DEFB1 gene polymorphisms and tuberculosis in a Northeastern Brazilian population

Abstract

β-Defensin-1, an antimicrobial peptide encoded by the DEFB1 gene, is known to play an important role in lung mucosal immunity. In our association study we analyzed three DEFB1 functional polymorphisms -52G>A (rs1799946), -44C>G (rs1800972) and -20G>A (rs11362) in 92 tuberculosis patients and 286 healthy controls, both from Northeast Brazil: no association was found between the studied DEFB1 polymorphisms and the disease. However we cannot exclude that this lack of association could be due to the low number of subjects analyzed, as suggested by the low statistical power achieved for the three analyzed SNPs (values between 0.16 and 0.50).

Keywords Tuberculosis; Innate immunity; DEFB1; Genotyping; Polymorphism

Introduction

Mycobacterial diseases are a major health concern, affecting approximately one-third of the world's population.¹ The most common clinical form of the disease is pulmonary tuberculosis, a granulomatous disease of the lungs caused by Mycobacterium tuberculosis. Extrapulmonary tuberculosis may occur when the bacillus infects other organs.^2–4

The general risk of infected individuals to develop the disease during their lifetime ranges between 5 and 10%.^2,3 A possible genetic predisposition to tuberculosis (TB) has been suggested in several studies.^5–8 Moreover evidence exist that extrapulmonary TB is the result of an underlying immune defect,⁴ however, the functional correlation between genetic polymorphisms and immunologic defect should be yet unraveled.⁷

The mucosal immunity of air epithelia is complex, and the innate immune system, mainly by the mucosal secretion of antimicrobial peptides gradient, is thought to play an important role in defense against M. tuberculosis, although the exact mechanisms are yet poorly understood.^9,10

The human β-defensin-1 (hBD-1) is a small (29–47 amino acids) cationic antimicrobial peptide,¹¹ constitutively expressed in airway epithelia that plays an important role in mucosal immunity in the lung.^12–15 This peptide is encoded by DEFB1 gene, located at chromosome 8p22-23.¹⁶

Three single nucleotide polymorphisms (SNPs), located at -52G>A (rs1799946), -44C>G (rs1800972) and -20G>A (rs11362) position in the 5′ untranslated region (UTR) of the DEFB1 gene have been reported to promote functional alteration of hBD-1 expression in different cell models.^17,18 These SNPs were firstly described by Dörk and Stuhrmann¹⁹ and have been then associated with chronic obstructive pulmonary disease,²⁰ increased levels of Candida albicans in the oral cavity,²¹ increased risk for Human Immunodeficiency Virus 1 infection,^22–25 susceptibility to mycobacterial diseases^8,26 and other diseases.^27–31

Considering the possible role of the innate immune system in the risk of tuberculosis infection, this study aims to examine whether functional polymorphisms at 5′UTR in DEFB1 gene are associated with the occurrence of active tuberculosis.

Patients and methods

Study population

Ninety-two TB patients were enrolled from five reference center for TB follow up in Recife (Pernambuco, Northeast Brazil): Hospital das Clínicas – Universidade Federal de Pernambuco (HC-UFPE), Instituto de Medicina Integral Professor Fernando Figueira (IMIP), Hospital Barão de Lucena (HBL/US) and Hospital Otávio de Freitas. Of the 92 patients analyzed, 64 (69.6%) presented with pulmonary TB (36 males and 28 females, mean age 21 years ± 18.7) and 28 (30.4%) with extrapulmonary TB (17 males and 11 females, mean age 23 years ± 21.7). The diagnosis was based on clinical symptoms and radiographic findings, along with bacteriological confirmation (culture, smear and/or polymerase chain reaction) as described by the Diagnostic Standards and Classification of Tuberculosis in Adults and Children.³²

As control group, we enrolled 286 healthy individuals (163 females/123 males; mean age 27 years ± 20.3), unrelated to patients, with negative Mantoux test, showing no symptoms of tuberculosis or previous history of the disease. All patients and control subjects came from the same geographical area (Pernambuco, Brazil), were all HIV-1 negative and were not under immunosuppressive medication.

Since ethnicity is an important issue in genetic analyses, patients and controls reported their self-declared skin based ethnic origin. During ethnicity report a health professional accompanied the procedure guaranteeing the correctness of the procedure. After ethnicity assessment we verified the self-reported and health professional checked results, by analyzing in the controls 12 Ancestry informative markers (AIMs) as recently reported by Coelho et al.³³ Absence or extremely poor correlation between skin based ethnicity and the true genetic background, as revealed by the 12 AIMs analysis, was observed. So, we discarded the self-declared skin based ethnic origin and considered only the genetic AIMs findings. In fact, our studied population consists of a strong admixture of European-Caucasian, African and South American Amerindians population, leading the genetic North East Brazilian composition unique; consequently it is impossible to stratify our population for ethnicity, being an admixture of around 58% European, 24% African and 18% Amerindian as recently reported.³³

Written informed consent was obtained from the participants or their parents. The CPqAM/FIOCRUZ Ethics Committee (CEP Registration – 55/05) approved the study. Patients underwent a standardized clinical–epidemiological questionnaire.

DEFB1 genotyping

Genomic DNA was extracted from whole blood using QIAamp DNA Blood Kit according to manufacturer instructions (QIAamp DNA Blood Midi Kit, Qiagen).

The DEFB1 5′UTR polymorphisms -52G>A (rs1799946), -44C>G (rs1800972) and -20G>A (rs11362) were genotyped by direct sequencing, according to Braida et al.²² Sequences were handled using the 4Peaks (http://nucleobytes.com/index.php/4peaks) and CodonCode Aligner software (http://www.codoncode.com/aligner/).

Statistical analysis

Chi-square test was used to verify the Hardy–Weinberg equilibrium and the Fisher's exact test was employed for pair-wise comparison of allele, genotype and haplotype frequencies using contingency tables as appropriate; only p values < 0.05 were considered to be significant. All the statistical analyses were carried out using the open-source R package (http://www.r-project.org). The power of the study was checked by software G*power software version 3.1.³⁴ Haplotypes were computed using the Arlequin version 3.5.1.3³⁵ and Haploview software.³⁶

Results

The genotyping results of the three 5′UTR DEFB1 functional polymorphisms -52G>A (rs1799946), -44C>G (rs1800972) and -20G>A (rs11362) in TB patients and healthy controls, are shown in Table 1. All SNPs were in Hardy–Weinberg equilibrium for both TB patients and healthy controls.

Table 1
Alleles and genotypes distribution of SNPs – in the 5' untranslated region (UTR) of DEFB1 gene in tuberculosis patients (TB) and healthy controls (HC) from Northeast Brazil.

	SNPs	TB patients n = 92		Healthy controls n = 268	p -value, OR (95% CI)
		All	pTB		All vs. HC	pTB vs. HC
-52G>A (rs1799946)
	G	97 (0.53)	68 (0.54)	309 (0.54)	Reference	Reference
	A	87 (0.47)	58 (0.46)	263 (0.46)	p = 0.80, 1.05 (0.74–1.49)	p = 1.00, 1.00 (0.67–1.50)
	GG	29 (0.32)	19 (0.30)	89 (0.31)	Reference	Reference
	GA	39 (0.42)	30 (0.48)	131 (0.46)	p = 0.78, 0.91 (0.51–1.65)	p = 0.87, 1.07 (0.54–2.15)
	AA	24 (0.26)	14 (0.22)	66 (0.23)	p = 0.75, 1.11 (0.56–2.19)	p = 1.00, 0.99 (0.43–2.26)
	GA/AA	63 (0.68)	44(0.70)	197 (0.69)	p = 1.00, 0.98 (0.58–1.69)	p = 1.00, 1.05 (0.56–2.01)
−44C>G (rs1800972)
	C	153 (0.83)	102 (0.81)	491 (0.86)	Reference	Reference
	G	31 (0.17)	24 (0.19)	81 (0.14)	p = 0.40, 1.23 (0.75–1.96)	p = 0.17, 143 (0.82–2.40)
	CC	66 (0.72)	42 (0.67)	213 (0.74)	Reference	Reference
	CG	21 (0.23)	18 (0.29)	65 (0.23)	p = 0.89, 1.04 (0.56–1.88)	p = 0.29, 1.40 (0.71–2.69)
	GG	5 (0.05)	3 (0.05)	8 (0.03)	p = 0.32, 2.01 (0.50–7.25)	p = 0.40, 1.90 (0.31–8.32)
	CG/GG	26 (0.28)	21 (0.34)	73 (0.26)	p = 0.59, 1.15 (0.65–1.99)	p = 0.21, 1.46 (0.77–2.71)
−20G>A (rs11362)
	G	118 (0.64)	82 (0.65)	361 (0.63)	Reference	Reference
	A	66 (0.36)	44 (0.35)	211 (0.37)	p = 0.86, 0.96 (0.66–1.37)	p = 0.76, 0.92 (0.60–1.40)
	GG	39 (0.42)	26 (0.41)	115 (0.40)	Reference	Reference
	GA	40 (0.43)	30 (0.48)	131 (0.46)	p = 0.70, 0.90 (0.52–1.54)	p = 1.00, 1.01 (0.54–1.90)
	AA	13 (0.14)	7 (0.11)	40 (0.14)	p = 1.00, 0.96 (0.42–2.06)	p = 0.66, 0.77 (0.26–2.02)
	GA/AA	53 (0.57)	37 (0.59)	171 (0.60)	p = 0.72, 0.91 (0.55–1.52)	p = 0.89, 0.96 (0.53–1.74)

• pTB, pulmonary tuberculosis; OR, odds ratio; 95% CI, 95% confidence interval; SNP, single nucleotide polymorphism. Reference indicates the most frequent alleles and genotypes.

DEFB1 SNPs allele and genotype frequencies were similar in TB patients and healthy controls and no significant difference has been found (p > 0.05), also when TB patients were stratified according to the pulmonary form of disease.

We computed the haplotypes and observed that the three SNPs in the 5′UTR of DEFB1 are in strong linkage disequilibrium (D ′ > 0.9), forming four haplotypes (Table 2). However, no significant differences were found when comparing TB patients and healthy controls (Table 2).

Table 2
Haplotypes frequency distribution of 5'UTR DEFB SNPs in tuberculosis patients (TB) and healthy controls (HC) from Northeast Brazil.

Haplotypes (−52/−44/−20)	TB patients		Healthy controls 2 n = 572	p -value, OR (95% CI)
	All 2 n = 184	pTB 2 n = 126		All vs. HC	pTB vs. HC
ACG	86 (0.47)	58 (0.46)	264 (0.46)	Reference	Reference
GCA	64 (0.35)	42 (0.34)	210 (0.37)	p = 0.78, 0.94 (0.63–1.38)	p = 0.74, 0.91 (0.57–1.44)
GGG	30 (0.16)	22 (0.18)	80 (0.14)	p = 0.61, 1.15 (0.68–1.91)	p = 0.47, 1.25 (0.68–2.23)
GCG	4 (0.01)	4 (0.02)	18 (0.03)	p = 0.61, 0.68 (0.16–2.15)	p = 1.00, 1.01 (0.24–3.23)

• pTB, pulmonary tuberculosis; OR, odds ratio; 95% CI, 95% confidence interval; nc, not calculable. Reference indicates the most frequent haplotypes.

In addition, we estimated the linkage disequilibrium and the possible haplotypes, using Haploview software, for Brazilian healthy controls (BRA) and healthy individuals of African (AFR), Caucasian (CEU) and Ad Mixed American (AMR) origin, provided by 1000 Genome Project Database (www.1000genomes.org). We noted that DEFB-1 SNPs were in perfect linkage disequilibrium (D′ = 1.0) and have different haplotype distributions in the studied populations (Table 3).

Table 3
Haplotypes frequencies in healthy individual of different ethnic origin.

Haplotypes	Frequencies
	Brazilian (BRA)	African (AFR)	Caucasian (CEU)	Ad Mixed American (AMR)
ACG	0.460	0.588	0.399	0.297
GCA	0.369	0.300	0.354	0.438
GGG	0.142	0.064	0.000	0.000
GCG	0.030	0.049	0.247	0.261

The post hoc analysis of the study through the G*power software, indicate the following power for three SNPs: 0.30 (-52G>A), 0.5 (-44C>G) and 0.16 (-20G>A).

Discussion

In the current study, no association was found between -52G>A (rs1799946), -44C>G (rs1800972), -20G>A (rs11362) DEFB1 polymorphisms and TB in a Brazilian population, also when considering the pulmonary form of the disease.

In a case–control study, Wu et al.⁸ analyzed DEFB1 5′UTR polymorphisms in 102 patients with pulmonary TB and 148 healthy controls of Chinese Han population observing that the DEFB1 -44C/C genotype was associated with pulmonary TB susceptibility (OR = 2.096) and the GGG (-52/-44/-20) haplotype was associated with the pulmonary TB protection (OR = 0.348).

In a recent work, performed on a limited number of Mexican patients (27 with pulmonary TB and 16 with extrapulmonary TB) and controls (n = 49), López Campos et al.³⁷ described an association between DEFB1 -44C>G SNP (CG genotype in a codominant genetic model) and susceptibility to extrapulmonary TB, but not with the pulmonary TB, hypothesizing a role for hBD-1 in the fight against M. tuberculosis outside the lung.

Our findings have discrepancies with the results obtained by Wu et al.⁸ on Han population and by Lopez Campos et al.³⁷ in Mexicans: we can hypothesize that the differences could be, at least in part, due to ethnicity, since the frequencies DEFB1 SNPs vary in different populations. According to the NCBI and HapMap databases,³⁸ for -52 SNP (rs1799946) the frequency of the ancestral allele is 0.64 in European (CEU), 0.56 in Sub-Sahara African (YRI) and 0.40 in Asian (HCB); the frequency of ancestral allele of -44 SNP (rs1800972) is 0.26 in European (CEU), 0.04 in Sub-Sahara African (YRI) and 0.12 in Asian (HCB); for -20 SNP (rs11362), the frequency is 0.61 in European (CEU), 0.60 in Sub-Sahara (YRI) and 0.71 in Asian (HCB). In addition, the SNPs analyzed in Brazilian healthy individuals and healthy individual of African (AFR), in Caucasian (CEU) and Ad Mixed American (AMR) origin, were in perfect linkage disequilibrium (D′ = 1.0) and presented different haplotypic distribution (Table 3), suggesting that ethnics background may be responsible for the differences observed in populations that have been studied. Additionally, in the Han Chinese population no linkage disequilibrium was detected (D′ < 0.5)⁸ and as a consequence comparing DEFB1 haplotype results with our findings has not been possible. Moreover, when looking at the Mexican population it has to be remarked that the study of Lopez Campos et al.³⁷ just analyzed one DEFB1 SNP -44C>G and other SNPs in cathelicidin encoding CAMP gene. So, it was not possible to compare our DEFB1 haplotypes with the results reported for Mexican patients. Furthermore, DEFB1 rs1800972 allele and genotype frequencies reported in the Mexican population (not in Hardy–Weimberg Equilibrium) were significantly different (0.0002) for the C allele to Brazilian ones.

However, we cannot exclude, that the lack of association reported in our study, could be due to the limited sample size; the power of our study (post hoc Power ≤ 0.5) indicates that not finding an association between SNPs with TB development, could be due to a type II error. This is a strong limitation of our study, and seeing that also the works of Wu et al.⁸ and Lopez Campos et al.³⁷ suffer the same sample size problem, we are convinced that any further study dealing with association between defensins gene polymorphism and TB, should enroll greater number of patients, may be joining several research groups, in order to achieve better statistical power to provide a definitive answer on the role of defensins in susceptibility to TB.

All this considered, even if our results seem to exclude an association between DEFB1 polymorphisms and TB development in Northeast Brazilian population, since DEFB1 could be a potential candidate gene for susceptibility to TB and only very few studies have addressed this topic achieving controversial results, we do think that further studies on a larger number of patients and in other populations are needed to disclose the role of DEFB1 variations in tuberculosis development.

• Associate Editor: Jorge Luiz Mello Sampaio

Acknowledgments

The authors thank all the health professionals from the hospitals involved in the study that collaborated with the research. We thank the Laboratory of Immunoepidemiology – Centro de Pesquisas Aggeu Magalhões/Fundação Oswaldo Cruz (CPqAM/FIOCRUZ), the Laboratory of Immunopathology Keizo Asami, the Department of Genetics, Federal University of Pernambuco, the Graduate Program in Genetics for supporting physical and scientific, as well as Fundação de Amparo à Ciência e Tecnologia do Estado de Pernambuco (FACEPE), Coordenação de Aperfeiçoamento Pessoal de Nível Superior (CAPES) and Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), for financial support.

References

• ¹ WHO. Global Tuberculosis Report; 2012. http://www.who.int/tb/publications/global_report/en/ Accessed 30.07.13.
  » http://www.who.int/tb/publications/global_report/en/
• ² Ducati RG, Ruffino-Netto A, Basso LA, Santos DS. The resumption of consumption – a review on tuberculosis. Mem Inst Oswaldo Cruz 2006;101:697-714.
• ³ Clark-Curtiss JE, Haydel SE. Molecular genetics of Mycobacterium tuberculosis pathogenesis. Annu Rev Microbiol 2003;57:517-549.
• ⁴ Gonzalez OY, Adams G, Teeter LD, Bui TT, Musser JM, Graviss EA. Extra-pulmonary manifestations in a large metropolitan area with a low incidence of tuberculosis. Int J Tuberc Lung Dis 2003;7:1178-1185.
• ⁵ Bellamy R, Ruwende C, Corrah T, McAdam KP, Whittle HC, Hill AV. Variations in the NRAMP1 gene and susceptibility to tuberculosis in West Africans. N Engl J Med 1998;338:640-644.
• ⁶ Bellamy R, Ruwende C, Corrah T, et al. Tuberculosis and chronic hepatitis B virus infection in Africans and variation in the vitamin D receptor gene. J Infect Dis 1999;179:721-724.
• ⁷ Berrington WR, Hawn TR. Mycobacterium tuberculosis, macrophages, and the innate immune response: does common variation matter?. Immunol Rev 2007;219:167-186.
• ⁸ Wu X, Gong L, Lin J, Wang H. Association between human beta defensin-1 single nucleotide polymorphisms and susceptibility to pulmonary tuberculosis. Zhonghua Yu Fang Yi Xue Za Zhi 2012;46:912-915.
• ⁹ van Crevel R, Ottenhoff THM, van der Meer JWM. Innate immunity to Mycobacterium tuberculosis Adv Exp Med Biol 2003;531:241-247.
• ¹⁰ Rivas-Santiago B, Sada E, Tsutsumi V, Aguilar-Leon D, Contreras JL, Hernandez-Pando R. beta-Defensin gene expression during the course of experimental tuberculosis infection. J Infect Dis 2006;194:697-701.
• ¹¹ Pazgier M, Hoover DM, Yang D, Lu W, Lubkowski J. Human beta-defensins. Cell Mol Life Sci 2006;63:1294-1313.
• ¹² McCray PB, Bentley L. Human airway epithelia express a beta-defensin. Am J Respir Cell Mol Biol 1997;16:343-349.
• ¹³ Bals R. Epithelial antimicrobial peptides in host defense against infection. Respir Res 2000;1:141-150.
• ¹⁴ Klotman ME, Chang TL. Defensins in innate antiviral immunity. Nat Rev Immunol 2006;6:447-456.
• ¹⁵ Wallace AM, He J-Q, Burkett KM, et al. Contribution of alpha- and beta-defensins to lung function decline and infection in smokers: an association study. Respir Res 2006;7:76
• ¹⁶ Linzmeier RM, Ganz T. Human defensin gene copy number polymorphisms: comprehensive analysis of independent variation in alpha- and beta-defensin regions at 8p22-p23. Genomics 2005;86:423-430.
• ¹⁷ Milanese M, Segat L, Crovella S. Transcriptional effect ofDEFB1 gene 5′ untranslated region polymorphisms. Cancer Res 2007;67:5997, author reply 5997.
• ¹⁸ Sun CQ, Arnold R, Fernandez-Golarz C, et al. Human beta-defensin-1, a potential chromosome 8p tumor suppressor: control of transcription and induction of apoptosis in renal cell carcinoma. Cancer Res 2006;66:8542-8549.
• ¹⁹ Dörk T, Stuhrmann M. Polymorphisms of the human beta-defensin-1 gene. Mol Cell Probes 1998;12:171-173.
• ²⁰ Matsushita I, Hasegawa K, Nakata K, Yasuda K, Tokunaga K, Keicho N. Genetic variants of human beta-defensin-1 and chronic obstructive pulmonary disease. Biochem Biophys Res Commun 2002;291:17-22.
• ²¹ Jurevic RJ, Bai M, Chadwick RB, White TC, Dale BA. Single-nucleotide polymorphisms (SNPs) in human beta-defensin 1: high-throughput SNP assays and association with Candida carriage in type I diabetics and nondiabetic controls. J Clin Microbiol 2003;41:90-96.
• ²² Braida L, Boniotto M, Pontillo A, Tovo PA, Amoroso A, Crovella S. A single-nucleotide polymorphism in the human beta-defensin 1 gene is associated with HIV-1 infection in Italian children. AIDS 2004;18:1598-1600.
• ²³ Milanese M, Segat L, Pontillo A, Arraes LC, de Lima Filho JL, Crovella S. DEFB1 gene polymorphisms and increased risk of HIV-1 infection in Brazilian children. AIDS 2006;20:1673-1675.
• ²⁴ Freguja R, Gianesin K, Zanchetta M, De Rossi A. Cross-talk between virus and host innate immunity in pediatric HIV-1 infection and disease progression. New Microbiol 2012;35:249-257.
• ²⁵ Freguja R, Gianesin K, Del Bianco P, et al. Polymorphisms of innate immunity genes influence disease progression in HIV-1-infected children. AIDS 2012;26:765-768.
• ²⁶ Prado-Montes de Oca E, Velarde-Félix JS, Ríos-Tostado JJ, Picos-Cárdenas VJ, Figuera LE. SNP 668C (−44) alters a NF-kappaB1 putative binding site in non-coding strand of human beta-defensin 1 (DEFB1) and is associated with lepromatous leprosy. Infect Genet Evol 2009;9:617-625.
• ²⁷ Zanin V, Segat L, Bianco AM, et al. DEFB1 gene 5′ untranslated region (UTR) polymorphisms in inflammatory bowel diseases. Clinics (Sao Paulo). 2012;67:395-398.
• ²⁸ Sandrin-Garcia P, Brandão LAC, Guimarães RL, et al. Functional single-nucleotide polymorphisms in the DEFB1 gene are associated with systemic lupus erythematosus in Southern Brazilians. Lupus 2012;21:625-631.
• ²⁹ Tiszlavicz Z, Somogyvári F, Szolnoki Z, et al. Genetic polymorphisms of human β-defensins in patients with ischemic stroke. Acta Neurol Scand 2012;126:109-115.
• ³⁰ Crovella S, Segat L, Amato A, et al. A polymorphism in the 5′ UTR of the DEFB1 gene is associated with the lung phenotype in F508del homozygous Italian cystic fibrosis patients. Clin Chem Lab Med 2011;49:49-54.
• ³¹ Nurjadi D, Herrmann E, Hinderberger I, Zanger P. Impaired β-defensin expression in human skin links DEFB1 promoter polymorphisms with persistent Staphylococcus aureus nasal carriage. J Infect Dis 2013;207:666-674.
• ³² Diagnostic Standards and Classification of Tuberculosis in Adults and Children. This official statement of the American Thoracic Society and the Centers for Disease Control andPrevention was adopted by the ATS Board of Directors, July 1999. This statement. Am J Respir Crit Care Med 2000;161(Pt1):1376–1395.
• ³³ Coelho A, Moura R, Cavalcanti C, et al. A rapid screening of ancestry for genetic association studies in an admixed population from Pernambuco, Brazil. Genet Mol Res 2015;14:2876-2884.
• ³⁴ Faul F, Erdfelder E, Lang A-G, Buchner A. G*Power 3: a flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behav Res Methods 2007;39:175-191.
• ³⁵ Excoffier L, Laval G, Schneider S. Arlequin (version 3.0): an integrated software package for population genetics data analysis. Evol Bioinform Online 2005;1:47-50.
• ³⁶ Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics 2005;21:263-265.
• ³⁷ López Campos GN, Velarde Félix JS, Sandoval Ramírez L, et al. Polymorphism in cathelicidin gene (CAMP) that alters Hypoxia-inducible factor (HIF-1α:: ARNT) binding is not associated with tuberculosis. Int J Immunogenet 2014;41:54-62.
• ³⁸ The International HapMap Consortium. A haplotype map ofthe human genome. Nature 2005;437:1299–1320.

Publication Dates

• Publication in this collection
  Apr-Jun 2016

History

• Received
  17 Sept 2014
• Accepted
  5 Sept 2015