Targeting HIF-1α abrogates PD-L1–mediated immune evasion in tumor microenvironment but promotes tolerance in normal tissues
Christopher M. Bailey, … , Yang Liu, Yin Wang
Christopher M. Bailey, … , Yang Liu, Yin Wang
Published March 3, 2022
Citation Information: J Clin Invest. 2022;132(9):e150846. https://doi.org/10.1172/JCI150846.
View: Text | PDF
Research Article Therapeutics

Targeting HIF-1α abrogates PD-L1–mediated immune evasion in tumor microenvironment but promotes tolerance in normal tissues

Abstract

A combination of anti–CTLA-4 plus anti–PD-1/PD-L1 is the most effective cancer immunotherapy but causes high incidence of immune-related adverse events (irAEs). Here we report that targeting of HIF-1α suppressed PD-L1 expression on tumor cells and tumor-infiltrating myeloid cells, but unexpectedly induced PD-L1 in normal tissues by an IFN-γ–dependent mechanism. Targeting the HIF-1α/PD-L1 axis in tumor cells reactivated tumor-infiltrating lymphocytes and caused tumor rejection. The HIF-1α inhibitor echinomycin potentiated the cancer immunotherapeutic effects of anti–CTLA-4 therapy, with efficacy comparable to that of anti–CTLA-4 plus anti–PD-1 antibodies. However, while anti–PD-1 exacerbated irAEs triggered by ipilimumab, echinomycin protected mice against irAEs by increasing PD-L1 levels in normal tissues. Our data suggest that targeting HIF-1α fortifies the immune tolerance function of the PD-1/PD-L1 checkpoint in normal tissues but abrogates its immune evasion function in the tumor microenvironment to achieve safer and more effective immunotherapy.

Authors

Christopher M. Bailey, Yan Liu, Mingyue Liu, Xuexiang Du, Martin Devenport, Pan Zheng, Yang Liu, Yin Wang

×

Figure 8

Echinomycin induces PD-L1 to counter ipilimumab-induced GI irAEs by an IFN-γ–dependent mechanism.

Options: View larger image (or click on image) Download as PowerPoint
Echinomycin induces PD-L1 to counter ipilimumab-induced GI irAEs by an I...
(A) Experimental design. Ipilimumab (Ipi.) was used to induce GI irAEs in CTLA4h/h pups (detailed in Methods). Single dose of ipilimumab and other agents are indicated in the diagram by arrows (red, mAbs; blue, vehicle/LEM). ICIs, immune checkpoint inhibitors (mAbs). (B and C) Representative immunofluorescence images showing T cell infiltration (B) or PD-L1 expression (C) in jejunum of vehicle- or ipilimumab-treated mice. (D) Association of intestinal PD-L1 expression with GI irAEs determined by FITC-dextran assay. Intestinal PD-L1 expression was scored as negative/low (n = 18) or high (n = 21) based on immunofluorescence, and serum FITC-dextran intensity is presented as the mean ± SEM for each group, analyzed by Student’s t test. Aggregate data shown from 4 experiments. (E–G) Effects of LEM, anti–PD-1 (RMP1-14), and anti–IFN-γ (XMG1.2) in the GI irAE model. Mice were grouped as follows to receive therapies based on the experimental design depicted in A: vehicle (n = 33), LEM (n = 27), ipilimumab (n = 33), ipilimumab plus LEM (n = 26), ipilimumab plus RMP1-14 (n = 27), ipilimumab plus LEM plus RMP1-14 (n = 28), and ipilimumab plus LEM plus XMG1.2 (n = 15). (E) Serum FITC-dextran intensity shown as mean ± SEM for individual mice pooled from 3 independent experiments. GI irAE incidence corresponding to each group is annotated (percentages); dotted line represents the threshold for the determination of GI irAEs. Statistics were determined by 2-tailed, unpaired Student’s t test. (F) Representative H&E images from intestines of mice receiving different therapies. Panel iv shows cellular debris and necrosis in lamina propria and epithelium (arrow). (G) Representative immunofluorescence images showing PD-L1 staining in jejunum of mice from different treatment groups. Scale bars: 50 μm (B, C, and G). (H) Flow cytometry analysis of PD-L1 expression in intestinal epithelial cells (gated on live CD45–cytokeratin+ singlets) from mice treated with ipilimumab (n = 6) or ipilimumab plus LEM (n = 8). Data shown as mean ± SEM of the PD-L1 MFI for each mouse, and were analyzed by 2-tailed, unpaired Student’s t test. *P < 0.05; **P < 0.01; ***P < 0.001.