Sphingomyelin Synthase 1 (SMS1) Downregulation Is Associated With Sphingolipid Reprogramming and a Worse Prognosis in Melanoma

Abstract

Sphingolipid (SL) metabolism alterations have been frequently reported in cancer including in melanoma, a bad-prognosis skin cancer. In normal cells, de novo synthesized ceramide is mainly converted to sphingomyelin (SM), the most abundant SL, by sphingomyelin synthase 1 (SMS1) and, albeit to a lesser extent, SMS2, encoded by the SGMS1 and SGMS2 genes, respectively. Alternatively, ceramide can be converted to glucosylceramide (GlcCer) by the GlcCer synthase (GCS), encoded by the UGCG gene. Herein, we provide evidence for the first time that SMS1 is frequently downregulated in various solid cancers, more particularly in melanoma. Accordingly, various human melanoma cells displayed a SL metabolism signature associated with (i) a robust and a low expression of UGCG and SGMS1/2, respectively, (ii) higher in situ enzyme activity of GCS than SMS, and (iii) higher intracellular levels of GlcCer than SM. SMS1 was expressed at low levels in most of the human melanoma biopsies. In addition, several mutations and increased CpG island methylation in the SGMS1 gene were identified that likely affect SMS1 expression. Finally, low SMS1 expression was associated with a worse prognosis in metastatic melanoma patients. Collectively, our study indicates that SMS1 downregulation in melanoma enhances GlcCer synthesis, triggering an imbalance in the SM/GlcCer homeostasis, which likely contributes to melanoma progression. Evaluating SMS1 expression level in tumor samples might serve as a biomarker to predict clinical outcome in advanced melanoma patients.

Keywords: cancer; ceramide; glucosylceramide; prognosis biomarker; sphingolipids.

Figure 1

Sphingomyelin synthase 1 (SMS1) is frequently downregulated in melanoma. (A) cDNA samples isolated from normal (N) and tumor (T) tissues from the same patient were compared. Expression of SGMS1 (left panel) and ubiquitin (right panel). (B) The SGMS1 expression was normalized to ubiquitin and expressed for each pair in normal skin and melanoma samples. (C) SGMS1 expression was analyzed in 3 different cohorts from Oncomine in normal Skin (n = 4), primary (PM; n = 14), and metastatic (MM; n = 39) melanoma (Ricker’s cohort) (left panel); in nevus (n = 9), primary (PM; n = 6), and metastatic (MM; n = 19) melanoma (Haqq’s cohort) (middle panel); in nevus (n = 18) and primary melanoma (PM; n = 45) (Talantov’s cohort) (right panel). (D) The expression of SGMS1 was analyzed in various cancer type cohorts from cbioportal. (E) The expression of UGCG, SGMS1, and SGMS2 was analyzed in melanoma samples from the TCGA metastatic melanoma patients (n = 342). (F) A set of melanoma cell lines (n = 10) was analyzed for the expression of UGCG, SGMS1, and SGMS2 by RT-qPCR (n = 10). Data from at least two independent experiments are means ± SEM. ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001, and ^∗∗∗p < 0.0001.

Figure 2

Sphingomyelin synthase 1 downregulation is associated with a worse prognosis in advanced melanoma patients. (A,B) A set of melanoma cell lines (n = 10) was analyzed for SLs by mass spectrometry (A) and GCS and SMS enzyme activities (B). Data from one experiment representative of three independent experiments are means ± SEM. (C) SGMS1 expression in melanoma samples from the TCGA melanoma cohort (n = 342) (left panel) and overall survival of patients exhibiting low (n = 68), medium (n = 206), and high (n = 68) SGMS1 expression (right panel). Cox model: SGMS1^low (Reference), SGMS1^medium: HR = 0.62 [95% CI = 0.44; 0.88] p = 0.007; SGMS1^high: HR = 0.48 [95% CI 0.31; 0.76] p = 0.002. ^∗p < 0.05, ^∗∗p < 0.01, and ^∗∗∗p < 0.001.