Mapping of Quantitative Trait Loci Determining NK Cell-Mediated Resistance to MHC Class I-Deficient Bone Marrow Grafts in Perforin-Deficient Mice This information is current as of September 29, 2021. Maria H. Johansson, Mesha A. Taylor, Maja Jagodic, Katalin Tus, John D. Schatzle, Edward K. Wakeland and Michael Bennett J Immunol 2006; 177:7923-7929; ; doi: 10.4049/jimmunol.177.11.7923 Downloaded from http://www.jimmunol.org/content/177/11/7923

References This article cites 63 articles, 29 of which you can access for free at: http://www.jimmunol.org/ http://www.jimmunol.org/content/177/11/7923.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists by guest on September 29, 2021 • Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2006 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Mapping of Quantitative Trait Loci Determining NK Cell-Mediated Resistance to MHC Class I-Deficient Bone Marrow Grafts in Perforin-Deficient Mice1

Maria H. Johansson,2,3*† Mesha A. Taylor,3* Maja Jagodic,‡ Katalin Tus,§ John D. Schatzle,* Edward K. Wakeland,§ and Michael Bennett*

NK cells reject allogeneic and MHC class I-deficient bone marrow (BM) grafts in vivo. The mechanisms used by NK cells to mediate this rejection are not yet thoroughly characterized. Although perforin plays a major role, perforin-independent mech- anisms are involved as well. C57BL/6 mice deficient in perforin (B6 perforin knockout (PKO)) reject class I-deficient TAP-1 KO BM cells as efficiently as normal B6 mice. In contrast, perforin-deficient 129S6/SvEvTac mice (129 PKO) cannot mediate this rejection while normal 129 mice efficiently reject. This suggests that in 129, but not in B6, mice, perforin is crucial for NK Downloaded from cell-mediated rejection of MHC class I-deficient BM grafts. To identify loci linked to BM rejection in perforin-deficient mice, we ؋ generated backcross 1 progeny by crossing (129 B6)F1 PKO mice to 129 PKO mice. In transplantation experiments, >350 backcross 1 progeny were analyzed and displayed a great variation in ability to reject TAP-1 KO BM grafts. PCR-based mic- rosatellite mapping identified four quantitative trait loci (QTL) on 2, 4, and 8, with the QTL on 8 showing the highest significance, as well as a fifth epistatic QTL on chromosome 3. This study describes the first important step toward identifying BM graft resistance (s). The Journal of Immunology, 2006, 177: 7923–7929. http://www.jimmunol.org/

atural killer cells are the primary effector cells mediating CD3␨,orFc␧RI␥ (2). The adapter molecules contain ITAMs that acute rejection of bone marrow (BM)4 grafts (1). They become phosphorylated upon cross-linking of the receptor, allow- N also carry out cytotoxic and cytokine-mediated activity ing recruitment of downstream signaling molecules like SYK and against tumor cells and cells infected with virus, intracellular bac- Zap-70 (3). However, recognition of MHC class I-deficient tumor teria, or parasites. These activities are strictly regulated by acti- cells such as the T cell lymphoma RMA-S can occur in the absence vating and inhibitory receptor-ligand interactions. Inhibitory re- of SYK and Zap-70, (4). The activating NK cell receptor NCR-1 ceptors on NK cells prevent killing of healthy endogenous cells via (the mouse homolog of human NKp46) has recently been shown to by guest on September 29, 2021 recognition of self MHC molecules, but allow killing of MHC participate in recognition of RMA-S cells (5). However, this re- class I-deficient cells. In contrast, activating receptors on NK cells ceptor was not essential for rejection and it is thus still not clear trigger lysis of target cells expressing ligands induced by tumor which activating receptors and ligands mediate the recognition of transformation and/or infection. Most activating receptors associ- these cells, as well as of MHC class I-deficient BM grafts in the ate with small adapter molecules e.g., DAP10, KARAP/DAP12, mouse. At least two mechanisms for NK cell-mediated lysis have been defined in recent years, calcium-dependent release of perforin and *Department of Pathology, University of Texas Southwestern Medical Center, Dallas, granzymes and interaction between death receptors and their li- TX 75390; †Department of Microbiology, Tumor and Cell Biology and Strategic Research Center for Studies of Integrative Recognition in the Immune System, Karo- gands. The function of perforin in concert with granzymes appears linska Institutet, Stockholm, Sweden; ‡Center for Molecular Medicine, Department of to be a primary cytotoxic mechanism of NK cells. Conjugate for- Clinical Neuroscience, Neuroimmunology Unit, Karolinska Institutet, Stockholm, mation occurs between the effector and the target cell resulting in Sweden; and §Center for Immunology, University of Texas Southwestern Medical Center, Dallas, TX 75390 degranulation and release of perforin and granzymes which coop- Received for publication October 18, 2005. Accepted for publication September erates in inducing DNA fragmentation and apoptosis of the target 19, 2006. cell (6, 7). Perforin as a key mediator of NK cell-mediated cytol- ϩ The costs of publication of this article were defrayed in part by the payment of page ysis was first demonstrated by the lack of killing of Fas YAC-1 charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. lymphoma cells, which are highly NK cell sensitive, by NK cells derived from perforin-deficient (perforin knockout (PKO)) mice 1 This work was supported by National Institutes of Health Grants CA36922, CA70134, and AI38938. M.H.J. was supported by a postdoctoral fellowship from the (8). Thus, perforin deficiency results in effector cells with de- Swedish Cancer Society and by grants from Karolinska Institutet, the Åke Wiberg creased ability to reject certain tumors and clear virus infections (9, Foundation, the Magnus Bergwall Foundation, the Royal Swedish Academy of Sci- ences, and the Swedish National Board of Health and Welfare. 10). Alternatively, engagement of death receptors on the target cell 2 Address correspondence and reprint requests to Dr. Maria H. Johansson, Depart- leads to apoptosis mediated via death domain-containing ment of Microbiology, Tumor and Cell Biology, Karolinska Institutet, Box 280, SE- and caspase 8. Thus, NK cells trigger target cell apoptosis via 17177 Stockholm, Sweden. E-mail address: [email protected] death receptor ligands such as FasL and TRAIL (11–14). Interest- 3 M.H.J. and M.A.T. contributed equally to the study. ingly, it has been demonstrated that NK cells have the ability to 4 Abbreviations used in this paper: BM, bone marrow; KO, knockout; PKO, perforin up-regulate cell surface levels of Fas on previously negative target ␤ ␤ KO; wt, wild type; QTL, quantitative trait ; 2m, 2-microglobulin; BC, back- cells before Fas-dependent lysis in vitro and in vivo (15). The cross; LOD, logarithm of odds; bmgr, bone marrow graft rejection; NKC, NK gene complex; PLC, phospholipase C; PKC, kinase C; MMP, matrix metallopro- mechanisms behind this regulation are unknown. It has also been teinase; GBP, guanylate-binding protein. shown that TRAIL is important for NK cell-mediated protection

Copyright © 2006 by The American Association of Immunologists, Inc. 0022-1767/06/$02.00 7924 GENETIC MAPPING OF PERFORIN-INDEPENDENT BM REJECTION LOCI

FIGURE 1. Distribution of 125I-labeled UdR uptake ϫ in F1 129 PKO mice suggests polygenic influence on BM resistance. Lethally irradiated hosts were inoculated i.v. with 3.5 ϫ 106 donor BM cells. Host spleens were harvested after 5 days to assess hemopoietic cell re- population. Growth of transplants is represented as the percentage of 125I-labeled UdR uptake. In the graph, individual host mice are represented by circles and geo- metric mean values by horizontal lines. A, TAP-1 KO, ϫ B6 PKO, 129:B6 PKO, and (129:B6 B6)F1 PKO mice were grafted with TAP-1 KO donor BM. F1 mice were significantly different from 129:B6 PKO and TAP-1 KO, but not from B6 PKO (p Ͻ 0.05). One rep- resentative experiment of nine is shown. B, TAP-1 KO ϭ ϭ ϫ ϭ (n 54), 129 PKO (n 48), (129 B6)F1 PKO (n 55), and (F ϫ 129) PKO (BC-1 PKO; n ϭ 357) mice

1 Downloaded from were grafted with TAP-1 KO donor BM. The figure shows a summary of 12 experiments. C, The graph shows the distribution of 125I-labeled UdR uptake (%) in different mouse types. Horizontal bars below the ϫ graph denotes the (F1 129) PKO individuals selected for genotypic analysis. http://www.jimmunol.org/

from metastasis of certain tumors and that this protection is de- nisms involved in BM graft rejection. However, they do not ex- pendent on IFN-␥-induced up-regulation of TRAIL on NK cells plain the difference in rejection capacity between mouse strains. (16). In addition, TRAIL is responsible for NK cell-mediated elim- Besides the above-described strain differences in perforin-indepen- ination of immature dendritic cells in vivo showing that also non- dent BM rejection under conventional conditions, the genetic by guest on September 29, 2021 transformed cells can be killed via TRAIL (17). NK cell-mediated background has also proven to be important in mice kept under cytolysis dependent on membrane-bound or secreted TNF has also specific pathogen-free conditions. Although 129:B6 PKO mice been reported (18). Although these pathways represent important were unable to reject B6.lprTAP-1 KO BM grafts (negative for mechanisms of cytotoxicity, much remains to be studied when it Fas), B6 PKO mice readily rejected such grafts (21). Furthermore, comes to their role in NK cell-mediated BM rejection. in another in vivo assay measuring rejection of large numbers The role of perforin in NK cell-mediated BM rejection has been ␤ ␤ Ϫ/Ϫ of fluorescence-labeled 2-microglobulin ( 2m) spleen cells, somewhat controversial. Previous reports have suggested that per- wild-type (wt) 129 mice displayed impaired rejection compared forin is not necessary for rejection of BM grafts from allogeneic with wt B6 mice (M. H. Johansson, unpublished data). These re- mice (19). However, using BM cells that were TAP-1 deficient, sults emphasize the importance of genetic differences between the and thereby highly NK susceptible, we have shown that NK cells two mouse strains in regulation of rejection of hemopoietic cells, from perforin-deficient (PKO) mice of the 129:B6 strain have a also under specific pathogen free conditions. The functional dif- decreased ability to reject these grafts. In contrast, B6 PKO ferences between B6 PKO and 129 PKO mice may be explained mice rejected the grafts. Interestingly, this was only true when by genetic influence directly on killing mechanisms alternative to the mice were housed in a conventional facility: perforin-deficient perforin. Alternatively, the explanation could lie in a differential mice kept under specific pathogen-free conditions were able to reject regulation of NK cell recognition or activity in a more general MHC class I-negative BM grafts regardless of genetic background (20). This result emphasized two factors that affect BM rejection in way, but with changes subtle enough to result in functional dif- PKO mice: housing conditions and genetic background. Regarding ferences only under suboptimal conditions like in the absence of the housing conditions, we recently demonstrated involvement of perforin. Therefore, we set out to investigate how the genetic back- the Fas/FasL pathway in resistance to BM grafts and that dimin- ground of the mouse controls NK cell-mediated BM graft rejection ished BM rejection capacity during conventional housing condi- in the absence of perforin and under conventional housing tions may be explained by down-regulation of NK cell FasL conditions. expression (21). This down-regulation could be induced by We performed a whole genome scan of backcross mice display- pathogens infecting the animals (22), since sentinel mice in our ing extreme BM graft rejection phenotypes and describe four conventional animal facility occasionally carry viral infections, quantitative trait loci (QTLs) mapping to chromosomes 8, 2, and 4. e.g., mouse hepatitis virus, mouse parvovirus, and sendai virus. In A fifth epistatic QTL, located on chromosome 3 and dependent on addition, the inhibitor cellular FLIP, overexpressed in susceptible the genotype at the QTL on chromosome 4, is also described. By hemopoietic stem cells, was shown to convey partial engraftment this study, we have taken the first important steps toward identi- in B6 PKO mice, suggesting involvement of death receptors in BM fying involved in resistance to MHC class I-deficient BM rejection (23). These results suggest a redundancy in the mecha- grafts. The Journal of Immunology 7925

Materials and Methods Mice All mice were bred and maintained in a conventional mouse colony at the University of Texas Southwestern Medical Center as previously described (20). C57BL/6 (B6) and 129S6/SvEvTac (129) mice both carry the H-2b MHC haplotype. Perforin-deficient B6 PKO mice were originally pur- chased from The Jackson Laboratory. 129 PKO mice were derived from 129:B6 PKO mice and were backcrossed to 129S6/SvEvTac mice seven generations. Perforinϩ/Ϫ mice were then intercrossed, offspring tested for perforin, and PKO homozygous mice were bred and used for experiments. TAP-1 KO mice have a mixed background of 129 and B6 genes. Founder mice of the 129 background were crossed once to B6 and the mice have since then been inbred (Ͼ10 generations). All work using animals was reviewed and approved by the Institutional Animal Care and Research Advisory Committee at the University of Texas Southwestern Medical Center. BM transplantation Transplantation was done as previously described (20, 24, 25). Experimen- tal groups in general contained five mice with only a few exceptions where

no less than three mice were included in a group. Briefly, recipient mice Downloaded from were lethally irradiated (850 cGy) and injected i.v. with 3.5 ϫ 106 TAP-1 KO donor BM cells. Five days after cell transfer, growth of donor-derived BM cells was assessed by spleen incorporation of 125I-labeled UdR (Am- ersham Biosciences) which is a specific DNA precursor and thymidine analog. The results are expressed as the geometric mean (95% confidence level) percentage uptake of 125I-labeled UdR. A high percentage of 125I- labeled UdR uptake represents growth of the BM graft and a low percent-

125 http://www.jimmunol.org/ age of I-labeled UdR uptake denotes rejection. Using log10 values, para- metric and nonparametric statistical analyses were performed to determine significance of differences between geometric mean values. The Student t test was used to compare two groups and the Newman-Keuls multiple comparison test was used to determine significantly different groups at p ϭ FIGURE 2. Multiple QTLs influence BM graft rejection in PKO mice. 0.05 level for experiments that consist of multiple recipient groups for a A, The whole-genome scan for perforin-independent BM rejection per- Ͻ ϫ given donor. Values significantly ( p 0.05) different from another group formed in 137 (F1 129) PKO mice. The y-axis indicates LOD scores and by nonparametric and parametric analysis are indicated in the figure the x-axis chromosomal positions. Vertical lines on the x-axis represent legends. positions of microsatellite markers used for genotyping. Horizontal dashed Ͻ Ͻ Genetic mapping and statistical analysis lines indicate significant (p 0.05) and suggestive (p 0.10) threshold levels determined by permutation tests. Three significant QTL on chromo- by guest on September 29, 2021 Genomic DNA was isolated from tail tips from backcross-1 (BC-1) prog- somes 2 and 8 and one suggestive QTL on chromosome 4 were detected. eny and control mice. Genotypes were determined by PCR amplification of B, Four independent QTLs, bmgr1–4, regulate BM rejection. Gray bars loci polymorphic between 129 and B6 that contain simple sequence length represent 95% confidence intervals for chromosomal location of QTLs as polymorphisms, i.e., microsatellite markers. Primers were obtained from determined by bootstrapping. The y-axis indicates LOD scores and the Invitrogen Life Technologies/ResGen. Standard PCR was performed at experimentally determined optimal annealing temperature for each primer x-axis chromosomal positions. pair. PCR products were visualized with ethidium bromide on 5% agarose gels to determine the genotype at each marker position (129 homozygous experiment to which it is applied (27). The 95th percentile of the distri- or 129/B6 heterozygous). Marker positions were obtained from The Jack- bution of maximum LOD scores serves as a genome-wide threshold for ͳ ʹ son Genome Database ( www.informatics.jax.org ) us- significance, while the 90th percentile is a threshold for suggestive QTLs. ing positions from the Whitehead Institute Genetic Map. In total, 78 mark- Ninety-five percent confidence intervals, i.e., the most probable genomic ers distributed throughout the 19 autosomes were analyzed (see Fig. 2). location for each QTLs were determined using a bootstrap resampling ϳ These markers allowed a genome coverage of 75%, taking into account method (28). This method uses repeated resampling with replacement an area of 10 cM on each side of a marker. (i.e., in each resampling random animals from the data set are duplicated All linkage analysis and permutation tests were performed using R/qtl and others are removed), each resample is then subjected to linkage analysis. software (26). BC-1 mice with the most extreme rejection phenotypes, 46 The 95% confidence interval for each QTL is determined from the resulting mice in each group, were selected for initial scan of the genome analyzing distribution of maximum LOD score positions. Possible interactions be- 61 markers distributed throughout the 19 autosomes. We analyzed BM tween genes were tested by implementing a two-dimensional scan with a rejection as a binary trait: rejecting mice were scored as 0 and accepting two-QTL model using a multiple imputation method that examines all pairs mice as 1. Loci of interest after the initial linkage analysis were more of genetic markers and intermarker positions for association with ability to carefully investigated by PCR analysis of additional markers and including reject BM grafts. To confirm the identified QTLs and interactions, we all 137 mice displaying the extreme phenotypes. We conducted linkage implemented a fit multiple-QTL model that is based on the creation of an analysis with the aim to localize genomic regions linked to phenotype by initial model of phenotypic variance comprising all identified QTLs and analyzing coinheritance of genetic markers and phenotype. Linkage anal- their interactions. Each QTL or QTL combination is then excluded from ysis was performed using a multiple imputation method that uses infor- the initial model in subsequent steps and the resulting influence on phe- mation in the marker genotype to simulate multiple versions of complete notypic variance is determined. A significant difference ( p Ͻ 0.05) be- genotype information. This is a powerful tool to analyze sample material tween the influence on phenotypic variance caused by the whole set of where part of the material is only analyzed for genotype in certain regions. QTLs and the influence when one has been excluded, confirms the rele- Threshold levels for logarithm of odds (LOD) score significance of puta- vance of the excluded component. tive QTLs were determined by a permutation test procedure (27). In this procedure, the set of data is subjected to repeated random shuffling of phenotypes among the individual genotypes and the maximum LOD score Results obtained in each shuffle is recorded. These reflect maximum LOD scores A wide range of rejection capacities in BC-1 mice suggests that can occur by chance, because the relationship between the phenotype polygenic control of BM graft rejection and genotype is broken in the shuffle (i.e., each animal has been assigned the wrong phenotype while retaining its genetic map). This test for signif- We have shown that perforin can play a crucial role in BM graft icance is thus empirical and reflects the characteristics of the particular rejection. Perforin-deficient mice on the 129:B6 background were 7926 GENETIC MAPPING OF PERFORIN-INDEPENDENT BM REJECTION LOCI

Table I. Summary of QTLs linked to BM graft rejection

QTL Chromosome Peak Marker LOD Score Rejection Allele Candidate Genes

bmgr1 8 D8Mit249 3.87 B6 MMP2, -15, IL-12R␤, IL-15, PLC-␥2, JAK3 bmgr2 2 D2Mit237 2.48 B6 IL-15R␣, Vav2, PKC␪, GATA-3 bmgr3 2 D2Mit291 2.55 129 MMP-9, -24, NFAT1, PLC-␥1 bmgr4 4 D4Mit145 2.01 B6 IFN-␣ family, IFN-␤, JAK1 ebmgr1 3 D3Mit163 epistasis 129 with bmgr4 B6 GBP1, -2, -5 not able to mediate rejection of TAP-1 KO BM, while B6 mice can also originate from the strain that allows acceptance of BM with the same defect rejected such BM (Fig. 1A) (20). To elucidate and highlights the fact that each mouse strain contains a mixture of genetic differences that may be responsible for the differential abil- rejection- and acceptance-predisposing alleles. The final outcome, ity to reject class I-deficient BM, we set out to map genetic loci rejection, or acceptance of BM, will be a net effect of multiple associated with BM rejection. First, 129:B6 PKO mice were alleles. A suggestive QTL on chromosome 4, designated bmgr4, ϫ crossed to B6 PKO to generate (129:B6 B6)F1 PKO progeny. conferred rejection by the B6 allele (Fig. 2A; Table I.). We used a Transplantation of TAP-1 KO BM to perforin-deficient F1 progeny bootstrap resampling method to determine chromosomal locations demonstrated that rejection was as efficient as in B6 PKO recipi- of the identified QTLs. The 95% confidence intervals for QTL ents (Fig. 1A). This suggests that rejection may be governed by locations are shown in Fig. 2B. Downloaded from dominant alleles of the B6 genotype. 129:B6 PKO mice were backcrossed to the 129 background for Epistatic interactions affect BM graft rejection further studies (see Materials and Methods). The resulting 129 To test for possible gene-gene interactions regulating BM graft ϫ PKO were crossed to B6 PKO to generate (129 B6)F1 PKO rejection, we performed an analysis using a two QTL model that progeny. 129 PKO mice were unable to reject TAP-1 KO BM examines all pairs of genetic markers and intermarker positions for ϫ while (129 B6)F1 PKO progeny rejected, confirming previous association with ability to reject BM. The analysis enabled iden- http://www.jimmunol.org/ results with 129:B6 background mice (Fig. 1B). A larger variation tification of a fifth QTL located close to D3Mit163 at the distal end in the rejection by F1 PKO on the 129 background (Fig. 1B) than of chromosome 3. This QTL did not show any effect on BM re- F1 PKO on the 129:B6 background (Fig. 1A) is most likely due to jection when analyzed by itself (Figs. 2A and 3A). However, a 129 the large mouse number in the latter case. However, genetic factors homozygous genotype at D3Mit163 seemed to enhance the effect cannot be formally excluded. Importantly the F1 mice on both of the genotype at the suggestive QTL bmgr4 (compare Fig. 3, B backgrounds rejected TAP-1 KO BM. Based on the dominant and C), while a 129/B6 genotype at D3Mit163 eradicated or even characteristic of BM graft rejection in this setting, we crossed F1 reversed the effect of bmgr4 (Fig. 3D). This indicates epistatic mice to 129 PKO to generate BC-1 progeny, which were subjected to TAP-1 KO BM transplantation. As shown in Fig. 1B, there was by guest on September 29, 2021 a wide variability in rejection capacity among the BC-1 progeny. Notably, the distribution was unimodal rather than bimodal, sug- gesting involvement of several independent loci in BM rejection (Fig. 1C).

Microsatellite mapping reveals multiple loci that regulate BM graft rejection Of the 357 BC-1 progeny tested, we selected mice with extreme BM graft rejection phenotypes, 49 mice were defined as accepting the BM grafts and 89 mice were defined as rejecting (Fig. 1C). Forty-six mice from each extreme phenotype group were analyzed in an initial genome scan. Loci of interest (data not shown) were more carefully investigated by analyzing additional microsatellite markers and including all mice displaying the extreme phenotypes. Linkage analysis and permutation tests were performed imple- menting multiple imputation methods and analyzing BM rejection as a binary trait (rejecting vs accepting mice). Threshold levels for significant linkage were determined by a permutation test proce- dure, which is empirical and reflects the characteristics of the par- ticular experiment to which it is applied. This whole-genome scan revealed multiple QTLs that displayed linkage to BM graft rejec- FIGURE 3. Epistatic interactions between QTLs on chromosome 3 and tion in PKO mice (Fig. 2A). The QTL showing the strongest link- 4 affect BM graft rejection. The figure shows allele effects and interaction age to rejection was identified on chromosome 8. We designated it plots for QTLs detected on chromosomes 3 and 4. The relative I UdR BM graft rejection (bmgr) 1 (Fig. 2A). The B6 allele at the peak uptake (on a scale from 0 to 10 where 0 is rejection and 10 is acceptance) is shown on the y-axis as mean Ϯ SEM. A, Influence of genotype at the marker marker (D8Mit249) conferred rejection (Table I.). Furthermore, (D3Mit163) closest to the epistatic QTL (ebmgr1) on chromosome 3 alone. B, two significant QTLs, bmgr2 and bmgr3, mapped on chromosome Influence of genotype at the marker (D4Mit145) closest to the epistatic QTL 2 in vicinity of D2Mit237 and D2Mit291, respectively (Fig. 2A). (bmgr4) on chromosome 4 alone. C, Influence of genotype at D4Mit145 in Interestingly, whereas the B6 allele at bmgr2 conferred rejection, animals homozygous for the 129 genotype at D3Mit163. D, Influence of ability to reject BM grafts was linked to a 129 homozygous ge- genotype at D4Mit145 in animals with a heterozygous genotype (129/B6) at notype at bmgr3 (Table I.). This result shows that rejection alleles D3Mit163. The Journal of Immunology 7927 interactions between the two QTLs and the QTL at chromosome 3 NKG2D cannot be excluded, we find it unlikely that the strain was thus designated epistatic bmgr1 (ebmgr1; Table I). Analysis difference in rejection capacity would be due to differential recog- using a fit-multiple QTLs model confirmed that the phenotype is nition of NKG2D ligands. First, NKG2D-induced BM rejection regulated by these four independent QTLs and the fifth ebmgr1 in was MHC class I independent (38), second, an in vivo assay mea- concert with bmgr4 (data not shown). Thus, we demonstrate that suring rejection of large numbers of fluorescence-labeled ␤ Ϫ/Ϫ multiple QTLs, including complex gene-gene interactions, regu- B6 2m spleen cells revealed a similar difference where wt 129 late BM graft rejection in PKO mice. mice displayed impaired rejection compared with wt B6 mice (M. H. Johansson, unpublished data). In the latter case, the B6 Discussion genetic background of the donor cells (lacking NKG2D ligand The recognition strategies and effector mechanisms involved in expression) argue against the importance of differences dependent NK cell-mediated BM rejection have not been thoroughly eluci- on NKG2D. dated. In this study, we have initiated a quest for genes affecting Interestingly, mice with a nonfunctional NK cell receptor (ncr1; BM resistance in the absence of perforin. We used the fact that mouse homolog of human NKp46) gene displayed a reduced re- perforin-deficient 129 mice housed in a conventional animal facil- jection of the TAP-2-deficient tumor cells RMA-S compared with ity are unable to reject TAP-1 KO BM grafts, while mice of the B6 wild-type mice, but only in mice of the 129/Sv background. In genetic background under the same conditions readily reject such contrast, rejection in B6 mice was independent of NCR1 function grafts in an NK cell-dependent manner. (5). This suggests that B6 NK cells may express additional acti- A number of differences between 129 and B6 mouse strains vating receptor(s) recognizing RMA-S cells, receptors that may be have been reported with regards to NK cell function. 129 mice are lacking on 129/Sv NK cells. Such putative receptors are still un- Downloaded from d not able to reject BM from BALB/c mice (H-2 ), while B6 mice identified, but could be important also for the recognition of TAP-1 readily reject such cells. This variance has been attributed to dif- knockout BM cells. ferences in the NK gene complex (NKC) on chromosome 6 (29). Our previous data suggested that death receptors may be in- Indeed, the NKC is highly polymorphic between the two strains. volved in BM rejection (21, 23). It has also been shown that For example, NK1.1 (NKRP1-C), a commonly used marker ex- TRAIL is important for NK cell-mediated protection from metas- pressed exclusively on NK cells and a small subset of T cells, and tasis of certain tumors and mediates resistance to certain virus http://www.jimmunol.org/ encoded in the NKC, is present in B6, but not in 129 mice (30). infections. Furthermore, IFN-␣/IFN-␤ and IFN-␥ can induce up- The Ly49 gene family, encoding MHC class I-reactive activating regulation of TRAIL on NK cells (16, 39, 40). Thus, it is not and inhibitory receptors expressed on subpopulations of NK cells, unlikely that TRAIL may be involved in rejection of BM grafts is also located within the NKC. In three recent reports, the Ly49 and may play a role in the rejection investigated in the present gene family of the 129 strain was characterized and shown to be study. Regulatory T cells have been demonstrated to influence NK highly different from B6 (31–33). Importantly, we found no link- cell mediated rejection of allogeneic BM grafts and grafts lacking age in our study to the NKC. Moreover, 129:B6 PKO mice car- ϩ one MHC haplotype (hybrid resistance). Depletion of CD25 T rying the B6 NKC were as deficient as pure 129 PKO mice in

cells greatly enhanced rejection of these grafts (41). It may be by guest on September 29, 2021 TAP-1 KO BM graft rejection, further refuting influence of the possible that strain-dependent differences in this regulatory activity NKC in this setting (data not shown). could be involved in the results described here. A reported defect in 129/J NK cell function is aberrant DAP12 As a first step toward identifying genetic factors in perforin- signaling. DAP12 is an adapter molecule used by activating Ly49 and some other receptors. McVicar et al. (34) showed that signal- deficient mice, we performed a genome-wide linkage analysis for ing downstream of DAP12 is defective in 129/J mice, a defect of rejection of MHC class I-deficient BM grafts on progeny from a ϫ unknown molecular nature. However, lysis of the prototypic NK backcross of (B6 129) PKO F1 mice to 129 PKO. This analysis target line YAC-1 was normal, showing that all activating path- revealed that the phenotype in 129 PKO mice is regulated by mul- ways were not destroyed. The signaling defect may indeed explain tiple QTLs. Although some of the QTLs are independent, others the poor resistance to H-2d grafts described above; this rejection is display complex gene-gene interactions. Three independent QTLs mediated via activating Ly49 receptors (Ly49D in B6 and Ly49R were significantly linked to BM graft rejection with the strongest in 129) and dependent on DAP12 (34–37). However, these dif- linkage to a locus on chromosome 8 (bmgr1). There were two loci ferences are unlikely explanations in our system. Activating Ly49 on chromosome 2. Interestingly, at bmgr2 BM graft rejection was receptors are not involved in the rejection of MHC class I-deficient linked to the B6 allele, while at bmgr3 a 129 homozygous geno- BM grafts and DAP12-loss-of function mice are fully able to reject type conferred rejection. In addition, bmgr4 on chromosome 4 was ␤ Ϫ/Ϫ d suggestively linked to BM graft rejection. 2m spleen cells (37). Furthermore, rejection of H-2 grafts was largely perforin dependent (21), while the present study in- Because a whole genome scan, like the one performed, does not vestigates perforin-independent functions. However, it cannot be allow determination of the exact locations of QTLs, the chromo- excluded that one of the putative receptors triggering reactivity somal regions where the bmgr QTLs may be located are large and against MHC class I-deficient BM grafts could be associated with consist of vast numbers of genes. Nevertheless, we want to bring DAP12 or use the same signaling pathway. Even so, none of the the attention to a number of candidate genes in these loci. The previously reported differences between B6 and 129 NK cell func- genes encoding matrix metalloproteinase (MMP)-2 and MMP-15, tion can sufficiently explain the differential rejection patterns ex- are located in bmgr1 on chromosome 8 and MMP-9 and MMP-24 plored in this study. are located in bmgr3. Expression of MMP-2, -9, and -15 have been A recent study has identified NKG2D ligands as recognition demonstrated in mouse NK cells (42). MMPs have been implicated ϫ targets in BM graft rejection. (BALB/c C57BL/6)F1 mice re- in regulation of NK cell function: first, they degrade extracellular jected BALB/c BM grafts in an NKG2D-dependent manner and matrix molecules and may promote extravasation of NK cells (42); expression of Rae-1 family members and H60 was detected on BM second, MMP-9 cleaves the adhesion molecule ICAM-1 from the cells from BALB/c but not C57BL/6 or F1 mice (38). It is not cell surface of tumor cells, protecting such cells from NK cell lysis known whether NKG2D-mediated recognition is involved in the (43); third, MMPs cleave FasL and TNF-␣ from the cell surface rejection of TAP-1 KO BM grafts. Although participation of and may be able to regulate lysis by NK cells (44, 45). 7928 GENETIC MAPPING OF PERFORIN-INDEPENDENT BM REJECTION LOCI

Genes encoding cytokines, cytokine receptors, or downstream and may be regulating death receptor-mediated killing by NK signaling molecules affecting NK cell function are encoded within cells, or other more general NK cell functions. QTLs defined in the bmgr loci. IL-12R␤1 and IL-27R␣ are encoded in bmgr1 and studies like the one presented here could provide information use- are both involved in NK cell responses involving IFN-␥ produc- ful in studies of other complex traits, thus revealing links between tion and proliferation (46, 47). However, both 129 and B6 mice these and NK cells or perforin-independent rejection mechanisms. seem to have intact responses to IL-12, arguing against polymor- Further studies will include generation and analysis of congenic phisms in the IL-12R␤1 gene between these strains (48). IL-15 mouse strains allowing dissection of loci and hopefully the iden- (bmgr1) and its receptor (IL-15R␣; bmgr2) are essential for mat- tification of genes involved in NK cell-mediated BM resistance. uration, homeostasis, and proliferation of NK cells and can induce TRAIL expression. Deficiencies in any of these genes result in Acknowledgments mice lacking mature NK cells (49, 50). JAK3 (bmgr1) associates We thank Silvio and Maria Pen˜a for excellent animal husbandry and Klas with the common ␥-chain of the IL-2, 4, 7, 9, 15, and 21 receptors. Ka¨rre for critical reading of the manuscript. Deficiency in JAK3 in mice results in a lack of T, B, and NK cells. The NK cell deficiency is probably due to lack of IL-15 signaling Disclosures (51, 52). Interestingly, one report on patients with JAK3 deficiency The authors have no financial conflict of interest. describes a phenotype with mature T cells and only mild immu- nodeficiency. However, the JAK3-deficient T cells did not up-reg- References ulate FasL in response to IL-2 (53). JAK1 (bmgr4) is involved in 1. Yu, Y. Y., V. Kumar, and M. Bennett. 1992. Murine natural killer cells and the signaling downstream of JAK3, as well as in the signaling in marrow graft rejection. Annu. Rev. Immunol. 10: 189–213. Downloaded from ␣ 2. Lanier, L. L. 2005. NK cell recognition. Annu. Rev. Immunol. 23: 225–274. response to several other cytokines. The IFN- gene family and 3. Zompi, S., and F. Colucci. 2005. Anatomy of a murder—signal transduction IFN-␤ are located within bmgr4 on chromosome 4. These cyto- pathways leading to activation of natural killer cells. Immunol. Lett. 97: 31–39. kines are known to stimulate NK cell cytotoxicity by stimulating 4. Colucci, F., E. Schweighoffer, E. Tomasello, M. Turner, J. R. Ortaldo, E. Vivier, V. L. Tybulewicz, and J. P. Di Santo. 2002. Natural cytotoxicity uncoupled from expression of up-regulation of perforin and TRAIL (40). Multiple the Syk and Zap-70 intracellular kinases. Nat. Immunol. 3: 288–294. genes in the IFN-␣ gene family have allelic differences between 5. Gazit, R., R. Gruda, M. Elboim, T. I. Arnon, G. Katz, H. Achdout, J. Hanna, U. Qimron, G. Landau, E. Greenbaum, et al. 2006. Lethal influenza infection in 129/Sv and B6 strains (54). However, NK cells in both mouse the absence of the natural killer cell receptor gene Ncr1. 7: 517–523. http://www.jimmunol.org/ strains respond to type I IFN during virus infections (48) and it is 6. Lowin, B., M. C. Peitsch, and J. Tschopp. 1995. Perforin and granzymes: crucial not clear whether allelic differences in the IFN-␣ family may in- effector molecules in cytolytic T lymphocyte and natural killer cell-mediated cytotoxicity. Curr. Top. Microbiol. Immunol. 198: 1–24. fluence activation of NK cells. 7. Lieberman, J. 2003. The ABCs of granule-mediated cytotoxicity: new weapons in A number of signaling molecules important for NK cell activa- the arsenal. Nat. Rev. Immunol. 3: 361–370. tion may be involved in the BM graft rejection. Phospholipase C 8. van den Broek, M. F., D. Kagi, R. M. Zinkernagel, and H. Hengartner. 1995. Perforin dependence of natural killer cell-mediated tumor control in vivo. Eur. (PLC)-␥ isoforms couple immune receptors to downstream cal- J. Immunol. 25: 3514–3516. cium signaling. It was recently shown that mouse NK cells pre- 9. Kagi, D., B. Ledermann, K. Burki, P. Seiler, B. Odermatt, K. J. Olsen, dominantly express PLC␥-2 (bmgr1) while PLC␥-1 (bmgr3)is E. R. Podack, R. M. Zinkernagel, and H. Hengartner. 1994. Cytotoxicity medi- ated by T cells and natural killer cells is greatly impaired in perforin-deficient by guest on September 29, 2021 expressed at minimal levels. Deficiency in PLC␥-2 resulted in im- mice. Nature 369: 31–37. paired cytotoxicity in response to ITAM- and DAP10-dependent 10. Smyth, M. J., K. Y. Thia, E. Cretney, J. M. Kelly, M. B. Snook, C. A. Forbes, and A. A. Scalzo. 1999. Perforin is a major contributor to NK cell control of tumor signaling (55, 56). In addition, the guanine nucleotide exchange metastasis. J. Immunol. 162: 6658–6662. factor vav2 (bmgr2) has been implicated in regulation of FcR␥ and 11. Mori, S., A. Jewett, K. Murakami-Mori, M. Cavalcanti, and B. Bonavida. 1997. DAP12-dependent signaling in NK cells (57). Several intracellular The participation of the Fas-mediated cytotoxic pathway by natural killer cells is tumor-cell-dependent. Cancer Immunol. Immunother. 44: 282–290. signaling proteins are important in regulation of death receptor 12. Arase, H., N. Arase, and T. Saito. 1995. Fas-mediated cytotoxicity by freshly ligand expression. Protein kinase C (PKC)-␪ (bmgr2) has been isolated natural killer cells. J. Exp. Med. 181: 1235–1238. implicated in regulation of TCR/CD3-induced FasL expression in 13. Zamai, L., M. Ahmad, I. M. Bennett, L. Azzoni, E. S. Alnemri, and B. Perussia. 1998. Natural killer (NK) cell-mediated cytotoxicity: differential use of TRAIL mouse CTL clones (58). NFAT (bmgr3) is probably important in and Fas ligand by immature and mature primary human NK cells. J. Exp. Med. the signaling pathway downstream of PKC␪ and has been sug- 188: 2375–2380. 14. Kayagaki, N., N. Yamaguchi, M. Nakayama, K. Takeda, H. Akiba, H. Tsutsui, gested to regulate CD16-induced FasL expression in human NK H. Okamura, K. Nakanishi, K. Okumura, and H. Yagita. 1999. Expression and cells (59). The transcription factor GATA-3 (bmgr 2) is essential function of TNF-related apoptosis-inducing ligand on murine activated NK cells. to T cell development, while GATA-3-deficient B and NK cells J. Immunol. 163: 1906–1913. 15. Screpanti, V., R. P. Wallin, H. G. Ljunggren, and A. Grandien. 2001. A central develop. GATA-3-deficient NK cells are fully cytotoxic but dis- role for death receptor-mediated apoptosis in the rejection of tumors by NK cells. play an immature phenotype and are unable to produce IFN-␥ (60). J. Immunol. 167: 2068–2073. We identified a locus in the distal end of chromosome 3 (eb- 16. Smyth, M. J., E. Cretney, K. Takeda, R. H. Wiltrout, L. M. Sedger, N. Kayagaki, H. Yagita, and K. Okumura. 2001. Tumor necrosis factor-related apoptosis-in- mgr1) showing epistatic interaction with bmgr4 on chromosome 4. ducing ligand (TRAIL) contributes to interferon ␥-dependent natural killer cell Interesting candidate genes in this locus may be the IFN-inducible protection from tumor metastasis. J. Exp. Med. 193: 661–670. 17. Hayakawa, Y., V. Screpanti, H. Yagita, A. Grandien, H. G. Ljunggren, M. J. guanylate-binding proteins (GBP)-1, -2, and -5. These proteins are Smyth, and B. J. Chambers. 2004. NK cell TRAIL eliminates immature dendritic induced by, e.g., LPS, Listeria monocytogenes infection, or hep- cells in vivo and limits dendritic cell vaccination efficacy. J. Immunol. 172: atitis C core protein via IFN-␥ and IFN-␣/IFN-␤. GBPs can inhibit 123–129. 18. Lee, R. K., J. Spielman, D. Y. Zhao, K. J. Olsen, and E. R. Podack. 1996. viral replication and have antiproliferative activity, but have also Perforin, Fas ligand, and tumor necrosis factor are the major cytotoxic molecules been reported to promote fibroblast growth (61–63). used by lymphokine-activated killer cells. J. Immunol. 157: 1919–1925. Most of the described candidate genes have important functions 19. Baker, M. B., E. R. Podack, and R. B. Levy. 1995. Perforin- and Fas-mediated cytotoxic pathways are not required for allogeneic resistance to bone marrow in NK cells and other cells of the mouse. Thus, a complete defi- grafts in mice. Biol. Blood Marrow. Transplant. 1: 69–73. ciency in these proteins is highly unlikely. Nevertheless, polymor- 20. Bennett, M., P. A. Taylor, M. Austin, M. B. Baker, L. B. Schook, M. Rutherford, V. Kumar, E. R. Podack, K. M. Mohler, R. B. Levy, and B. R. Blazar. 1998. phisms slightly changing their function or expression level may Cytokine and cytotoxic pathways of NK cell rejection of class I-deficient bone result in the phenotype observed in 129 PKO mice. In conclusion, marrow grafts: influence of mouse colony environment. Int. Immunol. 10: we have identified five loci responsible for BM graft rejection 785–790. 21. Taylor, M. A., B. Ward, J. D. Schatzle, and M. Bennett. 2002. Perforin- and (bmgr1–4 and ebmgr1) in the absence of perforin. There are sev- Fas-dependent mechanisms of natural killer cell-mediated rejection of incompat- eral possible candidate genes that may be involved in this rejection ible bone marrow cell grafts. Eur. J. Immunol. 32: 793–799. The Journal of Immunology 7929

22. Pinti, M., J. Pedrazzi, F. Benatti, V. Sorrentino, C. Nuzzo, V. Cavazzuti, 43. Fiore, E., C. Fusco, P. Romero, and I. Stamenkovic. 2002. Matrix metallopro- P. Biswas, D. N. Petrusca, C. Mussini, B. De Rienzo, and A. Cossarizza. 1999. teinase 9 (MMP-9/gelatinase B) proteolytically cleaves ICAM-1 and participates Differential down-regulation of CD95 or CD95L in chronically HIV-infected in tumor cell resistance to natural killer cell-mediated cytotoxicity. Oncogene 21: cells of monocytic or lymphocytic origin: cellular studies and molecular analysis 5213–5223. by quantitative competitive RT-PCR. FEBS Lett. 458: 209–214. 44. Mitsiades, N., W. H. Yu, V. Poulaki, M. Tsokos, and I. Stamenkovic. 2001. 23. Taylor, M. A., P. M. Chaudhary, J. Klem, V. Kumar, J. D. Schatzle, and Matrix metalloproteinase-7-mediated cleavage of Fas ligand protects tumor cells M. Bennett. 2001. Inhibition of the death receptor pathway by cFLIP confers from chemotherapeutic drug cytotoxicity. Cancer Res. 61: 577–581. partial engraftment of MHC class I-deficient stem cells and reduces tumor clear- 45. Schneider, P., N. Holler, J.-L. Bodmer, M. Hahne, K. Frei, A. Fontana, and ance in perforin-deficient mice. J. Immunol. 167: 4230–4237. J. Tschopp. 1998. Conversion of membrane-bound Fas(CD95) ligand to its sol- 24. Bennett, M. 1971. Graft-versus-host reactions in mice. I. Kinetic and immuno- uble form is associated with downregulation of its proapoptotic activity and loss genetic studies of alloantigen-sensitive units of lymphoid tissue. Transplantation of liver toxicity. J. Exp. Med. 187: 1205–1213. 11: 158–169. 46. Pflanz, S., J. C. Timans, J. Cheung, R. Rosales, H. Kanzler, J. Gilbert, L. Hibbert, 25. Davenport, C., V. Kumar, and M. Bennett. 1995. Rapid rejection of H2k and T. Churakova, M. Travis, and E. Vaisberg. 2002. IL-27, a heterodimeric cytokine ϩ H2k/b bone marrow cell grafts by CD8 T cells and NK cells in irradiated mice. composed of EBI3 and p28 protein, induces proliferation of naive CD4ϩ T cells. J. Immunol. 155: 3742–3749. Immunity 16: 779–790. 26. Broman, K. W., H. Wu, S. Sen, and G. A. Churchill. 2003. R/qtl: QTL mapping 47. Lucas, S., N. Ghilardi, J. Li, and F. J. de Sauvage. 2003. IL-27 regulates IL-12 in experimental crosses. Bioinformatics 19: 889–890. responsiveness of naive CD4ϩ T cells through Stat1-dependent and -independent 27. Churchill, G. A., and R. W. Doerge. 1994. Empirical threshold values for quan- mechanisms. Proc. Natl. Acad. Sci. USA 100: 15047–15052. titative trait mapping. Genetics 138: 963–971. 48. Nguyen, K. B., T. P. Salazar-Mather, M. Y. Dalod, J. B. Van Deusen, X.-q. Wei, 28. Visscher, P. M., R. Thompson, and C. S. Haley. 1996. Confidence intervals in F. Y. Liew, M. A. Caligiuri, J. E. Durbin, and C. A. Biron. 2002. Coordinated and QTL mapping by bootstrapping. Genetics 143: 1013–1020. distinct roles for IFN-␣␤, IL-12, and IL-15 regulation of NK cell responses to 29. Sentman, C. L., V. Kumar, G. Koo, and M. Bennett. 1989. Effector cell expres- viral infection. J. Immunol. 169: 4279–4287. sion of NK1.1, a murine natural killer cell-specific molecule, and ability of mice 49. Kennedy, M. K., M. Glaccum, S. N. Brown, E. A. Butz, J. L. Viney, M. Embers, to reject bone marrow allografts. J. Immunol. 142: 1847–1853. N. Matsuki, K. Charrier, L. Sedger, C. R. Willis, et al. 2000. Reversible defects 30. Brown, M. G., A. A. Scalzo, K. Matsumoto, and W. M. Yokoyama. 1997. The in natural killer and memory CD8 T cell lineages in interleukin 15-deficient mice. Downloaded from natural killer gene complex: a genetic basis for understanding natural killer cell J. Exp. Med. 191: 771–780. function and innate immunity. Immunol. Rev. 155: 53–65. 50. Lodolce, J. P., D. L. Boone, S. Chai, R. E. Swain, T. Dassopoulos, S. Trettin, and 31. Makrigiannis, A. P., A. T. Pau, A. Saleh, R. Winkler-Pickett, J. R. Ortaldo, and A. Ma. 1998. IL-15 receptor maintains lymphoid homeostasis by supporting lym- S. K. Anderson. 2001. Class I MHC-binding characteristics of the 129/J Ly49 phocyte homing and proliferation. Immunity 9: 669–676. repertoire. J. Immunol. 166: 5034–5043. 51. Thomis, D. C., and L. J. Berg. 1997. The role of Jak3 in lymphoid development, 32. Makrigiannis, A. P., A. T. Pau, P. L. Schwartzberg, D. W. McVicar, T. W. Beck, activation, and signaling. Curr. Opin. Immunol. 9: 541–547. and S. K. Anderson. 2002. A BAC contig map of the Ly49 gene cluster in 129 52. Notarangelo, L. D., S. Giliani, C. Mazza, P. Mella, G. Savoldi, C. Rodriguez-Perez, mice reveals extensive differences in gene content relative to C57BL/6 mice. E. Mazzolari, M. Fiorini, M. Duse, A. Plebani, et al. 2000. Of genes and phenotypes:

Genomics 79: 437–444. http://www.jimmunol.org/ the immunological and molecular spectrum of combined immune deficiency: defects 33. Makrigiannis, A. P., D. Patel, M.-L. Goulet, K. Dewar, and S. K. Anderson. 2005. of the gc-JAK3 signaling pathway as a model. Immunol. Rev. 178: 39–48. Direct sequence comparison of two divergent class I MHC natural killer cell 53. Frucht, D. M., M. Gadina, G. J. Jagadeesh, I. Aksentijevich, K. Takada, receptor haplotypes. Genes Immun. 6: 71–83. J. J. Bleesing, J. Nelson, L. M. Muul, G. Perham, G. Morgan, et al. 2001. Un- 34. McVicar, D. W., R. Winkler-Pickett, L. S. Taylor, A. Makrigiannis, M. Bennett, expected and variable phenotypes in a family with JAK3 deficiency. Genes Im- S. K. Anderson, and J. R. Ortaldo. 2002. Aberrant DAP12 signaling in the 129 mun. 2: 422–432. strain of mice: implications for the analysis of gene-targeted mice. J. Immunol. 54. van Pesch, V., H. Lanaya, J. C. Renauld, and T. Michiels. 2004. Characterization 169: 1721–1728. of the murine ␣ interferon gene family. J. Virol. 78: 8219–8228. 35. Raziuddin, A., D. L. Longo, L. Mason, J. R. Ortaldo, M. Bennett, and W. J. 55. Tassi, I., R. Presti, S. Kim, W. M. Yokoyama, S. Gilfillan, and M. Colonna. 2005. Murphy. 1998. Differential effects of the rejection of bone marrow allografts by ␥ the depletion of activating versus inhibiting Ly-49 natural killer cell subsets. Phospholipase C- 2 is a critical signaling mediator for murine NK cell activating J. Immunol. 160: 87–94. receptors. J. Immunol. 175: 749–754. 56. Caraux, A., N. Kim, S. E. Bell, S. Zompi, T. Ranson, S. Lesjean-Pottier, 36. George, T. C., L. H. Mason, J. R. Ortaldo, V. Kumar, and M. Bennett. 1999. by guest on September 29, 2021 ␥ Positive recognition of MHC class I molecules by the Ly49D receptor of murine M. E. Garcia-Ojeda, M. Turner, and F. Colucci. 2005. Phospholipase C- 2is NK cells. J. Immunol. 162: 2035–2043. essential for NK cell cytotoxicity and innate immunity to malignant and virally 37. Oberg, L., S. Johansson, J. Michaelsson, E. Tomasello, E. Vivier, K. Karre, and infected cells. Blood 107: 994–1002. P. Hoglund. 2004. Loss or mismatch of MHC class I is sufficient to trigger NK 57. Cella, M., K. Fujikawa, I. Tassi, S. Kim, K. Latinis, S. Nishi, W. Yokoyama, cell-mediated rejection of resting lymphocytes in vivo—role of KARAP/DAP12- M. Colonna, and W. Swat. 2004. Differential requirements for Vav proteins in dependent and -independent pathways. Eur. J. Immunol. 34: 1646–1653. DAP10- and ITAM-mediated NK cell cytotoxicity. J. Exp. Med. 200: 817–823. 38. Ogasawara, K., J. Benjamin, R. Takaki, J. H. Phillips, and L. L. Lanier. 2005. 58. Pardo, J., M. Buferne, M. J. Martinez-Lorenzo, J. Naval, A. M. Schmitt-Verhulst, Function of NKG2D in natural killer cell-mediated rejection of mouse bone mar- C. Boyer, and A. Anel. 2003. Differential implication of protein kinase C iso- row grafts. Nat. Immunol. 6: 938–945. forms in cytotoxic T lymphocyte degranulation and TCR-induced Fas ligand 39. Takeda, K., E. Cretney, Y. Hayakawa, T. Ota, H. Akiba, K. Ogasawara, expression. Int. Immunol. 15: 1441–1450. H. Yagita, K. Kinoshita, K. Okumura, and M. J. Smyth. 2005. TRAIL identifies 59. Furuke, K., M. Shiraishi, H. S. Mostowski, and E. T. Bloom. 1999. Fas ligand immature natural killer cells in newborn mice and adult mouse liver. Blood 105: induction in human NK cells is regulated by redox through a calcineurin-nuclear 2082–2089. factors of activated T cell-dependent pathway. J. Immunol. 162: 1988–1993. 40. Sato, K., S. Hida, H. Takayanagi, T. Yokochi, N. Kayagaki, K. Takeda, 60. Samson, S. I., O. Richard, M. Tavian, T. Ranson, C. A. Vosshenrich, F. Colucci, H. Yagita, K. Okumura, N. Tanaka, T. Taniguchi, and K. Ogasawara. 2001. J. Buer, F. Grosveld, I. Godin, and J. P. Di Santo. 2003. GATA-3 promotes Antiviral response by natural killer cells through TRAIL gene induction by IFN- maturation, IFN-␥ production, and liver-specific homing of NK cells. Immunity ␣/␤. Eur. J. Immunol. 31: 3138–3146. 19: 701–711. 41. Barao, I., A. M. Hanash, W. Hallett, L. A. Welniak, K. Sun, D. Redelman, 61. MacMicking, J. D. 2004. IFN-inducible GTPases and immunity to intracellular B. R. Blazar, R. B. Levy, and W. J. Murphy. 2006. Suppression of natural killer pathogens. Trends Immunol. 25: 601–609. cell-mediated bone marrow cell rejection by CD4ϩCD25ϩ regulatory T cells. 62. Miller, K., S. McArdle, M. J. Gale, Jr., D. A. Geller, B. Tenoever, J. Hiscott, Proc. Natl. Acad. Sci. USA 103: 5460–5465. D. R. Gretch, and S. J. Polyak. 2004. Effects of the hepatitis C virus core protein 42. Kim, M. H., R. P. Kitson, P. Albertsson, U. Nannmark, P. H. Basse, P. J. Kuppen, on innate cellular defense pathways. J. Interferon Cytokine Res. 24: 391–402. M. E. Hokland, and R. H. Goldfarb. 2000. Secreted and membrane-associated 63. Carter, C. C., V. Y. Gorbacheva, and D. J. Vestal. 2005. Inhibition of VSV and matrix metalloproteinases of IL-2-activated NK cells and their inhibitors. J. Im- EMCV replication by the interferon-induced GTPase, mGBP-2: differential re- munol. 164: 5883–5889. quirement for wild-type GTP binding domain. Arch. Virol. 150: 1213–1220.