Search

To obtain the quantitative and qualitative data available for estimating the intake of carotenoids from fruits in Japan, carotenoids were analyzed with reversed phase high-performance liquid chromatography (HPLC). Ten carotenoids were examined in 75 raw fruits and 15 processed fruits, all of which were harvested or purchased in Japan. Phytoene was detected in 58 of 90 fruit samples ; ζ-carotene, in 50 of 90 ; lycopene, in 13 of 90 ; α-carotene, in 18 of 90 ; lutein, in 56 of 90 ; β-carotene, in 80 of 90 ; β-cryptoxanthin, in 68 of 90 ; zeaxanthin, in 58 of 90 ; all-trans-violaxanthin, in 55 of 90 ; and 9-cis-violaxanthin, in 47 of 90 samples. Citrus fruits of the mandarin type (Satsuma mandarin and its hybrids, such as tangor) were rich in β-cryptoxanthin, β-carotene, all-trans-violaxanthin, and 9-cis-violaxanthin ; ‘Star ruby’ grapefruit in lycopene ; loquat, Japanese persimmon, and peach in β-cryptoxanthin, β-carotene, all-trans-violaxanthin, and 9-cis-violaxanthin ; mango in β-carotene, all-trans-violaxanthin, and 9-cis-violaxanthin ; acerolas in phytoene ; passion fruits in ζ-carotene. Carotenoid levels in common fruits, such as apple, grape, lemon, pear, strawberry, kiwifruit, cherry, pineapple, and banana, were low.

PD-1 is an immunoinhibitory receptor, which belongs structurally to the CD28 family. PD-1-deficient mice show breakdown of peripheral tolerance and manifest multiple autoimmune symptoms. We previously described expression of PD-1 on activated T and B lymphocytes and myeloid cells. However, little is known about the microanatomical distribution of PD-1 in lymphoid organs. In this study, we performed immunohistochemistry using monoclonal antibodies against human PD-1. In human tonsils, PD-1 was expressed on most of T cells and a small subset of centrocytes in the light zone of germinal centers (GCs), where clonal selection of centrocytes takes place. These results suggest that PD-1 may play an important role in GC reaction.

Dihydrochalcone derivative, 3',5'-di-C-beta-glucopyranosylphloretin (1), is present in the genus Fortunella, (F. crassifolia, F. japonica, F. margarita, F. polyandra and F. hindsii). These species accumulate a large quantity of 1 in their fruits (peel, 6.5-15.2 mg/g in dry wt; juice sac, 1.5-10.5 mg/g) and in their leaves (21.3-60.2 mg/g). Twenty-seven Tanaka's Citrus species examined lack 1, but C. madurensis and C. halimii contain 1 in large quantities in their peels (25.1 and 33.6 mg/g) and juice sacs (4.1 and 4.2 mg/g). Poncirus species do not contain 1. Fortunella-citrus hybrids, the Orangequat [C. unshiuxF. crassifolia], the Thomasville citrangequat [Fortunella sp.x(C. sinensisxPoncirus trifoliata)], and seven hybrid progenies [F. margaritaxC. junos], contain large amounts of 1 in their peels (17.0-7.9 mg/g) and juice sacs (2.0-9.9 mg/g). These facts suggest that accumulation of 1 is a generic trait of the genus Fortunella and that the inheritance of the trait among the intergeneric hybrids is controlled by a dominant allele. Thus C. madurensis and C. halimii are thought to originate from natural hybrids between the genera Citrus and Fortunella. Phloridzin, which has the same aglycon as 1, was not detected in the citrus plants examined.

Inhibitory receptors expressed on NK cells recognize MHC class I molecules and transduce negative signals to prevent the lysis of healthy autologous cells. The lectin-like CD94/NKG2 heterodimer has been studied extensively as a human inhibitory receptor. In contrast, in mice, another lectin-like receptor, Ly-49, was the only known inhibitory receptor until the recent discovery of CD94/NKG2 homologues in mice. Here we describe the expression and function of mouse CD94 analyzed by a newly established mAb. CD94 was detected on essentially all NK and NK T cells as well as small fractions of T cells in all mouse strains tested. Two distinct populations were identified among NK and NK T cells, CD94bright and CD94dull cells, independent of Ly-49 expression. The anti-CD94 mAb completely abrogated the inhibition of target killing mediated by NK recognition of Qa-1/Qdm peptide on target cells. Importantly, CD94bright but not CD94dull cells were found to be functional in the Qa-1/Qdm-mediated inhibition. In the presence of the mAb, activated NK cells showed substantial cytotoxicity against autologous target cells as well as enhanced cytotoxicity against allogeneic and “missing self” target cells. These results suggest that mouse CD94 participates in the protection of self cells from NK cytotoxicity through the Qa-1 recognition, independent of inhibitory receptors for classical MHC class I such as Ly-49.

The antiemetic ondansetron hydrochloride, a 5-hydroxytryptamine receptor antagonist, was administered to alleviate nausea and vomiting after transcatheter arterial embolization (TAE) in 19 patients with hepatocellular carcinoma (HCC). A total of 38 patients (30 men and 8 women), aged 43 to 78 years, with Child-A—type hepatitis C virus-antibody—positive hepatic cirrhosis and concurrent HCC participated in this study. Patients were randomly divided into two equal groups. All 38 patients underwent TAE with doxorubicin, mitomycin C, lipiodol, and gelatin sponge. In addition, patients in group A received three 5-mg doses of ondansetron hydrochloride intravenously, administered immediately after TAE and on post-TAE days 1 and 2. Patients in group B were controls. The attending physician and nurses evaluated patients' degree of nausea and vomiting using a nausea scale from 1 (no nausea) to 5 (nausea and vomiting). The incidence of vomiting in group A was significantly lower on the day of TAE than that in group B, 1 patient (5%) versus 6 patients (32%), respectively. The mean nausea score for patients in group A (1.79) was also significantly lower than that for patients in group B (2.58). The effectiveness of ondansetron hydrochloride compared with placebo was superior immediately after TAE and on post-TAE day 1; no patients in either group complained of nausea on post-TAE day 2. Our findings suggest that ondansetron hydrochloride alleviates post-TAE nausea and vomiting in patients with HCC.

PD-1, a member of the Ig superfamily, was previously isolated from an apoptosis-induced T cell hybridoma 2B4.11 by subtractive hybridization. Expresson of the PD-1 mRNA is restricted to thymus in adult mice. Using an anti-PD-1 mAb (J43), we examined expression of the PD-1 protein during differentiation of thymocytes in normal adult, fetal and RAG-2(-/-) mice with or without anti-CD3 mAb stimulation. While PD-1 was expressed only on 3-5% of total normal thymocytes, approximately 34% of the CD4(-)CD8(-) double-negative (DN) fraction are PD-1(+) cells with two distinct expression levels (low and high). PD-1(high) thymocytes belonged to TCR gammadelta lineage cells. In the DN compartment of the TCR alphabeta lineage, PD-1 expression started at the low level from the CD44(+)CD25(+) stage and the majority of thymocytes expressed PD-1 at the CD44(-)CD25(-) stage in which the thymocytes express TCR beta chains. The anti-CD3epsilon antibody administration augmented the PD-1 expression as well as the differentiation of the CD44(-)CD25(+) DN cells into the CD44(-)CD25(-) DN stage, not only in normal mice but also in RAG-2-deficient mice. The fraction of the PD-1(low) cells in the CD4(+)CD8(+) double-positive (DP) compartment was very small (

A mAb J43 has been produced against the product of the mouse PD-1 gene, a member of the Ig gene superfamily, which was previously isolated from an apoptosis-induced T cell hybridoma (2B4.11) by using subtractive hybridization. Analyses by flow cytometry and immunoprecipitation using the J43 mAb revealed that the PD-1 gene product is a 50-55 kDa membrane protein expressed on the cell surface of several PD-1 cDNA transfectants and 2B4.11 cells. Since the molecular weight calculated from the amino acid sequence is 29, 310, the PD-1 protein appears to be heavily glycosylated. Normal murine lymphoid tissues such as thymus, spleen, lymph node and bone marrow contained very small numbers of PD-1(+) cells. However, a significant PD-1(+) population appeared in the thymocytes as well as T cells in spleen and lymph nodes by the in vivo anti-CD3 mAb treatment. Furthermore, the PD-1 antigen expression was strongly induced in distinct subsets of thymocytes and spleen T cells by in vitro stimulation with either anti-CD3 mAb or concanavalin A (Con A) which could lead T cells to both activation and cell death. Similarly, PD-1 expression was induced on spleen B cells by in vitro stimulation with anti-IgM antibody. By contrast, PD-1 was not significantly expressed on lymphocytes by treatment with growth factor deprivation, dexamethasone or lipopolysaccharide. These results suggest that the expression of the PD-1 antigen is tightly regulated and induced by signal transduction through the antigen receptor and do not exclude the possibility that the PD-1 antigen may play a role in clonal selection of lymphocytes although PD-1 expression is not required for the common pathway of apoptosis.

Fas ligand (FasL) is a type II integral membrane protein homologous with tumor necrosis factor (TNF). Recent studies indicate that TNF is processed to yield the soluble cytokine by metalloproteinases at the cell surface of activated macrophages and T cells. In the present study, we investigated whether FasL is also released by metalloproteinases. Treatment with hydroxamic acid inhibitors of matrix metalloproteinases specifically led to accumulation of membrane-type FasL (p40) on the surface of human FasL cDNA transfectants and activated human T cells, as estimated by surface immunofluorescence and immunoprecipitation with newly established anti-human FasL monoclonal antibodies. This surface accumulation of mFasL was associated with the decrease of soluble FasL (p27) in the supernatant as estimated by quantitative ELISA and immunoprecipitation with anti-human FasL monoclonal antibodies. These results indicate that human FasL is efficiently released from the cell surface by metalloproteinases like TNF.

To investigate the mechanisms of T-cell-mediated cytotoxicity, we estimated the involvement of apoptosis-inducing Fas molecule on the target cells and its ligand on the effector cells. When redirected by ConA or anti-CD3 monoclonal antibody, a CD4+ T-cell clone, BK1, could lyse the target cells expressing wild-type Fas molecule but not those expressing death signaling-deficient mutants. This indicates the involvement of Fas-mediated signal transduction in the target cell lysis by BK1. Anti-CD3-activated but not resting BK1 expressed Fas ligand as detected by binding of a soluble Fas-Ig fusion protein, and the BK1-mediated cytotoxicity was blocked by the addition of Fas-Ig, implicating the inducible Fas ligand in the BK1 cytotoxicity. Ability to exert the Fas-mediated cytotoxicity was not confined to BK1, but splenic CD4+ T cells and, to a lesser extent, CD8+ T cells could also exert the Fas-dependent target cell lysis. This indicates that the Fas-mediated target cell lytic pathway can be generally involved in the T-cell-mediated cytotoxicity. Interestingly, CD4+ T cells prepared from gld/gld mice did not mediate the Fas-mediated cytotoxicity, indicating defective expression of functional Fas ligand in gld mice.

We have recently demonstrated that killer cells expressing a cytolytic factor, perforin, infiltrate the hearts of mice with acute viral myocarditis and may play an important role in the mechanism of myocardial damage. To clarify the mechanism of in vivo cardiac myocyte injury mediated by perforin, we investigated the release of perforin molecules from killer cells by immunoelectron microscopy and examined the circular lesions formed by perforin on the membrane of cardiac myocytes. We found that there was massive release of perforin molecules from the killer cells directly onto the surface of the cardiac myocytes. Furthermore, electron microscopy of ultrathin ventricular sections treated with trypsin revealed numerous circular lesions with the characteristics of perforin pores, in the membranes of cardiac myocytes. These findings provide the first direct evidence that killer cells injure cardiac myocytes by releasing perforin and may play a critical role in the myocardial damage occurring in acute viral myocarditis.

We have defined a population of CD3-, CD56+ small lymphocytes (SLs) that exhibit the same phenotype and lytic capacity as natural killer (NK) cells. NK cells characteristically express the surface markers CD16 and CD56, mediate non–major histocompatibility complex (MHC)–restricted lysis, and have been equated with CD3- large granular lymphocytes (LGLs). In the present study we extended the observation that CD3-, CD56+ SLs can mediate NK- and antibody-dependent cellular cytotoxicity activity by studying the activation signals and lytic mechanisms that might be utilized by CD3-, CD56+ SLs in comparison to CD3- CD56+ LGLs. Our results show that CD3- SLs, similar to CD3- LGLs, exhibited activated killing in response to interleukin-2 (IL-2). In addition, after IL-2 activation, the CD3- SLs exhibited morphologic changes, including increases in size and granularity, and both morphologically and phenotypically became virtually indistinguishable from CD3- LGLs. Similar to CD3- LGLs, CD3- SLs could be directly activated by IL-2 alone to secrete significant quantities of interferon-γ (IFN-γ) and to express IL-2 receptor (IL-2R) p55. Examination of serine esterases and pore-forming protein (PFP) demonstrated that these cells exhibited a cytoplasmic distribution of perforin, which, unlike that of CD3- LGLs, was not associated with dense cytoplasmic azurophilic granules. Serine esterase levels were similar. However, after IL-2 activation PFP was concentrated in dense cytoplasmic granules, similar or identical to the situation in CD3-, CD56+ LGLs. These CD3-, CD56+ subsets appear to represent a continuum of activated cells that might represent various states of maturation of NK cells.

We have previously detected perforin expression in a subpopulation of asialo GM1+ natural killer (NK) cells and CD8+ T lymphocytes in murine spleen cells by immunocytochemical staining with an anti-perforin monoclonal antibody. In the present study, more detailed analyses of perforin expression in murine cytotoxic lymphocyte subpopulations were performed. The expression of perforin in asialo GM1+ spleen cells was predominantly confined to the NK1.1+ subset, where all NK activity also resided. Perforin expression was also studied on alloreactive cytotoxic T lymphocyte (CTL) induced in vivo. The cells expressing perforin in peritoneal exudate lymphocytes predominantly resided in the CD8+ T cell subpopulation co-expressing asialo GM1 where an allospecific CTL activity also resided. Furthermore, the percentage of perforin-positive cells in this population was greatly reduced after stimulation with anti-CD3 or anti-T cell receptors antibodies, which induce serine esterase release from the cytoplasmic granules. These findings highly suggest that perforin is involved in in vivo NK cell- and CTL-mediated cytolysis.

To verify the physiological role of the pore-forming protein perforin in vivo, its expression in subpopulations of human peripheral blood lymphocytes was examined by immunocytochemical staining and their cytolytic potentials compared. In addition to NK cells and gamma delta T cells, which uniformly expressed abundant perforin in their cytoplasmic granules, only a small subpopulation of CD8+ alpha beta T cells contained perforin, namely the CD11b+ subset. However, in vitro activation with an anti-CD3 antibody and IL-2 induced perforin expression in approximately 50% of the CD8+CD11b- T cells and also in a small subset of CD4+ T cells. A distribution of perforin in CD8+ and CD4+ T cells, similar to in vitro activated T cells, was observed in fresh peripheral blood lymphocytes from infectious mononucleosis patients. In all instances, the expression of perforin correlated with the cytolytic potential of these subpopulations. The results strongly suggest that perforin plays a role in the manifestation of cytotoxic activity in vivo.

The cytotoxic activity and pore-forming protein (PFP) expression of human peripheral blood (PB) gamma/delta T cells were examined. Fresh gamma/delta T cells isolated from PB lymphocytes by fluorescence-activated cell sorting exhibited a substantial natural killer-like cytotoxic activity against K562 target cells and had a high cytotoxic potential triggered by anti-CD3 monoclonal antibody (mAb) against P815 target cells bearing Fc gamma R. Immunocytochemical staining with an anti-PFP mAb revealed that virtually all PB gamma/delta T cells are granular lymphocytes with abundant PFP in their cytoplasmic granules. Constitutive expression of PFP in PB gamma/delta T cells was also demonstrated by Northern blot analysis. These observations support the proposed role of gamma/delta T cells in cytolytic immune surveillance in vivo.

The killer cell characteristics of Thy-1-positive dendritic epidermal cells (Thy-1+ DEC) were examined. Four Thy-1+ DEC clones which were established from athymic nude mice exhibited spontaneous or lectin-redirectable cytotoxic activity against some murine tumor cell lines in a 4 h 51Cr-release assay. A colorimetric assay for benzyloxycarbonyl-L-lysine-thiobenzyl ester esterase revealed a strong serine esterase activity expressed in all cell clones. In addition, Northern blot analysis using a murine perforin cDNA probe revealed that all four Thy-1+ DEC clones expressed abundant mRNA for perforin, as do most killer T cells. More importantly, immunocytochemical staining with an anti-perforin monoclonal antibody revealed that not only all four Thy-1+ DEC clones but also a part of freshly isolated Thy-1+ DEC from normal and nude mice contained perforin. These results demonstrate that Thy-1+ DEC exhibit typical killer cell characteristics in vitro and in vivo. These data also suggest that Thy-1+ DEC may play a cytotoxic role in protecting the integrity of skin from infection or neoplastic transformation.

Perforin is one of the important cytolytic factors in cytotoxic T lymphocytes (CTL) and natural killer (NK) cells. In this paper, we report rat mAbs against mouse perforin established by immunization with a recombinant mouse perforin fragment. These mAbs reacted with purified mouse perforin prepared from cytoplasmic granules of an NK-like cell line in ELISA and Western blot analysis. However, none of these mAbs blocked the hemolytic activity of mouse perforin or absorbed it when fixed in the solid phase. These results indicate that all of these mAbs react with denatured but not with native mouse perforin. By using a combination of the mAbs, we established a sandwich ELISA, for quantitating the cellular contents of perforin. These mAbs were also useful for immunohistochemical staining analysis, and perforin was detected in the cytoplasmic granules of CTL and NK cell lines. Perforin was also detected in a minor population of lymphocytes of the spleen, liver, and lymph node. In normal spleen cells of 5- to 8-week-old mice, 12-15% of asialo GM1+ cells and 7-21% of CD8+ T cells were perforin-positive, but CD4+ T cells, B cells, and macrophages were totally negative. These data clearly show that perforin is expressed in cells of a cytotoxic character in normal mice, in the same way as in primed mice.

Objective We focused on a small New World monkey, the common marmoset ( Callithrix jacchus ), to establish a nonhuman primate model of the treatment of hematological disorders. In this study, we developed the first monoclonal antibodies (MAbs) against marmoset CD34 and tested the in vitro and in vivo hemopoietic activity of cell populations isolated using one of these MAbs. Methods and Results Marmoset cDNA encoding a human CD34 homologue was cloned from bone marrow (BM)–derived RNA using reverse transcription polymerase chain reaction and rapid amplification of cDNA ends. The amino acid sequence of the marmoset CD34 had 81% homology with the human sequence. Five mouse MAbs were raised against marmoset CD34 transfectant. One representative MAb, MA24 (IgM), reacted with approximately 0.5 to 1% of BM mononuclear cells (MNCs), where the colony-forming unit granulocyte/macrophage (CFU-GM) was enriched approximately 11- to 75-fold as compared with the whole BM MNCs. Multilineage differentiation of marmoset CD34 + cells in NOD/SCID mice was confirmed by flow cytometry 1 month after xenotransplantation. Conclusion These results demonstrated that MA24 is useful for the analysis and enrichment of hematopoietic progenitor cells in the marmoset model for preclinical experiments.

Fas ligand (FasL) has been implicated in cytotoxic T lymphocyte (CTL)- and natural killer (NK) cell-mediated cytotoxicity. In the present study, we investigated the localization of FasL in murine CTL and NK cells. Immunocytochemical staining showed that FasL was stored in cytoplasmic granules of CD8+ CTL clones and in vivo activated CTL and NK cells, where perforin and granzyme A also resided. Immunoelectron microscopy revealed that FasL was localized on outer membrane of the cytoplasmic granules, while perforin was localized in internal vesicles. Western blot analysis showed that the membrane-type FasL of 40 kDa was stored in CD8+ CTL clones but not in CD4+ CTL clones. By utilizing a granule exocytosis inhibitor (TN16), we demonstrated that FasL translocated onto cell surface upon degranulation of anti-CD3-stimulated CD8+ CTL clones. Moreover, TN16 markedly inhibited the FasL-mediated cytotoxicity by CD8+ T cell clones and NK cells. These results suggested a substantial contribution of FasL to granule exocytosis-mediated target cell lysis by CD8+ CTL and NK cells.

Cytotoxic activity of NK cells is regulated in delicate balance between activating signals and inhibitory signals mediated by functionally different receptor families, so-called activating receptors and inhibitory receptors. With the aim of identifying novel NK receptors, we generated mAbs by immunizing activated NK cells and selected a novel mAb designated Yuri3. Biochemical and molecular analysis indicated that Yuri3 recognizes mouse CD94. Here we explored the expression of mouse CD94 and evaluated its function on primary NK cells by using the mAb. CD94 was expressed on essentially all NK and NKT cells as well as some subsets of T cells. Two distinct populations were observed in NK and NKT cells, CD94bright and CD94dull cells, and the expression of CD94 was independent of the expression of Ly-49 family members. Addition of the anti- CD94 mAb abrogated the inhibition of the target cell lysis mediated by the NK recognition of Qa-1. Interestingly, CD94bright but not CD94dull population was responsible for Qa-1 recognition. Substantial cytotoxicity against autologous target cells as well as enhanced cytotoxicity against allogeneic target cells was observed in the presence of the mAb, suggesting that CD94 plays an important role in self protection from NK cytotoxicity and its inhibitory function is independent of inhibitory receptors for classical MHC class 1 such as Ly-49 family members.

Cell-mediated autoimmunity has been strongly implicated in the pathogenesis of vascular cell injury in Takayasu's arteritis. To clarify the immunological mechanisms involved, we examined the expression of a cytolytic factor, perforin in infiltrating cells of aortic tissue samples from seven patients with Takayasu's arteritis. We also examined the expression of a 65-kD heat-shock protein (HSP-65), human leukocyte antigen classes I and II, and intercellular adhesion molecule-1 in the aortic tissue. Immunohistochemical studies showed that the infiltrating cells mainly consisted of gamma delta T lymphocytes, natural killer cells, macrophages, cytotoxic T lymphocytes and T helper cells, and that perforin was expressed in gamma delta T lymphocytes, natural killer cells, and cytotoxic T lymphocytes. In situ hybridization analysis also revealed expression of perforin mRNA in the infiltrating cells. Immunoelectron microscopic studies demonstrated that the infiltrating cells released massive amounts of perforin directly onto the surface of arterial vascular cells. We also found that expression of HSP-65, human leukocyte antigen classes I and II, and intercellular adhesion molecule-1 was strongly induced in the aortic tissue and might facilitate the recognition, adhesion and cytotoxicity of the infiltrating killer lymphocytes. These findings provide the first direct evidence that the infiltrating cells in the aortic tissue mainly consist of killer cells, and strongly suggest that these killer cells, especially gamma delta T lymphocytes, may recognize HSP-65 and play a critical role in the vascular cell injury of Takayasu's arteritis by releasing perforin.

Despite a number of reports indicating that perforin, a pore-forming protein, is the primary effector molecule mediating specific target cell lysis by cytotoxic T lymphocytes (CTL), several lines of evidence suggest the existence of perforin-independent mechanisms. We established class II-restricted, soluble protein-specific CD4+ T cell clones with killing function which do not express a detectable amount of perforin and perforin mRNA. Nevertheless, these clones induced cytolysis and DNA fragmentation of target cells in a specific and highly directional manner which was not inhibitable by antibody against TNF/lymphotoxin. These data not only indicate the existence of cytotoxic T cell subsets which do not utilize perforin, but also suggest that perforin is not mandatory for specific target lysis by T cells.

Blastocyst MHC is a recently identified mouse MHC class Ib gene, which is selectively expressed in blastocyst and placenta, and may be the mouse homolog of HLA-G gene the products of which have been implicated in protection of fetal trophoblasts from maternal NK cells and evasion of some tumor cells from NK cell attack. In this study, we identified two blastocyst MHC gene transcripts encoding a full-length α-chain (bc1) and an alternatively spliced form lacking the α2 domain (bc2), which may be homologous to HLA-G1 and HLA-G2, respectively. Both placenta and a teratocarcinoma cell line predominantly expressed the bc2 transcript. When these cDNAs were expressed in TAP-deficient RMA-S or TAP-sufficient RMA cells, only bc1 protein was expressed on the surface of RMA cells, but both bc1 and bc2 proteins were retained in the cytoplasm of RMA-S cells. Significantly, the RMA-S cells expressing either bc1 or bc2 were protected from lysis by NK cells in vitro. This protection was at least partly mediated by up-regulation of Qa-1b expression on the surface of RMA-S cells, which engaged the CD94/NKG2A inhibitory receptor on NK cells. More importantly, the bc1- or bc2-expressing RMA-S cells were significantly protected from NK cell-mediated rejection in vivo. These results suggested a role for blastocyst MHC in protecting TAP-deficient trophoblasts and tumor cells from NK cell attack in vivo.