Inhibition of the chemokine receptors CXCR1 and CXCR2 synergizes with docetaxel for effective tumor control and remodeling of the immune microenvironment of HPV-negative head and neck cancer models

Abstract

Background: Relapsed head and neck squamous cell carcinoma (HNSCC) unrelated to HPV infection carries a poor prognosis. Novel approaches are needed to improve the clinical outcome and prolong survival in this patient population which has poor long-term responses to immune checkpoint blockade. This study evaluated the chemokine receptors CXCR1 and CXCR2 as potential novel targets for the treatment of HPV-negative HNSCC.

Methods: Expression of IL-8, CXCR1, and CXCR2 was investigated in HNSCC tissues and human cell line models. Inhibition of CXCR1/2 with the clinical stage, small molecule inhibitor, SX-682, was evaluated in vitro and in vivo using human xenografts and murine models of HNSCC, both as a monotherapy and in combination with the taxane chemotherapy, docetaxel.

Results: High levels of IL-8, CXCR1, and CXCR2 expression were observed in HPV-negative compared to HPV-positive HNSCC tumors or cell lines. Treatment of HPV-negative HNSCC cell lines in vitro with SX-682 sensitized the tumor cells to the cytotoxic activity of docetaxel. In vivo, treatment of HNSCC xenograft models with the combination of SX-682 plus docetaxel led to strong anti-tumor control resulting in tumor cures. This phenomenon was associated with an increase of microRNA-200c and a decreased expression of its target, tubulin beta-3, a protein involved in resistance to microtubule-targeting chemotherapies. In vivo treatment of a murine syngeneic model of HNSCC with SX-682 plus docetaxel led to potent anti-tumor efficacy through a simultaneous decrease in suppressive CXCR2⁺ polymorphonuclear, myeloid-derived suppressor cells and an increase in cytotoxic CD8⁺ T cells in the combination therapy treated tumors compared to controls.

Conclusions: This study reports, for the first time, mechanistic findings through which the combination of CXCR1/2 inhibition and docetaxel chemotherapy exhibits synergy in models of HPV-negative HNSCC. These findings provide rationale for the use of this novel combination approach to treat HPV-negative HNSCC patients and for future combination studies of CXCR1/2 inhibition, docetaxel, and immune-based therapies.

Keywords: CXCR1; CXCR2; Docetaxel; HNSCC; IL-8; Immunotherapy.

Fig. 1

Expression of IL-8 and CXCR1/2 in HNSCC. a-h Representative images of tissues from a head and neck cancer tumor tissue array stained for IL-8 mRNA (pink), pan-cytokeratin (CK, green), CXCR1 and CXCR2 mRNA (CXCR1/2, red), or CD45 (white). a, b Representative cases corresponding to normal pharyngeal mucosa (a) and normal tongue (b). c-h Representative cases of squamous cell carcinoma of larynx stage IVA (c, e), epiglottis stage III (d, f), two cases of squamous cell carcinoma of larynx stage III (g, h). i A case of squamous cell carcinoma of the laryngopharynx stage III stained with antibodies against CXCR1 (pink) and CK (green). j A case of nasopharynx stage I stained with antibodies against CXCR2 (pink) and CK (green). k RNAseq data from the TCGA head and neck cancer database analyzed for expression of CXCL8, CXCR1, and CXCR2 in HPV-negative (n = 410) and HPV-positive (n = 90) tumors. l, m Representative images of HPV-negative tumor tissues from a head and neck cancer tumor tissue array stained for IL-8 mRNA (pink), pan-cytokeratin (CK, green), and CD45 (white) corresponding to squamous cell carcinoma of pharynx stage I (l) and pharynx stage II (m). n IL-8 secreted by HPV-negative and HPV-positive HNSCC cell lines in 48-h culture supernatants determined by ELISA and normalized to cell counts at time of collection. o RT-PCR analysis of CXCR2 gene expression in HNSCC cell lines, separated by HPV status. p Expression of CXCR1 and CXCR2 proteins in HPV-positive and HPV-negative HNSCC cell lines via immunoblot analysis. GAPDH was detected on the same membranes as a loading control. Numbers below blots indicate expression of the protein of interest as a ratio compared to the corresponding GAPDH. q Representative images of HPV-negative HNSCC cell lines stained for IL-8 mRNA (pink), pan-cytokeratin (CK, green), and CXCR2 mRNA (white). DAPI (blue) was used to stain nuclei in all representative images. Values graphed in (n, o) represent the mean ± SEM of technical replicates (n = 2 for ELISA assay; n = 3 for RT-PCR). Results are representative of 2 independent experiments. * p < 0.05, ** p < 0.01, **** p < 0.0001 for Student’s t-test in (k, n)

Fig. 2

Inhibition of the CXCR1/2 pathway improves susceptibility of HPV-negative HNSCC to docetaxel in vitro. a, b Indicated cell lines were pre-treated with DMSO (Control) or SX-682 (10 µM) for 72 h, followed by treatment with DMSO or SX-682 and a range of cisplatin (a) or docetaxel (b) concentrations for 72 h. A CellTiterGlo assay was used to determine % survival of cells and GraphPad Prism was used to calculate an IC50 value. Values represent the mean ± SEM of technical replicates and the solid line is a calculated nonlinear regression. Dotted horizontal line plotted at 50% survival. c Proliferation of indicated cell lines pre-treated with DMSO, 1 µM SX-682 (orange), or 2.5 µM SX-682 (teal) and treated with either 0.1 ng/mL docetaxel (circles) or 0.5 ng/mL docetaxel (triangles). d Proliferation of indicated cell lines transfected with control siRNA (black) or a pool of CXCR1/2 siRNA (teal) and treated with either 0.1 ng/mL docetaxel (circles) or 0.5 ng/mL docetaxel (triangles). Values represent the mean ± SEM of technical replicates. Results in a, b, c, and d are representative of 2 independent experiments. IC50 values were compared using an extra sum-of-squares F test in (a, b). * p < 0.05, ** p < 0.01, **** p < 0.0001 for 2-way ANOVA in (c, d)

Fig. 3

SX-682 modulates the phenotype of HPV-negative HNSCC xenografts and upregulates miR-200. a UM-SCC-11B and UM-SCC-74B xenografts grown in NSG-MHC I/II knockout mice were evaluated via RNA in situ hybridization for expression of IL-8 or CXCR1/2 mRNA (red) and staining for CD45 (white). b Final tumor volumes of UM-SCC-11B and UM-SCC-74B xenografts on day 15 post-tumor implantation in mice receiving a control or a SX-682-containing feed starting on day 7 or day 6, respectively. c RNAseq analysis on control and SX-682 treated UM-SCC-74B xenografts. Shown are selected activated and deactivated GO, HALLMARK, and C3 pathways when comparing SX-682 to control. d GSEA Enrichment plot showing normalized enrichment score by ranked genes for HALLMARK Epithelial-Mesenchymal Transition pathway. e Heatmap of leading-edge genes identified in (d). Data represents n = 3 mice per group. f Cell lines treated with SX-682 (teal) or DMSO control (gray) for 24–48 h, followed by analysis of miR-200a, miR-200b, and miR-200c by RT-PCR. Expression was normalized to expression of snoRNA135 and graphed as relative to DMSO cells. g UM-SCC-22B and UM-SCC-74B cells were pre-treated with SX-682 and transfected with anti-miR200c or a control anti-miR. Cells were subsequently treated with docetaxel (0.75 ng/mL and 0.5 ng/mL respectively) with or without SX-682. Growth inhibition was determined after 72 or 48 h respectively. h-i Tumors from control and SX-682 treated UM-SCC-11B (h) and UM-SCC-74B xenografts (i) from (b) were collected at endpoint and stained for expression of tubulin beta-3 (red), HLA-A/B/C (white), and DAPI (blue) (n = 4 tumors per group). Quantification of the MFI of tubulin beta-3 expression within the regions of HLA-A/B/C expression is also shown; each dot represents an individual region of interest (ROI). Results in f and g are representative of 2 independent experiments. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001 for Student’s t-test in (f, h, i) and for 1-way ANOVA followed by Tukey’s post hoc test in (g)

Fig. 4

CXCR1/2 inhibition increases susceptibility of HNSCC xenografts to docetaxel in vivo. a, d, g Schematics depicting the timeline of therapy with the UM-SCC-11B, UM-SCC-74B, and UPCI-SCC-90 models, respectively. Control diet or SX-682 feed (200 mg/kg) was administered to mice starting on day 6 or 7, as indicated, for the duration of the experiment. Docetaxel injections (5 mg/kg) were delivered by intraperitoneal (i.p.) injection on indicated days. Average tumor growth and tumor volumes at the end of study, respectively, for the UM-SCC-11B (b, c), UM-SCC-74B (e, f), and UPCI-SCC-90 (h, i) models (n = 6–7 mice/group in all studies). Gray shaded area marks the administration of SX-682 while vertical dotted lines denote docetaxel injections. UM-SCC-11B and UM-SCC-74B data are representative of 1 of 2 independent experiments. Values graphed in (b, e, h) represent the mean ± SEM; * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001 for 2-way ANOVA in (b, e, h), and for 1-way ANOVA followed by Tukey’s post hoc test in (c, f, i)

Fig. 5

CXCR1/2 inhibition increases susceptibility of murine HNSCC models to docetaxel leading to enhanced anti-tumor immunity. a MOC1 tumors were evaluated via RNA in situ hybridization for the expression of murine IL-8 homologs CXCL1 (red) and CXCL2 (white) or the receptor CXCR2 (red) and immunofluorescence for pan-cytokeratin (CK, green) and CD45 (white). Nuclear staining was performed by DAPI (blue). b Schematic depicting the timeline of therapy. Control or SX-682 feed (200 mg/kg) was administered to mice starting on day 6 and remained for the duration of the experiment. Docetaxel (5 mg/kg) was delivered by i.p. injection on indicated days. c Average tumor growth (n = 7–8 mice/group). Data are representative of 1 of 2 independent experiments. d Flow cytometry analysis of indicated immune infiltrating cells in the tumors of MOC1 treated mice. e Representative scatter plots depicting the expression of CXCR2 on Ly6G⁺ tumor infiltrating cells in control and SX-682 plus docetaxel groups. f Orthotopic MOC1 tumors were evaluated via RNA in situ hybridization for the expression of CXCL1 (red, left) and CXCR2 (red, right) and immunofluorescence for pan-cytokeratin (CK, green) and CD45 (white). Nuclear staining was performed by DAPI (blue). g Schematic depicting the timeline of therapy for treatment of orthotopic MOC1 tumors with dosing as in (b). h Average tumor growth (n = 9–10 mice/group). i Tumor volumes recorded at the final timepoint of the experiment. j Flow cytometry analysis of indicated immune infiltrating cells in the tumors from (h, i). k Bulk RNAseq transcriptome analysis was performed on mRNA isolated from control, SX-682, docetaxel, and SX-682 plus docetaxel treated MOC1 tumors from (c); n = 5 mice per group. Bubble plots depicting the top activated and deactivated GO and HALLMARK pathways when comparing SX-682, docetaxel, and SX-682 plus docetaxel treated tumors to control tumors are shown. l Average tumor growth of MOC1 tumors untreated or treated with SX-682 plus docetaxel with or without administration of depleting antibodies for CD8⁺ or Ly6G.⁺ cells; n = 8 mice in control, 9 mice in SX-682 plus docetaxel, and 6 mice in both CD8 and Ly6G depletion groups. Values graphed in (c, h, l) represent the mean ± SEM; * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001 for 2-way ANOVA in (c, h, l), and for 1-way ANOVA followed by Tukey’s post hoc test in (d, i, j)