Role for the class A macrophage scavenger receptor in the ...

Proc. Natl. Acad. Sci. USA Vol. 93, pp. 12456-12460, October 1996 Immunology

Role for the class A macrophage scavenger receptor in the phagocytosis of apoptotic thymocytes in vitro NICK PLATT*t, HIROSHI SUZUKIt, YUKIKO KURIHARAt, TATSUHIKO KODAMAt, AND SIAMON GORDON* *Sir William Dunn School of Pathology, University of Oxford, South Parks Road, Oxford OX1 3RE, United Kingdom; and Department of Internal Medicine, University of Tokyo and Tokyo University Hospital, 7-3-1 Hongo, Tokyo 113, Japan

tThe Third

Communicated by Daniel Steinberg, Unversity of California at San Diego, La Jolla, CA, June 17, 1996 (received for review April 19, 1996)

sensitive terminal deoxynucleotidyltransferase-mediated UTP end labeling (TUNEL) technique to examine the spatial distribution of apoptotic thymocytes, found that the majority of immature thymocytes die in the cortical region of the thymus and that all, the detectable dead cells are localized within thymic M+. This confirmed that thymic MO are the principal stromal cells responsible for the phagocytosis of dying thymocytes and that the uptake is rapid, because it was rare to find "free" or uningested apoptotic thymocytes. While surface receptors of MO have been implicated in recognition of other apoptotic cells (4), receptors for apoptotic thymocytes are currently ill-defined (4). With particular relevance to this study, Sambrano and coworkers (12, 13) have shown that a receptor on peritoneal MO (PM) for oxidized low density lipoprotein (LDL) can mediate the binding of both oxidatively damaged red cells and apoptotic thymocytes in vitro. An analogous relationship has been reported for CD36, which is also a receptor for oxidized LDL (14) and recognizes apoptotic neutrophils (4). Class A MO scavenger receptors (SR-A) are trimeric membrane glycoproteins that bind acetylated LDL and a wide range of polyanionic ligands (15). On the basis of their binding properties, SR-A have been proposed to participate in a number of MO activities that include metabolism of modified lipoproteins, pattern recognition for host defense and cell adhesion (15-17). However, it is not known whether SR-A can act as a phagocytic receptor. Recently we isolated a specific rat mAb (2F8) that inhibited divalent cation-independent adhesion of MO in vitro, immunoprecipitated type I and II SR-A from MO (17), and have demonstrated specific antigen expression by MO in different tissues (18). Immunolocalization of SR-A in sections of normal mouse thymus revealed reactive MO in the medulla and cortex (18). 2F8+ cells also express another MO-specific marker, macrosialin (19), but are weakly positive or negative for the MO molecules, F4/80 (20) and sialoadhesin (21) (unpublished observations). SR-A is a known endocytic receptor (15), although its physiologic ligands are currently unknown, furthermore, a role for SR-A in phagocytosis remains to be determined. As SR-A is expressed on thymic MO and the thymus is a site of extensive T cell apoptosis, we have investigated the potential involvement of SR-A in apoptotic thymocyte clearance.

ABSTRACT Numerous immature thymocytes undergo apoptosis and are rapidly engulfed by phagocytic thymic macrophages. The macrophage surface receptors involved in apoptotic thymocyte recognition are unknown. We have examined the role of the class A macrophage scavenger receptor (SR-A) in the engulfmnent of apoptotic thymocytes. Uptake of steroid-treated apoptotic thymocytes by thymic and inflammatory-elicited SR-A positive macrophages is partially inhibited by an anti-SR-A mAb and more completely by a range of scavenger receptor ligands. Thymic macrophages from mice with targeted disruption of the SR-A gene show a 50% reduction in phagocytosis of apoptotic thymocytes in vitro. These data suggest that SR-A may play a role in the clearance of dying cells in the thymus.

Apoptosis or programmed cell death is now recognized as the physiologic mechanism by which large numbers of unwanted cells are deleted from the body (1). However, examination of tissues with ongoing programmed cell death highlights the scant evidence of dying cells in situ, a paradox that is explained by the existence of mechanisms for the specific and rapid removal of apoptotic cells by phagocytes. One of the earliest markers of commitment to undergo programmed cell death is phagocytosis of the apoptotic cell. In contrast to our knowledge of the process of apoptosis itself, relatively little is known of the clearance process by which dying cells are removed from the body in a manner that has no apparent inflammatory consequence. Examination of apoptotic cells in situ has shown that they are taken up and degraded by phagocytes, in particular macrophages (M+) (2). This activity can be observed in MO-like cells from species such as Drosophila, suggesting this is an important and conserved process (3) The process is specific, in that only cells committed to die are phagocytosed, and rapid, so that uptake is completed before the integrity of the apoptotic cell membrane is lost. Two essential components are required for the successful ingestion of the apoptotic cells: (i) specific receptor(s) on the surface of the phagocyte to mediate rapid recognition and ingestion and (ii) the presence of appropriate ligands on apoptotic cells that permit their distinction from healthy neighbors. The consensus that can be drawn from the relatively small number of reported studies on apoptotic cell recognition is that just like the regulation of apoptosis itself, it is complex and that there is not a single receptor-ligand system that explains apoptotic cell removal by all phagocytes (4). The thymus is the organ where the repertoire of mature T cells is selected from a much larger number of immature thymocytes and extensive apoptotic cell death occurs in the immature thymocyte populations. Although numerous immature thymocytes undergo apoptosis (5-7), few dead cells are observed in situ due to rapid engulfment by phagocytic M4 in the thymic stroma (8-10). A recent study (11), using the

MATERIALS AND METHODS Isolation and Cultivation of MO Populations. Thymic tissue from 4-week-old Balb/c mice was cut into small pieces, washed three times with RPMI medium 1640, and digested with 0.05% collagenase (Boehringer Mannheim) for 30 min at room temperature. The digest was washed, layered on fetal calf serum (FCS), and clusters were recovered by sedimentation under gravity before plating on glass coverslips in RPMI Abbreviations: MO, macrophage; PM, peritoneal MO; TPM, thioglycollate-elicited PM; SR-A, MO scavenger receptor class A; FCS, fetal calf serum; LDL, low density lipoprotein. tTo whom reprint requests should be addressed.

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

12456

Proc. Natl. Acad. Sci. USA 93 (1996)

Immunology: Platt et aL b

12457

phagocytosed labeled thymocytes scored. A minimum of 200 cells were counted in random fields from at least two independent coverslips. Comparable data were obtained in three or more independent experiments. Data are expressed as mean ± SD. TPM were also plated on coverslips coated with either BSA or maleylated BSA (29), cultured, and assayed as above. Immunocytochemistry. Immunostaining of thymic MO with 2F8 mAb was performed as described in ref. 18.

FIG. 1. Expression of SR-A by thymic MO. (a) Isolated thymic M4 rosetted with thymocytes. (b) Individual 2F8+ thymic MO (brown) with bound thymocytes (violet).

RESULTS AND DISCUSSION

medium 1640 with 10% FCS. Thioglycollate-elicited PMs (TPM) were harvested by peritoneal lavage 4 days after injection of 1 ml of Brewer's complete thioglycollate broth. Cells were washed with PBS, plated on glass coverslips, and cultivated overnight in RPMI medium 1640 with 10% FCS. Preparation of Labeled Apoptotic Thymocytes. Thymuses from 3- to 4-week-old Balb/c mice were mechanically dissociated to a single cell suspension, cultured in RPMI medium 1640 with 10% FCS and 1 AM dexamethasone (Sigma) for 6 hr, washed with PBS, and labeled with the fluorescent dye PKH2 according to the manufacturer's instructions (Sigma). In Vitro Phagocytosis Assay. Thymic M4s or TPM were incubated at 4°C for 30 min in the presence or absence of the following reagents: 2F8 (25 ,g/ml- 1), Campath-lG, rat IgG 2b isotype-matched control (22) (25 ,ug/ml-'), anti-CR3 mAb 5C6 (20 pug/ml-1), anti-vitronectin receptor mAb 2C11 (20 gg/ml-'; PharMingen), poly(G) (50 gg/ml-1; Pharmacia), poly(C) (50 ,ug/ml-1; Pharmacia), poly(A) (50 ,ug/ml-1; Pharmacia), dextran sulfate (100 pug/ml-1; Sigma), chondroitin sulfate (100 Ag/ml-1; Sigma), carrageenan (Sigma, 100 ,mg/ ml-'), polyvinyl sulfate (50 gg/ml-1; Sigma), fucoidan (100 ,tg/ml-1; Sigma), BSA (100 gg/ml-1; Boehringer Mannheim), and maleylated BSA (29) (100 gg/ml-), overlaid with a 30-fold excess of fluorescent apoptotic thymocytes and incubated at 37°C for 90 min. Nonphagocytosed thymocytes were removed by washing with PBS. Coverslips were examined by fluorescence microscopy and the percentage of MO that had

Thymic MO were isolated with minimal contamination by other stromal cell types, such as epithelia, by collagenase digestion, and by adherence to glass coverslips. Clusters of thymic MO with bound immature, nonapoptotic thymocytes were recovered (Fig. la). These MO expressed 2F8 (Fig. lb), class II MHC, macrosialin (19), and F4/80 (20) strongly, but were weakly positive or negative for sialoadhesin (21) and type 3 complement receptor (ref. 23 and data not shown). 2F8 staining was not restricted to the plasma membrane and was also present intracellularly. RT-PCR analysis (data not shown) suggests that thymic MO express mRNA for both type I and type II receptors (15). Thus, the phenotype of isolated thymic MO in culture was consistent with that observed in situ. We developed an in vitro phagocytosis assay to evaluate potential involvement of SR-A in clearance of apoptotic thymocytes. Cultured thymic MO were stripped free from bound thymocytes, incubated with fluorescently-labeled apoptotic thymocytes and uptake measured as the frequency of MO that have taken up dye. In the time frame of the assay, the internalized thymocytes are rapidly degraded and the lipophilic dye becomes incorporated into the membranes of the Mo. In experiments with freshly harvested thymocytes (which contained 70% apoptotic, >90% viable), showed evidence of ingestion (Fig. 2 Ab and B) and this was reduced -50% by saturating concentrations of intact 2F8 mAb (Fig. 2Ac and B). It is of interest that a comparable degree of inhibition was observed when 2F8 was used to block endocytic uptake of diI-labeled acetylated LDL by RAW MO (14). Rosetting of steroid treated thymocytes to thymic M4 was not apparently

affected by 2F8 (not shown), indicating that SR-A functions in uptake but not initial binding of apoptotic thymocytes. However inhibition of binding of apoptotic thymocytes by the mAb may be masked by the simultaneous rosetting of healthy, nonapoptotic cells. The binding of nonapoptotic thymocytes is not affected by preincubation with 2F8 (data not shown). Therefore, whether SR-A functions in the initial interaction of apoptotic cells with MO remains to be definitively established. Pretreatment with the SR-A ligands (25), polyguanylic acid, or

A B Wild Type

SR-A Null

20

40

60

80

% Thymic Macrophages Fluorescent

FIG. 4. (A) Uptake of fluorescent apoptotic thymocytes by thymic MO from SR-A-deficient mice. (a) Uptake by thymic MO from background strain 129. (b) Uptake by thymic M4 from SR-A null animals. 2F8 staining of thymic MO from 129 strain (c) and SR-A-deficient mice (d). (B) Histogram summarizing uptake of apoptotic thymocytes by thymic MO from 129 and SR-A-deficient animals.

Immunology: Platt et al. fucoidan also greatly inhibited phagocytosis of apoptotic thye and f and B), whereas the structurally related non-SR-A ligand polycytidylic acid and an isotypematched control mAb were without effect (Fig. 2 Ad and B). MO incubated in the presence of both polycytidilic acid and a ribonuclease inhibitor as a precaution against ribonucleotide degradation, also showed no effect on thymocyte uptake (not shown). The fact that the SR-A ligands inhibit to a greater extent than the 2F8 mAb suggests that other thymic MO receptors, to which these molecules may also bind, could be involved in the phagocytosis of apoptotic thymocytes. TPM that phagocytose apoptotic lymphocytes in vitro (26) express SR-A (18) and are readily available in high yield were used to test a broader range of known SR-A ligands (25). TPM were treated with potential inhibitors, incubated with labeled apoptotic thymocytes, and dye uptake was examined after 90 min. Under these conditions, -50% of the TPM engulf thymocytes (Fig. 3Aa and B). Preincubation with 2F8 reduced uptake by 50%, comparable inhibition to that observed in thymic MO (Fig. 3 Ad and B), whereas an isotype-matched control mAb (CP) and anti-type 3 complement receptor mAb (5C6) were ineffective (Fig. 3B). A mAb (2C11) specific for the vitronectin receptor, known to mediate uptake of apoptotic neutrophils by MO (27), did not reduce uptake, consistent with a previous study (28) that vitronectin receptor is not involved in clearance of dying lymphocytes by TPM. Because TPM that had been incubated with 2F8 showed greatly reduced ingestion of thymocytes, we confirmed the involvement of SR-A by testing two additional classes of molecules that competitively bind to SR-A and related compounds that do not (15). Although the degree of inhibition varied with the ligand (range 60-90%, Fig. 3B), all the known SR-A ligands (polyguanylic acid, polyinosinic acid, maleylated BSA, fucoidan, carrageenan, dextran sulfate, and polyvinyl sulfate) reduced uptake, whereas all those that are not ligands (polyadenylic acid, polycytidilic acid, BSA, and chondroitin sulfate) were without effect. SR-A are known to bind to polyribonucleotides, polysaccharides, and modified proteins, whereas class B scavenger receptors are reported to be only inhibited by modified proteins (15). The finding that the broader range of ligands were all capable of inhibiting apoptotic cell uptake is compatible with the blocking of SR-A and other scavenger receptors and not through other MO surface receptors. The most effective inhibitors, fucoidan and polyvinyl sulfate reduced thymocyte ingestion by almost 90%, indicating that either SR-A and other scavenger receptors are responsible for apoptotic cell ingestion and/or that it is an essential part of a multireceptor event that requires all components to be active. When TPM are cultured on glass coverslips coated with the ligand maleylated BSA, SR-A is redistributed to the basal surface and the apical surface is depleted of receptor, thus making it inaccessible to other ligands (29). TPM on this surface displayed reduced endocytosis of the modified lipoprotein ligand, acetylated LDL (not shown), and only minimal phagocytosis of apoptotic thymocytes (Fig. 3B). Cultivation on native BSA had no effect on either activity (Fig. 3B). To test directly the requirement for SR-A in the ingestion of apoptotic thymocytes we examined the phagocytic activity of thymic MO derived from mice lacking this receptor. Mice lacking both type I and type II SR-A were established by targeted disruption of exon 4 of SR-A gene in A3-1 ES cells by the method described (30). Immunostaining using 2F8 indicated that homozygote mice for SR-A deficiency completely lack both type I and type II receptor protein and PM obtained from mice indicated very low (18%) acetyl LDL mocytes (Fig. 2 A

degradation compared with their wild-type littermates (H.S.,

Y.K., and T.K., unpublished observation). They grow normally and are fertile. Their establishment and phenotype will be reported in full elsewhere. Thymic MO from SR-A-deficient

Proc. Natl. Acad. Sci. USA 93 (1996)

12459

mice, which have no discernible 2F8 immunoreactivity (Fig. 4A4d), showed a 50% reduction in their ability to phagocytose-

labeled apoptotic thymocyted relative to controls (Fig. 4 Ab and B). This degree of inhibition is comparable to that seen with the 2F8 mAb (Fig. 4B), demonstrating that this receptor accounts for -50% of the phagocytic activity required for the uptake of apoptotic thymocytes in vitro. When thymic MO are prepared from SR-A-deficient mice, they are recovered rosetted with immature, apparently healthy thymocytes, confirming that SR-A is not required for the binding of nonapoptotic thymocytes (not shown). These data provide the first evidence for the involvement of a known receptor, SR-A, in the clearance of apoptotic thymocytes. The residual phagocytic activity of these thymic MO suggest that at least one or more additional receptors are involved in this process that are partially sensitive to SR-A anionic ligands. In addition, there is also the possibility of heterogeneity in the ligands expressed by apoptotic thymocytes. The contribution of these additional MO receptors may account for the lack of a grossly altered thymic phenotype observed in a preliminary examination of the mutant mice (data not shown). We have not yet established if SR-A functions in the removal of other dying or damaged cells. A recent study has implicated a receptor for oxidized LDL in the recognition of oxidatively damaged erythrocytes by PM (31), but preincubation with 2F8 did not reduce binding, suggesting SR-A was not involved in this system. It is also of interest that the class B scavenger receptor CD36, can convey the capacity for enhanced phagocytosis of apoptotic cells when transfected into a deficient cell type (32). This receptor is not recognized by 2F8 (17) or apparently inhibited by the full range of polyanions used in this study (15) and would therefore not appear to be involved, but the relative nonspecificity of these inhibitory ligands does not exclude the possibility that it contributes to the residual SR-A-independent uptake. Recently, a new category of scavenger receptor, class C, that has different ligand specificities, has been isolated from Drosophila and shown to be expressed in MO-like cells that contained apoptotic bodies (33). As yet there are no data to support the existence or function for vertebrate homologs of this molecule, but they are candidate phagocytic receptors that deserve consideration. Importantly our evidence presented here shows that SR-A is essential for full phagocytic activity for which other receptor(s) cannot compensate in vitro. Further studies are required to identify the additional receptor(s) required for apoptotic cell clearance in the thymus, the nature of the ligands and consequences of SR-A-mediated phagocytosis of apoptotic thymocytes by thymic Mo. We thank Ruth Galily for help and important observations in the early stages of this work; Iain Fraser for sharing reagents; Liz Darley for histology; and Rosangela da Silva and Fran Platt for helpful discussions. N.P. is supported by the Medical Research Council, United Kingdom. 1. Raff, M. C. (1992) Nature (London) 356, 397-400. 2. Wyllie, A. H., Kerr, J. F. R. & Currie, A. R. (1980) Int. Rev. Cytol. 68, 251-306. 3. Tepass, U., Fessler, A. & Hartenstein, V. (1994) Development (Cambridge, U.K) 10, 1829-1837. 4. Savill, J., Fadok, V., Henson, P. & Haslett, C. (1993) ImmunoL Today 14, 131-136. 5. von Boehmer, H. (1990) Annu. Rev. Immunol. 6, 309-326. 6. McPhee, D., Pye, J. & Shortman, K. (1979) Thymus 1, 151-157. 7. Egerton, M., Scollay, R. & Shortman, K. (1990) Proc. Natl. Acad. Sci. USA 87, 2579-2582. 8. Duvall, E. & Wyllie, A. H. (1986) Immunol. Today 7, 115-119. 9. Duijvestijn, A. M. & Hoefsmit, E. C. M. (1981) Cell Tissue Res. 218, 279-292. 10. Kyewski, B. A., Rouse, R. V. & Kaplan, H. S. (1982) Proc. Natl. Acad. Sci. USA 79, 5646-5650.

12460

Immunology: Platt et al.

11. Surh, C. D. & Sprent, J. (1994) Nature (London) 372, 100-103. 12. Sambrano, G. R., Parthasarathy, S. & Steinberg, D. (1994) Proc. Natl. Acad. Sci. USA 91, 3265-3269. 13. Sambrano, G. R. & Steinberg, D. (1995) Proc. Natl. Acad. Sci. USA 92, 1396-1400. 14. Endemann, G., Stanton, L. W., Madden, K. S., Bryant, C. M., White, R. T. & Protter, A. A. (1993) J. Biol. Chem. 268, 1181111816. 12. Krieger, M. & Herz, J. (1994) Annu. Rev. Biochem. 63, 601-637. 13. Kodama, T., M., Rohrer, L., Zabrecky, J., Matsudaira, P. & Krieger, M. (1990) Nature (London) 343, 531-535 . 14. Fraser, I., Hughes, D. H. & Gordon, S. (1993) Nature (London) 364, 343-364. 15. Hughes, D. H., Fraser, I. & Gordon, S. (1995) Eur. J. Immunol. 25, 466-473. 16. Holness, C. L., da Silva, R. P., Fawcett, J., Gordon, S. & Simmons, D. L. (1993) J. Biol. Chem. 268, 9661-9666. 17. Austyn, J. M. & Gordon, S. (1981) Eur. J. Immunol. 11, 805-815. 18. Crocker, P. R., Mucklow, S., Bouckson, V., McWilliam, A., Willis, A. C., Gordon, S., Milon, G., Kelm, S. & Bradfield, P. (1994) EMBO J. 10, 1661-1665. 19. Hale, G. (1987) J. Immunol. Methods 103, 59-67. 20. Rosen, H. & Gordon, S. (1987) J. Exp. Med. 166, 1685-1701.

Proc. Natl. Acad. Sci. USA 93 (1996) 21. Lyons, A. B., Samuel, K., Sanderson, A. & Maddy, A. H. (1992) Cytometry 13, 809-821. 22. Krieger, M. (1992) Trends Biochem. Sci. 17, 141-146. 23. Duvall, E., Wyllie, A. H. & Morris, R. G. (1985) Immunology 56, 351-358. 24. Savill, J., Dransfield, I., Hogg, N. & Haslett, C. (1990) Nature (London) 343, 170-173. 25. Fadok, V., Savill, J. S., Haslett, C., Bratton, D. L., Doherty, D., Campbell, P. A. & Henson, P. M. (1992) J. Immunol. 149, 4029-4035. 26. El Khoury, J., Thomas, C. A., Loike, J. D., Hickman, S. E. M., Cao, L. & Silverstein, S. C. (1994) J. Biol. Chem. 269, 1019710200. 27. Kurihara, Y., Kurihara, H., Suzuki, H., Kodama, T., Maemura, K., Nagai, R., Oda, H., Kuwaki, T., Cao, W.-H., Kamata, N., Jishage, K., Ouchi, Y., Azuma, S., Toyoda, Y., Ishikawa, T., Kumada, M. & Yazaki, Y. (1994) Nature (London) 368,703-710. 28. Ottnad, E., Parthasarathy, S., Sambrano, G., Ramprasad, M. P., Quehenberger, O., Kondratenko, N., Green, S. & Steinberg, D. (1995) Proc. Natl. Acad. Sci. USA 92, 1391-1395. 29. Ren, Y., Silverstein, R. L., Allen, J. & Savill, J. (1995) J. Exp. Med. 181, 1857-1862. 30. Pearson, A., Lux, A. & Krieger, M. (1995) Proc. Natl. Acad. Sci. USA 92, 4056-4060.