Tumor Necrosis Factor-α, Interleukin-10, and α-Defensins in Plasma and Breast Milk of HIV-Infected Highly Active Antiretroviral Therapy-Treated and Untreated Pregnant Women in Mozambique

To the Editor: The risk of HIV transmission through breast milk has been shown to be dependent on the size of the inoculum,1 the amount of cell-free and cell-associated HIV,2 the presence of coinfections3 and of antiviral substances such as defensins,4 and the local specific immune response to HIV.5 Other yet undefined factors may be involved in determining the rate of mother-to-child transmission, however, and further studies on the analysis of breast milk components could help in clarifying the mechanisms that could favor or prevent transmission during the lactation period. In the present study, we evaluated the concentrations of tumor necrosis factor-α (TNFα), interleukin-10 (IL-10) and α-defensins in plasma and breast milk of highly active antiretroviral therapy (HAART)-treated (group A, n = 20) and untreated (group B, n = 34) HIV-positive mothers in Maputo, Mozambique. The design of the study has been extensively described elsewhere.6 Group A comprised pregnant women attending the Ante Natal Clinic in Matola and enrolled in the Drug Resource Enhancement Against AIDS and Malnutrition (DREAM) program who received antiretroviral regimens (generic formulation of zidovudine [or stavudine if hemoglobin [Hb] <8 g/dL] plus lamivudine and nevirapine) from 28 weeks of gestational age until 1 month postpartum. Group B women included mothers who were tested for HIV at delivery; if positive, they were asked to participate in the study. Women from both groups did not breast-feed their infants. Breast milk was expressed manually 5 times a day for 1 week, and 1-week breast milk samples were selected for testing. At the same time, blood samples were also collected and clinical examinations were performed to recognize the presence of mastitis or breast inflammation. TNFα and IL-10 were detected in plasma and skim milk samples by enzyme-linked immunosorbent assay (ELISA; Biosource Hu-TNF-α US and Hu-IL-10 US, respectively; BioSource International, Camarillo, CA). Concentrations of α-defensins were determined using a commercial ELISA kit recognizing human alpha defensins (HNP) 1 to 3 (hemoglobin [Hbt] Human HNP 1 to 3; Hycult Biotechnology, Uden, The Netherlands). All statistical analyses were performed using SPSS for Windows version 13.0 (SPSS, Chicago, IL). Twelve (60%) HAART-receiving women had detectable HIV RNA in plasma versus 100% of untreated mothers (P < 0.001); HIV RNA in breast milk was lower than the limit of detection in 45% of treated mothers, whereas it was detectable in 91.2% of untreated mothers (P = 0.005). The concentrations of HIV RNA in plasma and breast milk were significantly lower in HAART-treated women (P < 0.001), who also had better preservation of their peripheral blood CD4 cell count (P = 0.002) with respect to untreated mothers. HIV DNA in breast milk cells was detectable in 30% and 47.1% of group A and B women, respectively. In untreated women, significantly higher levels of TNFα, IL-10, and α-defensins were detected in breast milk with respect to plasma (P < 0.001, P = 0.035, and P = 0.001, respectively), whereas in HAART-treated women, similar levels of IL-10 and α-defensins were found in the 2 maternal districts (P = 0.309 and P = 0.445, respectively). Only TNFα concentration in breast milk significantly increased in comparison to plasma levels in both groups of mothers (P < 0.001; Fig. 1A). An analysis was also performed including samples of both groups to determine the correlation between TNFα, IL-10, and α-defensins and viral replication levels in plasma and in breast milk. TNFα positively correlated with HIV RNA viral load in plasma (r = 0.338, P = 0.012). Conversely, in plasma, no correlations were detected between HIV RNA and IL-10 (r = 0.232, P = 0.091) or α-defensins (r = 0.050, P = 0.724). In breast milk, we found a strong correlation between HIV RNA viral load and TNFα, IL-10, and α-defensins (see Fig. 1B).FIGURE 1: A, TNFα, IL-10, and α-defensin concentrations found in plasma (▪) and breast milk (□) samples collected 7 days after delivery in HAART-treated (group A) and naive (group B) mothers. Values are reported in pg/mL (TNFα and IL-10) and ng/mL (α-defensins). B, Correlations between soluble factor levels and HIV RNA viral load in breast milk. Correlations were investigated by using the Spearman correlation test. Gray and black triangles indicate group A and group B values, respectively. P < 0.05 was considered significant.This study suggests that cytokine production in plasma and in breast milk is independent and that HAART therapy influences breast milk cytokine profiles. In fact, naive mothers showed breast milk levels of TNFα, IL-10, and α-defensins from 2- to 6-fold higher than in plasma, whereas in HAART-treated mothers, only breast milk TNFα levels were significantly higher than the plasma levels. The high levels of these soluble factors in breast milk, where HIV RNA copies/mL are lower than in plasma, suggest that their concentrations are unlikely to be simply a reflection of those found in the plasma and support the hypothesis that breast milk cells productively release factors in response to stimuli, such as HIV replication. In this view, differently from plasma, TNFα, IL-10, and α-defensins in breast milk highly correlated with HIV RNA viral load in both groups of women. An active role of breast milk cells has been demonstrated by spontaneous release of IL-1β and TNFα7 and by the fact that cellular components of breast milk are able to produce substantial quantities of many known cytokines in response to various stimuli.8 Regarding α-defensins, their role in milk is controversial,5,9 but higher α-defensin concentrations found in the breast milk could be involved in processes reducing infectivity of the virus in breast milk or increasing infant resistance; their exact role in mother-to-child transmission needs to be explored in large clinical studies. A limitation of our study is that we could not ascertain whether the elevated concentrations of cytokines or soluble factors could be important in influencing postnatal transmission because these mothers did not breast-feed. Nevertheless, our study provides new data on the breast milk compartment of HIV-infected mothers and supports the hypothesis of a compartmentalization between plasma and breast milk and of a functional role of breast milk cells in an adaptive and innate immune response to HIV infection. ACKNOWLEDGMENTS The authors thank Patrizia Cocco, Daniela Diamanti, and Fernando Costa for technical support. They are grateful to Alessandra Mattei for administrative support. Silvia Baroncelli, BSc* Mauro Andreotti, BSc* Giovanni Guidotti, MD* Maria F. Pirillo, BSc* Susanna Ceffa, BSc†‡ Maria Grazia Mancini, MT* Paola Germano, MSc† Maria Cristina Marazzi, MD§ Stefano Vella, MD* Leonardo Palombi, MD∥ Marina Giuliano, MD* *Department of Drug Research and Evaluation Istituto Superiore di Sanità Rome, Italy †Community of Sant'Egidio DREAM Program Rome, Italy ‡University of Pisa Pisa, Italy §Libera Università Maria Ss. Assunta DREAM Program Rome, Italy ∥Department of Public Health University of Tor Vergata DREAM Program Rome Italy