B regulatory cells and the tumor-promoting actions of TNF-α during squamous carcinogenesis

Abstract

The inflammatory cytokine TNF-α has been recognized as a critical tumor promoter, but the effector cells that mediate its action have not been fully characterized. Because B cells regulate squamous and prostate carcinogenesis, and Tnf(-/-) mice harbor B-cell defects, we investigated the hypothesis that B cells are important effector cells for TNF-α-mediated promotion of cancer development. Using an adoptive transfer strategy and the 7,12-dimethylbenz[α]anthracene/terephthalic acid (DMBA/TPA) two-stage model of skin carcinogenesis, we found that both B cells and TNF-α are critical for the development of DMBA/TPA-induced papilloma. Transfer of B cells from DMBA/TPA-treated wild-type mice to Tnf(-/-) mice rescued papilloma development to a wild-type level, a result not observed when B cells from Tnf(-/-) mice were transferred to Rag2(-/-) mice or when TNF-α was eliminated selectively in B cells. Resistance to papilloma development in Tnf(-/-) mice was associated with increased IFN-γ and CD8(+) T cells in skin and a significant reduction in IL-10-producing B regulatory cells alongside an increase in IFN-γ-producing CD8(+) T cells in the spleen. These data indicate that during DMBA/TPA-induced squamous carcinogenesis TNF-α mediates tumor-promoting activity via regulatory B cells that repress antitumor immunity.

Fig. 1.

B cells are involved in DMBA/TPA squamous carcinogenesis. (A) Rag2^−/− mice are resistant to DMBA/TPA. ●, C57/Bl6 wild-type mice (n = 24); ■, C57/Bl6 Rag2^−/− mice (n = 23); P < 0.0001. This graph shows combined data from two independent experiments. (B) Mice deficient for B cells (Jh mice) are partially resistant to DMBA/TPA skin carcinogenesis. ●, BALB/c wild-type mice (n = 12); ▼, BALB/c Jh mice (n = 12); P < 0.05. (C) Wild-type B cells partially rescue the resistance to carcinogenesis of Rag2^−/− mice. ●, wild-type mice (n = 24); ■, Rag2^−/− mice (n = 16); ▲, mice with Rag2^−/− plus wild-type B cells (n = 21). Wild-type mice versus Rag2^−/− mice, P < 0.001; mice with Rag2^−/− plus wild-type B cells versus Rag2^−/− mice, P < 0.05; mice with Rag2^−/− plus wild-type B cells versus wild-type mice = no significant difference. This graph shows combined data from two independent experiments. (D) Wild-type B cells partially rescue the Tnf^−/−–resistant phenotype. ●, C57/Bl6 wild-type mice (n = 36); □, C57/Bl6 Tnf^−/− mice (n = 2); ▽, C57/Bl6 mice with Tnf^−/− plus wild-type B cells (n = 36). P < 0.05. Wild-type mice versus Tnf^−/− mice, P < 0.001; mice with Tnf ^−/− plus wild-type B cells versus Tnf^−/− mice, P < 0.05; mice with Tnf^−/− plus wild-type B cells versus wild-type mice = no significant difference. This graph shows combined data from three independent experiments. (E) Tnf^−/− B cells are not able to rescue the resistant phenotype of Rag2^−/− mice. ●, C57/Bl6 wild-type mice (n = 12); ♢, mice with Rag2^−/− plus Tnf^−/−B cells (n = 11); ■, C57/Bl6 Rag2^−/− mice (n = 10); Wild-type mice versus Rag2^−/− mice, P < 0.001; mice with Rag2^−/− plus Tnf^−/− B cells versus wild-type mice, P < 0.001; mice with Rag2^−/− plus Tnf^−/− B cells versus Rag2^−/− mice = no significant difference. (F) CD19-Cre/Tnf^f/f mice are partially resistant to DMBA/TPA. ●, C57/Bl6 wild-type mice (n = 12); △, C57/Bl6 CD19^Cre/Tnf^f/f mice (n = 12); difference between the groups, P < 0.002. Error bars indicate SE.

Fig. 2.

Characterization of the inflammatory status of wild-type and Tnf^−/− skin. (A) CD8⁺ cells are more numerous in Tnf^−/− mice than in wild-type mice (Left, wild-type, n = 4; Tnf^−/−, n = 4) and are reduced after transfer of wild-type B cells (Right, Tnf^−/− n = 3; Tnf^−/− plus wild-type B cells, n = 3). *P < 0.05. (B) F4/80⁺ cells are reduced in Tnf^−/− mice compared with wild-type mice (*P < 0.05) (Left, wild-type, n = 4; Tnf^−/−, n = 4) and are higher after transfer of wild-type B cells (Right, Tnf^−/−, n = 3; Tnf^−/− plus wild-type B cells, n = 3). (C) CCL5 mRNA expression. Results are from pooled mRNA from several mice for each group: wild-type and Tnf^−/−, n = 4, Tnf^−/−, n = 3; and Tnf^−/− plus wild-type B cells, n = 3. *P < 0.05, **P < 0.001. (D) IFN-γ mRNA expression. The results are from pooled mRNA from several mice for each group: wild-type and Tnf^−/−, n = 4; Tnf^−/−, n = 3; and Tnf^−/− plus wild-type B cells, n = 3. *P < 0.05. Error bars indicate SE.

Fig. 3.

Study of the activation of wild-type and Tnf^−/− B cells. (A) C1q/IgG immune complexes in serum of wild-type mice (n = 4), Tnf^−/− mice (n = 4), and in mice with Tnf^−/− and wild-type B cells (n = 3). Immune complexes are reduced in Tnf^−/− mice and are slightly increased after transfer of wild-type B cells but only at the 7-wk time point. *P < 0.05; **P < 0.001). (B) IL-10 production by lymphocytes from wild-type (n = 6) and Tnf^−/− (n = 6) inguinal lymph nodes. Mice were treated with DMBA and then TPA for 4 wk. Lymphocytes were kept in culture for 24 h, and cytokine expression was measured by electrochemiluminescence. *P < 0.05. (C) IL-10 and IL-12 total production by wild-type (n = 6) and Tnf^−/− (n = 6) B cells after 48 h incubation with LPS (10 μg/mL) and goat F(ab)2 anti-mouse IgM (10 μg/mL). *P < 0.05. Error bars indicate SE.

Fig. 4.

Characterization of B-cell subpopulations in wild-type and Tnf^−/− mice. (A) B cells (10⁶) from untreated and DMBA/TPA-treated (5 wk) spleens of wild-type and Tnf^−/− mice (n = 6) positive for IL-10. *P < 0.05. (B) IL-10 production by B cells from untreated and DMBA/TPA-treated (5 wk) spleens of wild-type and Tnf^−/− mice (n = 6) after 24 and 48 h in culture. *P < 0.05; **P < 0.001. ND, not detectable. (C) Percentage (Left) and absolute numbers (Right) of CD19⁺veCD21^hi cells from untreated and DMBA/TPA-treated (5 wk) spleens of wild-type and Tnf^−/− mice (n = 6). *P < 0.05. (D) Absolute numbers of CD3⁺ T cells expressing IFN-γ from untreated and DMBA/TPA-treated (5 wk) spleens of wild-type and Tnf^−/− mice (n = 6). *P < 0.05. (E) Absolute numbers of CD3⁺ T cells expressing IFN-γ from untreated and DMBA/TPA-treated (5 wk) inguinal lymph nodes of wild-type and Tnf^−/− mice (n = 6). *P < 0.05. (F) IL-10 production by B cells from untreated spleens of wild-type mice (n = 6) after 6 h in culture with medium only (Control), LPS (10 μg/mL), or LPS (10 μg/mL) plus anti–TNF-α (10 μg/mL). *P < 0.05. Error bars indicate SE.