Metformin increases natural killer cell functions in head and neck squamous cell carcinoma through CXCL1 inhibition

Abstract

Background: Metformin slows tumor growth and progression in vitro, and in combination with chemoradiotherapy, resulted in high overall survival in patients with head and neck cancer squamous cell carcinoma (HNSCC) in our phase 1 clinical trial (NCT02325401). Metformin is also postulated to activate an antitumor immune response. Here, we investigate immunologic effects of metformin on natural killer (NK) and natural killer T cells, including results from two phase I open-label studies in patients with HNSCC treated with metformin (NCT02325401, NCT02083692).

Methods: Peripheral blood was collected before and after metformin treatment or from newly diagnosed patients with HNSCC. Peripheral immune cell phenotypes were evaluated using flow cytometry, cytokine expression by ELISA and/or IsoLight, and NK cell-mediated cytotoxicity was determined with a flow-based NK cell cytotoxicity assay (NKCA). Patient tumor immune infiltration before and after metformin treatment was analyzed with immunofluorescence. NK cells were treated with either vehicle or metformin and analyzed by RNA sequencing (RNA-seq). NK cells were then treated with inhibitors of significant pathways determined by RNA-seq and analyzed by NKCA, ELISA, and western blot analyses.

Results: Increased peripheral NK cell activated populations were observed in patients treated with metformin. NK cell tumor infiltration was enhanced in patients with HNSCC treated with metformin preoperatively. Metformin increased antitumorigenic cytokines ex vivo, including significant increases in perforin. Metformin increased HNSCC NK cell cytotoxicity and inhibited the CXCL1 pathway while stimulating the STAT1 pathway within HNSCC NK cells. Exogenous CXCL1 prevented metformin-enhanced NK cell-mediated cytotoxicity. Metformin-mediated NK cell cytotoxicity was found to be AMP-activated protein kinase independent, but dependent on both mechanistic target of rapamycin and pSTAT1.

Conclusions: Our data identifies a new role for metformin-mediated immune antitumorigenic function through NK cell-mediated cytotoxicity and downregulation of CXCL1 in HNSCC. These findings will inform future immunomodulating therapies in HNSCC.

Keywords: cytotoxicity, immunologic; head and neck neoplasms; immunity, cellular; immunotherapy; killer cells, natural.

Conflict of interest statement

Competing interests: None declared.

© Author(s) (or their employer(s)) 2022. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ.

Figures

Figure 1

Reduced NK cell populations and impaired IFN-γ production in peripheral blood of patients with locally advanced HNSCC. (A) Scatterplot representation of CD56+ NK cell populations in healthy control versus. (B) Patient with HNSCC. (C) Dot plots indicating per cent of CD56brightCD16–, CD56dimCD16+, and NKT cells for healthy controls versus patients with HNSCC. n=5 for controls, n=8 for patients with HNSCC (all human papilloma virus negative with advanced stage but heterogenous site of disease). (D) IFN-γ production of NK cells isolated from healthy controls primed with low dose IL-2, and activated with IL-12+IL-18. (E) IFN-γ production of NK cells isolated from patients with HNSCC primed with low dose IL-2, and activated with IL-12+IL-18. (F) Summary data of (C) and (D). Patients were heterogenous using two sites and three stages of HNSCC. Data in (C) were analyzed by an unpaired t-test with Welch’s correction; data in (D), (E), and (F) were analyzed by a paired Student’s t-test and two-way analysis of variance. HNSCC, head and neck are squamous cell carcinoma; IFN, interferon; IL, interleukin; NK, natural killer; NKT natural killer T cells.

Figure 2

Metformin treatment increases NK cell populations and increases expression of NKG2D in patients with locally advanced HNSCC. (A) Per cent positive for CD56brightCD16– and CD56dimCD16+ NK cell populations from patients with HNSCC before versus after metformin treatment. (B) Ratio of CD56dim/CD56bright NK cell populations in patients with HNSCC before versus after metformin treatment. (C) Per cent positive for CD56+CD3+ NKT cells from healthy controls and patients with HNSCC before and after metformin treatment. Horizontal black bars represent mean±SEM. (D) Per cent expression of NKG2D gated on CD56+CD3– NK cells from healthy controls and patients with HNSCC before and after metformin treatment. (E) Representative flow cytometry histograms showing NKG2D expression in CD56brightCD16– and CD56dimCD16+ NK cell populations from patients with HNSCC before versus after metformin treatment. (F) Mean fluorescence intensity of NKG2D measured in CD56brightCD16– (red) and CD56dimCD16+ (blue) NK cell populations from patients with HNSCC. n=5 for healthy controls, n=8 for patients with HNSCCs. Statistical significance indicated by p value. Data in (A) and (F) were analyzed by a paired Student’s t-test; data in (C) and (D) were analyzed by two-way analysis of variance (p=0.010 for (C) and p=0.05 for (D)). HNSCC, head and neck are squamous cell carcinoma; NK, natural killer.

Figure 3

Metformin increases infiltration, perforin production, and cytotoxicity of HNSCC NK cells. (A) Immunofluorescence images of HNSCC tumor biopsy before treatment and resection after metformin treatment. Red=pan-cytokeratin, pink=NKG2D, orange=CD3, and blue=DAPI nuclear stain. (B) Quantification of total and infiltrating NK cells within the stained tumor tissue. Patients sourced from Amin et al. (C) Bulk cytokine analysis by IsoLight 24 hours after treatment with vehicle or 12 mM metformin treatment. (D) Cytotoxicity of HNC NK cells co-cultured with target cells treated with 12 mM metformin for 24 hours. Target cells included the established cell lines, CAL-27 which are HPV–, and UMSCC47 which are HPV+, and primary cells are derived from matched HNC tumor tissue from patients. Peripheral NK cells are derived from blood while infiltrating NK cells are derived from tumor-infiltrating lymphocytes. (E) Cytotoxicity of HNC cells compared with healthy donor (NML). (B), (C) and (D) were analyzed by paired student’s t-test. (E) was analyzed by two-way analysis of variance. HNC, head and neck cancers; HNSCC, head and neck are squamous cell carcinoma; HPV, human papilloma virus; IFN, interferon; IL, interleukin; NK, natural killer.

Figure 4

Metformin reduces CXCL1, activates pSTAT1, and inhibits pSTAT3. (A) RNA sequencing heat map of patient NK cells treated with vehicle or 12 mM metformin for 24 hours resulted in four most significantly downregulated and three most significantly upregulated genes (p value1.5). Red indicates upregulated genes while blue indicates downregulated. (B) CXCL1 expression by ELISA in supernatant of HNC peripheral NK cells treated with vehicle or 12 mM metformin for 24 hours. (C) Cytotoxicity assay of HNC peripheral NK cells treated with 12 mM metformin, 50 ng CXCL1, or 10 µM CXCR2 inhibitor navarixin for 24 hours and subsequently co-cultured with CAL-27 target cells at 5:1 ratio. (D) Perforin in supernatant of experiment (C) analyzed by ELISA. (E) Western blot analysis of NK92 cells treated with 12 mM metformin, 50 ng CXCL1, or 10 µM CXCR2 inhibitor navarixin for 24 hours and probed for STAT1, STAT3, and actin. (F) Quantification of blot (E) using ImageJ. (B), (C), and (D) were analyzed by matched one-way analysis of variance. HNC, head and neck cancers; NK, natural killer.

Figure 5

Metformin-mediated NK cellular cytotoxicity is AMPK independent, pSTAT1 dependent. (A) NKCA of peripheral HNC NK cells treated with 12 mM metformin, 50 ng CXCL1, or 10 µM pSTAT1 inhibitor for 24 hours against target cells CAL27 at target to effector ratio 1:5. (B) Perforin ELISA of supernatant from (A). (C) NKCA of peripheral HNC NK cells treated with 12 mM metformin, 50 ng CXCL1, 10 µM pSTAT3 inhibitor, or 10 µM pSTAT1 inhibitor for 24 hours against target cells CAL27 at target to effector ratio 1:5. (D) Perforin ELISA of supernatant from (C). (E) NKCA of peripheral HNC NK cells treated with 12 mM metformin or 10 µM AMPK inhibitor for 24 hours against target cells CAL27 at target to effector ratio 1:5.(F) Perforin ELISA of supernatant from (E). (G) NKCA of peripheral HNC NK cells treated with 12 mM metformin, 10 µM mTOR activator, or 10 µM mTOR inhibitor for 24 hours against target cells CAL27 at target to effector ratio 1:5. (H) Perforin ELISA of supernatant from (G). All figures analyzed by matched one-way analysis of variance. AMPK, AMP-activated protein kinase; HNC, head and neck cancers; mTOR, mechanistic target of rapamycin; NKCA, natural killer cytotoxicity assay; NK, natural killer.

Figure 6

Proposed mechanism for metformin in head and neck cancer natural killer cells. Figure created in BioRender, license agreement YR246ADK5E.