Serine hydroxymethyl transferase 1 stimulates pro-oncogenic cytokine expression through sialic acid to promote ovarian cancer tumor growth and progression

Abstract

High-grade serous (HGS) ovarian cancer accounts for 90% of all ovarian cancer-related deaths. However, factors that drive HGS ovarian cancer tumor growth have not been fully elucidated. In particular, comprehensive analysis of the metabolic requirements of ovarian cancer tumor growth has not been performed. By analyzing The Cancer Genome Atlas mRNA expression data for HGS ovarian cancer patient samples, we observed that six enzymes of the folic acid metabolic pathway were overexpressed in HGS ovarian cancer samples compared with normal ovary samples. Systematic knockdown of all six genes using short hairpin RNAs (shRNAs) and follow-up functional studies demonstrated that serine hydroxymethyl transferase 1 (SHMT1) was necessary for ovarian cancer tumor growth and cell migration in culture and tumor formation in mice. SHMT1 promoter analysis identified transcription factor Wilms tumor 1 (WT1) binding sites, and WT1 knockdown resulted in reduced SHMT1 transcription in ovarian cancer cells. Unbiased large-scale metabolomic analysis and transcriptome-wide mRNA expression profiling identified reduced levels of several metabolites of the amino sugar and nucleotide sugar metabolic pathways, including sialic acid N-acetylneuraminic acid (Neu5Ac), and downregulation of pro-oncogenic cytokines interleukin-6 and 8 (IL-6 and IL-8) as unexpected outcomes of SHMT1 loss. Overexpression of either IL-6 or IL-8 partially rescued SHMT1 loss-induced tumor growth inhibition and migration. Supplementation of culture medium with Neu5Ac stimulated expression of IL-6 and IL-8 and rescued the tumor growth and migratory phenotypes of ovarian cancer cells expressing SHMT1 shRNAs. In agreement with the ovarian tumor-promoting role of Neu5Ac, treatment with Neu5Ac-targeting glycomimetic P-3Fax-Neu5Ac blocked ovarian cancer growth and migration. Collectively, these results demonstrate that SHMT1 controls the expression of pro-oncogenic inflammatory cytokines by regulating sialic acid Neu5Ac to promote ovarian cancer tumor growth and migration. Thus, targeting of SHMT1 and Neu5Ac represents a precision therapy opportunity for effective HGS ovarian cancer treatment.

Figure 1

SHMT1 is necessary for ovarian cancer tumor growth. (a) Schematic diagram showing the functional validation pipeline for identifying the metabolic genes promoting ovarian cancer tumor growth. (b) The indicated ovarian cancer cell lines expressing control nonspecific (NS) or SHMT1 shRNAs were analyzed for their ability to grow in an anchorage-independent manner in a soft-agar assay. Representative soft-agar assay images are shown (left) and relative colony numbers and average colony size are shown (right). (c) The indicated ovarian cancer cell lines expressing NS or SHMT1 shRNAs were subcutaneously injected into the flanks of athymic nude mice and analyzed for tumor-forming ability. Average tumor volumes at indicated time points (n=5) are shown for ovarian cancer cell lines PEO4, COV504 and COV413B. Data are presented as mean±s.e.m.; *P<0.05 and **P<0.005.

Figure 2

SHMT1 knockdown inhibits ovarian cancer cell migration. (a) PEO4 cells expressing control nonspecific (NS) or SHMT1 shRNAs were analyzed for invasive ability using Matrigel-based Boyden chamber assay. Representative images are shown (left) and quantitation is presented (right). (b) PEO4 cells expressing the indicated shRNAs were analyzed for migratory potential in a wound-healing assay. (Left) Representative images at days 0 and 4 are shown. (Right) Percent migration relative to day 0 is plotted. Data are presented as mean±s.e.m.; **P<0.005.

Figure 3

Transcription factor WT1 regulates SHMT1 transcription in ovarian cancer cells. (a) RT–qPCR analysis of SHMT1 mRNA expression and immunoblot analysis of WT1 and SHMT1 levels in the indicated ovarian cancer cell lines expressing either control nonspecific (NS) or WT1 shRNAs. SHMT1 mRNA level was normalized to that of β-ACTIN, and β-ACTIN served as a loading control on blots. (b) PEO4 cells expressing NS or WT1 shRNAs were analyzed for WT1 protein enrichment on SHMT1 promoter using chromatin immunoprecipitation assay. β-ACTIN and GAPDH promoter regions were used as negative controls. Percent WT1 enrichment relative to input for indicated conditions for each promoter locus (SHMT1, β-ACTIN and GAPDH) is shown. (c) PEO4, COV504, and COV413 cells expressing indicated NS or WT1 shRNAs were analyzed for ability to grow in an anchorage-independent manner in a soft-agar assay. Representative soft-agar assay images are shown on left and relative colony numbers and average colony size are shown on right. (d) PEO4 cells expressing NS or WT1 shRNAs with either an empty vector or SHMT1 cDNA were analyzed for the ability to grow in an anchorage-independent manner in a soft-agar assay. Representative soft-agar assay images are shown on the left, and relative colony numbers and average colony size are shown on the right. Data are presented as mean±s.e.m.; **P<0.005.

Figure 4

SHMT1 loss results in reduced amino sugar and nucleotide sugar metabolic pathway intermediates in ovarian cancer cells. (a) Key steps of the amino sugar and nucleotide sugar metabolic pathway. (b) Relative concentrations of indicated metabolites in SHMT1 knockdown PEO4 cells compared with cells expressing nonspecific (NS) shRNA.

Figure 5

SHMT1 loss results in reduced levels of pro-oncogenic inflammatory cytokines that are necessary for SHMT1-mediated ovarian cancer growth. (a–c) PEO4 (a), COV504 (b) and COV413B (c) cells expressing control nonspecific (NS) or SHMT1 shRNAs were analyzed for the expression of IL-6 and IL-8 mRNA by RT–qPCR (left) and of IL-6, IL-8, SHMT1 and β-ACTIN by immunoblotting (right). (d) PEO4 cells expressing NS or SHMT1 shRNAs alone or expressing IL-6 or IL-8 cDNA were analyzed for anchorage-independent growth in a soft-agar assay. Representative images under indicated conditions (top) and colony number and size relative to those of cells expressing NS shRNA (bottom) are shown. (e) PEO4 cells expressing NS or SHMT1 shRNAs alone or expressing IL-6 or IL-8 cDNA were injected subcutaneously into the flanks of athymic nude mice. Average tumor volumes (n=5) at indicated times are shown. Data are presented as mean±s.e.m.; *P<0.05 and **P<0.005.

Figure 6

Neu5Ac supplementation stimulates IL-6 and IL-8 and rescues growth of SHMT1 knockdown ovarian cancer cells (a) and (b) PEO4 cells expressing indicated shRNAs without or with Neu5Ac were analyzed for expression of IL-6 and IL-8 mRNA by RT–qPCR (a) and of IL-6, IL-8, SHMT1 and β-ACTIN by immunoblotting (b). (c) PEO4 cells expressing indicated shRNAs without or with Neu5Ac were tested for the ability to grow in an anchorage-independent manner in a soft-agar assay. Representative images (top) and colony number and size relative to those of cells expressing nonspecific (NS) shRNA (bottom) are shown. (d) PEO4 cells expressing indicated shRNAs without or with Neu5Ac were tested for their ability to migrate in wound-healing assay. Migration on day 4 relative to that on day 0 under indicated conditions is shown. (e) Indicated ovarian cancer cells were treated with vehicle (1 × PBS) or Neu5Ac-targeting glycomimetic P-3Fax-Neu5Ac and analyzed for expression of IL-6 and IL-8 mRNA by RT–qPCR. Expression relative to that in control vehicle-treated cells is shown (f). Indicated ovarian cancer cells were treated with vehicle or P-3Fax-Neu5Ac and analyzed for ability to grow in an anchorage-independent manner in a soft-agar assay. Representative soft-agar assay images for indicated cell lines treated with vehicle or P-3Fax-Neu5Ac are shown (left) and relative colony number and average colony size are shown (right). Data are presented as mean±s.e.m.; *P<0.05.

Figure 7

Model. SHMT1, by regulating nucleotide and amino sugar metabolism, regulates pro-oncogenic inflammatory cytokine expression through the sialic acid Neu5Ac, which in turn facilitates ovarian cancer cell growth and migration.