Elucidating Gene-by-Environment Interactions Associated with Differential Susceptibility to Chemical Exposure

Abstract

Background: Modern societies are exposed to vast numbers of potentially hazardous chemicals. Despite demonstrated linkages between chemical exposure and severe health effects, there are limited, often conflicting, data on how adverse health effects of exposure differ across individuals.

Objectives: We tested the hypothesis that population variability in response to certain chemicals could elucidate a role for gene-environment interactions (GxE) in differential susceptibility.

Methods: High-throughput screening (HTS) data on thousands of chemicals in genetically heterogeneous zebrafish were leveraged to identify a candidate chemical (Abamectin) with response patterns indicative of population susceptibility differences. We tested the prediction by generating genome-wide sequence data for 276 individual zebrafish displaying susceptible (Affected) vs. resistant (Unaffected) phenotypes following identical chemical exposure.

Results: We found GxE associated with differential susceptibility in the sox7 promoter region and then confirmed gene expression differences between phenotypic response classes.

Conclusions: The results for Abamectin in zebrafish demonstrate that GxE associated with naturally occurring, population genetic variation play a significant role in mediating individual response to chemical exposure. https://doi.org/10.1289/EHP2662.

Figure 1.

Study Design. (A) Chemical selection from HTS data: Example concentration–response curves from 1,060 chemicals interrogated for adverse morphological end points. Each panel represents a test chemical where the proportion of individuals displaying adverse morphological development (vertical axis) is plotted against the tested concentrations (horizontal axis). The curve with the asterisk in the upper left-hand corner of the panel represents a chemical response suggestive of differential population susceptibility, whereas all other curves depict steeper toxic points of departure (i.e., less spread in the range of concentrations eliciting effects across the population) or lack of response. (B) Rangefinders: Successive screens to find the critical concentration as that at which approximately 50% incidence is observed. The heatmaps show horizontal blocks (separated by whitespace) of identical concentrations, whose height corresponds to the number of zebrafish tested. Within each concentration block, each row is the vector of observed morphological end points (17 columns; see “Methods” section) for an individual. As per the legend at the lower right, blue represents no end point incidence, red represents incidence of an end point, and grey cross-hatching represents mortality. (C) Critical concentration exposure: Example of a single exposure plate, where 72 individuals (in single wells) were exposed to $0.6\mathrm{\mu }\mathrm{M}$ Abamectin at 6 hpf, plus 24 individuals exposed to vehicle [dimethylsulfoxide (DMSO)] controls. Developmental morphology screening was performed at 120 hpf to identify Affected individuals with the phenotype of altered eye (EYE), snout (SNOU), jaw (JAW), pericardial edema (PE), yolk sac edema (YSE), and axis development (AXIS) vs. Unaffected individuals with no observed defects. (D) Individual DNA extraction: Individuals classified as Affected and Unaffected were randomly selected for whole-genome sequencing.

Figure 2.

Genome-wide association study (GWAS) results for Abamectin. The Manhattan plot shows the genomic coordinate for each SNP on the horizontal axis (grouped into chromosomes) vs. its association with phenotypic status on the vertical axis (as the negative logarithm of p-value). The horizontal line indicates the Bonferroni-adjusted significance threshold. The dots above this line indicate candidate SNPs (cfap on chromosome 8, erf on chromosome 19, sox7 on chromosome 20) for validation as genetic factors associated with differential susceptibility (i.e., Affected vs. Unaffected phenotypes) to Abamectin exposure.

Figure 3.

Functional Validation of sox7. (A) Depiction of the sox7 transcript, gene expression primer locations, and frequency sequence logos for the region surrounding the significant SNP (20:19,166,444) in Affected and Unaffected individuals from the genome-wide association study (GWAS). Sequence logos are centered at the SNP site, denoted as position 0. The relative letter height corresponds to the frequency of the base at that position. (B) Notched boxplots showing the distribution of ${\log }_2$ (fold change) of sox7 expression for Control (unexposed), Unaffected (exposed), and Affected (exposed) groups. The boxplots show the median (thick horizontal lines), the upper and lower quartiles (top and bottom of boxes, respectively), 1.5 times the interquartile range (whiskers), and any outlier samples outside the whisker range of the observed expression for each group.