Fatty acid transport protein 2 reprograms neutrophils in cancer

Abstract

Polymorphonuclear myeloid-derived suppressor cells (PMN-MDSCs) are pathologically activated neutrophils that are crucial for the regulation of immune responses in cancer. These cells contribute to the failure of cancer therapies and are associated with poor clinical outcomes. Despite recent advances in the understanding of PMN-MDSC biology, the mechanisms responsible for the pathological activation of neutrophils are not well defined, and this limits the selective targeting of these cells. Here we report that mouse and human PMN-MDSCs exclusively upregulate fatty acid transport protein 2 (FATP2). Overexpression of FATP2 in PMN-MDSCs was controlled by granulocyte-macrophage colony-stimulating factor, through the activation of the STAT5 transcription factor. Deletion of FATP2 abrogated the suppressive activity of PMN-MDSCs. The main mechanism of FATP2-mediated suppressive activity involved the uptake of arachidonic acid and the synthesis of prostaglandin E₂. The selective pharmacological inhibition of FATP2 abrogated the activity of PMN-MDSCs and substantially delayed tumour progression. In combination with checkpoint inhibitors, FATP2 inhibition blocked tumour progression in mice. Thus, FATP2 mediates the acquisition of immunosuppressive activity by PMN-MDSCs and represents a target to inhibit the functions of PMN-MDSCs selectively and to improve the efficiency of cancer therapy.

Extended data Figure 1.. Lipid accumulation and expression of lipid transporters in MDSC

a. Lipid accumulation measured by BODIPY staining in PMN-MDSC isolated from spleen and tumors of indicated tumor models. Each group included 4–8 mice. Each circle represents an individual mouse. Mean ± SD are shown. Inset - confocal image representative of 2 independent experiments. b. Lipid accumulation in PMN generated from BM HPC with GM-CSF and TES. (n=3–5). c. LC/MS analysis of TG in PMN from control mice and PMN-MDSC from of EL4 TB mice (n=4). d. Lipid accumulation measured by BODIPY staining in M-MDSC isolated from spleen and tumor of indicated tumor models (n=10). Each circle represents an individual mouse. Mean ± SD are shown. e. Lipid accumulation in DC and MDSC generated from CD204 KO HPC in presence of TES (n=3). Mean ± SD are shown. f. Lipid accumulation in PMN-MDSC from spleen of tumor bearing WT and CD204 KO mice (n=3). Mean ± SD are shown. g. Suppressive activity of PMN-MDSC from spleen of tumor bearing WT and CD204 KO mice. Representative of 4 experiments each performed in triplicates. Mean ± SD are shown. h. Expression of msr1, fabps, slc27a (1–5), cd36 in control PMN and PMN-MDSC isolated from spleen and tumor of EL4 TB mice (n=4–5). Mean ± SD are shown. In all panels, p values were calculated in unpaired two-sided Student’s t-test: *P<0.05; ^**P<0.01; ^***P<0.001; ^****P<0.0001.

Extended data Figure 2.. Expression of gene involved in lipogenesis in PMN-MDSC.

a. FATP2 protein in control PMN and PMN-MDSC from spleen of TB mice. Representative of 2 experiments. b. FATP2 in PMN generated in vitro from BM HPC. Representative of 2 experiments. For gel source data, see Supplementary Figure 1. c. F244 tumor growth in WT and FATP2 KO SV129veve mice (n=4). d. Verification of correct targeting of FATP2 by RT-qPCR and WB in PMN-MDSC isolated from spleen of slc27a2^fl/fl x s100a8-cre⁻ and slc27a2^fl/fl x s100a8-cre⁺ TB mice. For gel source data, see Supplementary Figure 1. e. IFNγ production by CD8⁺ T cells and CD4 and CD8 T-cell proliferation (n=3) in WT and FATP2KO mice. Mean ± SD are shown. f. Suppressive activity of M-MDSC isolated from WT or FATP2 KO TB mice. Dashed line shows T cell proliferation without MDSC. Four experiments with similar results were performed. g. Suppressive activity of TAM from WT or FATP2 KO TB mice. Dashed line shows T-cell proliferation without macrophages. 3 independent experiments with similar results were performed. In all panels mean and SD are shown. h. Growth of EL4 tumors in WT and FATP4 KO mice (n=4). Representative of 2 experiments. Mean ± SD are shown. i. Suppressive activity of PMN-MDSC isolated from spleen or tumor of WT or FATP4 KO mice. Representative of 2 independent experiments performed in triplicates. Mean ± SD are shown. Dotted line – control values of T cell proliferation without presence of PMN-MDSC. j. Growth of LLC tumor in WT and CD36 KO mice, depleted of CD8⁺ T cells (n=3). Mean ± SD are shown. k. Lipid accumulation measured by BODIPY staining in PMN-MDSC and M-MDSC isolated from spleen and tumor of CD36 KO mice (n=3). Mean ± SD are shown. In all experiments p values were calculated in unpaired two-sided Student’s t-test.

Extended data Figure 3.. Effect of FATP2 KO on mRNA gene expression.

a. Expression heatmap for genes affected at least 5-fold, b. Number of significantly affected genes (FDR<5%) for different fold change thresholds. c. List of upstream regulators whose targets were found by Ingenuity Pathway Analysis (IPA) as significantly enriched among genes affected by FATP2 KO. n=number of affected targets, p=enrichment pvalue, Z=activation z-scores calculated by IPA represent predicted regulator state based on the known effect on target and direction of mRNA change. Negative activation z-scores predict inhibition and positive z-scores – activation of the regulator in the FATP2 KO mice.

Extended data Figure 4.. LC-MS analysis of lipids from WT and FATP2 KO PMN-MDSC.

a. TGs molecular species containing LA (18:2), docosapentaenoic acid (22:5), and docosahexaenoic acid (22:6) (n=7). b. Total CE and CE (20:4) molecular species (n=7). c. LA (18:2), docosapentaenoic acid (22:5), and docosahexaenoic acid (22:6) fatty acids (n=12). d. Distribution of major PLs in FATP2 KO and WT PMN-MDSC samples. e. Content of PLs containing AA in PE, PC, PI, and PS (n=12). f. Content of AAd11 labelled phospholipids (PI, PG, PA and PS), n=5. Statistical analysis was performed using unpaired two-sided Student’s t-test *P<0.05; ^**P<0.01 (each circle indicates an individual mouse, Mean ± SD).

Extended data Figure 5.. Metabolomic analysis and expression of FAO related genes in PMN-MDSC.

a. Oxygen consumption rate (OCR) (top panel) and basal OCR (bottom panel) of WT and FATP2 KO PMN-MDSC. Representative of 2 experiments is shown in top panel (n=3–4). Bottom panel cumulative results are shown. Each circle indicates an individual mouse (n=7). Mean and SD are shown. P values were calculated using unpaired two-sided Student’s t-test. b. Extracellular acidification rate (ECAR) (top panel) and basal ECAR (bottom panel) of WT and FATP2 KO PMN-MDSC. Top panel - representative of 2 independent experiments (n=3–4). Bottom panel – cumulative results. Each circle indicates an individual mouse (n=7). Mean and SD are shown. Statistical analysis (unpaired two-sided Student’s t-test) was performed. NS, not significant; *P<0.05. c. Carbon-13 labeling of TCA cycle intermediates and associated amino acids. Ex vivo MDSC were cultured in physiological-like medium supplemented with BSA-conjugated 13C16-palmitate and GM-CSF for 18 hours. Metabolites were then extracted and analyzed by high-resolution LC-MS. Carbon-13 isotopologs (M+x) for each metabolite are represented as normalized stacked bars. Data presented as Mean and SD of 3 biological replicates. Statistical analysis (unpaired two-sided Student’s t-test) was performed. d. Expression of genes involved in FAO. RT-qPCR analysis of cpta1, acadm, hadha expression in control PMN and PMN-MDSC isolated from spleen and tumor of TB mice. Each group included 3–6 mice and shown as Mean ± SD

Extended data Figure 6.. Exchange of nutrients with the media.

Ex vivo MDSCs were cultured in physiological-like medium supplemented with GM-CSF for 18 hours. Metabolites were then extracted from the media and analyzed by LC-MS. Upward bars represent efflux from the cells into the media, and downward bars represent uptake (or depletion) from the media by the cells. Data are normalized to protein content after extraction. Data are presented as a mean and SD (n=3).

Extended data Fig. 7.. Effect of AA on PGE2 production and suppressive activity of PMN-MDSC.

a. LC/MS analysis of PGE2 in PMN from control mice and PMN-MDSC from EL4 and CT26 TB mice (n=3). Mean ± SD are shown. b. PGE2 release (ELISA) by control PMN (n=4), PMN-MDSC from WT (n=11), and FATP2 KO (n=8) LLC TB mice. c. Expression of ptges in PMN-MDSC isolated form spleen of EL4 TB mice (n=13–15), KPC (n=3), RET (n=3–6). Mean ± SD are shown. d. Expression of ptgs2 and ptges in PMN-MDSC (qRT-PCR) (n=6). e. Expression of arg1 and nos2 (qRT-PCR) in spleen PMN-MDSC from WT and FATP2 KO EL4 TB mice (n=3–5). Mean ± SD are shown. f. Flow cytometry of myeloid cells differentiated from HPC in presence of AA. Representative of 3 experiments. g. Expression of arg1, nos2 and nox2 in PMN isolated from HPC cultures with AA. Data are pooled from 6 independent experiments and depicted as mean ± SD. h. pSTAT5 by flow cytometry at different time points in mouse PMN isolated from BM treated with different amounts of GM-CSF. Representative of 3 independent experiments. i. LLC tumor growth (n=4) in stat5^fl/fl:cre⁻ and stat5^fl/fl:cre⁺ mice. j. Slc27a2 expression (RT-qPCR) in PMN-MDSC from spleen of WT and KO TB mice (n=4). Statistical analysis - unpaired two-sided Student’s t-test: NS, not significant; *P<0.05; ^**P<0.01

Extended data Figure 8.. Lipid accumulation in MDSC from cancer patients.

a. Amount of lipids (BODIPY staining) in M-MDSC isolated from blood of cancer patients or healthy individuals. Each circle indicates an individual and Mean ± SD are also shown. b. Amount of lipids (BODIPY staining) in M-MDSC from blood and tumor tissue of cancer patients. Each circle indicates an individual (n=5). c. RNAseq analysis of genes involved in lipid accumulation in human LOX1⁺ PMN-MDSC and LOX1⁻ PMN (n=4). d. PTGES expression in LOX1⁺ and LOX1⁻ PMN from blood of cancer patients. Fold change over LOX1⁻PMN (n=3). e. Slc27a2 expression in M-MDSC and monocytes isolated from blood of cancer patients and healthy donors, respectively. Each circle indicates an individual (n=4–6). Mean ± SD are shown. f. pSTAT5 by flow cytometry at different time points, in human PMN isolated from blood of healthy donor and treated with different amounts of GM-CSF. g. FATP2 in PMN isolated from blood of healthy donors and treated with GM-CSF. Representative of 3 independent experiments is shown. For gel source data, see Supplementary Figure 1. h. Content of total PE and AA-containing PE species in PMN-MDSC isolated from lung cancer patients or healthy donors. Each circle indicates an individual; mean ± SD (n=4). Statistical analysis - unpaired two-sided Student’s t-test: *P<0.05; ^**P<0.01; ^***P<0.001; ^****P<0.0001.

Extended data Figure 9.. Effect of lipofermata treatment on tumor-bearing mice.

a. MTT assay after 3-day incubation of tumor cells with indicated concentration of lipofermata. b. Percentage and absolute number of tumor-associated antigen (E7-derived peptide) specific CD8⁺ T cells in draining lymph nodes of mice bearing TC-1 tumor and treated with lipofermata (n=3). Means and SD are shown. P values were calculated in two-sided Student’s t-test: *P<0.05; c. Growth of TC-1 tumors in mice treated with CTLA4 antibody and lipofermata (n=5). Mean and SD are shown. d. CD8⁺ T cell infiltration of TC-1 tumors in mice treated with CTLA4 antibody and lipofermata. Typical staining of 2 different mice is shown. Scale bar = 50 μm. Bottom – the number of CD8+ T cells per mm2 (n=2). e. Growth of TC-1 in mice treated with PD1 antibody and lipofermata (n=5). Mean ± SD are shown. P values are calculated in two-way ANOVA test with correction for repeated measurements. *P<0.05; ^**P<0.01; ^***P<0.001; ^****P<0.0001.

Figure 1.. The effect of FATP2 deletion on tumor growth and PMN-MDSC function.

a. Slc27a2 expression in control PMN and PMN-MDSC from TB mice. b. Slc27a2 expression in M-MDSC and PMN-MDSC from spleen of TB mice. c. Slc27a2 expression in indicated cells in EL4 TB mice. In a-c results of individual mice are shown. N=4–5. d. EL4 or LLC tumor growth in C57BL/6 mice (n=4–5). Representative of 2 experiments. e. EL4 tumor in mice reconstituted with WT or FATP2 KO BM cells. N=4–5. Representative of 2 experiments. f. EL4 or LLC tumors in WT and FATP2 KO mice depleted of CD8⁺ T cells. Representative of 2 experiments (n=4–5). g. LLC tumors in mice with FATP2 KO targeted to PMN (S100A8-cre). N=4. h. Suppression of T-cell proliferation by PMN-MDSC isolated from WT or FATP2 KO TB mice. Four experiments with similar results were performed. Dashed line shows T cell proliferation without MDSC. In d-g p values were calculated in two-way ANOVA test. In all other panels p values were calculated using unpaired two-sided Student’s t-test. In all panels *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001 between control and test samples.

Figure 2.. Mechanism of FATP2 mediated suppression by PMN-MDSC.

a. TG in PMN-MDSC from spleens of EL4 TB WT and FATP2 KO mice (n=7). b. FA in PMN-MDSC from spleen of TB WT (n=12) and FATP2 KO mice (n=11). c. Phospholipid species containing AA residues in PMN-MDSC from spleen of TB WT (n=12) and FATP2 KO mice (n=10). d. AAd11, and PGE2d11 in PMN-MDSC from spleen of TB WT and FATP2 KO mice (n=5). e. AAd11 labelled PE and PC in PMN-MDSC from WT (n=5) and FATP2 KO (n=4) TB mice. f. PGE2 (LS-MS) in PMN-MDSC from WT and FATP2 KO mice (n=6). g. Expression of slc27a2 (qRT-PCR) in PMN generated from HPC transduced with lentivirus expressing FATP2 or GFP (n=4). h. PGE2 release from cells described in g. N=4. In g,h - fold increase over GFP⁻ cells after transduction. i. Suppressive activity (in triplicates) of PMN differentiated from HPC in the presence of AA. Representative of 3 experiments is shown. Dashed line shows T cell proliferation without MDSC. j. PGE2 production by PMN differentiated from HPC in the presence of AA (n=5). Fold changes over the control. k. PGE2 production by PMN differentiated from Ptgs2 KO HPC in the presence of AA. N=4. Fold changes over control. l. Suppressive activity (in triplicates) of PMN differentiated from Ptgs2 KO HPC in the presence of AA. Two independent experiments were performed. Dashed line shows T cell proliferation without MDSC. In all experiments Mean ± SD are shown and p values were calculated using unpaired two-sided Student’s t-test. *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001 between control and test samples.

Figure 3.. Regulation of FATP2 in PMN-MDSC.

a. FATP2 in PMN treated with GM-CSF. Representative of 3 experiments. For gel source data, see Supplementary Figure 1. b. ChIP assay with STAT5 antibody in PMN from BM treated with GM-CSF. Triplicate measurements of representative of 2 experiments are shown c. FATP2 in stat5^fl/fl:s100a8-cre PMN treated with GM-CSF. Representative of 3 experiments. d. Amount of lipid (BODIPY staining) in PMN-MDSC isolated from blood healthy individuals (n=9) or patients with head and neck cancer (n=11), non-small cell lung cancer (n=6), or breast cancer (n=5). e. Amount of lipid (BODIPY staining) in PMN-MDSC isolated from blood and tumor tissue of cancer patients with NSCLC (n=4). f. Expression of SLC27A2 by RT-qPCR in PMN-MDSC isolated from blood of cancer patients and healthy donors. Fold change over control PMN (n=6). g. FATP2 in PMN-MDSC isolated from blood of cancer patients or healthy individuals. Representative of 3 experiments. h. SLC27A2 expression by RT-qPCR in LOX1⁺ and LOX1⁻ PMN from blood of cancer patients. Fold change over LOX1⁻ PMN (n=8) i. LS/MS lipidomics of TG in PMN from healthy donors and PMN-MDSC from cancer patients. N=4. j. LS/MS lipidomics of free AA, LA, and DHA in PMN from healthy donors and PMN-MDSC from cancer patients (n=4). k. LS/MS lipidomics of PGE2 in PMN from healthy donors and PMN-MDSC from cancer patients (n=4). In all panels Mean ± SD are shown. Statistical analysis was performed using Two-Way ANOVA (d). In all other panels using unpaired two-sided Student’s t-test: NS, not significant; *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001.

Figure 4.. Therapeutic effect of targeting FATP2.

Treatments with lipofermata (2 mg/kg twice per day s.c.) started 8–10 days after tumors injections. CTLA4 antibody (200 μg/mouse i.p.) was administrated at day 7 and day 11 after tumor injection. CSF1R antibody (300 μg/mouse; every other day). a. Growth of indicated tumors in C57BL/6 mice treated with lipofermata. Representative of 2 independent experiments (n= 4–5 mice per group) are shown. b. Growth of LLC tumors in NOD-SCID mice treated with lipofermata (n=5). c. Growth of LLC tumor in mice depleted of CD8 T cells and treated with lipofermata (n=5). d. Growth of LLC tumor in mice treated with CTLA4 antibody and lipofermata (n=5). e. Growth of LLC tumor in mice treated with CSF1R inhibitor and lipofermata (n=5). In all experiments Mean ± SEM are shown and statistical analysis was performed in two-way ANOVA test with corrections for multiple comparison *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001 differences from untreated cells and between treated groups.