The role of mononuclear phagocytes in cardiac allograft rejection in the rat

CELLULAR

69, 27 I-280

IMMUNOLOGY

The

Role

( 1982)

of Mononuclear Phagocytes in Cardiac Rejection in the Rat

II. Characterization

of Mononuclear Phagocytes Rat Cardiac Allografts

S. E. CHRISTMAS’ Sir William

Dunn

School

Received

of Parhology. February

Extracted

from

G. G. MACPHERSON~

AND South

3, 1982;

Allograft

Parks

accepted

Road, February

Oxford

OX1 3RE,

England

27, 1982

A method was developed for the extraction of leukocytes infiltrating rat cardiac allografts. Mononuclear phagocytes (MNP) comprised 52.4 f 5.5% of the cells extracted from allografts at the time of rejection (Day 7). Day 4 allografts and Day 7 syngeneic grafts yielded considerably fewer MNP although numbers of lymphoid cells were similar in all three groups. Allograft MNP were phagocytic for latex particles but only very low numbers were adherent to a variety of surfaces, About 50% were positive for attachment and internalization of opsonized sheep red cells via the Fc receptor. However, fewer cells were able to internalize sheep red cells than were able to bind them when complement receptor-mediated phagocytosis was investigated. Large amounts of plasminogen activator were secreted by allograft MNP while cells from syngeneic grafts produced very little. The possible participation of MNP in the effector phase of a mechanism for allograft rejection similar to delayed-type hypersensitivity is discussed. INTRODUCTION Many workers have noted an accumulation of mononuclear phagocytes ( MNP)3 within allografts during rejection (l-7). Also, the numbers of MNP in lymph draining renal allografts in sheep increased greatly as the end-point of rejection approached (8). The function of these cells may merely be the phagocytosis and digestion of effete cells and interstitial debris (2). Alternatively, macrophages may actively participate in the effector phase of the rejection process. There is much evidence that macrophages can, following activation, effect the elimination of neoplastic cells (9-13) and there are also reports of killing of allogeneic cells by ’ Present address: Rheumatic Diseases Centre, Clinical Sciences Building, Hope Hospital, Eccles Old Road, Salford, Manchester M6 8HD, England. ’ To whom correspondence should be addressed. 3 Abbreviations used: DTH, delayed type hypersensitivity; FCS, fetal calf serum; MEM, eagle’s minimal essential medium; MNP, mononuclear phagocytes; NSE-I, nonspecific (a-naphthylbutyrate) esterase; PA, plasminogen activator; PBS, phosphate-buffered saline; SRBC, sheep red blood cells; E, sheep erythrocytes; EA, antibody-coated sheep erythrocytes; EAC, antibody and complement-coated sheep erythrocytes; ip, intraperitoneal; HBSS, Hanks’ balanced salt solution. 271 0008-8749/82/080271-10$02.00/O Copyright 0 1982 by Academic Press. Inc. All rights of reproduction in any form reserved.

272

CHRISTMAS

AND

MAC

PHERSON

sensitized macrophages (14, 15). While the actual mechanism of macrophage-mediated cytotoxicity is unclear it seems likely that secretory products are important. Macrophages synthesize and secrete a wide variety of enzymes and low-molecularweight substances following activation (16, 17). Nathan et al. (18) found that there was a close correlation between hydrogen peroxide secretion and tumor cell killing by activated macrophages whereas Adams et al. (19) obtained a similar correlation between neutral protease secretion and tumor cytotoxicity. There are few reports of macrophage activity within allografts. but a progressive increase in acid phosphatase activity was found in macrophages infiltrating mouse skin allografts (20). Strom et al. (21) used deoxyribonuclease to extract infiltrating cells from rat cardiac allografts. More recently, von Willebrand et al. (22) have reported a technique utilizing collagenase and deoxyribonuclease to purify the cellular infiltrate from rat kidney allografts. The present communication describes a method for the removal of infiltrating cells from rat cardiac allografts which produced a consistently high yield of viable cells. The mononuclear phagocyte component of such infiltrates has been studied with reference to surface markers and functional characteristics (phagocytosis, adherence, and plasminogen activator secretion). MATERIALS

AND

METHODS

Animals. Rats of the inbred strains PVG/c (Hooded; RT-1”) and A0 (Albino Oxford: RT-1Y) were bred in a barrier maintained unit at the Sir William Dunn School of Pathology, Oxford. DBA/2 mice, a strain deficient in c’5, were also bred at the Dunn School. Animals were maintained on normal laboratory diet and water ad libitum. Heart grafts.

Heterotopic cardiac transplantation was performed using a modification of the method of Ono and Lindsey (32). The donor aorta was anastomosed end-to-side to the recipient abdominal aorta and the donor pulmonary artery was joined side-to-side to the recipient inferior vena cava. The beating status of grafts was assessed by palpation through the flanks of recipients. Extraction of cells from cardiac grafts. A transplanted heart was removed from the recipient and perfused ex vivo with PBS at 4°C for about 5 min until the perfusate was clear. The organ was then perfused at a pressure of 100 mm Hg with a mixture of 0.05% collagenase (Sigma, crude type I) and 0.02% deoxyribonuclease (bovine pancreatic, Sigma) in HBSS minus divalent cations, pH 7.2, for 15 min at 37°C. The perfusate was collected and kept on ice. The ventricles of the heart were then dissected free and the thrombus which had usually formed in the left ventricle was carefully removed as this may have been a nonspecific source of inflammatory cells. The ventricles were then minced using scalpel blades and incubated in the same enzyme mixture for a further 15 min at 37°C. The crude extract was filtered through cotton wool to remove undigested fragments and the filtrate was added to the perfusate previously collected. The cells were washed once and their viability tested using trypan blue (33). Cytochemistry and identification of cell types. Smears were made by spinning onto microscope slides lo6 cells in 1 ml FCS using a cytocentrifuge (Shandon Elliott Ltd.) for 5 min at 500 rpm. The slides were air dried and fixed in formaldehyde vapor for 1 min at room temperature. Cells were stained for the presence of anaphthylbutyrate esterase (nonspecific esterase; NSE- 1) using a modification (van

MACROPHAGES

IN CARDIAC

ALLOGRAFT

REJECTION

273

Furth, personal communication) of the method of Scher et al. (34). Smears were stained for 30 min at room temperature in a moist atmosphere and counterstained with 10% Giemsa for 30 sec. Mononuclear phagocytes showed intense, reddish-brown cytoplasmic staining for NSE-1. Most lymphoid cells were negative but some showed one or two positive cytoplasmic granules. A small number (<5%) of neutrophils were also positive, but these were readily identified by their characteristic nuclear morphology. Myocardial cells were identified on the basis of their large size and the presence of myofibrils. Latex phagocytosis. Cells extracted from a cardiac allograft were resuspended in MEM + 10% FCS at a concentration of lO’/ml. Latex particles (Difco Ltd., 1 pm diameter) were added at 109/ml and the cells were incubated for an hour at 37°C. After washing the cells thoroughly to remove unbound latex, cytocentrifuge smears were made and stained for NSE- 1. Mononuclear cells were scored for the presence or absence of internalized latex particles and esterase staining. Adherence of cells extracted from allografts. Cells were resuspended in MEM + 10% FCS at a concentration of 106/ml or lO’/ml. They were then plated out into Linbro plastic tissue culture wells alone or containing 1 l-mm glass coverslips, or were added to Teflon-bottomed “Petriperm” dishes (Heraeus). After incubating for 1 or 18 hr at 37°C nonadherent cells were gently washed off with warm MEM and the adherent preparations were air dried, fixed, and stained for NSE-I. In order to deplete cells extracted from allografts of adherent cells a nylon wool column was used. Fenwal nylon wool (0.3 g) (Leuko-pak Ltd.) was placed in a glass syringe and washed thoroughly with PBS. The column was then incubated with MEM + 10% FCS for 30 min and 2 X 10’ cells were added in 1 ml of medium. After 45 min incubation at 37°C nonadherent cells were eluted with 30 ml of warm MEM + 10% FCS; cell viabilities were always >95%. Fc receptor assays. Sheep red cells (SRBC) were washed three times in MEM and made up to a 5% (packed cell volume) suspension in MEM. Five percent SRBC (0.5 ml) was mixed with 0.4 ml of MEM and 0.1 ml of a 1 in 10 dilution of rabbit IgG anti-SRBC (Nordic) or of a rat antiserum (prepared by repeated ip injections of IO8 SRBC) and incubated for 30 min at 37°C. Opsonized SRBC were then washed twice to remove unbound antibody and made up to 10 ml with MEM. SRBC incubated in the absence of antibody served as a control. Glass coverslips bearing adherent cells extracted from a Day 7 allograft were prepared by incubating them with 10’ cells in 1 ml of MEM for 1 hr at 37°C and rinsing gently with warm MEM to remove nonadherent cells. The adherent cell preparations were then incubated for 90 min at 37°C with opsonized or uncoated SRBC and were rinsed with warm MEM to remove unattached red cells. Some of the preparations were dipped in distilled water for 15 set to lyse any noninternalized SRBC. The coverslips were then dipped in FCS, air dried, fixed in methanol for 5 min, and stained with 10% Giemsa. One hundred cells on duplicate coverslips were scored according to the number of SRBC bound or internalized. Complement receptor assays. Opsonized SRBC were prepared as above but using an IgM anti-SRBC antibody. This was a gift from Dr. C. Bianco and contained no 7 S component. After washing twice, the opsonized or uncoated red cells were incubated for a further 30 min in the presence or absence of 20% DBA/2 serum and washed twice in ice cold MEM. (DBA/2 serum is deficient in C’S.) As before, adherent cell preparations were incubated for 90 min at 37°C with the opsonized

274

CHRISTMAS

AND

MAC

PHERSON

and control SRBC but after rinsing the cells were incubated for a further 30 min in MEM + 30% FCS to promote internalization of the complement-coated red cells (36). The coverslips were then prepared and scored as above. Plasminogen activator secretion assays. These were carried out using the method of Gordon (35). Briefly, Linbro plates coated with r2’I-labeled fibrinogen were incubated for 2-16 hr with MEM + 10% FCS at 37°C to convert the fibrinogen to fibrin. The plates were then washed twice with PBS. Mononuclear leukocytes extracted from cardiac grafts were prepared by density gradient centrifugation. Cells (5 X 10’) in 20 ml PBS were carefully layered over 20 ml Ficoll-Paque (Pharmacia, Uppsala, Sweden) density 1.07 1 g/ml, and spun at 400g for 40 min. The cells remaining at the interface were essentially free of contaminating red cells, granulocytes, and myocardial cells. Mononuclear cells were washed and resuspended in MEM + 5% acid-treated dog serum (35). The cells were then added in triplicate to wells containing 12SI-labeled fibrin and incubated at 37°C. Several different cell concentrations were assayed and acid-treated dog serum was omitted from some wells to test for plasmin-independent degradation of fibrin. Samples (0.1 ml) were removed at intervals and the solubilization of ‘251-labeled fibrin was measured by counting with a Packard Auto-Gamma 5210 scintillation spectrometer. Control wells were incubated with medium alone to test for spontaneous solubilization of 1251-labeled fibrin. Some wells were incubated with 0.125% trypsin to obtain an estimate of the total releasable counts, and results were expressed as the percentage of total counts released. Cell counts. Nucleated cell counts were performed using a Coulter Counter, Model Fn, with a lOO-pm-diameter aperture (Coulter Electronics Ltd., Harpenden). Twenty microliters of the cell suspension to be counted was diluted with 20 ml Isoton, and three drops of Zaponin (Coulter) was added to lyse contaminating red cells. RESULTS Recovery

of Cells Injiltrating

Cardiac

Allografts

The mean rejection time for 10 A0 - PVG/c cardiac allografts was 7.3 +- 0.4 days. By Day 7 all allografts were swollen and indurated and retained little or no TABLE Numbers

of Nucleated

Total cells (X10’) Day 4 allografts

(n = 3)

Day 7 allografts

(n = 18)

Day 7 (n = 2) syngcneic grafts a Results are expressed in the extracts.

Cells

+ 0.25

0.50 (24.7

5.4

+ 1.4 -

2.48 AI 0.26 (52.4 -+ 5.5)

3.2, 3.8

Extracted

Mononuclear phagocytes

2.12

0.58,

f 0.13 + 6.4)

0.87

(24 29) as mean

f SD and figures

1

in parentheses

from

Cardiac

Lymphoid cells

Grafts” Myocardial Granulocytes

CdlS

1.26 f 0.24 (62.4 + 11.9)

0.26 f 0.37 (12.9 2 18.3)

0.70

f 0.20 -

1.67 k 0.25 (35.5 + 5.3)

0.58 + 0.24 (12.1 f 5.1)

0.65

+ 0.24 -

1.6, 1.5

(66.49) are the percentages

0.23, 0.68 (10, 2-7) of the different

0.80,

0.76 -

leukocyte

types

MACROPHAGES

IN CARDIAC

ALLOGRAFT

TABLE NSE-1 NSE-I+

Staining

and Latex

NSE-1’

latex+ 55%

’ Two hundred

Phagocytosis

2

by Mononuclear

latex-

7.5%

275

REJECTION

NSE-I

Cells

from

a Day

latex’ I .5%

7 Allograft”

NSE-

1~ latex 36%

cells counted

palpable beat. The composition of the cellular extract from three Day 4 allografts and eighteen Day 7 allografts is shown in Table 1, together with the results for two Day 7 syngeneic grafts. Despite perfusing the grafts with saline before enzymatic extraction there was significant red cell contamination of the extracts. Viabilities were always ~80% and many of the dead cells were probably myocardial cells as judged by their large size. The most striking difference between groups was the much larger numbers of mononuclear phagocytes (MNP) from Day 7 allografts compared with Day 4 allografts and syngeneic grafts. Numbers of the other cell types were rather similar in all three groups.

Latex Phagocytosis

by Cells Extracted from

Cardiac Allografts

Table 2 shows the results of NSE-1 staining and latex phagocytosis of cells extracted from a Day 7 cardiac allograft. Most NSE-l-positive cells had phagocytosed latex particles and these were presumably macrophages, but a few NSEl-positive cells appeared to be nonphagocytic.

Adherence of MNP

Extracted from Cardiac Allografts

Very few MNP extracted from Day 7 allografts were adherent to plastic, glass, or Teflon and overnight culture did not increase the number of adherent cells. When the same numbers of resident peritoneal cells were plated out under the same conditions about 10 times as many cells adhered as did cells extracted from allografts. This was probably not a result of the enzymatic treatment during extraction of the cells as incubation of resident peritoneal cells under the same conditions had no effect at all on their adhesiveness. Allograft MNP were also poorly adherent to nylon wool columns. Table 3 shows the composition of the cell populations before and after passage through a nylon wool column compared with the corresponding figures for peritoneal cells from PVG/c rats injected with IO8 A0 spleen cells ip 5 days previously.

Fc Receptors on Allograft

MNP

Aliquots of 10” cells extracted from Day 7 cardiac allografts were incubated overnight at 37°C over glass coverslips. After rinsing off nonadherent cells, the remaining adherent cells were >90% NSE-I positive. Table 4 shows the results of assays for Fc receptor-mediated attachment and internalization of opsonized SRBC by Day 7 allograft adherent cells. Resident peritoneal cells prepared in the same way served as controls. About 50% of Day 7 allograft adherent cells were positive for attachment and phagocytosis of SRBC with the rabbit IgG and slightly more

276

CHRISTMAS

AND

MAC

TABLE Adherent

Cell

Depletion

PHERSON

3

Using

a Nylon

Wool

Column”

Percentage Mononuclear phagocytes

nucleated Lymphoid cells

Other

Day 7 allograft cells’ (n = 4)

Unseparated Eluted

48.6 f 4.7 18.3 f 5.3

35.3 f 56.8 +

Allogeneically primed peritoneal cells’ (n = 5)

Unseparated Eluted

52.6 + 7.8 4.8 + 3.0

38.9 + 11.4 79.7 f 11.3

a Results are expressed as means + SD. b “Other cells” comprised neutrophils and myocardial ’ “Other cells” comprised mast cells and eosinophils.

cells

7.1 7.6

cells

18.1 k 24.9 f

7.5 4.1

8.5 f 3.6 15.5 + 14.2

cells.

were positive with the rat antiserum. However, resident peritoneal cells were considerably less efficient at Fc receptor-mediated phagocytosis than they were at binding opsonized SRBC. Complement

Receptors

on Allograft

h4NP

Adherent preparations of Day 7 allograft cells, resident peritoneal cells, and peritoneal cells from a Day 7 allograft recipient were assayed for complementmediated attachment and internalization of opsonized SRBC. As shown in Table 5 fewer allograft MNP were positive for complement-mediated phagocytosis than for attachment but this difference was not as pronounced as that found with resident peritoneal macrophages. However, amongst peritoneal macrophages from a Day 7 allograft recipient all cells with c’ receptors were able to internalize opsonized SRBC although there was a high “background” due to phagocytosis of autologous red cells. Plasminogen

Activator

Secretion

by Allograft

MNP

Mononuclear cells from Day 5 and Day 7 cardiac allografts and Day 7 syngeneic grafts were purified using density gradient centrifugation. The cell populations TABLE Fc Receotors

on Adherent

Cells

from

Day

7 Allografts

4 and on Adherent Percentage

Cell

source

Day 7 allograft Resident oeritoneal

E

cells

3.0 I.5

’ One hundred cells scored on duplicate b Fifteen second dip in distilled H20. ’ ND, not done.

EA (rabbit) 49.5 74.5 preparations.

Resident

Peritoneal

Cells

cells positive”

EA (rabbit) + HXO’ 48.5 48.0

EA (rat)

EA (rat) + H206

53.0 ND’

56.5 ND

MACROPHAGES

IN CARDIAC

ALLOGRAFT

TABLE Complement

Receptors

on Adherent and Peritoneal

Populations Ceils from

5

of Day 7 Allograft a Day 7 Allograft Percentage

Cell

source

Day 7 allograft Resident peritoneal Allograft peritoneal

277

REJECTION

EA

EAC

0.5 0.5 14.0

41.0 64.0 80.5

Cells, Resident Recipient

Peritoneal

Cells,

positive” EAC

+ H20b 29.5 23.5 78.0

y One hundred cells scored on duplicate preparations. b Fifteen second dip in distilled H,O. ’ RBC already present in peritoneal cells.

obtained comprised 50-60s MNP, the remainder of the cells being lymphoid cells with ~1% contamination with neutrophils or myocardial cells. These were then assayed for plasminogen activator (PA) secretion as shown in Fig. 1; resident peritoneal cells were also tested. Mononuclear cells from Day 5 and Day 7 allografts gave high levels of PA secretion; this was not due to plasmin-independent fibrinolysis as omission of plasminogen greatly decreased fibrin solubilization. Day 7 syngeneic graft cells and resident peritoneal cells both gave very low levels of PA secretion. In a separate experiment mononuclear cells from a Day 7 allograft were tested for PA secretion at different cell concentrations and the dose-response obtained is shown in Fig. 2. Although fibrin solubilization was not strictly proportional to cell numbers, there was a greater secretion of PA with increasing cell concentration. It was possible that the enzymatic treatment of allograft MNP during extraction had affected their PA secretion. In order to investigate this, resident peritoneal

Time

in hours

FIG. 1. Plasminogen activator secretion by mononuclear cell populations: +, medium control (no cells); 0, resident peritoneal cells; 0, Day 5 allograft cells; n , Day 7 allograft cells; 8, Day 7 syngeneic graft cells. Means of triplicate samples k SD; lo6 cells per well.

278

CHRISTMAS

AND

MAC

PHERSON

50 x .z .N

40

I 2 c ;: f ?A

30

20

8 10

0

1

2

3 Time

4

5

in hours

FIG. 2. Dose-response of plasminogen activator secretion by Day 7 allograft mononuclear cells: +, medium control (no cells); 0, 3.25 X lo5 cells per well; 8, 6.5 X lo5 cells per well; q , 1.3 X lo6 cells per well. Means of triplicate samples f SD.

cells were incubated for 30 min in the same collagenase and deoxyribonuclease mixture used to extract allograft cells. This had no effect at all upon the basal levels of PA secretion by resident peritoneal cells (data not shown). DISCUSSION An extraction method similar to that used by von Willebrand et al. (22) resulted in a yield of infiltrating leukocytes from Day 7 cardiac allografts which comprised 52.4 it_ 5.5% MNP. This was comparable to the figure of 47% found in kidney allograft extracts (22) but was considerably higher than that found by Strom et al. (21) with heart allografts. The latter used only deoxyribonuclease to extract cells and may have failed to remove macrophages which were strongly adherent to the myocardium. Von Willebrand et al. (22) confirmed that the extracts were similar in cellular composition to sections of intact, rejected allografts and this was also shown in the present study (37). Table 1 shows that syngeneic grafts yielded low numbers of MNP and that in allografts lymphoid cells predominated earlier but MNP increased considerably toward the end-point of rejection. Allograft MNP were apparently phagocytic for latex particles although it was difficult to ascertain whether the latex had actually been internalized by the cells. However, as essentially all the cells associated with latex particles were positive for NSE-1 it seemslikely that these were phagocytic macrophages. The NSE- lpositive nonphagocytic population may have contained dendritic-like cells and fibroblasts as the latter may stain for NSE-1 in sections of cardiac allografts (37). However, only small numbers of allograft MNP were adherent to glass, plastic, Teflon, or nylon wool, and overnight culture did not increase the number of adherent cells. This was probably not due to the action of the enzymes used in the extraction process and the reason for the poor adherence properties of allograft MNP is unknown. About 50% of adherent allograft MNP were positive for binding and internalization of opsonized SRBC via the Fc receptor. However, fewer allograft MNP

MACROPHAGES

IN CARDIAC

ALLOGRAFT

279

REJECTION

were able to internalize opsonized SRBC via the complement receptor than were positive for complement receptor-mediated binding (Table 5). Resident peritoneal cells showed a greater discrepancy between binding and phagocytosis while adherent peritoneal cells from an allograft recipient showed no such difference. It has been reported that complement receptor-mediated ingestion is a marker for macrophage activation (23-25) and by this criterion some allograft MNP were activated. It may be that as many of the allograft MNP were newly arrived from the blood they had not been present in the graft long enough for activation to take place. Alternatively, as such a low percentage of allograft MNP were adherent to glass the coverslip preparations used in these experiments may have constituted an unrepresentative subpopulation of macrophages. Allograft mononuclear cells (presumably MNP) secreted large amounts of PA, while cells from syngeneic grafts and resident peritoneal cells produced very small quantities. Plasminogen activator is a product of activated macrophages (26) and activated T cells can induce macrophages to secrete PA although nonimmunologic stimuli may also be effective (27, 28). It has been proposed that a role of PA secretion by macrophages would be to facilitate their movement into inflammatory sites in which fibrin deposition has occurred (29). This may be relevant in the cardiac allograft model in view of the large amounts of fibrin deposited interstitially immediately prior to rejection (Christmas and Jasani, in preparation). In addition to lysing fibrin, plasmin can activate complement components which may further augment the inflammatory process (17). As well as plasminogen, PA may have other substrates (17) and PA or other neutral proteases (19) may play a part in allograft destruction. In adoptive transfer experiments it has been found that T cells alone were sufficient to bring about cardiac allograft rejection in rats (30). If cytotoxic T cells were responsible, macrophage activation would not be necessary for rejection to take place. However, if a mechanism akin to delayed-type hypersensitivity (DTH) were operative then macrophages would be expected to participate in the effector phase of the response. Although cytotoxic T cells have been isolated from rejecting allografts (21) recent work has indicated that cytotoxic T cells are not required for skin allograft rejection in mice (31). It was found that Lyt-1+,2-,3T cells alone were sufficient to restore the ability to reject skin allografts to irradiated recipients; cytotoxic T cells and their precursors are contained in the Lyt-1+,2+,3+ subpopulation. A similar result has been obtained in the rat (Mason, Dallman, and Webb, in preparation). Thus the demonstration of macrophage activation within rejecting allografts is consistent with a DTH-like mechanism being important in cardiac allograft rejection. ACKNOWLEDGMENTS We are grateful for the skilled technical assistance of Mr. Mrs. P. R. Woodward. S.E.C. was the holder of an M.R.C.

C. D. Jenkins and the excellent Research Studentship.

REFERENCES 1. 2. 3. 4.

Dempster, Dempster, Dempster, Nabarra,

W. J., W. J., W. J., B., and

Brit. J. Exp. Pathol. 58, 418, 1917. Brit. J. Exp. Pathol. 58, 504, 1977. Brit. J. Exp. Pathol. 58, 683, 1917. Descamps, B., Clin. Exp. Immunof. 24, 300,

1976.

typing

of

280

CHRISTMAS

AND

MAC

PHERSON

5. Porter, K. A., in “Pathology of the Kidney” (R. H. Heptinstall, Ed.), 2nd Ed., p. 977. Little Brown, Boston, 1977. 6. Rothwell, T. L. W., and Papadimitriou, J. M., J. Pathol. 107, 235, 1972. 7. Tilney, N. L., &it. J. Surg. 60, 314, 1973. 8. MacPherson, G. G., Murphy, M. J., and Morris, B., Transplantation 24, 16, 1977. 9. Evans, R., and Alexander, P., Immunology 23, 615, 1972. 10. Hibbs, J. B., Lambert, L. H., and Remington, J. S., Nature (London) 235, 48, 1972. 11. Keller, R., J. Exp. Med. 138, 625, 1973. 12. Meltzer, M. S., Tucker, R. W., Sanford, K. K., and Leonard, E. J., J. Nut. Cancer Inst. 54, 1177, 1975. 13. Freedman, V. H., Calvelli, T. A., Sigali, S., and Silverstein, S. C., J. Exp. Med. 152, 657, 1980. 14. Gallily, R., Cell. Immunol. 15, 419, 1975. 15. Debray-Sachs, M., Dy, M., Aurengo, H., and Descamps, B., C.R. Acad. Sci. Paris 286, 249, 1978. 16. Davies, P., and Bonney, R. J., J. Reticuloendothel. Sot. 26, 37, 1979. 17. Nathan, C. F., Murray, H. W., and Cohn, Z. A., N. Engl. J. Med. 303, 622, 1980. 18. Nathan, C. F., Silverstein, S. C., Brukner, L. H., and Cohn, Z. A., J. Exp. Med. 149, 84, 1979. 19. Adams, D. O., Kao, J-J., Farb, R., and Pizzo, S. V., J. Immunol. 124, 293, 1980. 20. Poulter, L. W., Bradley, N. J., and Turk, J. L., Transplantarion 12, 40, 1971. 21. Strom, T. B., Tilney, N. L., Paradysz, J. M., Bancewicz, J., and Carpenter, C. B., J. Immunol. 118, 2020, 1977. 22. von Willebrand, E., Soots, A., and Hayry, P., Cell. Immunol. 46, 309, 1979. 23. Bianco, C., Griffin, F. M., and Silverstein, S. C., J. Exp. Med. 141, 1278, 1975. 24. Mantovani, B., J. Immunol. 126, 127, 1981. 25. Morland, B., and Kaplan G., Exp. Cell Res. 108, 279, 1977. 26. Gordon, S., Fed. Proc. 37, 2754, 1978. 27. Vassalli, J-D., and Reich, E., J. Exp. Med. 145, 429, 1977. 28. Nogueira, N., Gordon, S., and Cohn, Z., J. Exp. Med. 146, 172, 1977. 29. Snyderman, R., and McCarty, G. A., Clin. Rheum. Dis. 4, 507, 1978. 30. Hall, B. M., Dorsch, S. E., and Roser, B., J. Exp. Med. 148, 878, 1978. 31. Loveland, B. E., Hogarth, P. M., Ceredig, R., and McKenzie, I. F. C., J. Exp. Med. 153, 1044, 1981. 32. Ono, K., and Lindsey, E. S., J. Thoruc. Cardiovasc. Surg. 57, 225, 1969. 33. Gorer, P. A., and O’Gorman, P., Trunspl. Bull. 3, 142, 1956. 34. Scher, W., Ansley, H., and Ornstein, L., J. Cell Biol. 70, 254, 1976. 35. Gordon, S., in “Handbook of Experimental Immunology” (D. M. Weir, Ed.), 3rd ed. Blackwell, Oxford, 1978. 36. Munthe-Kaas, A. C., Kaplan, G., and Seljelid, R., Exp. Cell Res. 103, 201, 1976. 37. Christmas, S. E., and MacPherson, G. G., Cell. Immunol. 69, 248-270, 1982.