SPHK1 Is a Novel Target of Metformin in Ovarian Cancer

Abstract

The role of phospholipid signaling in ovarian cancer is poorly understood. Sphingosine-1-phosphate (S1P) is a bioactive metabolite of sphingosine that has been associated with tumor progression through enhanced cell proliferation and motility. Similarly, sphingosine kinases (SPHK), which catalyze the formation of S1P and thus regulate the sphingolipid rheostat, have been reported to promote tumor growth in a variety of cancers. The findings reported here show that exogenous S1P or overexpression of SPHK1 increased proliferation, migration, invasion, and stem-like phenotypes in ovarian cancer cell lines. Likewise, overexpression of SPHK1 markedly enhanced tumor growth in a xenograft model of ovarian cancer, which was associated with elevation of key markers of proliferation and stemness. The diabetes drug, metformin, has been shown to have anticancer effects. Here, we found that ovarian cancer patients taking metformin had significantly reduced serum S1P levels, a finding that was recapitulated when ovarian cancer cells were treated with metformin and analyzed by lipidomics. These findings suggested that in cancer the sphingolipid rheostat may be a novel metabolic target of metformin. In support of this, metformin blocked hypoxia-induced SPHK1, which was associated with inhibited nuclear translocation and transcriptional activity of hypoxia-inducible factors (HIF1α and HIF2α). Further, ovarian cancer cells with high SPHK1 were found to be highly sensitive to the cytotoxic effects of metformin, whereas ovarian cancer cells with low SPHK1 were resistant. Together, the findings reported here show that hypoxia-induced SPHK1 expression and downstream S1P signaling promote ovarian cancer progression and that tumors with high expression of SPHK1 or S1P levels might have increased sensitivity to the cytotoxic effects of metformin. IMPLICATIONS: Metformin targets sphingolipid metabolism through inhibiting SPHK1, thereby impeding ovarian cancer cell migration, proliferation, and self-renewal.

Figure 1.. Metformin regulates the sphingolipid rheostat and represses S1P-driven migration in OvCa.

(A) Schematic of the key regulatory steps regulating the sphingolipid rheostat and the resultant bioactive lipids. (B) S1P levels in serum of age matched patients with stage III/IV OvCa using metformin for diabetes (n=9) compared to control counterparts (n=10). (C) Lipidomic profiling measuring ceramide, sphingosine, and S1P concentration in CAOV3 OvCa cells after metformin treatment (1mM, 72h) (n=3). (D) Lipidomic profiling showing the effects of metformin on DH-S1P, DH-ceramide and DH-sphingosine in CAOV3 OvCa cells (n=3). (E) Wound healing assay in TYKnu and CAOV3 cells with exogenous treatment with S1P (10 and 100nM) following metformin pretreatment (1mM, 24 hours) (n=6). (F) Transwell invasion assay in HeyA8 OvCa cells using S1P (100nM in serum-free media) as a chemoattractant with concurrent metformin treatment (1mM, 15h) (n=3). FBS (10%) as a chemoattractant was used as a positive control. Data represent mean value ± S.D. *p < 0.05, **p < 0.01, ***p < 0.001, ****p< 0.0001 of one-way ANOVA with post-hoc Tukey test. N.S.; not significant.

Figure 2.. SPHK1 is a novel target of metformin.

(A) SPHK1 mRNA expression was measured by qRT-PCR in TYKnu (left) and CAOV3 (right) OvCa cells treated with metformin or phenformin at the indicated doses for 24h. Results expressed as relative quantity over GAPDH housekeeping gene (n=3) (B) SPHK1 protein expression was assessed by immunoblot in TYKnu (left) and CAOV3 (right) cells treated with metformin or phenformin at the indicated doses for 72h, (n=3). (C) Wound healing assay in OVCAR5 cells stably overexpressing SPHK1 (SPHK1-OE) (n=6). (D) MTT assay measuring proliferation in SPHK1 overexpressing and control transfected OVCAR5 cells (n=6). (E) Immunoblot of AKT activation (S473 phosphorylation) in SPHK1 overexpressing and control transfected OVCAR5 cells (n=3). (F) Colony formation assay showing number of colonies formed after 10 days in SPHK1 overexpressing and control transfected OVCAR5 cells (n=3). (G) SOX2 mRNA expression was measured by qRT-PCR in OVCAR5 cells overexpressing SPHK1 or control vector. Results expressed as relative quantity over GAPDH housekeeping gene (n=3). (H) SOX2 protein expression was assessed by immunoblot in OVCAR5 cells overexpressing SPHK1 or control vector. (I) Xenograft mouse model with OVCAR5 cells overexpressing of SPHK1 was compared to vector control OVCAR5 cells (Control n=9, SPHK1-OE n=10). X-axis is expressed as fold change in mean tumor weight (g) over control. IHC analysis of tumors from SPHK1 overexpressing group and control group measuring expression of Ki67 (J) and SOX2 (K). Data represent mean value ± S.D. *p < 0.05, **p < 0.01, ***p < 0.001, ****p< 0.0001 of one-way ANOVA with post-hoc Tukey test. N.S.; not significant.

Figure 3.. Metformin inhibits hypoxia-induced expression of SPHK1 in OvCa.

(A) SPHK1 protein expression was assessed by immunoblot in OVCAR5, Kuramochi and TYKnu cells overexpressing constitutively stable HIF1α with P402A/P564A (left) or HIF2α with P405A/P531A (right) mutations (n=3). (B) SPHK1 protein expression was analyzed by immunoblot after metformin treatment (1mM, 72h) in OVCAR5 and Kuramochi cells overexpressing constitutively stable HIF1α with activating substitution P402A/P564A (left) or HIF2α with P405A/P531A (right) (n=3). (C) HIF1α, HIF2α and SPHK1 protein expression were assessed by immunoblot in response to hypoxia (1% O₂, 5% CO₂) with metformin concurrent treatment (1mM, 48h) (n=3). (D) HIF1α and HIF2α protein expression and localization in OVCAR5 cells was measured by immunoblot following cellular fractionation. HIF expression was stabilized using CoCl₂ (100μM) and concurrently treated with metformin (1mM, 24h). Histone H3 and βTubulin were used to normalize nuclear and cytosolic localization, respectively (n=3). (E) HIF1α and HIF2α protein expression and localization in OVCAR5 cells was measured by immunoblot following cellular fractionation following transfection with HIF1α or HIF2α constitutively stable mutants and subsequent treatment with metformin (1mM, 24h). Histone H3 and βTubulin were used to confirm nuclear and cytosolic fractions. (F) HIF transcriptional activity was measured in HeyA8 cells transiently co-transfected with the 5HRE reporter system in conjunction with either constitutively active HIF1α or HIF2α plasmids. Cells were then treated ± metformin (1mM, 24h) (n=3). Data represent mean value ± S.D. *p < 0.05, **p < 0.01 of one-way ANOVA with post-hoc Tukey test. N.S.; not significant.

Figure 4.. SPHK1 expression is associated with metformin sensitivity in OvCa cells.

(A) MTT viability assay of six OvCa cell lines analyzing sensitivity to metformin using a dose range up to 40mM in 5mM glucose DMEM (n=6). SPHK1 and SGPL1 mRNA (B) and protein (C) expression was assessed in multiple OvCa cell lines (n=3). (D) MTT viability assay IC₅₀ values for metformin (mM) in TYKnu, OVCAR5 and Kuramochi cells overexpressing SPHK1 or control vector (n=6). (E) Expression of cleaved caspase 3 was assessed through immunoblot in OVCAR5 and Kuramochi cells overexpressing SPHK1 or control vector with concurrent treatment with metformin (1mM, 48h) (n=3). (F) MTT viability assay IC₅₀ values for metformin (mM) in TYKnu cells with transient knockdown of SPHK1 using siRNA, in either normoxia or hypoxia (1% O₂, 5% CO₂) (n=6). (G) Knockdown of SPHK1 mRNA expression was confirmed by qRT-PCR following knockdown using siRNA in HeyA8 cells (n=3). (H) Western blot of SPHK1 and cleaved caspase 3 in TYKnu cells following knockdown of SPHK1 using siRNA, with or without metformin treatment (1mM, 48h) (n=3). Data represent mean value ± S.D.

Figure 5.. Metformin inhibits SPHK1-dependent tumorigenicity.

(A) IHC staining of tumor microarrays of common anatomical sites in OvCa for expression of SPHK1 (n=389). Representative images (20x magnification). Scale bar represents 100μm. (B) Colony formation assay showing number of colonies formed in 10 days in SPHK1 overexpressing and control transfected OVCAR5 (left) or Kuramochi (right) cells in the presence of metformin (1mM) (n=3). (C) Colony formation assay over 10 days in TYKnu cells with SPHK1 knockdown using siRNA, with or without metformin treatment (1mM) (n=3). (D) SOX2 and Snail protein expression were assessed by immunoblot in OVCAR5 and Kuramochi cells overexpressing SPHK1 or control vector with metformin treatment (1mM, 72h) (n=3). (E) Transwell invasion assay in HeyA8 cells with transient knockdown of SPHK1 using siRNA. Cells were treated with hypoxia (1% O₂) in the presence or absence of metformin (1mM) during the invasion assay. 10% FBS was used as a chemoattractant below the transwell inserts, and cells were allowed to invade for 16h following seeding. (F) Schematic depicting inhibition of SPHK1 by metformin, resulting in a shift away from tumorigenic S1P signaling. Data represent mean value ± S.D. *p < 0.05, **p < 0.01, ***p < 0.001 of Student’s t Test, or one-way ANOVA with post-hoc Tukey test.