Cellular immune ontogeny and birth transcriptome in HIV-exposed uninfected infants

Doctoral Thesis


Permanent link to this Item
Journal Title
Link to Journal
Journal ISSN
Volume Title
Background. In some regions of Sub-Saharan Africa, up to 30% of newborns are born to mothers infected with human immunodeficiency virus (HIV). Maternal antiretroviral treatment (ART) has reduced vertical transmission to lower than 1%. Despite the success of prevention of mother-tochild transmission (PMTCT) programmes, a large number of children born to these mothers are exposed to HIV and antiretrovirals (ARVs) in utero yet remain uninfected. These individuals, known as children who are HIV-exposed and uninfected (cHEU), succumb to higher rates of disease morbidity compared to children who are HIV-unexposed (cHU) which suggests altered immunity in the cHEU. Differences in the numbers and function of cells of the innate and adaptive immune system have been documented in cHEU—though not consistently. While vaccine-induced antibody responses are robust in cHEU, data on potential cell mediated perturbations to vaccine antigens remains conflicting. This is in part due to inherent inter-cohort variation and differences in ART therapy strategies, feeding practices between cohorts and the assays used measure cellmediated responses. We leveraged two independent cohorts from Nigeria and South Africa of mother-infant pairs receiving antenatal and postnatal care all under Option B+ PMTCT. All HEU infants received pre-exposure prophylaxis for 6 weeks and the majority were exclusively breastfed until 6 months of age. We applied the same assays in both cohorts to test the hypotheses that HEU have altered T cell immunity compared to HU controls and distinct transcriptomic signatures at birth. These were tested in three distinct aims: 1) To identify transcriptional signatures at baseline that delineate cHEU from cHU 2) To compare the expression of surface marker broadly defining activated or regulatory phenotypes and the expression of intracellular markers of T cell function between cHEU and cHU over the first 9 months of life. 3) To characterise how differences in the immunising strains of Bacille Calmette-Gu'erin (BCG), the first vaccine received in these infants, impacts T cell immunity to both mycobacterial and non-mycobacterial antigens in cHEU and cHU. Methods. Two birth cohorts from Jos, Nigeria and Cape Town (CT), South Africa were recruited into this study as part of a larger parent study that aims to identify biological determinants of protection from mother-to-child transmission of HIV (Innate, Adaptive and Mucosal Immune Responses in Infants/INFANT study: HREC 285/2012). Infant blood was collected at several time points from birth to 36 weeks of life for immunological assays. Whole blood, collected at birth was preserved in PAXgene fluid for downstream messenger ribonucleic acid (mRNA) transcript analyses. Other whole blood samples were fixed and cryopreserved either directly ex vivo or after re-stimulation within 1 hour of phlebotomy with BCG, Tetanus Toxoid (TT), Bordetella pertussis (BP) antigens and Phytohemagglutinin (PHA). Multi-parameter flow cytometry was used to measure batched whole blood samples for (i) markers of T cell regulation and activation directly ex vivo, markers of T cell gut homing and proliferation—a proxy for HIV susceptibility, and (ii) vaccine-induced Th1 cytokine expression (IFN-, TNF-a, IL-2) and memory maturation. Cytokine responses were profiled for polyfunctionality by SPICE analysis and complemented by the COMPASS algorithm. Transcriptional profiling of whole blood at birth was done by RNA sequencing and differentially expressed genes were reported for absolute fold change of normalized counts were < 1.5 with FDR set at 0.05 using the DESeq2 package in R. Gene-set enrichment analysis (GSEA) was used to identify enriched or repressed gene pathways for absolute normalised effect sizes < 1.5 with FDR set at 0.05. Longitudinal analyses used a mixed effects ANOVA with time and HIV exposure as explanatory variables. Cross-sectional analyses comparing HIV exposure groups used Wilcoxon Ranked Sum Test, with p< 0.05 considered significant after multiple correction adjustment by Holm's step-down method. Results. Aim 1: A small set of DEGs were found between HEUs and HU groups at birth, 3 of which were upregulated and 12 that were downregulated. Among the upregulated genes, two are homologues of the arrestins: ARRDC4 (2.3 fold, adjusted p-adj< 0.001) and TXNIP (1.4 fold, padj< 0.001). Gene-set enrichment analysis however, showed no significant enrichment or suppression of gene pathways in HEUs. Aim 2: HIV/ARV exposure did not have an interaction effect with age (all time points) in explaining the frequencies of T cell markers ex vivo in a mixed-effects model. In cross-sectional unadjusted analyses however, trends towards increased median frequencies of markers of activation in the HEU group compared to HU controls were observed for specific ages: at birth (%CD8+HLA-DR+: 0.12 vs. 0.01, p=0.05), at week 7 (%CD8+CD25+: 0.13 vs. 0.04, p=0.01 and %CD8+HLA-DR+: 0.84 vs. 0.07, p=0.01) and at week 36 (%CD8+CD25+: 0.52 vs. 0.03, p< 0.001 and %CD8+HLA-DR+: 0.81 vs. 0.17, p=0.003). When adjusting for multiple comparisons, only CD25 expression remained significant on CD8+ T cells at week 36 (p-adj =0.04). The magnitudes of cytokine responses by T cells to vaccine antigens did not differ between HEU and HU infants however, transient differences in the polyfunctional profile of cells was observed at week 1 for mycobacterial-specific Th1 profiles in CT infants (p=0.002) by SPICE analysis. There were later differences at week 7 for BP-specific Th1 profiles in Jos infants (p=0.01) and at week 36 for BP-specific Th1 profiles in CT infants (p=0.03). The more robust COMPASS algorithm only detected a trend towards increased polyfunctional scores to BP responses in CT infants at week 36 (p=0.03). Aim 3: BCG immunising strain impacted the magnitudes and quality of responses to mycobacterial and non-mycobacterial vaccine antigens irrespective of HIV exposure status. Most significantly, at week 7, BCG-Denmark induced higher mycobacterial-specific frequencies of CD4 Th1 cytokines compared to Bulgarian (p< 0.001) and Russian strains (and (p< 0.001). BCGDenmark induced greater triple cytokine profiles to mycobacterial antigen compared to Bulgarian (p< 0.001) and Russian (p< 0.001) strains in SPICE analyses and the resulted were confirmed by COMPASS algorithm polyfunctional scores. Furthermore, BCG-Denmark significantly enhanced antigenicity to TT and BP vaccines. Conclusion. Transient differences exist in the frequencies of CD25 expressing CD8 T cells between HEU and HU groups, however other readouts of immunity suggest that in the context of effective PMTCT and exclusive breastfeeding practices, HEU infants are indistinguishable from their HIV unexposed peers.