PPARδ-87T/C plays a critical role in the development of colorectal cancer

To the Editor: The role of the peroxisome proliferator-activated receptor (PPAR)δ in colorectal cancer (CRC) is controversial. Indeed, PPARβ/δ can inhibit cancer by inhibiting metastasis and promote cancer by promoting metastasis and angiogenesis.[1,2] The PPARδ-87T/C (also referred to as +294 T/C [rs2016520]) single-nucleotide polymorphism (SNP) is associated with the binding ability of the transcription factor specificity protein-1 (Sp-1).[3,4] Sp-1 is an important transcription factor in CRC tumorigenesis and progression.[5] Therefore, Sp-1 is a potential molecular marker of CRC progression and a promising therapeutic target in this malignancy. The purpose of this study was to investigate the impact of PPARδ-87T/C on transcription factor binding. This study assessed the relationship between the PPARδ-87T/C polymorphism, CRC risk, and key clinicopathological parameters in a Western Chinese Han population. This case-control study was approved by the Institutional Review Board of the West China Hospital (29-11-2021/No. 1308) and was registered (#ChiCTR2100054751, www.chictr.org.cn) on December 28, 2021. Written informed consent was obtained from the participants. Peripheral blood samples were collected from 410 patients with primary CRC and 496 controls. Genomic DNA was extracted for genotyping PPARδ-87T/C. The PPARδ-87T/C polymorphism nucleotide sequences were used to predict the differences in binding transcription factors of this SNP [Figure 1A and B]. Gene Ontology (GO) functional enrichment analysis was conducted to assign various biological processes (BPs), molecular functions (MFs), and cellular components (CCs). Sp-1 was overexpressed in HT29 cells. A dual-luciferase reporter assay was performed for PPARδ and Sp-1.Figure 1: Detailed experimental procedures. (A) Genotyping of the PPARδ-87T/C polymorphism. Lane M, 2000 bp marker ladder; lanes 1–3, TT genotype (bands at 239 bp); lanes 4–5, TC genotype (bands at 239, 201, and 38 bp); and lanes 6 CC genotype (bands at 201 and 38 bp). (B) A partial PPARδ-87T/C polymorphism nucleotide sequence in colorectal cancer (CRC). (a) Rare homozygous CC genotype; (b) heterozygous TC genotype; (c) common homozygous TT genotype. (C) Differences in transcriptional activity between the T and C alleles of the -87T/C polymorphism by dual-luciferase reporter assay. Data are mean ± SE. **P <0.01; ***P <0.001. The 1001 bp DNA constructs of the T or C genotypes, respectively, were transfected into human CRC HT29 cells. An empty plasmid or a plasmid containing Sp-1 was also transfected. Sp-1 overexpression was mainly performed to compare the fluorescence values of a single promoter and/or a single transcription factor, plus the overexpression vector. Fluorescence was higher for the C allele with SP-1 compared with the C allele construct (n = 5, t = 9.662, P <0.001). Fluorescence elevation for the T allele with Sp-1 was compared with the activity of the T allele (n = 5, t = 12.962, P <0.001). (D) The -87T/C polymorphism of PPARβ/δ influences the binding of the transcription factor Sp-1. EMSA of nuclear extracts from human CRC HT29 cells was performed using double-strand 27-mer oligonucleotides corresponding to the sequence (5′-CTGATGGGAACCAT/CCCTGTAGAGGTCC-3′) of the common T allele and the rare C allele. Lane 1, -87C, without extract. Lanes 2 and 3, with human CRC HT29 cell nuclear extract in the absence of the competitor. Lanes 4–6 and 8–10, with human CRC HT29 cell nuclear extract in the presence of a 150-fold excess of cold DNA as a competitor. The competitors included the -87C probe (lanes 4 and 8, indicated with C), the -87T probe (lanes 5 and 9, indicated with T), and a nonspecific probe (lanes 6 and 10, indicated with X). Lane 7, with human CRC HT29 cell nuclear extract in the presence of antibodies targeting Sp-1. EMSA: Electromobility shift assay; GO: Gene Ontology; PPAR: Peroxisome proliferator-activated receptor; SE: Standard Error; Sp-1: Specificity protein-1.According to the The Cancer Genome Atlas (TCGA) COAD/READ bulk transcriptome RNA-sequencing data, PPARδ was significantly downregulated in primary tumor tissue (P = 8.3E−16, log2|Fold change| = −0.775) [Supplementart Figure 1, https://links.lww.com/CM9/C650], and its expression level was not significantly correlated with overall survival (OS) [Supplementary Figure 2, https://links.lww.com/CM9/C650]. In addition, the relationships between the PPARδ gene expression and sex, age, primary size, lymphatic invasion, microsatellite instability, tumor stage, and Tumor-Node-Metastasis (TNM) stage were assessed among the patients [Supplementary Figure 1 and 2, https://links.lww.com/CM9/C650]. There were no significant relationships between PPARδ gene expression and all clinical pathological variables except microsatellite instability (all P values >0.05). Specifically, when a specific age threshold was set, there might be significant differences in the expression of this gene between high and low age groups [Supplementart Figure 1, https://links.lww.com/CM9/C650]. The prediction results from TFBIND indicated that the PPARδ-87T/C polymorphic nucleotide sequences can bind to different transcription factors [Supplementary Table 1, https://links.lww.com/CM9/C650]. In particular, it was found that it can bind to the transcription factor Sp-1 when the C allele is present; however, the presence of the T allele does not cause it to bind to Sp-1. It may suggest that the SNP may affect the interaction with Sp-1, thereby affecting CRC risk. Therefore, the GO analysis of PPARδ and Sp-1 was performed. It was found that PPARδ and Sp-1 could be significantly enriched simultaneously in 13 biological processes and four molecular functions (P <0.05) [Supplementart Figure 3, https://links.lww.com/CM9/C650], including RNA polymerase II repressing transcription factor binding (P = 3.54E–06), cholesterol metabolic process (P = 6.20E–05), epithelial cell migration (P = 3.6E–04), and regulation of epithelial cell proliferation (P = 4.5E–04). The dual-luciferase reporter assay was used to evaluate allele-specific differences in transcriptional activity. In this study, human CRC HT29 cells were transfected with 1001 bp DNA constructs carrying the -87T or C allele. As shown in Figure 1C, the fluorescence of HT29 cells transfected using PPARδ-87C + PCDNA3.1-SP1 was significantly enhanced compared with the cells transfected with PPARδ-87C + pcDNA3.1 (n = 5, t = 9.662, P <0.001). The fluorescence of HT29 cells transfected by PPARδ-87T + PCDNA3.1-SP1 was significantly enhanced compared with that in PPARδ-87T + pcDNA3.1 group (n = 5, t = 12.962, P <0.001). Simultaneously, transfection of HT29 cells with PPARδ-87C + pcDNA3.1-SP1 resulted in significantly enhanced fluorescence compared with PPARδ-87T + pcDNA3.1-SP1 (n = 5, t = −4.621, P = 0.004). Transfection of HT29 cells with PPARδ-87C + pcDNA3.1 also resulted in significantly enhanced fluorescence compared with the PPARδ-87T + pcDNA3.1 group (n = 5, t = 5.802, P <0.001) [Figure 1C]. These results indicated that PPARδ-87C has higher transcriptional activity than PPARδ-87T. The electrophoretic mobility shift assay was performed to assess whether the -87T/C polymorphism influences the binding to nuclear proteins. As depicted in Figure 1D, nuclear proteins complexed with the tested DNA probes (5′-CTGATGGGAACCAT/CCCTGTAGAGGTCC-3′) slowed their mobility. In contrast to the frequent -87T-probe, these proteins are only bound to the uncommon -87C-probe [Figure 1D, lanes 2 and 3]. Competition assays using a 150-fold excess of cold probes revealed that protein complexes with the fragment harboring the C allele had sequence specificity. Nuclear proteins from human CRC HT29 cells interacted with both the -87C-probe and -87T-probe [Figure 1D], and Sp-1 antibodies identified these proteins as the Sp-1 protein. The shifted band was unchanged upon adding the T probe or an unspecific probe as a competitor, while diminishing almost completely with the C probe as a competitor [Figure 1D, lanes 4–6]. When supershift analyses were carried out with antibodies targeted against Sp-1 [Figure 1D, lane 7], supershift bands representing DNA-protein complexes were shifted after incubation with anti-Sp1 antibodies. No specific binding to the T allele was detected by competition experiments [Figure 1D, lanes 8–10]. The findings mentioned earlier suggested that Sp-1 binds specifically to the rare C allele and not to the common T allele. The PPARδ-87T/C polymorphism frequencies in controls did not deviate from the Hardy–Weinberg equilibrium (P = 0.087). The study population characteristics were compiled (P >0.05) [Supplementary Table 2, https://links.lww.com/CM9/C650]. The patients were 59.1 ± 13.3 years (range, 17.0–87.0 years), and the controls were 59.5 ± 10.3 years (range, 18.0–91.0 years). The sex ratios of the patients and controls (male-to-female) were 1.43:1 and 1.61:1, respectively. There were no significant differences in sex (P = 0.372) or age (P = 0.283) between the two groups, consistent with successful mapping. In patients with CRC, 58.8% (241/410) were male, 74.6% (306/410) were afflicted by rectal cancer, and 78.3% (321/410) had stage T3 and T4 CRC. In most cases, patients exhibited moderate tumor differentiation (59.3%, 243/410), with poor differentiation evident in 30.7% (126/410). The PPARδ-87T/C genotype and allele frequencies in the controls and patients were compiled in Supplementary Table 3, https://links.lww.com/CM9/C650. In the controls, PPARδ-87T/C TT, TC, and CC genotypes were evident in 65.5%, 32.3%, and 2.2%, respectively, and in patients with CRC, these frequencies were 61.7%, 32.2%, and 6.1%, respectively (P = 0.011). In the controls, the T and C alleles exhibited frequencies of 81.7% and 18.3%; in patients with CRC, 77.8% and 22.2%, respectively (P = 0.042). Binary logistic regression analyses were used to probe the relationship between PPARδ-87T/C polymorphisms and CRC risk and revealed that relative to the most common TT genotype, those carrying the CC genotype were at an elevated risk of CRC (odds ratio [OR] = 2.835 and 95% confidence interval [CI] = 1.366–5.884; P = 0.005). No relationships between sex or age and CRC risk were detected (P >0.05). The relationship between the PPARδ-87T/C genotype and sex, age, tumor location, tumor size, T-stage, lymph node status, distant metastases, and differentiation was assessed among the patients [Supplementary Table 4, https://links.lww.com/CM9/C650]. The PPARδ-87T/C genotype was associated with the tumor location in CRC (P = 0.037) but showed no significant associations with other clinicopathological variables (all P values >0.05). According to the Hardy–Weinberg equilibrium analysis [Supplementary Table 5, https://links.lww.com/CM9/C650], there was a genetic imbalance in the genotype frequency in CRCs, and the analysis showed that there was a significant difference between TT and TC (OR = 0.520, 95% CI: 0.308–0.880, P = 0.015), while there were no statistically significant differences between the TT and CC genotypes (OR = 1.160, 95% CI: 0.480–2.807, P = 0.741) or between the T and C alleles (OR = 0.757, 95% CI: 0.510–1.124, P = 0.167). It indicates that the samples were uneven and that there was selection bias. In conclusion, this study strongly suggests that the CC genotype of PPARδ-87T/C appears to be associated with a higher CRC susceptibility and tumor location among Chinese Han individuals compared with the common TT genotype. Additional studies are warranted to confirm these results in human specimens and to examine the various genes and proteins involved. In addition, future large-scale multicenter studies will be necessary to understand how PPARδ transcriptional activity in response to the Wnt/β-catenin signaling pathway and PPARδ-87T/C polymorphism is linked to colorectal tumorigenesis. Funding This work was supported by grants from Provincial Natural Science Foundation of Shanxi (No. 201901D111440); PhD start-up fund, Shanxi Provincial People’s Hospital (2018). Conflicts of interest None.