Discovery of genomic intervals that underlie nematode responses to benzimidazoles

Abstract

Parasitic nematodes impose a debilitating health and economic burden across much of the world. Nematode resistance to anthelmintic drugs threatens parasite control efforts in both human and veterinary medicine. Despite this threat, the genetic landscape of potential resistance mechanisms to these critical drugs remains largely unexplored. Here, we exploit natural variation in the model nematodes Caenorhabditis elegans and Caenorhabditis briggsae to discover quantitative trait loci (QTL) that control sensitivity to benzimidazoles widely used in human and animal medicine. High-throughput phenotyping of albendazole, fenbendazole, mebendazole, and thiabendazole responses in panels of recombinant lines led to the discovery of over 15 QTL in C. elegans and four QTL in C. briggsae associated with divergent responses to these anthelmintics. Many of these QTL are conserved across benzimidazole derivatives, but others show drug and dose specificity. We used near-isogenic lines to recapitulate and narrow the C. elegans albendazole QTL of largest effect and identified candidate variants correlated with the resistance phenotype. These QTL do not overlap with known benzimidazole target resistance genes from parasitic nematodes and present specific new leads for the discovery of novel mechanisms of nematode benzimidazole resistance. Analyses of orthologous genes reveal conservation of candidate benzimidazole resistance genes in medically important parasitic nematodes. These data provide a basis for extending these approaches to other anthelmintic drug classes and a pathway towards validating new markers for anthelmintic resistance that can be deployed to improve parasite disease control.

Fig 1. Discovery of benzimidazole response QTL in C. elegans.

(A) Results of C. elegans linkage mapping experiments are shown for the five drug-dose conditions tested. QTL peak markers (circles) and confidence intervals (lines) are depicted. Fill color corresponds to the QTL significance (LOD) score. Overlapping QTL for a given condition are represented by the trait with the highest significance score. In total, 15 non-overlapping QTL were identified across these conditions. (B) Linkage mapping plots for the QTL of highest significance for each drug condition. The plots show genomic position on the x-axis and significance (maximum LOD score across mapping iterations) on the y-axis. QTL confidence intervals (e.g., chromosome IV: 15.47–15.91 Mb for 12.5 μM albendazole) are shaded in gray and the red dashed line represents the genome-wide significance threshold for the first mapping iteration (LOD score for 12.5 μM albendazole QTL peak marker = 19.93; LOD threshold = 2.99). (C) Tukey box plots showing the phenotypic split of RIAILs that have either the N2 or CB4856 genotype at the peak QTL marker position(s).

Fig 2. Narrowing of the albendazole QTL interval using near-isogenic lines.

(A) NIL genotypes and corresponding Tukey box plots of NIL phenotypes are shown with respect to their parental strain. NILs constructed via introgression of N2 into the CB4586 background can be classified into albendazole-sensitive (LJ1206 and LJ1207) and resistant groups (LJ1204, LJ1209, LJ1210, and LJ1212). This phenotypic separation allows for the tentative narrowing of the QTL interval (chromosome IV: 15.47–15.91 Mb) to a much smaller region (chromosome IV: 15.57–15.65 Mb) that contains one or more causative variants. Solid and dashed-vertical lines are used to mark the original QTL interval and the NIL-narrowed interval, respectively. Statistically significant differences among NILs in the resistant group suggest that additional variants outside of this narrowed interval contribute to phenotypic variance through epistatic interactions. Comparison of the N2 parental strain and LJ1203 are consistent with at least one causal variant contained within the narrowed interval. ANOVA with a post hoc Tukey HSD test was used to compare NILs used in these assays (p-values reported in S9 Table). (B) Variants distinguishing N2 and CB4856 within the NIL-narrowed interval are highlighted with respect to gene annotations. Candidate albendazole resistance variants include nonsynonymous and splice-donor variants in protein-coding genes (blue circles). No variants occur within genes encoding many non-coding RNA biotypes (tRNAs, miRNAs, and snoRNAs), though at least 276 variants occur within annotated piRNAs within the narrowed interval (S3 Table, not plotted).

Fig 3. Discovery of benzimidazole response QTL in C. briggsae.

(A) Results of C. briggsae linkage mapping experiments are shown for three tested drugs as grouped by trait category. QTL peak markers (circles) and confidence intervals (lines) are depicted. Fill color corresponds to the QTL LOD score. In total, four non-overlapping QTL were identified across conditions. (B) Linkage mapping plots are shown for the QTL of highest significance for each drug, including a major QTL associated with the effect of fenbendazole on brood size, and to a lesser extent, length and pumping traits. Plots show genomic position along the x-axis and significance (maximum LOD score across mapping iterations) along the y-axis. The QTL confidence intervals (chromosome IV: 2.56–3.45 Mb for 30 μM fenbendazole) are shaded in gray and the red dashed line represents the genome-wide correction significance threshold for the first mapping iteration (LOD score for fenbendazole QTL peak marker = 9.97; LOD threshold = 2.75). The fenbendazole QTL explains 18% of trait variation (effect size = 0.70). (C) Tukey box plots showing the phenotypic split of RIAILs that have either the AF16 or HK104 genotype at the peak QTL marker positions. AF16 animals are significant more resistant to fenbendazole exposure than HK104 animals, exhibiting much larger brood sizes in the presence of drug.

Fig 4

(A) Synteny between major-effect C. elegans albendazole QTL and C. briggsae fenbendazole QTL on chromosome IV. Circos [55] plot showing synteny of orthologous gene pairs on chromosome IV. The outer track shows linkage plots for the C. elegans albendazole and C. briggsae fenbendazole mappings (red line: LOD scores; blue line: QTL confidence intervals; black line: significance thresholds). Both QTL intervals fall within major piRNA clusters in both species (piRNA cluster beginning and end sites shaded in blue). Inner green links represent all orthologs that occur outside of the QTL intervals, yellow links represent protein-coding orthologs with at least one member inside either QTL interval but with no detected variants (58 pairs), and black links represent orthologs with at least one member inside either QTL interval and with a detected variant (38 pairs). No orthologs were identified across species that both contained variants and fell within their respective QTL confidence intervals. This result suggests that different genes and mechanisms account for the drug effects mapped to these loci. (B) Conservation of candidate C. elegans benzimidazole resistance genes in the parasitic nematodes S. ratti and B. malayi. Venn diagram showing that a substantial proportion of the QTL-contained protein-coding genes with predicted functional variants (5,434 total) are conserved in representative clade IV and clade III parasites. Approximately 40% of the candidate resistance genes have identifiable orthologs in the clade III human filarial parasite Brugia malayi (2,163 of 5,434) and the clade IV model gastro-intestinal parasite Strongyloides ratti (2,149 of 5,434), and one-third are conserved across all three species (1,685 of 5,434).