BACKGROUND : Subclinical mastitis (SCM) is relatively common in lactating women and may be associated with HIV
shedding in breast milk. The potential association between HIV infection and breast milk immunologic factors and
immune response to SCM needs to be addressed.
METHODS : In this cross-sectional study, SCM (Na/K ratio > 1) was tested in 165 mature breast milk samples collected
from 40 HIV-infected women who didn’t transmit HIV to their child by breastfeeding and 43 HIV-uninfected women
enrolled in an interventional cohort in South-Africa (Vertical Transmission Study). The level of 33 immune markers
related to Th1/Th2 related response, inflammation and bacterial exposure were compared in ART-naive HIV-infected
versus HIV-uninfected women. The associations between HIV infection and SCM on the concentration of immune
factors were tested separately by Wilcoxon rank-sum test and corrected for false discovery rate. To control for potential
confounder effects and take into account the clustering of breast milk samples from a single woman, multivariate
mixed linear models adjusted on child age at the time of sampling were performed for each immune factor.
RESULTS : Subclinical mastitis was detected in 15 (37.5%) HIV-infected women and 10 (23.3%) HIV-uninfected women.
In the absence of SCM, the breast milk levels of IP-10 and MIG were higher and IL1-RA lower in HIV-infected women
than in HIV-uninfected women (respectively p < 0.001, p = 0.001, p = 0.045). In HIV-uninfected women, SCM was characterized
by a robust immune response with higher concentrations of a broad panel of Th1 and inflammatory related
immune markers than in samples without SCM. By contrast, in HIV-infected women a limited number of immune
markers were increased and lower increases were observed in samples with SCM than without SCM.
CONCLUSION : HIV infection in ART-naïve women was associated with elevated breast milk levels of IP-10 and MIG,
which areTh1-related cytokines induced by IFN-γ. During SCM, a lower and narrower immune response was observed
in HIV-infected than HIV-uninfected women, suggesting that HIV infection affects the capacity of the mammary gland
to respond to SCM.
Additional file 1: Table S1. Detection rates of immune factors in samples
with and without sub-clinical mastitis, by HIV group. This table compares
the detection rates of immune factors measured in mature breast milk
samples with and without sub-clinical mastitis, separately for samples
from HIV-infected and HIV-uninfected women. There were no detection
rate differences between samples with and without SCM in samples from
HIV-infected women. In comparison, in samples from HIV-uninfected
women, interferon-α, interferon-γ, interleukin-4, α-defensin, Tumor Necrosis
Factor-α and interleukin-6 were more frequently detected in samples
with SCM compared to samples without SCM.
Additional file 2: Table S2. Multivariate models assessing the effect
of HIV on immunologic factor concentration in samples without SCM.
This table indicates the adjusted regression coefficients and associated
p-values of multivariate mixed linear models assessing the effect of HIV
infection on each breast milk soluble immunologic factor concentration,
in samples without sub-clinical mastitis, adjusted on child age at the time
of sampling. HIV infection was associated with an increase of monokine
induced by gamma interferon, inflammatory protein-10, C-reactive protein and ß2-microglobuline, and a decrease of receptor antagonist of
interleukin 1ß and soluble CD14.
Additional file 3: Table S3. Correlations between breast milk immune
factor concentration, HIV plasma parameters and breast milk HIV RNA. This
table indicates the Spearman’s non parametric correlations and associated
p-values between breast milk soluble immunologic factor concentration
and plasma HIV viral load (approximately 6 months after delivery), plasma
CD4 count (approximately 6 months after delivery) and breast milk HIV
RNA (at the time of breast milk sampling), in samples from HIV-infected
women only. There are only a few weak correlations (maximum |ρ| = 0.44).
Additional file 4: Table S4. Breast milk immune factor comparisons
between samples with sub-clinical mastitis from HIV-uninfected and
HIV-infected women. This table compares the concentration of immune
factors measured in mature breast milk samples with sub-clinical mastitis,
between samples from HIV-infected and HIV-uninfected women. There
were no statistically significant concentration differences in samples with
sub-clinical mastitis between HIV-infected and HIV-uninfected women.
Additional file 5: Table S5. Multivariate models assessing the effect
of SCM on immunologic factor concentration, by HIV group. This table
indicates the adjusted regression coefficients and associated p-values of
multivariate mixed linear models assessing the effect of sub-clinical mastitis
on each breast milk soluble factor concentration, adjusted on child
age at the time of sampling, separately for samples from HIV-infected and
HIV-uninfected women. Sub-clinical mastitis was associated with and
increase of 9/13 immune factors analyzed in samples from HIV-uninfected
women compared to 7/17 immune factors analyzed in samples from