NK lymphocytes lyse certain xenogeneic cells without prior sensitization. The receptors by which NK cells recognize xenogeneic targets are largely uncharacterized but have been postulated to possess broad specificity against ubiquitous target ligands. However, previous studies suggest that mouse NK cells recognize xenogeneic targets in a strain-specific manner, implicating finely tuned, complex receptor systems in NK xenorecognition. We speculated that mouse Ly-49D, an activating NK receptor for the MHC I ligand, H2-Dd, might display public specificities for xenogeneic target structures. To test this hypothesis, we examined the lysis of xenogeneic targets by mouse Ly-49D transfectants of the rat NK cell line RNK-16 (RNK.Ly-49D). Of the xenogeneic tumor targets tested, RNK.Ly-49D, but not untransfected RNK-16, preferentially lysed tumor cells derived from Chinese hamsters and lymphoblast targets from rats. Ly-49D-dependent recognition of Chinese hamster cells was independent of target N-linked glycosylation. Mouse Ly-49D also specifically stimulated the natural killing of lymphoblast targets derived from wild-type and MHC-congenic rats of the RT1lv1 and RT1l haplotypes, but not of the RT1c, RT1u, RT1av1, or RT1n haplotypes. These studies demonstrate that Ly-49D can specifically mediate cytotoxicity against xenogeneic cells, and they suggest that Ly-49D may recognize xenogeneic MHC-encoded ligands.
The mechanisms of cell-mediated xenograft rejection are not fully understood and remain an obstacle to xenogeneic transplantation (1). Successful strategies for the prevention of hyperacute xenograft rejection have fostered renewed interest in dissecting the mechanisms of delayed cellular responses to xenogeneic grafts (1). At least one component of this cellular response is mediated by NK lymphocytes. NK cells can lyse certain xenogeneic target cells in vitro, and NK cells form a large component of the cellular infiltrate associated with xenograft rejection in vivo (2, 3, 4, 5, 6). Depletion of NK cells from recipients of xenografts results in prolongation of graft survival in some experimental systems (7, 8, 9). Interspecies NK cell rejection can be most clearly demonstrated in xenogeneic bone marrow transplantation models. While the successful engraftment of allogeneic marrow can be achieved using a specialized conditioning regimen of selective T cell depletion, the successful engraftment of xenogeneic bone marrow requires additional NK cell ablation (10, 11, 12). “Xenogeneic resistance” to rat bone marrow grafts by mice is genetically regulated by dominantly inherited, strain-specific genetic elements (13, 14, 15, 16). These strain-specific elements involved in xenogeneic resistance appear to control graft rejection by radioresistant mouse NK cells (16, 17, 18, 19).
There is only limited knowledge regarding activating NK cell receptors involved in the recognition of xenogeneic cells (20). We have previously demonstrated that the rat NKR-P1A receptor is specifically required for lysis of the mouse tumor cell line IC-21 (21). The IC-21 target structure recognized by rat NKR-P1A has not yet been defined. Recent studies by Idris et al. have shown that the differential lysis of Chinese hamster ovary (CHO)3 cells by NK cells from C57BL/6 and BALB/c mice is controlled by a polymorphic dominant genetic locus that they have termed Chok (22). This locus has been mapped to the NK complex, or NKC, on mouse chromosome 6, which encodes a superfamily of structurally related lectin-like receptors implicated in the control of NK cell cytotoxicity (23). These genetic data suggest that an NKC-encoded activating receptor may stimulate the cytotoxicity of CHO targets by mouse NK cells. Known activating NK cell receptors in this region include NKR-P1C (NK1.1 Ag), Ly-49D, and Ly-49H (23). The strain-specific cytotoxicity of CHO cells described by Idris did not correlate with strain-specific differences in NKR-P1C expression on NK cells, and neither anti-Ly-49D nor anti-Ly-49H Abs had any effect on the cytotoxicity of IL-2-activated NK cells against CHO targets (22).
The mouse Ly-49D receptor has previously been shown to activate NK cytotoxicity (24). We and others have demonstrated that the activating Ly-49D receptor is specific for the mouse classical MHC class I Ag H2-Dd, but not H2-Db, Kd, or Kk (25, 26). We speculated that mouse Ly-49D might display public specificities for xenogeneic, as well as allogeneic, target ligands.
In the current study, we examined the role of mouse Ly-49D in the recognition of xenogeneic cells. We examined the lysis of xenogeneic targets by Ly-49D transfectants of the rat NK tumor line RNK-16. In our studies, we demonstrate that Ly-49D recognizes a target structure on Chinese hamster lines, as well as an MHC-encoded target structure on lymphoblasts from RT1l and RT1lv1 haplotype rats. These studies demonstrate that the NK cell activating receptor Ly-49D can specifically mediate cytotoxicity against certain xenogeneic cells, and they suggest that Ly-49D may recognize xenogeneic MHC ligands.
Materials and Methods
RNK-16, a spontaneous NK cell leukemia from F344 rats, was the gift of Craig Reynolds (National Cancer Institute, Frederick, MD) and was adapted for in vitro growth in RPMI 1640 supplemented with 10% heat-inactivated FCS, 25 μM 2-ME, 2 mM l-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin (cRPMI) (27). Tumor target cell lines CHO; the glycosylation-deficient mutants of CHO: Lec 1, Lec 2, and Lec 8 (28); BHK-21 (Syrian hamster kidney epithelium); CHL/IU (Chinese hamster lung); DDT1MF-2 (Syrian hamster leiomyosarcoma); HaK (Syrian hamster kidney); BL-3 (bovine lymphosarcoma); 104 C1 (guinea pig, transformed fetus); Daudi (human Burkitt lymphoma); K562 (human erythroleukemia); Jurkat (human T cell leukemia); ST (porcine, fetal testis); and YB2/0 (rat plasmacytoma hybrid, RT1u) were obtained from American Type Culture Collection (Manassas, VA). Tumor targets were maintained in cRPMI. The mouse Ly-49D transfectant of RNK-16, RNK.Ly-49D, and the Ly-49A transfectant of RNK-16, RNK.Ly-49A, have been previously described and were maintained in cRPMI supplemented with 1 mg/ml of G418 (Boehringer, Indianapolis, IN) (25, 29). Transfected RNK-16 effectors were grown in cRPMI without G418 for at least 2 days before functional assays.
Abs and flow cytometry
Monoclonal Abs to mouse Ly-49D ((12A8 (cross reactive with Ly-49A), rat IgG2a)), OVA (2C7, rat IgG2a), H2-Db (B22.249), H2-Dd (34-5-8S, mouse IgG2a), and phosphotyrosine (APT, 4G10, mouse IgG2b) were produced from their respective hybridoma lines. Abs were partially purified from ascites by ammonium sulfate precipitation as previously described. F(ab′)2 were generated by pepsin digestion, and completion of digestion was verified by SDS-PAGE and silver staining (30). For fluorescence analysis, mAbs were used at a concentration of 1 μg/106 cells. Routine analysis was performed by using a FACScan (Becton Dickinson, San Diego, CA).
Specific lysis of NK targets was determined by using a standard 4-h 51Cr release assay as previously described (31). Briefly, target cells were harvested and labeled for 1 h at 37°C with 200 μCi of [51Cr]sodium chromate (Amersham, Arlington Heights, IL) in cRPMI. Labeled target cells were washed and resuspended at 105 cells/ml, and 0.1 ml of this cell suspension was added to each well of 96-well plates containing 0.1 ml of effector cells at the indicated E:T ratios. Plates were incubated at 37°C for 4 h, then centrifuged for 5 min. Then, 100 μl of supernatant was counted in a gamma counter and the specific cytotoxicity was calculated as described (31). All assays were performed in triplicate. For Ab inhibition studies, effector cells were preincubated for 15 min at room temperature with F(ab′)2 at a concentration of 25 μg/106 effectors or with intact Ab at a concentration of 10 μg/106 effectors before the addition of targets. RNK-16 effectors at 5 × 106 cells/ml were preincubated in cRPMI containing intact Ab at 50 μg/ml. Lymphokine-activated killer (LAK) cell effectors at 2 × 106 cells/ml were preincubated in cRPMI containing intact Ab at 20 μg/ml. Ab was not removed before addition of target cells. Effectors were then serially diluted to enable stated E:T ratios.
C57BL/6 and BALB/c mice were obtained from The Jackson Laboratory (Bar Harbor, ME) and were used at 6–8 wk of age. F344, PVG, DA, AO, BN, and LOU rats were purchased from Harlan Sprague-Dawley (Indianapolis, IN). F344 and LEW rats were obtained from Harlan U.K. (Bicester, U.K.). MHC-congenic PVG rats PVG.1LV1, PVG.1L, PVG.1AV1, and PVG.1U were obtained from the Babraham Institute (Cambridge, U.K.). All animals were housed in approved animal care facilities in accordance with institutional guidelines.
IL-2-activated NK cells
IL-2-activated NK cells were prepared as previously described from fresh C57BL/6 and BALB/c mouse splenocytes (32). 12A8-positive and -negative IL-2-activated NK cells were isolated by panning, using a variation of a previously described separation protocol (33). Briefly, day 6 IL-2-activated NK cells were panned with the anti-Ly-49D mAb, 12A8. Nonadherent cells were then treated with rabbit anti-rat Ig (Cappel, Malvern, PA), followed by treatment with rabbit complement (Cedarlane, Westbury, NY) for 1 h at 37°C. Ly-49D− and Ly-49D+ cell populations were then cultured overnight in cRPMI supplemented with 1000 U/ml human IL-2 (National Cancer Institute). Cells were washed extensively with HBSS with 3% FCS on day 7, replated, and used for assays on day 9. This resulted in populations of NK cells that were >95% pure as assessed by their expression of NK1.1 (not shown).
Con A blast preparation
Con A-stimulated blasts were prepared from rat splenocytes, using methods previously described (34). Briefly, rat spleens were harvested aseptically and separated into single-cell suspensions. Following lysis of RBC, splenocytes were washed in cRPMI and placed in culture at a density of 1 × 106 cells/ml in cRPMI, supplemented with 3 μg/ml Con A (Sigma, St. Louis, MO). Following 48 h of culture at 37°C, cells were harvested, purified over Ficoll-Hypaque, washed twice in cRPMI, and labeled for use as targets in cytotoxicity assays.
Target cell stimulation, immunoprecipitation, and Western blotting
For studies of NK cell activation by targets, 9 × 107 effector cells were mixed with 1.8 × 108 target cells in a total volume of 1 ml cRPMI in microcentrifuge tubes. Cell suspensions were immediately centrifuged for 30 s at 50 × g, then incubated at 37°C for the indicated time. Cells were then centrifuged at 500 × g for 10 s, and cell pellets were resuspended in cold lysis buffer (20 mM triethanolamine, pH 7.8, 150 mM NaCl, 1 mM EDTA, 1 mM MgSO4, 2.5 mM CaCl2, 0.01% azide, 1 mM sodium orthovanadate, and protease inhibitors) with 1% digitonin and 0.12% Triton-X100 as described (35). For pervanadate stimulation, 3 × 107 cells/sample were incubated for 5 min at 37°C in cRPMI with 0.03% H2O2 and 100 mM sodium orthovanadate (pervanadate), and cell pellets were resuspended in 1 ml cold lysis buffer, as previously described (29). Lysates were precleared with protein A-Sepharose beads (Pharmacia, Piscataway, NJ) that had been previously coated with rabbit anti-rat Ab (Cappel, Malvern, PA) and control mAb 2C7 (rat IgG2a) for 2 h at 4°C, then immunoprecipitated overnight at 4°C with protein A beads coated with rabbit anti-rat Ab and anti-Ly-49D mAb 12A8. After washing with cold wash buffer (50 mM Tris, pH 8, 150 mM NaCl, 5 mM 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate and protease inhibitors), immunoprecipitates were resolved by 15% SDS-PAGE under reducing conditions and transferred to polyvinylidene difluoride membranes (Immobilon-P; Millipore, Marlborough, MA). After blocking with TBST (10 mM Tris pH 8, 150 mM NaCl, 0.05% Tween 20) containing 10% horse serum, the membranes were incubated with 1 μg/ml of anti-phosphotyrosine mAb (4G10) in TBST with 10% horse serum for 1 h at room temperature. Following extensive washing in TBST, blots were incubated in HRP-conjugated rabbit anti-mouse Ig (Amersham) and developed using chemiluminescence Super Signal (Pierce, Rockford, IL), according to the manufacturers’ instructions.
Lysis of xenogeneic target cells by RNK.Ly-49D effectors
To study xenogeneic target recognition by the mouse Ly-49D receptor in the absence of other activating and inhibitory mouse NK receptors, we used our previously described mouse Ly-49D transfectant of RNK-16 (25). We examined natural killing of a panel of xenogeneic target cells, comparing the cytotoxicity of RNK.Ly-49D to that of wild-type RNK-16 or RNK-16 transfected with the inhibitory mouse alloreceptor, Ly-49A (RNK.Ly-49A) (29). As shown in Table I, RNK.Ly-49D cells lysed two of the tested xenogeneic targets: hamster CHO cells and F344 strain rat Con A blasts. Neither of these xenogeneic targets were lysed by RNK-16 or by RNK.Ly-49A cells in at least four duplicate experiments, while these effectors readily lysed standard YAC-1 tumor targets. As previously published, RNK.Ly-49D effectors do not lyse YAC-1 targets (25). RNK.Ly-49D effectors failed to lyse other tested xenogeneic tumor targets derived from cows, pigs, guinea pigs, or humans. Thus, the Ly-49D-associated activation of cytotoxicity appears to be species and/or cell-line specific.
|Species .||Target Cell .||% Cytotoxicitya .||.||.||.||.||.|
|.||.||RNK-16 (E:T) .||.||RNK.Ly-49A (E:T) .||.||RNK.Ly-49D (E:T) .||.|
|.||.||(50) .||(12.5) .||(50) .||(12.5) .||(50) .||(12.5) .|
|ConA blasts F344||1||1||0||1||11||9|
|Species .||Target Cell .||% Cytotoxicitya .||.||.||.||.||.|
|.||.||RNK-16 (E:T) .||.||RNK.Ly-49A (E:T) .||.||RNK.Ly-49D (E:T) .||.|
|.||.||(50) .||(12.5) .||(50) .||(12.5) .||(50) .||(12.5) .|
|ConA blasts F344||1||1||0||1||11||9|
Standard 4-h cytotoxicity assays were performed using RNK-16, RNK.Ly-49A, or RNK.Ly-49D effectors against the target cells listed in column 2. Percent cytoxicity at an E:T ratio of 50:1 and 12.5:1 is shown for each effector cell line.
Lysis of CHO cells by RNK.Ly-49D cells is mediated by the Ly-49D receptor
To demonstrate that lysis by RNK.Ly-49D effectors is mediated through the transfected Ly-49D receptor, we examined the effect of anti-Ly-49D mAb blockade on cytotoxicity against CHO target cells. As shown in Fig. 1 A, cytotoxicity by RNK.Ly-49D was blocked by anti-Ly-49D mAb 12A8 but was unaffected by isotype-matched control mAb 2C7. Lysis by RNK-16 was not induced by the presence of either Ab. These studies confirm that lysis of CHO target cells is specifically dependent on the Ly-49D receptor.
We next examined the ability of CHO target cells to stimulate signaling events activated by the Ly-49D receptor. Previous studies by Smith et al. and Mason et al. have shown that Ly-49D stimulation leads to the tyrosine phosphorylation of the associated signaling chain DAP12 (36, 37). McVicar et al. demonstrated that mAb stimulation of Ly-49D in RNK.Ly-49D transfectants leads to tyrosine phosphorylation of DAP12, which can be coprecipitated with the Ly-49D receptor (38). Therefore, we examined the effect of CHO target stimulation on the tyrosine phosphorylation of Ly-49D-associated DAP12 in RNK.Ly-49D transfectants. Fig. 1 B shows an anti-phosphotyrosine Western blot of anti-Ly-49D (12A8) immunoprecipitates from RNK.Ly-49D cells stimulated with pervanadate or with CHO target cells for the indicated durations. Consistent with previously published studies, lane 1 shows coprecipitation of a 12- to 16-kDa tyrosine-phosphorylated protein (DAP12) with Ly-49D from pervanadate-stimulated RNK.Ly-49D cells (38). In lanes 5–7, tyrosine-phosphorylated DAP12 coprecipitates with Ly-49D in lysates from RNK.Ly-49D effectors stimulated with CHO targets for 1, 2, or 5 min. CHO-induced DAP12 tyrosine phosphorylation peaks at 2 min and diminishes by 15 min (lane 8). In unstimulated RNK.Ly-49D cells (lane 3), tyrosine-phosphorylated DAP12 does not coprecipitate with Ly-49D. Tyrosine phosphorylation of DAP12 is minimal, but detectable, in RNK.Ly-49D cells lysed immediately after the addition of CHO targets (0 time point, lane 4), which indicates extremely rapid DAP12 phosphorylation following receptor stimulation. The ability of CHO targets to induce the brisk tyrosine phosphorylation of Ly-49D-associated signaling molecules indicates that Ly-49D serves not merely as an adhesion molecule, but rather specifically activates NK cells upon the recognition of a target structure on CHO.
Ly-49D recognition of CHO cells is not affected by alterations in target cell glycosylation
The recognition of polymorphic carbohydrate differences between xenogeneic cells is important in some mechanisms of xenogeneic rejection, particularly hyperacute rejection (1). Because Ly-49D contains a putative carbohydrate-binding lectin domain, we speculated that this mouse alloreceptor might bind carbohydrate ligands on xenogeneic CHO targets. To examine whether surface glycosylation of CHO targets was critical for their interaction with Ly-49D, we examined the Ly-49D-dependent killing of three glycosylation-mutants of CHO, Lec 1, Lec 2, and Lec 8 (Fig. 2). Lec 1 is a CHO cell line mutated in GlcNAc-glycosyl transferase, resulting in no detectable complex type N-linked oligosaccharides, Lec 2 is defective in the transport of CMP-sialic acid resulting in sialic acid-deficient cell-surface glycoproteins, and Lec 8 is deficient in the translocation of UDP-galactose, resulting in the defective expression of galactose-specific glycoforms (28). As seen in Fig. 2, RNK.Ly-49D effectors lyse CHO, Lec 1, Lec 2, and Lec 8 target cells equivalently. Lysis of all four cell lines is blocked by anti-Ly-49D mAb 12A8, but not by isotype-matched control mAb 2C7. Therefore, these three CHO mutants lines are all recognized by the Ly-49D receptor, indicating that none of the alterations in glycosylation substantially affects target recognition by Ly-49D.
We and others have previously demonstrated that Ly-49D recognizes the murine MHC I allele H2-Dd (25, 26). Thus, we speculated that Ly-49D might recognize a related hamster MHC I ligand on CHO cells. Direct testing of this hypothesis was difficult because of the lack of available hamster MHC I cDNAs and a paucity of anti-hamster MHC mAbs. A murine anti-Db/Ld mAb, B22.249, cross-reacts with CHO cells and is presumed to bind to hamster MHC I (Dr. Keith Gould, unpublished observations). However, F(ab′)2 B22.249 failed to block the lysis of CHO by RNK.Ly-49D cells (data not shown). It is possible that the B22.249 mAb does not identify the ligand on CHO cells recognized by Ly-49D. However, the lack of B22.249 mAb blockade does not completely rule out a possible role for hamster MHC, as it may be that the B22.249 epitope is not directly involved in Ly-49D recognition. MHC polymorphism is reported to be limited in Syrian hamsters (Mesocricetus auratus), but the MHC has not been extensively examined in Chinese hamsters (Cricetulus griseus) (39). We examined Ly-49D-mediated lysis of several additional hamster lines, CHL-IU, a lung carcinoma from Chinese hamsters, and several lines from Syrian hamsters: BHK-21, a kidney epithelial line, DDT1MF-2, a leiomyosarcoma line, and HaK, a kidney epithelial line. CHO and CHL-IU cells demonstrated similar levels of B22.249 expression by FACS, while BHK-21, DDT1MF-2, and HaK cells showed no B22.249 mAb staining, suggesting that Syrian and Chinese hamster MHC I are serologically divergent (data not shown). RNK.Ly-49D effectors were able to lyse CHO cells and to lyse CHL-IU weakly, and lysis of both Chinese hamster lines was blocked by anti-Ly-49D mAb (12A8). In contrast, RNK.Ly-49D failed to lyse BHK-21, DDT1MF-2, or HaK targets from Syrian hamsters (data not shown). Although Ly-49D appeared to discriminate between tumors derived from Chinese and Syrian hamsters, the species-specific xenogeneic ligand recognized by Ly-49D could not be identified from these experiments.
CHO targets are lysed by Ly-49D+ and by Ly-49D− NK cells from C57BL/6 mice
We next examined the role of Ly-49D in the lysis of CHO targets by IL-2-activated mouse NK cells. We separated NK cells from C57BL/6 mice into Ly-49D+ and Ly-49D− populations. Completely pure populations were difficult to obtain, and Ly-49D-enriched populations routinely contained 10–15% Ly-49D− cells, while Ly-49D-depleted populations contained 3–5% Ly-49D+ cells (data not shown). We examined the lysis of CHO cells by these separated effectors. For comparison, bulk C57BL/6 and BALB/c IL-2-activated NK effectors were also tested, as these have been previously shown to lyse CHO targets differentially (22). As shown in Fig. 3, Ly-49D+ effectors killed CHO more efficiently than Ly-49D− effectors, but Ly-49D− NK cells still lysed CHO cells. The addition of blocking anti-Ly-49D mAb 12A8 had no effect on the lysis of CHO cells by either the 12A8+ or 12A8− effectors (data not shown). 12A8 mAb also failed to influence the lysis of CHO by bulk C57BL/6 or BALB/c NK effectors (data not shown). These experiments demonstrate that mouse NK cells expressing Ly-49D have higher levels of cytotoxicity against CHO than do Ly-49D− NK cells, which nonetheless kill CHO with moderate efficiency. These studies highlight the difficulty in working with purified NK cell populations, which may express activating and inhibitory receptors in mixed combinations. Because the 12A8 mAb effectively blocks CHO lysis by RNK.Ly-49D effectors, but not by Ly-49D+ IL-2-activated NK cells, it appears that mouse NK cells from C57BL/6 mice may possess CHO-specific activating receptors in addition to Ly-49D.
Ly-49D mediates lysis of rat Con A blasts from specific rat strains
In addition to CHO, the xenogeneic Con A lymphoblasts from F344 rats were preferentially lysed by RNK.Ly-49D (Table I). To further investigate this finding, we examined the strain specificity of Ly-49D by using Con A lymphoblast targets isolated from seven different rat strains. As shown in Fig. 4, RNK.Ly-49D effectors lysed Con A blasts from F344 (MHC haplotype RT1lv1) and LEW (RT1l) rats, but not from PVG strain (RT1c) rats. Lysis of F344 and LEW Con A blasts required the Ly-49D receptor, as lysis was blocked by anti-Ly-49D mAb but not by control Ab. RNK.Ly-49D failed to lyse Con A blasts isolated from LOU (RT1u), DA (RT1av1), AO (RT1u), or BN (RT1n) rats (data not shown). These data show that mouse Ly-49D augments the lysis of rat lymphoblasts in a strain-specific manner.
Ly-49D recognizes an MHC-encoded ligand on rat lymphoblasts
As shown in Fig. 4, Con A blasts from PVG (haplotype RT1c) rats are not lysed by RNK.Ly-49D, while blasts from F344 (RT1lv1) and LEW (RT1l) are killed by RNK.Ly-49D. Because Ly-49D recognizes a specific mouse MHC Ag, we speculated that mouse Ly-49D may also recognize a polymorphic MHC-encoded structure on rat lymphoblasts. Therefore, we tested control RNK-16 and RNK.Ly-49D effectors against Con A blasts from MHC-congenic resistant rats on the PVG background. As seen in Fig. 5, Con A blasts isolated from the MHC-congenic strains PVG.1LV1 (RT1lv1) and PVG.1L (RT1l) rats were readily lysed by Ly-49D-transfected RNK-16, and cytotoxicity was specifically blocked by anti-Ly-49D mAb. In contrast, Con A blasts from PVG.1AV1 (RT1av1) or from PVG.1U (RT1u) rats were not killed by RNK.Ly-49D. These studies provide clear evidence that xenogeneic MHC-controlled ligands common to RT1lv1 and RT1l rat lymphoblasts are recognized by mouse Ly-49D. Notably, these haplotypes carry the same alleles at the classical MHC Ia region RT1.A, the MHC II region RT1.B/D and the TAP loci, while they differ at the nonclassical MHC class Ib RT1.C region. Thus, Ly-49D recognizes either an RT1.A-B/D-encoded ligand or possibly an RT1.C-encoded target structure common to the RT1lv1 and RT1l haplotypes. Therefore, these experiments demonstrate that the mouse Ly-49D alloreceptor can function as a receptor for xenogeneic rat MHC-encoded Ags.
We and others have previously shown that mouse Ly-49D is an activating alloreceptor for the classical mouse MHC I allele H2-Dd (25, 26). In the current studies, we demonstrate that Ly-49D can also activate NK cell cytotoxicity in response to xenogeneic ligands from selected species of hamsters and strains of rats. Our studies using hamster target lines demonstrate that Ly-49D recognizes a ligand from Chinese, but not Syrian, hamsters and that Ly-49D recognition of Chinese hamster cells is independent of N-linked target glycosylation. These data expand on previous genetic studies by Idris et al., who examined the differential killing of CHO by NK cells from high-responder C57BL/6 mice and low-responder BALB/c mice. In back-crossed mice, the cosegregation of CHO killing (the Chok locus) with polymorphisms within the NKC suggested that the Chok gene product might be an activating member of the NK lectin receptor superfamily (22). Ly-49D is identified by the mAb 12A8 and 4E5 in high-responder C57BL/6, but neither of these mAb bind to LAK cells from the low-responder BALB/c, mice (our unpublished observations). Though BALB/c mice may express a Ly-49D-like gene, it does not appear to share ligand specificity for a CHO target structure. Our studies show that the Ly-49D lectin-like receptor is a likely Chok gene product, but it is not the only receptor controlling CHO lysis. CHO cells are effectively lysed by Ly-49D− NK cells, which suggests that other activating receptors on C57BL/6 NK cells may participate in CHO killing. Unexpectedly, anti-Ly-49D mAb 12A8 failed to even partially block CHO killing by Ly-49D+ NK cells. This may have been due to the possibility that blocking of Ly-49D-mediated lysis could not be observed in the presence of strong lysis mediated through other receptors; however, we cannot rule out that either inadequate Ab was present to completely block all Ly-49D-mediated interactions because the 12A8 Ab would also interact with Ly-49A receptors or that blocking of Ly-49A receptors also affected lysis of CHO cells. A paucity of specific hamster reagents has prevented the localization of the specific ligand on CHO cells recognized by Ly-49D.
Our studies also demonstrate that mouse Ly-49D activates NK cytotoxicity against F344 and LEW rat lymphoblasts, but not against lymphoblasts from PVG, DA, LOU, AO, or BN rats. The availability of MHC-congenic rat strains has enabled us to identify the Ly-49D rat ligand as an MHC-encoded structure common to RT1lv1 and RT1l rat strains. These data fully support early xenogeneic transplantation studies of LEW rat bone marrow into supralethally irradiated mice. In these studies, F344 or LEW bone marrow was rejected by irradiated (Ly-49D+) C57BL/6 mice, but not by (Ly-49D−) C3H, CBA, BALB/c, or A/Sn mouse strains (13, 14, 15, 16). Also, bone marrow from BN rats was similarly rejected by irradiated A/Sn and BALB/c mice but not by irradiated C57BL/6, C3H, or CBA mice (16). These strain-specific responses were dominantly inherited, and in retrospect it is likely that radioresistant Ly-49D+ NK cells were mediators of “xenogeneic resistance” against F344 and LEW blasts (16, 40). Moreover, these early data suggest that NK cells from A/Sn and BALB/c mice might express activating receptors specific for BN lymphoblasts.
The high lytic capacity of NK cells for xenogeneic targets initially led investigators to the hypothesis that NK cell killing was not restricted by MHC (41). It has been proposed that NK cells express multiple activating receptors against ubiquitous protein or carbohydrate ligands on susceptible targets (42). The cytolytic capacity of these activating receptors would be regulated by inhibitory receptors for self-MHC. Because these inhibitory receptors should not bind to xenogeneic MHC, foreign targets should be lysed (43). However, other studies indicate that recognition of xenogeneic targets is more specific. Our previous studies using an NKR-P1-loss mutant of rat RNK-16 showed that the lectin-like rat NKR-P1A receptor can specifically induce the killing of some, but not all, xenogeneic mouse targets (21). However, the mouse target ligands recognized by rat NKR-P1A have not been identified. This study implied that the recognition of xenogeneic targets by NK cells might be mediated by multiple ligand-specific activating receptors on NK cells. The xenogeneic target spectrum of NK cells would thus be determined by the composite effects of specific activating NK receptors with divergent specificities for xenogeneic target structures.
Our current studies demonstrate that, in at least some species and strain combinations, an activating receptor on NK cells displays strain specificities for polymorphic xenogeneic ligands, and that one xenogeneic ligand for Ly-49D is encoded within the rat MHC. Though these studies may suggest that Ly-49D can directly interact with xenogeneic MHC Ags, it is also possible that other xenogeneic MHC-encoded ligands are involved in recognition by Ly-49D. Like NK cell allorecognition, NK cell recognition of xenogeneic cells is not likely to be entirely dependent on target cell MHC class I expression. In a previous study, xenorecognition of mouse and human cells by rat NK cells showed lysis of xenogeneic cells deficient in MHC class I (44).
Recognition of rat or hamster ligands by mouse Ly-49D could reflect a coincidental cross-reactivity of xenogeneic MHC with H2-Dd, the murine ligand for Ly-49D. Alternatively, activating NK receptors for xenogeneic MHC might be evolutionarily conserved molecules that promote the rapid rejection of an acquired inoculum of foreign cells. Aside from the fundamental role of the immune system to discriminate self from nonself in higher organisms, NK cells might be physiologically important in the rapid elimination of xenogeneic targets. Xenogeneic cells may indeed find their way into the bloodstream of higher organisms as a result of interspecies predation or defensive behaviors. Because T cells are educated to recognize foreign Ags in the context of self-MHC, they might not directly recognize xenogeneic MHC. Thus, the elimination of xenogeneic targets by T cells might require xenoantigen presentation followed by T cell expansion. In contrast, NK cells could eliminate xenogeneic cells rapidly and efficiently through specialized activating receptors for foreign MHC.
Regardless of their role in nature, xenogeneic receptors on NK cells are likely to be important in the cellular responses during xenotransplantation, particularly bone marrow transplantation. Our studies indicate that xenoreceptors on NK cells likely exhibit a high degree of species and strain specificity that has not been previously recognized. Moreover, they suggest that NK xenorecognition may involve the entire complement of host activating NK receptors, restricted, in part, by xenogeneic MHC ligands (3).
While this paper was in revision, Idris et al. (45) also identified Ly-49D as the Chok gene product.
This work was supported by the U.S. Veterans Administration, the Norwegian Cancer Society, the Norwegian Research Council and the International Human Frontiers in Science Program (HFSPO RG-309). M.C.N. is supported by National Institutes of Health Grant K11 AR01927, the Arthritis Foundation, and the American Cancer Society. W.E.S is the recipient of National Institutes of Health Grant RO1 CA69299, and J.C.R. is the recipient of National Institutes of Health Grant R29 CA60944. G.W.B. is supported by grants from the Biotechnology and Biological Sciences Research Council through the Babraham Institute.
Abbreviations used in this paper: CHO, Chinese hamster ovary; NKC, NK complex; cRPMI, complete RPMI 1640 medium; LAK, lymphokine-activated killer.