DEFB1 gene polymorphisms and increased risk of HIV-1... : AIDS (original) (raw)

Milanese, Michelea; Segat, Ludovicaa; Pontillo, Alessandraa; Arraes, Luiz Claudiob,c; de Lima Filho, José Luizc; Crovella, Sergioa,c

Beta defensins are small secreted antimicrobial peptides that are components of innate immunity [1]. A role for beta defensins in contrasting infection from HIV-1 has been reported by Quiñones-Mateu et al. [2], who suggested that human beta defensin (hBD) 2 and hBD-3 are able to protect GHOST X4/R5 cells from being infected, by directly binding to the viral envelope. A similar effect, in which beta defensins interact with surface carbohydrates, has been shown by Leikina et al. [3]. Moreover Sun et al. [4] hypothesized an involvement of beta-defensins in HIV-1 oral transmission, emphasizing their protective role in the oral mucosa.

Despite the fact that Quiñones-Mateu et al. [2] were not able to prove a role for hBD-1 in protecting cells from infection by HIV-1, Braida et al. [5] recently reported an association between the −44(C/G) single nucleotide polymorphism (SNP) in the 5′ untranslated region (UTR) of the DEFB1 gene, encoding for the hBD-1 peptide, and HIV-1 infection in Italian children.

In our study we analysed the same SNPs in the 5′ UTR of the DEFB1 gene, −52(G/A), −44(C/G) and −20(G/A) (rs1799946, rs1800972, and rs11362, respectively), as reported by Braida et al. [5] in three groups of northeastern Brazilian children in order to assess the role of these polymorphisms in HIV-1 infection when looking at ethnic groups other that the Caucasians studied by Braida et al. [5].

We enrolled 128 HIV-1 perinatally infected children (average age 4 years, range 0–13) and 60 exposed uninfected children (average age 7 years, range 0–15) born to HIV-1-positive mothers who did not receive antiretroviral therapy during gestation, and did not undergo caesarean section to prevent vertical transmission. We also recruited 115 uninfected and unexposed children (average age 8 years, range 0–19) from the same population as the healthy controls. All children came from the Instituto Materno Infantil do Pernambuco (Recife, Brazil), and informed consent was obtained from their parents.

DNA extraction and DEFB1 SNP genotyping were performed according to Braida et al. [5].

The significance of differences in allelic and genotype frequencies was calculated by chi-square test. Yates' continuity correction and Fisher's exact test were employed when appropriate. The correction of P values for multiple tests was performed automatically using the R software (www.r-project.org) according to Benjamini and Hochberg [6] in order to keep the false discovery rate under 5%.

The allelic and genotype frequencies for 5′ UTR DEFB1 SNPs in the three groups studied are shown in Table 1. The genotype frequencies follow the Hardy–Weinberg equilibrium in the three populations, with the exception of the polymorphism −20(G/A) in HIV-1-infected children.

Table 1:

Frequencies of DEFB1 polymorphisms in children perinatally exposed to HIV-1 (infected and uninfected) and healthy controls.

DEFB1 genotype −52 A/A was significantly less frequent in HIV-1 perinatally infected children when compared with healthy controls and exposed uninfected children (corrected P values are 0.023 and 0.025, respectively), whereas frequencies did not vary between HIV-1-exposed uninfected children and healthy controls. The allelic frequencies of the −52(G/A) SNP in infected children varied significantly when compared with those of healthy controls (corrected P value 0.030), but when compared with exposed uninfected children were not statistically significant (corrected P value 0.055). Nevertheless, the odd ratios calculated comparing HIV-positive children with exposed uninfected children and healthy controls were not statistically different when tested using bootstrapping techniques, suggesting that the −52 A allele has a similar protective effect in the two groups.

A weaker correlation was found for the −44(C/G) polymorphism in which the P values for genotype frequencies did not remain statistically significant after multiple test correction. Allelic frequencies of the −44 G allele were similar between healthy controls and HIV-1-exposed uninfected children, and although a lower frequency of the −44 G allele was found in HIV-1-infected children, the difference was not statistically significant. A stronger association was observed in the −20(G/A) SNP, in which the frequency of the −20 G/G genotype was significantly lower in HIV-1-positive children than in exposed uninfected children and healthy controls (corrected P values both < 0.005). Allelic frequencies of the −20 G/A SNP were significantly different between HIV-1-positive children and healthy controls (corrected P value 0.015), whereas the significance was not achieved between HIV-1-infected and exposed uninfected children as a result of multiple test correction.

As we did not find any significant difference between the exposed uninfected children and the healthy controls, we pooled the two groups to perform a more powerful analysis using the Haploview software [7], built to perform single SNP and haplotype association tests. We confirmed the results found with the chi-square test, but despite the proximity of the three SNPs and a co-segregation rate of 93% between the −20(G/A) and −44(C/G) SNPs, we did not find evidence of the presence of any haploblock. Multiple regression analysis performed with the R software on the new dataset did not show any significant interaction between the different SNPs.

We have demonstrated that significant correlation exists between SNPs in the 5′ UTR of the DEFB1 gene and the risk of being infected with HIV-1 in Brazilian children. We found a significant increase of the −52 A/A and −20 G/G genotypes among HIV-1-infected children, when compared with healthy controls, and the odd ratios of carrying the −52 A or −20 G allele were, respectively, 1.688 (95% confidence limits 1.151–2.482) and 1.814 (95% confidence limits 1.243–2.655), suggesting a role for these polymorphisms in increasing the susceptibility to infection. We also found a sensible, even if not significant, reduction of the frequency of the −44 G allele: the frequency of this polymorphism was very low in the Brazilian population when compared with other populations [8], and this fact could account for the lack of statistical significance.

Finally, a particular consideration is deserved for the departure of the −20 G/A allele from the Hardy–Weinberg equilibrium. The excess of G/A heterozygous subjects in the exposed-infected children could possibly be explained by hypothesizing a dominant effect for the A allele, able to increase the susceptibility to HIV-1 infection even if present in a single dose.

In conclusion, our results obtained on Brazilian children confirm that the SNPs located in the 5′ UTR of the DEFB1 gene can be employed as a marker of risk for HIV-1 infection.

References

1. Kagan BL, Ganz T, Lehrer RI. Defensins: a family of antimicrobial and cytotoxic peptides. Toxicology 1994; 87:131–149.

2. Quiñones-Mateu ME, Lederman MM, Feng Z, Chakraborty B, Weber J, Rangel HR, et al. Human epithelial beta-defensins 2 and 3 inhibit HIV-1 replication. AIDS 2003; 17:F39–F48.

3. Leikina E, Delanoe-Ayari H, Melikov K, Cho MS, Chen A, Waring AJ, et al. Carbohydrate-binding molecules inhibit viral fusion and entry by crosslinking membrane glycoproteins. Nat Immunol 2005; 6:995–1001.

4. Sun L, Finnegan CM, Kish-Catalone T, Blumenthal R, Garzino-Demo P, La Terra Maggiore GM, et al. Human beta-defensins suppress human immunodeficiency virus infection: potential role in mucosal protection. J Virol 2005; 79:14318–14329.

5. 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.

6. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc 1995; 57:289–300.

Cited Here

7. Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics 2005; 21:263–265.

8. Jurevic RJ, Chrisman P, Mancl L, Livingston R, Dale BA. Single-nucleotide polymorphisms and haplotype analysis in beta-defensin genes in different ethnic populations. Genet Test 2002; 6:261–269.