Brequinar induces mRNA expression of antigen presentation pathway genes and upregulates cell surface MHC-I in diverse cancer cell lines.

A) Schematic of RNA sequencing experiment for panels B-D, with de novo pyrimidine pathway shown to highlight the role of DHODH B) Normalized enrichment scores for gene sets commonly enriched (FDR q < 0.25) in S2-013 (500nM) and CFPAC-1 (250nM) cells following two-week BQ treatment, as assessed by gene set enrichment analysis. C) GSEA plots for indicated gene sets following two-week BQ treatment of CFPAC-1 (left) or S2-013 (right) cells at the indicated doses D) Heatmap showing log2 fold change mRNA expression of APP genes in CFPAC-1 cells treated with BQ for indicated dose and duration. E) Heatmap showing log2 fold change mRNA expression for APP genes in A375 melanoma cells treated with the DHODH inhibitor teriflunomide (25µM) for indicated durations, data extracted from 37. F) RT-qPCR quantification of HLA-A, HLA-B, and HLA-C mRNA levels in cancer cell lines after 24-hour BQ treatment. Numbers represent fold change relative to vehicle control for each cell line. Data are representative of at least 3 independent experiments. HLA-B was not detectable in MiaPaCa2 cells. G-H) Flow cytometry analysis of cell surface MHC-I in live CFPAC-1 (G) or B16F10 (H) cells following 10-day treatment with BQ (250nM for CFPAC-1 and 10µM for B16F10). I-J) LC-MS/MS metabolomics quantification of de novo pyrimidine pathway metabolites in CFPAC-1 (I) or B16F10 (J) cells following 8-hour BQ treatment at indicated doses. Data represent mean +/- SD of four (CFPAC-1) or six (B16F10) biological replicates. # indicates p < 0.0001 by two-way ANOVA with Bonferonni’s post-comparison test.

BQ-mediated APP induction requires pyrimidine nucleotide depletion.

A) Dose-response cell viability experiment in B16F10 cells treated with BQ +/- uridine (100µM) for 72 hours. Data represent mean +/- SEM of three biological replicates. One representative result of three independent experiments is shown. B) Quantification of pyrimidine metabolites following 24-hour treatment of B16F10 cells with vehicle, BQ (10µM), or BQ + uridine (100µM). Data represent mean +/- SEM of six biological replicates. # indicates p < 0.0001 by two-way ANOVA with Bonferonni post-comparison test. C) RT-qPCR of indicated genes in B16F10 cells following 24-hour treatment with BQ or teriflunomide (100µM) +/- uridine (100µM) as in (B). Data represent mean +/- SEM of four technical replicates. One representative result of three independent experiments is shown. D) Left: flow cytometry analysis of cell surface MHC-I (H2-Db) on live B16F10 cells following 24-hour treatment with indicated agents (BQ 10µM, teriflunomide 100µM, uridine 1mM). Right: quantification of H2-Db mean fluorescence intensity normalized to vehicle control. Data represent mean +/- SEM of three independent experiments. # indicates p < 0.0001 with two-way ANOVA with Bonferroni post-comparison test. E) RT-qPCR analysis of indicated genes in S2-013 cells with DHODH knockout (sgDHODH) or non-targeting control vector (sgNT) treated with indicated agents for 72 hours. Data represent mean +/- SD of four determinations. One representative result of three independent experiments is shown. F) RT-qPCR analysis of indicated genes in CFPAC-1 cells following 72-hour treatment with indicated agents. Numbers in the heatmap represent mean fold change versus vehicle control with four determinations.

IKK2 inhibitor BMS-345541 abrogates BQ-mediated APP induction in an IKK2-independent manner.

A-B) HEK-293T cells were treated with indicated agents for 24 hours and then subjected to RT-qPCR analysis for indicated genes. Numbers in the heatmap represent mean of four determinations. C-E, G) B16F10 (C), CFPAC-1 (D), HCT116 (E) or MiaPaCa2-IKK2-KO (G) cells were treated with indicated agents for 24 hours and subjected to RT-qPCR analysis of indicated genes. Data in D,E, and G represent mean +/- SD of three independent experiments. * indicates p < 0.05, ** p < 0.01, *** p < 0.001, and # p < 0.0001 with two-way ANOVA with Bonferroni post-comparison test. For C, numbers in the heatmap represent mean fold change versus vehicle with three determinations. Representative result of three independent experiments is shown. F) Flow cytometry analysis of cell surface MHC-I in HCT116 cells treated with indicated agents for 24 hours.

P-TEFb inhibitor flavopiridol abrogates APP induction downstream of nucleotide depletion.

A) Plot of percent inhibition (10µM treatment) vs -log(dissociation constant) for kinases bound by BMS-345541 in KINOMEscan assays 46. Each data point represents an individual kinase. B-RT-qPCR analysis for indicated genes in HEK-293T cells treated with indicated agents for 24 hours. Data represent mean +/- SD of three independent experiments. # indicates p < 0.0001 by two-way ANOVA with Bonferonni’s post-comparison test D) Western blot analysis for CDK9 performed on HEK-293T cells treated with CDK9-targeted PROTAC (PROTAC2) and/or pomalidomide (POM) for 24 hours. Beta actin was used as a loading control. E) Visualization of Pol II occupancy after 48-hour treatment with teriflunomide (red) or vehicle control (blue) in A375 melanoma cells reported by Tan et al 37.

BQ inhibits tumor growth, increases tumor MHC-I, and enhances immune checkpoint blockade efficacy in B16F10 murine melanoma model.

A-D) B16F10 cells were injected subcutaneously into syngeneic hosts and mice were treated with brequinar (BQ; 10mg/kg IP daily) or vehicle control starting at day 7 post implantation. A) Longitudinal measurement of B16F10 subcutaneous tumors with BQ (10mg/kg IP daily) or vehicle treatment. B) Weight (left) and volume (right) of tumors at necropsy. C) Quantification of metabolites from B16F10 tumors harvested at necropsy. D) RT-qPCR analysis of APP genes from tumors harvested at necropsy. For (A-D), data represent mean +/- SD of n = 5 mice per group (n = 4 for BQ group in C). * indicates p < 0.05, ** p < 0.01, *** p < 0.001, and **** p < 0.0001 by unpaired t-test. E) Kaplan-Meier survival analysis for mice implanted with B16F10 tumors as in (A-D) and treated with indicated regimens; see Fig S5D for treatment timeline. * indicates p < 0.05, # p < 0.0001 by Mantel-Cox logrank test. For vehicle, immune checkpoint blockade (ICB; Anti-CTLA-4 and anti-PD-1; 100µg/mouse each, IP twice per week), and BQ (10mg/kg IP daily) + concurrent ICB, n = 15. For BQ, n = 7. For BQ + delayed ICB, n = 8.

BQ treatment upregulates APP genes and depletes pyrimidine nucleotides.

A) Heatmap showing log2 fold change mRNA expression of APP genes in S2-013 cells treated with BQ for indicated dose and duration. B-C) RT-qPCR quantification of APP genes after two-week BQ treatment of CFPAC-1 (250nM) (B) or B16F10 (10µM) (C) cells. D-E) Quantification of pyrimidine metabolites in CFPAC-1 (C) or B16F10 (D) cells treated with BQ for 8 hours at indicated doses. Data represent mean +/- SEM of four (CFPAC-1) or six (B16F10) biological replicates. * indicates p < 0.05, ** p < 0.01, *** p < 0.001, and # p < 0.0001 by unpaired t-test (B, C) or two-way ANOVA with Bonferonni’s post-comparison test (D, E).

Uridine rescues B16F10 cells from teriflunomide toxicity but does not alter APP expression by itself.

A) Dose-response cell viability experiment as in Fig 2A but with teriflunomide (Ter) instead of BQ. B) RT-qPCR analysis of indicated genes +/- single agent uridine treatment for 24 hours. Data represent mean +/- SEM of 4 determinations. One representative result of three independent experiments is shown. C) Flow cytometry analysis of cell surface MHC-I (H2-Db) following 24-hour uridine (1mM) treatment. Data represent mean +/- SEM of three independent experiments. D) RT-qPCR analysis of indicated genes in B16F10 cells following 24-hour treatment with MPA (5µM). Data represent mean +/- SEM of four determinations. One representative result of three independent experiments is shown.

A) RT-qPCR time course analysis for MHC-I genes in HEK-293T cells treated for indicated times with BQ (10µM). Data represent mean +/- SD of four determinations. B) RT-qPCR analysis for Nlrc5 (left) or Tap1 (right) in B16F10 cells treated for 24 hours with indicated agents. Data represent mean +/- SD of four determinations. # indicates p < 0.0001 with two-way ANOVA with Bonferroni post-comparison test. Representative results for one of three independent experiments are shown. C) Heatmap indicating RT-qPCR analysis for indicated genes in wild-type or IKK2-KO MiaPaCa2 cells treated with indicated agents for 24 hours. Numbers in the heatmap represent mean of four determinations.

A) RT-qPCR analysis for indicated genes in HCT116 cells treated with indicated agents in the presence or absence of flavopiridol (1µM). Numbers in the heatmap represent mean of three determinations.

A) Overall survival in melanoma patients (SKCM) from the Cancer Genome Atlas (TCGA) with above (indicated in red) and below median (indicated in blue) mRNA expression of indicated genes. B-C) Principal component analysis (PCA) plot (B) and unsupervised hierarchical clustering (C) from metabolomics analysis of B16F10 tumors at necropsy (Fig 5C). Treatment regimen of mice from Fig 5E.