Sphingosine kinase 1 contributes to the metastatic potential of epithelial ovarian cancer to the adipocyte-rich niche

Abstract

Unlike many solid tumors, epithelial ovarian cancer (EOC) has a clear metastatic predilection to the adipocyte-rich niche, especially the omentum. However, the underlying mechanism driving this process remains incomplete. Here we show that SphK1 is over-expressed in omental metastases compared with ovarian primary tumors in EOC patients. In vitro, inhibition of SphK1 suppressed the metastatic ability of EOC induced by adipocytes. In vivo, blockage of SphK1 could attenuate the omental metastasis of EOC. Importantly, SphK1 modulates adipocyte-induced E/N-cadherin switch through Twist1, a key process in EOC metastasis. Our study reveals a previously unrecognized role of SphK1 in modulating the metastatic tropism of EOC to the adipocyte-rich niche, suggesting a new target for EOC therapy.

To the editor,

Epithelial ovarian cancer (EOC) is the deadliest gynecological malignancy [1]. Unlike many solid tumors, EOC tends to metastasize to the adipocyte-rich microenvironment, especially the omentum [2, 3]. This metastatic tropism of EOC can lead to extensive pelvic and abdominal metastasis, which is the major cause of death in EOC patients. Therefore, identifying the mechanisms driving this process is crucial for disease-specific therapeutic intervention. Sphingosine kinase 1 (SphK1) is a key enzyme with a well-established role in the regulation of sphingolipid metabolism [4]. SphK1 catalyzes the phosphorylation of sphingosine to produce sphingosine-1-phosphate (S1P), which can act through S1P receptor (S1PR) pathway and intracellular pathway [4]. We previously showed that SphK1 was over-expressed in EOC tissue and was involved in EOC growth and angiogenesis [5,6,7]. We recently found adipocytes could activate SphK1 in EOC cells [6]. In this work, we tested the role of SphK1 in EOC metastasis induced by adipocyte-rich niche.

We first performed SphK1 immunohistochemistry (IHC) on paired samples from ovarian primary tumor and omental metastatic tissue of 20 high-grade serous ovarian cancer patients (Additional file 1: Table S1). Quantification confirmed that SphK1 expression was significantly increased in omental metastatic deposits compared with the primary tumors (Fig. 1a, b). Because adipocytes, major components of the omentum, promotes the metastasis of EOC [8], we next tested whether SphK1 was involved in EOC metastasis induced by adipocytes from human omentum (Additional file 2: Fig. S1A). Markedly, blockage of SphK1 by either siRNA or selective inhibitor PF543 significantly inhibited SKOV3 cell migration and invasion induced by adipocyte culture medium (CM) (Fig. 1c, d; Additional file 3: Fig. S2A–D). We used Hey cells, another EOC cell line, to confirm our results (Additional file 2: Fig. S1B–E, Additional file 3: Fig. S2A–D). To test whether SphK1 is required for the omental metastasis, we established xenograft mouse models of EOC by intraperitoneally injecting human EOC cell line SKOV3. Then the mice were injected with PF543 intraperitoneally twice a week. We found that the largest tumor is often formed in the omentum. PF543 treatment resulted in less tumor burden in the omentum and other metastatic sites compared with the control group (Fig. 1e, f). We confirmed this observation by measuring and calculating the tumor weight and the tumor number of metastatic nodules (Fig. 1g, h). We also used SphK1 knockdown EOC cells to confirm our results (Additional file 6: Fig. S5A–E). Together, these results suggested that SphK1 contributed to EOC metastasis to the adipocyte-rich niche.

Fig. 1

Inhibition of SphK1 suppressed EOC metastasis to the adipocyte-rich niche. a SphK1 IHC of paired samples from ovarian primary tumors and their corresponding metastatic omental deposits from 20 patients. Representative images are shown. The scale bar represents 60 μm. b Quantitative analysis of IHC staining for SphK1 of ovarian primary tumors and their corresponding metastatic omental deposits from 20 patients. **P < 0.01; two-tailed Student’s t test. C SKOV3 cells were transfected with negative control siRNA (si-Ctrl) or SphK1 siRNA (si-SphK1). 24 h after transfection, mRNA level of SphK1 was determined by quantitative reverse transcription polymerase chain reaction (qRT-PCR) and normalized to GAPDH. 48 h after transfection, protein level of SphK1 was determined by Western blot and normalized to GAPDH. Data are mean ± SD (n = 3). **P < 0.01; two-tailed Student’s t test. d Representative images of migration and invasion assays (200×). Transfected SKOV3 cells were serum starved overnight and metastasized towards serum free medium (SFM) or adipocyte CM. Migrated cells and invaded cells were photographed and calculated. Experiments have been carried out with adipocytes from 2 to 3 independent donors. Data are mean ± SD (n = 3). *P < 0.05; two-tailed Student’s t test. e Representative images of omentum in healthy mouse and representative images of EOC xenograft mouse model in the Ctrl and PF543 group. Black arrows indicated the location of the metastatic nodules. f Representative images of omental metastatic tumors in EOC mouse models of the Ctrl and PF543 group. g Quantification of tumor weight in different metastatic sites. Data are mean ± SD (n = 6). *P < 0.05 versus Ctrl group; ns, no significance; two-tailed Student’s t test. h Quantification of tumor number in different metastatic sites. Data are mean ± SD (n = 6). *P < 0.05 versus Ctrl group; ns, no significance; two-tailed Student’s t-test

Epithelial-mesenchymal transition (EMT) was recognized as a key process in tumor metastasis [9]. EMT was characterized by decreased expression of epithelial cadherin (E-cadherin) and increased expression of neural cadherin (N-cadherin), referred to as ‘E/N-cadherin switch’ [10]. In EOC, E/N-cadherin switch is associated with EMT and invasive phenotype acquisition [11]. Because adipocytes promoted EOC metastasis, we investigated whether adipocytes regulated E/N-cadherin switch. As expected, we found that adipocyte CM induced the gain of N-cadherin accompanied by the loss of E-cadherin in EOC (Fig. 2a). Having shown the function of SphK1 in adipocyte-induced metastasis of EOC, potential roles of SphK1 in adipocyte-induced E/N-cadherin switch was suggested. Indeed, SphK1 inhibition significantly attenuated the level change of E-cadherin and N-cadherin induced by adipocytes (Fig. 2a). These results indicated an important role of SphK1 in regulating adipocyte-induced E/N-cadherin switch. E-cadherin and N-cadherin expression is notably controlled by zinc-finger transcriptional factors, including Snail, Slug, Twist1 and ZEB [12]. Therefore, we further invested the role of adipocytes in these factors. Expression level of Twist1 (Fig. 2b), but not Snail, Slug, ZEB1 or ZEB2 (Additional file 4: Fig. S3A, B), was enhanced in EOC by adipocyte CM co-culture, which indicated that adipocyte may affect E/N-cadherin switch in EOC through Twist1 activation. Moreover, SphK1 blockage significantly attenuated the adipocyte-induced Twist1 expression in EOC (Fig. 2b). Furthermore, suppression of Twist1 by siRNA could significantly inhibit EOC cell migration and invasion induced by adipocyte CM (Additional file 5: Fig. S4A–D). In addition, SphK1 blockage enhanced E/N-cadherin switch and inhibited Twist1 expression in mouse omental metastases models (Fig. 2c, d; Additional file 6: Fig. S5F, G). Together, these results suggest that SphK1 is involved in adipocyte-induced Twist1 activation, which subsequently drives the E/N-cadherin switch.

Fig. 2

SphK1 modulates adipocyte-induced E/N-cadherin switch through Twist1. a SKOV3 and HEY cells were transfected with negative control siRNA or SphK1 siRNA. Cells were serum starved overnight and cultured in SFM or adipocyte CM for 48 h. Expression levels of SphK1, E-cadherin and N-cadherin were determined by Western blot and normalized to GAPDH. Densitometric analyses of E-cadherin and N-cadherin were shown on the right. Experiments have been carried out with adipocytes from 2 to 3 independent donors. Data are mean ± SD (n = 3). *P < 0.05 versus control; ^#P < 0.05 versus adipocyte CM alone; two-tailed Student’s t test. b SKOV3 and HEY cells were transfected with negative control siRNA or SphK1 siRNA. Cells were serum starved overnight and cultured in SFM or adipocyte CM for 48 h. Expression level of SphK1 and Twist1 was determined by Western blot and normalized to GAPDH. Densitometric analysis of Twist1 was shown on the right. Experiments have been carried out with adipocytes from 2 to 3 independent donors. Data are mean ± SD (n = 3). *P < 0.05 versus control; ^#P < 0.05 versus adipocyte CM alone; two-tailed Student’s t test. C IHC staining of E-cadherin, N-cadherin and Twist1 in the omental metastatic tumor tissue of mouse models. The scale bar represents 20 μm. Statistical analysis of integrated optical density (IOD)/area was shown on the right. Data are mean ± SD (n = 6). *P < 0.05 versus Ctrl group; two-tailed Student’s t-test. d Expression levels of SphK1, E-cadherin, N-cadherin and Twist1 in omental metastatic tumor tissue of mouse models were determined by Western blot and normalized to GAPDH. Densitometric analyses of E-cadherin, N-cadherin and Twist1 were shown on the right. Data are mean ± SD (n = 6). *P < 0.05 versus Ctrl group; two-tailed Student’s t test. e Model of adipocyte-induced omental metastasis of EOC cells. Adipocytes are capable of activating SphK1, leading to the release of S1P. SphK1/S1P signaling may subsequently activate Twist1 through S1P receptor (S1PR) pathway and/or intracellular pathway, which drives E/N-cadherin switch

In conclusion, our study provided mechanistic data delineating a new mechanism that mediated the metastatic potential of EOC to the adipocyte-rich niche via SphK1 signaling (Fig. 2e), suggesting a new target for EOC therapy.