The relationship between IGF2BP2 and PPARG polymorphisms and susceptibility to esophageal squamous-cell carcinomas in the eastern Chinese Han population

Abstract

The aim of this case–control study was to assess whether PPARG and IGF2BP2 polymorphisms confer susceptibility to esophageal squamous-cell carcinoma (ESCC). A total of 507 patients pathologically confirmed for ESCC and 1,496 age-, sex-, and residence-matched healthy individuals were enrolled. The PPARG rs1801282 C>G and rs3856806 C>T and IGF2BP2 rs1470579 A>C and rs4402960 G>T polymorphisms were selected and genotyped by SNPscan genotyping assays. Multivariable logistic analysis suggested that the PPARG rs3856806 C>T polymorphism might increase the risk of ESCC. In different stratified analyses, there were significant associations between PPARG rs3856806 C>T and risk of ESCC in female, never-smoking, drinking, and never-drinking subgroups. In addition, we also found that PPARG rs1801282 C>G increased ESCC risk in the never-smoking subgroup. There was significant difference in C_rs1470579G_rs4402960C_rs1801282C_rs3856806-haplotype distribution among ESCC cases and control subjects. In conclusion, our findings highlight that PPARG rs1801282 C>G and rs3856806 C>T polymorphisms are candidates for susceptibility to ESCC in the eastern Chinese Han population. The C_rs1470579G_rs4402960C_rs1801282C_rs3856806 haplotype is associated with susceptibility to ESCC.

Keywords:

• PPARG
• IGF2BP2
• polymorphism
• risk
• ESCC

Introduction

Esophageal cancer (EC) is a complex disease characterized by progressive dysphagia and emaciation. Because of aging and unhealthy lifestyles (eg, low intake of fruit and vegetables, the rising prevalence of smoking and drinking), EC constitutes a burden worldwide. Esophageal squamous-cell carcinoma (ESCC) is the most common subtype of EC in China.^1,2 The potential risk factors driving the high incidence of ESCC are not well understood. It is thought that poor nutritional status, insufficient fruit/vegetables intake, smoking, and drinking beverages at very high temperatures may be involved in the development of ESCC, though these potential risk factors cannot explain the total etiology of ESCC. Nowadays, it is considered that genetic variants may influence the risk of ESCC.

PPARs comprise a group of nuclear transcription factors, which are classified into three subtypes: PPARα, PPARβ, and PPARγ.³ PPARγ is also named PPARG. In humans, the PPARG gene is located on chromosome 3p25. PPARG interacts with the retinoid X receptor, constructs a dipolymer, and then regulates its target genes, which are involved in cellular differentiation and metabolism of carbohydrates and lipids.⁴ Polymorphisms in the PPARG gene are assumed to influence the development of malignancies and metabolism-related diseases. Pro12Ala (rs1801282 C>G) and His449His (rs3856806 C>T) polymorphisms are the two most common single-nucleotide polymorphism (SNPs) in the PPARG gene. Recently, a case–control study was conducted to assess the relationship of PPARG rs3856806 C>T with susceptibility to EC. The results indicated that PPARG rs3856806 C>T might be associated with the risk of EC.⁵ In addition, the association between PPARG rs1801282 C>G polymorphism and EC risk was unknown.

IGF2BP2 binds to the 5′UTR of IGF2 mRNA and affects its translation.⁶ Barghash et al reported that IGF2BP2 expression correlated with poor survival in patients with esophageal adenocarcinoma and ESCC.⁷ Case–control studies have indicated that IGF2BP2 rs4402960 G>T might be associated with the risk of breast cancer⁸ and colorectal cancer.⁹ In addition, it has been reported that IGF2BP2 rs1470579 A>C was associated with the risk of type 2 diabetes.^10,11 However, the association between IGF2BP2 polymorphisms and EC risk was unclear.

The aim of this case–control study was to explore the potential relationship of genetic variations in PPARG and IGF2BP2 with risk of ESCC in the eastern Chinese Han population. PPARG rs1801282 C>G and rs3856806 C>T and IGF2BP2 rs1470579 A>C and rs4402960 G>T polymorphisms were selected and genotyped by SNPscan genotyping assays in 507 patients with ESCC and 1,496 controls.

Materials and methods

Subjects

A total of 507 patients pathologically confirmed for ESCC from the Affiliated People’s Hospital of Jiangsu University and the Affiliated Union Hospital of Fujian Medical University (mean age 62.77±8.01 years) were recruited in our study. The noncancer controls consisted of 1,496 age-, sex-, and residence-matched healthy individuals (mean age 62.77 ±8.84 years) without any cancer history or autoimmune diseases. All participants were enrolled between August 2013 and December 2016. EDTA-anticoagulated peripheral blood was collected after written consent had been signed. A questionnaire was used to obtain participants’ risk factors and demographic variables. A body-mass index (BMI) ≥24 kg/m² was accepted as the criterion of obesity and overweight.^12,13 This study was approved by the institutional review boards of Jiangsu University (Zhenjiang, China) and Fujian Medical University (Fuzhou, China).

DNA extraction and genotyping

Genomic DNA was extracted from whole blood using a DNA kit (Promega, Madison, WI, USA). PPARG rs1801282 C>G and rs3856806 C>T and IGF2BP2 rs1470579 A>C and rs4402960 G>T genotypes were determined by double ligation and multiplex-fluorescence polymerase chain reaction (SNPscan; Genesky Biotechnologies, Shanghai, China).¹⁴ For quality control, 80 samples (4%) were randomly selected from the 2,003 DNA samples and genotyped again by another technician. Genotypes of PPARG rs1801282 C>G and rs3856806 C>T and IGF2BP2 rs1470579 A>C and rs4402960 G>T polymorphisms were confirmed.

Statistical analysis

Continuous variables (age, BMI, height, and weight) are expressed as means ± SD. Comparisons of these continuous variables between two groups were performed using Student’s t-test. The χ² test was used to compare categorical variables (PPARG and IGF2BP2 genotypes, BMI, sex, age, and smoking status and alcohol use). We checked the deviations for Hardy–Weinberg equilibrium in normal controls with an Internet-based calculator (http://ihg.gsf.de/cgi-bin/hw/hwa1.pl).^15–21 Statistical significance was defined as P<0.05 (two-tailed). The relationships of PPARG rs1801282 C>G and rs3856806 C>T and IGF2BP2 rs1470579 A>C and rs4402960 G>T polymorphisms with ESCC susceptibility were determined by crude odds ratios (ORs) and 95% CIs. Adjusted for BMI, age, sex, alcohol use, and smoking status, multivariate linear regression was used to assess the potential association further among these polymorphisms and susceptibility to ESCC. SAS 9.4 software for Windows (SAS Institute, Cary, NC, USA) was used to analyze the data. SHEsis software (http://analysis.bio-x.cn/myanalysis.php; Bio-X, Shanghai, China) was used online to construct the haplotypes.^22–24

Results

Baseline characteristics

Characteristics of 507 ESCC cases and 1,496 controls included in this case–control study are presented in Table 1. ESCC cases and controls were well matched on age and sex, as shown by χ² tests (P=0.994 and P=0.406, respectively). As shown in Table 1, significant differences were found on smoking status and alcohol use between cases and controls (P<0.001). The primary information for PPARG rs1801282 C>G and rs3856806 C>T and IGF2BP2 rs1470579 A>C and rs4402960 G>T SNPs is shown in Table 2. For these four genotyped SNPs, the successful ratio was 99.45%–99.5% in all 2,003 DNA samples. The concordance rates of quality-control testing were 100%. Minor allele frequency of PPARG rs1801282 C>G and rs3856806 C>T and IGF2BP2 rs1470579 A>C and rs4402960 G>T SNPs in controls was close to the minor allele-frequency data for Chinese (Table 2). In controls, the genotype frequencies for PPARG rs1801282 C>G and rs3856806 C>T polymorphisms were in Hardy–Weinberg equilibrium (Table 2).

Table 1 Distribution of selected demographic variables and risk factors in ESCC cases and controls

	Cases (n=507)		Controls (n=1,496)		P-valuea
	n	%	n	%
Age (years), mean ± SD	62.77±8.01		62.77±8.84		0.994
Age (years)					0.225
<63	271	53.45	753	50.33
≥63	236	46.55	743	49.67
Sex					0.406
Male	377	74.36	1,084	72.46
Female	130	25.64	412	27.54
Tobacco use					<0.001
Never	247	48.72	1,090	72.86
Ever	260	51.28	406	27.14
Alcohol use					<0.001
Never	341	67.26	1,329	88.84
Ever	166	32.74	167	11.16
Height (cm)	166±7.29		166.1±7.08		0.743
Weight (kg)	61.54±9.83		66.11±9.92		<0.001
BMI (kg/m^2), mean ± SD	22.27±2.90		23.91±3.03		<0.001
BMI (kg/m^2)					<0.001
<24	370		779
≥24	137		717

Note:

a Two-sided χ² test and Student’s t-test.

Abbreviations: ESCC, esophageal squamous-cell carcinoma; BMI, body-mass index.

Table 2 Primary information for PPARG rs1801282 C>G, rs3856806 C>T, and IGF2BP2 1470579 A>C, rs4402960 G>T polymorphisms

Genotyped SNPs	Chromosome	Chromosome position (NCBI build 38)	MAF for Chinese in database	MAF in our controls (n=1,496)	P-value for HWE test in our controls	Genotyping method	Genotyping value (%)
PPARG rs1801282 C>G	3	12351626	0.07	0.05	0.911	SNPscan	99.5
PPARG rs3856806 C>T	3	12434058	0.25	0.22	0.083	SNPscan	99.5
IGF2BP2 rs1470579 A>C	3	185811292	0.27	0.25	0.002	SNPscan	99.5
IGF2BP2 rs4402960 G>T	3	185793899	0.26	0.25	0.002	SNPscan	99.45

Abbreviations: SNPs, single-nucleotide polymorphisms; MAF, minor allele frequency; HWE, Hardy–Weinberg equilibrium.

Association of PPARG rs1801282 C>G and rs3856806 C>T and IGF2BP2 rs1470579 A>C and rs4402960 G>T polymorphisms with ESCC risk

The genotypes of PPARG rs1801282 C>G and rs3856806 C>T and IGF2BP2 rs1470579 A>C and rs4402960 G>T polymorphisms are summarized in Table 3. In single-locus analyses, the genotype frequencies of PPARG rs3856806 C>T were 54.56% (CC), 39.09% (CT), and 6.35% (TT) in ESCC patients and 59.7% (CC), 36.13% (CT), and 4.16% (TT) in controls. When the PPARG rs3856806 CC homozygote genotype was used as the reference group, the PPARG rs3856806 CT genotype was correlated with a significantly increased risk of ESCC (CT vs CC, adjusted OR 1.28, 95% CI=1.02–1.6; P=0.033). When the PPARG rs3856806 CC homozygote genotype was used as the reference group, the PPARG rs3856806 TT genotype was correlated with a borderline significantly increased risk of ESCC (TT vs CC, adjusted OR 1.55, 95% CI=0.96–2.50; P=0.074). In the recessive model, when the PPARG rs3856806 CC/CT genotypes were used as the reference group, the PPARG rs3856806 TT homozygote genotype was not associated with susceptibility for ESCC (adjusted OR 1.41, 95% CI=0.88–2.26; P=0.153). In the dominant model, PPARG rs3856806 CT/TT genotypes were associated with an increased risk of ESCC compared with the PPARG rs3856806 CC genotype (adjusted OR 1.31, 95% CI=1.06–1.63; P=0.014) (Table 3). Logistic regression analyses showed that PPARG rs1801282 C>G and IGF2BP2 rs1470579 A>C, rs4402960 G>T polymorphisms were not correlated with the susceptibility for ESCC (Table 3).

Table 3 Logistic regression analyses of association between PPARG rs1801282 C>G, rs3856806 C>T and IGF2BP2 1470579 A>C, rs4402960 G>T polymorphisms and risk of ESCC

Genotype	ESCC cases (n=507)		Controls (n=1,496)		Crude OR (95% CI)	P-value	Adjusted ORa (95% CI)	P-value
	n	%	n	%
PPARG rs1801282 C>G
CC	440	87.3	1,334	89.59	1
GC	63	12.5	151	10.14	1.26 (0.92–1.73)	0.144	1.24 (0.88–1.73)	0.219
GG	1	0.20	4	0.27	0.76 (0.08–6.79)	0.804	1.08 (0.11–10.5)	0.950
GC+GG	64	12.7	155	10.41	1.25 (0.92–1.71)	0.156	1.23 (0.88–1.72)	0.217
CC+GC	503	99.8	1,485	99.73	1		1
GG	1	0.20	4	0.27	0.74 (0.08–6.62)	0.786	1.05 (0.11–10.26)	0.966
G allele	65	6.45	159	5.34
PPARG rs3856806 C>T
CC	275	54.56	889	59.7	1
CT	197	39.09	538	36.13	1.18 (0.96–1.46)	0.125	1.28 (1.02–1.6)	0.033
TT	32	6.35	62	4.16	1.66 (1.06–2.6)	0.026	1.55 (0.96–2.5)	0.074
CT+TT	229	45.44	600	40.30	1.23 (1.01–1.51)	0.043	1.31 (1.06–1.63)	0.014
CC+CT	472	93.65	1,427	95.84	1		1
TT	32	6.35	62	4.16	1.56 (1.01–2.42)	0.047	1.41 (0.88–2.26)	0.153
T allele	261	25.89	662	22.23
IGF2BP2 1470579 A>C
AA	280	55.56	855	57.42	1		1
AC	194	38.49	517	34.72	1.14 (0.92–1.41)	0.218	1.09 (0.87–1.37)	0.453
CC	30	5.95	117	7.86	0.78 (0.51–1.19)	0.252	0.78 (0.5–1.22)	0.282
AC+CC	224	44.44	634	42.58	1.08 (0.88–1.32)	0.465	1.04 (0.83–1.29)	0.748
AA+AC	474	94.05	1,372	92.14	1		1
CC	30	5.95	117	7.86	0.74 (0.49–1.12)	0.159	0.76 (0.49–1.17)	0.213
C allele	254	25.20	751	25.22
IGF2BP2 rs4402960 G>T
GG	294	58.45	872	58.56	1		1
GT	179	35.59	506	33.98	1.04 (0.84–1.29)	0.698	0.99 (0.78–1.24)	0.904
TT	30	5.96	111	7.45	0.8 (0.52–1.22)	0.295	0.83 (0.53–1.29)	0.402
GT+TT	209	41.55	617	41.44	1.01 (0.82–1.23)	0.694	0.96 (0.77–1.2)	0.737
GG+GT	473	94.04	1,378	92.55	1		1
TT	30	5.96	111	7.45	0.79 (0.52–1.19)	0.261	0.84 (0.54–1.29)	0.418
T allele	239	23.76	728	24.45

Note:

a Adjusted for age, sex, BMI, alcohol use, and smoking status.

Abbreviations: ESCC, esophageal squamous-cell carcinoma; BMI, body-mass index; OR, odds ratio.

Association of PPARG rs1801282 C>G and rs3856806 C>T and IGF2BP2 rs1470579 A>C and rs4402960 G>T polymorphisms with ESCC risk in different stratification groups

To determine the potential effects of PPARG rs1801282 C>G genotypes on ESCC risk in different subgroups according to BMI, age, sex, and smoking and drinking status, we carried out stratified analyses (Table 4). In the never-smoking subgroup, after adjustment for sex, age, BMI, and alcohol use, we found that the PPARG rs1801282 C>G polymorphism increased ESCC risk in two genetic models (CG vs CC, adjusted OR 1.54, 95% CI 1.01–2.35, P=0.047; CG/GG vs CC, adjusted OR 1.54, 95% CI 1.01–2.34, P=0.044 [Table 4]).

Table 4 Stratified analyses between PPARG rs1801282 C>G polymorphism and ESCC risk by sex, age, BMI, smoking status, and alcohol consumption

	PPARG rs1801282 C>G (case/control)a			Adjusted ORb (95% CI); P-value
	CC	CG	GG	CC	CG	GG	CG/GG	GG vs (CG/CC)
Sex
Male	328/963	47/112	0/3	1	1.21 (0.81–1.79); P=0.351	–	1.18 (0.8–1.75); P=0.407	–
Female	112/371	16/39	1/1	1	1.55 (0.81–2.99); P=0.188	3.96 (0.25–63.97); P=0.332	1.62 (0.86–3.08); P=0.138	3.77 (0.23–60.91); P=0.349
Age, years
<63	207/679	28/66	1/2	1	1.43 (0.85–2.41); P=0.179	1.78 (0.13–24.51); P=0.665	1.43 (0.86–2.39); P=0.172	1.71 (0.13–23.44); P=0.687
≥63	233/655	35/85	0/2	1	1.14 (0.73–1.78); P=0.555	–	1.14 (0.73–1.77); P=0.571	–
Smoking status
Never	210/976	34/106	1/4	1	1.54 (1.01–2.35); P=0.047	1.34 (0.14–12.59); P=0.796	1.54 (1.01–2.34); P=0.044	1.28 (014–11.97); P=0.830
Ever	230/358	29/45	0/0	1	0.92 (0.54–1.57); P=0.757	–	0.92 (0.54–1.56); P=0.746	–
Alcohol consumption
Never	296/1,186	41/134	1/3	1	1.20 (0.82–1.76); P=0.348	1.82 (0.18–18.63); P=0.614	1.22 (0.84–1.78); P=0.305	1.78 (0.17–18.25); P=0.626
Ever	144/148	22/17	0/1	1	1.41 (0.67–3); P=0.369	–	1.32 (0.63–2.77); P=0.457	–
BMI (kg/m^2)
<24	319/695	47/78	1/1	1	1.32 (0.87–1.99); P=0.193	2.3 (0.14–37.56); P=0.558	1.34 (0.89–2.02); P=0.165	2.23 (0.14–36.38); P=0.574
≥24	121/639	16/73	0/3	1	1.13 (0.63–2.03); P=0.691	–	1.09 (0.61–1.96); P=0.775	–

Notes:

a For PPARG rs1801282 C>G, genotyping was successful in 507 (99.41%) ESCC cases and 1,496 (99.53%) controls;

b adjusted for multiple comparisons (age, sex, BMI, smoking status, and alcohol consumption [besides stratified factors accordingly]) in a logistic regression model.

Abbreviations: ESCC, esophageal squamous-cell carcinoma; BMI, body-mass index; OR, odds ratio.

Table 5 shows genotype frequencies of PPARG rs3856806 C>T in different subgroups. Significantly increased susceptibility to ESCC associated with the PPARG rs3856806 C>T polymorphism was found among several subgroups (Table 5). In the female subgroup after adjustment for BMI, age, and smoking and drinking status, the PPARG rs3856806 CT/TT genotypes were associated with increased ESCC risk compared with the PPARG rs3856806 CC genotype (CT/TT vs CC, adjusted OR 1.55, 95% CI 1.02–2.35; P=0.041 [Table 5]). In the never-smoking subgroup after adjustment for BMI, age, sex, and drinking status, we found that PPARG rs3856806 CT/TT genotypes increased ESCC risk compared with the PPARG rs3856806 CC genotype (CT/TT vs CC, adjusted OR 1.37, 95% CI 1.03–1.82; P=0.032 [Table 5]). In the drinking subgroup after adjustment for BMI, age, sex, and smoking status, significantly increased risk of ESCC associated with the PPARG rs3856806 C>T polymorphism was also found (TT vs CC, adjusted OR 3.36, 95% CI 1.05–12.74, P=0.041; TT vs CT/CC, adjusted OR 3.58, 95% CI 1.04–12.29, P=0.043 [Table 5]). In the never-drinking subgroup after adjustment for BMI, age, sex, and smoking status, significantly increased risk of ESCC associated with the PPARG rs3856806 C>T polymorphism was also found (CT vs CC, adjusted OR 1.37, 95% CI 1.06–1.77, P=0.015; CT/TT vs CC, adjusted OR 1.37, 95% CI 1.07–1.75, P=0.013 [Table 5]). In addition, there was no significant risk of ESCC correlated with the IGF2BP2 rs1470579 A>C and rs4402960 G>T polymorphisms evident among any subgroup (data not shown).

Table 5 Stratified analyses between PPARG rs3856806 C>T polymorphism and ESCC risk by sex, age, BMI, smoking status, and alcohol consumption

	PPARG rs3856806 C>T (case/control)a			Adjusted ORb (95% CI); P-value
	CC	CT	TT	CC	CT	TT	CT/TT	TT vs (CT/CC)
Sex
Male	206/632	144/403	25/43	1	1.22 (0.93–1.59); P=0.146	1.57 (0.89–2.77); P=0.116	1.26 (0.98–1.63); P=0.078	1.46 (0.84–2.54); P=0.185
Female	69/257	53/135	7/19	1	1.53 (0.99–2.36); P=0.054	1.57 (0.61–4.03); P=0.351	1.55 (1.02–2.35); P=0.041	1.33 (0.53–3.37); P=0.543
Age, years
<63	131/457	89/259	16/31	1	1.35 (0.96–1.9); P=0.081	1.6 (0.78–3.28); P=0.199	1.37 (0.99–1.9); P=0.061	1.42 (0.7–2.87); P=0.331
≥63	144/432	108/279	16/31	1	1.26 (0.93–1.72); P=0.136	1.53 (0.79–2.96); P=0.209	1.31 (0.98–1.77); P=0.073	1.4 (0.73–2.68); P=0.306
Smoking status
Never	129/645	103/396	13/45	1	1.35 (1.01–1.82); P=0.044	1.35 (0.7–2.6); P=0.37	1.37 (1.03–1.82); P=0.032	1.2 (0.63–2.29); P=0.578
Ever	146/244	94/142	19/17	1	1.21 (0.84–1.73); P=0.302	1.82 (0.86–3.84); P=0.117	1.27 (0.9–1.79); P=0.171	1.69 (0.81–3.52); P=0.162
Alcohol consumption
Never	179/793	142/472	17/58	1	1.37 (1.06–1.77); P=0.015	1.24 (0.7–2.21); P=0.458	1.37 (1.07–1.75); P=0.013	1.10 (0.63–1.94); P=0.740
Ever	96/96	55/66	15/4	1	1.03 (0.62–1.72); P=0.901	3.66 (1.05–12.74); P=0.041	1.19 (0.73–1.94); P=0.48	3.58 (1.04–12.29); P=0.043
BMI (kg/m^2)
<24	207/469	135/268	25/37	1	1.23 (0.93–1.62); P=0.156	1.53 (0.86–2.72); P=0.146	1.27 (0.97–1.67); P=0.08	1.42 (0.81–2.51); P=0.222
≥24	68/420	62/270	7/25	1	1.41 (0.96–2.07); P=0.080	1.66 (0.68–4.05); P=0.268	1.43 (0.98–2.07); P=0.063	1.43 (0.6–3.44); P=0.423

Notes:

a For PPARG rs3856806 C>T, genotyping was successful in 507 (99.41%) ESCC cases and 1,496 (99.53%) controls;

b adjusted for multiple comparisons (age, sex, BMI, smoking status, and alcohol consumption [besides stratified factors accordingly]) in a logistic regression model.

Abbreviations: ESCC, esophageal squamous-cell carcinoma; BMI, body-mass index; OR, odds ratio.

SNP haplotypes

Using the SHEsis software,²² we constructed eight haplotypes (Table 6). There were significant differences in the CGCC haplotype of the order rs1470579 A>C, rs4402960 G>T, rs1801282 C>G and rs3856806 C>T polymorphism distribution among ESCC cases and the control subjects (OR 2.23, 95% CI=1.09–4.59; P=0.025 [Table 6]).

Table 6 Haplotype frequencies (%) in cases and controls and risk of ESCC

	Cases (n=1,006)		Controls (n=2,978)		Crude OR (95% CI)	P-value
	n	%	n	%
AGCC	543	53.98	1,679	56.38	Reference
CTCC	174	17.30	565	18.97	0.95 (0.78–1.16)	0.624
AGCT	164	16.30	430	14.44	1.18 (0.96–1.45)	0.113
CTCT	45	4.47	107	3.59	1.3 (0.91–1.87)	0.153
AGGT	36	3.68	79	2.65	1.41 (0.94–2.12)	0.096
CTGT	13	1.29	34	1.14	1.18 (0.62–2.26)	0.611
CGCC	13	1.29	18	0.60	2.23 (1.09–4.59)	0.025
AGGC	9	0.89	30	1.01	0.93 (0.44–1.97)	0.845
Others	9	0.89	36	1.21	0.77 (0.37–1.62)	0.492

Note: With order of rs1470579 A>C, rs4402960 G>T, rs1801282 C>G and rs3856806 C>T in gene position.

Abbreviations: ESCC, esophageal squamous-cell carcinoma; OR, odds ratio.

Discussion

In this case–control study, we explored the associations between the PPARG rs1801282 C>G and rs3856806 C>T and IGF2BP2 rs1470579 A>C and rs4402960 G>T SNPs and risk of ESCC in the eastern Chinese Han population. Multivariable logistic analysis suggested that PPARG rs3856806 C>T might be associated with an increased risk of ESCC. In different stratified analyses, there were significant associations between this polymorphism and risk of ESCC in the female, never-smoking, drinking, and never-drinking subgroups. In addition, we also found that PPARG rs1801282 C>G increased ESCC risk in the never-smoking subgroup. To the best of our knowledge, this is the first study to identify a potential association between PPARG rs1801282 C>G and rs3856806 C>T polymorphisms and increased risk of ESCC in Asians.

PPARG is a member of the nuclear hormone-receptor superfamily and may possess anti-inflammatory properties.²⁵ PPARG also plays an important role in cell proliferation/differentiation, which affects the development and progression of cancer.^26,27 The PPARG rs1801282 C>G polymorphism is located in the exon B region of the PPARG gene. Deeb et al reported that this SNP was associated with decreased transactivation activity and lower BMI and promoted insulin sensitivity.²⁸ A recent meta-analysis suggested that PPARG rs1801282 C>G polymorphism is a candidate for susceptibility to Asians.²⁹ The association between PPARG rs1801282 C>G and risk of ESCC has not been studied before. In this study, we found that the PPARG rs1801282 CG genotype was more frequent in ESCC patients in the never-smoking subgroup, which was in accordance with the results of the meta-analysis just mentioned. The function of the PPARG rs1801282 C>G SNP remains to be investigated in ESCC patients.

There was a difference in genotype distribution of the PPARG rs3856806 C>T polymorphism between ESCC patients and controls. The PPARG rs3856806 CT and TT/CT genotypes were more frequent in ESCC patients compared with healthy controls, suggesting that the PPARG rs3856806 TT/CT and CT genotypes might contribute to the development of ESCC. The PPARG rs3856806 C>T polymorphism is located in the exon of the PPARG gene. It is difficult to illustrate the exact function of a synonymous SNP. It is proposed that PPARG rs3856806 C→T substitution may disrupt the splice site,³⁰ and then affect the expression of PPARG. A meta-analysis suggested that PPARG rs3856806 C>T is marginally associated with cancer susceptibility,³¹ and our results were similar.

In the present investigation, we constructed eight haplotypes to study inherited patterns. We found that the C_rs1470579G_rs4402960C_rs1801282C_rs3856806 haplotype was associated with susceptibility for ESCC. Comparing the CGCC with the AGCC haplotype in the order of rs1470579 A>C, rs4402960 G>T, rs1801282 C>G, and rs3856806 C>T polymorphisms, we found that A→C variation in the rs1470579 A>C locus led to susceptibility of the haplotype to ESCC. Several case–control studies have reported that IGF2BP2 rs1470579 A>C was associated with type 2 diabetes mellitus.^11,32–34 However, a potential association between IGF2BP2 rs1470579 A>C polymorphism and ESCC risk was not found in our case–control study. In the future, more case–control studies with large samples and detailed risk factors should be carried out to confirm or refute our findings.

There were some limitations in our study. Firstly, this case–control study was limited by the moderate sample size of ESCC patients, which might lead to suboptimal power to identify true associations in the stratified analyses. Secondly, the controls were recruited from two local hospitals, and might not represent the general Chinese population well; this possible bias should not be ignored. Thirdly, only some functional SNPs in the PPARG and IGF2BP2 genes were selected. The relationship of PPARG and IGF2BP2 variants was not fully explored. In the future, a fine-mapping study should be conducted. Fourthly, detailed information on metastasis and survival of ESCC was not available at the time of research, which restricted further analysis on the potential role of PPARG and IGF2BP2 variants in ESCC progression and prognosis. Finally, for lack of some environmental risk factors, such as lifestyle and intake of fruit/vegetables, the interaction of gene variants with environmental risk factors was not considered.

Conclusion

Our findings highlight that PPARG rs1801282 C>G and rs3856806 C>T polymorphisms are candidates for susceptibility to ESCC in the eastern Chinese Han population. A fine-mapping study is required to confirm these preliminary findings.

Acknowledgments

We appreciate all subjects who participated in this study. We wish to thank Dr Yan Liu (Genesky Biotechnologies Inc, Shanghai, China) for technical support. This study was supported in part by the Natural Science Foundation of Universities and Colleges of Jiangsu (grant 16KJB310002), Senior Talents Scientific Research Foundation of Jiangsu University (grant 16JDG066). The Natural Science Foundation of Fujian Province (Grant No 2015J01435, 2017J01259), and the National Clinical Key Specialty Construction Program, Young and Middle-Aged Talent Training Project of the Health Development Planning Commission in Fujian (2016-ZQN-25 and 2014-ZQN-JC-11), Medical Innovation Project of Fujian (2014-CX-15 and 2014-CX-18), Nursery Garden Project of Fujian Medical University (2015MP020), and Science and Technology Project of Fujian (2060203).

Disclosure

The authors report no conflicts of interest in this work.