Inhibition of xCT suppresses the efficacy of anti-PD-1/L1 melanoma treatment through exosomal PD-L1-induced macrophage M2 polarization

Abstract

Tumor cells increase glutamate release through the cystine/glutamate transporter cystine-glutamate exchange (xCT) to balance oxidative homeostasis in tumor cells and promote tumor progression. Although clinical studies have shown the potential of targeting programmed cell death 1 (PD-1)/programmed death ligand 1 (PD-L1) signaling in melanoma, response rates are low. However, it remains unclear how glutamate metabolism affects anti-PD-1/PD-L1 treatment efficacy in melanoma. Here, we demonstrated that although inhibition of xCT either by pharmacological inhibitor (sulfasalazine [SAS]), approved by US Food and Drug Administration (FDA) for inflammatory diseases, or genetic knockdown induced reactive oxygen species (ROS)-related death in melanoma cells, inhibition of xCT significantly reduced the efficacy of anti-PD-1/PD-L1 immune checkpoint blockade through upregulating PD-L1 expression via the transcription factors IRF4/EGR1, as a consequence, exosomes carrying relatively large amounts of PD-L1 secreted from melanoma cells resulted in M2 macrophage polarization and reduced the efficacy of anti-PD-1/PD-L1 therapy in melanoma. Taken together, our results reveal that inhibition of xCT by SAS is a promising therapeutic strategy for melanoma; on the other hand, SAS treatment blunted the efficacy of anti-PD-1/PD-L1 via exosomal PD-L1-induced macrophage M2 polarization and eventually induced anti-PD-1/PD-L1 therapy resistance.

Trial registration: ClinicalTrials.gov NCT04205357.

Keywords: PD-1/PD-L1; exosome; macrophages; melanoma; xCT.

Conflict of interest statement

Declaration of interests The authors declare no competing interests.

Copyright © 2021 The American Society of Gene and Cell Therapy. Published by Elsevier Inc. All rights reserved.

Figures

Graphical abstract

Figure 1

Significant elevation of glutamate levels in melanoma patient plasma and the effect of SAS on melanoma (A) The ROC curve for glutamate (left panel) is shown, and glutamate levels in melanoma patient plasma (n = 20) and normal plasma (n = 20) were measured by metabolomics (right panel). Data are presented as the mean ± SD. (B) Glutamate levels in melanoma patient plasma (n = 15) and normal plasma (n = 20) were measured by ELISA. Data are presented as the mean ± SD. (C) B16F10 cells were treated with 1 mM SAS for 24−48 h, and Sk-Mel-5/Sk-Mel-28 cells were treated with 500 μm SAS for 24−48 h. The cell cycle distribution was detected by flow cytometry. Data are presented as the mean (n = 3) ± SD. (D) B16F10 cells were treated with 1 mM SAS for 24−48 h, and Sk-Mel-5/Sk-Mel-28 cells were treated with 500 μm SAS for 24−48 h. The extent of apoptosis was determined by flow cytometry. Data are presented as the mean (n = 3) ± SD. (E) Tumor growth curves for B16F10 tumors from C57BL/6 mice receiving the indicated treatments were constructed (left panel). Tumor growth curves for B16F10 tumors from BALB/c nude mice receiving the indicated treatments were constructed (right panel). Data are presented as the mean (n = 5) ± SD. (F) The body weight of C57BL/6 (left panel) and BALB/c nude (right panel) mice receiving the indicated treatments is shown. Data are presented as the mean (n = 5) ± SD. Asterisks indicate significant differences: ∗p 

Figure 2

Inhibition of xCT led to anti-PD-1 therapy resistance in B16F10 tumor-bearing C57BL/6 mice (A) Schematic of the treatment plan. (B) Photographs of tumor samples isolated from C57BL/6 mice receiving the indicated treatments. (C) Tumor growth curves for tumors receiving the indicated treatments. (D) The proportions of CD3+CD8+ T cells determined by fluorescence-activated cell sorting (FACS) after the indicated treatments. (E) The proportions of total CD3+CD4+ and CD3+CD8+ T cells determined by FACS after the indicated treatments. (F) The ratio of CD3+CD8+/CD3+CD4+ T cells determined by FACS after the indicated treatments. (G) The proportions of CD3+CD8+IFN-γ+ T cells determined by FACS after the indicated treatments. (H) The proportions of CD3+CD8+granzyme B+ T cells determined by FACS after the indicated treatments. (I) The proportions of F4/80+CD11B+MHC class II+ macrophages determined by FACS after the indicated treatments. (J) The proportions of F4/80+CD11B+CD206+ macrophages determined by FACS after the indicated treatments. Data from multiple experiments are expressed as the mean (n = 5) ± SD. Asterisks indicate significant differences: ∗p < 0.05.

Figure 3

Inhibition of xCT upregulated the expression of PD-L1 in melanoma cells (A) Cell lysates and mRNA were extracted from SAS-treated melanoma cells as indicated, and immunoblotting (upper panels) and quantitative real-time PCR (lower panels) were then performed. Data are presented as the mean ± SD. (B) Cell lysates and mRNA were extracted from lentivirus-infected melanoma cells as indicated, and immunoblotting (upper panels) and quantitative real-time PCR (lower panels) were then performed. Data from multiple experiments are expressed as the mean (n = 3) ± SD. Asterisks indicate significant differences: ∗p 

Figure 4

Overexpression of IRF4/EGR1 upregulated the expression of PD-L1 in melanoma cells (A−F) mRNA and cell lysates were extracted from EGR1/IRF4-transfected melanoma cells as indicated, and quantitative real-time PCR (A−D) and immunoblotting (E and F) were then performed. Data from multiple experiments are expressed as the mean (n = 3) ± SD. Asterisks indicate significant differences: ∗p 

Figure 5

The xCT-IRF4/EGR1 axis is required for PD-L1 regulation (A) IRF4 (left panel) and EGR1 (right panel) binding sites predicted on website. (B) PD-L1 transcriptional activity in IRF4 (left panel)- or EGR1 (right panel)-transfected 293T cells detected by dual luciferase reporter assays. (C) Designed primers for ChIP experiments. (D) ChIP-qPCR assays for IRF4 (left panel) or EGR1 (right panel) binding to the PD-L1 promoter sites in Sk-Mel-5 cells with or without SAS treatment. Data from multiple experiments are expressed as the mean (n = 3) ± SD. Asterisks indicate significant differences: ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001.

Figure 6

SAS treatment of melanoma leads to M2 macrophage polarization (A) A schematic of the conditioned medium collection protocol is shown. (B and C) mRNA was extracted from conditioned medium-treated macrophages as indicated, and quantitative real-time PCR was then performed. (D) A schematic of the coculture plan is shown. (E) SAS-treated melanoma cells were cocultured with macrophages as indicated, and the proportions of F4/80+CD11B+CD206+ macrophages were analyzed by FACS. (F) Cell lysates were extracted from conditioned medium-treated macrophages as indicated, and immunoblotting was then performed. Data from multiple experiments are expressed as the mean (n = 3) ± SD. Asterisks indicate significant differences: ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001.

Figure 7

Exosomal PD-L1 secreted by SAS-treated melanoma cells leads to M2 macrophage polarization (A) Schematic of exosome isolation and transmission electron microscopy (TEM)/ELISA analysis. (B) A representative TEM image of purified exosomes from SAS-treated melanoma cell conditioned medium. (C) PD-L1 concentration on the surface of exosomes in SAS-treated melanoma cell conditioned medium measured by ELISA. Data are presented as the mean (n = 3) ± SD. (D) PD-L1 concentration on the surface of exosomes in SAS-treated mouse serum measured by ELISA. Data are presented as the mean (n = 3) ± SD. (E and F) mRNA was extracted from exosome-treated macrophages as indicated, and quantitative real-time PCR was then performed. Data are presented as the mean (n = 3) ± SD. (G) Cell lysates were extracted from exosome-treated macrophages as indicated, and immunoblotting was then performed. (H) Tumor growth curves of mice receiving the indicated treatments. Data are presented as the mean (n = 5) ± SD. (I) The proportions of F4/80+CD11B+CD206+ macrophages determined by FACS after the indicated treatments. Data are presented as the mean (n = 5) ± SD. Asterisks indicate significant differences: ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001.