Immunology Letters, 32 (1992) 31 - 34
Elsevier IMLET 01750
Isotypic and allotypic analysis with monoclonal antibodies and jacalin of 309 serum monoclonal IgA from French and Japanese myeloma patients P i e r r e A u c o u t u r i e r 1, Lucile Musset 2, Yoshihisa Itoh 3, Yue C h a u Ko 3, Christine Silvain I a n d J e a n Louis P r e u d ' h o m m e 1 ILaboratoire d'Immunologie et Immunopathologie, CNRS URA 1172, CHU, Poitiers, France; 2Laboratoire d'Immunochimie, Centre Hospitalier Pitid-Salpdtri~re, Paris, France; and 3Jichi Medical School, Department of Clinical Pathology, Minamikawachi-MachL Kawach#Gun, Tochigi-Ken, Japan
(Receivedand accepted 18 November 1991)
1. Summary The subclass and allotype distribution of serum monoclonal IgA from myeloma patients was determined by ELISA with monoclonal antibodies in two French and one Japanese laboratory. In addition, the French sera were tested for their reactivity with the lectin jacalin. No significant difference in the isotypic distribution between French and Japanese series could be demonstrated: r / h ratios were 0.99 and 1.17, and the IgA1 subclass accounted for 93.9°70 and 91070 of cases in the French and Japanese studies, respectively. Five out of 7 myeloma IgA2 from Japan and only one of the 12 IgA2 from France belonged to the A2m(2) allotype ( P < 0.01). All 219 IgA1 tested reacted with jacalin by immunoelectrophoresis (IEP), although with variable intensities. Among IgA2 proteins, only one (of the A2m(1) allotype) yielded a precipitating line with jacalin by IEP. Molecular analysis demonstrated that this protein was an IgAl-IgA2 hybrid bearing most of the A2m(1) epitopes. 2. Introduction H u m a n IgA includes two subclasses, IgA1 and Key words: IgA; Subclass; Isotype; Allotype; Jacalin Correspondence to: P. Aucouturier, Laboratoire d'Immunologie et Immunopathologie, CNRS URA 1172, CHU, BP 577, 86021 Poitiers, France.
IgA2, with well-defined structural and functional differences (reviewed in [1]). In addition, a gene conversion in the 0~2 constant region gene segment has resulted in two allotypic variants of IgA2, A2m(1) and A2m(2) [2]; these allotypes are differently distributed across to ethnic populations [3]. Although the bulk of IgA is secreted by the mucosal immune system, a significant amount is produced in the bone marrow and in various peripheral lymphoid organs, accounting for most serum IgA molecules. Contrary to secretory IgA, serum IgA is mostly monomeric, and IgAl subclass content is higher [4, 5]. The study of IgA subclasses and allotypes has long been hampered by the scarcity of potent specific antibodies, and there is only one published study of a large series of serum monoclonal IgA, which was performed using polyclonal antisera [6]. In the latter study, the IgA2 subclass was found to be under-represented by comparison with the normal subclass distribution of IgA secreting plasma cells. In a preliminary study of 176 IgG and 62 IgA myeloma proteins with monoclonal antibodies (mAbs), similar findings led us to suggest that the malignant transformation of B cell clones is not a random process [7]. Using specific monoclonal antibodies, we have now determined the isotypic and allotypic distribution of 309 serum myeloma IgA from French and Japanese patients. In addition, we examined the reactivity of the lectin jacalin with these monoclonal IgA by gel double diffusion.
0165-2478 / 92 / $ 5.00 © 1992 Elsevier SciencePublishers B.V. All rights reserved
31
3. Materials and Methods
3.2. Japanese study
3.1. French study
Sera from 78 patients (34 women and 44 men) with IgA myeloma were analysed by IEP. Subclasses were determined by a sandwich ELISA as previously described [10]. Briefly, a rabbit anti-human c~-chain antibody (Dako, Glostrup, Denmark) diluted 1:104 or 1:103 was coated onto microtitration plates; after incubation of the appropriately diluted samples or standard solutions, the bound IgA was revealed by incubation for 1.5 h at room temperature with monoclonal anti-subclass antibodies (anti-IgA1 clone 3C3-3A at 1.5/zg/ml and anti-IgA2 clone IF8-58 at 2/zg/ml, prepared in our laboratory [10]) and peroxidase-conjugated rabbit anti-mouse Ig (Dako). IgA2 allotypes were determined by immunoblotting after sodium dodecyl sulphate-polyacrylamide gel electrophoresis under non-reducing conditions.
Sera from 231 myeloma patients containing a monoclonal IgA at a concentration _>3 mg/ml were consecutively collected in two laboratories: 125 (59 women and 66 men) at the Piti6-Salp6tribre University Hospital, Paris, and 106 (48 women and 58 men) at the Poitiers University Hospital. Immunoglobulin (Ig) heavy chain classes, light chain types and jacalin reactivity were tested by standard immunoelectrophoretic analysis (IEP) using monospecific anti-lg sera (Silenus, Hawthorn, Australia) and jackfruit crude extract (JCE) from seeds collected in La R6union, as previously described [8]. When monoclonal IgA failed to precipitate with JCE concentrated 3 mg/ml, they were also further tested against 1 and 9 mg/ml of JCE. Serum IgA subclass levels were determined by a competitive immunoenzymatic assay according to a previously described method [9]. Briefly, polystyrene microtitration plates were coated with 200 tzl per well of a purified monoclonal IgA of the relevant subclass at 2/~g/ml; 100/xl of each serum diluted at 1:100 in 0.15 M NaC1/0.01 M phosphate buffer, pH 7.4, containing 2% bovine serum albumin (PBS-BSA) and 100 ~tl of the relevant mAb at the appropriate dilution were successively added and incubated for 1 h at 37 °C. After washing, bound antibodies were revealed with a peroxidase conjugated anti-mouse IgG antibody (Immunotech, Luminy, France) diluted 1: 15 000 in PBS-BSA for 1 h at room temperature. The anti-IgA1 mAb (clone M4D8, the Binding Site Ltd, Birmingham, U.K.) was used diluted 1:10000; the anti-IgA2 and anti-IgA2m(2) mAbs (clone HP6109, diluted 1:35 000 and clone 194-3.1, diluted 1: 15 000, respectively) were kindly provided by Dr. J. Mestecky (Birmingham, AL). IgA subclass concentrations in sample sera were calculated by comparison with a standard curve obtained with purified monoclonal IgA at 0 . 5 - 62.5/zg/ml. In a few cases, the subclasses and allotypes were confirmed by haemagglutination inhibition with polyclonal antisera (kindly performed by Dr. G. de Lange, Amsterdam).
32
3.3. Stat&tical analysis Percentages were compared using the
X2 t e s t .
4. Results
The isotypic and allotypic distributions of myeloma IgA in the Japanese and French series are shown in Table 1. The incidence of the different isotypes did not significantly differ according to the patients' geographical origin (Paris, Poitiers or Tochigi) and sex, with an IgA1 predominance and a total K/X ratio close to one. The r/X ratio was not significantly different from one subclass to another one. In contrast, the A2m(2) allotype was the most frequent in the Japanese patients (5 out of 7 IgA2) whereas A2m(1) was the predominant allotype of the monoclonal IgA2 in the French series (11 of 12 IgA2, P < 0.01). All 219 IgA1 tested reacted with jacalin by IEP. However, some monoclonal IgA yielded faint precipitation lines (see patient 2, Fig. 1), and only with one jacalin concentration. Among 12 IgA2 tested, only one reacted with jacalin. The complete primary structure of this protein was deduced from cDNA sequencing and proved to correspond to an IgAl-IgA2m(1) hybrid molecule containing the al hinge region and most IgA2m(1) epitopes (Silvain et al., in preparation).
TABLE 1 Isotypic and allotypic distribution of serum IgA myeloma proteins in France and Japan (data on the allotypes are not pooled since there is a difference between the two countries). IgA1
IgA2
IgA2m(1) IgA2m(2) Total
French study K K/~, ratio total (°70)
111 108 1.03 219 (94.8)
4 8 0.50 12 (5.2)
3 8 0.38 11 (4.6)
1 0 1 (0.5)
115 116 0.99 231
38 33 1.15 71 (91)
4 3 1.33 7 (9)
0 2 0 2 (2.5)
4 1 4 5 (6.4)
42 36 1.17 78
149 141 1.06 290 (93.9)
8 11 0.73 19 (6.1)
Japanese study ~, K/X ratio total (070)
Whole study K K/;~ ratio total (°7o)
157 152 1.03 309
Fig. 1. Immunoelectrophoretic analysis of two representative patients' sera, showing a weaker reactivity of patient 2's monoclonal IgA with jacalin (1 and 5 p.l of each serum deposited in external and central wells, respectively).
5. Discussion We report on the first large series of serum monoclonal IgA characterized with respect to isotypy and allotypy using mAbs, and on a comparison of their distribution in a European (French) and an Asian (Japanese) population. There was no signifi-
cant difference between the two countries with respect to the isotypes of the myeloma IgA, with a slight overrepresentation of the IgA1 subclass (94°70 of the IgA studied). Indeed, the observed subclass distribution is different from that expected on the basis of that of either normal bone marrow plasma cells (88o-/o IgA1 [4], P=0.015) or normal serum IgA molecules (83.7O-/o IgA1, mean of the 7 studies quoted in [1], P < 1 0 - 4 ) , thus confirming a previous study with polyclonal antisera [6]. As in the latter study, we found a K/~, ratio close to one. The emergence of myeloma requires previous stimulation of the transformed clones by antigens, possibly including autoantigens. There is indeed a strikingly high frequency of monoclonal Ig of the various classes, including IgA, which display antibody activities against autoantigens and even more often the idiotypes of autoantibodies (reviewed in ref. [11]) and at least certain such myeloma proteins with autoantibody specificity express highly mutated variable genes, in agreement with the hypothesis of an antigen-driven process [12]. The isotypic distribution of myeloma IgA reported here might reflect a special pattern of antigenic stimulation. That it is similar in Caucasian and Asian populations, which are partly exposed to different environmental stimuli, is possibly an argument for the role of autoantigens. In contrast to the isotypes, the allotypes of the monoclonal IgA differed, with a predominance of IgA2m(1) in France and of IgA2m(2) in Japan, which merely reflects the wellknown difference of the incidence of Ig alleles in these parts of the world [3]. Typing of the IgA subclasses with mAbs and with jacalin was in perfect agreement, except for a single IgA which turned out to be an IgA1/IgA2 hybrid, thus confirming the IgA1 specificity of the lectin in precipitating assays and its interest in immunochemistry [13]. It is worth noting that the reactivity of jacalin with different IgA1 varies (possibly reflecting a glycosylation heterogeneity of the hinge region) and that using several concentrations of jacalin may be required to determine IgA subclasses by immunoelectrophoretic analysis.
Acknowledgements We thank Dr. G. de Lange for her participation in the study. 33
References [1] Mestecky, J. and Russel, M. W. (1986) Monogr. Allergy 19, 277. [2] Flanagan, J . G . , Lefranc, M.P. and Rabbitts, T. H. (1984) Cell 36, 681. [3] De Lange, G. G. (1988) PhD thesis, London. [4] Skvaril, F. and Morell, A. (1974) Adv. Exp. Med. Biol. 45, 433. [5] Delacroix, D. L., Dive, C., Rambaud, J. C. and Vaerman, J. P. (1982) Immunology 47,383. [6] Vaerman, J. P., Hereman, J. F. and Laurell, C. B. (1968) Immunology 14, 425. [7] Aucouturier, P. and Preud'homme, J. L. (1987) Immunol. Lett. 16, 55. [8] Aucouturier, P., Mihaesco, E., Mihaesco, C. and
34
Preud'homme J. L. (1987) Mol. Immunol. 24, 503. [9] Aucouturier, P., Danon, F., Daveau, M., Guillou, B., Sabbah, A., Besson, J. and Preud'homme, J. L. (1984) J. Immunol. Methods 74, 151. [10] Ko, Y. C., ltoh, Y., Monden, S., Kobayashi, K. and Kawai, T. (1991) in: Frontiers of Mucosal Immunology, Vol. 1 (M. Tsuchiya, H. Nagura, T. Hibi and I. Moro, Eds.) pp. 215 218, Excerpta Medica, Amsterdam. [11] Bataille, R., Durie, B. G. M. and Klein, B. (1989) Autoimmunity 4, 181. [12] Davidson, A., Manheimer-Lory, A., Aranow, C., Peterson, R., Hannigan, N. and Diamond, B. (1989) J. Clin. Invest. 85, 1401. [13] Aucouturier, P., Pineau, N., Brugier, J. C., Mihaesco, E., Duarte, F., Skvaril, F. and Preud'homme, J. L. (1989) J. Clin. Lab. Anal. 3,244.
Elsevier IMLET 01750
Isotypic and allotypic analysis with monoclonal antibodies and jacalin of 309 serum monoclonal IgA from French and Japanese myeloma patients P i e r r e A u c o u t u r i e r 1, Lucile Musset 2, Yoshihisa Itoh 3, Yue C h a u Ko 3, Christine Silvain I a n d J e a n Louis P r e u d ' h o m m e 1 ILaboratoire d'Immunologie et Immunopathologie, CNRS URA 1172, CHU, Poitiers, France; 2Laboratoire d'Immunochimie, Centre Hospitalier Pitid-Salpdtri~re, Paris, France; and 3Jichi Medical School, Department of Clinical Pathology, Minamikawachi-MachL Kawach#Gun, Tochigi-Ken, Japan
(Receivedand accepted 18 November 1991)
1. Summary The subclass and allotype distribution of serum monoclonal IgA from myeloma patients was determined by ELISA with monoclonal antibodies in two French and one Japanese laboratory. In addition, the French sera were tested for their reactivity with the lectin jacalin. No significant difference in the isotypic distribution between French and Japanese series could be demonstrated: r / h ratios were 0.99 and 1.17, and the IgA1 subclass accounted for 93.9°70 and 91070 of cases in the French and Japanese studies, respectively. Five out of 7 myeloma IgA2 from Japan and only one of the 12 IgA2 from France belonged to the A2m(2) allotype ( P < 0.01). All 219 IgA1 tested reacted with jacalin by immunoelectrophoresis (IEP), although with variable intensities. Among IgA2 proteins, only one (of the A2m(1) allotype) yielded a precipitating line with jacalin by IEP. Molecular analysis demonstrated that this protein was an IgAl-IgA2 hybrid bearing most of the A2m(1) epitopes. 2. Introduction H u m a n IgA includes two subclasses, IgA1 and Key words: IgA; Subclass; Isotype; Allotype; Jacalin Correspondence to: P. Aucouturier, Laboratoire d'Immunologie et Immunopathologie, CNRS URA 1172, CHU, BP 577, 86021 Poitiers, France.
IgA2, with well-defined structural and functional differences (reviewed in [1]). In addition, a gene conversion in the 0~2 constant region gene segment has resulted in two allotypic variants of IgA2, A2m(1) and A2m(2) [2]; these allotypes are differently distributed across to ethnic populations [3]. Although the bulk of IgA is secreted by the mucosal immune system, a significant amount is produced in the bone marrow and in various peripheral lymphoid organs, accounting for most serum IgA molecules. Contrary to secretory IgA, serum IgA is mostly monomeric, and IgAl subclass content is higher [4, 5]. The study of IgA subclasses and allotypes has long been hampered by the scarcity of potent specific antibodies, and there is only one published study of a large series of serum monoclonal IgA, which was performed using polyclonal antisera [6]. In the latter study, the IgA2 subclass was found to be under-represented by comparison with the normal subclass distribution of IgA secreting plasma cells. In a preliminary study of 176 IgG and 62 IgA myeloma proteins with monoclonal antibodies (mAbs), similar findings led us to suggest that the malignant transformation of B cell clones is not a random process [7]. Using specific monoclonal antibodies, we have now determined the isotypic and allotypic distribution of 309 serum myeloma IgA from French and Japanese patients. In addition, we examined the reactivity of the lectin jacalin with these monoclonal IgA by gel double diffusion.
0165-2478 / 92 / $ 5.00 © 1992 Elsevier SciencePublishers B.V. All rights reserved
31
3. Materials and Methods
3.2. Japanese study
3.1. French study
Sera from 78 patients (34 women and 44 men) with IgA myeloma were analysed by IEP. Subclasses were determined by a sandwich ELISA as previously described [10]. Briefly, a rabbit anti-human c~-chain antibody (Dako, Glostrup, Denmark) diluted 1:104 or 1:103 was coated onto microtitration plates; after incubation of the appropriately diluted samples or standard solutions, the bound IgA was revealed by incubation for 1.5 h at room temperature with monoclonal anti-subclass antibodies (anti-IgA1 clone 3C3-3A at 1.5/zg/ml and anti-IgA2 clone IF8-58 at 2/zg/ml, prepared in our laboratory [10]) and peroxidase-conjugated rabbit anti-mouse Ig (Dako). IgA2 allotypes were determined by immunoblotting after sodium dodecyl sulphate-polyacrylamide gel electrophoresis under non-reducing conditions.
Sera from 231 myeloma patients containing a monoclonal IgA at a concentration _>3 mg/ml were consecutively collected in two laboratories: 125 (59 women and 66 men) at the Piti6-Salp6tribre University Hospital, Paris, and 106 (48 women and 58 men) at the Poitiers University Hospital. Immunoglobulin (Ig) heavy chain classes, light chain types and jacalin reactivity were tested by standard immunoelectrophoretic analysis (IEP) using monospecific anti-lg sera (Silenus, Hawthorn, Australia) and jackfruit crude extract (JCE) from seeds collected in La R6union, as previously described [8]. When monoclonal IgA failed to precipitate with JCE concentrated 3 mg/ml, they were also further tested against 1 and 9 mg/ml of JCE. Serum IgA subclass levels were determined by a competitive immunoenzymatic assay according to a previously described method [9]. Briefly, polystyrene microtitration plates were coated with 200 tzl per well of a purified monoclonal IgA of the relevant subclass at 2/~g/ml; 100/xl of each serum diluted at 1:100 in 0.15 M NaC1/0.01 M phosphate buffer, pH 7.4, containing 2% bovine serum albumin (PBS-BSA) and 100 ~tl of the relevant mAb at the appropriate dilution were successively added and incubated for 1 h at 37 °C. After washing, bound antibodies were revealed with a peroxidase conjugated anti-mouse IgG antibody (Immunotech, Luminy, France) diluted 1: 15 000 in PBS-BSA for 1 h at room temperature. The anti-IgA1 mAb (clone M4D8, the Binding Site Ltd, Birmingham, U.K.) was used diluted 1:10000; the anti-IgA2 and anti-IgA2m(2) mAbs (clone HP6109, diluted 1:35 000 and clone 194-3.1, diluted 1: 15 000, respectively) were kindly provided by Dr. J. Mestecky (Birmingham, AL). IgA subclass concentrations in sample sera were calculated by comparison with a standard curve obtained with purified monoclonal IgA at 0 . 5 - 62.5/zg/ml. In a few cases, the subclasses and allotypes were confirmed by haemagglutination inhibition with polyclonal antisera (kindly performed by Dr. G. de Lange, Amsterdam).
32
3.3. Stat&tical analysis Percentages were compared using the
X2 t e s t .
4. Results
The isotypic and allotypic distributions of myeloma IgA in the Japanese and French series are shown in Table 1. The incidence of the different isotypes did not significantly differ according to the patients' geographical origin (Paris, Poitiers or Tochigi) and sex, with an IgA1 predominance and a total K/X ratio close to one. The r/X ratio was not significantly different from one subclass to another one. In contrast, the A2m(2) allotype was the most frequent in the Japanese patients (5 out of 7 IgA2) whereas A2m(1) was the predominant allotype of the monoclonal IgA2 in the French series (11 of 12 IgA2, P < 0.01). All 219 IgA1 tested reacted with jacalin by IEP. However, some monoclonal IgA yielded faint precipitation lines (see patient 2, Fig. 1), and only with one jacalin concentration. Among 12 IgA2 tested, only one reacted with jacalin. The complete primary structure of this protein was deduced from cDNA sequencing and proved to correspond to an IgAl-IgA2m(1) hybrid molecule containing the al hinge region and most IgA2m(1) epitopes (Silvain et al., in preparation).
TABLE 1 Isotypic and allotypic distribution of serum IgA myeloma proteins in France and Japan (data on the allotypes are not pooled since there is a difference between the two countries). IgA1
IgA2
IgA2m(1) IgA2m(2) Total
French study K K/~, ratio total (°70)
111 108 1.03 219 (94.8)
4 8 0.50 12 (5.2)
3 8 0.38 11 (4.6)
1 0 1 (0.5)
115 116 0.99 231
38 33 1.15 71 (91)
4 3 1.33 7 (9)
0 2 0 2 (2.5)
4 1 4 5 (6.4)
42 36 1.17 78
149 141 1.06 290 (93.9)
8 11 0.73 19 (6.1)
Japanese study ~, K/X ratio total (070)
Whole study K K/;~ ratio total (°7o)
157 152 1.03 309
Fig. 1. Immunoelectrophoretic analysis of two representative patients' sera, showing a weaker reactivity of patient 2's monoclonal IgA with jacalin (1 and 5 p.l of each serum deposited in external and central wells, respectively).
5. Discussion We report on the first large series of serum monoclonal IgA characterized with respect to isotypy and allotypy using mAbs, and on a comparison of their distribution in a European (French) and an Asian (Japanese) population. There was no signifi-
cant difference between the two countries with respect to the isotypes of the myeloma IgA, with a slight overrepresentation of the IgA1 subclass (94°70 of the IgA studied). Indeed, the observed subclass distribution is different from that expected on the basis of that of either normal bone marrow plasma cells (88o-/o IgA1 [4], P=0.015) or normal serum IgA molecules (83.7O-/o IgA1, mean of the 7 studies quoted in [1], P < 1 0 - 4 ) , thus confirming a previous study with polyclonal antisera [6]. As in the latter study, we found a K/~, ratio close to one. The emergence of myeloma requires previous stimulation of the transformed clones by antigens, possibly including autoantigens. There is indeed a strikingly high frequency of monoclonal Ig of the various classes, including IgA, which display antibody activities against autoantigens and even more often the idiotypes of autoantibodies (reviewed in ref. [11]) and at least certain such myeloma proteins with autoantibody specificity express highly mutated variable genes, in agreement with the hypothesis of an antigen-driven process [12]. The isotypic distribution of myeloma IgA reported here might reflect a special pattern of antigenic stimulation. That it is similar in Caucasian and Asian populations, which are partly exposed to different environmental stimuli, is possibly an argument for the role of autoantigens. In contrast to the isotypes, the allotypes of the monoclonal IgA differed, with a predominance of IgA2m(1) in France and of IgA2m(2) in Japan, which merely reflects the wellknown difference of the incidence of Ig alleles in these parts of the world [3]. Typing of the IgA subclasses with mAbs and with jacalin was in perfect agreement, except for a single IgA which turned out to be an IgA1/IgA2 hybrid, thus confirming the IgA1 specificity of the lectin in precipitating assays and its interest in immunochemistry [13]. It is worth noting that the reactivity of jacalin with different IgA1 varies (possibly reflecting a glycosylation heterogeneity of the hinge region) and that using several concentrations of jacalin may be required to determine IgA subclasses by immunoelectrophoretic analysis.
Acknowledgements We thank Dr. G. de Lange for her participation in the study. 33
References [1] Mestecky, J. and Russel, M. W. (1986) Monogr. Allergy 19, 277. [2] Flanagan, J . G . , Lefranc, M.P. and Rabbitts, T. H. (1984) Cell 36, 681. [3] De Lange, G. G. (1988) PhD thesis, London. [4] Skvaril, F. and Morell, A. (1974) Adv. Exp. Med. Biol. 45, 433. [5] Delacroix, D. L., Dive, C., Rambaud, J. C. and Vaerman, J. P. (1982) Immunology 47,383. [6] Vaerman, J. P., Hereman, J. F. and Laurell, C. B. (1968) Immunology 14, 425. [7] Aucouturier, P. and Preud'homme, J. L. (1987) Immunol. Lett. 16, 55. [8] Aucouturier, P., Mihaesco, E., Mihaesco, C. and
34
Preud'homme J. L. (1987) Mol. Immunol. 24, 503. [9] Aucouturier, P., Danon, F., Daveau, M., Guillou, B., Sabbah, A., Besson, J. and Preud'homme, J. L. (1984) J. Immunol. Methods 74, 151. [10] Ko, Y. C., ltoh, Y., Monden, S., Kobayashi, K. and Kawai, T. (1991) in: Frontiers of Mucosal Immunology, Vol. 1 (M. Tsuchiya, H. Nagura, T. Hibi and I. Moro, Eds.) pp. 215 218, Excerpta Medica, Amsterdam. [11] Bataille, R., Durie, B. G. M. and Klein, B. (1989) Autoimmunity 4, 181. [12] Davidson, A., Manheimer-Lory, A., Aranow, C., Peterson, R., Hannigan, N. and Diamond, B. (1989) J. Clin. Invest. 85, 1401. [13] Aucouturier, P., Pineau, N., Brugier, J. C., Mihaesco, E., Duarte, F., Skvaril, F. and Preud'homme, J. L. (1989) J. Clin. Lab. Anal. 3,244.
