Evidence that Gsta4 modifies susceptibility to skin tumor development in mice and humans

Abstract

Background: The incidence of nonmelanoma skin cancer (NMSC) is equivalent to that of all other cancers combined. Previously, we mapped the 12-O-tetradecanoylphorbol-13-acetate (TPA) skin tumor promotion susceptibility locus, Psl1, to distal chromosome 9 in crosses of sensitive DBA/2 mice with relatively resistant C57BL/6 mice. Here, we used the mouse two-stage skin carcinogenesis model to identify the gene(s) responsible for the effects of Psl1.

Methods: Interval-specific congenic mouse strains (n ≥ 59 mice per strain) were used to more precisely map the Psl1 locus. Having identified glutathione S-transferase α4 (Gsta4) as a candidate tumor promotion susceptibility gene that mapped within the delimited region, we analyzed Gsta4-deficient mice (n = 62) for susceptibility to skin tumor promotion by TPA. We used quantitative polymerase chain reaction, western blotting, and immunohistochemistry to verify induction of Gsta4 in mouse epidermis following TPA treatment and biochemical assays to associate Gsta4 activity with tumor promotion susceptibility. In addition, single-nucleotide polymorphisms (SNPs) in GSTA4 were analyzed in a case-control study of 414 NMSC patients and 450 control subjects to examine their association with human NMSC. Statistical analyses of tumor studies in mice were one-sided, whereas all other statistical analyses were two-sided.

Results: Analyses of congenic mice indicated that at least two loci, Psl1.1 and Psl1.2, map to distal chromosome 9 and confer susceptibility to skin tumor promotion by TPA. Gsta4 maps to Psl1.2 and was highly induced (mRNA and protein) in the epidermis of resistant C57BL/6 mice compared with that of sensitive DBA/2 mice following treatment with TPA. Gsta4 activity levels were also higher in the epidermis of C57BL/6 mice following treatment with TPA. Gsta4-deficient mice (C57BL/6.Gsta4(-/-) mice) were more sensitive to TPA skin tumor promotion (0.8 tumors per mouse vs 0.4 tumors per mouse in wild-type controls; difference = 0.4 tumors per mouse; 95% confidence interval = 0.1 to 0.7, P = .007). Furthermore, inheritance of polymorphisms in GSTA4 was associated with risk of human NMSC. Three SNPs were found to be independent predictors of NMSC risk. Two of these were associated with increased risk of NMSC (odds ratios [ORs] = 1.60 to 3.42), while the third was associated with decreased risk of NMSC (OR = 0.63). In addition, a fourth SNP was associated with decreased risk of basal cell carcinoma only (OR = 0.44).

Conclusions: Gsta4/GSTA4 is a novel susceptibility gene for NMSC that affects risk in both mice and humans.

Figure 1

Susceptibility of C57BL.Psl1^dba congenic mouse strains to 12-O-tetradecanoylphorbol-13-acetate (TPA) skin tumor promotion. A) Map of the distal end of mouse chromosome 9. Microsatellite markers and the position of glutathione S-transferase α4 (Gsta4) are shown above the chromosomal map. The black lines below the chromosome represent regions of chromosome 9 that were introgressed from DBA/2 mice onto the C57BL/6 genetic background to generate the four subcongenic strains. Gray lines represent regions that were inherited either from C57BL/6 or DBA/2 mice. B) Time course of tumor development in C57BL.Psl1^dba congenic mice after promotion of skin tumors using TPA. Skin tumors in shaved female mice (C57BL/6, n = 60; C57BL/6.Psl1A^dba, n = 60; C57BL/6.Psl1B^dba, n = 59; C57BL/6.Psl1E^dba, n = 60; C57BL/6.Psl1F^dba, n = 62) were initiated by topical application of 2.5 μmol N-methyl-N′-nitro-N-nitrosoguanidine over the shaved dorsal skin. For tumor promotion, the mice were administered topical applications of 13.6 nmol TPA as a tumor promoter twice weekly. Tumors were counted weekly by palpation and tumor multiplicity was determined by dividing the total number of tumors by the number of mice at risk when the first tumor was observed. Similar results were obtained from two separate experiments and have been combined. The 95% confidence interval of the mean tumors per mouse is displayed for weeks 36–40. The data were analyzed using a one-sided Mann–Whitney U test.

Figure 2

Expression of glutathione S-transferase α4 (Gsta4) mRNA in the epidermis of various inbred mouse strains following treatment with 12-O-tetradecanoylphorbol-13-acetate (TPA). A) Time course of Gsta4 mRNA accumulation in TPA promotion resistant C57BL/6 mice vs sensitive DBA/2 mice. Groups of three female C57BL/6 or DBA/2 mice were treated once topically with 6.8 nmol TPA or acetone (vehicle) and killed by cervical dislocation at the indicated time points. The dorsal skin was removed and total RNA was harvested from epidermal scrapings from individual mice. The levels of Gsta4 and Hras1 mRNA were assessed by quantitative polymerase chain reaction (qPCR). For all qPCR experiments, Gsta4 mRNA levels were normalized to Hras1 expression and presented as arbitrary units (AU). Means and 95% confidence intervals are presented. B) Gsta4 mRNA levels in C57BL/6.Psl1A^dba mice (DBA/2 allele of Gsta4) vs C57BL/6.Psl1F^dba mice (C57BL/6 allele of Gsta4) following treatment with TPA. Groups of three female C57BL/6.Psl1A^dba or C57BL/6.Psl1F^dba mice were topically treated once with 6.8 nmol TPA or acetone and killed 18 hours later. Epidermal mRNA was harvested from the dorsal skin of individual mice and analyzed for Gsta4 message level by qPCR as described in (A). These experiments were performed in duplicate with similar results. All P values were based on two-sided Student t tests.

Figure 3

Glutathione S-transferase α4 (Gsta4) protein expression and enzymatic activity in mouse skin following treatment with 12-O-tetradecanoylphorbol-13-acetate (TPA). A) Gsta4 protein levels in C57BL/6 vs DBA/2 epidermis following TPA treatment. Groups of three female C57BL/6 or DBA/2 mice were treated topically once with 6.8 nmol TPA or acetone. Mice were killed at the indicated time points and dorsal skin was removed. Epidermal cells were harvested by scraping over a chilled glass plate and then placed in radioimmunoprecipitation assay buffer before being homogenized using an 18-gauge needle and syringe. The homogenates were cleared by centrifugation and the supernatant was analyzed for protein content. Then, equal amounts of supernatant protein were subjected to western blot analysis of Gsta4 expression. Actin protein levels were also probed as a loading control. These results are representative of three independent studies. B) Immunohistochemical staining of Gsta4 in skin sections from DBA/2 and C57BL/6 mice. Mice were treated twice weekly for 2 weeks with either 6.8 nmol TPA or acetone and killed 24 hours after the final treatment. Dorsal skin was removed and samples were formalin fixed and embedded in paraffin before staining using hemotoxylin counterstain (blue) and a rabbit polyclonal anti-Gsta4 antibody and an horseradish peroxidase–conjugated anti-rabbit secondary antibody. The peroxidase reaction was assayed using diaminobenzidine as the chromagen (brown). Representative photomicrographs are presented (scale bar = 50 μm). C) 4-hydroxy-2(E)-nonenal (4-HNE) glutathione conjugation activity in epidermal GST preparations from TPA-treated C57BL/6 and DBA/2 mice. Groups of six C57BL/6 and DBA/2 mice were killed at the indicated time points following a single topical application of 6.8 nmol TPA, and the dorsal skins were removed. The epidermis was harvested and homogenized, and then cytosolic glutathione S-transferases were purified from the epidermal lysates using a glutathione–agarose column. Conjugation activity toward 4-HNE was assessed spectrophotometrically and normalized to protein content. The mean specific activity and 95% confidence interval are presented. These experiments were performed in duplicate with similar results.

Figure 4

Epidermal 4-hydroxy-2(E)-nonenal (4-HNE) conjugation capacity in glutathione S-transferase α4 (Gsta4)-deficient vs wild-type mice. A) Western blot analysis of Gsta4 protein expression in Gsta4-deficient vs wild-type mice. Gsta4-deficient mice and littermate controls were treated with 6.8 nmol 12-O-tetradecanoylphorbol-13-acetate (TPA) or vehicle and killed at various time points. The dorsal skin was removed and the epidermis was scraped into radioimmunoprecipitation assay buffer before being homogenized with an 18-gauge needle and syringe. The homogenates were cleared by centrifugation and the supernatant was assayed for protein content. Then equal amounts of supernatant protein were analyzed for Gsta4 protein expression by western blot. Actin protein expressed is presented as a loading control. This experiment was performed in triplicate with similar results. B) Immunohistochemical analysis of Gsta4 protein expression in Gsta4-deficient vs wild-type mice. Gsta4-deficient and wild-type mice (n = 3 per strain) were treated topically with TPA twice weekly for 2 weeks and 48 hours after the final treatment the mice were killed and the dorsal skin removed. Skin sections were formalin-fixed, embedded, and examined for Gsta4 expression by immunohistochemistry as described above (see Figure 3, B). C) 4-HNE glutathione conjugation activity in epidermal GST preparations from Gsta4-deficient vs wild-type mice. Groups of six Gsta4-deficient and wild-type mice were killed at the indicated time points following topical treatment with TPA and the dorsal skins were removed. The epidermis was harvested, pooled, and homogenized, and then cytosolic glutathione S-transferases were purified from the epidermal lysates using a glutathione–agarose column. Conjugation activity toward 4-HNE was assessed spectrophotometrically and normalized to protein content. The mean and 95% confidence interval are displayed. This experiment was performed in duplicate with similar results. D) Glutathione-conjugated metabolites of 4-HNE in epidermis of Gsta4-deficient and wild-type mice following TPA treatment. Gsta4-deficient mice and littermate controls were treated twice weekly for 2 weeks with 3.4 nmol TPA, and 24 hours following the final treatment, epidermal scrapings were harvested from at least three individual mice per group. Glutathione-conjugated metabolites of 4-HNE in the epidermal samples were detected by Liquid Chromatography-Tandem Mass Spectrometry. The mean and 95% confidence are presented. This experiment was performed in triplicate with similar results and the data have been combined. All P values were based on two-sided Student t tests. 4-HNE-SG = glutathione conjugated 4-HNE.

Figure 5

Susceptibility of C57BL/6.glutathione S-transferase α4 deficient (Gsta4^−/−) vs wild-type mice to tumor promotion by TPA. Groups of at least 24 mice of each strain were initiated by topical application of 100 nmol 7,12-dimethylbenz(a)anthracene to the shaved dorsal skin and promoted twice weekly with 3.4 nmol TPA for 38 weeks. Tumors were counted weekly by palpation and tumor multiplicity was determined by dividing the total number of tumors by the number of mice at risk when the first tumor was observed. Two independent experiments were performed with similar results; therefore, the data have been combined (C57BL/6.Gsta4^−/−, n = 62; C57BL/6.Gsta4^+/+, n = 53). A) Time course of tumor development in Gsta4-deficient and wild-type mice. Tumor multiplicity was statistically significantly higher in Gsta4-deficient mice than in wild-type mice at 38 weeks (P = .007, one-tailed Mann–Whitney U test). The mean and 95% confidence interval for tumor multiplicity at 36 and 38 weeks are presented. B) Tumor latency in Gsta4-deficient and wild-type mice. Tumor latency was decreased in Gsta4-deficient mice (P = .006, Gehan–Breslow–Wilcoxon). The number of mice at risk for weeks 0, 20, and 38 were 53, 53, 36 for wild-type and 62, 54, and 31 for Gsta4-deficient mice.