(C) PLOS One
This story was originally published by PLOS One and is unaltered.
. . . . . . . . . .
Quantifying the fitness effects of resistance alleles with and without anthelmintic selection pressure using Caenorhabditis elegans [1]
['Amanda O. Shaver', 'Molecular Biosciences', 'Northwestern University', 'Evanston', 'Illinois', 'United States Of America', 'Dept. Of Biology', 'Johns Hopkins University', 'Baltimore', 'Maryland']
Date: 2024-05
Albendazole (a benzimidazole) and ivermectin (a macrocyclic lactone) are the two most commonly co-administered anthelmintic drugs in mass-drug administration programs worldwide. Despite emerging resistance, we do not fully understand the mechanisms of resistance to these drugs nor the consequences of delivering them in combination. Albendazole resistance has primarily been attributed to variation in the drug target, a beta-tubulin gene. Ivermectin targets glutamate-gated chloride channels (GluCls), but it is unknown whether GluCl genes are involved in ivermectin resistance in nature. Using Caenorhabditis elegans, we defined the fitness costs associated with loss of the drug target genes singly or in combinations of the genes that encode GluCl subunits. We quantified the loss-of-function effects on three traits: (i) multi-generational competitive fitness, (ii) fecundity, and (iii) development. In competitive fitness and development assays, we found that a deletion of the beta-tubulin gene ben-1 conferred albendazole resistance, but ivermectin resistance required the loss of two GluCl genes (avr-14 and avr-15). The fecundity assays revealed that loss of ben-1 did not provide any fitness benefit in albendazole conditions and that no GluCl deletion mutants were resistant to ivermectin. Next, we searched for evidence of multi-drug resistance across the three traits. Loss of ben-1 did not confer resistance to ivermectin, nor did loss of any single GluCl subunit or combination confer resistance to albendazole. Finally, we assessed the development of 124 C. elegans wild strains across six benzimidazoles and seven macrocyclic lactones to identify evidence of multi-drug resistance between the two drug classes and found a strong phenotypic correlation within a drug class but not across drug classes. Because each gene affects various aspects of nematode physiology, these results suggest that it is necessary to assess multiple fitness traits to evaluate how each gene contributes to anthelmintic resistance.
Control of parasitic nematodes often depends on mass-drug administration (MDA) programs, where combinations of anthelmintics are distributed to at-risk populations. Two commonly co-administered anthelmintic drugs in MDA programs are albendazole and ivermectin, and resistance to both drugs has emerged. Although the mechanism of resistance (MoR) to albendazole has been primarily attributed to variation in a beta-tubulin gene, the MoR to ivermectin remains unknown. Ivermectin acts through pentameric glutamate-gated chloride channels (GluCls). However, it is unclear whether genes that encode GluCls are involved in ivermectin resistance in parasitic nematodes. Using Caenorhabditis elegans, we quantified the fitness costs associated with deletions of the beta-tubulin gene ben-1 and three genes encoding GluCl subunits avr-14, avr-15, and glc-1 on three traits: (i) multi-generational competitive fitness, (ii) fecundity, and (iii) development. We found different anthelmintic responses across strains and traits but no evidence of multi-drug resistance. Our results suggest that multiple traits should be considered to understand resistance comprehensively and that the determination of whether a gene plays a role in anthelmintic resistance depends on the trait measured. Understanding the quantitative effects and fitness-associated costs of each GluCl subunit in C. elegans can help explain the costs of mutations in these subunits in parasites.
Using C. elegans, we defined the fitness costs associated with the loss of ben-1, avr-14, avr-15, and glc-1 and the loss of combinations of GluCl subunits all in a controlled genetic background on nematode resistance to albendazole and ivermectin. We focused on the three genes encoding GluCl subunits that have been most implicated in ivermectin resistance to date: avr-14, avr-15, and glc-1 [ 64 , 66 , 67 , 75 – 77 ]. We measured three fitness components: (i) multi-generational competitive fitness, (ii) fecundity, and (iii) development. First, in the competitive fitness assay, we found that loss of ben-1 conferred albendazole resistance, and loss of GluCl subunits did not confer albendazole resistance. We found that loss of avr-15 carried significant fitness consequences when not under drug selection pressure. Under constant ivermectin exposure, loss of both avr-14 and avr-15 and three GluCl subunits (avr-14, avr-15, and glc-1) caused strong selective advantages compared to the wild-type strain. Second, in the fecundity assays, we found that loss of ben-1 did not confer any advantage in the presence of albendazole and that all strains with an avr-15 deletion had reduced fecundity in all conditions compared to the wild-type strain. Third, in our assessment of development, we found that loss of ben-1 conferred resistance to albendazole, and loss of both avr-14 and avr-15 or three GluCl subunits (avr-14, avr-15, and glc-1) conferred ivermectin resistance. Fourth, we sought to identify any evidence of cross-resistance between albendazole and ivermectin by comparing the fitness costs of each deletion mutation at ben-1, avr-14, avr-15, and glc-1 in both drugs. Across the three fitness traits we assessed, we found that the ben-1 deletion mutant strain did not confer resistance in the presence of ivermectin, nor did the GluCl deletion mutant strains display resistance in the presence of albendazole. Fifth, we assessed the development of 124 C. elegans wild strains across six BZs and seven MLs to identify evidence of cross-resistance between the two drug classes in natural populations. We found a strong correlation with phenotype within a drug class but not across drug classes, which indicates that phenotypic responses to the two drug classes are distinct, likely because they target different aspects of nematode development. Here, we present a comprehensive study that assessed the quantitative effects that ben-1 and GluCl mutations have on various aspects of nematode fitness in the presence of albendazole or ivermectin. These results suggest that conclusions about a gene’s involvement in anthelmintic resistance depend on the trait assessed and that multiple fitness traits must be considered to understand resistance comprehensively.
By comparing the Dent study with our current investigation, it is essential to note several key methodological differences. The Dent study performed EC 37 estimates, which encompassed a range of ivermectin concentrations from 0.6 ng/ml (0.5 nM) to 13.8 ng/ml (12 nM) for strains without Dyf mutations. Our plate-based assays were conducted at 1.5 nM ivermectin, which does fall within the range of ivermectin concentrations used in the Dent assays. However, the Dent study allowed embryos to grow to adulthood for two weeks, which would allow developmentally delayed animals to reach maturity as C. elegans completes its life cycle in four days. Moreover, assessing multiple traits to evaluate ivermectin resistance is important because GluCls are widely expressed across several tissue types in the C. elegans nervous system and pharynx [ 67 , 75 ]. Survival is an aggregate trait that integrates developmental rate, fecundity, numerous behaviors, and metabolism. To adequately evaluate the MoR to ivermectin, fitness, development, and fecundity should be used to understand how GluCl subunit genes interact and play a role in resistance and, ultimately, assess how they could affect the spread of resistance alleles in parasite populations.
In C. elegans, GluCls are thought to be homopentameric or heteropentameric transmembrane complexes where six genes encode GluCl subunits: avr-14, avr-15, glc-1, glc-2, glc-3, and glc-4 [ 63 – 66 ]. Although it is established that GluCl subunits are the main targets of ivermectin in C. elegans, null mutations in avr-14, avr-15, or glc-1 individually do not cause ivermectin resistance [ 66 ]. However, mutagenesis studies of C. elegans showed that a triple GluCl mutant strain (avr-14; avr-15 glc-1) displayed greater than 4000-fold resistance and that a double GluCl mutant strain (avr-14; avr-15) had intermediate levels of resistance as compared to the wild-type strain [ 66 ]. This study demonstrated that mutations in multiple GluCl subunit genes can cause high-level ivermectin resistance. Nevertheless, it is essential to note that the mutagenesis studies in C. elegans were not performed in a controlled background and assessed only one trait, survival, as measured by placing embryos on agar plates with ivermectin and observing the number of embryos that grew to adulthood [ 64 , 66 ]. Because many genes, as shown by others [ 66 – 74 ] and outlined below, can cause ivermectin resistance, a controlled genetic background is essential. The Dent study performed screens in 10 ng/ml of ivermectin on Nematode Growth Medium (NGM) plates with animals in an avr-15 mutant background.
Ivermectin acts as a positive allosteric modulator that selectively opens inhibitory glutamate-gated chloride channels (GluCls) in the membranes of pharyngeal muscles, motor neurons, female reproductive tracts, and the excretory/secretory pores [ 59 – 62 ]. However, the relationship between GluCls and the MoR of ivermectin is poorly understood. To date, mutations in GluCl subunits have been infrequently detected and associated with ivermectin-resistant parasitic nematodes. Our ability to find GluCl mutations in nature can be complicated by the variation in GluCl gene family number across species, interactions between genes outside the GluCl family, poor nematode genome quality leading to false negatives, and the locations of GluCl tissue expression (see Discussion ).
Both the laboratory-adapted strain, N2, and C. elegans wild strains have facilitated the identification and characterization of the beta-tubulin gene ben-1 as the primary target of albendazole and other BZs [ 34 , 35 , 45 , 46 , 49 ]. Loss-of-function mutations in ben-1 have been identified in C. elegans strains resistant to BZs [ 50 ]. Furthermore, resistance alleles corresponding to point mutations in ben-1 homologs in parasitic nematode populations continue to be identified [ 35 , 49 , 51 , 52 ]. Notably, the redundancy among the six beta-tubulin genes in C. elegans allows strains with a non-functional ben-1 gene to develop normally [ 50 ]. To date, beta-tubulins have been the most well characterized anthelmintic target across nematodes [ 50 , 53 – 58 ].
Two of the most commonly used anthelmintic drugs in MDA programs are albendazole, a benzimidazole (BZ), and ivermectin, a macrocyclic lactone (ML) [ 23 – 26 ], where research suggests that these two drugs can be co-administered as a chemotherapeutic treatment for intestinal helminths and lymphatic filariasis [ 27 – 30 ]. Because albendazole and ivermectin are two drugs on the World Health Organization (WHO) Model Lists of Essential Medicines [ 27 , 30 – 32 ], it is critical to identify the MoR for both drugs to inform appropriate administration and slow the development of resistance. However, anthelmintic resistance in parasites can be difficult to disentangle because of multiple factors, including a lack of access to relevant life stages and in vitro culture systems, dependence on vertebrate hosts, and a limited molecular toolkit (e.g., gene knockouts and induced mutations cannot be used to study genes associated with resistance in parasites) [ 33 ]. With its ease of growth, outstanding genetic tractability, and molecular toolkits, the free-living nematode Caenorhabditis elegans has contributed to the identification and characterization of the Mode of Action (MoA) and MoR of all major anthelmintic drug classes [ 33 – 40 ]. Additionally, wild C. elegans strains from the Caenorhabditis Natural Diversity Resource (CaeNDR) [ 41 ] have been used to explore anthelmintic resistance in natural populations and to uncover novel MoR [ 36 , 41 – 48 ].
Simultaneous treatment with two or more drugs (i.e., co-administration) from different anthelmintic classes is one method proposed to slow the development of resistance [ 12 – 15 ]. Anthelmintic rotation, another suggested method to slow resistance, uses the periodic switching of drug classes to alleviate selection pressures on one drug class and prolong drug lifespan and efficacy [ 16 – 18 ]. Despite the implementation of both strategies, co-administration and anthelmintic rotation have failed to control resistance and pose the risk of multi-drug resistance [ 9 , 19 , 20 ], particularly if a shared mechanism confers resistance to both drugs (e.g., drug export) [ 21 ], a phenomenon known as cross-resistance [ 22 ]. Although empirical evidence for cross-resistance is lacking, its potential occurrence could increase the development of resistance to both drugs. To date, various accounts of multi-drug resistance have been reported in veterinary medicine [ 21 – 24 ]. Therefore, it is critical to define and address the mechanisms of resistance (MoR) for each drug in a treatment plan to ensure that drug efficacy can be reached and to prevent multi-drug resistance.
Parasitic nematodes are some of the most abundant, diverse, and destructive parasites of humans that cause significant socio-economic and health impacts, including the collective loss of over eight million disability-adjusted life years (DALYs) [ 1 – 3 ]. Control of these parasites often depends on mass-drug administration (MDA) programs, where anthelmintics are distributed in combinations to at-risk populations. However, anthelmintic resistance has emerged in veterinary helminths, with reports of reduced drug efficacy against nematodes in humans, which threatens our ability to control parasitic nematode infections [ 4 – 8 ]. In veterinary medicine, overuse of anthelmintics has placed strong selective pressures on parasites, which has led to the evolution of resistance to all major drug classes [ 9 – 11 ] and highlights the potential for a similar pattern of anthelmintic resistance to spread throughout human parasitic nematode populations.
Spearman-rank correlations between 124 wild isolates exposed to BZs and MLs. Spearman-rank correlations and significance testing were performed between 124 wild isolates exposed to six BZs and seven MLs. The dendrograms were constructed using Euclidean distance and complete linkage metrics and then split into their two largest clusters to show the relationships of similarities between the 124 wild strains exposed to the two major anthelmintic classes. A correlation of 1 signifies the strongest phenotypic correlation (i.e., identical median animal lengths) and a correlation of -0.2 signifies the weakest phenotypic correlation (i.e., dissimilar median animal lengths). Significant correlations (p < 0.05) were recorded ( S3 Table ).
Our fitness assays showed how ben-1, avr-14, avr-15, and glc-1 respond under drug pressure, but we know that these genes do not account for all of the albendazole or ivermectin resistance found across the C. elegans species [ 49 , 86 ]. Therefore, we performed an HTA (see Methods ) to assess the nematode development of 124 wild strains in the presence of six BZs and seven MLs, which included albendazole and ivermectin. We used a Spearman’s Rank correlation test to identify any evidence of multi-drug resistance among the BZs and MLs. We find much stronger phenotypic correlations of responses within the same drug class than we do between drug classes ( Fig 5 ), and no significant correlations exist across the two drug classes ( S3 Table ), which suggests that we did not detect evidence of multi-drug resistance between the two drug classes.
Because albendazole and ivermectin can be distributed together to at-risk populations to treat lymphatic filariasis and has been proposed as a strategy to treat soil-transmitted helminth infections [ 28 – 30 , 84 , 85 ], it is critical to ensure that the two drugs do not have the same MoR to avoid the possibility of cross-resistance. Because we compared the response of the ben-1 deletion mutant strain in ivermectin and all GluCl deletion mutant strains in albendazole for all fitness assays, we obtained a comprehensive picture of how these genes interact in the presence of a drug that is not designed to affect their given target. In the competitive fitness and HTAs, the loss of ben-1 did not cause ivermectin resistance. Additionally, none of the GluCl deletion mutant strains conferred resistance to albendazole across the competitive fitness and HTAs, as compared to a loss of ben-1. However, it is important to note that a loss of ben-1 did confer a slight advantage compared to the N2 strain in the ivermectin fecundity assay ( Fig 3C ). All GluCl deletion mutant strains, except the glc-1 deletion mutant strain, conferred a slight advantage compared to the N2 strain in the albendazole HTA ( Fig 4A ). However, it is essential to note that these slight advantages across conditions could be caused by genes targeted by both drugs expressed in the same tissues or cells. Importantly, the competitive fitness assays did not display any evidence of cross-resistance ( Fig 2 ).
High-throughput assays for each mutant strain in the presence of albendazole and ivermectin. The regressed median animal length values for populations of nematodes growth in either (A) 30 μM albendazole or (B) 250 nM ivermectin are shown on the y-axis. Each point represents the normalized median animal length value of a well containing approximately 5–30 animals. Data are shown as Tukey box plots with the median as a solid horizontal line, and the top and bottom of the box representing the 75th and 25th quartiles, respectively. The top whisker is extended to the maximum point that is within the 1.5 interquartile range from the 75th quartile. The bottom whisker is extended to the minimum point that is within the 1.5 interquartile range from the 25th quartile. Significant differences between the wild-type strain and all other deletions are shown as asterisks above the data from each strain (p > 0.05 = ns, p < 0.001 = ***, p < 0.0001 = ****, Tukey HSD).
As previously reported, the N2 strain was developmentally delayed in albendazole, where animals were shorter than in control conditions, demonstrating sensitivity to albendazole [ 66 ]. By contrast, the loss of ben-1 caused albendazole resistance, as demonstrated by longer animal length than observed in the N2 strain ( Fig 4A ). Although each GluCl deletion mutant strain had significantly longer animal lengths than the N2 strain, none of the seven GluCl deletion mutant strains conferred albendazole resistance as observed in the ben-1 deletion mutant ( Fig 4A ). In ivermectin, the N2 strain had the greatest delay in development. Although, a loss of avr-14 or avr-15 alone or a loss of avr-14 and glc-1 or avr-15 and glc-1 in combination had significantly longer median animal lengths than the N2 strain in ivermectin ( Fig 4B ). However, the loss of avr-14 and avr-15 or the loss of avr-14, avr-15, and glc-1 caused quantitative ivermectin resistance as compared to the N2 strain. This ivermectin resistance confirms previous findings [ 66 ]. However, we did not see a significant difference in median animal length between the double GluCl mutant avr-14 and avr-15 or the triple GluCl mutant, as reported previously [ 49 , 75 ]. A higher concentration of ivermectin (500 nM) was also measured and confirmed the same patterns described above ( S7 Fig ).
We then performed high-throughput assays (HTAs) to measure nematode length, a proxy for development, in strains with a loss of ben-1 or loss of GluCl subunit genes ( S2 Table ) in response to drug treatment. The assay included 72 replicates per strain with 5–30 animals per replicate in each drug or control condition. The reported nematode length of each strain is the delta between animal lengths in control and drug conditions to obtain normalized animal length and assess drug effects. Longer median animal length (i.e., larger animals) than the N2 strain corresponds to increased resistance to the tested drugs, and shorter median animal length (i.e., smaller animals) than the N2 strain corresponds to increased sensitivity to the tested drugs. Strains varied in length after growth for 48 hours in control conditions, but the loss of avr-14 and avr-15 caused the most significant delays in development ( S6 Fig ). Despite substantial variation among strains in control (DMSO) conditions, animal measurements were categorized as the L4 larval stage by our custom CellProfiler worm models [ 82 , 83 ], indicating all strains underwent normal development in control conditions.
By evaluating fecundity under drug conditions, we can uncover the fitness effects caused by mutations in ben-1 or the GluCl subunit genes and determine how these mutations under drug pressure could affect the spread of potential resistance alleles in a population. In the presence of albendazole, the loss of ben-1 did not cause a significant effect on lifetime or daily fecundity as compared to the N2 strain, which suggests that ben-1 does not confer albendazole sensitivity by alteration of nematode fecundity ( Figs 3B and S4 ). However, in the presence of ivermectin, the loss of ben-1 had a significant increase in lifetime fecundity compared to the control strain, exhibiting the same pattern observed in control conditions ( Figs 3C and S5 ). The loss of glc-1 and the loss of both avr-14 and glc-1 did not cause significant differences in lifetime or daily fecundity compared to the N2 strain in albendazole or ivermectin, which suggests that glc-1 alone or in combination with avr-14 are necessary for normal fecundity production under drug pressure. By contrast, a loss of avr-14 alone caused significantly reduced fecundity compared to the control strain in ivermectin, which indicates that avr-14 is necessary for fecundity in ivermectin. Additionally, the loss of avr-15 alone, the loss of both avr-14 and avr-15 or avr-15 and glc-1, or the loss of all three GluCl subunits caused significantly reduced lifetime fecundity compared to the control strain, a trend observed across all conditions, which suggests that avr-14 and avr-15 alone or in combination are necessary for normal fecundity. Although the loss of avr-15 in the single, double, or triple mutant strains caused a reduction in fecundity across conditions, the daily fecundity patterns varied across the three conditions ( S3 – S5 Figs ). In all three conditions, strains with a loss of avr-15 had a reduction in daily fecundity between days two and three compared to the N2 strain. In DMSO and ivermectin, strains with a loss of avr-15 had an increase in daily fecundity between days five and seven at the end of the assay ( S4 Fig ). Because strains with a loss of both avr-14 and avr-15 and a loss of all three GluCl subunits have significantly reduced fecundity across all conditions, it would be unlikely for animals in nature to acquire loss-of-function mutations in these genes that cause detrimental fitness consequences.
Fecundity directly impacts population growth rate and survival. Measuring fecundity provides insights into reproductive success, which is fundamental to the assessment of animal fitness and resilience when exposed to environmental changes. Here, the fecundity assays enabled us to focus on offspring production both in control conditions and under constant drug selection pressure. By assessing fecundity in control (DMSO) conditions, we can discern the fitness effects caused by loss of ben-1 or by the loss of GluCl subunit genes when not under drug selection pressure. The loss of ben-1 caused increased fecundity ( Fig 3A ), which suggests that ben-1 limits fecundity in control conditions. The loss of avr-15 alone and loss of both avr-15 and glc-1 did not affect lifetime fecundity as compared to the N2 strain. However, the loss of avr-15 in combination with avr-14 or in combination with avr-14 and glc-1 caused significantly reduced lifetime fecundities as compared to the N2 strain, which suggests that in combination avr-14 and avr-15 or avr-14, avr-15, and glc-1 are necessary for normal fecundity in control conditions ( Fig 3 ). By contrast, the loss of avr-14 or glc-1 and the loss of both genes caused significantly increased lifetime fecundities as compared to the N2 strain, which suggests that avr-14 and glc-1 limit fecundity in control conditions.
Variation in lifetime fecundity of beta-tubulin and GluCl mutants in the presence of DMSO, albendazole, or ivermectin. Bar plots for lifetime fecundity, y-axis, for each deletion strain on the x-axis in (A) DMSO (1%), (B) albendazole (1.25 μM), and (C) ivermectin (1.5 nM) are shown. Error bars show the standard deviation of lifetime fecundity among 7–10 replicates. The laboratory reference strain, N2, is colored orange. Other strains are colored by genotype. Comparisons of lifetime fecundity between the laboratory reference strain, N2, and all deletion strains are shown. Statistical significance was calculated using Tukey HSD. Significance of each comparison is shown above each comparison pair (p > 0.05 = ns, p < 0.05 = *, p < 0.01 = **, p < 0.001 = ***, p < 0.0001 = ****, Tukey HSD).
To dissect the genetic basis of anthelmintic resistance, we must identify the roles of ben-1 and the GluCl subunit genes in nematode fecundity and their potential influence on the spread and persistence of resistance alleles in a population. As a measure of relative fitness, fecundity refers to the number of offspring produced by an organism [ 78 , 79 ]. To compare the effects on fecundity caused by the loss of ben-1 and the GluCl subunit genes, we measured lifetime fecundity and daily fecundity of the nine C. elegans strains (see Methods ). Single L4 larval stage hermaphrodites were placed on NGMA plates under control (DMSO), albendazole, or ivermectin conditions [ 81 ]. Hermaphrodites were transferred every 24 hours for five days and maintained under standard laboratory conditions. After five days, hermaphrodites were transferred for the final time to a new NGMA plate for 48 hours. We manually counted the offspring from images of assay plates from single hermaphrodites. The results showed considerable differences in lifetime fecundity among the nine strains across the three conditions ( Fig 3 ).
In drug conditions, we can assess whether loss of any of these genes causes resistance. In albendazole conditions, the loss of ben-1 caused a strong fitness advantage over the control strain and swept to fixation in the population by the third generation ( Fig 2C and 2D ). Notably, the loss of one, two, or all three GluCl genes caused a reduction of fitness in albendazole conditions, and each strain displayed a similar competitive fitness response in albendazole as observed in control conditions ( Fig 2C and 2D ). In ivermectin conditions, the loss of ben-1 or glc-1 caused no significant differences in competitive fitness as compared to the control strain. The loss of avr-14 or avr-15, the loss of both avr-14 and glc-1, or the loss of both avr-15 and glc-1 caused significantly reduced competitive fitness in ivermectin compared to the control strain ( Fig 2E and 2F ). By contrast, the loss of both avr-14 and avr-15 and the loss of all three GluCl genes (avr-14, avr-15, and glc-1) caused strains to sweep to fixation in the presence of ivermectin by the third generation, indicating that they had significantly improved fitness as compared to the control strain ( Fig 2E ). Results from these competitive fitness assays suggest that ivermectin resistance is only observed when both avr-14 and avr-15 functions are lost. The fitness disadvantages of losing avr-14 or avr-15 alone outweigh any ivermectin resistance that could be present. The loss of glc-1 in addition to loss of both avr-14 and avr-15 caused the triple mutant strain to sweep to fixation faster than the avr-14; avr-15 double mutant strain. However, these two strains did not significantly differ in competitive fitness.
Competitive fitness assays across seven generations in DMSO, albendazole, and ivermectin. A barcoded N2 wild-type strain, PTM229, was competed with strains that have deletions in either one, two, or three genes that encode GluCl channels or the beta-tubulin gene ben-1 in (A) DMSO (1%), (C) albendazole (1.25 μM), and (E) ivermectin (1.5 nM). Generation is shown on the x-axis, and the relative allele frequencies of the nine strains with genome-edited alleles and N2 are shown on the y-axis. The log 2 -transformed competitive fitness of each allele is plotted in (B) DMSO (1%), (D) albendazole (1.25 μM), and (F) ivermectin (1.5 nM). The gene tested is shown on the x-axis, and the competitive fitness is shown on the y-axis. Each point represents a biological replicate of that competition experiment. Data are shown as Tukey box plots with the median as a solid horizontal line, and the top and bottom of the box representing the 75th and 25th quartiles, respectively. The top whisker is extended to the maximum point that is within the 1.5 interquartile range from the 75th quartile. The bottom whisker is extended to the minimum point that is within the 1.5 interquartile range from the 25th quartile. Significant differences between the wild-type N2 strain and all the other alleles are shown as asterisks above the data from each strain (p > 0.05 = ns, p < 0.001 = ***, p < 0.0001 = ****, Tukey HSD). Because two DMSO competitive fitness assays were performed, results from both DMSO assays are reported ( S2 Fig ).
In control conditions, the wild-type strain, N2, showed no differences in competitive fitness compared to the barcoded wild-type control strain, PTM229, as expected. The loss of ben-1, avr-14, or glc-1 in single deletion mutants, along with the loss of both avr-14 and glc-1 in the double mutant strain, did not have significant differences in competitive fitness as compared to the control strain, which suggests that a deletion in these genes, in control conditions, did not cause fitness consequences ( Figs 2A, 2B and S2 ). Notably, all strains with a loss of avr-15, whether it be a single, double, or triple GluCl mutant, were vastly unfit and ceased to exist in the population by the third generation ( Fig 2A ). These results suggest that, when not under drug selection pressure, loss of avr-15 is severely detrimental to animal fitness. Hence, we hypothesize that GluCl loss-of-function alleles are not anticipated to occur within parasite populations.
The competitive fitness assays enabled us to focus on two key traits critical to nematode fitness: time to reproduction and reproductive rate. These assays allow us to observe small effects on nematode fitness over multiple generations. If an allele confers a deleterious fitness effect compared to the wild-type allele, then that strain will decrease in frequency over the generations. Conversely, if an allele confers a beneficial effect compared to the wild-type allele, then that strain will increase in frequency over the generations. Finally, if an allele has no difference in effect when compared to the wild-type allele, then the two strains will be found at approximately equal frequencies throughout the competitive fitness assay.
CRISPR-Cas9 genome editing was performed to generate four deletion strains that each cause loss of function. Each strain contains a single deletion in either the beta-tubulin gene, ben-1, or the GluCl subunit genes avr-14, avr-15, or glc-1 ( Fig 1 , S1 Table ). Next, to comprehensively assess the role that avr-14, avr-15, and glc-1 play in ivermectin resistance, we created double mutants of each combination of GluCl deletion alleles and a triple mutant (avr-14; avr-15 glc-1) by crossing the single deletion strains (see Methods ). We performed competitive fitness assays to determine the selective advantages or disadvantages of alleles in control or drug treatment conditions. Fitness involves the ability of an organism, or population, to survive and reproduce in its environment [ 78 , 79 ]. In these assays, a query strain was competed against the barcoded wild-type strain PTM229 ( S1 Table ), which contains a synonymous change in the dpy-10 locus in the N2 background and does not cause any fitness effects compared to the normal laboratory N2 strain [ 80 ] in the presence of dimethyl sulfoxide (DMSO), albendazole, or ivermectin ( S1 Fig ) . Because all query strains contain the wild-type dpy-10 locus, allele frequencies of dpy-10 between PTM229 and each query strain were measured for each generation to quantify relative fitness.
Discussion
Here, we assessed multiple fitness traits to understand how ben-1 and three genes that encode GluCl subunits contribute to albendazole and ivermectin resistance. Additionally, the quantitative assessment of avr-14, avr-15, and glc-1 in ivermectin response is critical to understand how GluCl subunits affect fitness in nematode populations. Because C. elegans shares the major characteristics of the parasitic nematode body plan, such as the cuticle and organization of the nervous system, along with a conserved neuromuscular system and neurotransmitters [33,87], the traits that we assessed can help us better understand how resistance alleles could spread in parasitic nematode populations.
Our competitive fitness and development assay results confirm previous findings, which showed that a loss of ben-1 confers albendazole resistance and that a loss of both avr-14 and avr-15 are necessary to confer ivermectin resistance [75,77,78]. Animals with the loss of both avr-14 and avr-15 or the loss of three GluCl subunits (avr-14, avr-15, and glc-1) did not significantly differ in ivermectin response as previously reported in mutagenesis studies [82,84,85]. It is also critical to note that strains with a loss of both avr-14 and avr-15 or a loss of three GluCl subunits (avr-14, avr-15, and glc-1) have significant fitness consequences when not under ivermectin exposure, rendering it unlikely that loss-of-function mutations will occur in these genes and confer ivermectin resistance in nature.
On the contrary, animals under strong selection because of prolonged ivermectin exposure might select resistance alleles in GluCl genes, which would allow mutations in GluCl subunits to spread throughout populations. However, parasitic nematodes must remain under constant ivermectin exposure at higher than therapeutic doses to overcome the severe fitness consequences that mutations in GluCl subunits carry for selection to remain significant. Because human drug treatment programs apply ivermectin once to three times annually [88], it is unlikely that ivermectin persists in humans between treatments at concentrations that would cause constant drug selection pressure and subsequent selection of resistance alleles in parasitic nematodes. Although veterinary medicine prescribes ivermectin more frequently, ivermectin again is unlikely to remain at concentrations high enough to cause selective pressures, especially during winter months when treatment subsides [89,90].
Prior to this study, minimal research had been performed to assess the quantitative contributions of GluCl subunits on different traits involved in nematode fitness. To identify why GluCls in parasites lack mutations in ivermectin-resistant populations [91] and why we have yet to validate ivermectin-resistant genes, we tested the contributions of three genes that encode GluCl subunits on C. elegans fitness. We found that the competitive fitness and development assays produced similar results, but the fecundity assays deviated from the results of these two assays. We have two hypotheses to explain this discrepancy between assays. First, the competitive fitness assay focuses on two traits involved in nematode fitness: development (i.e., time to reproduction) and reproductive rate. Because competitive fitness encompasses development, we can expect results from the competitive fitness assay and development assay to be more comparable than the results of the fecundity assay. In the fecundity assay, we measured the daily and total fecundity of individual animals over seven days and did not focus on reproductive rate. Therefore, it is difficult to compare the daily fecundity of animals for the duration of one generation to the competitive fitness assay, where animals have offspring over multiple generations. Second, it is possible that mutations in ben-1 or the GluCl subunit genes do not affect fecundity to the same extent that they affect development, rendering the fecundity results less clear to interpret compared to the competitive fitness and development assays.
We found that a loss of avr-14 alone or in combination with glc-1 has significant fitness consequences in ivermectin but not when in albendazole or control conditions. Overall, a loss of avr-15 has profoundly detrimental effects compared to the control strain across all conditions. An animal with a loss of avr-15 in combination with avr-14 must remain under constant ivermectin pressure to exhibit any fitness benefit compared to the control strain. Finally, a loss of glc-1 does not affect fitness across conditions, indicating perhaps why mutations of this gene are found in C. elegans wild populations [86,92,93]. Our investigation did not reveal any indication that glc-1 is responsible for conferring resistance to ivermectin, which contradicts earlier findings [92,93]. However, we focused on the effects of muti-generational competitive fitness, fecundity, and development when nematodes are exposed to ivermectin. By contrast, previous studies measured nematode response against different macrocyclic lactones (i.e., abamectin), focused on different traits such as body bends, paralysis, and gene expression, and used different genetic approaches to assess resistance [92,93]. The discrepancy between our findings and earlier research underscores the necessity of evaluating multiple traits, as the response to anthelmintic treatment can vary depending on the trait measured.
Several reasons might explain the infrequent detection of GluCl subunits mutated in parasitic nematode isolates that have ivermectin resistance, including variation in GluCl gene family number, interactions between genes outside the GluCl family, poor nematode genome quality, and the locations of GluCl tissue expression. First, we do not fully understand the composition of GluCls. The GluCls are members of the pentameric ligand-gated ion channel family and, similar to other members of this family, the functional channels formed in vivo could be homomeric or heteromeric [89]. More research needs to be done to elucidate the subunit composition of GluCls across parasitic nematode species. In addition, no auxiliary proteins, analogous to the genes that influence the trafficking and assembly of nematode nicotinic acetylcholine receptors, have been reported for the GluCls, and we do not know if any such proteins exist [91]. Furthermore, although avr-14, avr-15, glc-1, glc-2, glc-3, and glc-4 are predicted to be conserved widely across nematode species [94], avr-14 is currently the only GluCl gene found to be highly conserved throughout Nematoda [26,92–94]. A polymorphism in an avr-14 ortholog has been identified in Cooperia oncophora [57]. Although polymorphisms in several genes (Hco-glc-3 and Hco-glc-5) have been identified in resistant isolates of Haemonchus contortus [26,95–97], recent genome-wide analyses have not identified mutations in GluCl subunit associated in ivermectin resistant in H. contortus [98]. Importantly, these results are correlative and do not show a causal connection between ivermectin resistance and GluCl genes. Additionally, research in both C. elegans and parasitic nematodes has led to the suggestion that ivermectin resistance might be polygenic [94], so combinations of genes must also be considered.
Second, the absence of mutations in GluCl genes in ivermectin-resistant parasites has led researchers to search for additional genes outside this gene family involved in the MoR of ivermectin [91]. It is unknown why, despite GluCls being understood as the MoA and confirmation of association with resistance to ivermectin in C. elegans, the GluCl subunits have not been widely associated with resistance in parasite populations. As illustrated here, the fitness disadvantages of losing avr-14 or avr-15 alone or in combination outweigh any ivermectin resistance. Additionally, the loss of glc-1 alone or in combination with avr-14 or avr-15 had no discernible impact on fitness across conditions but also did not confer ivermectin resistance. Hence, unlike alleles that confer albendazole resistance, alleles associated with ivermectin resistance exhibit pronounced fitness consequences. Consequently, susceptible alleles would be more likely to endure within parasite populations because they confer fitness advantages in the absence of ivermectin. Our findings provide insights into why GluCls have not been clearly associated with resistance in parasites.
To date, ivermectin resistance across Nematoda appears to involve various genes and mechanisms [66,68–70]. Additional genes outside of those that encode for GluCl subunits are implicated in the MoR of ivermectin, which includes Dyf genes and genes involved in neuronal development and function, such as unc-7, unc-9, unc-38, and unc-63 [98–101]. Genes involved in ivermectin metabolism, such as ATP-binding cassette (ABC) transporters, cytochrome P450 enzymes, GABA receptors, P-glycoproteins, and other signaling proteins, have been implicated in ivermectin resistance though much more research needs to be done to determine their role in the MoR of ivermectin [67,72–74]. Ivermectin also affects some nicotinic receptors and acts as a positive allosteric modulator of the 57 neuronal nicotinic acetylcholine receptor [99]. Overall, additional genes outside of the GluCl family could be involved in the MoR of ivermectin. However, functional studies, such as those conducted in the present study, are essential to elucidate the role each gene plays in ivermectin resistance.
Third, it is important to highlight that GluCl subunits have been well characterized in only a few nematode species, partly because of poor-quality genomes. Recent efforts have been made to generate high-quality reference parasitic nematode genomes [100–103]. WormBase Parasite [104,105] serves as the main repository for these data, which now hosts a collection of 240 genomes, representing 181 species. Recently, 864 total GluCl gene predictions were categorized across 125 species into orthologous groups, which suggests that additional GluCl subunits across Nematoda have yet to be discovered [94]. As our genome assemblies, technologies, and analytical techniques improve, so will our ability to search for and identify GluCl genes.
Fourth, although albendazole and ivermectin target different genes and have different MoR, it is conceivable that the genes targeted by both drugs can be expressed in the same tissues or cells. To date, our understanding of the MoR and tissue-specific susceptibility for most anthelmintic drugs across nematode species is not well known. Previous research has shown that ben-1 function in neurons, specifically in cholinergic neurons, underlies susceptibility to BZs [105]. However, analogous experiments are imperative to ascertain in which tissues GluCl subunit genes underlie ivermectin susceptibility. Expression data from the Complete Gene Expression Map of the C. elegans Nervous System (CeNGEN) [106] shows an overlap in ben-1 and GluCl subunit gene expression in neurons (S8 Fig). In particular, a pronounced overlap between ben-1 and avr-14 expression in cholinergic neurons is observed, which suggests albendazole and ivermectin could target the same tissues. The overlap in ben-1 and GluCl gene expression in the same neuronal cell subtypes could explain the small advantages that ben-1 conferred in ivermectin and the GluCls in albendazole in the HTA (Fig 4). The co-expression of two genes in the same neurons implies a potential functional relationship between the genes, which could collaborate to regulate specific neural functions associated with the neurotransmitter. A shared expression between genes could lead to unexpected consequences, such as changes in sensitivity to other drugs or alterations in neural processes beyond resistance to the targeted drug. Given that multiple GluCls are present in C. elegans, functional redundancy is conceivable. Because neurotransmitters play essential roles in physiological processes, including behavior, locomotion, and sensory perception, it is imperative to delineate which neurons are implicated in GluCl expression and consequently influences ivermectin susceptibility. Fully understanding the MoR of each drug class is complicated by the implication that genes associated with drug resistance overlap in expression within the same neuronal pathways. Finally, it is important to note that all the expression data discussed here have been performed in C. elegans and not parasitic nematode models, so differences among nematodes might not be captured entirely by research on this free-living nematode species.
In summary, our experiments suggest that loss of ben-1 confers albendazole resistance and that multiple mutations in GluCl genes are required to obtain ivermectin resistance. Nevertheless, our understanding of ivermectin’s MoR remains incomplete because we have not identified all the genes involved in ivermectin resistance in C. elegans nor identified the genes involved in ivermectin resistance in parasitic nematodes. To solve this problem, we need to conduct additional experiments that quantitatively assess the fitness effects of all six GluCl subunit genes singly and in combination, both in control and ivermectin conditions. Moreover, to identify in which tissues GluCl function underlies ivermectin susceptibility, transgenic strains that express each GluCl subunit genes in different tissues will determine the tissue-specific susceptibility of ivermectin. Finally, when considering the opportunity for multi-drug resistance to occur, either by cross-resistance or independent selection, our data suggest that cross-resistance is unlikely because ben-1 and the GluCl subunits do not appear to target the same MoR in C. elegans. In studies of H. contortus, multi-drug resistant isolates carry independent mutations in genes specific to a given resistance phenotype, suggesting that albendazole and ivermectin target different genes [98]. Studies across C. elegans and H. contortus suggest that animals acquire multi-drug resistance independently and not by cross-resistance.
[END]
---
[1] Url:
https://journals.plos.org/plospathogens/article?id=10.1371/journal.ppat.1012245
Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.
via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/
