsy procedures or surgical treatment of kidney donors and included skin, cerebellum, thyroid, peripheral nerve and temporal artery. The following fresh animal tissues were also used: mouse liver and stomach, rat pylorus and kidney, rabbit parotid, and calf adrenal gland. Tissue sections were fixed on glass slides by immersion in cold acetone (-20°C) for 2 min. Undiluted culture supernatants were incubated with each tissue for 30 min at 37°C. After careful washing in PBS, the sections were incubated for 30 min at 37°C with goat antibodies against human Ig p chain conjugated to fluorescein diluted U10 (Kallestadt, Biolyon, France). Tissue preparations were washed and mounted under a cover slip with buffered glycerin and read using a Leitz fluorescence microscope. Study of Autoantibody Activity by ELlSA Noncompetitive Enzyme Immunoassay. Polystyrene flat-bottom NUNC plates were coated with various antigens as previously described [l2]. Balblc muscle actin [20] and myosin [21] and pig brain tubulin [22], human cytokeratins [23] and thyroglobulin [24] were prepared according to classical methods. Native double-stranded DNA and ovalbumin (OVA) were purchased from Sigma Chemical (St. Louis, Mo., USA) and BSA from IBF (Villeneuve-la-Garenne, France). All antigens were tested by sodium-dodecyl-sulfate-polyacrylamide gel electrophoresis, and each of them was found to be free of contamination by the others. Trinitrophenyl BSA (TNP-BSA) and trinitrophenyl OVA (TNP-OVA) were prepared according to Little and Eisen [25]. Nitroiodophenyl-BSA (NIP-BSA) was kindly donated by Dr. A. Coutinho (Institut Pasteur, Paris, France). The antigen-coated plates were thoroughly washed with PBS containing 0.1% fieen-20 (PBS-Tween) and were incubated with each
198
Blancher/Roubinet/O ksmadTern y nck/Broly/ChevaleyreiVezon/Ducos
Table 1. IgG anti-Rh mAbs: reactivity on tissues C
Allotype
d
Titer'
IgG2
LCL LCL LCL H LCL H
D D D D D D D D D D D
13 11 14 19 13 13 11 9 12 13 13 10
0 1991 S. Karger AG, Basel 0042-900719Y0613-0196 $2.7510
Polyreactivity of Human Monoclonal Antibodies: Human Anti-Rh Monoclonal Antibodies of IgM Isotype Are Fkequently Polyreactive A . Blanchera, E Roubineta, E Oksmana, I: Ternynckb, H . Broly', J. Chevaleyred, G. Vezond,J . D U C O S ~ 'Laboratoire Central d'Immunologie, C.H.U. Purpan, Toulouse; bInstitut Pasteur, Unit6 d'Immunocytochimie, Paris; 'C.R.T.S., Lille: dC.R.T.S., Bordeaux, France
Abstract. The specific aim of this study was to characterize human anti-Rh monoclonal antibodies cross-reacting with self-antigens. We studied supernatants from man-mouse hybridomas and from lymphoblastoid cell lines. Man-mouse hybridomas were established by fusion of peripheral blood lymphocytes from healthy individuals recently immunized against Rh alloantigens, with mouse myeloma (or man-mouse heteromyeloma) cell lines. Lymphoblastoid cell lines were produced by Epstein-Barr virus induction of lymphocytes from identical sources. Of the 55 monoclonal alloantibodies studied, 11also reacted with intracellular self-antigens as demonstrated by immunofluorescence assay on cryostat sections of human tissues. This cross-reactivity was mainly a property of monoclonal alloantibodies belonging to the IgM isotype (among the 11cross-reacting mAbs 10 were IgM). The cross-reactivities of these monoclonal antibodies were ascertained by absorption of alloreacting antibodies with red blood cells. Similar results were obtained on a panel of purified cellular antigens by ELISA. The results confirm that during an immune response against a foreign antigen (alloantigen), B cells that produce polyreactive antibodies are not excluded from the pool of responding cells. Therefore, polyreactive autoantibodies present in sera from healthy individuals may be the result of an immune response against foreign antigens.
Introduction A variety of studies have documented the presence of autoreactive B cell clones in healthy humans [l-31, mice [2, 41 and rats [5]. Hybridomas derived from preimmune healthy mice were shown to secrete monoclonal autoantibodies [6, 71. Most of these natural monoclonal antibodies (mAbs) are polyspecific [&lo] ; they react with more than two self- or nonself-antigens. In mice, all of these polyreactive natural autoantibodies were found belong to the IgM isotype [4], even though certain antibodies of IgG isotype are polyreactive [ll]. The presence of natural autoantibodies in hyperimmunized mice has also been demonstrated [12]. It has been hypothesized that cells carrying polyspecific natural antibodies as receptors are capable of clonal expansion after antigenic stimulation and that cells producing highly specific antibodies for epitopes of a given antigen are derived from them by somatic mutations. Thus polyreactive B cells are permanently renewed by medullary lymphopoiesis and could serve as precursors for highly specific B cells recruited during peripheral immunization [13].
Thus, it should be theoretically possible to isolate clones representative of an intermediate stage of the immune response which would produce antibodies specific for the immunizing antigen and simultaneously conserve part of their primordial auto- and polyreactivity. The aim of the present study was to verify whether, among alloreactive human monoclonal antibodies, some conserve their polyreactivity and yet cross-react with self-antigens. To achieve this, we chose the immune response against the Rh antigens because of their dependence on T-cell regulation rather than the usual T-independent responses to carbohydrate antigens [14]. Also, the frequency of naturally occurring antibodies to such protein antigens is extremely low [15, 161. We herein describe a study of 55 anti-Rh human mAbs. Their cross-reactivity with a panel of cellular antigens was studied by indirect immunofluorescence assay on cryostat tissue sections and by ELISA. We found that monoclonal anti-Rh antibodies of IgM isotype frequently cross-react with self-intracellular antigens. On the contrary, anti-Rh mAbs of IgG isotype exceptionally exhibit such cross-reactivity. The polyreactivity of
Polyreactivity of Human Monoclonal Anti-D Antibodies
anti-Rh of IgM isotype was compared with that of human IgM unreactive with Rh antigens. The data demonstrate that B cells producing polyreactive antibodies are not excluded from the responding pool in immune response against foreign antigens.
Material and Methods Human Monoclonal Anti-Rh Antibodies Anti-Rh mAbs (tablesl, 2) were produced in four laboratories: Toulouse (this study), Lille (this study), Bordeaux (this study), Bristol [17] from 25 donors. We studied 51 mAbs with specificity for Rh D and 4 mAbs with specificity for Rh E. Among the 55 mAbs studied, 39 were of IgG isotype and the remaining 13 of IgM isotype. The mAbs were derived either from heterohybridoma cell lines or from Epstein-Barr virus (EBV)-transformed lymphoblastoid cell lines (see tables 1and 2). Cell Fusion Peripheral blood mononuclear cells (PBMC) from individuals immunized against Rh antigen were isolated on MSL (Eurobio) and fused as previously described [18] with cells of murine myeloma cell line X 63 Ag 8.653 [I91or with cell of heteromyeloma cell lines derived from X 63 Ag 8.653. Transformation of B Lymphocytes by EBV PBMC were separated on MSL (Eurobio), incubated with EBV (1 ml culture supernatant from filtered mycoplasma-free B95-8 cell line per lo7 PBMC) at 37°C for 1 h and washed. Aliquots were plated at 0.5 . lo6cellslml in 100-pI wells in culture medium RPMI 1640 +lo% fetal calf serum (FCS) and O.Spg/ml ciclosporin A (Sandoz). The cultures were thereafter maintained by weekly renewal of half of the medium. Cloning Cells lines were cloned by limiting dilution. Cells were plated with or without feeder cells (mouse thymocytes). Cultures were fed once a week with RPMI 1640 +20% FCS or lymphoblastoid cell lines for with ISCOVE's medium +20% FCS for hybridomas. Gm Allotype Determination Sensitized red cells with each mAb were resuspended at 3% in PBS. Equal volume (50pl) of the red-cell suspensions and Gm reagents (human anti-sera kindly provided by J.M. Dugoujon, CNRS 8291, Toulouse) were incubated in round-bottomed wells of microtiter plates for 2 h at room temperature with gentle alternative stirring (maximum slope 30", frequency 2 min). The plates were placed on a 30" slope allowing the reading of agglutination. Serology The supernatants were tested against a panel of D-positive and D-negative red cells using the conventional saline, albumin displacement, two-stage enzyme (bromelin and papain) or indirect antiglobulin techniques. In all cases, 3% cell suspensions in glass tubes at a 1:l ratio of supernatantkells were used. Titrations of anti-D mAbs in microtiter plates (round-bottomed wells) were performed by adding 50 pI of 3% 0 RIRlpapain-treated red cells to 50 PI of doubling dilutions of supernatants in PBS.
197
After incubation at room temperature for 60 min the plates were read macroscopically. The titer was taken to be the reciprocal of the final dilution at which agglutination was observed. Anti-E mAbs were studied on a panel of E-positive and E-negative red cells and were titrated on 0 R1R2(CDelcDE) papain-treated red cells. Screening Human Immunoglobulins in the Culture Supernatants by Immunoenzyme Assay The plates were coated overnight (4°C) with rabbit antibodies directed against human x and h chains (DAKO) diluted U5,OOO in carbonate buffer (0.1M ; pH 9.6). Culture supernatants diluted U5 in phosphate-buffered saline (PBS) containing 3% bovine serum albumin (BSA) and 0.1% fieen-20, were distributed into the wells (100pY well), as well as serial dilutions of a reference human serum (Behring). After a 3-h incubation at 37", the plates were extensively washed. For each well, 100 pl of alkaline phosphatase (AP)-conjugated rabbit antihuman Ig antibodies were added (Behring), either anti-p chain-AP diluted U100 or anti-y chain diluted U160 in PBS-Tween). After a 90-min incubation, plates were again washed extensively. The reaction was revealed by the addition of 200 p1 of a substrate solution (p-nitrophenyl phosphate (PNP): 1mg/ml in 1M diethanolamine buffer, pH9.8, supplemented with 0.5 mM MgCI2). The optical density (OD) was read at 405 nm using a Titertek multiskan (Flow laboratories). Human Ig dosages in supernatants were confirmed by immunonephelometry (Beckman). The limit of detection for human IgG and IgM was 4 pg/ml. Indirect Immunofluorescence Assay (IFA) The search for antitissue autoantibodies was carried out using thin sections (5 pm) of frozen tissues. Fresh human tissues were obtained at biopsy procedures or surgical treatment of kidney donors and included skin, cerebellum, thyroid, peripheral nerve and temporal artery. The following fresh animal tissues were also used: mouse liver and stomach, rat pylorus and kidney, rabbit parotid, and calf adrenal gland. Tissue sections were fixed on glass slides by immersion in cold acetone (-20°C) for 2 min. Undiluted culture supernatants were incubated with each tissue for 30 min at 37°C. After careful washing in PBS, the sections were incubated for 30 min at 37°C with goat antibodies against human Ig p chain conjugated to fluorescein diluted U10 (Kallestadt, Biolyon, France). Tissue preparations were washed and mounted under a cover slip with buffered glycerin and read using a Leitz fluorescence microscope. Study of Autoantibody Activity by ELlSA Noncompetitive Enzyme Immunoassay. Polystyrene flat-bottom NUNC plates were coated with various antigens as previously described [l2]. Balblc muscle actin [20] and myosin [21] and pig brain tubulin [22], human cytokeratins [23] and thyroglobulin [24] were prepared according to classical methods. Native double-stranded DNA and ovalbumin (OVA) were purchased from Sigma Chemical (St. Louis, Mo., USA) and BSA from IBF (Villeneuve-la-Garenne, France). All antigens were tested by sodium-dodecyl-sulfate-polyacrylamide gel electrophoresis, and each of them was found to be free of contamination by the others. Trinitrophenyl BSA (TNP-BSA) and trinitrophenyl OVA (TNP-OVA) were prepared according to Little and Eisen [25]. Nitroiodophenyl-BSA (NIP-BSA) was kindly donated by Dr. A. Coutinho (Institut Pasteur, Paris, France). The antigen-coated plates were thoroughly washed with PBS containing 0.1% fieen-20 (PBS-Tween) and were incubated with each
198
Blancher/Roubinet/O ksmadTern y nck/Broly/ChevaleyreiVezon/Ducos
Table 1. IgG anti-Rh mAbs: reactivity on tissues C
Allotype
d
Titer'
IgG2
LCL LCL LCL H LCL H
D D D D D D D D D D D
13 11 14 19 13 13 11 9 12 13 13 10
