The Effect of AMBP SNPs,Their Haplotypes,and Gene-Environment Interactions on the Risk of Atherothrombotic Stroke Among the Chinese Population

Abstract

Background/Objectives:

Ischemic stroke (IS) is a severe and complex disorder with high morbidity and mortality rates and it has been associated with both environmental and genetic predisposing factors. The purpose of this study was to evaluate the association of the alpha-1-microglobulin/bikunin precursor (AMBP) gene polymorphisms with IS and any possible interactions between specific AMBP alleles and traditional risk factors among a Han Chinese cohort.

Materials and Methods:

We conducted a candidate gene study designed to characterize nine (9) single nucleotide polymorphisms (SNPs) of the AMBP gene among 195 patients with atherothrombotic stroke (ATS) (a major subtype of IS) and 184 nonstroke controls. Allelic and genotypic frequency differences were evaluated using a logistic regression model. False discovery rate (FDR) correction for multiple comparisons was used. The interactional analyses were performed using the multifactor dimensionality reduction test.

Results:

We found an association between the rs2567698 CC genotype (odds ratio [OR], 95% confidence interval [CI]: 2.176, 1.159-4.086) and the T allele (OR, 95% CI: 0.654, 0.446-0.960) with risk of ATS in men. However, these associations did not survive FDR correction. In haplotype analyses, the GCCCCCCCC haplotype had a higher frequency (OR, 95% CI: 2.191, 1.048-4.580) in ATS in the ≥45 years of age subgroup, whereas the GCCTCCCCC haplotype decreased the risk for ATS (OR, 95% CI: 0.543, 0.345-0.853) in men. In addition, we also found interactions for ATS risk between SNPs in the AMBP gene and modifiable risk factors for ATS, including: rs11788411 and hypertension in the overall population and women; rs2251680 and hypertension in subjects aged 45 years and older, as well as the interaction among hypertension and the rs2567698 and rs10817564 genotypes in men.

Conclusion:

Our results show a possible association between AMBP SNP haplotypes and gene-environment interactions with ATS susceptibility in a Han Chinese cohort.

Introduction

Stroke is associated with high morbidity, mortality, and disability-adjusted life years, as well as a high economic burden, and it is the primary neurological cause of acquired disability in adults and a leading cause of death (Wang et al., 2016; Hay et al., 2017). In China, the cerebrovascular diseases (CVDs) were recognized as the first leading cause of death in 2013 (Zhou et al., 2016), and the adjusted annual incidence and mortality rates of stroke were 246.8 and 114.8 per 100,000 inhabitants (Wang et al., 2017). The total medical cost for inpatients with stroke was about 75.6 billion Yuan (RMB) in 2015 (Chen et al., 2017). What's more, the hospitalization expenses are projected to increase significantly in the future (Huo et al., 2017).

Ischemic stroke (IS) is the most prevalent type of stroke, accounting for 43.7-78.9% of all cases in China, while the proportion is 67.3-80.5% in developed countries (Liu et al., 2007). IS is a complex disorder associated with modifiable and genetic predisposing factors (Dichgans, 2007). Thrombosis of brain arteries secondary to atherosclerosis is considered one of the major pathophysiological mechanisms and stroke subtypes of IS (Humphries and Morgan, 2004). The genetic susceptibility to atherosclerosis has been extensively studied.

In 2009, scientists reported a novel stroke susceptibility locus in a pedigree from Northern Sweden. They found a highly significant linkage of stroke to chromosome 9q31-q33 by the genome-wide scan (Janunger et al., 2009). Of this region, a gene called the alpha-1-microglobulin/bikunin precursor (AMBP) could be implicated in the atherosclerosis. It encodes both α-1-microglobulin (A1M) and bikunin (urinary trypsin inhibitor [UTI]), which exhibit immunosuppressive functions and anti-inflammatory effects and, thus, play a significant role in inflammatory and atherosclerosis processes (Kobayashi, 2006). Dysfunction in the gene could affect vascular integrity, then resulting in the increased susceptibility to IS in carriers (Kobayashi, 2006).

Genetic susceptibility is hypothesized to play a critical role in the pathogenesis of IS patients aged <70 years in comparison to older counterparts (Dichgans, 2007), and most of population in China is the Han nationality. Therefore, our objective was to evaluate the association of AMBP gene with IS, and the interaction between genetic variants and traditional risk factors for IS using single nucleotide polymorphisms (SNPs) as genetic marker in a Han Chinese cohort aged 19-70 years.

Materials and Methods

Subjects

The detailed original samples have been described previously (Liu et al., 2013). Briefly, all IS patients in the study were consecutively recruited between August 2010 and December 2011 from the affiliated Nanchong Central Hospital of North Sichuan Medical College. IS was diagnosed using the World Health Organization (WHO) criteria (Asplund et al., 1988), and we recruited only patients with atherothrombotic stroke (ATS), defined using modified “The Trial of Org 10172 in Acute Stroke Treatment” (TOAST) classification of Han (KmTOAST) in 2007 (Han et al., 2007). We excluded the IS patients secondary to other diseases, such as immunological disorders, coagulopathies, thyroid disease, tuberculosis, or malignant tumor, as well as subjects with severe liver/kidney function failure, thrombolytic therapy, or pregnancy.

Nonstroke Han Chinese individuals without blood ties to a case were selected as controls, including inpatients with minor illness from other departments of the same hospital or visitors to patients. Controls were free of any vascular or neurological diseases and follow the same exclusion criteria as the cases.

The study was reviewed and approved by the local ethics committee. All subjects provided written informed consent.

Baseline examination

The demographic characteristics and possible stroke risk factors, as well as medical history, were collected from each individual. Cigarette smoking was defined as having smoked consecutively at least one cigarette per day for 1 year or more (Kelly et al., 2008), and the presence of alcohol consumption was defined as drinking alcohol at least 12 times during the past year (Kelly et al., 2008). The diagnosis of hypertension was based on the WHO/International Society of Hypertension guidelines (blood pressure ≥140/90 mmHg) or the use of antihypertensive drugs (WHO et al., 1999). Diabetes mellitus was diagnosed based on high fasting plasma glucose (≥7.0 mM) or an oral hypoglycemic agent or insulin treatment history according to WHO criteria (Alberti and Zimmet, 1998). Dyslipidemia was defined based on Chinese guidelines (Joint Committee for Developing Chinese guidelines on Prevention and Treatment of Dyslipidemia in Adults, 2007). The overweight was presented if the body mass index is 24 kg/m² or more according to the Chinese criteria (Chen and Lu, 2004). Notably, because of the ethnic differences between Chinese and Euro-American people, the definition of overweight in China is different from WHO.

Picking SNPs and genotyping

We analyzed the tagging SNPs from the National Center for Biotechnology Information Bulletin 36 (NCBI) database and HapMap database (HapMap Data Phase III/Rel#2, Feb09, on NCBI36 B36 assembly, dbSNP b126) (International HapMap Consortium, 2005), and we used the genotype data from Han Chinese. SNPs in the AMBP gene with an r² > 0.8 and a minor allele frequency >0.05 were selected as genetic marker for further association analyses, and we also included the SNPs located in 5′- and/or 3′-flanking or untranslated regions (UTRs). A total of nine tagging SNPs (rs11788411, rs3762058, rs2251680, rs2567698, rs16912311, rs12377342, rs10817564, rs10817563, and rs3860175) were identified within a 26-kb region spanning AMBP (including 5′- and/or 3′-flanking or UTRs; chromosome 9, positions 115860000 … 115883000).

Genomic DNA of all samples was obtained as described elsewhere (Liu et al., 2013). The selected SNPs were genotyped in all subjects using SNaPshot method from Applied Biosystems Company (ABI, Foster City, CA) with the technical support from Shanghai Generay Biotech.

A brief description of the programs is shown below. The polymerase chain reaction (PCR) was first performed containing 10 × buffer, 25 mM MgCl₂, dNTPs, and Taq polymerase (R0192, EP0406; Thermo Fisher Scientific (China) Co., Ltd, Beijing, China), genomic DNA, and primers (Table 1). The program included the first step of 95°C for denaturation for 3 min, then followed by the first 11 cycles (94°C, 15 s; 60°C, 15 s (the temperature was decreased by 0.5°C in each consecutive cycle); 72°C, 30 s), followed by the second 24 cycles (95°C, 15 s; 54°C, 15 s; 72°C, 30 s), and a final extension step (72°C, 3 min). And then, the purifying of the PCR products was performed containing ExoI (EN0582), FastAP^™ thermosensitive alkaline phosphatase (EF0652; Thermo Fisher Scientific (China) Co., Ltd), and ExoI buffer following the program (37°C, 15 min; 80°C, 15 min).

Table 1.

The Primers in Original PCR and the Extension Reaction

SNPs	Primer name	Primer sequence (5′-3′)
rs11788411	F	AAATCATCCTACTAAGCCCGC
	R	CAACATCCTGGACTTATCCTC
	EP	TTTTTTTTTTTTTTTTTTTTTTTTTTTTACCTTCATGTTGGGCCTGC
rs3762058	F	TCACACTCTTGCCAGCAATAC
	R	GTTTCTGCACTGACTACATGG
	EP	TTTTTTTTTTTTTTTAGATTTAACCACTGGAGAA
rs2251680	F	GATCCTTTGAAGGTACAGCTC
	R	TCATTAGAGAAGGAAGGGCTG
	EP	TTTTTTTTTTTTTTTTTTTTTTATAGAATGTCAGCCAAGAG
rs2567698	F	GGATACATGAGAGAGCAAGAC
	R	GTGGATTCTTCAGAACTCAGG
	EP	TTTTTTTTTTTTTTTTTTTTTTACCTGCCTGCCTGACCCAC
rs16912311	F	TGAGAAGTTGACTAGCCAGAG
	R	TACCTGGCTGCATAAGGATG
	EP	TTTTTTTTTTTTTTTTTTTCAGTGGGGAGGCACGAGCA
rs12377342	F	CAAAAACACCACCTGGCAAAG
	R	GAAACTCTGTGGCAACTCTTC
	EP	TTTTTTTTTTTTTTTTTTTATTATGGAGACAGAGAAGA
rs10817564	F	TGTCATCAGTGTCACTCAAGG
	R	TCCTCTGTCAGGAAAGGAAAG
	EP	TTTTTTTTTTTTTTTTTTTTTTTTTACACAAGACATTTTGGAAC
rs10817563	F	ACGGGAAATAGAAGTGTGCTG
	R	TCCTTGAGTGACACTGATGAC
	EP	TTTTTTTTTTTTTTTTTTTTTTTTTATTTCTTTATCAATGTCAA
rs3860175	F	TCCTCAAATCCTTGTCTCCAG
	R	CCAGCTTTGTGTGTTCTGAAG
	EP	TTTTTTTTTTTAGAAGAGGAGGTAGGGCCC

EP, extension primer in the extension reaction; F, forward primer for PCR; PCR, polymerase chain reaction; R, reverse primer for PCR; SNP, single nucleotide polymorphism.

Third, the extension reaction was performed following the ABI SNaPshot protocol, containing purified PCR product, SNaPshot multiplex mix containing Taq polymerase and fluorescently labeled dideoxy-NTPs (ABI), mixed extension primer (Table 1), and ddH₂O. The program included the initial denaturation (96°C, 1 min) and the following 28 cycles (96°C, 10 s; 52°C, 5 s; 60°C, 30 s). After that, the extension product was incubated with FastAP (EF0652) following the program (37°C, 15 min; 80°C, 15 min) to degrade the fluorescent dNTPs unincorporated. Finally, the purified extension product with deionized formamide and electrophoresed on an ABI 3730XL genetic analyzer (ABI) using the parameters from the SNaPshot protocol, and then, we used the GeneMapper Software v4.0 (ABI) for analyzing the separated products.

Statistical analyses

Continuous variables were expressed as means and standard deviations, and the differences of variables between groups were assessed using t-test or the Mann-Whitney U test. The chi-square test (χ²) was applied for evaluating proportions of clinical and environmental variables, Hardy-Weinberg equilibrium (HWE), as well as subsequent genetic association analyses were performed in three genetic models (dominant, recessive, and allelic comparison) (Liu et al., 2013). Univariate and multivariate logistic regression analysis with adjustment of traditional risk factors for stroke (such as demography variable, medical history, as well as habits and customs) was performed to obtain the crude and adjusted odds ratios (ORs) and 95% confidence intervals (CIs) for the risk genotypes. Variables with a p-value <0.1 in univariate analysis were entered into the multivariate models. The false discovery rate (FDR) method by Benjamini-Hochberg (Benjamini and Hochberg, 1995) was used for multiple testing correction.

Haplotype was constructed using the genetic statistical software SHEsis (Shi and He, 2005). Multifactor dimensionality reduction (MDR) software was used for detecting the possible interactions between genetic and environmental risk factors (Hahn et al., 2003) with parameters, including cross-validation consistency (CVC), the testing balanced accuracy (TBA), and the sign test. The final candidate model of interactions was selected based on maximum CVC and maximum TBA, in which p-values were estimated by a permutation test under 1,000 times. Logistic regression was applied for evaluating the significant interaction in ATS risk.

Analyses were stratified by gender and age (<45 or ≥45 years) for detecting subpopulation specific variants (Pezzini, 2012). SPSS software version 19.0 (SPSS, Inc., Chicago, IL) was applied to all statistical analyses unless indicated otherwise. A p-value threshold of 0.05 was considered significant.

Results

Participant characteristics

During the study period, 823 IS patients in total were admitted to our hospital, and 431 (52.4%) were classified as ATS. Of these, we excluded 229 with history of CVD, thrombolytic therapy, severe liver/kidney disease, pregnancy, or other severe general diseases. A total of 202 cases and 188 controls were enrolled into our study based on the inclusion and exclusion criteria. Genotyping failure was seen in seven SNPs in cases and four in controls and these were excluded. The final analyses include 195 cases with ATS and 184 control subjects. Background characteristics of the subjects are given in Table 2.

Table 2.

Baseline Demographic and Clinical Characteristics of the Studied Population

Variables	ATS (195)	Control (184)	OR, 95% CI, p
Variables	ATS (195)	Control (184)	Univariable analysis		Multivariable analysis
Age (years)	60.63 ± 8.18	56.32 ± 11.44	—	—	—	—
Age^a (years)	62 (56, 68)	58 (48, 66)	—	0.001^b	—	—
Age: ≤44/45-64/≥65 years	10 (5.1)/104 (53.4)/81 (41.5)	32 (17.4)/95 (51.6)/57 (31.0)	—	0.000^b	—	0.020^b
Sex: men/women	125 (64.1)/70 (35.9)	97 (52.7)/87 (47.3)	1.602, 1.061-2.417	0.025^b	1.634, 0.967-2.762	0.067
Body mass index (kg/m²)	23.52 ± 2.56	23.33 ± 2.19	—	0.421	—	—
Overweight	77 (41.8)	60 (32.6)	1.349, 0.885-2.055	0.164	—	—
Smoking	61 (31.2)	37 (20.1)	1.809, 1.130-2.896	0.013^b	1.219, 0.664-2.238	0.523
Alcohol	47 (24.1)	29 (15.8)	1.697, 1.014-2.840	0.043^b	1.727, 898-3.321	0.101
Hypertension	117 (60.0)	35 (19.0)	6.386, 4.005-10.181	0.000^b	6.337, 3.823-10.505	0.000^b
Diabetes	41 (21.0)	10 (5.4)	4.632, 2.245-9.560	0.000^b	3.374, 1.526-7.457	0.003^b
Dyslipidemia	74 (37.9)	60 (32.6)	1.264, 0.828-1.929	0.277	—	—
CHD	14 (7.2)	4 (2.2)	3.481, 1.124-10.777	0.022^b	6.282, 1.785-22.105	0.004^b

Unless stated otherwise, data are presented as n (%) or mean ± standard deviation values.

Showing the median and interquartile range due to the non-normal distribution of the data.

Statistically significant.

ATS, atherothrombotic stroke; CHD, coronary heart disease; CI, confidence interval; OR, odds ratio.

As were shown in Table 2, the cases with ATS had a mean age of 60.63 ± 8.18 years (the median was 62 years), and the controls had age of 56.32 ± 11.44 years (median, 58 years). After adjustment of confounders, it was found that the subjects were older and presented higher frequency of hypertension, diabetes, and coronary heart disease (CHD) in the ATS group than in controls.

Individual SNP analysis

All SNP frequencies followed the HWE in controls (p > 0.05) and met quality control criteria. Only in men, we observed an increased risk for ATS in carriers with the CC genotype compared with those carrying CT and TT genotype combined (CC vs. CT+TT genotype) in rs2567698 (adjusted OR = 2.176; 95% CI: 1.159-4.086; p = 0.016), and the risk for ATS decreased in subjects with the T allele (T vs. C allele) with an adjusted OR (95% CI) of 0.654 (0.446-0.960) (p = 0.030). In addition, the A allele of rs10817563 also showed a marginal significant association with ATS risk in men (adjusted OR, 95% CI: 1.485, 1.006-2.192; p = 0.046). However, all significant associations did not survive FDR correction. In addition, we did not find any relationship of all nine SNPs to ATS in overall samples or women or ≥45-year subjects for all genetic comparisons. These data are summarized in Table 3.

Table 3.

Frequency of the Alleles and Genotypes from the SNPs in Cases and Controls and Results of Genetic Association Analyses

SNPs	Populations	Cases				Controls				Genetic models (OR, 95% CI, p^a/p-_FDR)
		Genotypes			MA	Genotypes			MA	D	R	A
		CC	CT	TT	T	CC	CT	TT	T	D	R	A
rs2251680	All	52	96	47	190	42	102	40	182	1.155, 0.668-1.998, 0.606/0.676	1.427, 0.821-2.480, 0.208/0.607	0.971, 0.730-1.291, 0.839/0.944
	≥45	43	93	49	191	34	82	36	154	1.062, 0.606-1.861, 0.833/0.869	1.324, 0.744-2.355, 0.339/0.624	1.039, 0.767-1.408, 0.803/0.803
	Women	19	32	19	70	13	51	23	97	1.867, 0.735-4.741, 0.189/0.810	1.185, 0.518-2.711, 0.688/0.695	0.794, 0.508-1.240, 0.310/0.848
	Men	33	64	28	120	29	51	17	85	0.789, 0.405-1.538, 0.486/0.729	1.704, 0.811-3.579, 0.159/0.636	1.184, 0.812-1.725, 0.380/0.570
rs3860175	All	98	80	17	114	104	70	10	90	0.773, 0.482-1.239, 0.285/0.641	1.775, 0.705-4.468, 0.223/0.607	1.276, 0.924-1.762, 0.139/0.401
	≥45	93	76	16	108	84	60	8	76	0.858, 0.529-1.391, 0.534/0.869	1.787, 0.680-4.693, 0.239/0.624	1.237, 0.878-1.743, 0.224/0.572
	Women	31	33	6	45	44	37	6	49	0.862, 0.412-1.802, 0.693/0.953	1.514, 0.394-5.826, 0.546/0.695	1.208, 0.744-1.962, 0.444/0.848
	Men	67	47	11	69	60	33	4	41	0.716, 0.391-1.310, 0.278/0.500	2.621, 0.736-9.336, 0.137/0.636	1.423, 0.914-2.214, 0.117/0.211
rs12377342	All	160	32	3	38	152	28	4	36	0.877, 0.475-1.622, 0.676/0.676	0.460, 0.066-3.219, 0.434/0.607	0.996, 0.616-1.609, 0.986/0.986
	≥45	152	31	2	35	122	26	4	34	1.054, 0.567-1.957, 0.869/0.869	0.243, 0.032-1.869, 0.174/0.624	0.830, 0.504-1.366, 0.462/0.594
	Women	58	11	1	13	72	12	3	18	0.782, 0.304-2.012, 0.610/0.953	0.617, 0.055-6.878, 0.695/0.695	0.887, 0.419-1.880, 0.754/0.848
	Men	102	21	2	25	80	16	1	18	0.896, 0.411-1.954, 0.782/0.934	0.524, 0.021-13.178, 0.694/0.821	1.086, 0.574-2.055, 0.799/0.899
rs16912311	All	67	95	33	161	76	80	28	136	0.736, 0.452-1.198, 0.218/0.641	1.050, 0.556-1.981, 0.880/0.880	1.199, 0.895-1.607, 0.223/0.401
	≥45	64	92	29	150	63	65	24	113	0.781, 0.475-1.285, 0.331/0.869	1.030, 0.533-1.990, 0.931/0.931	1.152, 0.844-1.574, 0.372/0.572
	Women	27	30	13	56	31	42	14	70	1.041, 0.492-2.204, 0.917/0.953	1.366, 0.525-3.557, 0.523/0.695	0.990, 0.629-1.560, 0.967/0.967
	Men	40	65	20	105	45	38	14	66	0.539, 0.290-1.002, 0.051/0.230	1.007, 0.436-2.328, 0.987/0.987	1.404, 0.952-2.072, 0.087/0.196
rs2567698	All	79	85	31	147	61	89	34	157	1.493, 0.916-2.434, 0.108/0.641	0.801, 0.429-1.497, 0.487/0.607	0.813, 0.608-1.087, 0.163/0.401
	≥45	73	84	28	140	50	75	27	129	1.440, 0.871-2.382, 0.155/0.869	0.828, 0.430-1.595, 0.573/0.764	0.826, 0.606-1.125, 0.225/0.572
	Women	23	35	12	59	33	39	15	69	0.958, 0.452-2.030, 0.910/0.953	0.699, 0.263-1.859, 0.472/0.695	1.108, 0.705-1.743, 0.656/0.848
	Men	56	50	19	88	28	50	19	88	2.176, 1.159-4.086, 0.016^b/0.144	0.749, 0.341-1.646, 0.472/0.821	0.654, 0.446-0.960, 0.030^b/0.196
rs3762058	All	141	46	8	62	125	53	6	65	1.337, 0.799-2.236, 0.269/0.641	1.545, 0.452-5.281, 0.488/0.607	0.881, 0.602-1.290, 0.515/0.662
	≥45	134	44	7	58	101	45	6	57	1.385, 0.820-2.340, 0.224/0.869	1.154, 0.331-4.030, 0.822/0.931	0.806, 0.539-1.204, 0.291/0.572
	Women	53	14	3	20	58	26	3	32	1.586, 0.699-3.601, 0.270/0.810	1.928, 0.305-12.193, 0.486/0.695	0.740, 0.402-1.360, 0.331/0.848
	Men	88	32	5	42	67	27	3	33	1.074, 0.561-2.056, 0.830/0.934	1.580, 0.319-7.837, 0.575/0.821	0.985, 0.597-1.625, 0.953/0.953
rs10817564	All	70	81	44	169	78	73	33	139	0.824, 0.509-1.334, 0.431/0.647	1.264, 0.701-2.279, 0.436/0.607	1.260, 0.942-1.685, 0.119/0.401
	≥45	66	77	42	161	63	63	26	115	0.878, 0.534-1.443, 0.608/0.869	1.364, 0.740-2.513, 0.320/0.624	1.266, 0.929-1.726, 0.135/0.572
	Women	28	25	17	59	33	38	16	70	0.978, 0.464-2.059, 0.953/0.953	1.793, 0.738-4.355, 0.197/0.695	1.082, 0.689-1.701, 0.732/0.848
	Men	42	56	27	110	45	35	17	69	0.614, 0.333-1.134, 0.119/0.268	1.151, 0.536-2.471, 0.718/0.821	1.423, 0.968-2.093, 0.072/0.196
		GG	GA	AA	A	GG	GA	AA	A
rs11788411	All	178	15	2	19	171	13	0	13	0.680, 0.279-1.659, 0.397/0.647	—	1.399, 0.681-2.874, 0.359/0.539
	≥45	169	15	1	17	140	12	0	12	0.851, 0.352-2.060, 0.721/0.869	—	1.172, 0.551-2.493, 0.680/0.765
	Women	66	3	1	5	83	4	0	4	0.412, 0.089-1.919, 0.259/0.810	—	1.574, 0.415-5.976, 0.502/0.848
	Men	112	12	1	14	88	9	0	9	0.990, 0.358-2.732, 0.984/0.984	—	1.219, 0.516-2.879, 0.650/0.836
		CC	CA	AA	A	CC	CA	AA	A
rs10817563	All	70	91	34	159	76	83	25	133	0.880, 0.543-1.428, 0.606/0.676	1.231, 0.642-2.360, 0.531/0.607	1.216, 0.907-1.631, 0.191/0.401
	≥45	66	87	32	151	60	70	22	114	0.956, 0.581-1.575, 0.861/0.869	1.337, 0.690-2.591, 0.390/0.624	1.149, 0.842-1.569, 0.381/0.572
	Women	30	28	12	52	32	42	13	68	1.216, 0.579-2.553, 0.605/0.953	1.326, 0.492-3.570, 0.577/0.695	0.921, 0.582-1.457, 0.725/0.848
	Men	40	63	22	107	44	41	12	65	0.594, 0.320-1.104, 0.100/0.268	1.458, 0.625-3.402, 0.383/0.821	1.485, 1.006-2.192, 0.046^b/0.196

Genetic models (D) mean dominant, (R) for recessive, and (A) for allelic comparison model. That is to say, for an SNP with MA a and major allele A, D model means A/A versus both a/A and a/a combined, R means a/a versus both a/A and A/A genotypes combined, and A means a versus A allele.

Adjusted p-value for D and R (adjusted for age, gender, smoking, alcohol, hypertension, diabetes, and CHD).

Statistically significant.

FDR, false discovery rate; MA, minor allele; p-_FDR, the p-value with the FDR correction.

Haplotype analysis

All nine loci (the order of SNPs listed being rs11788411-rs3762058-rs2251680-rs2567698-rs16912311-rs12377342-rs10817564-rs10817563-rs3860175) were used for haplotype construction (the frequency of haplotypes constructed is at least 0.03) in the overall population, women, men, or subjects aged ≥45 years.

In ≥45 years group, it was found that the frequency of the GCCCCCCCC haplotype was higher in cases than in controls (6.8% vs. 3.5%, OR, 95% CI: 2.191, 1.048-4.580; p < 0.05), and the frequency of the GCCTCCCCC haplotype had a marginal p-value between two groups (OR, 95% CI: 0.693, 0.482-0.998, p = 0.048), showing a trend of protection against ATS. However, in men, the frequency of the GCCTCCCCC haplotype was lower in the cases (OR = 0.543, 95% CI: 0.345-0.853; p < 0.05), demonstrating a protective effect (Table 4). No significant differences were found between cases and controls in haplotype distribution in either the overall population or other subpopulation groups. The following Table 4 showed the haplotypes only with the p-value <0.10.

Table 4.

Frequency of Constructed Haplotypes in Cases and Controls

Populations	Haplotypes	ATS (freq.)	Control (freq.)	Chi²	Fisher's p	Pearson's p	OR (95% CI)
Overall population	GCCTCCCCC	82.57 (0.212)	109.17 (0.297)	3.494	0.061662	0.061615	0.715 (0.503-1.017)
	GCCTCCCCT	14.01 (0.036)	6.94 (0.019)	2.936	0.086700	0.086641	2.205 (0.874-5.563)
	GCTCTCTAT	70.60 (0.181)	57.99 (0.158)	2.724	0.098887	0.098823	1.395 (0.939-2.072)
≥45 years	GCCCCCCCC	25.25 (0.068)	10.66 (0.035)	4.531	0.033331^a	0.033302^a	2.191 (1.048-4.580)
	GCCTCCCCC	82.32 (0.222)	92.38 (0.304)	3.902	0.048278^a	0.048240^a	0.693 (0.482-0.998)
	GCCTCCCCT	13.25 (0.036)	4.23 (0.014)	3.708	0.054207	0.054165	2.843 (0.938-8.617)
Women	GCCTCCCCT	7.80 (0.056)	4.00 (0.023)	2.770	0.096089	0.096026	2.746 (0.800-9.427)
Women	GTTCTCTAC	7.68 (0.055)	19.78 (0.114)	2.756	0.096926	0.096863	0.482 (0.201-1.157)
Men	GCCCTCTAC	7.65 (0.031)	1.22 (0.006)	3.534	0.060193	0.060148	5.247 (0.767-35.907)
Men	GCCTCCCCC	52.61 (0.210)	63.55 (0.328)	7.082	0.007811^a	0.007804^a	0.543 (0.345-0.853)

The haplotype is listed in the order rs11788411-rs3762058-rs2251680-rs2567698-rs16912311-rs12377342-rs10817564-rs10817563-rs3860175, only the haplotypes with p < 0.10 were shown; ATS (freq.) or Control (freq.): it means the number of cases or controls estimated to have haplotypes constructed, and this haplotype's frequency in case or controls; analogous meaning for other cells.

Statistically significant (p < 0.05).

Interaction of AMBP gene and modifiable risk factors with risk of ATS

We used the MDR software to detect the best interaction combinations among SNPs of AMBP gene and modifiable risk factors for ATS, which with a p ≤ 0.10 will be entered into the analyses. We observed a significant interaction between rs11788411 and hypertension (p < 0.05) or among rs2251680, rs3860175, rs10817563, and diabetes (p < 0.05) in overall population, and the interactions between rs11788411 and hypertension (p < 0.05) in women or among rs2567698, rs10817564 and hypertension (p < 0.05) in men, or between rs2251680 and hypertension (p < 0.05) in ≥45 years group. We summarized the best interaction model in Table 5.

Table 5.

The Interactions Between AMBP Gene and Modifiable Risk Factors in Different Populations by MDR

Populations	Interactions	Best model ^a	Training BA	Testing BA	CVC	p	p′
Overall population	Gene-H	Hypertension-rs11788411	0.7076	0.6995	9/10	0.0125	0.0000-0.0010^b
Overall population	Gene-D	Diabetes-rs2251680-rs3860175-rs10817563	0.6671	0.6058	9/10	0.1748	0.0150-0.0160^b
≥45 years	Gene-H	Hypertension-rs2251680	0.6925	0.6762	7/10	0.0398	0.0000-0.0010^b
≥45 years	Gene-D	Diabetes-rs2567698-rs10817564	0.6286	0.5502	5/10	0.5481	0.4220-0.4230
Women	Gene-H	Hypertension-rs11788411	0.7289	0.715	10/10	0.0874	0.0000-0.0010^b
Men	Gene-H	Hypertension-rs2567698-rs10817564	0.7363	0.678	8/10	0.0955	0.0020-0.0030^b
	Gene-D	Diabetes-rs2567698	0.6233	0.5384	8/10	0.7199	0.6240
	Gene-A	Age-rs2251680-rs3860175-rs2567698-rs3762058-rs10817564	0.7691	0.5291	10/10	0.7841	0.6980

p′: with a 1,000 times permutation test.

Only best models available were shown.

Statistically significant (p < 0.05).

A, age; AMBP, alpha-1-microglobulin/bikunin precursor; BA, balanced accuracy; CVC, cross-validation consistency; D, diabetes; H, hypertension; MDR, multifactor dimensionality reduction.

Logistic regression analysis was used to assess the role of the significant interactions in ATS risk. After adjusting covariates, the carriers with G-allele (GG plus AG genotype) of rs11788411 and hypertension had a 6.249 times risk for ATS than subjects carrying AA genotype alone with no hypertension (OR = 6.249, 95% CI: 3.770-10.360; p < 0.001) in overall population; however, no significant differences were found for interaction model (rs2251680, rs3860175, rs10817563 and diabetes) between the high risk block and low risk one in MDR chart. In people aged 45 years or older, the subjects carrying the C-allele (CC+CT genotype) of rs2251680 with hypertension had 4.154-fold increased risk for ATS than one carrying TT genotype alone without hypertension (OR = 4.154, 95% CI: 2.421-7.130; p < 0.001). In women, we found that the carriers with rs11788411 GG genotype and hypertension had a higher risk for ATS (OR = 6.754, 95% CI: 3.112-14.654; p < 0.001) than subjects carrying A-allele (GA plus AA genotype) with no hypertension. In men, we only estimated the impact of hypertension and rs2567698 plus rs10817564 on ATS in recessive genetic model (CC+CT vs. TT genotype), and the data showed that the combination of hypertension with the CC plus CT genotype of rs2567698 and rs10817564 increased the susceptibility to ATS compared to nonhypertension with the TT genotype alone (OR = 3.828, 95% CI: 1.816-8.071; p < 0.001).

Discussion

This study aimed to evaluate the role of the AMBP gene, as well as its interaction with modifiable risk factors in ATS risk in a Han Chinese cohort. We observed that the CC genotype of rs2567698 increased the ATS risk, whereas the T allele played a protective role in rs2567698 in men. However, the SNP did not surpass FDR correction for multiple testing. Meanwhile, the GCCCCCCCC haplotype exhibited an elevated risk for ATS among those aged ≥45 years, and the GCCTCCCCC haplotype showed an opposite effect in men.

The AMBP gene has been localized to chromosome 9q32-q34 (Salier et al., 1992), consisting of 10 exons. It encodes a precursor for two polypeptides, A1M by the first six of exons and UTI by the last four (Vetr and Gebhard, 1990).

A1M is a low-molecular weight protein acting as a physiological antioxidant with powerful cell- and tissue-protective properties that may be implicated with atherosclerosis (Olsson et al., 2013; Åkerström and Gram, 2014). Furthermore, A1M also can protect low density lipoprotein against oxidative damage caused by myeloperoxidase (MPO) and hydrogen peroxide (Cederlund et al., 2015). The MPO and its intermediates have been found in atherosclerotic lesions and to be associated with atherosclerosis (Daugherty et al., 1994).

UTI is another protein encoded by the AMBP gene that may have anti-inflammatory properties. It can prevent the release of pro-inflammatory cytokines in macrophages stimulated with lipopolysaccharide (LPS) (Wakahara et al., 2005) and LPS-induced neutrophil activation and cytokine (Kanayama et al., 2007). Moreover, UTI can also inhibit calcium influx and extracellular signal-regulated kinase signaling using LPS receptors and/or other UTI signaling receptors. Deficits in the signaling cascades can downregulate cytokine expression, which render macrophages/neutrophils more inactive, and impair inflammatory processes (Kobayashi, 2006).

These findings hinted that the AMBP gene might be implicated in the pathophysiology of atherosclerosis and then playing a role in atherosclerotic diseases, such as ATS, coronary heart disease, and so on.

We found interactions in ATS susceptibility between SNPs from AMBP gene and modifiable risk factors for ATS among different populations. Gene-environment interactions are increasingly hypothesized as a driving force for IS risk across different populations. Many methods were used to detect the significance to explore this hypothesis. MDR, enjoying great popularity since its first introduction in 2001, was used to detect the interactions in our study (Ritchie et al., 2001). The AMBP gene or the encoded protein is significantly associated with vascular risk factors such as diabetes and hypertension.

Previous studies showed that urinary A1M was related to the duration, severity, and control of diabetes (Fiseha and Tamir, 2016; Saif and Soliman, 2017), as well as is also an important marker of inflammation associated with hypertension (Vyssoulis et al., 2007). Moreover, it also emerged as a prognostic marker for cardiovascular (CV) events in nondiabetic hypertensive participants followed up for 42.5 months (Schrader et al., 2006), and also significantly related (p < 0.05) to CV risk scores (Framingham and Heart Score) both at baseline and at the end of follow-up (Liakos et al., 2016). It is also found that the higher urine UTI levels may represent a useful marker of chronic inflammatory condition in type 1 and 2 diabetes patients (Lepedda et al., 2013). And therefore, based on these theories, it is more easy for us to understand the interactions between AMBP gene and diabetes or hypertension in ATS risk.

A few limitations in the present study need to be noted. First, we had a small sample size, and we were likely underpowered for other possible important results. Second, it was found that the mean age was younger in the control group such that they had not yet developed the chance to have a stroke despite being at higher risk. Few studies on the association between the AMBP gene and risk of ATS have been performed, especially in a Han Chinese population. Genetic studies across ethnically different populations should be performed for clarifying the role of variants and their interaction with modifiable risk factors. Such studies could add important information to the efforts of precision medicine for IS subtypes in diverse populations. In addition, the future research should focus on the functional role of the AMBP gene SNPs, which is helpful for elucidating the etiology and pathogenesis of IS.

Conclusion

Our findings suggested a possible association of the AMBP SNP haplotypes and gene-environment interactions with ATS risk in a Han Chinese population. The knowledge on the genetic risk effect on the ATS might provide crucial information on the identification of molecular mechanisms related to IS. Studies with larger samples should be conducted among different ethnic populations in the future.

Key message

(1)

AMBP gene GCCCCCCCC haplotype was associated with an elevated risk of ATS.

(2)

The contribution of interactions was also found in ATS risk between AMBP SNPs and hypertension or diabetes.

Footnotes

Acknowledgments

This work was supported by the grants from Department of Science and Technology of Sichuan province (2015JY0122). The authors are very grateful to Dr. Joshua Z. Willey for his revision of the article.

Author Disclosure Statement

No competing financial interests exist.

References

Åkerström

, Gram

(2014) A1M, an extravascular tissue cleaning and housekeeping protein. Free Radic Biol Med, 74:274-282.

Alberti

, Zimmet

(1998) Definition, diagnosis and classification of diabetes mellitus and its complications. Part 1: diagnosis and classification of diabetes mellitus provisional report of a WHO consultation. Diabet Med, 15:539-553.

Asplund

, Tuomilehto

, Stegmayr

, et al. (1988) Diagnostic criteria and quality control of the registration of stroke events in the MONICA project. Acta Med Scand Suppl, 728:26-39.

Benjamini

, Hochberg

(1995) Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Statist Soc B, 57:289-300.

Cederlund

, Deronic

, Pallon

, et al. (2015) A1M/α1-microglobulin is proteolytically activated by myeloperoxidase, binds its heme group and inhibits low density lipoprotein oxidation. Front Physiol, 6:11.

Chen

, Lu

, Department of Disease Control Ministry of Health

China (2004) The guidelines for prevention and control of overweight and obesity in Chinese adults. Biomed Environ Sci, 17:1-36.

Chen

, Gao

, Liu

, et al. (2017) Outline of the report on cardiovascular disease in China, 2016 (in Chinese). Chin Circ J, 32:521-530.

Daugherty

, Dunn

, Rateri

, et al. (1994) Myeloperoxidase, a catalyst for lipoprotein oxidation, is expressed in human atherosclerotic lesions. J Clin Invest, 94:437-444.

Dichgans

(2007) Genetics of ischaemic stroke. Lancet Neurol, 6:149-161.

10.

Fiseha

, Tamir

(2016) Urinary markers of tubular injury in early diabetic nephropathy. Int J Nephrol, 2016:4647685.

11.

Hahn

, Ritchie

, Moore

(2003) Multifactor dimensionality reduction software for detecting gene-gene and gene-environment interactions. Bioinformatics, 19:376-382.

12.

Han

, Kim

, Lee

, et al. (2007) A new subtype classification of ischemic stroke based on treatment and etiologic mechanism. Eur Neurol, 57:96-102.

13.

Hay

, Abajobir

, Abate

, et al. (2017) Global, regional, and national disability-adjusted life-years (DALYs) for 333 diseases and injuries and healthy life expectancy (HALE) for 195 countries and territories, 1990-2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet, 390:1260-1344.

14.

Humphries

, Morgan

(2004) Genetic risk factors for stroke and carotid atherosclerosis: insights into pathophysiology from candidate gene approaches. Lancet Neurol, 3:227-235.

15.

Huo

, Jiang

, Chen

, et al. (2017) Difference of hospital charges for stroke inpatients between hospitals with different levels and therapeutic modes in Beijing, China. Int J Neurosci, 127:752-761.

16.

International HapMap Consortium (2005) A haplotype map of the human genome. Nature, 437:1299-1320.

17.

Janunger

, Nilsson-Ardnor

, Wiklund

, et al. (2009) A novel stroke locus identified in a northern Sweden pedigree: linkage to chromosome 9q31-33. Neurology, 73:1767-1773.

18.

Joint Committee for Developing Chinese Guidelines on Prevention and Treatment of Dyslipidemia in Adults (2007) [Chinese guidelines on prevention and treatment of dyslipidemia in adults]. Zhonghua Xin Xue Guan Bing Za Zhi, 35:390-419.

19.

Kanayama

, Yamada

, Onogi

, et al. (2007) Bikunin suppresses expression of pro-inflammatory cytokines induced by lipopolysaccharide in neutrophils. J Endotoxin Res, 13:369-376.

20.

Kelly

, Gu

, Chen

, et al. (2008) Cigarette smoking and risk of stroke in the Chinese adult population. Stroke, 39:1688-1693.

21.

Kobayashi

(2006) Endogenous anti-inflammatory substances, inter-a-inhibitor and bikunin. Biol Chem, 387:1545-1549.

22.

Lepedda

, Nieddu

, Rocchiccioli

, et al. (2013) Development of a method for urine bikunin/urinary trypsin inhibitor (UTI) quantitation and structural characterization: application to type 1 and type 2 diabetes. Electrophoresis, 34:3227-3233.

23.

Liakos

, Vyssoulis

, Markou

, et al. (2016) Twenty-four-hour urine α1-Microglobulin as a marker of hypertension-induced renal impairment and its response on different blood pressure-lowering drugs. J Clin Hypertens (Greenwich), 18:1000-1006.

24.

Liu

, Huang

, Yang

, et al. (2013) MRAS genetic variation is associated with atherothrombotic stroke in the Han Chinese population. J Clin Neurol, 9:223-230.

25.

Liu

, Wu

, Wang

, et al. (2007) Stroke in China: epidemiology, prevention, and management strategies. Lancet Neurol, 6:456-464.

26.

Olsson

, Rosenlöf

, Kotarsky

, et al. (2013) The radical-binding lipocalin A1M binds to a complex I subunit and protects mitochondrial structure and function. Antioxid Redox Signal, 18:2017-2028.

27.

Pezzini

(2012) Genetic determinants of juvenile stroke. Thromb Res, 129:330-335.

28.

Ritchie

, Hahn

, Roodi

, et al. (2001) Multifactor-dimensionality reduction reveals high-order interactions among estrogen-metabolism genes in sporadic breast cancer. Am J Hum Genet, 69:138-147.

29.

Saif

, Soliman

(2017) Urinary α1 -microglobulin and albumin excretion in children and adolescents with type 1 diabetes. J Diabetes, 9:61-64.

30.

Salier

, Simon

, Rouet

, et al. (1992) Homologous chromosomal locations of the four genes for inter-alpha-inhibitor and pre-alpha-inhibitor family in human and mouse: assignment of the ancestral gene for the lipocalin superfamily. Genomics, 14:83-88.

31.

Schrader

, Lüders

, Kulschewski

, et al. (2006) Microalbuminuria and tubular proteinuria as risk predictors of cardiovascular morbidity and mortality in essential hypertension: final results of a prospective long-term study (MARPLE Study). J Hypertens, 24:541-548.

32.

Shi

, He

(2005) SHEsis, a powerful software platform for analyses of linkage disequilibrium, haplotype construction, and genetic association at polymorphism loci. Cell Res, 15:97-98.

33.

Vetr

, Gebhard

(1990) Structure of the human alpha 1-microglobulin-bikunin gene. Biol Chem Hoppe Seyler, 371:1185-1196.

34.

Vyssoulis

, Tousoulis

, Antoniades

, et al. (2007) Alpha 1-microglobulin as a new inflammatory marker in newly diagnosed patients. Am J Hypertens, 20:1016-1021.

35.

Wakahara

, Kobayashi

, Yagyu

, et al. (2005) Bikunin suppresses lipopolysaccharide-induced lethality through down-regulation of tumor necrosis factor- alpha and interleukin-1 beta in macrophages. J Infect Dis, 191:930-938.

36.

Wang

, Naghavi

, Allen

, et al. (2016) Global, regional, and national life expectancy, all-cause mortality, and cause-specific mortality for 249 causes of death, 1980-2015: a systematic analysis for the Global Burden of Disease Study 2015. Lancet, 388:1459-1544.

37.

Wang

, Jiang

, Sun

, et al. (2017) Prevalence, incidence, and mortality of stroke in china: results from a nationwide population-based survey of 480 687 adults. Circulation, 135:759-771.

38.

World Health Organization-International Society of Hypertension Guidelines for the Management of Hypertension, Guidelines Subcommittee (1999). J Hypertens, 17:151-183.

39.

Zhou

, Wang

, Zhu

, et al. (2016) Cause-specific mortality for 240 causes in China during 1990-2013: a systematic subnational analysis for the Global Burden of Disease Study 2013. Lancet, 387:251-272.