The underlying mechanism of the protective and suppressive role of NKT cells in human tumor immunosurveillance remains to be fully elucidated. We show that the frequencies of CD8+ NKT cells in patients with EBV-associated Hodgkin's lymphoma or nasopharyngeal carcinoma are significantly lower than those in healthy EBV carriers. These CD8+ NKT cells in tumor patients are also functionally impaired. In human-thymus-severe combined immunodeficient (hu-thym-SCID) chimeras, EBV challenge efficiently promotes the generation of IFN-γ–biased CD8+ NKT cells. These cells are strongly cytotoxic, drive syngeneic T cells into a Th1 bias, and enhance T-cell cytotoxicity to EBV-associated tumor cells. Interleukin-4–biased CD4+ NKT cells are predominately generated in unchallenged chimeras. These cells are noncytotoxic, drive syngeneic T cells into a Th2 bias, and do not affect T-cell cytotoxicity. In humanized xenogeneic tumor-transplanted hu-thym-SCID chimeras, adoptive transfer with EBV-induced CD8+ NKT cells significantly suppresses tumorigenesis by EBV-associated malignancies. EBV-induced CD8+ NKT cells are necessary and sufficient to enhance the T-cell immunity to EBV-associated malignancies in the hu-thym-SCID chimeras. CD4+ NKT cells are synergetic with CD8+ NKT cells, leading to a more pronounced T-cell antitumor response in the chimeras cotransferred with CD4+ and CD8+ NKT cells. Thus, immune reconstitution with EBV-induced CD8+ NKT cells could be a useful strategy in management of EBV-associated malignancies. [Cancer Res 2009;69(20):7935–44]
- NKT cells
- EBV-associated malignancies
- SCID mice
NKT cells are unconventional glycolipid-reactive T cells that bridge innate and adaptive immunity ( 1– 3). Unlike T cells, NKT cells recognize CD1d-presented antigenic glycolipids through their semi-invariant Vα24-Jα18/Vβ11 (Vα14-Jα18/Vβ8,7,2 in mouse) TCR ( 1– 6). On encountering glycolipids, NKT cells promptly become activated to provide protection against infection and to regulate tumor immunity and autoimmunity ( 5, 6). There are two subsets of CD1d-dependent NKT cells based on whether the cells express semi-invariant Vα24-Jα18 (Vα14-Jα18 in mice) TCR (called type I and II NKT cells; refs. 1, 2, 7). IFN-γ–mediated antitumor activity of type I NKT cells on α-GalCer stimulation have been identified ( 8– 10). The type II NKT cells play an immunosuppressive role, inhibiting tumor immunosurveillance following release of interleukin (IL)-13 ( 11– 13). The type I NKT cells can also suppress tumor immunosurveillance based on their IL-13 production ( 14– 16). Whether these protective and suppressive subsets can cross-regulate each other and form a new immunoregulatory axis remains to be fully elucidated, particularly in humans.
EBV causes asymptomatic lifelong infection in ∼90% of adult populations. The proliferation of EBV-infected cells in healthy virus carriers is controlled by both humoral and cellular immune responses. Both NK-cell–mediated nonspecific and T-cell–mediated EBV-specific responses play important roles during primary infection ( 17). EBV-specific T cells are involved in restraining the proliferation of EBV-infected cells during latent infection ( 18, 19). EBV is associated with several malignancies including Hodgkin's lymphoma and nasopharyngeal carcinoma (NPC; ref. 19). EBV-specific T cells recognize Hodgkin's lymphoma and NPC malignant cells to control the oncogenic capacity of EBV ( 20). Adoptive transfer of in vitro activated and expanded EBV-specific T cells can prevent, and even cure, EBV-associated malignancies ( 21). However, a role of regulatory T cells in the suppression of EBV-specific cell-mediated antitumor immunity in patients with Hodgkin's lymphoma is well documented. Selective impairment of EBV-specific responses to the EBV-positive Hodgkin's lymphoma at the tumor site promotes tumor cell escape ( 22– 24). Although immunosuppression is well known to contribute to the pathogenesis of immunoblastic lymphoma, the failure of EBV-specific immunity in other EBV-associated malignancies is not well understood, because the affected patients are usually found to be relatively immunocompetent.
This study shows that CD8+ NKT cells are impaired in EBV-associated Hodgkin's lymphoma and NPC patients. In EBV-challenged human-thymus-severe combined immunodeficient (hu-thym-SCID) chimera, EBV promotes the generation of IFN-γ–biased CD8+ NKT cells that are highly cytotoxic to EBV-associated tumor cells in vitro. Adoptive transfer with EBV-induced CD8+ NKT cells suppresses tumorigenesis by EBV-associated malignancies in vivo humanized xenogeneic tumor-transplanted hu-thym-SCID (HXTT H-T-S) chimeras.
Materials and Methods
Patients, cells, epitope peptides, tetramers, and EBV stock. The Hodgkin's lymphoma patients were diagnosed according to WHO criteria and staged according to the Ann Arbor classification ( 25). The NPC and thyroid carcinoma (THC) patients were diagnosed and staged according to International Union against Cancer and American Joint Committee on Cancer staging manuals ( 26). Histopathology was based on the WHO international histologic classification. The latent EBV infection (LEI) or control normal (CN) subjects were healthy EBV-seropositive or EBV-seronegative individuals ( 17). For diagnosis of EBV association, EBV infection status of Hodgkin's lymphoma, NPC, and THC, patients were screened by serologic assay defined by positive in situ hybridization for EBV-encoded EBER plus immunohistochemistry for latent member protein 1 (LMP1) in tumor (Materials and Methods in Supplementary Data). On top of these assays, EBV infection possibility in THC patients was further ruled out by quantitative PCR (Q-PCR) in tumor (Supplementary Table S2). Clinical information on all patients and subjects was listed in Supplementary Table S1. Human fetal thymic cells and peripheral blood mononuclear cells were anonymously obtained from voluntarily elective pregnancy terminations [<24 weeks of gestation; typing-matched HLA-A2 and HLA-DRB1(*03), mismatched HLA-A11 and HLA-DQ5]. The possibility of EBV infection and human T-cell leukemia virus type I (HTLV-I) infection in the mothers was ruled out by serologic determination and Q-PCR ( 27). For transplantation and functional studies, NKT cells were depleted or purified from thymic cells, peripheral blood mononuclear cells, or organs by MACS beads after staining with α-GalCer–loaded CD1d tetramers (ProImmune; refs. 28– 30). Synthesized peptides used in the study were EBV epitopes GLCTLVAML in BMLF1 (GLC; HLA-A2–restricted), AVFDRKSDAK in EBNA3B (AVF; HLA-A11–restricted), TSLYNLRRGTAL in EBNA1 (TSL; HLA-DRB1–restricted), and SDDELPYIDPNM in EBNA3C (SDD; HLA-DQ5–restricted; ref. 18, 31). To prepare viral stocks, EBV producer cell line P3HR-1 (American Type Culture Collection) was treated with 12-O-tetradecanoylphorbol-13-acetate (30 ng/mL) for 14 days. The virus was then pelleted from the supernatant. In controls, 12-O-tetradecanoylphorbol-13-acetate (at concentration equal to the residual compound in viral suspension for final use) had no obvious NKT cell proliferation effect in vitro and in vivo hu-thym-SCID chimeras.
Hu-thym-SCID chimeras and HXTT H-T-S chimeras. To establish the hu-thym-SCID chimeras, human fetal thymic cells, consisting of NKT cell–depleted ( 28, 29) thymocytes and thymic epithelial cells and other stromal cells (∼1:1 ratio), were transplanted into thymi of irradiated (300 cGy/mouse) and anesthetized 8-week-old female SCID mice (NOD/LtSz-prkdcscid/prkdcscid strain; The Jackson Laboratory). Syngeneic human fetal liver tissue (containing ∼1 × 107 liver cells) was simultaneously implanted under the mouse kidney capsule. The chimeras were intrathymically challenged with EBV (107 plaque-forming units; ref. 17, 32) or HTLV-I (107 plaque-forming units) and rechallenged after 6 days. The hu-thym-SCID chimeras were maintained for 4 weeks unless otherwise stated ( 33). In some cases, the thymic CD4+ and/or CD8+ NKT cells (1 × 105 per mouse) from EBV-challenged hu-thym-SCID chimeras were mixed with syngeneic spleen T cells (2 × 106 per mouse) and antigen-presenting cells (APC; dendritic cells; 0.5 × 106 per mouse), adoptively transferred i.v. into primary SCID mice. At day 4 after cell transfer, a Hodgkin's lymphoma–derived EBV-associated cell line (L540CY; 1 × 106 cells per chimera, Cancer Centre, Sun Yet-sat University) or a EBV-encoded LMP1-transfected NPC cell line (CNE1-LMP1; 1 × 106 cells per chimera; Cancer Centre, Sun Yet-sat University) was implanted s.c. into the flanks of EBV-sensitized hu-thym-SCID chimeras (called HXTT H-T-S chimeras) and maintained as long as indicated. For detecting therapeutic effects of EBV-induced NKT cells on EBV-associated malignancies, tumor-implanted SCID mice (at day 21 post-implantation) were adoptively transferred i.v. with EBV-challenged CD4+ and/or CD8+ NKT cells (each 1 × 105 per mouse) plus syngeneic spleen T cells (2 × 106 per mouse) and dendritic cells (0.5 × 106 per mouse). The mice were housed in a pathogen-free environment in the Animal Research Institute, Wuhan University. The Wuhan University Ethical Committee approved all protocols in accordance with the current Chinese laws.
Flow cytometry. The α-GalCer–loaded CD1d tetramer was used to define total NKT cells. The appropriate isotype-matched control antibodies and fluorochrome-conjugated empty CD1d tetramer were used to gate out nonspecifically stained cells. For tetramer staining, the cells were incubated with the fluorochrome-labeled tetramer at 37°C for 15 min. For analysis of cellular proliferation, cells were labeled with carboxyfluorescein diacetate succimidyl ester. The intensity of carboxyfluorescein diacetate succimidyl ester fluorescence was analyzed ( 34). The average number of cell divisions was calculated by the equation: average number of divisions = log2(A/B), where A is the mean intensity of the nondividing group identified in the cells without antigen stimulation and B is the mean intensity of the test NKT cell population. Acquisitions were done with a flow cytometer (FACSCalibur; BD Biosciences) and analyzed with CellQuest software.
Real-time Q-PCR. All Q-PCRs were done as described elsewhere ( 35, 36). Briefly, total RNA from purified cells (1 × 104; purity >99%) or cell lines was prepared. The Q-PCR was done in a 96-well microtiter plate (Applied Biosystems) with an ABI PRISM 7700 Sequence Detector Systems (Applied Biosystems). Primers used in Q-PCR were listed in Supplementary Table S2.
Cytokine assays. For intracellular cytokine detection, Cytofix/Cytoperm Plug with Golgiplug kit (BD Pharmingen) was used. Cells were stimulated with various stimuli in the presence of Golgiplug and APC at 37°C for indicated duration. The intracellular stained cells were analyzed by flow cytometry. For ELISPOT cytokine detection, stimulated cells were stained with streptavidin-horseradish peroxidase (Mabtech) and diluted at 1:100 plus Nova Red substrate according to the manufacturer's instructions.
Cytotoxicity assay. Purified NKT cells were prepared as effector cells in a final volume of 1 mL complete RPMI 1640 in 24-well plates (0.8 × 104 per well). Target cells were prepared by pulsing with antigen (1 mL/well) for 24 h preincubation and added at various E:T ratios. Cytotoxicity was measured using the CytoTox 96 nonradioactive cytotoxicity assay kit (Promega).
Statistical analysis. Statistical analyses were done by paired or unpaired Student's t tests as appropriate. P values < 0.05 were considered statistically significant.
CD8+ NKT cells from patients with EBV-associated malignancies are impaired. To examine the role of NKT cell immunity against EBV-associated malignancies, we studied hospitalized Hodgkin's lymphoma, NPC, and THC patients ( 25, 26) and LEI and CN subjects (six of each; patient clinical information is shown in Supplementary Table S1 and Supplementary Fig. S1 for ALC). All Hodgkin's lymphoma, NPC, and THC patients in this study were newly diagnosed individuals. None of the Hodgkin's lymphoma, NPC, and THC patients had any previous treatments before entry into the study.
Frequencies of total and CD8+ NKT cells in peripheral blood mononuclear cells from EBV+ Hodgkin's lymphoma and NPC patients were significantly lower than that from LEI and CN subjects ( Fig. 1A ). About 15% to 30% of NKT cells from LEI and CN subjects were CD8+, whereas CD8+ NKT cells was <3% in Hodgkin's lymphoma and NPC patients ( Fig. 1A). By contrast, the frequency of CD4+ NKT cells was comparable among various patients and subjects as was the frequency of CD4−CD8− (DN) NKT cells, although this was low in all groups ( Fig. 1A). Frequencies of total and CD8+ NKT cells in EBV− Hodgkin's lymphoma and EBV− THC patients were comparable with that from LEI and CN subjects (data not shown).
To investigate cell functions, we measured cytokine expression by NKT cells in response to α-GalCer. CD4+ NKT cells in peripheral blood mononuclear cells of EBV+ Hodgkin's lymphoma and NPC patients and LEI and CN subjects produced similar low levels of IFN-γ and high levels of IL-4, IL-13, IL-10, and transforming growth factor-β1 as assessed by Q-PCR and ELISPOT ( Fig. 1B and C; some data not shown). CD8+ NKT cells from LEI and CN subjects expressed very high levels of IFN-γ and IL-2 but low levels of IL-4, IL-13, IL-10, and transforming growth factor-β1 ( Fig. 1B and C; some data not shown). CD8+ NKT cells from EBV+ Hodgkin's lymphoma and NPC patients produced much less IFN-γ than cells from the control subjects ( Fig. 1B and C). DN NKT cells of the patients and controls expressed comparable levels of IFN-γ and IL-4 as assessed by Q-PCR and ELISPOT (data not shown). In an analysis of cytotoxic activity, CD8+ NKT cells from LEI and CN subjects efficiently killed α-GalCer–loaded EBV-associated Hodgkin's lymphoma and NPC cell lines but not nontumor cells ( Fig. 2A ). By contrast, CD8+ NKT cells from EBV+ Hodgkin's lymphoma and NPC patients displayed only moderate cytotoxicity against these α-GalCer–loaded EBV-associated cell lines ( Fig. 2A). Particularly, CD8+ NKT cells from LEI subjects strongly targeted α-GalCer–loaded Hodgkin's lymphoma and NPC cell lines, resulting in a high level of cell killing ( Fig. 2A). CD4+ NKT cells from various patients and subjects showed no or very weak cytotoxicity both to α-GalCer–loaded tumor cell lines and to nontumor cells ( Fig. 2A). In proliferation studies, α-GalCer provoked a vigorous response by CD8+ NKT cells from CN and LEI subjects but not by CD8+ NKT cells from EBV+ Hodgkin's lymphoma and NPC patients ( Fig. 2B and C). CD4+ NKT cells from control subjects and patients all showed moderate proliferation in response to α-GalCer ( Fig. 2B and C). The functions of CD4+ NKT cells in EBV− Hodgkin's lymphoma and THC patients were comparable with that in LEI and CN subjects in terms of cytokine production and cell proliferation (data not shown). Interestingly, the functions of CD8+ NKT cells in EBV− Hodgkin's lymphoma and THC patients were not significantly changed in terms of IFN-γ production, cytotoxicity, and cell proliferation compared with that in LEI and CN subjects (data not shown). We monitored the frequencies and functions of different subsets of NKT cells in EBV− and EBV+ Hodgkin's lymphoma patients who were at different stages or at different time intervals of the treatments (1, 3, and 6 months). There was no clear evidence showing the correlation between NKT cell function and the disease stage or constitutional symptoms in Hodgkin's lymphoma patients during the follow-up (data not shown), which needed to be further investigated and clarified.
EBV promotes generation of IFN-γ–biased CD8+ NKT cells in hu-thym-SCID chimeras. To further explore the cellular basis of the pathogenesis of EVB-associated malignancies, we established hu-thym-SCID chimeras ( 33) and examined the NKT cell development and functions in EBV-challenged or HTLV-I–challenged hu-thym-SCID chimeras. In EBV-challenged chimeras, the frequency of total NKT cells in various organs was significantly higher than that in unchallenged or HTLV-I–challenged chimeras ( Fig. 3A ). In unchallenged and HTLV-I–challenged chimeras, >95% of NKT cells were CD4+, whereas CD8+ NKT cells were below detectable levels in the various organs ( Fig. 3B). By contrast, frequency of CD8+ NKT cells was greatly increased in all organs of EBV-challenged chimeras (30-38% of total NKT cells; Fig. 3C). Particularly, the CD8+ populations in the peripheral blood and lymph nodes reached ∼70% of total NKT cells at weeks 4 and 5 after EBV exposure. The frequencies of DN NKT cells were low and comparable in various organs in different hu-thym-SCID chimeras (data not shown).
We next examined cytokine production by NKT cells from different hu-thym-SCID chimeras in response to α-GalCer. Consistent with the human data, thymic CD4+ NKT cells from unchallenged and EBV-challenged hu-thym-SCID chimeras produced low levels of IFN-γ but high levels of IL-4, IL-13, IL-10, and transforming growth factor-β1. Thymic CD8+ NKT cells from EBV-challenged hu-thym-SCID chimeras expressed very high levels of IFN-γ and IL-2 and low levels of IL-4, IL-13, IL-10, and transforming growth factor-β1 as assessed by Q-PCR and ELISPOT ( Fig. 4A ; some data not shown). Control studies showed that there was no preactivation of EBV-specific NKT cells by the purification procedure (Supplementary Fig. S2). The cytokine production patterns by CD4+ and CD8+ NKT cells from spleens, livers, peripheral blood, and lymph nodes were similar to that in the thymus from the various hu-thym-SCID chimeras as assessed by Q-PCR and ELISPOT (data not shown).
We further examined the interaction between NKT cells and syngeneic T cells ex vivo the hu-thym-SCID chimeras. Thymic CD4+ or CD8+ NKT cells from EBV-challenged hu-thym-SCID chimeras were cocultured with syngeneic spleen T cells in the presence of syngeneic dendritic cells and examined cytokine expression by T cells 12 h later. In this system, EBV epitopes (GLC or TSL) were used as stimuli for EBV-specific CD4+ and CD8+ T cells, respectively. Thymic CD8+ NKT cells from EBV-challenged hu-thym-SCID chimeras could significantly drive spleen T cells to express Th1 cytokine (IFN-γ and IL-2) as assessed by Q-PCR and flow cytometry ( Fig. 4B; some data not shown). Thymic CD4+ NKT cells from unchallenged and EBV-challenged hu-thym-SCID chimeras could drive spleen T cells to express Th2 cytokines (IL-4 and IL-13) as well as IL-10 and transforming growth factor-β1 ( Fig. 4B; some data not shown). A similar activity of the CD4+ and CD8+ NKT cells from spleen, liver, peripheral blood, and lymph nodes from both types of chimeras was also confirmed in this coculture system (data not shown).
EBV-induced CD8+ NKT cells are cytotoxic to EBV-associated tumor cells in vitro. We next examined on the cytotoxicity of chimeric NKT cells. The thymic and spleen CD8+ NKT cells from EBV-challenged hu-thym-SCID chimeras efficiently killed α-GalCer–primed EBV-associated tumor cell lines but not nontumor cells ( Fig. 4C and D). CD4+ NKT cells from unchallenged and EBV-challenged hu-thym-SCID chimeras showed no cytotoxicity toward either EBV-associated tumor cells or nontumor cells ( Fig. 4C and D). The differential cytotoxicity of CD4+ versus CD8+ NKT cells from livers, peripheral blood, and lymph nodes of different chimeras were also confirmed (data not shown). We further examined the specific cytotoxicity of different T cells. We established a killing system involving coculture of NKT cells with syngeneic CD3+CD56−CD161− T cells ( Fig. 4E). The spleen T cells from EBV-challenged hu-thym-SCID chimeras alone moderately killed EBV-associated tumor cells (L540CY and CNE1-LMP1; data not shown). In coculture with thymic CD4+ NKT cells from the unchallenged or EBV-challenged hu-thym-SCID chimeras, spleen T cells from EBV-challenged hu-thym-SCID chimeras were moderately cytotoxic toward EBV-associated tumor cells but not nontumor cells ( Fig. 4E). In coculture with thymic CD8+ NKT cells from EBV-challenged hu-thym-SCID chimeras, the spleen T cells from EBV-challenged hu-thym-SCID chimeras very efficiently killed EBV-associated tumor cell lines but not nontumor cells ( Fig. 4E).
The α-GalCer provoked a vigorous proliferation of thymic CD8+ NKT cells from EBV-challenged hu-thym-SCID chimeras (Supplementary Fig. S3A and B). By contrast, the thymic CD4+ NKT cells from both unchallenged and EBV-challenged hu-thym-SCID chimeras only showed moderate proliferative capacity in response to α-GalCer (Supplementary Fig. S3A and B). The different proliferative capacities of CD4+ and CD8+ NKT cells from spleens, livers, peripheral blood, and lymph nodes in both chimeras were also confirmed (data not shown).
EBV-induced CD8+ NKT cells suppress tumorigenesis by EBV-associated malignancies in vivo. To investigate the protective role of EBV-induced CD8+ NKT cells against EBV-associated tumors in vivo, various HXTT H-T-S chimeras were established by transfer with different combinations of the EBV-sensitized or EBV-specific immune cells. The tumor growth curves ( Fig. 5A and C ) and the animal survival rates ( Fig. 5B and D) were monitored ( 37, 38). In HXTT H-T-S chimeras transferred with EBV-challenged thymic CD8+ NKT cells (plus spleen syngeneic T cells and dendritic cells, called +T+DC), the local growth of both tumors was efficiently inhibited. In HXTT H-T-S chimeras transferred with unchallenged or EBV-challenged thymic CD4+ NKT cells, the local size of both tumors was enlarged rapidly ( Fig. 5A). Unexpectedly, the transfer with combination of EBV-challenged thymic CD4+ and CD8+ NKT cells (+T+DC) could very efficiently inhibit local Hodgkin's lymphoma and NPC tumor growth in HXTT H-T-S chimeras. The tumor size was maintained at 3 to 10 mm3 for 18 weeks ( Fig. 5A). In HXTT H-T-S chimeras transferred with EBV-challenged thymic CD8+ NKT cells (+T+DC), the survival rates were significantly increased (∼30% of animals survived for 18 weeks in both tumor models). In contrast, in HXTT H-T-S chimeras transferred with unchallenged or EBV-challenged thymic CD4+ NKT cells (+T+DC), the survival time was much shorter (all animals died within 9 or 10 weeks; Fig. 5B; some data not shown). All Hodgkin's lymphoma and NPC HXTT H-T-S chimeras that were transferred with combinations of EBV-challenged thymic CD4+ and CD8+ NKT cells (+T+DC) survived even longer than 20 weeks, indicating a synergistic effect of the two subsets of NKT cells ( Fig. 5B). No anti-EBV-associated malignancy effect was observed in the hu-thym-SCID chimeras transferred with CD4+, CD8+ NKT cells, EBV-specific CD3+CD56−CD161− T cells, or dendritic cells alone ( Fig. 5A and B; data not shown). In an eradication model of HXTT H-T-S chimeras, the therapeutic effects of EBV-challenged NKT cells on EBV-associated malignancies were also observed. The transfer with EBV-challenged thymic CD8+ NKT cells (+T+DC) efficiently inhibited the growth of Hodgkin's lymphoma and NPC and increased animal survival, whereas CD4+ NKT cells (+T+DC) had no such function ( Fig. 5C and D). The synergistic effect of EBV-challenged thymic CD4+ and CD8+ NKT cells (+T+DC) in suppressing EBV-associated malignancies was also confirmed ( Fig. 5C and D). No eradicative effect on EBV-associated malignancies was observed in the hu-thym-SCID chimeras transferred with CD4+, CD8+ NKT cells, EBV-specific CD3+CD56−CD161− T cells, or dendritic cells alone ( Fig. 5C and D; data not shown). No effects on either tumor growth or survival were found in hepatocellular carcinoma cell-implanted (HepG2; American Type Culture Collection) HXTT H-T-S chimeras transferred with EBV-challenged thymic CD4+ and/or CD8+ NKT cells (+T+DC; data not shown).
Thus, EBV-induced CD8+ NKT cells were necessary and sufficient for enhancing the T-cell immunity against EBV-associated malignancies. CD4+ NKT cells were synergistic, but not required, for CD8+ NKT cells to enhance the T-cell immunity.
The distinctive type I and II NKT cells play positive and negative regulatory roles in tumor immunity, respectively ( 39). The type I NKT cells mediate strong antitumor responses on α-GalCer–induced IFN-γ production ( 9– 11). The type II NKT cells suppress tumor immunity through production of IL-13 ( 40). We have found that EBV induces the generation of IFN-γ–biased and cytotoxic CD8+ NKT cells that drive T cells into Th1-bias. IL-4–biased and noncytotoxic CD4+ NKT cells are predominately generated in unchallenged hu-thym-SCID chimeras, which drive T cells into Th2 bias. Adaptive transfer with EBV-induced CD8+ NKT cells significantly suppresses tumorigenesis by EBV-associated malignancies in vivo. Beyond type I versus II NKT cells, distinctive human CD4+ and (EBV-induced) CD8+ NKT cells are important effectors differentially modifying the pathogenesis process of EBV-associated malignancies.
CD4+ T cells are well-known required for the optimal development, survival, and functional responsiveness of memory CD8+ T cells in an effective immune response against infections and tumors ( 41– 43). Several factors including IL-2 ( 44), CD40-CD40L ( 45), and TRAIL ( 46) are involved in this complex interaction between CD4+ and CD8+ T cells. We show that EBV-induced CD8+ NKT cells are required for enhancing the EBV-specific T-cell immunity against EBV-associated malignancies. CD4+ NKT cells are synergistic, but not required, for the CD8+ NKT cells to enhance the EBV-specific T-cell antitumor immunity. A rather complex picture begins to take shape in which the optimal antitumor immune response of EBV-specific T cells against EBV-associated malignancies is CD8+ NKT cell-dependent but can be enhanced by CD4+ NKT cells. The detailed mechanism of the synergy between CD4+ and CD8+ NKT cells in antitumor immunity will require further investigation. In this study, all fetal samples are collected from the pregnancies in non-EBV-infected mothers. EBV challenge reflects the direct viral effects on promotion of CD8+ NKT cells in the chimeras. However, EBV is one of eight known human herpesviruses ( 17); the possibility that EBV-seronegative humans are infected with other human herpesviruses, such as herpes simplex viruses or cytomegalovirus, cannot be ruled out in this study. The question of whether infection with other human herpesviruses affects the generation of CD8+ NKT cells should be the subject of future investigations.
Notably, both in vitro and in vivo, CD8+ NKT cells exert effects against the Hodgkin's lymphoma–derived EBV-positive cell line (L540CY); those pattern of viral gene expression is markedly different from that seen in tumor diagnostic specimens in EBV+ Hodgkin's lymphoma patients. The CD8+ NKT cells also exert immune function against NPC-derived EBV-encoded LMP1-transfected cell line (CNE1-LMP1) that does not harbor entire EBV. Therefore, CD8+ NKT cell functions on EBV-associated tumors need to be further investigated, such as which viral gene expression is essential to trigger EBV-induced CD8+ NKT cells to enhance human tumor immunosurveillance.
Beyond pilot trials of immunotherapy with EBV-specific CTL cells to prevent and treat EBV-associated malignancies ( 47, 48), in accordance with the in vitro antitumor activity of NKT cells in mice and in humans, several clinical trials of α-GalCer have been performed in cancer patients ( 49). Adoptive transfer of autologous in vitro α-GalCer–activated and IL-2–activated NKT cells or dendritic cells has also been tested in a phase I clinical trial in cancer patients ( 50). In contrast to the promising results in mice, few clinical trials have succeeded in achieving significant efficacy against human tumors in vivo. We show that CD8+ NKT cells are functionally impaired in EBV-associated malignancies. EBV-induced CD8+ NKT cells are highly cytotoxic to EBV-associated tumor cells in vitro and also capable of enhancing conventional T-cell cytotoxicity and can suppress tumorigenesis by EBV-associated malignancies in vivo. Given the high prevalence of lifelong persistent EBV infection in the adult population, this principle should be considered when designing clinical trials involving NKT cell transfer-based immunotherapy for human EBV-associated malignancies. The reconstitution of immunity with EBV-induced and in vitro activated CD8+ NKT cells could be a useful strategy in the management of EBV-associated malignancies. Whether there are any species-related differences in NKT cell activity or differences in the experimental design of the clinical treatments between advanced cancer patients and adoptively tumor-transplanted mice remains an open question. Furthermore, both protective and suppressive roles of NKT cells in tumor immunosurveillance have been reported ( 8– 16). The nature of NKT cell ligand involved in EBV infection remains unknown. These unanswered questions could be potential hurdles for transferring CD8+ NKT cells into patients with EBV-associated malignancies.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Grant support: National Natural Science Foundation of China grants 30730054, 30572119, 30670937, 30971279, and 30901363; Hi-tech Research and Development Program of China from Ministry of Science and Technology grant 2007AA02Z120; Ministry of Education grant 20060486008; National Innovation Experiment Program for College Students grant WU2007061; Provincial Department of Science and Technology of Hubei grant 2007ABC010; and Chang Jiang Scholars Program from Ministry of Education and Li Ka Shing Foundation (Chang Jiang Scholar T. Jinquan).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).
H. Yuling, X. Ruijing, L. Li, J. Xiang, Z. Rui, W. Yujuan, Z. Lijun, and D. Chunxian contributed equally to this work.
- Received March 9, 2009.
- Revision received June 2, 2009.
- Accepted July 29, 2009.
- ©2009 American Association for Cancer Research.