DNA methylation patterns associated with konzo in Sub-Saharan Africa

Comparative analysis of normalized intensities derived from interrogation of over 850,000 methylation probe sites between konzo cases and age- and sex-matched healthy controls suggests that there are 117 differentially methylated probes (DMPs) significantly associated with konzo-affected individuals (Fig. 1A). Of the 117 total sites of differential methylation, 99 DMPs were hypomethylated, while 18 sites were hypermethylated compared to controls (Additional file 1: Table S4). The EPIC array data were validated using next-generation targeted bisulfite sequencing, which confirmed our findings at 33 of the most differentially methylated sites (Additional file 1: Table S3). Unsupervised hierarchical clustering of DMP intensities revealed that samples cluster strongly by cohort, except for one sample (K1), which appears to cluster more with the control cohort than konzo (Fig. 1A). Without other clinical information, it is not possible to determine if this sample is clustering differently due to other phenotypic associations (i.e., disease onset or severity) or a true outlier of the study. We did however rule out the influence of immune-cell-type contributions, which were not observed to be significantly different in our analysis (Fig. 1B).

Since there are apparent sex-specific differences in the manifestation and presentation of konzo, where females of child-bearing age appear to be more vulnerable than males (although this may be attributed to social differences), we interrogated konzo males versus konzo females and did not identify significantly differentially methylated loci attributed to sex (data not shown) [7]. However, the analyzed cohort is small and may not have had enough statistical power to uncover sex-specific differences in DNA methylation associated with disease [7].

Of the 117 DMPs between konzo and controls, there were 2 genes with multiple differentially methylated probes that were significantly differentially methylated in the konzo cohort, ZNF718 and AKAP12 (Additional file 1: Table S4). All sites associated with the genes ZNF718 and AKAP12 were significantly hypermethylated (FDR p value ≤ 0.05 and log2 fold-change in DNA methylation intensity ≤ 1) in konzo cases, compared to controls (Additional file 1: Table S4). 46 of these sites associated with genes were identified in the promoter region (TSS or 5′ UTR). These 46 sites were analyzed for gene ontology (GO) enrichment at a statistical threshold of p ≤ 0.05 and we were able to ascertain that the konzo cohort was enriched for biological processes relevant to konzo etiology and potentially relevant pathways (Fig. 1C, Additional file 1: Table S7). For example, among the top enriched terms, we noted regulation of skeletal muscle contraction (GO: 0014819), which may be directly relevant to the spastic movements and paraparesis that are characteristically associated with the konzo phenotype (Fig. 1C) [2]. Additionally, using the Online Mendelian Inheritance of Man database, we identified that the associated gene for this GO term, KCNJ2, has been implicated in other disorders of periodic paralysis, such as Andersen Syndrome, and thus may be directly relevant to konzo disease presentation [11].

Of additional interest, we noted significant enrichment for the biological process queuosine metabolic process (GO: 0046116). Queuosine is a modified nucleoside present in certain mammalian tRNAs and its abundance has been linked to the presence of micronutrients derived from the gut microbiome and directly links to transcriptional regulation [12, 13]. As konzo onset is linked not only to dietary exposure to cyanogenic glucosides, but also to SAA deficiency, an adjunct role of the gut microbiome could also play into the disease phenotype and be linked to changes in DNA methylation, transcription, and metabolic processes. The literature suggests that DNA methylation is strongly influenced by the environment, so changes in diet that are known to be associated with disruptions in molecular processes and the gut microbiome, like queuosine metabolism, may be of interest for elucidating the complex mechanisms associated with konzo disease onset and progression.

Overall, while enrichment of these biological processes may suggest a role for modifications to DNA methylation in disease phenotype, these 117 sites of differential methylation may serve as biomarkers for monitoring populations at-risk for konzo. Functional validation is critical to further explore these findings and understand the impact of these differentially methylated sites in the context of dietary cyanogenic glucoside exposure and konzo presentation, as well as determine if there are specific epigenetic markers associated with susceptibility or risk to developing a clinical phenotype, as konzo does not present in all who are exposed to the same, homogenous diet of cyanogenic cassava.

A limitation of this study is the small sample size (n = 32) that was used. While we were able to determine statistically significant DNA methylation differences between our cohorts, future studies should look to expand on the size of the cohort used, increase the age range, and include numerous disease severity levels, and ensure that a sufficient sample size is used for a well-powered study. Additionally, by correlating the level of in vivo cyanogenic glucoside metabolites in serum or urine at the time of collection could provide invaluable information regarding the level exposure of each individual to cyanogenic cassava through the diet and further be associated with the DNA methylation changes present. As previously mentioned, children are also at high-risk groups for developing konzo. In this cohort, the median age of recruited konzo and healthy control individuals was approximately 13 years old. As such, we are unable to draw conclusions regarding the DNA methylation patterns associated with pediatric versus adolescent ages groups. Future research should aim to observe longitudinal progression of this disease and consider timing of onset and severity of these age groups, which may elucidate the molecular underpinnings of the sudden and irreversible phenotype associated with konzo.

This study has provided the first analysis of epigenetic changes associated with clinical diagnosis of konzo. Future experiments should focus on further identifying biomarkers of low-dose dietary cyanide exposure and identifying factors of konzo disease susceptibility and pathobiology through other molecular approaches.

留言 (0)

沒有登入
gif