Immunogenicity of premalignant lesions is the primary cause of general cytotoxic T lymphocyte unresponsiveness

Abstract

Cancer is sporadic in nature, characterized by an initial clonal oncogenic event and usually a long latency. When and how it subverts the immune system is unknown. We show, in a model of sporadic immunogenic cancer, that tumor-specific tolerance closely coincides with the first tumor antigen recognition by B cells. During the subsequent latency period until tumors progress, the mice acquire general cytotoxic T lymphocyte (CTL) unresponsiveness, which is associated with high transforming growth factor (TGF) beta1 levels and expansion of immature myeloid cells (iMCs). In mice with large nonimmunogenic tumors, iMCs expand but TGF-beta1 serum levels are normal, and unrelated CTL responses are undiminished. We conclude that (a) tolerance to the tumor antigen occurs at the premalignant stage, (b) tumor latency is unlikely caused by CTL control, and (c) a persistent immunogenic tumor antigen causes general CTL unresponsiveness but tumor burden and iMCs per se do not.

Figure 1.

Sporadic immunogenic tumors induce CTL unresponsiveness to unrelated antigens. CTL activity against the minor histocompatibility antigens (HY) and the LCMV peptide epitope gp33 was analyzed in vivo. (A and B) For analysis of HY-specific CTLs, female and male spleen cells (10⁷ each) labeled with different amounts of CFSE were injected into the indicated mice, and 18 h later the ratio between both populations in the spleen was determined by flow cytometry. Immunization was performed by a single i.p. injection of 5 × 10⁶ male C57BL/6 spleen cells 2 wk before the assay. The percentage of specific killing of male cells is indicated. One representative example per experimental group is shown in A, and all data are shown in B (horizontal bars indicate mean values). Serum anti-Tag antibody titers (mg/ml) are given for tumor-bearing mice. Genotype (B6, C57BL/6; Tg, LoxP-Tag), age, cancer status (−, no tumor; LTB), treatment (N/−, naive; I/+, immunized), and gender of mice (f, female; m, male) are indicated. Experimental groups were young naive female C57BL/6 mice (B6; 2–3 mo old; n = 5), young immunized male C57BL/6 mice (B6; 2–3 mo old; n = 3), young immunized female LoxP-Tag mice (Tg; 2–3 mo old; n = 7), old immunized female C57BL/6 mice (B6; 15 mo old; n = 4), and LTB female LoxP-Tag mice (Tg; 7–19 mo old; n = 6). n.a., not analyzed. (C and D) For analysis of LCMV-gp33 peptide–specific CTLs, nonloaded and LCMV-gp33 peptide–loaded spleen cells (10⁷ each) labeled with different amounts of CFSE were injected into the indicated mice, and 18 h later the ratio between both populations was determined by flow cytometry. Mice were immunized by a single s.c. injection of 10⁷ cells of tumor line MC57-gp33-Hi 8–10 d before the assay. The percentage of specific killing of peptide loaded cells is indicated. Representative examples are shown in C, and all data are shown in D (horizontal bars indicate mean values). Genotype (B6, C57BL/6; Tg, LoxP-Tag), age, cancer status (−, no tumor; LTB), and treatment (N/−, naive; I/+, immunized) are indicated. Experimental groups were young naive C57BL/6 mice (B6; 2–3 mo old; n = 5), young immunized C57BL/6 mice (B6; 2–3 mo old; n = 6), and LTB LoxP-Tag mice (Tg; 19 mo old; n = 4).

Figure 2.

Tag-specific CTL cell tolerance occurs almost simultaneously with the development of Tag-specific antibodies. (A) To identify LoxP-Tag mice that had most recently immunologically recognized PMLs (open symbols; n = 6), anti-Tag IgG antibodies were measured as indicated in individual mice. As a control, mice that already had detectable anti-Tag antibodies for several months (LTB mice; closed symbols; n = 4) were used. (B) CTL activity against the Tag-specific peptide IV was analyzed in vivo in PML mice (6–14 mo old) and LTB mice (16–19 mo old) shown in A. Therefore, nonloaded and peptide IV–loaded splenocytes labeled with different amounts of CFSE were injected into the respective mice and as a control into young naive (2–3 mo old; n = 2) and immunized LoxP-Tag transgenic mice (2–3 mo old; n = 3), and aged C57BL/6 mice (B6; 6–19 mo old; n = 5). 18 h later the ratio between both populations was determined by flow cytometry of spleen cells. The percentage of specific killing of peptide-loaded cells is indicated (horizontal bars indicate mean values). Genotype (B6, C57BL/6; Tg, LoxP-Tag), age, cancer status (−, no tumor; PML; LTB), and treatment (−, naive; +, immunized) are indicated. (C) LoxP-Tag mice at ∼6–9 mo of age have developed Tag-specific tolerance. 10⁶ cells of Tag-expressing tumor line 16.113 were s.c. injected into 3-mo-old Rag-2^−/− mice (n = 4), as well as 3-mo-old (n = 4), 6-mo-old (n = 4), 9-mo-old (n = 4), and 13-mo-old (n = 3) LoxP-Tag mice, and tumor growth was observed. Error bars represent SD. The age and number of mice with tumor per number of mice in the experiment are shown in parenthesis.

Figure 3.

Tag-specific tolerance and general CTL unresponsiveness are induced consecutively. (A) HY-specific CTL activity 2 wk after immunization of female LoxP-Tag mice of different ages (7, 10–12, and 17 mo) with 5 × 10⁶ male spleen cells was determined by in vivo kill assay, as described in Fig. 1. Female and male spleen cells were injected into the indicated mice, and 18 h later the ratio between both populations in the spleen was determined by flow cytometry. The percentage of specific killing of male cells is indicated (horizontal bars indicate mean values), and numbers give the anti-Tag IgG antibody titers (mg/ml) in the serum of corresponding mice. Anti-Tag IgG antibody titers in the serum (mg/ml) are indicated if mice were positive. n.a., not analyzed. (B) 6–13-mo-old LoxP-Tag mice were first immunized with 5 × 10⁶ male spleen cells and 1 wk later challenged s.c. with 10⁶ Tag-expressing 16.113 cells. Mice in which challenge tumors were not rejected and that had a tumor size of 8–10 mm in diameter (open symbols; n = 8) were boosted with 5 × 10⁶ male spleen cells and subjected to an HY-specific in vivo kill assay 2 wk later. Identically treated 2-mo-old C57BL/6 mice (B6; n = 4) and LoxP-Tag mice (Tg; n = 2) were used as controls (closed symbols). The percentage of specific killing of male cells is indicated (horizontal bars indicate mean values). Anti-Tag IgG antibody titers in the serum (mg/ml) are indicated if mice were positive. n.d., not detectable. (C) 6- and 12-mo-old LoxP-Tag mice reject regressor tumor cells expressing gp33. 10⁶ cells of gp33-expressing tumor line MC57-gp33-Hi were s.c. injected into 3-mo-old Rag-2^−/− (n = 3), 6-mo-old LoxP-Tag (n = 3), and 12-mo-old LoxP-Tag (n = 3) mice, and tumor growth was observed. The age and number of mice with tumor per number of mice in the experiment is shown in parenthesis. (D) LoxP-Tag mice shown in C that rejected MC57-gp33-Hi regressor tumor cells were immunized with Tag⁺ 16.113 cells, and 9 d later were analyzed for CTL activity against the Tag-specific peptide IV in vivo. Therefore, nonloaded and peptide IV–loaded splenocytes labeled with different amounts of CFSE were injected into the respective mice and as a control into young naive and immunized C57BL/6 mice (2 mo old; n = 2, respectively). 18 h later, the ratio between both populations was determined by flow cytometry of spleen cells. The percentage of specific killing of peptide-loaded cells is indicated (horizontal bars indicate mean values).

Figure 4.

Primary but not memory CTL responses are suppressed in old LoxP-Tag mice. (A) HY-immunized mice shown in Fig. 3 (A and B) were analyzed for the presence of UTY-specific T cells by staining spleen cells with anti-CD8 mAbs and H-2D^b/UTY tetramers. One representative example of flow cytometric analysis for each experimental group is shown. The percentage of UTY-specific tetramer-positive CD8⁺ T cells is indicated. (B) The percentage of HY-specific killing is plotted against the percentage of UTY-specific tetramer-positive CD8⁺ T cells from experiments shown in Fig. 3 (A and B). (C) Memory CTL responses against unrelated antigens are not generally inhibited in aged LoxP-Tag mice. 2-mo-old female LoxP-Tag and C57BL/6 mice were immunized three times with 5 × 10⁶ male spleen cells (prime). At the age of 18–19 mo, these mice were reimmunized with 5 × 10⁶ male spleen cells (boost), and HY-specific CTL activity was determined 2 wk later by in vivo kill assay, as described in Fig. 1. CFSE-labeled female and male spleen cells were injected into the indicated mice, and as a control into 2-mo-old naive C57BL/6 (B6; n = 2), immunized C57BL/6 mice (B6; n = 2), and immunized 18-mo-old LoxP-Tag mice (Tg; n = 2). 18 h later, the ratio between both populations was determined by flow cytometry. The percentage of specific killing of male cells is indicated (horizontal bars indicate mean values). Genotype (B6, C57BL/6; Tg, LoxP-Tag), age, cancer status (−, tumor-free; PML), and treatment (−, naive; +/−, immunized once at the indicated age; +/+, primed at 2 mo/boosted at the indicated age) are indicated.

Figure 5.

Detection of tumors and PMLs in LoxP-Tag mice. (A) Tumors in LTB mice were detected by MRI. Shown are representative transversal and coronal MR images of a kidney tumor (#1) and a tumor in the bone (#2); tumors are indicated by white circles. (B) Immunohistology of tissue sections in LTB mice. Tissues were stained for Tag, Ki-67, CD3 (red), FoxP3 (brown), and F4/80 (red) expression and counterstained with hematoxylin. Bar, 200 μm. (C) Representative PMLs in mice that recently developed anti-Tag antibodies but without macroscopically detectable tumors. Kidneys of 6–7-mo-old LoxP-Tag mice, which were shown to have developed tumor-induced tolerance, were cross sectioned in 4-μm steps and examined for the presence of PMLs by staining with hematoxylin and eosin (not depicted). Consecutive sections of identified PMLs were stained with mAbs as in B. In four out of seven mice analyzed, PMLs were detected in the kidney. Lymph nodes obtained from tumor-free mice served as a positive control for FoxP3 staining (Fig. S2, available at http://www.jem.org/cgi/content/full/jem.20072016/DC1). Staining of additional PMLs is shown in Fig. S1. Bar, 100 μm.

Figure 6.

Tumors in LoxP-Tag mice induce an increase in iMCs and serum TGF-β1 levels. (A and B) For analysis of iMCs, single-cell preparations of spleen and tumor were double stained with antibodies against Gr-1 and CD11b or Gr-1 and c-kit, respectively, to assess CD11b⁺/Gr-1⁺ (top) and c-kit⁺/Gr-1⁺ (bottom) cells in spleens of 2–3-mo-old LoxP-Tag mice (young Tg), 14–17-mo-old wild-type mice (old B6), 6–13-mo-old tumor-bearing LoxP-Tag mice with PMLs, 7–19-mo-old LTB LoxP-Tag mice, and 3–4-mo-old LoxP-Tag × Alb-Cre double-transgenic mice (TB Tg × Alb-Cre) with large liver tumors. Numbers show the percentage of double-positive cells of nonlymphocytes. One representative example per experimental group is shown in A, and all data for splenic CD11b⁺/Gr-1⁺ and c-kit ⁺/Gr-1⁺ cells (•) and tumor-infiltrating CD11b⁺/Gr-1⁺ cells (▴) are shown in B (horizontal bars indicate mean values). (C) Serum obtained from individual young LoxP-Tag (Tg; 3 mo old; n = 5), old C57BL/6 mice (B6; 12–23 mo old; n = 5), LoxP-Tag transgenic mice with PMLs (PML Tg; 6–13 mo old; n = 4), LTB LoxP-Tag transgenic mice (LTB Tg; 7–19 mo old; n = 4), and LTB LoxP-Tag × Alb-Cre double-transgenic mice (LTB Tg × Alb-Cre; 3–4 mo old; n = 4) was analyzed for TGF-β1 by ELISA. Error bars represent SD. (D) Perihilar lymph nodes of kidneys obtained from young tumor-free LoxP-Tag mice and mice with PMLs were double stained for TGF-β1 (red) and IgG (green); nuclei were counterstained with DAPI. Cells that are double positive for IgG antibodies and TGF-β1 are shown in orange (merge). Additional stainings for PML and LTB mice are shown in Fig. S3 (available at http://www.jem.org/cgi/content/full/jem.20072016/DC1). Bar, 20 μm.

Figure 7.

Large nonimmunogenic tumors in LoxP-Tag × Alb-Cre double-transgenic mice do not induce general CTL unresponsiveness. (A) LoxP-Tag × Alb-Cre double-transgenic mice are tolerant for Tag. LoxP-Tag × Alb-Cre double-transgenic mice (1–2 mo old; n = 3) were immunized s.c. with 10⁶ Tag⁺ 16.113 cells and subjected to a Tag-specific peptide IV in vivo kill assay 2 wk later. As a control, naive mice (2 mo old; n = 2) and immunized wild-type mice (2 mo old; n = 2) were used. The percentage of specific killing of Tag-specific peptide IV–loaded spleen cells is shown (horizontal bars indicate mean values). (B) LoxP-Tag × Alb-Cre double-transgenic mice (Tg × Alb-Cre) develop large tumors in the liver at 3–4 mo of age. MRI analysis revealed livers to be four to five times larger in size in comparison to livers from 3–4-mo-old LoxP-Tag mice (Tg). (C) Immunohistochemical analysis using Tag-specific antibody shows strong Tag expression in liver tumors that consist of hepatocellular and cholangiocellular cell carcinoma (HCC and CCC, respectively). Bar, 200 μm. (D) LoxP-Tag × Alb-Cre mice with large liver tumors at 3–4 mo of age (n = 3) were immunized with 5 × 10⁶ male spleen cells and subjected to an HY-specific in vivo kill assay 2 wk later, as described in Fig. 1. Naive and HY-immunized C57BL/6 (B6) mice served as controls. The percentage of specific killing of male cells is shown (horizontal bars indicate mean values).