NKG2D-independent suppression of T cell proliferation by H60 and MICA

Abstract

The activating receptor NKG2D recognizes a wide range of different ligands, some of which are primarily expressed in "stressed" tissues or on tumor cells. Until now, similar stimulatory effects on natural killer and CD8+ T cells have been described for all NKG2D ligands, and the NKG2D receptor/ligand system has therefore been interpreted as a sensor system involved in tumor immune surveillance and activation of immune responses. We show here that the NKG2D ligands H60 and MIC class 1 chain-related protein A (MICA) can also mediate strong suppressive effects on T cell proliferation. Responsiveness to H60- and MICA-mediated suppression requires IL-10 and involves a receptor other than NKG2D. These findings might provide explanations for the observation that strong in vivo NKG2D ligand expression, such as that on tumor cells, sometimes fails to support effective immune responses and links this observation to a distinct subgroup of NKG2D ligands.

Fig. 1.

Generation of soluble NKG2D ligands. (A) Schematic structure of NKG2D ligands from N to C terminus showing leader sequence (LS), extracellular domain (ED), transmembrane region (TM), and cytosolic domain (CD) or glycosylphosphatidylinositol-anchor (GPI). cDNAs were mutagenized at the positions indicated in single-letter amino acid code, and a biotinylation sequence (BS) was added. (B) Expression of recombinant proteins after transformation of expression vectors into BL21(DE3) hosts. Samples from each culture before (0) and 3 h after (3) isopropyl β-d-thiogalactoside induction were run over a 10% SDS/PAGE gel. Murine (C) and human (D) antigen-specific CD8⁺ T cell lines were stained with multimers of NKG2D ligands, as indicated in the legends. Gray-filled histograms show background control staining using streptavidin PE (SA-PE), and dotted gray lines represent control staining with the PE-conjugated isotype-specific mAb used to visualize anti-NKG2D (mAb C7) staining on murine cells.

Fig. 2.

Influence on different NKG2D ligands on effector function. SIINFEKL-specific T cells were incubated in the presence of target cells (EL-4) and different SIINFEKL peptide concentrations (x axis) with or without addition of 10 μg/ml soluble Rae1 (filled circles) or H60 (open circles) multimers. Specific lysis was determined in a standard chromium release assay. Filled triangles indicate control values in the absence of any soluble NKG2D ligand. Data are representative of at least three independent experiments. Assays were done in triplicate, and standard deviations are indicated.

Fig. 3.

H60-specific inhibition affects CD8⁺ as well as CD4⁺ T cells. (A and B) CFSE-labeled splenocytes (BALB/c) were analyzed 65 h after stimulation with anti-CD3 in the presence or absence of different soluble NKG2D ligand multimers. For analysis, cells were gated on CD8⁺ (A) or CD4⁺ (B) T cells. Fluorescence profiles represent cell divisions in the presence of Rae1 (bold line), MULT1 (dashes; sometimes hidden behind the bold line), H60 (filled gray), or no additional ligand as a control; the dotted line indicates the CFSE profile in the absence of any stimulation. (C and D) Highly purified (>95%) CD8⁺ (C) and CD4⁺ (D) T cells were stimulated with plate-bound anti-CD3 under the same experimental setting as in A and B. (E) Expression of NKG2D-S and -L splice variants was examined in highly purified NK cells and CD8⁺ and CD4⁺ T cells by RT-PCR. (F) Splenocytes from C57BL/6 mice were CSFE-labeled and stimulated with anti-CD3 in the presence of irradiated RMA cells overexpressing H60 (gray) or Rae1(black line); mock-transfected cells (RMA-pIG) served as negative controls (dashed line).

Fig. 4.

H60-mediated T cell inhibition is independent of naturally arising Tregs and NKG2D. CFSE proliferation profiles of CD8⁺ T cells upon anti-CD3 (A-C, E, and F) or peptide stimulation (D) in the presence or absence of soluble NKG2D ligands were determined as described for Fig. 3 (same symbols used for ligands and controls). Splenocytes were derived from different gene-deficient (A-C and E)or transgenic mice (D), as indicated near the top of each histogram. L9.6/RAG^-/- cells (D) were stimulated in the presence of 1 μM p60_217-225 peptide. (F) Summary of data obtained for splenocytes derived from wild-type BALB/c mice stimulated in the presence of high concentrations of blocking NKG2D mAb (clone C7).

Fig. 5.

IL-10 is required for H60-mediated inhibition. (A Left) Freshly isolated CFSE-labeled human PBMC were stimulated with anti-CD3 (OKT3) in the presence or absence of soluble MICA (allele 8; bold line) or murine NKG2D ligands (H60, filled gray; Rae1, thin dark line), as described in Fig. 4 legend. As controls, cells stimulated only with anti-CD3 (thin gray line) or maintained in the absence of any stimulation (dotted line) are shown. (Right) Murine (BALB/c) splenocytes were used as responder cells under the same experimental settings. (B) Splenocytes derived from IL10^-/- (Left) or wild-type (Right) mice were stimulated in the presence or absence of H60; neutralizing anti-IL10 antibody was added to wild-type cultures, as indicated. (C Upper) Splenocytes from wild-type mice were analyzed after short-term in vitro stimulation with anti-CD3. Dot plots of live lymphocytes are shown, with staining for CD8 (x axis) and NKG2D ligand/streptavidin-APC multimers (y axis, as indicated). As a control, cells were stained with streptavidin-APC and CD8. (Lower) Under the same experimental setting as in A, splenocytes were labeled with CFSE to monitor their proliferation profile. The dot plots (gated on live lymphocytes) show the CFSE profile (x axis) of cells stained with MICA multimers (y axis, Left), H60 (Center), or streptavidin-APC as control (Right). (D) Splenocytes derived from IL-10^-/- mice were treated as described under C, stainings for CD8 vs. MICA-8 (Upper) or streptavidin-APC control (Lower) are shown.

Fig. 6.

Inhibition by native H60 expression and in vivo effects of NKG2D ligand application on tumor growth. Cohorts of mice were s.c. inoculated with aggressively growing tumor cells and subsequently treated with a three-step therapy to target H60 or Rae1 in vivo to the tumor cells (for details see Materials and Methods). The volumes of growing tumors were monitored over time.