Integrin αvβ8–expressing tumor cells evade host immunity by regulating TGF-β activation in immune cells
Naoki Takasaka, … , Jody L. Baron, Stephen L. Nishimura
Naoki Takasaka, … , Jody L. Baron, Stephen L. Nishimura
Published October 18, 2018
Citation Information: JCI Insight. 2018;3(20):e122591. https://doi.org/10.1172/jci.insight.122591.
View: Text | PDF
Research Article Immunology

Integrin αvβ8–expressing tumor cells evade host immunity by regulating TGF-β activation in immune cells

Abstract

TGF-β is a promising immunotherapeutic target. It is expressed ubiquitously in a latent form that must be activated to function. Determination of where and how latent TGF-β (L-TGF-β) is activated in the tumor microenvironment could facilitate cell- and mechanism-specific approaches to immunotherapeutically target TGF-β. Binding of L-TGF-β to integrin αvβ8 results in activation of TGF-β. We engineered and used αvβ8 antibodies optimized for blocking or detection, which — respectively — inhibit tumor growth in syngeneic tumor models or sensitively and specifically detect β8 in human tumors. Inhibition of αvβ8 potentiates cytotoxic T cell responses and recruitment of immune cells to tumor centers — effects that are independent of PD-1/PD-L1. β8 is expressed on the cell surface at high levels by tumor cells, not immune cells, while the reverse is true of L-TGF-β, suggesting that tumor cell αvβ8 serves as a platform for activating cell-surface L-TGF-β presented by immune cells. Transcriptome analysis of tumor-associated lymphoid cells reveals macrophages as a key cell type responsive to β8 inhibition with major increases in chemokine and tumor-eliminating genes. High β8 expression in tumor cells is seen in 20%–80% of various cancers, which rarely coincides with high PD-L1 expression. These data suggest tumor cell αvβ8 is a PD-1/PD-L1–independent immunotherapeutic target.

Authors

Naoki Takasaka, Robert I. Seed, Anthony Cormier, Andrew J. Bondesson, Jianlong Lou, Ahmed Elattma, Saburo Ito, Haruhiko Yanagisawa, Mitsuo Hashimoto, Royce Ma, Michelle D. Levine, Jean Publicover, Rashaun Potts, Jillian M. Jespersen, Melody G. Campbell, Fraser Conrad, James D. Marks, Yifan Cheng, Jody L. Baron, Stephen L. Nishimura

×

Figure 2

Expression of β8 increases in vivo tumor growth.

Options: View larger image (or click on image) Download as PowerPoint
Expression of β8 increases in vivo tumor growth. (A) LLC cells were tran...
(A) LLC cells were transfected with itgb8. Stable pools were established and sorted for uniform expression and C6D4 (filled histogram, 1 μg/ml), and stained cells were compared with mock LLC (unfilled histogram). (B) β8-expressing LLC cells were tested for their ability to support TGF-β activation in the presence of C6D4 (filled boxes) or isotype controls SV5 (open boxes) or B5 (closed circles), n = 4. (C) Tumor growth of s.c. injected β8 LLC cells compared with mock LLC cells. Shown is a representative experiment (n = 14–16, repeated 3 times). (D) Tumor weight from individual mice bearing mock or β8 LLC tumors harvested at day 14. Open boxes, β8 LLC; filled boxes, mock LLC. (E and F) Spider plots of tumor cell growth in individual mice followed until day 19 after injection with β8 LLC cells. Mice were treated with isotype control (E) or C6D4 (F). Arrows indicate treatments (7 mg/kg i.p.). n = at least 9/group from a representative experiment of 3. (G) Average tumor volumes and (H) weights from tumors harvested at day 19 in E and F. Open boxes, isotype control; filled boxes, C6D4. (I–N) Established β8 LLC tumors were treated with isotype control (I), C6D4 (J), anti–PD-1 (RMP1-14) (K), or C6D4 + RMP1-14 (L). Arrows indicate injection time points. (M) Average tumor volumes and (N) weights from tumors harvested at day 19 in I–L. n = 10/ group. (O) B16 melanoma cells, which normally do not express αvβ8, were stably transfected with itgb8, sorted for uniform expression, and stained with C6D4 (1 μg/ml) and compared with mock-transfected B16 cells. Shown is a representative experiment of 4. (P) β8 B16 cells were tested for their ability to support TGF-β activation in the presence of isotype (filled bar) or C6D4 (open bar) (10 μg/ml) or were compared with mock-transfected B16 cells. n = 3. (Q) β8 or mock-transfected B16 cells were injected (i.v.), and 14 days later, lungs were morphometrically assessed for metastatic burden. n = 9–10/group. Shown is a representative experiment repeated twice. (R–T) β8 B16 cells were injected (i.v.), and mice were treated with isotype control or C6D4 (7 mg/kg i.p.) at day 0 and 7 and assessed for metastasis at day 14. In Q, open boxes, β8 B16; filled boxes, mock-transfected B16. In R, open boxes, isotype; filled boxes, C6D4. Shown are representative micrographs of lungs taken at 40× magnification from isotype (S) or C6D4-treated mice (T). Scale bar: 200 μm. Significance was determined by unpaired Student’s t test or ANOVA for multiple comparisons followed by Tukey’s post-hoc test to find where the difference lay. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001