Requirement for Vα14 NKT Cells in IL-12-Mediated Rejection of Tumors

See allHide authors and affiliations

Science  28 Nov 1997:
Vol. 278, Issue 5343, pp. 1623-1626
DOI: 10.1126/science.278.5343.1623


A lymphocyte subpopulation, the Vα14 natural killer T (NKT) cells, expresses both NK1.1 and a single invariant T cell receptor encoded by the Vα14 and Jα281 gene segments. Mice with a deletion of the Jα281 gene segment were found to exclusively lack this subpopulation. The Vα14 NKT cell–deficient mice could no longer mediate the interleukin-12 (IL-12)–induced rejection of tumors. Although the antitumor effect of IL-12 was thought to be mediated through natural killer cells and T cells, Vα14 NKT cells were found to be an essential target of IL-12, and they mediated their cytotoxicity by an NK-like effector mechanism after activation with IL-12.

Vα14 NKT cells, originally defined as NK1.1+ T cells (1), are a distinct lymphoid lineage expressing several characteristics that demarcate them from T cells, natural killer (NK) cells, and B cells. First, Vα14 NKT cells develop outside of the thymus at an early stage of embryogenesis (2) and belong to the CD4CD8 or CD4+CD8 populations that express heat-stable antigen and the B220+, Mac-1+, CD44+, CD45RBdim, and Mel-14dimphenotypes (3-5). Second, Vα14 NKT cells express a single invariant T cell receptor α chain (TCRα) encoded by the Vα14 and Jα281 segments (4, 6, 7), mostly in association with Vβ8 (3, 5), which is not expressed on conventional T cells (8). Because of the unique expression of invariant Vα14 TCR, transgenic (tg) mice that only express the invariant Vα14 TCR and no endogenous TCRα (TCRα / Vα14tg) preferentially develop Vα14 NKT cells and block conventional αβ+ T cell development (8). Therefore, we speculated that disruption of the genes encoding the invariant Vα14 TCR would generate mice lacking Vα14 NKT cells while leaving the other lymphoid lineages intact.

IL-12 is known to mediate antitumor effects and was originally hypothesized to interact with NK cells or CD8+ cytotoxic T lymphocytes (CTLs) (9). However, recent studies indicate that the systemic administration of IL-12 activates NK1.1+TCRintermediate cells (which are similar to Vα14 NKT cells) and inhibits the hepatic metastasis of tumor (10). It is thus unclear whether IL-12 interacts with NK, T, or Vα14 NKT cells. It is possible that Vα14 NKT cells may be responsible for functions that have been reported for NK cells, because purifications were difficult and the studies were done with a mixture of NK and Vα14 NKT cells. We investigated the antitumor effects of IL-12 to determine if the functional target of the cytokine in vivo is NK cells, CD8+ CTLs, or Vα14 NKT cells.

To address this question, we generated mice that lacked invariant Vα14 Jα281 TCR expression by specific deletion of the Jα281 gene segment using homologous recombination and aggregation chimera techniques (11) (Fig. 1, A and B). The loss of Jα281 gene expression in Jα281 / mice was confirmed by ribonuclease (RNase) protection assay (12) (Fig. 1C). Thus, the invariant Vα14 Jα281 TCR is not expressed in Jα281 / mice. The development of the lymphoid organs in Jα281 / mice was macroscopically normal, and the numbers of lymphocytes were almost the same as in wild-type mice. However, the relative numbers of NK1.1+ TCRβ+ cells in the thymus, spleen, bone marrow, and liver were reduced in Jα281 / mice (Fig.1D). On the other hand, the total number of other hematopoietic cells were not affected (see legend to Fig. 1). Thus, the defects in hematopoietic cell development were subtle, and Vα14 NKT cells were exclusively affected. This confirms the requirement of invariant Vα14 TCR for the development of Vα14 NKT cells (8).

Figure 1

Generation and analysis of Jα281–/–(Vα14 NKT–deficient) mice. (A) Targeting construct. Targeting vector pJ1 (top) was designed to replace the Cla I fragment (1.0 kb) containing the Jα281 gene segment (middle) with a diphtheria toxin A (DTA) cassette and a neomycin resistant gene cassette (designated Neo) derived from pPGK-neo r-poly A (PGK, 3-phosphoglycerate kinase). Sal I–Cla I (1.2 kb) and Cla I (12 kb) genomic fragments of 129Sv/J origin were used for the short and long arms, respectively. The targeting vector was introduced into R1 embryonic stem cells of 129 mice by electroporation. Germline chimeric mice were generated by aggregation methods (11). Mutant mice were backcrossed three generations with C57BL/6 mice. Mutant alleles (bottom) were detected by Southern (DNA) blot analysis with probes A and B. X, Xho I; S, Sal I; C, Cla I; H, Hind III; N, Not I. (B) Southern blot analysis. Wild-type and mutant alleles were detected by Hind III digestion as 7.2-kb and 3.6-kb bands, respectively. (C) Ribonuclease protection assay with the Jα281Cα probe (12), which detects the Jα281Cα (227 bp) and Cα (211 bp) bands. (D) Fluorescence-activated cell sorter (FACS) profiles. The percent of the gated populations analyzed by FACS (23) is indicated. Total numbers of hematopoietic cells in Jα281–/– mice were as follows: 2.2 × 106 NK1.1+, 3.0 × 107 αβ+ T, 1 × 106 γδ+ T, 3.0 × 107 IgM+ B, 6.5 × 106 Mac−1+, and 1.5 × 106 Gr-1+ cells.

To elucidate the functional defects in vivo that arose as a result of the selective loss of the Vα14 NKT cell population, we investigated the antitumor activity induced by IL-12. One model was intrasplenic injection of FBL-3 erythroleukemia (Fig. 2A) or B16 melanoma (Fig. 2B), which preferentially metastasize to the liver (13). Intraperitoneal administration of IL-12 after the intrasplenic graft of FBL-3 or B16 suppressed metastasis in the liver of Jα281+/+ wild-type but not Jα281 / mice (Fig.2, A and B). Because the mutant mice exclusively lack Vα14 NKT cells, with the other lymphoid populations left intact (Vα14 NKT–deficient mice) (Fig. 1D), this implies that Vα14 NKT cells are a primary functional target of IL-12 in vivo. To exclude the possibility that a T cell population other than Vα14 NKT cells is the primary target of IL-12, we examined antitumor cytotoxicity in transgenic (Vα14tgVβ8.2tg) mice lacking recombination activating gene (RAG) (RAG / Vα14tgVβ8.2tg), which preferentially generate Vα14 NKT cells but block the development of any other lymphocyte lineages, including NK, B, and T cells (Vα14 NKT mice) (14). In these mice, FBL-3 or B16 tumor cells were rejected as in wild-type mice (Fig.2, A and B), indicating that NK and T cells are not required for IL-12–mediated tumor rejection. Similar results were also obtained with other tumor systems, such as subcutaneous growth of B16 (Fig. 2C) and pulmonary metastases of B16 (Fig. 2D) or Lewis lung carcinoma (LLC) cells (Fig. 2E). These results confirm the previous observation of the protective effects of IL-12 on those tumors in vivo (15-17). In addition, RAG–/– mice having only NK cells could not reject metastasis (18), and IL-12 is effective in NK cell–deficient beige mice or anti-asialoGM1-treated NK-depleted mice (17), indicating that NK cells are not a primary target of IL-12. Thus, conventional T and NK cells seem to be unnecessary for the IL-12–mediated rejection of tumors.

Figure 2

Inhibition of tumor growth or metastasis in IL-12–treated mice. Protocols are in (24). (A) Number of metastatic nodules of FBL-3 in the liver. (B) Amounts of GM3 melanoma antigens (Ag) in the metastatic liver and their representative photographic views of B16 melanoma. Methods for measurement of melanoma antigens are in (22). Each group had five to seven mice; cpm, counts per minute. (C) Subcutaneous growth of B16. Symbols for mice treated with PBS (open) or IL-12 (closed) are indicated. (D) Number of metastatic nodules of B16 in the lung. (E) Number of metastatic nodules of LLC in the lung and their representative photographic views. Each group had three mice.

We further examined IL-12–induced antitumor activity in vitro (Fig. 3). The induction of antitumor cytotoxicity was impaired in Jα281–/– mice (Fig. 3B), whereas B16 was killed in RAG–/–Vα14tgVβ8.2tg mice to a similar extent as in Jα281+/+ mice (Fig. 3, A and C), again confirming that Vα14 NKT cells are responsible for IL-12–mediated cytotoxicity. The most important observation on the effector mechanisms is that Vα14 NKT cells kill the target through direct contact, because their cytotoxicity was specifically inhibited by unlabeled B16 (Fig. 3D). However, the molecules responsible for the killing interactions have not been identified yet because none of the monoclonal antibodies (mAbs) so far tested successfully blocked Vα14 NKT cell–mediated cytotoxicity (Fig.3E). In addition, the cytotoxicity was abrogated by treatment with concanamycin A (CMA) (Fig. 3F). Because CMA is known to be a specific inhibitor of vacuolar-type H+-dependent adenosine triphosphatase and it inhibits the activity of perforin (19), the effector function is likely to be mediated by NK-like mechanisms. Similar results were also observed with other tumors, including FBL-3 and LLC (Fig. 3, G and H).

Figure 3

Effector mechanisms of Vα14 NKT cells. Twenty-four hours after the intraperitoneal injection of IL-12 (2400 U/mouse) into the Jα281+/+ (A), Jα281–/– (B), or RAG–/– Vα14tgVβ8.2tg (C) mice, splenocytes were subjected to a 51Cr-release assay on B16 (A to F), FBL-3, and LLC (G and H) as described (25). Blocking experiments were carried out by using unlabeled (cold) targets and various mAbs (26). For antibody blocking, target or Vα14 NKT cells from RAG–/– Vα14tgVβ8.2tg mice were incubated with mAb to FcγRII/III (anti-FcγRII/III) (50 μg/ml) to avoid nonspecific killing. The 51Cr-labeled B16 and Vα14 NKT cells were incubated with the indicated numbers of unlabeled tumor cells as inhibitors as described (27) at the indicated ratios (D) and with mAbs (26) at the indicated concentrations (E). In some experiments, Vα14 NKT cells were incubated with CMA (Wako Pure Chemical Industries, Osaka) or control vehicle (dimethyl sulfoxide; DMSO) (F) at the indicated concentrations as described (19). IL-12–activated Vα14 NKT cells were also tested for their cytotoxicity on FBL-3 and LLC (G), whose activities were abrogated by cold targets (H) but not by mAbs (26). The assays were performed at an effector to target (E/T) ratio of 25:1 in the blocking experiments. The data are expressed as the mean of three cultures ± SD.

Interferon-γ (IFN-γ) is reported to be important in the IL-12–mediated cytotoxicity because it is essential for CTL generation and NK cell activation (20). The depletion of CD8+ CTLs and the blocking of IFN-γ function with mAb to IFN-γ reduce the efficacy of IL-12 (16, 17). Although Vα14 NKT cells are the major source of IFN-γ (14), IFN-γ may not be important in the effector phase. This is because high doses of mAbs to IFN-γ do not inhibit Vα14 NKT cell–mediated cytolysis (Fig. 3E).

We examined the potential activity of NK and T cells in vitro in the mutant mice; both NK-mediated and T cell–mediated killing functions were potent. The NK activity induced by polyinosinic-polycytidylic acid [poly(I:C)] was as potent in Jα281–/–mice as in wild-type mice (Fig. 4A). Similarly, significant allospecific CTL activity was detected on P815 (H-2d) and BALB/c (H-2d) concanavalin A (ConA) blasts, but not on EL-4 (H-2b), in Jα281–/– mice to the same extent as in wild-type mice (Fig. 4B). Thus, NK and conventional T cells in Jα281–/– mice are functionally active, yet not indispensable for tumor rejection upon IL-12 stimulation.

Figure 4

Potential cytotoxicity of NK and conventional T cells in Jα281–/– mice. (A) NK activity. Poly(I:C) (150 μg per mouse; Pharmacia) or PBS (control) was injected intraperitoneally into Jα281+/+and Jα281–/– mice. After 24 hours, spleen cells were tested on YAC-1 cells. (B) Allospecific CTL activity. On day 0, P815 mastocytoma (3 × 107; H-2d) or PBS was injected intraperitoneally (primed and unprimed, respectively) into Jα281+/+ and Jα281–/– mice. Spleen cells prepared on day 14 were assayed on P815, ConA-activated BALB/c splenocytes (H-2d), and EL-4 tumor cells (H-2b). The data are expressed as the mean of three cultures ± SD.

The primary effect of IL-12 on Vα14 NKT cells is also supported by the fact that IL-12 causes an increase in the actual numbers of Vα14 NKT cells (about a fourfold increase; 1.5 × 105 to 6 × 105) and in their cell volume (1.3 to 2.5-fold increase) (21). It is now clear that a reevaluation of NK and T cell functions in the absence of Vα14 NKT cells may alter our understanding of the functions of various subsets of lymphocytes in vivo.

  • * These authors contributed equally to this work.

  • To whom correspondence should be addressed: E-mail: taniguti{at}


Stay Connected to Science

Navigate This Article