macrophage derived cells in normal and transplanted human kidneys

CLINICAL

IMMUNOLOGY

AND IMMUNOPATHOLOGY

Monocyte/Macrophage Transplanted CHARLES E. ALPERS* Department

36, 129-140

(1985)

Derived Cells in Normal Human Kidneys’ ANDJAY

H.

and

BECKSTEAD~

of Pathology. University of California at San Francisco. San Francisco, California 94143

The existence of HLA-DR/Ia-like antigen (Ia)-bearing cells of the mononuclear phagocyte system, macrophages (Mac), and/or interdigitating cells (IDC), in the normal kidney is controversial. If present, such cells may be important in renal transplant rejection. We performed enzyme histochemistry using u-naphthyl acetate/butyrate esterases (aNAE, uNBE), 5’-nucleotidase (5’N), acid phosphatase (AcP), alkaline phosphatase (AlkP), and ATPase (ATP) as well as immunoperoxidase staining for Ia and lectin binding (Ulex europaeus I; UEA) on plastic-embedded tissue sections of normal kidneys and rejected renal allografts. Plastic embedding provides clear visualization of histologic detail and allows specific identification of immunoperoxidase-stained cells. Mac and IDC (shown to be la+, aNAE+, AcP+, ATP+ in other sites) could not be demonstrated in normal renal interstitium. IDC and Mac were not generally identified in normal mesangium, although they could not be altogether distinguished from Ia’ endothelial cells. Focal mesangial staining for aNAE but not oNBE was present. Rejected kidneys showed increased numbers of uNAE+ cells in glomeruli. These cells were frequently Ia negative and often appeared to be blood monocytes present in capillary lumens. Peritubular capillaries and glomerular endothelium stained strongly for UEA, 5’N, and Ia. Our results suggest that previous reports of the presence of IDC in renal tissue on the basis of staining for Ia on frozen tissue may be due to staining of compressed or obliquely o 1985 Academx sectioned vascular structures that were not adequately visualized. Press, Inc.

INTRODUCTION A series of studies has been recently performed in which monoclonal antibodies conjugated to fluorescent or peroxidase chromogens have been used to demonstrate the presence of interdigitating cells (IDC) in the interstitium of kidneys of humans and other mammalian species (l-4). The nature of these cells is incompletely understood. Current studies have focused on their possible derivation from monocyte precursors and on their role as antigen processing/presenting cells (57). As such, these cells might be expected to participate in immunologically mediated tubulointerstitial nephropathies and in renal allograft rejection. The immunohistochemical evidence for the presence of these cells in human kidneys is controversial (8, 9). In part, this may be attributable to a certain lack of histologic detail inherent in the use of immunofluorescence techniques and/or ’ Presented in part at the International Academy of Pathology, U.S.-Canadian Division, 73rd Annual Meeting, San Francisco Calif., March, 1984. Supported in part by USPHS Grant CA-14264. 2 Present address: Department of Pathology, Brigham and Women’s Hospital, 75 Francis St.. Boston, Mass. 02115. 3 To whom correspondence should be sent: Department of Pathology, HSW 501. University of California School of Medicine, San Francisco, Calif. 94143. 129 0090-1229/85 $1.50 Copyright 0 1985 by Academic Press, Inc. All right\ of reproduction in any form reserved.

130

ALPERS

AND

BECKSTEAD

the frozen tissues utilized for these studies. Plastic embedding has been shown to allow good preservation of histologic detail with preservation of endogenous enzymatic activity and antigenic expression (IO- 12), thus obviating many of the above-noted difficulties in morphologic visualization. We utilized a series of histochemical procedures which have been shown to be useful in the identification of cells of monocyte/macrophage lineage, cY-naphthyl acetateibutyrate esterases (oNAE/cxNBE), and acid phosphatase (AcP), as well as immunohistochemical stains for HLA-DRiIa-like antigen (la). We also utilized histochemical procedures useful in the identification of vascular endothelium, including ATPase (ATP). alkaline phosphatase (AlkP), 5’-nucleotidase (5’N), and tissue binding by the plant lectin Ulex europaeus Z (UEA) (12-16). In this manner, we attempted to characterize the phenotypic properties of any interstitial cell population which maybe present in normal and transplanted human kidneys. Our findings do not support the concept that IDC are normally present in human renal interstitium. We also provide evidence against the hypothesis that there is a resident population of macrophage-derived cells in the normal human glomerulus. MATERIALS

AND METHODS

Renal tissue was obtained from routinely accessioned surgical nephrectomy specimens. We studied 19 cases of severe chronic transplant rejection. Samples of normal renal tissue were obtained from uninvolved areas of 24 kidneys removed for carcinoma. The procedures used for fixation, processing, and embedding of tissue sections in glycol methacrylate have been previously reported in detail (8, 9). Briefly, thin tissue slices (approximately 2 mm) were fixed in cold (4°C) 4% paraformaldehyde in 0.1 M PO, buffer. After processing and embedding, 1- to 3-km sections were stained with a modified Maximow’s stain (hematoxylin-eosin-azure) for study of routine morphology. Histochemical procedures (10) with subsequent moditications (11) for alkaline phosphatase, acid phosphatase, a-naphthyi acetate esterase, o-napthyl butyrate esterase, ATPase, and 5’-nucleotidase were performed as previously described. Immunohistochemical and lectin histochemical procedures on plastic sections were performed in the following sequence (unless otherwise specified, all incubations were carried out at room temperature): (1) digestion with 0.25% trypsin for 30 min at 37°C. followed by phenylhydrazine (pH 7.1) for 1 hr at 37°C; (2) incubation with 3% normal horse serum or 1% ovalbumin (for lectins) in Ca*+- and Mg*+-free phosphate-buffered saline (CMF-PBS) for 30 min at 37°C; (3) incubation with the following sequences of antisera or lectin conjugate (all dilutions were made in CMF-PBS with 3% horse serum or 0.1% ovalbumin (for lectins), and sections were rinsed between incubations with the same solution) (a) mouse anti-human Ia (Coulter, I,; 1:500) overnight at 4°C; biotinylated horse anti-mouse IgG (Vector; 1:300) for 1 hr: avidin-biotin-peroxidase complex (Vector, 1:80) for 1 hr. (b) U. europaeus Z lectin directly conjugated to peroxidase (E-Y Labs; 1:50) for 2 hr:

MONOCYTElMACROPHAGE

DERIVED

CELLS

IN KIDNEY

131

(4) development of peroxidase reaction by preincubation with 0.05% 3,3-diaminobenzidine (DAB) for 10 min, following by incubation with 0.05% DAB with 0.1% H,O, and 0.1 M imidazole (Sigma) for 7 min (the reaction is further enhanced by incubation with 0.5% C&O, for Smin); (5) counterstaining with Gills hematoxylin #2 for 1.5 min; blueing with Scott’s water; drying and mounting with Permount. An expanded description of antigen labeling using monoclonal antibodies in plastic sections is reported in a separate publication (Beckstead, J., in press). Control procedures included the substitution of irrelevant antibodies or ascites fluid for the primary antibodies and elimination of primary or secondary reagents. RESULTS HLA-DRlia-like

Antigen

Staining of normal kidneys for Ia was prominent in endothelium of glomerular capillaries, peritubular capillaries, and larger blood vessels (Fig. 1). These reactions were equally strong in normal and severely damaged kidneys (Fig. 2). In the glomerulus, the intense staining observed appeared to be confined to capillary endothelium, but because of the attenuated nature of the normal capillary endothelial cells we could not always be certain that lesser, focal areas of mesangial

FIG. 1. Normal kidney stained for Ia antigen. There is prominent staining of glomerular and peritubular capillary endothelium. The pattern of peritubular capillary staining, particularly when capillaries are collapsed or sectioned tangentially (arrows), could be misconstrued as staining of interstitial interdigitating cells if adequate visualization is not achieved t x 490).

132

ALPEKS

AND BEC‘KS’IXAI>

FIG. 2. Rejected renal allograft stained for la antigen. Many interstitial capillaries remain Ia positive n the presence of a mononuclear cell infiltrate ( x 200).

staining were not also present. Of particular note was the frequent finding of obliquely sectioned positively stained capillaries in the interstitium, which at times assumed a stellate appearance. No population of extravascular interstitial cells staining for this antigen was observed. Focal staining of cell borders of tubular epithelium in normal and damaged kidneys was observed. Rejected renal allografts frequently contained monocytes/histiocytes within glomerular capillary lumens and mesangium. Only a proportion of these cells, estimated at one-third to one-half of the total, stained positively for the Ia antigen. Endothelial staining was well preserved in glomeruli and interstitium even in areas of prominent inflammatory injury. Mononuclear cell infiltrates usually contained a subpopulation of Ia+ cells. c+Naphthyl Acetate EsteraselwNapthyl

Butyrate Esterase

Staining of normal glomeruli for aNAE frequently showed fine granular staining of mesangium that could not always be clearly localized to specific cells, but most often appeared present in mesangial cells (Fig. 3a). The number of positively staining cells per glomerulus showed considerable variability and ranged from 0 to 10 in normal kidneys. The extent of mesangial staining for olNAE generally appeared moderately increased in rejected allografts (Fig. 4), although the variability observed in normal glomeruli was also seen in this group of kidneys. Normal glomeruli were entirely negative for CXNBE (Fig. 3b). Rejected allografts

MONOCYTE/MACROPHAGE

DERIVED

CELLS

IN KIDNEY

133

FIG. 3. Normal kidney stained for aNAE (a) and uNBE (b). There is fine, focal granular staining of mesangium with aNAE (arrowheads), and diffuse cytoplasmic staining of tubules for both aNAE and aNBE. There is no positively staining population of interstitial cells for either enzyme ((a) x 348; (b) x360).

134

ALPERS AND BECKSTEAD

FIG. 4. Rejected renal allograft stained for IxNAE. There is an influx of positively staining mononuclear cells into the glomerulus (arrowheads). often in capillary lumens. as well as a population of negatively stained mononuclear cells ( x448).

often contained monocytes/histiocytes within capillary lumens and within mesangium that were strongly positive for both aNAE and CYNBE. Vascular endothelium did not stain for (wNAE or oNBE in glomeruli or interstitium. There was no population of interstitial cells that stained for CXNAE or aNBE in normal kidneys. Inflammatory cell infiltrates in rejected renal allografts usually contained a subpopulation of monocyteslhistiocytes that stained positively for both enzymes. Acid Phosphatase

There was rare weak staining of glomerular mesangial cells in normal kidney, but most often there was no glomerular staining. Rejected kidney allografts often showed an increase in glomerular AcP which could be attributed to a population of monocytes/histiocytes with weak cytoplasmic staining infiltrating capillary lumens and mesangium. No population of AcP-staining cells was seen in normal renal interstitium; such cells were seen only as a component of an inflammatory cell infiltrate. Alkaline Phosphatase

There was no staining of glomeruli in normal or rejected kidneys for this enzyme. Renal interstitium showed strong endothelial staining in both large and small arterial and venous blood vessels and peritubular capillaries. No population

MONOCYTEiMACROPHAGE

DERIVED

CELLS

IN

KIDNEY

135

of positively staining interstitial cells was observed. Obliquely sectioned interstitial capillaries having a stellate configuration showed positive staining similar to that seen with ATP, 5’N, UEA, and Ia antigen. 5’-Nucleotidase

Endothelial cells lining glomerular capillaries showed prominent staining for 5’N in both normal and rejected kidneys. No mesangial staining was observed. Staining of infiltrating inflammatory cells was not observed in glomeruli of rejected allografts. Strong staining of endothelial cells lining large and small blood vessels and peritubular capillaries, similar to that seen for AIkP, ATP, UEA, and Ia, was present in both normal and rejected kidneys. No population of positive-staining extravascular interstitial cells was observed in normal kidneys. Occasional interstitial mononuclear inflammatory cells showed positive staining in rejected allografts. ATPase

There was no staining of glomeruli for ATP. The endothelium of the entire interstitial vasculature, including peritubular capillaries, showed strong linear staining in both normal and rejected kidneys which was identical to the findings for AlkP, 5’N, UEA, and Ia. No population of interstitial cells staining for ATP was identified in normal kidneys. Inflammatory cell infiltrates in rejected allografts frequently contained scattered mononuclear cells with focal cytoplasmic staining. Ulex europaeus

I

Staining of normal kidneys for binding by the lectin UEA was identical to that observed for Ia (Fig. 5). There was prominent staining of all renal endothelium. These reactions were equally strong in normal and severely damaged kidneys. Moderately strong staining of collecting ducts was occasionally present; no other portions of the nephron demonstrated staining in either normal or diseased kidneys. Obliquely sectioned peritubular capillaries with stellate configurations like those described under Ia, AlkP, ATP, and 5’N, above, also stained positively for UEA. DISCUSSION

An important and unanswered question in understanding the rejection process in renal transplantation involves the mechanism of local recognition of donor antigens that may mediate immunologic recognition and activation of the host’s inllammatory response. A currently attractive hypothesis has focused on the possible role of a population of interstitial interdigitating cells referred to by some investigators as dendritic cells, that may be a normal constituent of human kidney (l-4). Interdigitating cells are presently thought to represent a family of cells of apparent bone marrow origin that have been best studied in lymph node and spleen and which have been ascribed an important role in antigen presentation

136

ALPERS

AND BECKSTEAD

FIG. 5. Normal kidney stained for UEA binding. There is strong staining of glomerular and peritubular capillaries. Peritubular capillary staining is comparable to that seen with ATPase and la antigen (X315).

to lymphocytes (5-7). Accordingly, these cells are invariably la positive; they additionally share some features of macrophages in humans including the presence by enzyme histochemistry of ATPase and the variable presence of orNAE/ aNBE and AcP (6, 7, 11). Cells with similar morphologic and immunohistochemical (Ia+) properties have been demonstrated in a wide variety of tissue sites, including skin, liver, heart, ureter, and testis (5, 17-20) where it has been presumed that they may also mediate local immunologic reactions. Although there is as yet no definitive evidence that such cells are necessary for the development of rejection of human organ transplants, the presence of such cells in normal kidney would support the hypothesis that they have an important role in antigen presentation in transplant rejection and other immunologically mediated nephropathies. A number of investigators have utilized indirect immunofluorescence and/or immunoperoxidase techniques, usually on frozen tissue sections, to provide evidence for a dendritic cell population in normal kidney. Such cells apparently can be demonstrated in the rat (21) while their presence in human kidneys is less certain. The findings of this study are in agreement with a portion of each of those studies that have examined human tissue (l-5, 8, 9, 22-24). We observed extensive staining of the endothelial cells of peritubular capillaries and larger blood vessels for Ia. The cytochemical features of the la+ endothelial cells in renal interstitium were similar to those seen in endothelium of other sites, i.e..

MONOCYTElMACROPHAGE

DERIVED

CELLS

IN

KIDNEY

137

prominent binding to UEA and the presence of stainable S’N, ATPase, and AlkP, but not aNAE, aNBE, or AcP (lo- 16). However, like Hancock et al., and Hinglais et al., we were unable to identify a second interstitial cell population of Iaf cells that was not of endothelial origin and which might correspond to a normal resident population of interdigitating cells (8, 9). Rather, the improved visualization of Ia+ structures provided by the plastic embedding techniques revealed in every case stellate staining patterns which might be suggestive of dendritic cells if seen only with the limited resolution provided by immunofluorescence or frozen section techniques, were in fact portions or peritubular capillaries that had been obliquely sectioned or sectioned on end. Confirmation of this interpretation is provided by the histochemical findings that staining patterns for Ia were paralleled by those of UEA, 5’N, ATP, and less regularly by AlkP, all features of renal interstitial endothelium (12-14). Alternatively, staining for aNAE, aNBE, and AcP, each of which might be seen in cells of monocyte lineage, failed to stain a nonendothelial population of interstitial ceils. The strongest evidence for a renal population of Ia+ interdigitating cells have come from immunofluorescence and immunoperoxidase studies on frozen tissue by Raftery et al. (3, 4) and Hart and co-workers (I, 2). The studies of Raftery et al. noted the cytochemical features of ATP +, ACP- staining in the Ia+ cells that they recognized as interstitial dendritic cells, features compatible with either endothelial (lo- 12) or interdigitating cell origin (6). However, because of their morphology, reactivity with monoclonal antibodies said to recognize IDC, and failure to stain for Factor VIII-related antigen, these investigators interpreted the histochemical findings as indicative of dendritic cells. Hart et al. similarly provide evidence for the existence of renal interdigitating cells through the use of multiple antisera. They, too, noted a proportion of interstitial Ia+ cells also appeared to stain with a monoclonal antibody said to be specific for cells of leukocyte origin, thus indicating a presumed monocyte rather than endothelial lineage for these cells. Our studies did not identify the presence of an interstitial population of Ia+ cells distinct from endothelium in normal kidneys. However, we have not utilized the antisera employed by these other investigators and are therefore uncertain of the significance of their findings. Antisera to Factor VIII-related antigen has been shown in some studies to be somewhat insensitive to cells of endothelial origin, particularly those lining small capillaries, as compared to UEA binding (14, 15); therefore the failure of some cells to stain for this reagent need not exclude an endothelial origin for any given cell. Given the uncertainty as to the ontogeny of interdigitating cells, and that the phenotypic similarity between these cells and monocytes may be only partial at best (6, 7), the significance of staining by the common leukocyte or monocyte antibodies is also uncertain. It is possible that some antibodies against monocytes may also react with some kinds of endothelium. A particular concern is that by using immunofluorescence techniques or immunoperoxidase techniques with imperfect visualization of frozen tissues, incidental circulating leukocytes transiently present in the interstitial space might be inadvertently stained and misinterpreted. Based upon the improved resolution of the plastic-embedding technique, we remain unconvinced that a population of

138

ALPERS

AND

BECKSTEAD

interdigitating cells in normal human renal interstitium has been clearly demonstrated. Rather, we believe this study further illustrates the considerable caution that is required in the interpretation of immunostained tissue specimens when histologic detail is not optimally preserved. Our studies of the interstitial cell population in severe allograft rejection showed that endothelial staining for la, UEA, 5’N. AlkP, and ATP was generally retained except in areas of extreme inflammatory injury. In most cases there was a patchy influx of mononuclear inflammatory cells into the interstitium. Within this influx was an admixture of lymphocytes and macrophages, a portion of which showed some staining for Ia. However, this staining does not distinguish between an influx of peripherally derived la+ monocytes and lymphocytes from a small local population of interstitial cells that show induced expression of Ia in the setting of transplant rejection. This study also provides evidence against the hypothesis that there is a resident population of monocyte/macrophage-derived cells present in the normal human glomerulus. Schreiner et al. have shown that a population of such cells can be identified in glomeruli of the rat, where they comprise an estimated l-2% of the glomerular cell population (25, 26). In the rat, these cells have demonstrated phagocytic properties, Fc’ receptors, and the ability to promote lymphocyte activation in vitro. Such cells, if also present in human glomeruli, might serve as important mediators of various glomerulonephridites. We reasoned that monocyte-derived cells in the glomerulus should bear considerable similarity to the phenotype of monocyte-derived cells elsewhere, e.g., the presence of cytoplasmic (wNAE, oNBE, AcP, as well as possible expression of Ia (10, 11). We were unable to demonstrate a population of cells with these characteristics in the normal glomerulus. Efforts by other groups to identify a population of these cells in the normal human glomerulus have utilized glomerular tissue cultures, enzyme histochemistry, and ultrastructural studies (9, 27, 28). Such studies have demonstrated phagocytic properties in some mesangial cells (27). However, in contrast to various proliferative glomerulonephridites and transplant rejection, where influx of monocyte-derived cells into the glomerulus is now well accepted (27, 29-36), the evidence for a resident population in the normal human glomerulus remains inconclusive. As noted in previous studies, we too have found an influx of mononuclear cells within glomerular capillary lumens and mesangium in transplant rejection (36). These cells have the phenotypic appearance of monocytes with positive staining for aNAE, oNBE, AcP, and less frequently for Ia. While our studies do not clarify the function of these cells in the graft rejection process, these findings support a role for glomerular injury in the pathophysiology of rejection in addition to the better known and easily recognized interstitial and vascular processes usually associated with this type of injury. In summary, it seems likely that phenotypic analysis of the cellular constituents of normal and diseased kidneys is likely to provide better understanding of the mechanisms of renal injury. Using a technique which optimally preserves histologic detail of immunostained tissue, we have been unable to demonstrate either a population of antigen presenting interdigitating cells in the renal interstitium or

MONOCYTE/MACROPHAGE

DERIVED

CELLS

IN KIDNEY

139

a population of monocyte-derived cells in normal mesangium. However, the phenotypic expression of interdigitating cells in nonlymphoid sites remains poorly characterized at present, and it is likely that additional markers for these cells will be discovered. Additional studies of human kidneys using probes for such markers should provide a more definite answer concerning the putative existence of these cells. REFERENCES I. Williams, K. A., Hart, D. N. J., Fabre, J. W., and Morris, P. J., Transplantation 29, 274, 1980. 2. Hart, D. N. J.. Fuggle, S. V.. Williams. K. A., Fabre, J. W., Ting, A., and Morris, P. J.. Trunsplantation 31, 428, 1981. 3. Raftery. M. J., Poulter, L. W., Janossy, G.. Sweny, P.. Fernando, 0. N., and Moorhead, J. F., J. Clin. fatho/. 36, 734, 1983. 4. Raftery. M. J., Poulter, L. W., Sweny, P.. Fernando, 0. N., Janossy. G.. and Moorhead, J. F., Transplant. Proc. 15, 1781, 1983. 5. Natali, P. G., Russo, C., Ng, A.-K., Giacomini, P., Indiveri. F., Pellegrino, M. A., and Ferrone, S.. Tissue distribution of human Ia-like antigens. In “Ia Antigens: Man and Other Species” (S. Ferrone. C. S. David, Eds.), Vol. II, pp. 81-110. CRC Press, Boca Raton. FL, 1982. 6. Thorbecke, G. J., Silberberg-Sinakin, I., and Flotte, T. J., J. Invest. Dermutol. 75, 32, 1983. 7. Flotte. T. J.. Springer, J. A., and Thorbecke, G. J., Amer. J. Pathol. 111, 112. 1983. 8. Hancock, W. W., Kraft, N., and Atkins, R. C., Pathology 14, 409, 1982. 9. Hinglais. N., Katzatchkine, M. D.. Charron, D. J.. Appay. M.-D., Mandet, C., Paing. M.. and Bariety, J., Kidney Znt. 25, 544. 1984. IO. Beckstead, J. H.. Halverson. P. S.. Ries, C. A., and Bainton. D. E. Blood 57, 1088. 1981. 11. Beckstead, J. H.. Amer. J. C/in. Pathol. 80, 131, 1983. 12. Alpers, C. E., and Beckstead, J. H., Amer. /. C/in. Puthol., 1985, in press. 13. Holthofer, H., Virtanen, I., Pettersson. E.. Tornroth, T.. Alfthan, 0.. Linder, E., and Miettinen, A., Lab. Invest. 45, 391, 1981. 14. Holthofer, H., Virtanen, I., Karinemi, A.-L., Hormia. M.. Linder. E.. and Miettinen. A.. Lab. Invest.

47, 60, 1982.

15. Miettinen, M., Holthofer, H., Lehto. V.-P., Miettinen. A,, and Virtanen. I.. Amer.

J. C/in. Pathol.

79, 32, 1982.

16. Tonezawa, S.. Nakamura, T., Irisa, S., Otsuji, Y., and Sato. E., Nephron 35, 187, 1983. 17. Wiman, K.. Curman, B., Forsum. U., Klareskog, L., Malmnas-Tjernlund, U.. Rask, L., Tragardh, L., and Peterson, P. A., Nature (London) 276, 711, 1978. 18. Pellegrino, M. A., Natali, P. G.. Ng. A.-K., Imai, K.. Russo, C., Bigotti, A., Wolf, P., and Ferrone, S., Unorthodox expression of Ia-like antigens on human tumor cells of nonlymphoid origin: Immunological properties. structural profile, and clinical relevance, In “Immunological Analysis: Recent Progress in Diagnostic Laboratory Immunology” (R. M. Nakamura, W. R. Dito. and E. S. Tucker III, Eds.), pp. lS7-173, Masson, New York, 1983. 19. Koyama, K., Fukunishi, T.. Barcos, M.. Tanigaki, N.. and Pressman, D.. Immunology 38, 333. 1979. 20. Daar, A. S., Fuggle, S. V., Hart, D. N. J.. Dalchav, R., Abdulaziz, Z., Fabre. J. W.. Ting, A., and Morris. P. J., Transplant. Proc. 1.5, 311, 1983. 21. Hart, D. N. J.. and Fabre, J. W., Transplantation 31, 318. 1981. 22. Hayry. P., Von Willebrand, E., and Andersson. L. C., Scund. J. Immunol. 11, 303, 1980. 23. Natali. P. G., De Martino, C., Marcellini, M., Quaranta. V.. and Ferrone. S., C/in. Immunol. Immunoputhol. 20, 1 I. 1981. 24. Paul. L. C.. Van Es, L. A.. and Baldwin, W. M., III. C/in. Immunol. Immunopathol. 19, 206, 1981. 25. Schreiner, G. F., Kiely, J.-M., Cotran, R. S., and Unanue. E. R.. J. C/in. Invest. 68, 920, 1981. 26. Schreiner, G. F., Cotran, R. S., and Unanue, E. R., Springer Semin. Immunoputhol. 5,251, 1982. 27. Atkins, R. C., Holdsworth, S. R., Hancock, W. W., Thomson. N. M., and Glasgow, E. F., Springer

Semin.

Immunoputhol.

5, 269, 1982.

ALPERS AND BECKSTEAD

140

28. Schreiner, G. F.. and Cotran, R. S., Fed. Proc. 42, 389. 1983. Cotran. R. S., J. Lab. C/in. Med. 92, 837, 1978. 30. Thomson, N. M., Holdsworth, S. R., Glasgow, E. E. and Atkins, R. C’.. Amer. 29.

J. Pu~/w/.

94,

223, 1979.

31. Monga, G.. Massucco, G., di Belgiojoso,

G. B.. and Busnach. G., Amer.

J. Parhol.

94, 271,

1979. 32.

33. 34. 35. 36.

Magil. A. B.. Wadsworth, L. D., and Loewen, M., Lab. fntresr. 44, 27, 1981, Magil, A. B., and Wadsworth, L. D.. Lab. Invest. 45, 77, 1981. Magil. A. B., and Wadsworth, L. D., Lab. Invest. 47, 160. 1982. Filit, H. M., and Zabriskie. J. B., Amer. J. Put/d. 109, 227, 1982. Harry, T. R., Coles, G. A., Davies, M., Bryant, D., Williams, G. T., and Griffin, P. J. A., Transplantation

37, 70. 1984.

Received October 29, 1984: accepted with revision February 25. 1985