Differences in multiple immune parameters between Indian and U.S. infants

Abstract

To compare immune phenotypes across two geographic and ethnic communities, we examined umbilical cord blood by flow cytometry and Luminex in parallel cohorts of 53 newborns from New Delhi, India, and 46 newborns from Stanford, California. We found that frequencies of a B cell subset suggested to be B-1-like, and serum IgM concentration were both significantly higher in the Stanford cohort, independent of differences in maternal age. While serum IgA levels were also significantly higher in the Stanford cohort, IgG1, IgG2, and IgG4 were significantly higher in the New Delhi samples. We found that neutrophils, plasmacytoid dendritic cells, CD8+ T cells, and total T cells were higher in the U.S. cohort, while dendritic cells, patrolling monocytes (CD14dimCD16+), natural killer cells, CD4+ T cells, and naïve B cells were higher in the India cohort. Within the India cohort, we also identified cell types whose frequency was positively or negatively predictive of occurrence of infection(s) in the first six months of life. Monocytes, total T cells, and memory CD4+ T cells were most prominent in having an inverse relationship with infection. We suggest that these data provide impetus for follow-up studies linking phenotypic differences to environmental versus genetic factors, and to infection outcomes.

Fig 1. Flow cytometric-gating strategies for cell lineages from Indian and US cord blood mononuclear cells (PBMCs).

After primary gating for CD45, various cell subsets were gated as shown and percentage values for each subset calculated. (A) First row: Gating for monocytes (CD45+CD66b-CD14+CD11c+), and among them, for patrolling monocytes (CD14dimCD16+), inflammatory monocytes (CD14+CD16+), and classical monocytes (CD14+CD16-). Second row: CD45+CD20+ B cells were further gated for B1 B cells (CD45+CD20+CD27+CD43+), and naïve B cells (CD45+CD20+CD27-). Naïve B cells were further gated as immature naïve B cells (CD45+CD20+CD27-CD19+CD10+). Third row: Gating for dendritic cells (DCs) (CD45+Lineage- [CD3-CD14-CD16-CD19-CD20-CD56-] HLA-DR+) and among them, for myeloid DCs (mDCs; CD45+lineage-HLADR+CD11c+CD123-) and plasmacytoid DCs (pDCs; CD45+lineage-HLA-DR+CD11c-CD123+). Fourth row: Gating for Monocytes (CD45+CD14+), NK (Natural Killer) cells (CD45+CD3-CD56+), B cells (CD45+CD19+) and Transitional B cells (CD45+CD19+CD24+CD38^bright). (B) CD45+CD3+ cells identified as T cells. First row: Classical T cells (CD45+CD3+CD25-CD56-TCRg/d-) which were sub-divided into CD4 and CD8 cells. CD4 and CD8 T cells were further characterized using CD45RA and CCR7 as naïve (CCR7+CD45RA+) cells, central memory (CCR7+CD45RA-) and TEMRA (CD8+CCR7-CD45RA+). Second row: CD45+CD3+ T cells were further gated for putative regulatory T cells (CD45+CD3+CD4+CD25+CD127-). Third row: Gating for gamma-delta T cells (CD45+CD3+TCRg/d+), NKT cells (CD45+CD3+CD56+) and invariant NKT (iNKT) cells (CD45+ CD3+V alpha 24-J alpha 18 TCR+).

Fig 2. Differences in 'B-1' B cells and IgM between Indian and U.S. cohorts.

B-1-like B cell percentages are shown in the left panel, with the blue dot indicating the mean, and whiskers extending from minimum to maximum values of observed data. B-1-like cells were gated as CD45+CD20+CD27+CD43+ and expressed as a percentage of total B cells (CD45+CD20+). IgM levels are shown in the right panel, as measured by Luminex assay, and expressed as ng/ml.

Fig 3. Differences in non-IgM isotypes between Indian and U.S. cohorts.

Ig isotype concentrations were determined by Luminex and compared between cohorts. The horizontal bar in each plot indicates the median, with whiskers extending from minimum to maximum values of observed data. Significant differences are noted. There was one U.S. sample below the lower limit of detection (LOD) for IgG2; 10 Indian and 10 U.S. samples above the upper LOD for IgG3; 3 Indian samples above the upper LOD for IgG4; and 2 Indian and 2 U.S. samples below the lower LOD for IgA. 00.

Fig 4. Differences in other cell subsets between Indian and U.S. cohorts.

Cell subset frequencies, reported as percent of the parent subset, were regressed on country as described in the Methods. A number of significant differences in country means were detected as shown, after correction for multiple comparisons [15]. Red asterisks indicate significantly higher mean in India cohort; black asterisks indicate significantly higher mean in U.S. cohort. Neither maternal age or the interaction of country and maternal age was a statistically significant predictor of these cell-subset frequencies. Blue dots indicate the mean, with whiskers extending from minimum to maximum values of observed data.

Fig 5. Association of cell subset frequencies with infection outcome in the Indian cohort.

Regression coefficients are shown for individual features. A multivariable regression model had estimated specificity of 89% and sensitivity of 94% in predicting infant infection during the first six months of life.