Inhibiting the biogenesis of myeloid-derived suppressor cells enhances immunotherapy efficacy against mammary tumor progression

Abstract

While immune checkpoint inhibitors (ICIs) have transformed the therapeutic landscape in oncology, they are effective in select subsets of patients. Efficacy may be limited by tumor-driven immune suppression, of which 1 key mechanism is the development of myeloid-derived suppressor cells (MDSCs). A fundamental gap in MDSC therapeutics is the lack of approaches that target MDSC biogenesis. We hypothesized that targeting MDSC biogenesis would mitigate MDSC burden and bolster tumor responses to ICIs. We tested a class of agents, dihydroorotate dehydrogenase (DHODH) inhibitors, that have been previously shown to restore the terminal differentiation of leukemic myeloid progenitors. DHODH inhibitors have demonstrated preclinical safety and are under clinical study for hematologic malignancies. Using mouse models of mammary cancer that elicit robust MDSC responses, we demonstrated that the DHODH inhibitor brequinar (a) suppressed MDSC production from early-stage myeloid progenitors, which was accompanied by enhanced myeloid maturation; (b) augmented the antitumor and antimetastatic activities of programmed cell death 1-based (PD-1-based) ICI therapy in ICI-resistant mammary cancer models; and (c) acted in concert with PD-1 blockade through modulation of MDSC and CD8+ T cell responses. Moreover, brequinar facilitated myeloid maturation and inhibited immune-suppressive features in human bone marrow culture systems. These findings advance the concept of MDSC differentiation therapy in immuno-oncology.

Keywords: Cancer immunotherapy; Immunology.

Figure 1. BRQ reverses the suppressive activity of MDSCs.

Female BALB/c mouse BM cells were cultured with 40 ng/mL recombinant mouse (rm) G-CSF plus rmGM-CSF for 96 hours with or without 1 μM BRQ (Tocris). (A) Flow cytometric analysis of CD11b and Gr-1 expression in cultures treated with vehicle (Veh) or BRQ. (B) Percentage of CD11b⁺Gr-1⁺ cells. (C) Percentage of viable cells as determined by trypan blue staining and live cell quantification. (D) Percentage of apoptotic cells, as determined by annexin V and DAPI staining of vehicle- or BRQ-treated MDSCs. (E) CD4⁺ and CD8⁺ T cell proliferation following coculture with MDSCs generated with or without 1 μM BRQ (from Clear Creek) or 25 μM Lef. Splenocytes from naive syngeneic mice were used as a source of T cells and were stimulated with 1 μg/mL anti-CD3 (αCD3) mAb for 72 hours. Cell proliferation was measured using CellTrace Violet. (F) CD4⁺ and CD8⁺ T cell proliferation following coculture with MDSCs with or without BRQ in the absence or presence of 200 �M uridine. Data are presented as the mean ± SEM of 5 separate experiments (B and C), 6 separate mice (D), or triplicate determination (E and F). **P < 0.01 and ****P < 0.0001, by unpaired t test (B–D).

Figure 2. BRQ induces myeloid cell maturation and reduces the expression of immune-suppressive genes in MDSCs.

(A) Cytospins of BM cultures were stained using Wright-Giemsa and analyzed for the indicated cell populations. Photomicrograph images show cells treated with vehicle or BRQ (from Tocris). Original magnification, ×1,000. The percentage of each cell type, shown in the graph, was quantified as follows: 300 cells/slide for each treatment condition were analyzed (100 cells/field × 3 fields) in biological duplicates. The average number of cells across those 6 fields covering the 2 separate slides possessing the indicated morphology was then recorded as a percentage of the total population reflecting those 3 scored cell types. (B–E) BM cells were cultured as in Figure 1 with or without BRQ (Clear Creek) or with 25 μM Lef. (B) CD11b⁺Ly6C^loLy6G⁺ and CD11b⁺Ly6C^hiLy6G^– cells were analyzed by flow cytometry for surface CD101 expression. Top: Histograms depict CD101 expression. Bottom: Percentage of CD101⁺ cells (left) and CD101 MFI of the indicated cell subset. (C) PMN-MDSCs were recovered after in vitro culturing by a positive magnetic bead selection method (Miltenyi) and analyzed by RT-qPCR for expression of the indicated genes. (D) PMN-MDSCs were recovered after in vitro culturing and analyzed by flow cytometry for VEGF-A and iNOS expression. (E) Bulk MDSCs were lysed after in vitro culturing for an arginase activity assay, as measured by urea production. Data in B are presented as the mean ± SEM or SD of 3 (BRQ) or 2 (Lef) separate experiments, respectively. Data in C are presented as the mean ± SEM of triplicate determinations and are representative of 2 independent experiments with similar results. Data in D and E are presented as the mean ± SEM of results involving 4–5 separate mice. *P < 0.05, **P < 0.01, and ***P < 0.001, by unpaired t test (B–E).

Figure 3. BRQ inhibits MDSC function in vivo.

(A) 4T1 tumor growth in mice treated with vehicle or daily injections of 10 mg/kg BRQ (from Clear Creek), starting on day 9 after implantation. (B) Flow cytometric analysis of Ly6C and Ly6G expression on the gated splenic CD11b⁺ cells. (C) Absolute numbers of splenic CD11b⁺Ly6C^loLy6G⁺ and CD11b⁺Ly6C^hiLy6G^– cells from B. (D) Numbers of WBCs, lymphocytes, monocytes, and granulocytes per microliter of peripheral blood. Boxed areas indicate the normal range (NR). (E) Expression of CD101 on the gated splenic CD11b⁺Ly6C^loLy6G⁺ cells. The percentage of CD101⁺ cells and CD101 MFI are shown. (F) Expression of Ly6G and Ly6C on the gated CD11b⁺Ly6C^loLy6G⁺ cells. (G) Proliferation of activated CD4⁺ or CD8⁺ T cells following coculturing with PMN-MDSCs isolated as in Figure 2C from 4T1-bearing animals. (H) Fold change in gene expression in purified MDSCs from the spleen (determined by RT-qPCR and normalized to PPIA). The Miltenyi Mouse Myeloid-Derived Suppressor Cell Isolation Kit was used for CD84, JAML, and TGFB1, and the STEMCELL EasySep Mouse MDSC isolation kit was used for ARG1, NOS2, S100A8, and S100A9. Data in all panels are presented as the mean ± SEM of the indicated data points. (A–G) n = 3–10 mice/group. Data in H are presented as the mean ± SEM of triplicate determinations and are representative of experiments using up to 3 separate mice, with similar results. *P < 0.05, **P < 0.01, and ***P < 0.001, by 2-sided Wald test (A) and unpaired t test (C–G).

Figure 4. BRQ enhances the antitumor efficacy of anti–PD-1 therapy and reduces spontaneous lung metastases.

(A) Scheme of the in vivo treatment of tumor-bearing mice with or without BRQ (from Clear Creek). (B) 4T1 and (C) E0771.ML-1 tumor growth in mice treated with vehicle, BRQ (10 mg/kg BRQ), anti–PD-1 (200 μg/injection), or the combination of BRQ and anti–PD-1, as shown in A. Tumors in the vehicle, BRQ alone, and anti–PD-1 alone treatment groups all had significantly higher growth rates than did tumors in the BRQ plus anti–PD-1 combination group. (D) Number of spontaneous lung metastases in 4T1-bearing mice receiving the treatments indicated in B. (E) Percentage of E0771.ML-1–bearing mice that had spontaneous lung metastases (n = 10 mice/group) after receiving the treatments indicated in C. (F) 4T1 tumor growth in mice treated with vehicle, BRQ (10 mg/kg), BRQ plus anti–PD-1 (200 μg/injection), anti–CTLA-4 (100 μg/injection), or BRQ plus anti–CTLA-4 as shown in A. Results for the combination treatment groups (BRQ plus anti–PD-1 and BRQ plus anti–CTLA-4) were not statistically different. (G) Number of spontaneous lung metastases in 4T1-bearing mice receiving the indicated treatments, as in F. Data are presented as the mean ± SEM of the indicated data points and represent 6–10 mice/group. *P < 0.05, **P < 0.01, and ****P < 0.0001, by 2-sided Wald test for the combination treatment group versus the vehicle control or the single-agent treatment groups for the tumor growth curves (B, C, and F) and by unpaired t test for lung metastasis data (D and G).

Figure 5. Inhibition of tumor growth by combined BRQ and anti–PD-1 therapy is dependent on depletion of uridine or MDSCs, and the presence of CD8⁺ T cells.

(A) Experimental scheme and 4T1 tumor growth rates in mice treated with vehicle, BRQ (from Clear Creek) plus anti–PD-1, or BRQ plus anti–PD-1 plus uridine. Uridine (300 mg/kg i.p.) was administered concomitantly with BRQ for the duration of the experiment. (B) Experimental scheme and 4T1 tumor growth rates in mice treated with vehicle, BRQ plus anti–PD-1, or BRQ plus anti–PD-1 plus CD11b⁺Gr-1⁺ MDSCs. MDSCs were flow-sorted from the spleens of syngeneic female Irf8^–/– mice (aged 10–12 weeks), and 1 × 10⁶ cells were administered i.v. on days 7 and 14. (C and D) Experimental schemes and 4T1 tumor growth rates in mice treated with vehicle, BRQ plus anti–PD-1, or BRQ plus anti–PD-1 plus anti-CD8–depleting antibody. Mice received anti-CD8–depleting antibody or isotype (400 μg/mouse i.p.) at the indicated time points. Mice in C received the first dose of anti-CD8–depleting antibody 3 days prior to 4T1 tumor implantation, and mice in D received the first dose of anti-CD8–depleting antibody 7 days after 4T1 tumor implantation. Data are presented as the mean ± SEM of multiple determinations (n = 5–10 mice/group). *P < 0.05, **P < 0.01, and ****P < 0.0001, by 2-sided Wald test for the combination treatment group versus the vehicle control or the specified experimental group in A–D.

Figure 6. BRQ increases the activation state of CD8⁺ T cells and the maturation of PMN-MDSCs within the TME.

4T1-bearing mice were treated with or without BRQ (Clear Creek), as shown in the preceding figures. At the experimental endpoint (average tumor volume = 600–800 mm³), tumors were removed and analyzed by flow cytometry. (A) Absolute numbers of CD45⁺ cells, (B) PMN-MDSCs (left), M-MDSCs (middle), and macrophages (far right) per gram of tumor tissue. (C) Expression of CD101 (percentage), Ly6C (MFI), and PD-L1 and PD-L2 (MFI) by gated PMN-MDSCs. (D) CD11b⁺Gr-1⁺ MDSCs were recovered from individual 4T1 tumors using the STEMCELL EasySep Mouse MDSC isolation kit and analyzed by RT-qPCR for the indicated genes. (E) Absolute number of CD8⁺ T cells per gram of tumor tissue. (F) Percentage of CD8⁺ T cells expressing PD-1, PD-1 and Ki-67, CD25, CD44, or ICOS. Data are presented as the mean ± SEM of multiple determinations (shown as individual data points; n = 3–5 mice/group). Data in D are presented as the mean ± SEM of triplicate determinations and are representative of 2–3 separate mice, with similar results. *P < 0.05, **P < 0.01, and ****P < 0.0001, by unpaired t test.

Figure 7. BRQ inhibits the development of MDSCs from BM myeloid progenitors.

(A) Proliferation of activated CD4⁺ or CD8⁺ T cells following coculturing with GMP-derived MDSCs. Unstim, unstimulated. (B) scRNA-Seq experiments were performed on c-Kit⁺ BM cells isolated from NTB mice or 4T1-bearing mice treated with vehicle or BRQ. For each experimental group, 3 biologic replicates were pooled. GSEA of the REACTOME UPR pathway comparing Veh-GMPs with NTB-GMPs (left) and BRQ-GMPs with Veh-GMPs (right). (C) Heatmap showing up- and downregulation of UPR pathways (P < 0.01, FDR < 0.25) in GMPs, based on the indicated comparisons. (D) GSEA of the Pathway Interaction Database (PID) Cdc42 signaling pathway comparing Veh-GMPs versus NTB-GMPs (left) and BRQ-GMPs versus Veh-GMPs (right). (E) Heatmap showing up- and downregulation of pathways related to RhoGTPase signaling (P < 0.01, FDR < 0.25) in GMPs, based on the indicated comparisons. (F) GSEA of the Kyoto Encyclopedia of Genes and Genomes (KEGG) leukocyte migration and the REACTOME neutrophil degranulation pathways comparing BRQ-GNs and Veh-GNs. Data in A are presented as the mean ± SEM of triplicate determinations from 2 separate mice. ***P < 0.001, by unpaired t test. In B, D, and F, the normalized enrichment score, FDR (q), and nominal P value are shown. NES, normalized enrichment score; Neg., negative; Pos., positive.

Figure 8. BRQ inhibits the development of an immunosuppressive phenotype in human myeloid cells.

Human BM cells were cultured for 96 hours with recombinant human GM-CSF (rhGM-CSF) and rhG-CSF (40 ng/mL each) plus 1 μM BRQ (from Clear Creek) or vehicle. (A) Flow cytometric analysis of CD33 expression and SSC properties (84%–88% CD33⁺ with or without BRQ treatment). FSC-A, forward scatter area; FSC-H, forward scatter height. (B) Percentage of CD33⁺ myeloid cells exhibiting high SSC (SSC^hi). (C) Histograms depicting CD101, CD11b, or HLA-DR expression by the gated CD33⁺ cells and quantification of the MFI values for CD101, CD11b, and HLA-DR by the gated CD33⁺ cells. Data are presented as the mean ± SEM of 3 separate donors. (D) Fold change in expression of ARG1, NOS2, and IL10 (determined by RT-qPCR analysis) in cultured human BM cells. Expression was normalized to PPIA and is depicted for each donor (n = 5 separate donors, including the 3 donors from A–C). The absence of a bar indicates no detectable signal for the expression of either ARG1, NOS2, or IL10. Donor 1: female, age 29 years; donor 2: female, age 48 years; donor 3: female, age 13 years; donor 4: male, age 7 years; donor 5: male, age 12 years. *P < 0.05, by paired t test (B and C).