Abstract
One in five children under the age of five worldwide suffers from stunting, a form of chronic undernutrition where children do not reach their full linear growth potential. Stunting is caused by multiple environmental factors and, when left untreated, leads to chronic health impairments. Unacceptably, current therapeutic interventions show limited efficacy, and rates of stunting remain alarmingly high. We seek to understand the molecular changes underlying cellular dysfunction in stunting to identify new therapeutic targets for treatment and prevention.
Using peripheral blood mononuclear cell (PBMC) samples from 18-week-old and one-year-old children and their mothers living in an urban slum in Bangladesh, we performed ChIP-seq targeting three histone marks: 1) an active promoter mark, histone 3 trimethylation on Lys-4 (H3K4me3); 2) an active enhancer mark, histone 3 acetylation on Lys-27 (H3K27ac); and 3) a mark associated with heterochromatin, histone 3 trimethylation on Lys-9 (H3K9me3).
To enhance the analysis of epigenetic landscapes in stunted children, we first introduce GenomicDistributions, a new computational tool for annotation and visualization of epigenetic regions that is then used throughout the rest of the dissertation.
Here we show that globally elevated H3K9me3 levels at 18 weeks of age were associated with poor linear growth between birth and one year. Many affected genes code for proteins targeting viral mRNA for degradation, and highly significant regions were enriched in transposon elements with potential regulatory roles in immune system activation and cytokine production. We thus speculate that high H3K9me3 levels may result in poor linear growth by repressing genes involved in immune system activation. At the same age, we further observed increased activation of H3K27ac regions with putative roles in stress responses and pathways active during viral infection, possibly due to facing a high pathogen load without a fully responsive immune system.
In one-year-old stunted children, we observed large-scale changes to activating histone marks, where H3K27ac was globally downregulated, and H3K4me3 redistributed from canonical sites near genes to distal ectopic loci. Both patterns indicate downregulated immune system function, possibly developed due to inadequate priming in early infancy or immune system exhaustion due to long-term pathogen exposure in an environment with limited nutrients. A second large category of genes associated with the downregulated H3K27ac regions at one year of age were metabolic genes, suggesting large-scale metabolic rewiring in stunted children. Notably, one-carbon metabolism was affected, possibly due to a lack of one-carbon precursors, including methionine, a metabolite whose relevance was not considered before with stunting. The H3K4me3 redistribution pattern provides further evidence for altered one-carbon metabolism in the one-year-old stunted children, as a similar pattern was previously observed in cells grown in methionine-limited conditions.
Analysis of maternal H3K4me3 data was not indicative of any intergenerational epigenetic relationships driving the stunted phenotype. And while maternal H3K9me3 profiles showed a similar trend to that observed in early infancy, the trend lacked statistical significance to infer an intergenerational relationship.
The studies presented in this dissertation provide the first comprehensive picture of epigenetic changes underlying stunting. Here we introduce increased H3K9me3 levels as a new early infancy biomarker of stunting and propose one-carbon metabolites, such as methionine and serine, as potential nutritional supplements for the treatment of stunting.
