TSPYL5 SNPs: association with plasma estradiol concentrations and aromatase expression

Abstract

We performed a discovery genome-wide association study to identify genetic factors associated with variation in plasma estradiol (E2) concentrations using DNA from 772 postmenopausal women with estrogen receptor (ER)-positive breast cancer prior to the initiation of aromatase inhibitor therapy. Association analyses showed that the single nucleotide polymorphisms (SNP) (rs1864729) with the lowest P value (P = 3.49E-08), mapped to chromosome 8 near TSPYL5. We also identified 17 imputed SNPs in or near TSPYL5 with P values < 5E-08, one of which, rs2583506, created a functional estrogen response element. We then used a panel of lymphoblastoid cell lines (LCLs) stably transfected with ERα with known genome-wide SNP genotypes to demonstrate that TSPYL5 expression increased after E2 exposure of cells heterozygous for variant TSPYL5 SNP genotypes, but not in those homozygous for wild-type alleles. TSPYL5 knockdown decreased, and overexpression increased aromatase (CYP19A1) expression in MCF-7 cells, LCLs, and adipocytes through the skin/adipose (I.4) promoter. Chromatin immunoprecipitation assay showed that TSPYL5 bound to the CYP19A1 I.4 promoter. A putative TSPYL5 binding motif was identified in 43 genes, and TSPYL5 appeared to function as a transcription factor for most of those genes. In summary, genome-wide significant SNPs in TSPYL5 were associated with elevated plasma E2 in postmenopausal breast cancer patients. SNP rs2583506 created a functional estrogen response element, and LCLs with variant SNP genotypes displayed increased E2-dependent TSPYL5 expression. TSPYL5 induced CYP19A1 expression and that of many other genes. These studies have revealed a novel mechanism for regulating aromatase expression and plasma E2 concentrations in postmenopausal women with ER(+) breast cancer.

Trial registration: ClinicalTrials.gov NCT00283608.

Figures

Figure 1.

A, Manhattan Plot of GWAS for plasma E2 Concentrations. The figure shows the results of conditional regression analysis adjusted for eight eigenvectors. B, The chromosome 8 region surrounding the TSPYL5 gene from the Manhattan plot showing a plot of −log10 (P value) for observed (circles), imputed (diamonds), as well as imputed and then genotyped (red circles) SNPs. The location of the TSPYL5 gene is also indicated. C, A Q-Q plot is shown of observed vs expected results under the null hypothesis of no SNP association with plasma E2 concentrations.

Figure 2.

A, Box Plots Of Log Plasma E2 Concentrations Expressed as Picograms per mL by rs1864729 SNP Genotype. The horizontal lines within each box show median values, whereas the red points are mean values for homozygous wild type (Wt/Wt), heterozygous (Wt/V), and homozygous variant SNP (V/V) SNP genotypes. Extreme values are also shown. The number of subjects in the study with each genotype is also listed. B, ChIP assay using ERα stably transfected LCLs with known genotypes for the rs2583506 SNP. Lanes 1–4 and 8–11 are PCR products for DNA that was immunoprecipitated with human ERα antibody. The “input” material for the variant and WT heterozygous (Wt/V) and WT homozygous (Wt/Wt) samples were PCR amplification products from pools of sheared DNA for the entire genome. Lanes 1–6 are ChIP assays for one Wt/Wt and Wt/V LCL pair, and lanes 8–13 are ChIP assays for another pair of LCLs. Lanes 7 and 14 are Invitrogen (Carlsbad, CA) 1-kb DNA ladders.

Figure 3.

A, Schematic diagrams of promoters for the CYP19A1 gene. The relative tissue specificity for each promoter is listed. The skin/adipose I.4 promoter is highlighted in red. B, TSPYL5 and CYP19A1 skin/adipose tissue promoter mRNA expression levels after TSPYL5 overexpression (OE) and knockdown (KD) in MCF-7 cells. *. P < .05; **, P < .01; and ***, P < .001 when compared with baseline. C, Western blot analyses of TSPYL5 and CYP19A1 protein levels after TSPYL5 was knocked down (KD) or overexpressed (OE) in MCF-7 cells. CON, control.

Figure 4.

SNP-Dependent Differences in TSPYL5 Expression in Response to Increasing Concentrations of E2 for Three LCLs Heterozygous (Wt/V, blue lines) for TSPYL5 SNPs, and Three LCLs Homozygous for Wt SNP Alleles (Wt/Wt, red lines). All LCLs were stably transfected with ERα. The upper left panel shows relative TSPYL5 mRNA expression, the upper right panel shows relative mRNA expression level for CYP19A1 due to the skin/adipose promoter, the lower left panel shows relative mRNA expression level for CYP19A1 due to the adipose/breast cancer promoter, and the lower right panel shows relative mRNA expression level for CYP19A1 due to the placental promoter. *. P < .05; **, P < .01 when compared with the alternative genotype at the same E2 concentration.

Figure 5.

A, TSPYL5 and CYP19A1 mRNA expression levels after TSPYL5 overexpression (OE) (red bars) and knockdown (KD) (blue bars) in HO3sD-96Wa human adipocytes. B, Western blot analysis of TSPYL5 and CYP19A1 protein levels after TSPYL5 OE or KD in HO3sD-96Wa human adipocytes. *, P < .05; **, P < .01. Neg, negative.

Figure 6.

ChIP Assays in MCF-7 Cells Using TSPYL5 Antibody. A, The upper panel shows PCR amplification after ChIP assay using different primer pairs within the skin/adipose promoter region of the CYP19A1 gene in MCF-7 cells. B, The lower panel shows qPCR quantification of the same ChIP assay shown in panel A. Ab, Antibody; w/o, without.

Figure 7.

Effect of TSPYL5A Overexpression (OE) and Knockdown (KD)on the Relative mRNA Expression of Genes That Contain the Putative TSPYL5 DNA Binding Motif in (A) MCF-7, (B) IMR-90, and (C) HEK293T Cells. Red bars represent data for TSPYL5 overexpression and blue bars represent data for TSPYL5 knockdown. The blue dash lines represent control levels of expression.

Figure 8.

ChIP Assays That Demonstrate the Binding of TSPYL5 to Areas of Genes That Contained the Putative TSPYL5 DNA Binding Motif. ChIP was performed in MCF-7 cells with anti-TSPYL5 antibody. In each case the top panel shows the results of qRT-PCR demonstrating the relative enrichment of TSPYL5 DNA motif sequences after immunoprecipitation with anti-TSPYL5 antibody (blue) as compared with no antibody (IgG only) (orange). Sequences of primers utilized to perform these assays are listed in Supplemental Table 2. At the bottom of each figure are ChIP results after PCR amplification shown relative to input on a 1.0% agarose gel. Some genes contained more than one TSPYL5 DNA binding motif, so they were labeled as “-1”, “-2”, “-3” etc.