Identification of VISTA regulators in macrophages mediating cancer cell survival

Abstract

Numerous human cancers have exhibited the ability to elude immune checkpoint blockade (ICB) therapies. This type of resistance can be mediated by immune-suppressive macrophages that limit antitumor immunity in the tumor microenvironment (TME). Here, we elucidate a strategy to shift macrophages into a proinflammatory state that down-regulates V domain immunoglobulin suppressor of T cell activation (VISTA) via inhibiting AhR and IRAK1. We used a high-throughput microfluidic platform combined with a genome-wide CRISPR knockout screen to identify regulators of VISTA levels. Functional characterization showed that the knockdown of these hits diminished VISTA surface levels on macrophages and sustained an antitumor phenotype. Furthermore, targeting of both AhR and IRAK1 in mouse models overcame resistance to ICB treatment. Tumor immunophenotyping indicated that infiltration of cytotoxic CD8⁺ cells, natural killer cells, and antitumor macrophages was substantially increased in treated mice. Collectively, AhR and IRAK1 are implicated as regulators of VISTA that coordinate a multifaceted barrier to antitumor immune responses.

Fig. 1.. A genome-wide CRISPR-Cas9 KO screen identifying VISTA regulators.

(A and B) Transduction of U937 cells by the TKOV3 library and puromycin selection. (C) Antibody surface labeling of VISTA followed by nanomagnetic beads binding before sorting the cells. (D) Sorting of the labeled pool of transduced cells to segregate different VISTA-expressing populations into low, medium, and high. (E) The isolation of the VISTA^low population DNA for next-generation sequencing (NGS), further validation, and characterization of druggable hits. FC, fold change.

Fig. 2.. Validation of the top hits identified from the screen.

(A) A plot showing the hits retrieved from sequencing the VISTA^low population sorted from a pool of transduced cells. Hits were identified using the DrugZ algorithms by identifying sgRNAs enriched in the VISTA^low population. (B) sgRNA distribution of the hits retrieved and carried forward in the study. Every gene is targeted by four sgRNAs. (C) GSEA of the top hits from the screen with an FDR below 20%. MyD88, myeloid differentiation primary response protein 88. (D) A bar plot showing the MFI after genetic perturbations of the selected genes using two independent sgRNA in U937 cells. Data are mean ± SD of n = 3 replicates. NTC, nontargeting control. (E) Representative histograms of the KO effects on VISTA surface levels in U937 cells. (F) Overview of the primary cells’ validation workflow in both human and murine samples. (G) Validation of the hits in human macrophages differentiated from monocytes isolated from four different healthy donors. Each donor is represented by the letter D. Data are mean ± SD of n = 3 replicates. Validation was carried out by using an independent sgRNA that has shown high efficiency in knocking out the respective gene. MFI is the median fluorescence intensity of the VISTA surface expression after knocking out the respective gene. (H) A plot showing the validation of the selected hits in bone marrow–derived monocytes (BMDMs) from WT Balb/C mice after differentiation into macrophages. MFI is the median fluorescence intensity of the VISTA surface expression after knocking out the respective gene (four mice). Data are mean ± SD of n = 3 replicates. ***P < 0.001, **P < 0.01, and *P < 0.05; not significant (n.s.): P > 0.05. Significance is relative to nontargeted control (NTC).

Fig. 3.. Functional characterization of hits reveals a preferential profinlamamtory phenotype.

(A) A heatmap depicting selected normalized transcript levels of the RNA-seq data from primary human macrophages with either no genetic modification, AhR KO, or IRAK1 KO. The analysis was done across six samples (two WT, two AhR KO, and two IRAK1 KO). (B) Heatmap of the enriched pathways from RNA-seq data from primary human macrophages showing enrichment of genes in a subset of biological processes involved in immune responses toward inflammation and cytokines regulation. (C) Cytokine profiling of the supernatant from primary human macrophages cultured in 96-well plates for 48 hours postfull differentiation (full differentiation takes 5 to 7 days). (D) ELISA of IL-10 from the supernatant of human primary macrophages. Data are mean ± SD of n = 3 replicates. (E) Western blot depicting the levels of the proinflammatory marker iNOS from primary macrophages. Histone H3 was used as a loading control. (F) A bar plot showing the MFI levels of CD86 and CD206 surface markers on the surface of macrophages. Data are mean ± SD of n = 3 replicates. (G) Overview of coculture assay of primary macrophages with primary CD8⁺/CD4⁺ T cells. (H) A bar plot showing the CFSE^low ratio of CD8⁺ cells after 72 hours of coculture with primary macrophages [the macrophages were differentiated from monocytes using macrophage colony-stimulating factor (M-CSF) + IL-4 + IL-10]. Data are mean ± SD of n = 3 replicates. (I) A bar plot showing the CFSE^low ratio of CD4⁺ cells after 72 hours of coculture with primary macrophages. CFSE^low corresponds to the ratio of cells in a proliferative state. Data are mean ± SD of n = 3 replicates. ****P < 0.0001; ***P < 0.001; **P < 0.01; *P < 0.05; P > 0.05.

Fig. 4.. AhR and IRAK1 regulate the expression of VISTA.

(A) A plot showing a correlation between NF-κB pathway mediators and T cell exhaustion markers of data retrieved from TCGA datasets across multiple cancer types. (B) A representative Western blot showing the levels of the indicated proteins of the NF-κB pathway in WT or IRAK1 KO U937 cells. β-Actin was used as loading control. (C) A heatmap depicting the genes involved in the NF-κB pathways retrieved from RNA-seq data of primary human macrophages performed earlier with either no genetic modification or IRAK1 KO. (D) A bar plot showing the MFI of the surface levels of VISTA protein after IRAK1 KO or tacrolimus (NF-κB inhibitor) or betulininc acid (NF-κB activator) treatment for 24 hours in U937-differentiated macrophages compared to WT control. (E) A bar plot depicting the binding of RelA (p65) to the promoter region of VISTA gene using ChIP–quantitative polymerase chain reaction (qPCR). (F) Representative immunofluorescence images of U937-differentiated macrophages either treated with dimethyl sulfoxide (DMSO) or an AhR inhibitor (inh.; 1 μM) for 48 hours post-PMA differentiation. (G) Representative Western blot of the nuclear fraction lysates of WT DMSO or Bay-218–treated (1 μM) differentiated U937 cells for 72 hours. Lamin A/C was used as the loading marker. (H) A box plot of AhR target genes from qPCR of primary human macrophages with either NTC or AhR KO. The values are displayed relative to unedited cells, and control are cells modified with a nontargeting sgRNA. (I) ChIP-seq of H3K27me3 or H3K9ac ChIP in either WT or AhR KO (n = 2 biological replicates) at the VISTA gene locus in human macrophages (the peaks were visualized using Integrative Genomics Viewer browser). ***P < 0.001; **P < 0.01; *P < 0.05; P > 0.05. IgG, immunoglobulin G; DAPI, 4′,6-diamidino-2-phenylindole.

Fig. 5.. Combination therapy IACT improved the survival and tumor control in CT26 syngeneic models.

(A) Treatment protocol for the combination therapy of the inhibitors and the antibodies. Triangles represent antibodies, and bottles represent inhibitors. Chemical structures of AhR (Bay-218) and IRAK1 (JH-X-119-01) inhibitors. Anti–CTLA-4 dose = 200 μg every 2 to 3 days for a max of five treatments, anti-VISTA dose = 300 μg every 2 to 3 days for a max of five treatments. q.o.d., Latin abbreviation for every other day. (B) Kaplan-Meier curves of the colorectal cancer CT-26 syngeneic models treated with indicated therapies. (C to F) Tumor volume burden curves (n = 10) of the colorectal cancer syngeneic models. R stands for (response to treatment). (G) Workflow of CT26 tumors immunophenotyping. (H to M) Quantitative assessment, obtained through FACS, of different immune cells per milligram of tumor tissue in control and treated CT26 tumors (n = 6). Error bars indicate means ± SD, and statistical significance was evaluated using a two-way analysis of variance (ANOVA). (N to Q) Quantitative evaluation, performed using FACS, of diverse immune myeloid cell populations per milligram of tumor tissue in control and treated CT26 tumors (n = 6). (R) A bar plot depicting the stained area percentage of immune cells in hematoxylin and eosin IHC samples isolated from lungs of mice inoculated with MB49 and treated with the indicated treatments (n = 2). (S) A heatmap exhibits the profiling of selected differential cytokines from pooled serum samples obtained from syngeneic mice (n = 3) inoculated with CT26 colorectal tumors after vehicle treatment or indicated treatments. Values represent the median of three mice, followed by z score scaling for each marker. ****P < 0.0001; ***P < 0.001; **P < 0.01; *P < 0.05; n.s., P > 0.05. GZMB, granzyme B.

Fig 6.. The use of AhR and IRAK1 inhibitors suppresses the expression of VISTA and potentiates ICB therapy outcomes.

This dual inhibition substantially suppresses VISTA which in turn embarks a more proinflammatory phenotype on the TME, shifting the macrophages toward the antitumor phenotype. Moreover, it shifts the chemokines and cytokines secreted from immunosuppressive to proinflammatory ones, such as IL-2, IL-12, etc., and augments intratumoral effector T cell infiltration and killing capacity.