Single-nucleotide polymorphisms in human [beta]-defensin-1 gene in Mozambican HIV-1-infected women and correlation with virologic parameters

We analyzed single nucleotide polymorphisms in the 5′-untranslated region (-44C/G and -52G/A) of the beta-defensin-1 gene in 78 Mozambican HIV-1-infected mothers. We observed significantly lower levels of HIV-1 RNA in breast milk, but not in plasma, in women with the -52GG genotype versus women with the -52GA and -52AA genotypes, supporting the hypothesis that different expression of β-defensins could have an impact on viral replication in breast milk.

Evidence suggests that β-defensin-1 (DEFB1) plays a role against infections ^[1,2] in inflammatory ^[3,4] and allergic processes ^[5,6]. Single nucleotide polymorphisms (SNPs) in the DEFB1 gene might be associated with alterations in β-defensin expression, and consequently with susceptibility to infections and mucosal disorders. In HIV-1 infection, significant correlations between the SNPs -44 G/C and -52 A/C in the 5′-untranslated region of the DEFB1 gene and a risk of vertical transmission were reported in Italian and Brazilian populations ^[7,8]. Recent in-vitro studies suggest that these SNPs possess a functional activity that leads to impaired gene expression ^[2,9].

We investigated the frequency of these SNPs in 78 Mozambican HIV-1-infected mothers and evaluated whether these polymorphisms were associated with virologic parameters. We studied two groups of women, highly active antiretroviral therapy (HAART)-treated and naïve, who had been included in a previous study within the DREAM Program (Drug Resource Enhancement against AIDS and Malnutrition program) aimed at evaluating effects of HAART on breast milk viral load, as described elsewhere ^[10]. Women from both groups did not breastfeed their infants. Breast milk was expressed manually five times a day for 1 week, and 1-week breast milk and plasma samples were selected for testing. Breast milk samples were centrifuged at 23 500 g for 1 h; cell pellets were washed in phosphate-buffered-saline and stored at -80°C. DNA was extracted from breast milk pellets using the QIAamp DNA Blood Mini kit (Qiagen, Hilden, Germany).

Polymorphic sites in genomic DNA were analyzed by the TaqMan allelic discrimination assay. Primers and probes specifically determine SNP site 668 (-44C/G) and 660 (-52G/A) were designed with the Primer Express software (version 1.5; Applied Biosystems, Foster City, California, USA) on the basis of the genomic DNA sequence of the DEFB1 gene (GenBank accession number U50930). The primers were: forward 5′-GAGGTTGTGCAATCCACCAGTCT-3′ and reverse 5′-GTTCTCATGGCGACTGGCA-3′. The probes were (allele-specific nucleotides are underlined): FAM-5′- AGCCAGCGTCTCCCCAGTTCC-3′-TAMRA (for 668G), VIC-5′-AGCCAGCCTCTCCCCAGTTCC-3′-TAMRA (for 668C), FAM-5′-GCTCAGCCTCCAAAGAAGCC-3′-TAMRA (for 660A), and VIC-5′-GCTCAGCCTCCAAAGGAGCC-3′-TAMRA (for 660G). PCR was performed in a thermal cycler (ABI Prism 7700, Applied Biosystems) in a reaction volume of 25 μl containing 600 nmol/l of each primer, 100 nmol/l of each probe, 12.5 μl of 2X TaqMan Universal PCR Master Mix (Applied Biosystems) and 1 ng of sample DNA. The thermal cycling conditions were 2 min at 50°C, 10 min at 95°C, and 45 cycles each of 95°C for 15 s and 60°C for 1 min. The genotypes were assigned using the Sequence Detection System software (version 1.9; Applied), analyzing the threshold cycle of amplification curves. Accuracy of genotyping was confirmed by known DNA samples of each genotype and by direct sequencing of randomly selected samples. Statistical analyses were performed using SPSS for Windows, version 13.0 (SPSS Inc., Chicago, Illinois, USA).

Median levels of HIV-1 RNA in plasma were 195 copies/ml (range 49-36 923) in HAART-treated, and 83 500 copies/ml (range 398-100 000) in naïve women. Corresponding values in breast milk were 91 copies/ml (range 49-2453) in HAART-treated, and 5334 copies/ml (range 49-100 000) in naïve women.

Analysis of the -44C/G polymorphism revealed that most mothers (87%) expressed the -44CC genotype, and that the -44GG genotype was absent from our population. This distribution is similar to that observed in Italian and Brazilian HIV-1-infected children ^[7,8]. The majority of HIV-1-infected Mozambican women (41.0%) expressed the -52GA genotype, 36% had the -52AA genotype, and 23% had the -52GG genotype. The proportion of Mozambican HIV-1-infected women with the -52AA genotype was higher than previously reported in Italian and Brazilian HIV-1-infected populations ^[7,8]. However, we do not know if this distribution of DEFB1 polymorphisms is also characteristic of HIV-1-negative Mozambican population.

No direct correlation between the different polymorphisms in DEFB1 with vertical transmission could be made, because data regarding maternal HIV-1 transmission were incomplete (i.e., most naïve mothers were lost to follow-up). However, we could investigate if the rate of viral replication was correlated with the -44C/G and -52G/A polymorphisms. The two groups of women were considered separately because of the significant differences in viral load in plasma and breast milk. In HAART-treated women, no correlation was made since most women had very low or undetectable HIV-1 RNA levels. Among naïve women, no correlation was found between the rate of viral replication in breast milk or in plasma and the distribution of the -44C/G polymorphism, probably as a result of the relative genetic homogeneity of our population, who for the most part expressed the -44CC genotype. Interestingly, analyzing the -52 SNP, we found that the subgroup presenting the -52GG genotype had lower levels of HIV-1 RNA in breast milk (1562 copies/ml, range 49-57 742) compared with women with the -52GA (19 545 copies/ml, range 49-100 000) and -52AA genotype (3575 copies/ml, range 159-100 000) (P = 0.03 and P = 0.250, respectively), whereas no significant differences were seen in HIV-1 RNA plasma levels.

We then analyzed in naïve women the distribution of genotypes, categorizing women according to the median value of HIV-1 RNA in plasma and in breast milk. We found that most women (78%) with the -52GG genotype had HIV-1 RNA levels below the median in breast milk (Fig. 1a). The difference was significant with the -52GA genotype (P = 0.01) but not with -52AA genotype subgroup (P = 0.418). Again, the same analysis in plasma (Fig. 1b) did not reveal any significant difference, suggesting a functional role of DEFB1 in breast milk only.

Fig. 1

It has been reported that breast milk has higher levels of DEFB1 than other compartments ^[11]. It is possible that high levels of hormones during pregnancy and lactation may up-regulate DEFB1 expression; thus, DEFB1 may be produced at higher concentrations by mammary gland epithelia than other mucosal surfaces ^[11].

In conclusion, for the first time, we report a significant association between viral load in breast milk and DEFB1 polymorphisms. This finding supports the hypothesis that, independent from other maternal compartments, breast milk cells produce factors in response to stimuli, and suggests a functional role of breast milk cells in the innate immune response to HIV-1 infection.

Acknowledgements

We thank Roberta Amici and Maria Grazia Mancini for their assistance, Patrizia Cocco, Fernando Costa, and Daniela Diamanti for technical support, and Andrea Cara for helpful advice. We are grateful to Stefania Donnini for administrative support.