Host-Residual Invariant NK T Cells Attenuate Graft-versus-Host Immunity1
http://www.100md.com
免疫学杂志 2005年第14期
Abstract
Invariant NK T (iNKT) cells have an invariant TCR- chain and are activated in a CD1d-restricted manner. They are thought to regulate immune responses and play important roles in autoimmunity, allergy, infection, and tumor immunity. They also appear to influence immunity after hemopoietic stem cell transplantation. In this study, we examined the role of iNKT cells in graft-vs-host disease (GVHD) and graft rejection in a mouse model of MHC-mismatched bone marrow transplantation, using materials including -galactosylceramide, NKT cells expanded in vitro, and J18 knockout mice that lack iNKT cells. We found that host-residual iNKT cells constitute effector cells which play a crucial role in reducing the severity of GVHD, and that this reduction is associated with a delayed increase in serum Th2 cytokine levels. Interestingly, we also found that host-residual iNKT cause a delay in engraftment and, under certain conditions, graft rejection. These results indicate that host-residual iNKT cells attenuate graft-vs-host immunity rather than host-vs-graft immunity.
Introduction
Natural killer T cells are a population of T cells that have NK cell markers such as NK1.1 (NKR-P1C) in mice or CD161 (NKR-P1A) in humans (1, 2). Some NK T cells use an invariant TCR- chain (V14-J18 in mice, V24-J18 in humans) paired with V8, V7, or V2 in mice or with V11 in humans (3, 4, 5, 6, 7), and are called invariant NKT (iNKT)5 cells. iNKT cells are activated by synthetic glycolipids such as -galactosylceramide (-GalCer) in a CD1d-restricted manner (1, 2, 8, 9). iNKT cells produce both Th1 (such as IFN- and TNF-) and Th2 (such as IL-4, IL-5, IL-10, and IL-13) cytokines (1, 2, 8, 9). It has been reported that iNKT cells control immune responses in some infections, tumors, autoimmune diseases, and allograft rejection (1, 2, 8, 9).
Graft-vs-host disease (GVHD) is one of the most serious complications in hemopoietic stem cell transplantation. It has been suggested in a mouse acute GVHD model that NK1.1+ T cells obtained from donor bone marrow can suppress GVHD induced by peripheral blood transplantation from the same donor (10). It has also been shown that a selected conditioning regimen, which preserves more host-residual NK1.1+ or DX5+ T cells than other T cells, is advantageous for reducing acute GVHD (11, 12). Furthermore, the suppressive effect of -GalCer on induced acute GVHD has been demonstrated in a mouse model(13, 14). We previously reported that the number of iNKT cells was lower in patients with GVHD than in those without GVHD (15) after hemopoietic stem cell transplantation, although the cause-effect relationship was not clear.
In this report, we provide direct evidence that host-residual iNKT cells reduce GVHD in a mouse model of MHC-mismatched bone marrow transplantation using J18 knockout mice that lack iNKT cells (16). Adoptively transferred iNKT cells with grafts also reduce GVHD, but, importantly, this effect is dependent on the presence of host-residual iNKT cells.
Results
Establishment of GVHD model mice
In the prototype transplantation from wild-type C57BL/6 to wild-type BALB/c, all of the recipients showed lethal GVHD as judged by the GVHD score. Histological examinations also confirmed GVHD in representative mice (data not shown). In this setting, we confirmed that full donor chimerism was achieved in all of the recipients (data not shown). We performed all of the transplantation experiments two to four times independently, and integrated all of the results from individual experiments to avoid experiment-to-experiment variation.
Administration of -GalCer prolonged survival of GVHD mice
First, we administered either -GalCer to activate iNKT cells or control vehicle every 4 days from the day of transplantation. -GalCer-treated mice survived significantly longer than control mice (p < 0.0001, Fig. 1A). The severity of GVHD in -GalCer-treated mice was milder than that in control mice (Fig. 1B). The GVHD score within the first 30 days after transplantation in the two groups was compared by repeated measures ANOVA and was found to be significantly lower in the -GalCer group (p < 0.0001).
We then examined whether -GalCer treatment influenced the number of iNKT cells in the liver, which contains the largest proportion of iNKT cells (1, 9) (14.2 ± 4.7% in our measurements; Fig. 1Di), after transplantation. In control recipients, the ratio of iNKT cells to lymphocytes decreased rapidly and iNKT cells became undetectable by approximately day 5 posttransplant (Fig. 1, C and D). The rate of decrease in the iNKT cells was delayed in -GalCer-treated mice, although iNKT cells were scarcely detectable after day 10 even in mice treated with -GalCer (Fig. 1C). We observed temporary donor-type iNKT cell chimerism before the disappearance, although the time course of the donor:recipient iNKT cell ratio was highly variable (data not shown).
Interestingly, we found that the engraftment of donor CD4+ T cells in -GalCer-treated mice was delayed compared with that in control mice during the early phase after transplantation, although complete donor chimerism was eventually achieved (Fig. 2A). Engraftment of Gr-1+ cells was also slightly delayed, but engraftment of other lineage cells was similar to that seen in vehicle-treated mice (data not shown). These findings encouraged us to compare various cytokine levels between -GalCer-treated and control mice, since iNKT cells produce high levels of both Th1 and Th2 cytokines and could influence Th1/Th2 polarization. Therefore, we examined the serum levels of IFN-, TNF-, IL–4, and IL-5 shortly after transplant and found that all of the cytokines were increased by -GalCer treatment (Fig. 2B, i-iv), reproducing previous reports that -GalCer has been shown to rapidly stimulate Th1 and Th2 cytokine production in vivo in the nontransplantation model (23, 24, 25). In contrast, it has been reported that in vivo -GalCer-primed splenocytes secrete less amounts of Th1 cytokines after further in vitro treatment with -GalCer compared with vehicle-primed splenocytes (23, 26). In our experimental system, we could exactly reproduce these findings (Fig. 2C, i-iv). Then we measured all four cytokines on day 5 or 6 posttransplant (1 or 2 days after the second -GalCer administration), both in -GalCer- and vehicle-treated mice. At this time point, GVHD signs were not been obvious yet, although the damage from radiation was inseparably measured "GVHD score" in Fig. 1B. IFN- and TNF- levels in -GalCer-treated mice were significantly lower than those in control mice, whereas levels of IL–4 and IL-5 in -GalCer-treated mice were significantly higher in -GalCer-treated mice (Fig. 2D, i-iv). These findings suggested that the administration of repeated -GalCer somehow influenced cytokine production by iNKT cells, which resulted in a difference in the engraftment of donor CD4+ T cells and a shift to the Th2 cytokine pattern early after transplantation.
In contrast, when J18–/– C57BL/6 mice were used as donors utilizing the same protocol as in the experiment shown in Fig. 1A (2 μg of -GalCer or control vehicle every 4 days from the day of transplantation), the survival of the recipient BALB/c mice was prolonged by -GalCer administration compared with vehicle administration (Fig. 4D), as was survival when wild-type C57BL/6 mice were used as donors (Fig. 1A). In this setting (donor: J18–/– C57BL/6), serum cytokine levels at 3–6 h posttransplant were significantly increased by -GalCer treatment, and the serum IL-4 and IL-5 levels on 5 or 6 days posttransplant in -GalCer-treated recipients were higher than those in vehicle-treated recipients and similar to those in wild-type recipients (data not shown).
]
These results collectively indicated that host-residual iNKT cells, rather than iNKT cells contained in the graft, are the major producers of various Th1 and Th2 cytokines shortly after transplant and key regulators of GVHD, and indeed are required for the regulation of GVHD by graft-contained and -adopted iNKT cells.
Combination of -GalCer pretreatment and use of iNKT cell-depleted grafts resulted in maximal GVHD reduction and graft rejection
In experiments in which J18–/– C57BL/6 mice were donors, we noticed that 2 of 25 recipients treated with -GalCer survived for >100 days, which was not the case if the wild-type C57BL/6 mice were donors (data not shown). Although this could represent a variation in the experimental conditions because there was no significant difference between the two groups, we expected that survival could be maximally improved if host-residual iNKT cells were stimulated before and after transplantation and iNKT cells were absent from the grafts. When we administered -GalCer on days –4, 0, and 4 of transplantation and transplanted the grafts from J18–/– C57BL/6 mice, 8 of 17 wild-type BALB/c recipient mice survived for >100 days without obvious GVHD (Fig. 5, A and B) as expected. Without -GalCer, there was no obvious difference in the survival of the recipients due to selection of the donor, i.e., wild-type or J18–/– C57BL/6 mice (cf Fig. 1A vs Fig. 5A and Fig. 1B vs Fig. 5B).
Among the eight long-term survivors described above in the setting of J18–/– C57BL/6 mice as donors and -GalCer started before transplantation, seven mice completely rejected the donor cells and the remaining one mouse exhibited mixed chimerism at 6 wk posttransplant (data not shown). Therefore, we examined the time course of donor cell chimerism early after transplantation in recipients with -GalCer that was started before transplantation. Both Gr-1+ and CD4+ T cells were engrafted in vehicle-treated mice (Fig. 5C). In contrast, both Gr-1+ and CD4+ T cells were rejected early after transplantation in -GalCer-treated mice. Particularly, Gr-1+ cells were never engrafted (Fig. 5C).
Since it is known that NK cells are major effectors in graft rejection (27) and play a role as effectors of iNKT cells in antitumor immunity by secreting IFN- (1, 9), we performed in vivo NK depletion by administering anti-asialo GM1 Ab. It is of note that iNKT was not depleted by this treatment (data not shown). However, the prolongation of survival (Fig. 6A) and graft rejection (Fig. 6B) by -GalCer were still observed as seen without anti-asialo GM1 Ab. For graft rejection, therefore, some targets other than NK cells should be considered as effectors of host-residual iNKT cells activated by -GalCer, particularly in the absence of donor iNKT cells.
Discussion
Many studies have suggested that an important physiological function of iNKT cells is to control immune responses against autoimmunity, infection, and tumors (1, 9). In transplantation immunity, iNKT cells are also thought to play a role in the induction of allograft (28, 29, 30) or xenograft tolerance (31). In this study, we examined the role of iNKT cells in GVHD mouse model systems, using an iNKT stimulator -GalCer, adoptive transfer of in vitro-expanded iNKT cells, and J18–/– mice (16), and found that host-residual iNKT cells can attenuate GVHD.
Some reports have suggested that both donor bone marrow-derived (10) and host-residual (11, 12) NKT cells (NK1.1+ or DX5+ T cells) may suppress acute GVHD. These NKT cell populations should overlap with the iNKT cell population that we describe here. Therefore, the attenuation of GVHD by adoptive transfer of in vitro-expanded iNKT cells described here is consistent with previous results.
Recently, a report has shown that -GalCer administration to recipients prolonged survival of GVHD mice and its administration to CD1d–/– mice did not prolong their survival (14). Our results show more direct evidence that such findings are caused by the functional activation of iNKT cells, with the use of J18–/– mice. Moreover, the need for host-residual iNKT cells was clearly shown by the result that adoptive iNKT cells from either strain did not attenuate GVHD if transferred to J18–/– BALB/c recipients.
Surprisingly, host-residual iNKT cells were maintained in the liver early after transplantation if C57BL/6 (donor strain)-derived iNKT cells were transferred, whereas very few host-residual iNKT cells were detected without adoptive iNKT transfer (Fig. 3C, i, iii, and iv). We could not distinguish the origin of H-2Dd+ iNKT cells detected in the recipient liver if BALB/c strain-derived iNKT cells were transferred (Fig. 3C, ii and iv). It is possible that they also represent host-residual rather than injected iNKT cells. Taken together, these findings suggest that the attenuation of GVHD by adoptively transferred iNKT cells is likely to occur through the maintenance of host-residual iNKT cells, although the precise mechanism remains to be elucidated. It should also be clarified whether injected and host-residual iNKT cells locally interact with each other in a specific tissue. Particularly, it would be highly desirable if we could visualize iNKT cells in the liver by marking their strain and origin, which would be possible only after technical advances are available for the specific staining of iNKT cells. Analyzing a direct interaction of liver-isolated iNKT cells and activated MHC-mismatched iNKT cells would be of great interest, but impracticable, given the extremely low proportion of liver iNKT cells. In addition, the -GalCer-loaded CD1d tetramer is the only tool for specifically staining iNKT cells, and the isolation procedure for iNKT cells might stimulate them and thus influence the results of in vitro analyses.
Other unexpected results include the delay in engraftment, the induction of mixed chimerism, and graft rejection by host-residual iNKT cells, particularly if -GalCer administration was started before transplantation and J18–/– C57BL/6 mice were used as donors. This setting conferred the maximal survival benefit to the recipients because of the mildness of GVHD, albeit this occurred in our meticulous experimental mouse model. Possibly, the activated host-residual iNKT cells may suppress donor CD4+ T cell function or stimulate host T cell function before total-body irradiation and transplantation. Given that mixed chimerism induces GVHD tolerance (32), host-residual iNKT cells may provide the attenuation of GVHD and the induction of mixed chimerism and graft rejection through a common mechanism that regulates graft-vs-donor immunity.
Induction of a Th2-dominant cytokine profile before the onset of obvious GVHD after transplantation (Fig. 2B) may, at least in part, be associated with such a mechanism, since many studies (33, 34, 35), with some conflicting reports (36, 37), have shown that Th2 cytokines protect against GVHD. In addition, some investigators have reported that IL-2, TNF-, and IFN- play important roles in the development of GVHD in vivo (35, 38, 39, 40, 41). Induction of cytokine production from residual iNKT cells by administration of -GalCer or iNKT a few hours after transplantation was obvious. Th1 as well as Th2 cytokine secretion at the very early stage of transplantation may be favorable for balancing the recipients’ and donors’ T cell function and, as a result, may suppress GVHD. Besides, repeated -GalCer administration may induce a Th2-dominant cytokine profile. These considerations, as well as previous reports that the GVHD-protective effect of NKT cells depends on IL-4 production from NKT cells (10, 11, 12, 14), support our speculation. Regarding other cytokines, several reports have noted that IL-12 (34, 42), IL-13 (43, 44), IL-15 (34, 38), and IL-18 (45) are also associated with GVHD. However, we could not obtain sufficient samples from sick mice to measure such various cytokines simultaneously. Although we were unable to evaluate all of these cytokines in this study, we hope we will be able to analyze a complete set of cytokines in future studies.
To further complicate this scenario, our findings also revealed that the absence of iNKT cells in the graft enhances the suppression of engraftment of donor cells. A simple interpretation of this result is that iNKT cells in the graft help donor cells engraft, which is apparently opposite to the effect of host-residual iNKT cells. While attenuating GVHD by inducing host-residual iNKT cells, graft-contaminated iNKT cells may suppress host-vs-graft immunity. We observed a delay of engraftment when we adoptively transferred iNKT cells expanded from the BALB/c strain while transplanting grafts from J18–/– C57BL/6 mice (data not shown). Furthermore, we observed high levels of serum cytokines after transplantation and in vitro-expanded iNKT cells only when recipients had iNKT cells. Therefore, these functional differences may simply depend on the place, tissue-residual iNKT cells or blood-borne iNKT cells that are pre-expanded in vitro. Under physiological conditions, the cytokine status could be created by iNKT cells in the liver to prevent autoreactivity, and this environment may be able to prevent GVHD and influence incoming iNKT cells to prevent the attack of the donor graft by the recipient lymphocytes. Alternatively, the suppression of graft-vs-host immunity by host-residual iNKT cells and the suppression of host-vs-graft immunity by graft-contaminated iNKT cells could also be explained by the recognition of non-self through members of Ly49 (46, 47, 48). At least NK cells, which were generally recognized as major effectors in graft rejection, are not the downstream effectors of iNKT cell-dependent graft rejection.
In the clinical setting, there is increasing interest in the kinetics of the establishment of donor chimerism because of the development of reduced intensity conditioning regimens for allogeneic stem cell transplantation. Our results suggest that recipient-residual iNKT cells play a role against the establishment of donor chimerism as well as in the prevention of GVHD. Warn us that -GalCer must be used carefully to prevent or treat GVHD are the facts that the combination of overstimulation of recipient iNKT cells before transplantation and that the lack of iNKT cells in grafts can cause graft rejection.
In conclusion, host-residual iNKT cells have a regulatory function in GVHD. -GalCer therapy has already been performed in clinical trials in cancer patients and was well tolerated (49). It may be attractive to use -GalCer or iNKT cells therapeutically for the prevention or treatment of GVHD. However, care must be taken in its clinical application because of the possibility that the activation of host-residual iNKT cells could also increase graft rejection.
Acknowledgments
We thank M. Kronenberg (La Jolla Institute for Allergy and Immunology, La Jolla, CA) for agreeing to provide us with -GalCer-loaded murine CD1d tetramer. We thank M. Harada (Chiba University, Chiba, Japan) for preparing J18–/– mice, T. Nakayama (Chiba University) for providing the Sf9 cell line and baculovirus-expressing mouse CD1d/2-microgloblin, and T. Ito (Chiba University) for considerable advice on the production of -GalCer-loaded CD1d tetramers. We also thank E. Nagata and Y. Sato for providing excellent technical assistance.
Disclosures
The authors have no financial conflict of interest.
Footnotes
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.
1 This work was supported in part by a Grant-in-Aid for Scientific Research, KAKENHI (14370300), and Research on Regulatory Sciences of Pharmaceuticals and Medical Devices. Health and Labour Sciences Research Grants. Ministry of Health, Labour and Welfare of Japan.
2 Current address: Division of Immunology & Rheumatology, Stanford University School of Medicine, Palo Alto, CA.
3 Hisamaru Hirai died on August 23, 2003.
4 Address correspondence and reprints requests to Dr. Shigeru Chiba, Department of Cell Therapy and Transplantation Medicine, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8655, Japan. E-mail address: schiba-tky@umin.ac.jp
5 Abbreviations used in this paper: iNKT, invariant NK T; -GalCer, -galactosylceramide; GVHD, graft-vs-host disease; CBA, cytometric bead array, cRPMI, complete RPMI.
Received for publication December 9, 2004. Accepted for publication April 21, 2005.
References
Godfrey, D., K. Hammond, L. Poulton, A. Baxter. 2000. NKT cells: Facts, functions and fallacies. Immunol. Today 21: 573-583.
Wilson, M. T., A. K. Singh, L. Kaer. 2002. Immunotherapy with ligands of natural killer T cells. Trends Mol. Med. 8: 225-231.
Budd, R., G. Miescher, R. Howe, R. Lees, C. Bron, H. MacDonald. 1987. Developmentally regulated expression of T cell receptor chain variable domain is immature thymocytes. J. Exp. Med. 166: 577-582.
Dellabona, P., E. Padovan, G. Casorati, M. Brockhaus, A. Lanzavecchia. 1994. An invariant V24-JQ/V11 T cell receptor is expressed in all individuals by clonally expanded CD4–CD8– cells. J. Exp. Med. 180: 1171-1176.
Porcelli, S., D. Gerdes, A. Fertig, S. Balk. 1996. Human T cells expressing an invariant V24-JQ TCR are CD4 –and heterogenous with respect to TCR expression. Hum. Immunol. 48: 63-67.
Lantz, O., A. Bendelac. 1994. An invariant T cell receptor chain is used by a unique subset of major histocompatibility complex class I-specific CD4+ and CD4–8– T cells in mice and humans. J. Exp. Med. 180: 1097-1106.
Taniguchi, M., M. Harada, S. Kojo, T. Nakayama, H. Wakao. 2003. The regulatory role of V14 NKT cells in innate and acquired immune response. Annu. Rev. Immunol. 21: 483-513.[Medline]
Joyce, S.. 2001. CD1d and natural T cells: How their properties jump-start the immune system. Cell Mol. Life Sci. 58: 442-469.(Kyoko Haraguchi*,, Tsuyos)
Invariant NK T (iNKT) cells have an invariant TCR- chain and are activated in a CD1d-restricted manner. They are thought to regulate immune responses and play important roles in autoimmunity, allergy, infection, and tumor immunity. They also appear to influence immunity after hemopoietic stem cell transplantation. In this study, we examined the role of iNKT cells in graft-vs-host disease (GVHD) and graft rejection in a mouse model of MHC-mismatched bone marrow transplantation, using materials including -galactosylceramide, NKT cells expanded in vitro, and J18 knockout mice that lack iNKT cells. We found that host-residual iNKT cells constitute effector cells which play a crucial role in reducing the severity of GVHD, and that this reduction is associated with a delayed increase in serum Th2 cytokine levels. Interestingly, we also found that host-residual iNKT cause a delay in engraftment and, under certain conditions, graft rejection. These results indicate that host-residual iNKT cells attenuate graft-vs-host immunity rather than host-vs-graft immunity.
Introduction
Natural killer T cells are a population of T cells that have NK cell markers such as NK1.1 (NKR-P1C) in mice or CD161 (NKR-P1A) in humans (1, 2). Some NK T cells use an invariant TCR- chain (V14-J18 in mice, V24-J18 in humans) paired with V8, V7, or V2 in mice or with V11 in humans (3, 4, 5, 6, 7), and are called invariant NKT (iNKT)5 cells. iNKT cells are activated by synthetic glycolipids such as -galactosylceramide (-GalCer) in a CD1d-restricted manner (1, 2, 8, 9). iNKT cells produce both Th1 (such as IFN- and TNF-) and Th2 (such as IL-4, IL-5, IL-10, and IL-13) cytokines (1, 2, 8, 9). It has been reported that iNKT cells control immune responses in some infections, tumors, autoimmune diseases, and allograft rejection (1, 2, 8, 9).
Graft-vs-host disease (GVHD) is one of the most serious complications in hemopoietic stem cell transplantation. It has been suggested in a mouse acute GVHD model that NK1.1+ T cells obtained from donor bone marrow can suppress GVHD induced by peripheral blood transplantation from the same donor (10). It has also been shown that a selected conditioning regimen, which preserves more host-residual NK1.1+ or DX5+ T cells than other T cells, is advantageous for reducing acute GVHD (11, 12). Furthermore, the suppressive effect of -GalCer on induced acute GVHD has been demonstrated in a mouse model(13, 14). We previously reported that the number of iNKT cells was lower in patients with GVHD than in those without GVHD (15) after hemopoietic stem cell transplantation, although the cause-effect relationship was not clear.
In this report, we provide direct evidence that host-residual iNKT cells reduce GVHD in a mouse model of MHC-mismatched bone marrow transplantation using J18 knockout mice that lack iNKT cells (16). Adoptively transferred iNKT cells with grafts also reduce GVHD, but, importantly, this effect is dependent on the presence of host-residual iNKT cells.
Results
Establishment of GVHD model mice
In the prototype transplantation from wild-type C57BL/6 to wild-type BALB/c, all of the recipients showed lethal GVHD as judged by the GVHD score. Histological examinations also confirmed GVHD in representative mice (data not shown). In this setting, we confirmed that full donor chimerism was achieved in all of the recipients (data not shown). We performed all of the transplantation experiments two to four times independently, and integrated all of the results from individual experiments to avoid experiment-to-experiment variation.
Administration of -GalCer prolonged survival of GVHD mice
First, we administered either -GalCer to activate iNKT cells or control vehicle every 4 days from the day of transplantation. -GalCer-treated mice survived significantly longer than control mice (p < 0.0001, Fig. 1A). The severity of GVHD in -GalCer-treated mice was milder than that in control mice (Fig. 1B). The GVHD score within the first 30 days after transplantation in the two groups was compared by repeated measures ANOVA and was found to be significantly lower in the -GalCer group (p < 0.0001).
We then examined whether -GalCer treatment influenced the number of iNKT cells in the liver, which contains the largest proportion of iNKT cells (1, 9) (14.2 ± 4.7% in our measurements; Fig. 1Di), after transplantation. In control recipients, the ratio of iNKT cells to lymphocytes decreased rapidly and iNKT cells became undetectable by approximately day 5 posttransplant (Fig. 1, C and D). The rate of decrease in the iNKT cells was delayed in -GalCer-treated mice, although iNKT cells were scarcely detectable after day 10 even in mice treated with -GalCer (Fig. 1C). We observed temporary donor-type iNKT cell chimerism before the disappearance, although the time course of the donor:recipient iNKT cell ratio was highly variable (data not shown).
Interestingly, we found that the engraftment of donor CD4+ T cells in -GalCer-treated mice was delayed compared with that in control mice during the early phase after transplantation, although complete donor chimerism was eventually achieved (Fig. 2A). Engraftment of Gr-1+ cells was also slightly delayed, but engraftment of other lineage cells was similar to that seen in vehicle-treated mice (data not shown). These findings encouraged us to compare various cytokine levels between -GalCer-treated and control mice, since iNKT cells produce high levels of both Th1 and Th2 cytokines and could influence Th1/Th2 polarization. Therefore, we examined the serum levels of IFN-, TNF-, IL–4, and IL-5 shortly after transplant and found that all of the cytokines were increased by -GalCer treatment (Fig. 2B, i-iv), reproducing previous reports that -GalCer has been shown to rapidly stimulate Th1 and Th2 cytokine production in vivo in the nontransplantation model (23, 24, 25). In contrast, it has been reported that in vivo -GalCer-primed splenocytes secrete less amounts of Th1 cytokines after further in vitro treatment with -GalCer compared with vehicle-primed splenocytes (23, 26). In our experimental system, we could exactly reproduce these findings (Fig. 2C, i-iv). Then we measured all four cytokines on day 5 or 6 posttransplant (1 or 2 days after the second -GalCer administration), both in -GalCer- and vehicle-treated mice. At this time point, GVHD signs were not been obvious yet, although the damage from radiation was inseparably measured "GVHD score" in Fig. 1B. IFN- and TNF- levels in -GalCer-treated mice were significantly lower than those in control mice, whereas levels of IL–4 and IL-5 in -GalCer-treated mice were significantly higher in -GalCer-treated mice (Fig. 2D, i-iv). These findings suggested that the administration of repeated -GalCer somehow influenced cytokine production by iNKT cells, which resulted in a difference in the engraftment of donor CD4+ T cells and a shift to the Th2 cytokine pattern early after transplantation.
In contrast, when J18–/– C57BL/6 mice were used as donors utilizing the same protocol as in the experiment shown in Fig. 1A (2 μg of -GalCer or control vehicle every 4 days from the day of transplantation), the survival of the recipient BALB/c mice was prolonged by -GalCer administration compared with vehicle administration (Fig. 4D), as was survival when wild-type C57BL/6 mice were used as donors (Fig. 1A). In this setting (donor: J18–/– C57BL/6), serum cytokine levels at 3–6 h posttransplant were significantly increased by -GalCer treatment, and the serum IL-4 and IL-5 levels on 5 or 6 days posttransplant in -GalCer-treated recipients were higher than those in vehicle-treated recipients and similar to those in wild-type recipients (data not shown).
]
These results collectively indicated that host-residual iNKT cells, rather than iNKT cells contained in the graft, are the major producers of various Th1 and Th2 cytokines shortly after transplant and key regulators of GVHD, and indeed are required for the regulation of GVHD by graft-contained and -adopted iNKT cells.
Combination of -GalCer pretreatment and use of iNKT cell-depleted grafts resulted in maximal GVHD reduction and graft rejection
In experiments in which J18–/– C57BL/6 mice were donors, we noticed that 2 of 25 recipients treated with -GalCer survived for >100 days, which was not the case if the wild-type C57BL/6 mice were donors (data not shown). Although this could represent a variation in the experimental conditions because there was no significant difference between the two groups, we expected that survival could be maximally improved if host-residual iNKT cells were stimulated before and after transplantation and iNKT cells were absent from the grafts. When we administered -GalCer on days –4, 0, and 4 of transplantation and transplanted the grafts from J18–/– C57BL/6 mice, 8 of 17 wild-type BALB/c recipient mice survived for >100 days without obvious GVHD (Fig. 5, A and B) as expected. Without -GalCer, there was no obvious difference in the survival of the recipients due to selection of the donor, i.e., wild-type or J18–/– C57BL/6 mice (cf Fig. 1A vs Fig. 5A and Fig. 1B vs Fig. 5B).
Among the eight long-term survivors described above in the setting of J18–/– C57BL/6 mice as donors and -GalCer started before transplantation, seven mice completely rejected the donor cells and the remaining one mouse exhibited mixed chimerism at 6 wk posttransplant (data not shown). Therefore, we examined the time course of donor cell chimerism early after transplantation in recipients with -GalCer that was started before transplantation. Both Gr-1+ and CD4+ T cells were engrafted in vehicle-treated mice (Fig. 5C). In contrast, both Gr-1+ and CD4+ T cells were rejected early after transplantation in -GalCer-treated mice. Particularly, Gr-1+ cells were never engrafted (Fig. 5C).
Since it is known that NK cells are major effectors in graft rejection (27) and play a role as effectors of iNKT cells in antitumor immunity by secreting IFN- (1, 9), we performed in vivo NK depletion by administering anti-asialo GM1 Ab. It is of note that iNKT was not depleted by this treatment (data not shown). However, the prolongation of survival (Fig. 6A) and graft rejection (Fig. 6B) by -GalCer were still observed as seen without anti-asialo GM1 Ab. For graft rejection, therefore, some targets other than NK cells should be considered as effectors of host-residual iNKT cells activated by -GalCer, particularly in the absence of donor iNKT cells.
Discussion
Many studies have suggested that an important physiological function of iNKT cells is to control immune responses against autoimmunity, infection, and tumors (1, 9). In transplantation immunity, iNKT cells are also thought to play a role in the induction of allograft (28, 29, 30) or xenograft tolerance (31). In this study, we examined the role of iNKT cells in GVHD mouse model systems, using an iNKT stimulator -GalCer, adoptive transfer of in vitro-expanded iNKT cells, and J18–/– mice (16), and found that host-residual iNKT cells can attenuate GVHD.
Some reports have suggested that both donor bone marrow-derived (10) and host-residual (11, 12) NKT cells (NK1.1+ or DX5+ T cells) may suppress acute GVHD. These NKT cell populations should overlap with the iNKT cell population that we describe here. Therefore, the attenuation of GVHD by adoptive transfer of in vitro-expanded iNKT cells described here is consistent with previous results.
Recently, a report has shown that -GalCer administration to recipients prolonged survival of GVHD mice and its administration to CD1d–/– mice did not prolong their survival (14). Our results show more direct evidence that such findings are caused by the functional activation of iNKT cells, with the use of J18–/– mice. Moreover, the need for host-residual iNKT cells was clearly shown by the result that adoptive iNKT cells from either strain did not attenuate GVHD if transferred to J18–/– BALB/c recipients.
Surprisingly, host-residual iNKT cells were maintained in the liver early after transplantation if C57BL/6 (donor strain)-derived iNKT cells were transferred, whereas very few host-residual iNKT cells were detected without adoptive iNKT transfer (Fig. 3C, i, iii, and iv). We could not distinguish the origin of H-2Dd+ iNKT cells detected in the recipient liver if BALB/c strain-derived iNKT cells were transferred (Fig. 3C, ii and iv). It is possible that they also represent host-residual rather than injected iNKT cells. Taken together, these findings suggest that the attenuation of GVHD by adoptively transferred iNKT cells is likely to occur through the maintenance of host-residual iNKT cells, although the precise mechanism remains to be elucidated. It should also be clarified whether injected and host-residual iNKT cells locally interact with each other in a specific tissue. Particularly, it would be highly desirable if we could visualize iNKT cells in the liver by marking their strain and origin, which would be possible only after technical advances are available for the specific staining of iNKT cells. Analyzing a direct interaction of liver-isolated iNKT cells and activated MHC-mismatched iNKT cells would be of great interest, but impracticable, given the extremely low proportion of liver iNKT cells. In addition, the -GalCer-loaded CD1d tetramer is the only tool for specifically staining iNKT cells, and the isolation procedure for iNKT cells might stimulate them and thus influence the results of in vitro analyses.
Other unexpected results include the delay in engraftment, the induction of mixed chimerism, and graft rejection by host-residual iNKT cells, particularly if -GalCer administration was started before transplantation and J18–/– C57BL/6 mice were used as donors. This setting conferred the maximal survival benefit to the recipients because of the mildness of GVHD, albeit this occurred in our meticulous experimental mouse model. Possibly, the activated host-residual iNKT cells may suppress donor CD4+ T cell function or stimulate host T cell function before total-body irradiation and transplantation. Given that mixed chimerism induces GVHD tolerance (32), host-residual iNKT cells may provide the attenuation of GVHD and the induction of mixed chimerism and graft rejection through a common mechanism that regulates graft-vs-donor immunity.
Induction of a Th2-dominant cytokine profile before the onset of obvious GVHD after transplantation (Fig. 2B) may, at least in part, be associated with such a mechanism, since many studies (33, 34, 35), with some conflicting reports (36, 37), have shown that Th2 cytokines protect against GVHD. In addition, some investigators have reported that IL-2, TNF-, and IFN- play important roles in the development of GVHD in vivo (35, 38, 39, 40, 41). Induction of cytokine production from residual iNKT cells by administration of -GalCer or iNKT a few hours after transplantation was obvious. Th1 as well as Th2 cytokine secretion at the very early stage of transplantation may be favorable for balancing the recipients’ and donors’ T cell function and, as a result, may suppress GVHD. Besides, repeated -GalCer administration may induce a Th2-dominant cytokine profile. These considerations, as well as previous reports that the GVHD-protective effect of NKT cells depends on IL-4 production from NKT cells (10, 11, 12, 14), support our speculation. Regarding other cytokines, several reports have noted that IL-12 (34, 42), IL-13 (43, 44), IL-15 (34, 38), and IL-18 (45) are also associated with GVHD. However, we could not obtain sufficient samples from sick mice to measure such various cytokines simultaneously. Although we were unable to evaluate all of these cytokines in this study, we hope we will be able to analyze a complete set of cytokines in future studies.
To further complicate this scenario, our findings also revealed that the absence of iNKT cells in the graft enhances the suppression of engraftment of donor cells. A simple interpretation of this result is that iNKT cells in the graft help donor cells engraft, which is apparently opposite to the effect of host-residual iNKT cells. While attenuating GVHD by inducing host-residual iNKT cells, graft-contaminated iNKT cells may suppress host-vs-graft immunity. We observed a delay of engraftment when we adoptively transferred iNKT cells expanded from the BALB/c strain while transplanting grafts from J18–/– C57BL/6 mice (data not shown). Furthermore, we observed high levels of serum cytokines after transplantation and in vitro-expanded iNKT cells only when recipients had iNKT cells. Therefore, these functional differences may simply depend on the place, tissue-residual iNKT cells or blood-borne iNKT cells that are pre-expanded in vitro. Under physiological conditions, the cytokine status could be created by iNKT cells in the liver to prevent autoreactivity, and this environment may be able to prevent GVHD and influence incoming iNKT cells to prevent the attack of the donor graft by the recipient lymphocytes. Alternatively, the suppression of graft-vs-host immunity by host-residual iNKT cells and the suppression of host-vs-graft immunity by graft-contaminated iNKT cells could also be explained by the recognition of non-self through members of Ly49 (46, 47, 48). At least NK cells, which were generally recognized as major effectors in graft rejection, are not the downstream effectors of iNKT cell-dependent graft rejection.
In the clinical setting, there is increasing interest in the kinetics of the establishment of donor chimerism because of the development of reduced intensity conditioning regimens for allogeneic stem cell transplantation. Our results suggest that recipient-residual iNKT cells play a role against the establishment of donor chimerism as well as in the prevention of GVHD. Warn us that -GalCer must be used carefully to prevent or treat GVHD are the facts that the combination of overstimulation of recipient iNKT cells before transplantation and that the lack of iNKT cells in grafts can cause graft rejection.
In conclusion, host-residual iNKT cells have a regulatory function in GVHD. -GalCer therapy has already been performed in clinical trials in cancer patients and was well tolerated (49). It may be attractive to use -GalCer or iNKT cells therapeutically for the prevention or treatment of GVHD. However, care must be taken in its clinical application because of the possibility that the activation of host-residual iNKT cells could also increase graft rejection.
Acknowledgments
We thank M. Kronenberg (La Jolla Institute for Allergy and Immunology, La Jolla, CA) for agreeing to provide us with -GalCer-loaded murine CD1d tetramer. We thank M. Harada (Chiba University, Chiba, Japan) for preparing J18–/– mice, T. Nakayama (Chiba University) for providing the Sf9 cell line and baculovirus-expressing mouse CD1d/2-microgloblin, and T. Ito (Chiba University) for considerable advice on the production of -GalCer-loaded CD1d tetramers. We also thank E. Nagata and Y. Sato for providing excellent technical assistance.
Disclosures
The authors have no financial conflict of interest.
Footnotes
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.
1 This work was supported in part by a Grant-in-Aid for Scientific Research, KAKENHI (14370300), and Research on Regulatory Sciences of Pharmaceuticals and Medical Devices. Health and Labour Sciences Research Grants. Ministry of Health, Labour and Welfare of Japan.
2 Current address: Division of Immunology & Rheumatology, Stanford University School of Medicine, Palo Alto, CA.
3 Hisamaru Hirai died on August 23, 2003.
4 Address correspondence and reprints requests to Dr. Shigeru Chiba, Department of Cell Therapy and Transplantation Medicine, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8655, Japan. E-mail address: schiba-tky@umin.ac.jp
5 Abbreviations used in this paper: iNKT, invariant NK T; -GalCer, -galactosylceramide; GVHD, graft-vs-host disease; CBA, cytometric bead array, cRPMI, complete RPMI.
Received for publication December 9, 2004. Accepted for publication April 21, 2005.
References
Godfrey, D., K. Hammond, L. Poulton, A. Baxter. 2000. NKT cells: Facts, functions and fallacies. Immunol. Today 21: 573-583.
Wilson, M. T., A. K. Singh, L. Kaer. 2002. Immunotherapy with ligands of natural killer T cells. Trends Mol. Med. 8: 225-231.
Budd, R., G. Miescher, R. Howe, R. Lees, C. Bron, H. MacDonald. 1987. Developmentally regulated expression of T cell receptor chain variable domain is immature thymocytes. J. Exp. Med. 166: 577-582.
Dellabona, P., E. Padovan, G. Casorati, M. Brockhaus, A. Lanzavecchia. 1994. An invariant V24-JQ/V11 T cell receptor is expressed in all individuals by clonally expanded CD4–CD8– cells. J. Exp. Med. 180: 1171-1176.
Porcelli, S., D. Gerdes, A. Fertig, S. Balk. 1996. Human T cells expressing an invariant V24-JQ TCR are CD4 –and heterogenous with respect to TCR expression. Hum. Immunol. 48: 63-67.
Lantz, O., A. Bendelac. 1994. An invariant T cell receptor chain is used by a unique subset of major histocompatibility complex class I-specific CD4+ and CD4–8– T cells in mice and humans. J. Exp. Med. 180: 1097-1106.
Taniguchi, M., M. Harada, S. Kojo, T. Nakayama, H. Wakao. 2003. The regulatory role of V14 NKT cells in innate and acquired immune response. Annu. Rev. Immunol. 21: 483-513.[Medline]
Joyce, S.. 2001. CD1d and natural T cells: How their properties jump-start the immune system. Cell Mol. Life Sci. 58: 442-469.(Kyoko Haraguchi*,, Tsuyos)