Th2-oriented profile of male offspring T cells present in women with systemic sclerosis and reactive with maternal major histocompatibility complex antigens.код для вставкиСкачать
ARTHRITIS & RHEUMATISM Vol. 46, No. 2, February 2002, pp 445–450 DOI 10.1002/art.10049 © 2002, American College of Rheumatology Published by Wiley-Liss, Inc. Th2-Oriented Profile of Male Offspring T Cells Present in Women With Systemic Sclerosis and Reactive With Maternal Major Histocompatibility Complex Antigens Cristina Scaletti,1 Alessandra Vultaggio,1 Stefania Bonifacio,2 Lorenzo Emmi,1 Francesca Torricelli,2 Enrico Maggi,1 Sergio Romagnani,1 and Marie-Pierre Piccinni1 Objective. To characterize the cytokine production profile of male-offspring T cells reactive against maternal major histocompatibility complex (MHC) antigens present in the peripheral blood and/or skin from women with systemic sclerosis (SSc). Methods. T cell clones were generated from peripheral blood and/or skin biopsy specimens from 3 women with SSc of recent onset and from peripheral blood from 3 healthy women, all of whom had 1 male child. All clones were screened for their proliferative response in vitro to maternal MHC antigens by measuring 3H-thymidine uptake and for their expression of Y chromosome by using fluorescence in situ hybridization. The concentrations of interferon-␥ and interleukin-4 (IL-4) released by T cell clones in response to maternal MHC antigens were evaluated in culture supernatants, using appropriate enzyme-linked immunosorbent assays. Results. Thirty-nine of 202 T cell clones generated from women with SSc and 11 of 312 from healthy women proliferated in vitro in response to maternal MHC antigens. Seven MHC-reactive T cell clones obtained from women with SSc and 1 obtained from healthy women exhibited the Y chromosome, thus indicating that the clones were derived from T cells of male offspring. All clones generated from male-offspring T cells of SSc women (but not from those of healthy women) produced significantly higher levels of IL-4 in response to stimulation with maternal MHC antigens than did all other clones generated from the same women. The other clones proliferated in response to maternal or allogeneic MHC antigens but did not exhibit the Y chromosome. Conclusion. Male-offspring T cells that are present in the blood and skin of women with SSc and react with maternal MHC antigens exhibit a Th2oriented profile, supporting the possibility that a chronic graft-versus-host reaction attributable to longterm microchimerism plays a pathogenic role in SSc. Systemic sclerosis (SSc) is a connective tissue disease characterized by inflammatory, microvascular, and fibrotic changes affecting the skin (scleroderma) and a variety of internal organs, including the gastrointestinal tract, lungs, heart, and kidney, and by the production of autoantibodies, most notably anticentromere, antitopoisomerase, and anti–RNA polymerase antibodies (1,2). In recent years, the pathogenic mechanisms responsible for the alterations seen in patients with SSc have been partially clarified. Increased numbers of CD4⫹ T cells have been found in skin lesions, as well as in other organs, of patients with early-stage SSc (1,2). Cytokines capable of altering endothelial cell function and/or inducing fibrosis have been detected in SSc sera and tissues (1), suggesting that cytokines secreted by T cells or other cells of the immune system may play an important pathogenic role in this disease (1,2). Recent advances in investigations of the immune response led to identification of subpopulations of CD4⫹ T helper cells, termed type 1 (Th1) and type 2 (Th2), based on their profiles of cytokine production, which are associated with different patterns of immuno- Supported by grants from the Italian Association for Cancer Research and the European Economic Community. 1 Cristina Scaletti, MD, Alessandra Vultaggio, MD, Lorenzo Emmi, MD, Enrico Maggi, MD, Sergio Romagnani, MD, Marie-Pierre Piccinni, PhD: University of Florence, Florence, Italy; 2Stefania Bonifacio, PhD, Francesca Torricelli, PhD: Azienda Ospedaliera di Careggi, Florence, Italy. Address correspondence and reprint requests to Sergio Romagnani, MD, Dipartimento di Medicina Interna, Viale Morgagni, 85, Florence 50134, Italy. E-mail: email@example.com. Submitted for publication September 26, 2000; accepted in revised form September 3, 2001. 445 446 logic reactions (3,4). Th1 cells produce interferon-␥ (IFN␥) and tumor necrosis factor ␤ (TNF␤) and are responsible for phagocyte-dependent host responses (characterized by delayed-type hypersensitivity reactions or a granulomatous pattern), which are involved in protection against intracellular parasites. Th2 cells, which produce interleukin-4 (IL-4), IL-5, and IL-13, are responsible for phagocyte-independent, antibodymediated responses, and are implicated in protection against gastrointestinal nematodes and in allergic conditions (3,4). We and other investigators have reported the existence of a predominant activation of IL-4–producing Th2-like T cells in patients with SSc, which may account for the major alterations that occur in this disease (5–8). The origin of the Th2-dominated immune response in SSc is still unclear. The clinical features of SSc are similar to those of chronic graft-versus-host disease (cGVHD) (9), a chimeric disorder that occurs in recipients of allogeneic stem cell transplants and in which Th2-type responses also predominate (10,11). Identification of fetal DNA and cells in skin lesions and blood from women with SSc has been previously reported (12), suggesting that a microchimerism established by fetal T cells and the activation of such cells may induce a graft-versus-host reaction manifesting as SSc (12,13). However, subsequent studies showed no difference in the frequency of microchimerism with fetal cells between healthy women and women with SSc; therefore, the causal link between microchimerism and disease pathogenesis remains uncertain (14). In the current study, we attempted to characterize the cytokine profile of male-offspring T cells reactive against maternal major histocompatibility complex (MHC) antigens present in the blood and skin of women with SSc. To do this, T cell clones were generated from peripheral blood (PB) and skin biopsy specimens from 3 women with SSc of recent onset and from blood from 3 healthy women; all 6 women had a male child. The maternal T cell clones were then screened for their ability to proliferate in vitro in response to autologous non–T cells, and proliferating clones were examined for expression of the Y chromosome. Seven maternal T cell clones derived from women with SSc and 1 derived from healthy women proliferated in response to autologous non–T cells and exhibited the Y chromosome, which demonstrated that they were derived from male-offspring T cells. Of note, all clones generated from male-offspring T cells of SSc women produced significantly higher levels of IL-4 in response to stimulation with maternal non–T cells than SCALETTI ET AL did all other clones generated from the same women. The other clones responded to the same stimulation but did not exhibit the Y chromosome. These data suggest that male-offspring T cells that are present in blood and/or skin of women with SSc and are reactive against maternal MHC antigens exhibit a Th2-oriented profile, thus supporting their possible role in the cGVHD that may occur in women with SSc. PATIENTS AND METHODS Patients. Three women (ages 38, 46, and 50 years) with recent-onset SSc participated throughout the study. Three healthy women of comparable ages (36, 42, and 52 years) served as controls. All 6 women were selected because each had 1 male child, and none had received either blood transfusions or allografts. The procedures followed in this study were in accordance with the ethical standards of the Regional Committee on Human Experimentation. Generation of T cell clones. T cell clones were generated from PB mononuclear cells (PBMC) and skin biopsy specimens from SSc patients, as described in detail elsewhere (5). Briefly, PBMC obtained from 10-ml samples were cultured for 3 days, and small (2-mm diameter) skin fragments were cultured for 7–10 days in RPMI 1640 supplemented with 2 mM glutamine, 20 M ␤-mercaptoethanol, and 10% fetal calf serum in the presence of recombinant IL-2 (rIL-2) (20 units/ ml). T cell blasts from both PB and skin lymphocyte cultures were then cloned under limiting dilution conditions (0.3 cell/ well) in the presence of phytohemagglutinin (1% volume/ volume), as described previously (5). Growing microcultures were expanded at weekly intervals, using rIL-2 (50 units/ml) and 105 irradiated (6,000 rads) allogeneic PBMC pooled from 3 different healthy women as feeder cells. Cell-surface marker analysis of T cell clones was performed on a Cytoron absolute cytofluorimeter (Ortho, Raritan, NJ), using fluorescein isothiocyanate– and phycoerythrin-conjugated anti-CD4 and anti-CD8 monoclonal antibodies (mAb) (Becton Dickinson, Mountain View, CA). Mixed lymphocyte cultures. The proliferative response of T cell clones to MHC antigens was assessed by mixed lymphocyte culture reaction. T cell blasts (5 ⫻ 104) were cultured for 5 days in 0.2 ml of medium in the absence or presence of 105 irradiated (6,000 rads) non–T cells, which were obtained by removing CD3⫹ cells, as described (5). Sixteen hours before harvesting, cultures were pulsed with 0.5 Ci of 3 H-thymidine (Amersham, Little Chalfont, UK), and radionuclide uptake was measured by scintillation counting, as previously reported. T cell clones were considered reactive to MHC antigens when the mitogenic index (ratio between counts per minute obtained in the presence and cpm in the absence of non–T cells) was ⬎10. Fluorescence in situ hybridization (FISH). The FISH technique was used to analyze T cell clones for the presence of Y chromosome–positive cells. To this end, T cell blasts for each clone (106/ml) were fixed in ethanol/acetic acid (3:1), according to standard procedures (15). They were denatured in a 70% formamide/2⫻ saline–sodium citrate (SSC) solution at 80°C for 2 minutes and then rehydrated with an increasing PATHOGENIC ROLE OF LONG-TERM FETAL MICROCHIMERISM IN SSc series of alcohol at ⫺20°C. Samples were then hybridized simultaneously with 2 probes: the one specific for the centromeric region of X chromosome was labeled with spectrum green, and the one specific for the centromeric region of Y chromosome was labeled with spectrum orange. Probe denaturation was performed at 73°C for 5 minutes. Hybridization was performed at 37°C in a moist chamber overnight. After hybridization, samples were washed twice in 0.1⫻ SSC/0.3% Nonidet P40 (NP40) at 73°C and then once in 2⫻ SSC/0.1% NP40 at room temperature. Samples were counterstained with 4⬘,6-diamidino-2-phenylindole (DAPI) (0.5 gm/ml) and mounted with antifade solution. Sample analysis was performed using a fluorescence microscope equipped with a triple-band filter for DAPI, spectrum green, and spectrum orange. For each sample, 80–100 nuclei were analyzed, as well as metaphases, if present. Detection of cytokine production by T cell clones. The ability of T cell clones to produce cytokines was evaluated after T cell blasts (106/ml) were stimulated for 36 hours with phorbol myristate acetate (PMA) (20 ng/ml; Sigma, St. Louis, MO) and anti-CD3 mAb (100 ng/ml; Ortho), or after T cell blasts (5 ⫻ 105) were incubated for 5 days with autologous or allogeneic non–T cells (5 ⫻ 105) in a total 0.2-ml volume. Cell-free supernatants were assessed for IFN␥ and IL-4 content by enzyme-linked immunosorbent assay (ELISA). The quantitative determination of IFN␥ and IL-4 was performed with ELISAs made in-house, as described elsewhere (5). Cytokine levels that were 5 SD above the mean cytokine levels of control supernatants (derived from irradiated feeder cells alone) were regarded as positive. MHC typing and statistical analysis. The sequencebased typing method (16) was used to perform MHC typing. The Mann-Whitney U nonparametric test and the chi-square test were used to analyze samples. RESULTS T cell clones were generated from both PB and skin biopsy specimens from 3 women with SSc of recent onset and from PB of 3 healthy women, all of whom had a male child. All clones were screened for their proliferative response to irradiated non–T cells from the same women. A total of 29 (18%) of 161 T cell clones generated from the PB and 10 (24%) of 41 from the skin of the 3 SSc women, but only 11 (3.5%) of 312 from the PB of healthy women, showed a proliferative response to autologous non–T cells. T cell clones could not be generated from irradiated feeder cells cultured alone, thus excluding the possibility that clones could be derived from T lymphocytes present in feeder cells. Table 1 shows the total numbers of T cell clones derived from PB and skin of the 6 women, as well as the number of clones that were reactive with autologous non–T cells from each woman. Next, the FISH technique was used to screen the reactive T cell clones for the presence of Y chromosome. 447 Table 1. Reactivity of T cell clones generated from peripheral blood (PB) or skin of women with systemic sclerosis (SSc) or from PB of healthy women to major histocompatibility complex (MHC) antigens from the same women* T cell clones, no. reactive/no. tested (%) SSc women 1 2 3 Total Healthy women 1 2 3 Total PB Skin 11/55 (20) 8/45 (18) 10/61 (16) 29/161 (18) 2/10 (20) 4/13 (30) 4/18 (22) 10/41 (24) 1/118 (0.8) 6/102 (5.8) 4/93 (4.3) 11/312 (3.5) ND ND ND * T cell blasts (5 ⫻ 104) for each clone were stimulated for 5 days with irradiated non–T cells (1 ⫻ 105) obtained from the same subject from which the T cell clones were generated, and their proliferative response was assessed by measuring the incorporation of labeled thymidine. T cell clones were considered reactive to MHC antigens when the mitogenic index was ⬎10, as described in Patients and Methods. ND ⫽ not done. All T cell blasts from 7 (18%) of 39 reactive clones from the SSc women (4 from the first, 2 from the second, and 1 from the third) and 1 (9%) of 11 from the healthy women showed Y chromosome (Figure 1). Because none of the women had previously received any transfusion or allograft, this finding indicated that these clones were derived from male-offspring T cells, and, therefore, that they were alloreactive clones specific for maternal MHC antigens rather than autoreactive clones. This hypothesis was confirmed by comparing the MHC sequence-based typing of 2 T cell clones from woman 1 (A*2601/0101;B*4501/0801;C*0602/0701;DRB1*0701/ 1302;DRB4*0101101; DRB3*0301;DQA1*0201/0102; DQB1*0202/0604;DPB1*⫺/02012) with her MHC typing (A*0101/2605;B*0801/0801;C*0701/0702; DRB1*1302/03011;DRB3*0301/0202;DQA1*0102/ 05011;DQB1*0604/0201;DPB1*02012/0401), and that of her son (A*2601/0101;B*4501/0801;C*0602/0701; DRB1*0701/1302;DRB4*0101101; DRB3*0301; DQA1*0201/0102;DQB1*0202/0604;DPB1*⫺/02012) and husband (A*03011/2601;B*1801/4501;C*0701/0602; DRB1*11011/0701;DRB3*0202; DRB4*0101101; DQA1*0505/0201;DQB1*0301/0202;DPB1*02012/⫺). In women with SSc, 4 of 7 clones were obtained from PB and 3 from skin, suggesting the presence of male-offspring–reactive T cells in both the circulation and target tissues. Moreover, the fact that these clones 448 SCALETTI ET AL Figure 1. Presence of Y chromosome in T cell clones generated from the skin of women with systemic sclerosis. T cell blasts from each clone were cohybridized simultaneously with 1 probe specific for the centromeric region of X chromosome (labeled with spectrum green) and the other specific for the centromeric region of Y chromosome (labeled with spectrum orange), distinguishing male cells (yellow ⫹ red signal) from female (2 yellow signals) cells. A, T cell blasts from 1 clone showing Y chromosome. B, A mitotic T cell blast from the same clone. C, T cell blasts from 1 clone showing no Y chromosome. D, Mitotic T cell blasts from the same clone. were truly reactive against maternal MHC antigens was fully supported by the observation that their in vitro proliferative response to maternal non–T cells could be completely blocked by the addition of anti–MHC class II mAb D.1/12 (data not shown). The ability of clones showing Y chromosome to produce IFN␥ and IL-4 in response to either polyclonal stimulation with PMA plus anti-CD3 mAb or specific stimulation with maternal MHC antigens (non–T cells) was also examined. As control, all reactive T cell clones showing no Y chromosome were tested for their ability to produce IFN␥ and IL-4 under the same experimental conditions, as well as in response to non–T cells from 10 different donors who had incompatible MHC antigens. Among the different groups of clones, there were no significant differences in the production of IFN␥ or IL-4 following polyclonal stimulation (data not shown). Following MHC stimulation, however, the clones derived from male-offspring T cells from SSc women produced reduced amounts of IFN␥ (1.8 ⫾ 0.8 ng/ml) compared with T cell clones that showed no Y chromosome from the same women (5.2 ⫾ 1.6 ng/ml), even though the difference was not statistically significant. However, the male-offspring clones produced significantly higher levels of IL-4 than did the clones showing no Y chromosome (3.3 ⫾ 1.1 versus 0.26 ⫾ 0.1 ng/ml) (Z, corrected for ties, ⫺3.8, P ⬍ 0.0001) when they were stimulated with either maternal MHC antigens (Figure 2) or allogeneic non–T cells (data not shown). In contrast, the only MHC-reactive clone generated from male-offspring PATHOGENIC ROLE OF LONG-TERM FETAL MICROCHIMERISM IN SSc 449 T cells of healthy women produced low concentrations of both IFN␥ and IL-4 in response to stimulation with maternal MHC antigens (Figure 2). DISCUSSION SSc is an autoimmune disease with a strong predilection in women, a peak incidence in the years after childbearing, and clinical and immunologic similarities to cGVHD (1,2). Both SSc and cGVHD are characterized by skin, lung, and esophageal involvement and intense fibrosis (1,2). Immunologic similarities include lymphocytic infiltration of affected tissues, the presence of Scl-70 and PM-Scl serum autoantibodies, and up-regulation of a series of cytokines (1,2,5–8). Moreover, several reports have indicated that IL-4 production by CD4⫹ T cells infiltrating SSc and cGVHD lesions is prominent in both animal models and humans, suggesting a Th2-polarized phenotype of the specific immune response in both conditions (5–8,10,11). Recent reports have suggested that SSc may be the result of a graft-versus-host reaction caused by persistent feto-maternal microchimerism, which has been detected in large numbers of childbearing women and their immunocompetent offspring, respectively (12,13). If this hypothesis is correct, the offspring T cells engrafted in the PB or skin of women with SSc should have a Th2-polarized phenotype, such as that responsible for cGVHD in bone marrow allograft recipients. To address this possibility, we generated T cell clones from both PB and skin of 3 women with recentonset SSc, as well as from the PB of 3 healthy women, all of whom had 1 male child. T cell clones that proliferated in response to maternal MHC antigens were then selected and screened for the presence of Y chromosome. T cell clones showing the Y chromosome were obviously derived from male-offspring–engrafted T cells, as was also demonstrated by their MHC typing. In contrast, the clones that were reactive with maternal non–T cells but showed XX chromosomes were the progeny of maternal autoreactive T cells. Of note, the number of clones that proliferated in response to autologous non–T cells was significantly higher in women with SSc than in healthy women, independent of whether they expressed the XY or the XX chromosome. This finding suggests the presence of an increased proportion of both autoreactive maternal and offspringengrafted T cells in women with SSc. The increased proportions of clones derived from male offspring found Figure 2. High interleukin-4 (IL-4) production by male-offspring T cells present in the peripheral blood (PB) or skin of women with systemic sclerosis (SSc) in response to maternal major histocompatibility complex (MHC) antigens. T cell clones were generated from PB and skin of women with SSc and from PB of healthy women, all of whom had male children, and were selected for their ability to proliferate in response to maternal MHC antigens. The ability of MHC-reactive T cell clones showing Y chromosome (E) and of those showing no Y chromosome (F) to produce IL-4 and interferon-␥ (IFN␥) following stimulation with irradiated non–T cells from the PB of the same women was compared. IL-4 production and IFN␥ production by individual T cell clones are shown. 450 in this study may be consistent with other studies that showed higher levels of fetal DNA in the blood or skin of women with SSc compared with healthy women (12,13). However, the mechanisms responsible for the increased proportions of autoreactive T cell clones in women with SSc remain unclear. When the ability of MHC-reactive clones to produce IFN␥ and IL-4 in response to polyclonal stimulation was assessed, no difference in the cytokine production profiles among the different groups of clones was found. However, clones derived from male-offspring T cells of women with SSc showed significantly higher IL-4 production in response to MHC stimulation than did maternal autoreactive clones, whereas the only clone derived from male-offspring T cells of healthy women produced low concentrations of both IFN␥ and IL-4. In male-offspring T cell clones derived from women with SSc, the high production of IL-4 in response to stimulation with maternal MHC antigens was not attributable simply to the different type of stimulation (autoreactive versus alloreactive). When T cell clones lacking the Y chromosome were stimulated with allogeneic cells, the amount of IL-4 produced was comparable with that produced by autoreactive maternal T cell clones but was significantly lower than that of male-offspring T cell clones showing alloreactivity toward maternal MHC antigens. Certainly, interpretation of these data is problematic because of the small number of subjects and the fact that the number of clones being compared (i.e., 7 from SSc women and 1 from healthy women) was different. However, the results suggest that maleoffspring T cells present in PB and/or skin of women with SSc exhibit a functional profile prevalently oriented toward Th2 rather than Th1 cytokine production in response to stimulation with maternal MHC antigens. This finding supports the concept that fetal cells found in the PB and/or skin of women with SSc are functionally similar to those responsible for cGVHD in experimental animal models (9,10). The selected CD4⫹ T cell population derived from fetal cell engraftment may be at least partially included in the population of T cells expressing IL-4 that are found in the skin of patients with SSc (5), and it is probably responsible for the enhanced levels of IL-4 in the biologic fluids of these patients (6–8). Thus, although the results of this study need further investiga- SCALETTI ET AL tion, they provide support for the hypothesis that a fetal antimaternal cGVHD may be an immunopathogenic mechanism in the development of SSc in some women. ACKNOWLEDGMENT The authors thank Roberto Accolla, University of Varese, Varese, Italy, for the kind gift of the D.1/12 anti–MHC class framework mAb. REFERENCES 1. White B. Pathogenesis: immune aspects. In: Clements PJ, Furst ED, editors. Systemic sclerosis. Baltimore: Williams and Wilkins; 1996. p. 229–50. 2. Galperin C, Gershwin ME. Systemic sclerosis (scleroderma). In: Rose NE, Mackay IR, editors. The autoimmune diseases. 3rd ed. San Diego: Academic Press; 1998. p. 317–42. 3. Mosmann TR, Coffman RL. TH1 and TH2 cells: different patterns of lymphokine secretion lead to different functional properties. Annu Rev Immunol 1989;7:145–73. 4. Romagnani S. Lymphokine production by human T cells in disease states. Annu Rev Immunol 1994;12:227–57. 5. Mavalia C, Scaletti C, Romagnani P, Carossino AM, Pignone A, Emmi L, et al. Type 2 helper T-cell predominance and high CD30 expression in systemic sclerosis. Am J Pathol 1997;151:1751–8. 6. Hasegawa M, Fujimoto M, Kikuchi K, Takehara K. Elevated serum levels of interleukin 4 (IL-4), IL-10, and IL-13 in patients with systemic sclerosis. J Rheumatol 1997;24:328–32. 7. Atamas SP, Yurovsky VV, Wigley FM, Goter Robinson CG, Henry P, et al. Production of type 2 cytokines by CD8⫹ lung cells is associated with greater decline in pulmonary function in patients with systemic sclerosis. Arthritis Rheum 1999;42:1168–78. 8. Ong C, Wong C, Roberts CR, Teh HS, Jirik FR. Anti-IL-4 treatment prevents dermal collagen deposition in the tight-skin mouse model of scleroderma. Eur J Immunol 1998;28:2619–29. 9. Chosidow O, Bagot M, Vernant JP, Roujeau JC, Cordonnier C, Kuentz M, et al. Sclerodermatous chronic graft-versus-host-disease: analysis of seven cases. J Am Acad Dermatol 1992;26:49–55. 10. De Wit D, van Mechelen M, Zanin C, Doutrelepont JM, Velu T, Gerard C, et al. Preferential activation of Th2 cells in chronic graft-versus-host reaction. J Immunol 1993;150:361–6. 11. Sakurai J, Ohata J, Saito K, Miyajima H, Hirano T, Kohsaka T, et al. Blockade of CTLA-4 signals inhibits Th2-mediated murine chronic graft-versus-host disease by an enhanced expansion of regulatory CD8⫹ T cells. J Immunol 2000;164:664–9. 12. Artlett CM, Smith JB, Jimenez SA. Identification of fetal DNA and cells in skin lesions from women with systemic sclerosis. N Engl J Med 1998;338:1186–91. 13. Nelson JL, Furst DE, Maloney S, Gooley T, Evans PC, Smith A, et al. Microchimerism and HLA-compatible relationship of pregnancy in scleroderma. Lancet 1998;351:559–62. 14. Evans PC, Lambert N, Maloney S, Furst DE, Moore JM, Nelson JL. Long-term fetal microchimerism in peripheral blood mononuclear cell subsets in healthy women and women with scleroderma. Blood 1999;93:2033–7. 15. Rooney DE, Crepulkowski BH. Human cytogenetics. Oxford: Oxford University Press; 1992. 16. Santamaria P, Lindstrom AL, Boyce-Jacino MT, Myster SH, Barbosa JJ, Faras AJ, et al. HLA class I sequence-based typing. Hum Immunol 1993;37:39–50.