Polymorphisms in the transforming growth factor-β1 gene and the risk of asthma: A meta-analysis

ABSTRACT

Background and objective:  Polymorphisms in the transforming growth factor-β1 (TGF-β1) gene have been implicated in susceptibility to asthma, but a large number of studies have reported inconclusive results. A meta-analysis was performed to investigate the association between polymorphisms in the TGF-β1 gene and asthma susceptibility.

Methods:  Searches were performed of Medline (Ovid), PubMed, the Chinese Biological Medicine Database (CBM), the Chinese Journals Full-text Database (CNKI), the Cochrane Library Database and the Web of Science, covering all papers published up to 30 April 2009. Statistical analysis was performed using Revman4.2.8 and STATA10.0 software.

Results:  Two polymorphisms (−509C/T and 915G/C(G25C)) were investigated in 14 studies, involving 2979 asthma patients and 4941 control subjects. The results showed that individuals carrying the −509T allele (TT+TC) had a 36% increased risk of asthma, when compared with homozygotes (−509CC) (OR 1.36, 95% CI: 1.12–1.65). However, there was no significant association with risk of asthma in carriers of the 915C allele (GC+CC) compared with 915GG homozygotes (OR 1.05, 95% CI: 0.65–1.70). In a subgroup analysis by ethnicity, the risk of asthma associated with the −509T allele was significantly elevated among Asians (OR 1.50, 95% CI: 1.04–2.17) but not Caucasians (OR 1.16, 95% CI: 1.00–1.36). In a subgroup analysis by age, the −509T allele was associated with a significantly elevated risk of asthma among adults (OR 1.45, 95% CI: 1.09–1.92) but not children (OR 1.19, 95% CI: 0.96–1.46).

Conclusions:  This meta-analysis suggested that the −509C/T polymorphism in the TGF-β1 gene may be a risk factor for asthma. To further evaluate gene–gene and gene–environment interactions between polymorphisms in the TGF-β1 gene and asthma susceptibility, more studies involving thousands of patients are required.

INTRODUCTION

Asthma is a chronic inflammatory disease that can affect both children and adults with airway hypersensitivity and atopy. It is characterized by periodic attacks of wheezing, shortness of breath and tightness of the chest. Acute asthma can be life-threatening and fatal.¹ Previous studies have indicated that asthma is a complex polygenetic disorder, resulting from a combination of genetic and environmental factors, including exposure to tobacco smoke, air pollution, allergens and infections, as well as genetic variation.^2–5 Numerous published studies have focused on the association between host genetic variants and asthma susceptibility, and the transforming growth factor-β1 (TGF-β1) gene has been extensively studied.

Transforming growth factor-β1 is an important fibrogenic and immunomodulatory factor that plays an important role in promoting the structural changes of airway remodelling. Increased expression of TGF-β1 mRNA was observed in bronchial biopsy specimens of asthmatic patients compared with normal subjects.⁶

The TGF-β1 gene is located on chromosome 19q13.1-13.2,⁷ a region at which several single nucleotide polymorphisms have been identified, including −509C/T (rs1800469), 915G/C(G25C) (rs1800471) and 869T/C(T10C) (rs1982073). These polymorphisms have been investigated as potential susceptibility factors for asthma. A large number of studies have reported associations between polymorphisms in the TGF-β1 gene and the risk of asthma, but the results have been inconclusive. Because a single study may lack the power to provide comprehensive and reliable conclusions, a meta-analysis of all eligible studies was performed, to investigate the association between TGF-β1 polymorphisms and asthma susceptibility. This is, to our knowledge, the first meta-analysis of genetic studies on the association between asthma susceptibility and TGF-β1 polymorphisms.

METHODS

Selection of studies

Two independent reviewers (Y.Z. and J.Z.) searched Medline (Ovid), PubMed, the Chinese Biological Medicine Database (CBM), the Chinese Journals Full-text Database (CNKI), the Cochrane Library Database and the Web of Science to identify studies that had investigated the association between asthma susceptibility and TGF-β1 polymorphisms, with the last updated search being performed on 30 April 2009. The keywords used were as follows: asthma, asthmatic, asthma genetics, and transforming growth factor beta, and polymorphism or variant or genotype or mutation. There was no language restriction. All retrieved titles and abstracts were scrutinized for relevance to the association between asthma and TGF-β1 polymorphisms. References cited in the retrieved articles were also assessed. Articles reporting on the association between asthma susceptibility and TGF-β1 polymorphisms were identified. Studies had to meet the following inclusion criteria: (i) a case–control study design, based on unrelated individuals; (ii) the outcome had to be asthma; and (iii) genotype distributions were available for both cases and controls. The following exclusion criteria were also applied: (i) study design based on family or sibling pairs; (ii) genotype frequency not reported; (iii) abstracts, reviews and repeat studies; and (iv) genotype distribution in the control population did not accord with Hardy–Weinberg equilibrium.

Extraction of data

Two independent reviewers (J.Z. and Y.Z.) collected the data and reached a consensus on all studies. In cases of disagreement, a third author (J.H.) also assessed those articles. A standardized data form was used and included: first author, year of publication, country of origin, ethnicity of the study population, sample size, genotyping methods, definition of asthma, total number of cases and controls, and genotype distribution in cases and controls.

Statistical methods

Any polymorphism that had been investigated in at least five studies was included in the meta-analysis. Two polymorphisms (−509C/T and 915G/C(G25C)) were finally identified. The genetic models evaluated for these two polymorphisms were dominant models (TT+TC vs CC for −509C/T, and CC+CG vs GG for 915G/C(G25C)), which were based on those mostly used in the primary studies. OR and the corresponding 95% CI were calculated in order to assess the strength of the association between TGF-β1 polymorphisms and asthma susceptibility. Heterogeneity was assessed by the χ²-based Q statistic test. A P-value > 0.10 for the Q test indicates a lack of heterogeneity among studies. The fixed-effect model was used to estimate the OR if a lack of heterogeneity was suggested by this test (P > 0.10); otherwise, a random-effect model was used, as this was more appropriate when there was significant heterogeneity. To evaluate ethnicity-specific and age-specific effects, subgroup analyses were performed for the −509C/T polymorphism, which was investigated in a sufficient number of studies. For both polymorphisms, other genetic models (−509C/T: TT vs TC+CC, T vs C, TT vs CC, TC vs CC; G915C(G25C): CC vs GC+GG, C vs G, CC vs GG, CG vs GG) were also used to assess the association with the risk of asthma.

Asymmetry of funnel plots was visually inspected to assess potential publication bias. Begg's test⁸ and Egger's test⁹ were also performed to assess publication bias statistically.

Hardy–Weinberg equilibrium was tested using a web-based programme (http://ihg2.helmholtz-muenchen.de/cgi-bin/hw/hwa1.pl). All statistical analyses were performed using Revman4.2.8 (The Cochrane Collaboration, http://www.cochrane.org) and STATA10.0 (http://www.stata.com) software.

RESULTS

Studies included in the meta-analysis

Sixty-one papers relevant to the search criteria were identified, and after screening titles and abstracts, 29 publications were selected for full-text review. An additional 15 articles were excluded. Five did not report case–control studies,^10–14 three were not asthma studies,^15–17 three were family-based studies,^18–20 two did not report useable data,^21,22 one was a repeat publication²³ and one reported on TGF-β2 polymorphisms rather than TGF-β1 polymorphism.²⁴ Therefore, 14 eligible articles reporting on TGF-β1 polymorphisms,^7,25–37 and meeting the inclusion criteria, were identified. One of these studies investigated asthma patients and control subjects in both North India and West India and the data were analysed separately for each group;⁷ these data were treated as two independent case–control studies. Another study investigated both Hispanic whites and non-Hispanic whites,³⁴ with separate analyses of the data, and these were also treated as two independent studies.

Only two polymorphisms (−509C/T and 915G/C(G25C)) were included in the meta-analyses, other polymorphisms such as 869T/C(T10C) and −800G/A being excluded because of insufficient numbers of studies on each polymorphism. In the present study, at least five single case–control studies were required for one meta-analysis on each polymorphism. All 16 case–control studies, comparing 2979 asthma patients and 4941 control subjects were used in the pooled analyses. There were 13 studies on −509C/T and five on 915G/C(G25C). Eight studies investigated Asian subjects, seven were performed on Caucasian subjects and one investigated a mixed population of Caucasians, mulattos and Negroids. Eleven studies were performed in adult populations, while five investigated children. The characteristics of these studies are shown in Table 1. Genotype and allele distributions for each polymorphism are shown in Table 2.

Table 1.  Details of the 16 case–control studies included in the meta-analysis

First author	Year of publication	Country	Ethnicity of population	Age of population (years)	Number of cases (n)	Number of controls (n)	Genotyping method
ARMS, amplification refractory mutation system; AS-PCR, allele specific PCR; H, Hispanic whites; N, North India; NH, non-Hispanic whites; PCR-SSP, PCR with sequence specific primers; RFLP, restriction fragment length polymorphism; W, West India.
Kumar A^28	2008	India	Asian	—	123	100	ARMS
Nagpal K (N)^7	2005	India	Asian	28.42 ± 0.97	187	187	RFLP
Nagpal K (W)^7	2005	India	Asian	34.22 ± 1.01	209	190	RFLP
Mak JCW^31	2006	China	Asian	41.0 ± 16.1	250	308	RFLP
Movahedi M^32	2008	Iran	Asian	—	60	140	PCR-SSP
Xia W^37	2006	China	Asian	15–60	60	30	RFLP
Lu JR^30	2004	China	Asian	1–13	98	52	RFLP
Jo MY^27	2008	Korean	Asian	17–74	108	162	RFLP
Silverman ES^35	2004	USA	Caucasian	33.7 ± 13.7	527	170	RFLP
Salam MT (NH)^34	2007	USA	Caucasian	—	295	1801	TaqMan
Salam MT (H)^34	2007	USA	Caucasian	—	113	814	TaqMan
Pulleyn LJ^33	2001	UK	Caucasian	44.8 and 27.9	213	202	RFLP
Wisniewski A^36	2009	Poland	Caucasian	39.9 ± 16.2	247	287	RFLP
Liebhart J^29	2008	Poland	Caucasian	51.0 ± 12.5	110	109	AS-PCR
Heinzmann A^26	2005	Germany	Caucasian	5–18	231	269	RFLP
de Faria ICJ^25	2008	Brazil	Mixed	10.3 ± 2.79	88	202	RFLP

Table 2.  Distribution of TGF-β1 genotypes among asthma patients and control subjects (n)

Study	Asthma			Control			Asthma		Control		Hardy–Weinberg equilibrium
H, Hispanic whites; N, North India; NH, non-Hispanic whites; W, West India.
−509C/T polymorphism	CC	CT	TT	CC	CT	TT	C	T	C	T
de Faria ICJ et al.^25	32	42	88	58	112	32	106	218	228	176	Yes
Heinzmann A et al.^26	94	108	29	121	111	37	296	166	343	185	Yes
Jo MY et al.^27	27	45	36	35	92	35	99	117	162	162	Yes
Kumar A et al.^28	44	55	24	43	42	15	143	103	128	72	Yes
Lu JR et al.^30	45	38	15	30	19	3	128	68	79	25	Yes
Mak JCW et al.^31	46	109	93	51	155	102	201	295	257	359	Yes
Nagpal K et al.^7 (N)	64	104	19	94	77	16	232	142	265	109	Yes
Nagpal K et al.^7 (W)	73	97	39	112	63	15	243	175	287	93	Yes
Salam MT et al.^34 (H)	25	60	28	215	411	188	110	116	841	787	Yes
Salam MT et al.^34 (NH)	137	121	37	813	799	189	395	195	2425	1177	Yes
Silverman ES et al.^35	212	246	69	83	76	11	670	384	242	98	Yes
Wisniewski A et al.^36	97	118	32	131	124	32	312	182	386	188	Yes
Xia W et al.^37	22	26	12	17	11	2	70	50	45	15	Yes
915G/C(G25C) polymorphism	GG	GC	CC	GG	GC	CC	G	C	G	C
Movahedi M et al.^32	32	16	0	119	17	2	80	16	255	21	Yes
Liebhart J et al.^29	95	14	1	91	18	0	204	16	200	18	Yes
Pulleyn LJ et al.^33	181	32	0	95	26	1	394	32	216	28	Yes
Heinzmann A et al.^26	192	37	2	230	37	2	421	41	497	41	Yes
Wisniewski A et al.^36	214	31	2	161	28	2	459	35	350	32	Yes

Quantitative synthesis of data

Transforming growth factor-β1 −509C/T polymorphism

Thirteen case–control studies on the relationship between the −509C/T polymorphism and the risk of asthma were included in the meta-analysis. As shown in Figure 1, the heterogeneity of TT+TC versus CC was analysed for all 13 studies and χ² was 30.74 with 12 degrees of freedom and P = 0.002 in a random-effect model. In addition, the I-square value, another index of heterogeneity, was 61.0%, suggesting moderate heterogeneity (Fig. 1). Therefore, the random-effect model was used for synthesis of the data. The overall OR for TT+TC genotypes versus CC genotypes was 1.36 (95% CI: 1.12–1.65) and the Z test for overall effect was 3.14 (P = 0.002). These results suggest that individuals who carry the T allele may have a 36% increased asthma risk compared with CC homozygotes. The funnel plots were almost symmetrical, and neither Egger's test nor Begg's test (Fig. 2) indicated any significant publication bias (all P > 0.05).

Figure 1. Meta-analysis of studies on the association between asthma risk and the −509C/T polymorphism in the transforming growth factor-β1 (TGF-β1) gene, using a random-effect model (TT+TC vs CC).

Figure 2. Begg's funnel plot for publication bias in selection of studies on the transforming growth factor-β1 (TGF-β1) −509C/T polymorphism.

Subgroup analyses were performed after stratification of the data by ethnicity and age. In the subgroup analysis by ethnicity, increased risks were identified among Asian (OR 1.50, 95% CI: 1.04–2.17; P = 0.003 for heterogeneity) and Caucasian subjects (OR 1.16, 95% CI: 1.00–1.36; P = 0.36 for heterogeneity) in a dominant model (TT+TC vs CC) (Fig. 3). Thus, Asian carriers of the T allele may have an increased risk of asthma. In the subgroup analysis by age, significantly increased risks were identified among adults (OR 1.45, 95% CI: 1.09–1.92 for a dominant model; P = 0.004 for heterogeneity) (Fig. 4). However, among children, there was no significant association with asthma risk in the dominant model (OR 1.19, 95% CI: 0.96–1.46, P = 0.25 for heterogeneity).

Figure 3. Subgroup analysis by ethnicity in the meta-analysis of studies on the association between asthma risk and the transforming growth factor-β1 (TGF-β1) −509C/T polymorphism, using a random-effect model (TT+TC vs CC).

Figure 4. Subgroup analysis by age in the meta-analysis of studies on the association between asthma risk and the transforming growth factor-β1 (TGF-β1) −509C/T polymorphism, using a random-effect model (TT+TC vs CC).

Transforming growth factor-β1 915G/C(G25C) polymorphism

Five case–control studies on the relationship between the 915G/C(G25C) polymorphism and the risk of asthma were included in the meta-analysis, and these studies included 849 cases and 829 controls.

As shown in Figure 5, the heterogeneity of CC+CG versus GG was analysed for all five studies and the value of χ² was 12.46 with 4 degrees of freedom and P = 0.01 in a random-effect model. The I-square was 67.9%, suggesting moderate heterogeneity. Therefore, the random-effect model was used for synthesis of the data. The overall OR for CC+CG genotypes versus GG genotypes was 1.05 (95% CI: 0.65–1.70) and the Z test for overall effect was 0.22 (P = 0.83). The funnel plots were symmetrical, and neither Egger's test nor Begg's test (Fig. 6) indicated any significant publication bias (all P > 0.05). A summary of the results from other comparative genetic models is presented in Table 3.

Figure 5. Meta-analysis of studies on the association between asthma risk and the transforming growth factor-β1 (TGF-β1) 915G/C(G25C) polymorphism, using a random-effect model (CC+CG vs CC).

Figure 6. Begg's funnel plot for publication bias in selection of studies on the transforming growth factor-β1 (TGF-β1) 915G/C(G25C) polymorphism.

Table 3.  Summary of results from different comparative genetic models

Polymorphism	Genetic model	Participants (n)	OR (95% CI)	Z	P-value	I-square(%)	Phet	Effect model
F, fixed-effect model; I-square, index of heterogeneity; Phet, P-value for heterogeneity; R, random-effect model; Z, test for overall effect.
−509C/T	TT+CT vs CC	7 180	1.36 (1.12, 1.65)	3.14	0.002	61.0	0.002	R
	TT vs CC+CT	7 180	1.68 (1.22, 2.32)	3.16	0.002	75.8	<0.00001	R
	TC vs CC	5 982	1.20 (0.97, 1.48)	1.71	0.09	62.1	0.002	R
	TT vs CC	3 919	1.77 (1.29, 2.41)	3.59	0.0003	66.3	0.0004	R
	T vs C	14 360	1.39 (1.17, 1.65)	3.74	0.0002	76.5	<0.00001	R
915G/C(G25C)	CC+GC vs GG	1 678	1.05 (0.65, 1.70)	0.22	0.83	67.9	0.01	R
	CC vs GG+GC	1 678	0.83 (0.29, 2.36)	0.35	0.73	0.0	0.81	F
	CG vs GG	1 666	0.77 (0.62, 0.96)	2.36	0.02	33.4	0.20	F
	CC vs GG	1 422	0.85 (0.30, 2.43)	0.30	0.76	0.0	0.81	F
	C vs G	3 356	1.03 (0.68, 1.54)	0.13	0.90	61.3	0.03	R

DISCUSSION

Recently, variants of the TGF-β1 gene have been suggested to be associated with the risk of asthma, but a number of studies have reported inconclusive results. Therefore, a meta-analysis was performed to assess these associations. In a meta-analysis, a small number of studies weaken the conclusions; therefore, only those polymorphisms that had been investigated in at least five case–control studies were included. Because age and ethnicity may influence the outcome of the results, stratified analyses were performed for the −509C/T polymorphism, for which there was a sufficiently large number of studies.

Two polymorphisms, −509C/T and 915G/C(G25C), which met the inclusion criteria for the meta-analyses, were finally identified. The results revealed that the −509C/T polymorphism was significantly associated with the risk of asthma, with carriers of the T allele having a nearly 36% increased risk. The risk appeared to be more evident in Asian rather than Caucasian subjects, suggesting a possible influence of different genetic backgrounds and environmental exposures. In the age-stratified analysis the −509C/T polymorphism was associated with a 45% increased asthma risk in adults, but not in children. This may reflect the fact that increased age modifies the associations. As for the 915G/C(G25C) polymorphism, there was no evidence of any association between this polymorphism and asthma susceptibility. One possible explanation is that the 915G/C(G25C) polymorphism may not influence the functional region of TGF-β1 gene that has an effect on the pathogenesis of asthma. This polymorphism may play only a minor role in the causation of asthma.

Meta-analysis is a useful method for investigating associations of asthma with genetic factors because it uses a quantitative approach by way of combining the results of different studies on the same topic, potentially providing more reliable conclusions. Several meta-analyses have been performed on the association of candidate variants with asthma susceptibility, including the −308G/C polymorphism in the tumour necrosis factor-α gene,³⁸ and the −589C/T polymorphism in the IL-4 gene.³⁹ In agreement with previous studies, the present meta-analysis revealed that the −509C/T polymorphism in the TGF-β1 gene may increase the risk of asthma. These results indicate that variants in genes involved in asthma pathogenesis significantly increase the risk of asthma, suggesting that these variants may be predictors of asthma occurrence or potential targets for the treatment of asthma.

There were several limitations that should be considered when interpreting the present results. First, only published studies that were included in Medline (Ovid), PubMed, CBM, CNKI, the Cochrane Library Database and Web of Science were identified; therefore, publication bias may have occurred. Second, populations under study were from different countries and control subjects were not uniform. Third, in the ethnicity-stratified analysis for the −509C/T polymorphism, the data were from Caucasian and Asian populations, and the results may therefore be applicable to these ethnic groups only. However, likelihood of bias was minimized by developing a detailed study protocol, meticulous search for published studies, and using explicit methods for study selection, data extraction and data analysis.

In conclusion, this meta-analysis indicated that the −509C/T polymorphism in the TGF-β1 gene is associated with an increased risk of asthma. In order to further clarify gene–gene and gene–environment interactions between polymorphisms in the TGF-β1 gene and asthma susceptibility, additional large case–control studies will be required.

ACKNOWLEDGEMENTS

This study was supported by grants #30470761 and #30871117 from National Natural Science Foundation of China.