Significant association between RETN genetic polymorphisms and alcohol-induced osteonecrosis of femoral head.

BACKGROUND: Alcohol-induced osteonecrosis of femoral head (ONFH) is a complex disease and genetic factors are one of the causes. The purpose of this study is to investigate the effects of RETN (resistin; OMIM: 605565) and LDLR (low density lipoprotein receptor; OMIM: 606945) polymorphisms on the risk of alcohol-induced ONFH in Chinese Han population.
METHODS: A case-control study including 201 patients and 201 controls was designed. Seven single nucleotide polymorphisms (SNPs) in RETN gene and four SNPs in LDLR gene were genotyped using Agena MassARRAY platform. In allele model and genetic model, chi-square test and logistic regression were used to study the associations between these SNPs and ONFH susceptibility. In addition, the relationships between these SNPs, clinical phenotypes, and blood lipid level with one-way analysis of variance were analyzed.
RESULTS: In the allele model, rs7408174 and rs3745369 in RETN were associated with increased risk of alcohol-induced ONFH, whereas rs34861192 and rs3219175 in RETN showed reduced risk of alcohol-induced ONFH. In the genetic model, rs7408174 was associated with increased risk of alcohol-induced ONFH in dominant model and log-additive model. Rs3745369 showed an increased risk in codominant model, recessive model, and log-additive model. Rs34861192 showed a decreased risk in codominant model, dominant model, and log-additive model, and rs3219175 showed a decreased risk in dominant model and log-additive model. The rs3745368 in RETN was associated with the clinical stage of the disease.
CONCLUSION: These results suggest that RETN genetic polymorphisms are associated with the susceptibility of alcohol-induced ONFH in Chinese Han population.

INTRODUCTION

Osteonecrosis of femoral head (ONFH) is a complicated disease in clinic and is usually divided into traumatic and nontraumatic types. Alcohol‐induced ONFH caused by excessive alcohol intake over a long period of time is a type of nontraumatic ONFH. The etiology of alcohol‐induced ONFH is complicated. Early diagnosis of this disease is difficult, and the complex pathological process is often manifested due to abnormal lipid metabolism and inflammation. Excessive alcohol drinking may result in dyslipidemia, abnormal differentiation of bone marrow mesenchymal stem cells (BMSCs) and bone metabolic disorders. Moreover, alcohol has a significant dose effect on bone homeostasis (Gaddini, Turner, Grant, & Iwaniec, 2016). However, we found in clinical work that only a portion of people who drank similar amounts of alcohol developed ONFH.

Some studies have suggested that the ONFH disease is caused by the interaction between genetic and environmental factors (Song et al., 2017; Wang, Azeddine, et al., 2018; Zhou, Qu, Lv, & Zhu, 2018). Therefore, genetic polymorphisms involved in alcohol metabolism, lipid metabolism, bone and circulatory homeostasis may lead to differences in susceptibility to alcohol‐induced ONFH (Cui, Kaisaierjiang, Cao, Wu, & Lv, 2014; Hadjigeorgiou et al., 2008).

RETN (resistin; OMIM: 605565), located on chromosome 19, encodes resistin. Resistin affects bone metabolism and in vitro studies have shown that it can promote bone remodeling (Thommesen et al., 2006). Some scholars believe that there is a negative correlation between resistin content and bone density (Oh et al., 2005; Pedone et al., 2013; Zhang et al., 2010). Plasma resistin is also correlated with insulin resistance, lower HDL‐C, and high hs‐CRP (Osawa et al., 2007). Studies have shown that the polymorphisms of RETN have significant effect on plasma resistin concentration (Asano et al., 2010). In recent years, some scholars have found that the polymorphism of human RETN is also associated with osteoarthritis and rheumatoid arthritis (Hamalainen, Solovieva, Vehmas, Hirvonen, & Leino‐Arjas, 2018; Junker et al., 2017; Wang, Tang, et al., 2018).

LDLR (low density lipoprotein receptor; OMIM: 606945) is located on chromosome 19, which plays a critical role in regulating the plasma cholesterol level. Mutations in LDLR result in elevated cholesterol (Hobbs, Brown, Russell, Davignon, & Goldstein, 1987). Cholesterol is one of the risk factors for osteoporosis and cholesterol metabolic disorders is detrimental to bone health (Li et al., 2018; Mandal, 2015). Alterations in the function of the LDLR affected bone development and homeostasis (Yang & Williams, 2017).

There are few studies on the association of RETN and LDLR with alcohol‐induced ONFH. This work studies the association between RETN and LDLR genetic polymorphisms and the susceptibility of alcohol‐induced ONFH in Chinese Han population, which can guide the identification of high‐risk alcohol‐induced ONFH patients.

MATERIALS AND METHODS

Ethics approval and consent to participate

This study was conducted under the approval of the Second Affiliated Hospital of Inner Mongolia Medical University of Inner Mongolia, China and Zhengzhou Traditional Chinese Medicine (TCM) Traumatology Hospital of Henan Province, China. Blood samples were collected at the time of initial diagnosis after informed consent was obtained from all participants.

Subjects

All the 402 individuals including 201 cases and 201 controls were male and members of Chinese Han population living in Henan Province in China. Individuals who disagree to participate in this study were excluded.

The cases in our research satisfy the following criteria: (a) Patients should have a history of alcohol intake >400 ml/week (320 g/week, any type of alcoholic beverage) of pure ethanol for more than 6 months; (b) ONFH should be diagnosed within 1 year after the alcohol intake with this dose; (c) Patients should not have direct trauma and other risk factors (such as history of taking corticosteroids, cardiovascular diseases, congenital diseases, human immunodeficiency virus infection, diabetes mellitus, renal dysfunction, cancers, and familial hereditary diseases); (c) The diagnosis and staging of alcohol‐induced ONFH was evaluated by X‐ray, computed tomography(CT), nuclear magnetic resonance imaging (MRI); The selection criteria for control: (a) The age of the control group was matched with that of the case group; (b) The controls should have a history of alcohol intake >400 ml/week (320 g/week, any type of alcoholic beverage) of pure ethanol for more than 6 months; (c) No ONFH occurred; (d) Other factors were excluded (history of taking corticosteroids, cardiovascular diseases, congenital diseases, human immunodeficiency virus infection, diabetes mellitus, renal dysfunction, cancers, and familial hereditary diseases).

SNP selection and genotyping

The GenBank reference sequence and version number: RETN (Reference Sequence and version number: NG_023447.1; accession: NG_023447), LDLR (Reference Sequence and version number: NG_009060.1; accession: NG_009060). All eleven SNPs in RETN and LDLR with minor allele frequencies >5% were selected from the 1,000 Genomes Project databases (http://www.internationalgenome.org/). Blood samples were collected in tubes containing ethylene diaminetetraacetic acid (EDTA) and stored at −80°C after centrifuging at 2,000 rpm for 10 min. Genomic DNA was extracted from the peripheral blood of the participants using the GoldMag whole blood genomic DNA purification kit (GoldMag Co. Ltd., Xi'an, China). DNA concentration was determined by using a NanoDrop 2000C spectrophotometer (Thermo Scientific, Waltham, MA). The genotyping primers were designed with the Agena MassARRAY Assay Design 3.0 Software. Agena Typer 4.0 Software was used for managing the related data and the Agena MassARRAY RS1000 was used for genotyping.

Statistical analyses

All statistical analyses were performed using Microsoft Excel, SPSS 19.0 (SPSS, Chicago, IL) and PLINK version 1.07 software. Two‐sided p‐values less than 0.05 were considered statistically significant. The alleles of cases and controls were tested by Chi‐square test. The genotype frequencies were tested by logistic regression and the control group was compared with expected frequencies to test the deviations from Hardy–Weinberg equilibrium (HWE). Associations between SNPs and the risk of alcohol‐induced ONFH were tested in four genetic models (codominant, dominant, recessive, and log‐additive) using PLINK version 1.07 software and determined by unconditional logistic regression adjusted for age. Linkage disequilibrium among polymorphic sites was assessed with Haploview software package (version 4.2).

Bioinformatics and expression analyses

The Genotype‐Tissue Expression (GTEx) project provides a scientific resource to study SNPs and gene expression levels. In this research, online database (http://www.gtexportal.org/) was used to investigate the association between the 11 selected SNPs and the expression of two genes.

RESULTS

This study involved 402 male subjects as shown in Table 1, including 201 cases and 201 controls. The mean ages were 42.68 ± 12.88 years for the cases and 43.80 ± 8.38 years for the controls. In the case group, there were 54 cases of stage II, 89 cases of stage III, 58 cases of stage IV, 44 cases of unilateral side and 157 cases of bilateral side. No significant differences in the distributions of age, TC, TG, HDL‐C, LDL‐C, TG/HDL‐C, and LDL‐C/HDL‐C between the cases and the controls were observed from the statistical analysis. However, there was a significant difference in TC/HDL‐C between the cases and the controls.

Table 1

Characteristics of the participants

Variables	Mean ± SD		p value
	Cases (n = 201)	Controls (n = 201)
Age (years)	42.68 ± 12.88	43.80 ± 8.38	0.302
TC (mmol/L)	4.65 ± 0.92	4.66 ± 0.88	0.938
TG (mmol/L)	1.89 ± 1.28	2.10 ± 1.12	0.084
HDL‐C (mmol/L)	1.04 ± 0.24	1.08 ± 0.19	0.099
LDL‐C (mmol/L)	2.73 ± 0.85	2.71 ± 0.74	0.869
TC/HDL‐C	4.62 ± 1.15	4.38 ± 0.81	0.017*
TG/HDL‐C	1.98 ± 1.58	2.02 ± 1.14	0.759
LDL‐C/HDL‐C	2.70 ± 0.86	2.56 ± 0.72	0.086
Clinical stages
Stage II	54
Stage III	89
Stage IV	58
Hip lesions
Unilateral	44
Bilateral	157

TC, total cholesterol; TG, triglycerides; LDL‐C, low‐density lipoprotein‐cholesterol; LDL‐C, high‐density lipoprotein‐cholesterol.

p value was calculated by Independent samples t test.

p < 0.05 indicates statistical significance.

The basic information of all SNPs is shown in Table 2. The genotype distributions were in Hardy–Weinberg equilibrium for the case and control groups (p > 0.05). Two SNPs, rs7408174 and rs3745369 in RETN, were associated with the increased risk of alcohol‐induced ONFH (OR = 1.43, 95% CI: 1.03–1.98, p = 0.032; OR = 1.35, 95% CI: 1.01–1.81, p = 0.045). On the other hand, rs34861192 and rs3219175 in RETN showed reduced risk (OR = 0.63, 95% CI: 0.43–0.92; p = 0.016; OR = 0.67, 95% CI: 0.46–0.97; p = 0.032) of alcohol‐induced ONFH.

Table 2

Basic information of candidate SNPs in this study

SNP	Gene	Chromosome	alleles	MAF		p value	ORs	95% CI		p value
			A/B	case	control	for HWE
rs7408174	RETN	19	C/T	0.271	0.206	1.000	1.43	1.03	1.98	0.032*
rs34861192	RETN	19	A/G	0.135	0.199	0.826	0.63	0.43	0.92	0.016*
rs3219175	RETN	19	A/G	0.144	0.201	0.826	0.67	0.46	0.97	0.032*
rs3745367	RETN	19	A/G	0.373	0.423	0.885	0.81	0.61	1.08	0.150
rs3745368	RETN	19	A/G	0.173	0.138	0.547	1.30	0.88	1.91	0.181
rs3745369	RETN	19	C/G	0.381	0.313	0.254	1.35	1.01	1.81	0.045*
rs1477341	RETN	19	A/T	0.526	0.463	0.887	1.29	0.97	1.71	0.075
rs12611067	LDLR	19	T/G	0.319	0.321	0.194	0.99	0.74	1.34	0.951
rs14158	LDLR	19	A/G	0.400	0.388	0.882	1.05	0.79	1.40	0.718
rs2738464	LDLR	19	G/C	0.278	0.294	0.125	0.93	0.68	1.26	0.630
rs2738465	LDLR	19	G/A	0.485	0.495	0.480	0.96	0.73	1.27	0.778

Reference Sequence and version number (RETN: NG_023447.1; LDLR: NG_009060.1).

Abbreviations: SNP, single nucleotide polymorphism; HWE, Hardy‐Weinberg equilibrium; OR, odds ratio; CI, confidence interval; MAF, minor allele frequency.

p was calculated by Chi‐squared test.

p < 0.05 indicates statistical significance.

Genetic models were used to compare the SNP genotypes and the risk of alcohol‐induced ONFH. The results of logistic regression analysis for each genetic model are shown in Table 3. Four SNPs in RETN had strong associations with alcohol‐induced ONFH in genetic models after they were adjusted by age. It was discovered that rs7408174 was associated with increased risk of alcohol‐induced ONFH in dominant model (OR = 1.57, 95% CI: 1.05–2.34, p = 0.028) and log‐additive model (OR = 1.47, 95% CI: 1.05–2.06, p = 0.024). On the other hand, rs3219175 showed a decreased risk in dominant model (OR = 0.64, 95% CI: 0.42–0.97, p = 0.036) and log‐additive model (OR = 0.65, 95% CI: 0.44–0.95, p = 0.025). In the other two SNPs, rs34861192 showed a reduced risk in codominant model (AG: OR = 0.62, 95% CI: 0.39–0.96, p = 0.032), dominant model (OR = 0.59, 95% CI: 0.38–0.91, p = 0.017), and log‐additive model (OR = 0.61, 95% CI: 0.41–0.90, p = 0.013). Rs3745369 showed an increased risk in codominant model (CC: OR = 2.50, 95% CI: 1.28–4.90, p = 0.007), recessive model (CC: OR = 2.55, 95% CI: 1.35–4.81, p = 0.004), and log‐additive model (OR = 1.36, 95% CI: 1.01–1.82, p = 0.042).

Table 3

Analysis of the association between SNPs and alcohol‐induced ONFH risk in males

SNP	Model	Genotype	case	control	OR (95% CI)	p value
rs7408174	Codominant	T/T	105	126	1.00	0.050
		C/T	83	67	1.51 (1.00–2.29)
		C/C	13	8	2.02 (0.80–5.08)
	Dominant	T/T	105	126	1.00	0.028*
		C/T‐C/C	96	75	1.57 (1.05–2.34)
	Recessive	T/T‐C/T	188	193	1.00	0.246
		C/C	13	8	1.71 (0.69–4.23)
	Log‐additive	—	—	—	1.47 (1.05–2.06)	0.024*
rs34861192	Codominant	G/G	146	128	1.00	0.032*
		A/G	47	66	0.62 (0.39–0.96)
		A/A	3	7	0.36 (0.09–1.43)
	Dominant	G/G	146	128	1.00	0.017*
		A/G‐A/A	50	73	0.59 (0.38–0.91)
	Recessive	G/G‐G/A	193	194	1.00	0.211
		A/A	3	7	0.42 (0.11–1.64)
	Log‐additive	—	—	—	0.61 (0.41–0.90)	0.013*
rs3219175	Codominant	G/G	146	127	1.00	0.066
		A/G	52	67	0.67 (0.43–1.03)
		A/A	3	7	0.36 (0.09–1.42)
	Dominant	G/G	146	127	1.00	0.036*
		A/G‐A/A	55	74	0.64 (0.42–0.97)
	Recessive	G/G‐G/A	198	194	1.00	0.197
		A/A	3	7	0.41 (0.10–1.60)
	Log‐additive	—	—	—	0.65 (0.44–0.95)	0.025*
rs3745369	Codominant	G/G	80	91	1.00	0.007*
		C/G	80	94	0.97 (0.63–1.48)
		C/C	34	16	2.50 (1.28–4.90)
	Dominant	G/G	80	91	1.00	0.403
		C/G‐C/C	114	110	1.19 (0.80–1.77)
	Recessive	G/G‐G/C	160	185	1.00	0.004*
		C/C	34	16	2.55 (1.35–4.81)
	Log‐additive	—	—	—	1.36 (1.01–1.82)	0.042*

Abbreviations: SNP, single nucleotide polymorphism; OR odds ratio; CI, confidence interval.

p value adjusted for age was calculated by logistic regression.

p < 0.05 indicates statistical significance.

Correlation analysis between the genotypes and hip lesions, as well as clinical stages are shown in Table 4. The rs3745368 in RETN shows association with the clinical stages (p = 0.022). Comparisons of lipid levels between each genotype are shown in Table 5. The blood lipid levels of different SNP genotypes were compared by Analysis of Variance (ANOVA), but no difference was found.

Table 4

The association of genotypes in RETN and LDLR genes with the clinical phenotypes

Gene	SNP	genotype	Hip lesions		p	Clinical stages			p
			Unilateral	Bilateral		Stage II	Stage III	Stage IV
RETN	rs7408174	CC	5	8	0.268	2	7	4	0.884
		CT	19	64		22	36	25
		TT	20	85		30	46	29
	rs34861192	AA	0	3	0.436	1	2	0	0.698
		AG	8	39		14	18	15
		GG	34	112		38	68	40
	rs3219175	AA	0	3	0.543	1	2	0	0.489
		AG	10	42		14	19	19
		GG	34	112		39	68	39
	rs3745367	AA	3	23	0.217	7	11	8	0.338
		AG	20	78		32	38	28
		GG	21	56		15	40	22
	rs3745368	AA	3	4	0.396	5	1	1	0.022*
		AG	12	43		10	31	14
		GG	29	109		39	56	43
	rs3745369	CC	5	29	0.547	10	14	10	0.976
		CG	19	61		20	35	25
		GG	18	62		22	36	22
	rs1477341	AA	14	40	0.243	13	21	20	0.125
		AT	15	76		31	39	21
		TT	12	32		7	21	16
LDLR	rs12611067	GG	20	68	0.739	27	40	21	0.306
		TG	19	72		19	41	31
		TT	5	12		6	5	6
	rs14158	AA	9	20	0.409	9	8	12	0.158
		AG	20	83		23	49	31
		GG	15	54		22	32	15
	rs2738464	CC	27	72	0.19	28	40	31	0.784
		GC	14	71		23	38	24
		GG	2	10		3	7	2
	rs2738465	AA	13	37	0.489	15	18	17	0.621
		GA	24	83		26	50	31
		GG	7	37		13	21	10

Reference Sequence and version number (RETN: NG_023447.1; LDLR: NG_009060.1).

p value was calculated by Chi‐squared test.

p < 0.05 indicates statistical significance.

Table 5

Comparison of lipids levels between each genotype

SNP	TC (mmol/L)	TG (mmol/L)	HDL‐C (mmol/L)	LDL‐C (mmol/L)	TC/HDL‐C	TG/HDL‐C	LDL‐C/HDL‐C
rs7408174
CC (n = 13)	4.49 ± 0.88	1.76 ± 0.89	1.02 ± 0.26	2.66 ± 0.87	4.52 ± 0.85	1.88 ± 1.12	2.61 ± 0.74
CT (n = 83)	4.66 ± 0.95	2.07 ± 1.60	1.03 ± 0.22	2.75 ± 0.94	4.64 ± 1.08	2.19 ± 1.99	2.74 ± 0.90
TT (n = 105)	4.66 ± 0.91	1.77 ± 1.01	1.06 ± 0.25	2.72 ± 0.78	4.61 ± 1.25	1.82 ± 1.21	2.68 ± 0.85
p	0.821	0.247	0.688	0.923	0.934	0.289	0.828
rs34861192
AA (n = 3)	4.68 ± 0.48	1.37 ± 0.42	1.08 ± 0.14	2.30 ± 0.85	4.36 ± 0.23	1.32 ± 0.56	2.22 ± 1.01
AG (n = 47)	4.67 ± 1.02	2.02 ± 1.73	1.07 ± 0.26	2.78 ± 0.91	4.52 ± 1.25	2.08 ± 2.08	2.67 ± 0.92
GG (n = 146)	4.65 ± 0.91	1.86 ± 1.12	1.04 ± 0.23	2.73 ± 0.84	4.64 ± 1.14	1.94 ± 1.38	2.72 ± 0.85
p	0.992	0.59	0.619	0.627	0.765	0.68	0.589
rs3219175
AA (n = 3)	4.68 ± 0.48	1.37 ± 0.42	1.08 ± 0.14	2.30 ± 0.85	4.36 ± 0.23	1.32 ± 0.56	2.22 ± 1.01
AG (n = 52)	4.68 ± 1.00	1.96 ± 1.59	1.07 ± 0.26	2.74 ± 0.89	4.58 ± 1.26	2.01 ± 1.89	2.65 ± 0.89
GG (n = 146)	4.64 ± 0.90	1.88 ± 1.17	1.04 ± 0.23	2.73 ± 0.84	4.63 ± 1.13	1.98 ± 1.47	2.73 ± 0.85
p	0.952	0.718	0.720	0.677	0.886	0.760	0.546
rs3745368
AA (n = 7)	4.53 ± 1.00	1.46 ± 0.73	1.02 ± 0.40	2.67 ± 0.89	4.85 ± 1.47	1.72 ± 1.31	2.79 ± 0.79
AG (n = 55)	4.51 ± 0.81	1.74 ± 1.09	1.03 ± 0.18	2.70 ± 0.82	4.50 ± 1.02	1.81 ± 1.30	2.67 ± 0.76
GG (n = 138)	4.72 ± 0.96	1.97 ± 1.37	1.05 ± 0.25	2.75 ± 0.87	4.65 ± 1.19	2.05 ± 1.69	2.70 ± 0.91
p	0.326	0.355	0.727	0.940	0.628	0.578	0.931
rs3745369
CC (n = 34)	4.53 ± 0.68	1.60 ± 0.87	1.04 ± 0.26	2.74 ± 0.63	4.56 ± 1.08	1.68 ± 1.03	2.74 ± 0.71
CG (n = 80)	4.58 ± 1.00	1.96 ± 1.30	1.02 ± 0.22	2.65 ± 0.89	4.61 ± 1.17	2.09 ± 1.66	2.64 ± 0.83
GG (n = 80)	4.76 ± 0.94	1.84 ± 1.20	1.07 ± 0.25	2.77 ± 0.92	4.64 ± 1.20	1.90 ± 1.42	2.72 ± 0.96
p	0.364	0.345	0.548	0.662	0.942	0.366	0.798

Abbreviations: TC, total cholesterol; TG, triglycerides; LDL‐C, low‐density lipoprotein‐cholesterol; LDL‐C, high‐density lipoprotein‐cholesterol.

The Linkage analysis showed that two SNPs (rs34861192, rs3219175) in RETN (Figure 1) and three SNPs (rs14158, rs2738464, rs2738465) in LDLR exhibited significant linkage disequilibrium (Figure 2).

Figure 1

Haplotype block map for the seven SNPs in the RETN gene. Block 1 includes rs34861192 and rs3219175

Figure 2

Haplotype block map for the four SNPs in the LDLR gene. Block 1 includes rs14158, rs2738464, and rs2738465

In Table 6, the risk alleles of rs34861192 (p = 4.0 × 10–14, p = 6.4 × 10–8) and rs3219175 (p = 1.6 × 10–14, p = 1.9 × 10–9) were associated with increased expression of RETN gene in the whole‐blood and muscle‐skeletal. In contrast, rs2738464 (p = 3.4 × 10–5) was associated with decreased expression of LDLR gene in muscle‐skeletal.

Table 6

GTEx results for three SNPs in genes expression in the most relevant tissue

SNP	Gene	Effect	p‐value	Tissue
rs34861192	RETN	1.00	4.00E‐14	Whole‐Blood
rs34861192	RETN	1.00	6.40E‐08	Muscle‐Skeletal
rs3219175	RETN	0.84	1.60E‐14	Whole‐Blood
rs3219175	RETN	1.00	1.90E‐09	Muscle‐Skeletal
rs2738464	LDLR	−0.26	3.40E‐05	Muscle‐Skeletal

DISCUSSION

In this research, it was discovered that RETN genetic polymorphisms were associated with alcohol‐induced ONFH risk among Chinese Han individuals. The rs3745368 was associated with the stages of the disease, and more patients with AA genotype were in Stage II than those in Stage III and Stage IV. However, more patients with AG and GG genotypes were in stage III. The SNPs (rs7408174, rs34861192, and rs3219175) are located in the upstream of RETN and rs3745369 is located in the downstream. The rs34861192 is associated with the level of serum insulin, glycemic index and cholesterol (Zhou, Chen, Ji, Luo, & Luo, 2018). The cholesterol of the case group was measured but no association with this SNP was found. In orthopedic diseases, individuals with the C allele of the SNP rs7408174 and the AG or A allele of the SNP rs3219175 are at a higher risk of developing rheumatoid arthritis compared with wild‐type (Wang, Tang, et al., 2018). Plasma resistin level is strongly affected by rs34861192, rs3219175, and rs3745368 (Asano et al., 2010; Nakatochi et al., 2015). Using GTEx portal, RETN and LDLR expressions in different genotype individuals were compared and it was found that the risk alleles of rs34861192 and rs3219175 were associated with increased expression of RETN gene in the whole‐blood and muscle‐skeletal. We observed from Tables 2 and 3 that the number of GG genotypes in rs34861192 and rs3219175 in the case group was significantly higher than that in the control group, while the number of AG/AA genotypes was lower than that in the control group. The SNPs rs34861192 and rs3219175 located in the promoter region of RETN were associated with resistin levels. Moreover, the number of minor alleles of the two SNPs was negatively associated with DNAm level at cg02346997 (Nakatochi et al., 2015). We also discovered these two minor alleles reduce the risk of ONFH. Therefore, we hypothesized that different genotypes in these two SNPs affected the expression of resistin in bones, which then had an impact on the occurrence of ONFH. It was also found that the risk alleles of rs2738464 decreased the expression of LDLR in muscle‐skeletal, but no significant difference was observed between this genotype and alcohol‐induced ONFH.

By comparing the blood lipid levels between the case group and the control group, we discovered that TC/HDL‐C and LDL‐C/HDL‐C in the case groups were much higher than those in the control group, which showed the disorder of lipid metabolism in alcohol‐induced ONFH. This result may be related to the decrease of HDL‐C in the case groups and studies have found that resistin affects HDL‐C levels (Osawa et al., 2007). To identify the effect of the five SNPs on the metabolic disorder of alcohol‐induced ONFH, we examined blood lipid levels and analyzed their associations. However, no significant association was found due to two possible reasons: (a) Dietary differences in the subjects may have an impact on lipid levels; (b) For this analysis, only 201 cases of the population were studied, which resulted in a small number of each genotype, which may affect the results.

CONCLUSION

In summary, this study suggested that polymorphisms of RETN were associated with the alcohol‐induced ONFH. The SNPs (rs7408174, rs34861192, rs3219175, and rs3745369) in RETN were associated with the risk of alcohol‐induced ONFH. The rs3745368 was associated with the stage of the disease. The levels of TC/HDL‐C in the case group was significantly lower than that in the control group. This study provide new insights to facilitate early diagnosis and early prevention of ONFH, as well as for new candidate gene studies. The sample size will be increased to stud the mechanism of RETN action in our future research.

CONFLICT OF INTEREST

The authors declare no conflict of interest.