1 Introduction

Parasitic nematodes cause chronic and debilitating illnesses in humans and animals, as well as economic losses in livestock and agriculture [1–3]. For example, soil-transmitted helminths, infect more than one billion people and cost roughly five million disability-adjusted life years reported in 2010 [4] and plant-parasitic nematodes affecting crops cause annual economic losses of >$80 billion globally [5]. Control strategies for parasitic nematodes in domesticated or livestock animals rely on anthelmintic drugs, which cost the European ruminant livestock industry an estimated €320 million per year [6]. The three major classes of anthelmintic drugs are benzimidazoles, tetrahydropyrimidines, and macrocyclic lactones (MLs). The MLs, such as Ivermectin (IVM) and Abamectin, are the most commonly used drug class in veterinary/human medicine and agriculture due to their high efficacy, low toxicity, and broad-spectrum nature [7]. Glutamine-gated chloride ion channels (GluCls) are the primary target of MLs in nematodes. This consequently leads to the depolarization of pharyngeal muscles and hyperpolarization of body wall muscles [8]. As a result, feeding becomes impaired, and the worm experiences flaccid muscle paralysis [9, 10]. In addition, γ-aminobutyric acid (GABA) receptors have been reported as secondary targets of IVM at concentrations [11] considered therapeutically insignificant.

However, due to the widespread use of anthelmintic drugs, extensive anthelmintic drug resistance has developed in many parasitic nematode species of livestock [12]. Anthelmintic resistance has been reported in Haemonchus contortus, Cooperia spp., and Ostertagia ostertagi infecting sheep in Europe [13]. In addition, anthelmintic-resistant parasitic nematodes in sheep, cattle and goats have been reported in a number of countries, including Brazil, Argentina, the United States, Australia, New Zealand, and South Africa (reviewed [12]). Furthermore, there are growing concerns that mass drug administration (MDA) programs are selecting for anthelmintic resistance among helminths that infect humans [14, 15]. In recent years, there have been emerging reports of IVM resistance in human parasite Onchocerca volvulus in Ghana [16] and Cameroon [17]. Evidently, understanding and identifying the genetic markers underlying resistance in parasitic nematodes is necessary.

However, identification of candidate genes or gene variants that drive anthelmintic resistance in parasitic nematodes can be challenging. This is partly due to the fact that parasitic nematodes have a host-dependent life cycle, are difficult to cultivate, maintain, manipulate, and study at scale, and often lack well-annotated genomes in full chromosomes. Until now, only a few parasitic nematodes, have relatively complete chromosomal-level genomes [18–21], allowing genetic and comparative genome analyses. In comparison, the free-living nematode Caenorhabditis elegans is biologically simple, has a rapid 3.5-day life cycle, genetically amenable, and molecular tools for manipulation are readily available [22]. In addition, the evolutionary relationship between C. elegans with parasitic nematodes [23], has led to the adoption of this free-living nematode as a model for parasitic nematodes [22, 24] because it allows for comparative studies and translatable results [25].

Studies in C. elegans have begun to uncover IVM resistance mechanisms. For example, simultaneous mutations in glc-1, avr-14, and avr-15 result in a ~4000-fold resistance to IVM in C. elegans [26]. These genes, together with glc-2, glc-3, and glc-4, code for the GluCls subunits, which are capable of forming diverse homomeric and heteromeric chloride channels [9, 10]. However, the discovery of IVM-resistant populations of the small-ruminant parasite H. contortus lacking mutations in the glc-1, avr-14, and avr-15 genes [19, 27] suggests the existence of alternative mechanisms of resistance to MLs. Several resistance mechanisms to MLs have been proposed, including changes in gene expression in GluCl and transport protein genes [28]. For example, as previously postulated, decreased expression of drug target genes may result in fewer drug receptors and thus reduced drug efficacy [29, 30]. Gene expression changes of transporter genes of the ATP-binding cassette family, particularly P-glycoproteins (Pgps), also known as efflux pumps, have been linked to anthelmintic drug resistance [31, 32]. This is because efflux pumps eliminate drugs from the cell, preventing them from reaching their target sites. Furthermore, increased expression of genes encoding drug metabolic enzymes, postulated to increase conversion of drugs into inactive metabolites, has been associated with resistance [33]. Collectively, this highlights the multigenic nature and multiple mechanisms of anthelmintic resistance. Although a previous microarray-based study [34] examined the effect of IVM on gene expression in C. elegans, the strain used was DA1316, which is highly resistant to IVM due to simultaneous mutations in three genes (glc-1, avr-14, and avr-15). As a result, the current study is premised on the lack of a previous transcriptomic study assessing the effect of IVM on gene expression in wild-type (N2) C. elegans.

In this study, we investigated gene expression in an adult C. elegans N2 strain exposed to IVM using an RNAseq transcriptomic approach in order to identify the underlying processes and genes involved in the response to IVM in a wild type strain. In addition, we compared our transcriptomic data with microarray data from the IVM-exposed C. elegans DA1316 strain [34] in order to identify genes that may be uniquely associated with IVM response in IVM-resistant strains. Additionally, we investigated the recently discovered Abamectin Quantitative Trait Loci (QTL) on chromosome V in C. elegans [35] to determine any potential associations between the identified genes and the QTL.

2 Materials and methods

3 Results

3.1 Transcriptomic variation amongst adult C. elegans worms exposed to Ivermectin

Total RNA (RIN > 8) was sequenced from C. elegans adults at maximum reproduction exposed to IVM 10⁻⁷ and 10⁻⁸ M yielding sequencing data of 21–60 million read-pairs/sample. The used concentrations of IVM treatment were selected according to [41]. However, due to the very low number of differentially expressed genes in IVM 10⁻⁹ M, only data from IVM 10⁻⁷ and 10⁻⁸ M was used in the current study. Following quality control and evaluation with SortMeRNA and Trimmomatic, 16–42 million read-pairs/sample were obtained with a mean phred score above 33, indicating high quality reads. The read-mapping rate to the custom-built PRJNA13758 C. elegans transcriptome (see Methods) and the C. elegans genome (PRJNA13758, release WS283) as a decoy, was 97% using Salmon.

According to principal component analysis of the VST read counts, worm pools (n = 300 worms per pool), both PC1 and PC2 explained the separation by treatment condition (Fig 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Variation in gene expression among worms exposed to IVM.

A PCA plot of C. elegans N2 strain based on variance stabilized read counts from worm pools (n = 300) after exposure to IVM 10⁻⁷ M, 10⁻⁸ M or control ^+DMSO. Pools are separated by both PC1 and PC2, which represent the largest variances in the data and are ellipsed by concentration/treatment.

https://doi.org/10.1371/journal.pone.0285262.g001

3.2 Differential gene expression

Following DESeq2 analysis, 615 genes were differentially expressed after IVM 10⁻⁷ M and/or 10⁻⁸ M exposure (S1 Table). Of these, 76% (468 genes) had known functional annotations including GO annotations according to WormBase (release WS283). Ninety five percent (582 genes) of DEGs were exclusively detected in 10⁻⁷ M, 1% (8 genes) in 10⁻⁸ M and 4% (25 genes) showed significant (p-value = 0.00001) overlap between the two concentrations (Fig 2 & S2 Table). About 33% (201 genes) of DEGs were altered more than 2-fold, with 85 genes upregulated and 116 genes downregulated. The top ten up- and downregulated DEGs consisted of a diverse set of genes or gene families including transmembrane transporters, and heat shock proteins (S3 Table).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Differentially expressed in IVM-exposed C. elegans N2 strain.

Venn diagrams showing the number of DEGs in C. elegans N2 strain between IVM concentrations 10⁻⁷ and 10⁻⁸ M based recommendations by Schurch et al. [54]. The number of genes differentially expressed above 2-fold are labelled in white.

https://doi.org/10.1371/journal.pone.0285262.g002

To gain insights into the biological pathways underlying the differentially expressed genes, we conducted over-representation analysis on both the upregulated (Fig 3A & S4 Table) and downregulated genes (Fig 3B & S4 Table). We found that the downregulated DEGs exhibited more enrichment terms (42) compared to the upregulated genes (17). The ORA on the upregulated DEGs revealed enrichment of terms related to response to external stimulus, ABC-transport activity, and Factor: elt-3; motif: TCTTATCA (TF: M07154) based on GO, KEGG pathway, and TRANSFAC databases, respectively (Fig 3A & S4 Table). In contrast, ORA on the downregulated genes showed enrichment of terms related to metabolism of glycans, amino acids, and organic compounds, Factor: elt-3; motif: TCTTATCA (TF: M07154) among others, based on the same databases (Fig 3B & S4 Table). Differentially expressed genes from ORA terms with putative involvement in drug metabolism were explored further.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3.

A. Biological processes underpinned by upregulated genes. An over-representation plot showing significantly enriched for terms among upregulated DEGs based on Gene ontology (molecular function, biological process and cellular component), Kyoto Encyclopedia of Genes and Genomes (KEGG), and TRANSFAC databases. The dots represent an enriched term. B. Biological processes underpinned by downregulated genes. An over-representation plot showing significantly enriched for terms among downregulated DEGs based on Gene ontology (molecular function, biological process and cellular component), Kyoto Encyclopedia of Genes and Genomes (KEGG), and TRANSFAC databases.

https://doi.org/10.1371/journal.pone.0285262.g003

3.2.1 Differentially expressed genes encoding drug receptors.

After exposure of N2 worms to IVM 10⁻⁷ M, the putative drug target glc-1, an alpha subunit of the GluCI receptor, glc-1 was downregulated 1.4-fold. In addition, the putative drug target, lgc-26, a possible nicotinic acetylcholine receptor was downregulated 1.6-fold after exposure to IVM 10⁻⁷ M (Table 1). However, in IVM 10⁻⁸ M concentration, differential expression of the putative drug receptors was not observed.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Upregulated genes putatively involved in IVM response in C. elegans N2 strain.

https://doi.org/10.1371/journal.pone.0285262.t001

3.2.2 Differentially expressed genes involved in Phase I and II xenobiotic metabolism.

There were 11 DEGs putatively classified to be involved in Phase I metabolism, of which eight were downregulated (Table 2) and three were upregulated in IVM 10⁻⁷ M concentration (Table 1). Eight genes were attributed to the cytochrome P450 family, of which three genes, cyp-33C8, cyp-34A8 and cyp-13B2 were on average upregulated 1.8-fold while the rest, cyp-33E1, cyp-13A5, cyp-25A1, cyp-35A2 and cyp-34A9 were downregulated between 1.5 and 2.2-fold. Other genes involved in drug metabolism with differential expression included Flavin-containing monoxygenase, fmo-2, short chain dehydrogenase, dhs-7 and aldehyde dehydrogenase adh-5, which were downregulated between 1.5 and 6-fold.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Downregulated genes putatively involved in IVM response in C. elegans N2 strain.

https://doi.org/10.1371/journal.pone.0285262.t002

Twelve DEGs putatively identified as UDP-glycosyltransferases and glutathione S-transferases involved in Phase II metabolism were enriched exclusively in IVM 10⁻⁷ M (Tables 1 & 2). UDP-glycosyltransferases, ugt-62 was upregulated (Table 1) above 1.8-fold while ugt-11, ugt-12, ugt-22, ugt-43 and ugt-44 were downregulated between 1.4 and 3-fold (Table 2). In addition, glutathione S-transferases, gst-4, gst-6, gst-20, gst-42 and gstk-2 were downregulated between 1.5 and 2-fold. Only one gene, ugt-22, appeared in 10⁻⁸ M, and was downregulated 1.5-fold.

3.2.3 Differentially expressed genes involved in ion transporter activity.

There were 15 DEGs putatively identified or confirmed as ion channels (Tables 1 & 2). Genes best-1, best-5, catp-1, cca-1, F47E1.4, kcc-2, ncx-2, nlr-1, tmc-2, Y70G10A.3, and ZK185.5, were downregulated between 1.4- and 2.4- fold (Table 2). Conversely, genes atp-6, best-21, F58G6.9 and kvs-5 were upregulated on average 2-fold (Table 1).

3.2.4 Differentially expressed genes involved in other solute transmembrane transporter activity.

There were 30 DEGs involved in transmembrane transport and comprised of various genes and gene families (Tables 1 & 2). The DEGs consisted predominantly of P-glycoprotein efflux pumps (3 genes), other solute transporters (27 genes). Efflux pumps included, pgp-5, pgp-6, and pgp-9, appeared exclusively in IVM 10⁻⁷ M and, were all upregulated above 2-fold (Table 1). Other solute transporters included folate transporter (folt-2), long-chain fatty acid transporter (pmp-1 and pmp-4), amino acid transporter (aat-3, slc-25A29, slc-36.5, pept-1, glt-5, and slc-28.1), myo-inositol transport (hmit-1.3), sugar transmembrane transporter (swt-1, K09C4.1) etc. Of these, only folt-2 and pmp-4 were upregulated (Table 1).

3.2.5 Differentially expressed genes involved in transcriptional regulation of xenobiotic metabolism.

The differential screening identified 35 transcriptional factors (Tables 1 & 2), of these, nine were nuclear hormone receptors (NHRs) and one was a GATA-type transcription factor. These ten have been postulated to be involved in xenobiotic metabolism (see review [61]). The NHRs, nhr-17, nhr-57 and nhr-115, nhr-137 were upregulated between 1.4 and 3.2-fold (Table 1) whereas nhr-21, nhr-55, nhr-99, nhr-119, nhr-121 and nhr-173 were downregulated between 1.6 and 2.8-fold. The GATA transcription factor, elt-2 was downregulated 1.6-fold (Table 2).

3.2.6 Differentially expressed genes involved in apoptosis.

The Ivermectin concentration of 10⁻⁷ M was found to exhibit a greater number of differentially expressed genes (ces-1, ces-2, lec-6, T05C3.6, and pal-1) involved in apoptosis compared to the concentration of 10⁻⁸ M (pal-1) (Tables 1 & 2). Four of these DEGs (ces-1, ces-2, lec-6, and T05C3.6) were observed to be downregulated between 1.4 and 2.3-fold at IVM 10⁻⁷ M (Table 2). C2H2-type zinc finger transcription factor ces-1 and basic region leucine-zipper (bZIP) transcription factor ces-2 have been previously reported as repressors of the gene egl-1, a key activator of apoptosis in C. elegans [62]. The presence of caudal-type homeodomain transcription factor pal-1, an activator of the apoptotic gene ced-3 (Fig 4), was detected in both IVM 10⁻⁷ M and 10⁻⁸ M and was found to be equally upregulated (~ 1.5-fold) in both concentrations (Table 1). This suggests that IVM (at 10⁻⁷ M and 10⁻⁸ M) partially activates the apoptotic machinery in the N2 C. elegans strain.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Adapted illustration of the genetic pathway of apoptosis from Conradt and Xue [62].

The boxes symbolize genes, where red and green boxes represent apoptotic genes that in the current study were downregulated or upregulated, respectively according to our thresholds (see Methods). Solid and dashed T-shaped lines represent repression of gene expression, whereas solid and dashed black arrows represent the opposite. Hypothetically, the green and red arrows represent the respective gene’s upregulation and downregulation consequent of T-lines and black arrows.

https://doi.org/10.1371/journal.pone.0285262.g004

3.3 Comparative analysis of gene expression in the C. elegans strains N2 and DA1316 following exposure to IVM

We compared our transcriptomic data from the C. elegans strain N2 to microarray data obtained from the IVM-resistant C. elegans strain DA1316 in a previous study [34]. Based on our preset criteria (see Methods), 152 genes were differentially expressed in the DA1316 strain after IVM exposure for 4 h. Ninety-one percent (139 genes) of DEGs were exclusively detected in 10⁻⁶ M, 7% (11 genes) in10^-7 M and 1% (2 genes) occurred in both concentrations (S1 Fig).

Overall, irrespective of IVM concentration, comparison of DEGs between DA1316 and N2 strains revealed that the N2 strain showed four times more differentially expressed genes. Thirty-one genes showed significant overlap (p-value < 0.00001) between the two strains (Fig 5 & Table 3). Of these, 61% (19 genes) showed an opposite expression i.e., upregulation in one strain and downregulation in another, notably folate transporter gene, folt-2 and putative transmembrane transporter, T22F3.11 (Table 3 & S5 Table). Gene folt-2 was downregulated 6-fold and 2-fold in DA1316: 10⁻⁶ and DA1316: 10⁻⁷, respectively, and upregulated 4-fold in N2:10⁻⁷ M concentration (S5 Table). Similarly, T22F3.11 was upregulated 3-fold in DA1316: 10⁻⁷ M and downregulated 7-fold in N2:10⁻⁷ M concentration. Eighteen of the 31 genes were enriched for Phase II metabolic, transmembrane transport and stress response processes. Twelve of the shared genes (acox-1.5, cpt-5, drd-5, F42A10.7, gst-4, pud-1.2, pud-2.1, pud-3, pud-4, sams-1, ugt-12 and ugt-22) were downregulated in both strains. The majority of the shared DEGs (22 genes) occurred in DA1316: 10⁻⁶ M and N2:10⁻⁷ M concentrations, whereas six genes (mtl-1, C23G10.11, scl-2, F21C10.10, T22F3.11 and adh-1) occurred in DA1316: 10⁻⁷ M and N2:10⁻⁷ M concentrations (Fig 5 & S5 Table).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Differentially expressed genes in IVM-exposed C. elegans N2 and DA1316 strains.

An UpSetR plot and Venn diagrams displaying the number of differentially expressed genes that overlap within and between C. elegans strains N2 and DA1316 [34] exposed to IVM 10⁻⁷ M, 10⁻⁸ M and 10⁻⁶ M, 10⁻⁷ M, respectively. The black boxes indicate the number of genes contained within each partition.

https://doi.org/10.1371/journal.pone.0285262.g005

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Overlapping differentially expressed genes between IVM-exposed C. elegans N2 and DA1316 strain.

https://doi.org/10.1371/journal.pone.0285262.t003

3.4 Ivermectin-induced DEGs are mapped within the Abamectin QTL on chromosome V

Evans et al. [35] identified three major QTL (VL, VR and VC) correlated to natural variation in Abamectin response, on chromosome V in C. elegans. We decided to use this information to see if any of the DEGs induced by IVM, map in the QTL regions. There was significant (adjusted p = 0.01) enrichment for DEGs in chromosomes X and V (S6 Table). A comparison of the DEGs of IVM-exposed C. elegans N2 adult worms to the QTL revealed that 45 DEGs mapped to the three QTL, 51% (23 genes) were upregulated, and the majority (71%) mapped to the VL locus (Fig 6 & S7 Table). Of the 45 DEGs, nine (folt-2, mtl-1, pud-1.2, pud-2.1, pud-3, pud-4, rips-1, T13F3.6 and T22F3.11) overlap with microarray data from IVM-resistant strain (DA1316) [34]. In addition, eight other DEGs from the DA1316 strain mapped to the QTL (S8 Table). The VL locus contained 32 DEGs (S7 Table) from N2 strain, which included seven metabolic genes (cyp-33C8, cyp-34A9, acs-3, T06A1.5, dach-1, alh-5, cyp-34A8), six transport-related genes (dod-3, T22F3.11, Y39H10B.2, folt-2, Y32G9B.1, slc-28.1), six stress response genes (hsp-16.2, irg-1, hsp-16.41, Y73C8C.3, irg-2, Y73C8C.8), two transcription factors (nhr-57, nhr-115), a muscle contraction regulator (tnt-4) and 11 others, including pud-genes (S7 Table). Genes cyp-33C8, cyp-34A9, dod-3, F40C5.2, hsp-16.2, hsp-16.41, pud-1.2, pud-2.1, pud-3, and pud-4 were the ten most significantly (adjusted p<0.001) expressed at the VL locus. Of the 10 genes, seven (cyp-34A9, dod-3, F40C5.2, pud-1.2, pud-2.1, pud-3 and pud-4) were downregulated between 1.6 and 4.4-fold while the rest were upregulated on average 2.3-fold (Fig 6). Six DEGs mapped to the VC locus (Fig 6), which comprised of three stress response genes (F31F7.1, mtl-1, clec-7), a transport-related gene (col-143), a metabolic gene (lipl-4), and an unknown gene (S7 Table). Genes F31F7.1, mtl-1, clec-7, lipl-4, and spp-27 were downregulated between 1.5 and 3.6-fold, with the exception of col-143, which was upregulated 2.4-fold (Fig 6 & S7 Table). Seven DEGs mapped to the VR locus (Fig 6) composed of a GluCl subunit (glc-1), a transporter (F11A5.9) and five unknown genes (S7 Table). The expression of the genes glc-1, F11A5.9, and oac-18 was decreased by 1.4-fold, 1.4-fold, and 2.6-fold, respectively. The remaining genes had 1.54- to 10-fold higher expression levels (Fig 6).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Differentially expressed genes of IVM-exposed C. elegans N2 strain map to Abamectin-QTL on chromosome V.

The DEGs are represented by green (upregulated) and red (downregulated) dots, with the size of the dot representing the -log10 (adj. p-value) of expression, where the larger the dot, the more significantly expressed. The x-axis represents chromosome V and the y-axis represents the Log2 fold change. The karyoplot shows DEGs (labeled in green and red) mapped to abamectin-QTL VL, VC, and VR represented by yellow, blue, and yellow boxes, respectively.

https://doi.org/10.1371/journal.pone.0285262.g006

A list of candidate genes for further study (S9 Table) has been created, including genes with an opposite expression between N2 and DA1316 C. elegans strain, genes within Abamectin-QTL, exclusive genes in DA1316 strain, and other putative drug-responsive genes.

4 Discussion

Parasitic nematodes pose a threat to human and animal health [2, 3], while the development of resistance against MLs such as IVM exacerbates the problem [16, 17, 63]. Therefore, it is crucial to understand the molecular mechanisms underlying ML resistance. In this study, we investigated gene expression in wild-type (N2) IVM-exposed C. elegans adults at maximum reproduction to understand the transcriptomic response to IVM exposure. Furthermore, the transcriptomic data was compared to microarray data from the IVM-resistant C. elegans (DA1316) strain [34], and the recently identified Abamectin-QTL [35]. We identified 615 differentially expressed genes in the wild-type C. elegans (N2) strain following exposure to IVM, of which 95% occurred in the IVM 10⁻⁷ M concentration. Only 31 N2 strain DEGs overlapped with the IVM-resistant C. elegans (DA1316) strain, 19 of which, including folate transporter FOLT-2 (folt-2) and transmembrane transporter T22F3.11 (T22F3.11), displayed an opposite expression in either strain. Eighteen of the overlapping genes were enriched for metabolic Phase II enzymes, transmembrane transport, and stress response processes. In addition, we identified 45 differentially expressed genes that mapped within the Abamectin-QTL on C. elegans chromosome V, nine of which overlapped with DEGs from the IVM-resistant C. elegans (DA1316) strain. Following comprehensive analysis we identified potential candidate genes, T-type calcium channel CCA-1 (cca-1), potassium chloride cotransporter KCC2 (kcc-2), folate transporter FOLT-2 (folt-2) and alpha subunit of glutamate-gated channel GluCl-1 (glc-1) for further investigation with possible involvement in IVM resistance.

The ML drug target glc-1 and the putative drug target lgc-26 encoding ion channel subunits were downregulated in the IVM 10⁻⁷ M concentration. The significance of glc-1 in ML response is evident through RNAi knockdown and loss-of-function-mutations [26, 36, 64, 65]. Heterologous expression of C. elegans glc-1 and formation of a functional chloride channel has been reported in Xenopus oocytes [66]. The role of lgc-26 in ML response remains unknown in C. elegans, but other cys-loop GABA receptor members have been implicated in the ML response of parasitic nematodes. For example, lgc-37 was upregulated in the horse roundworm parasite, Parascaris univalens after IVM exposure [41] and the lgc-54 in the sheep parasite Teladorsagia circumcincta was reported as potential candidate in IVM resistance through genome-wide studies [67]. However, Evans, Wit [35] recently challenged this claim, as they observed that lgc-54 mutants did not display a competitive advantage over wild-type in control conditions.

We observed a downregulation of the calcium channel alpha subunit (cca-1) and potassium chloride cotransporter (kcc-2) genes after IVM exposure. Gene cca-1 regulates pharyngeal pumping by facilitating the effective start of action potentials by influx of calcium ions in response to marginal cell motor neuron stimulation in C. elegans [68, 69]. A loss of function in cca-1 has been reported to significantly reduce pharyngeal pumping [69]. Furthermore it is established that IVM reduces pharyngeal pumping in C. elegans by allosterically modulating GluCl hence increased influx of chloride ions [9]. Hypothetically, a pharyngeal pump regulator such as cca-1 would be upregulated in response to decreased pharyngeal pumping rate, however, our results showed the opposite trend. Alternatively, the downregulation of cca-1 could be a downstream effect, a consequence of the negative membrane potential induced by the chloride ion influx, or a defense/ protective mechanism aimed at reducing the adverse effects of IVM. Nonetheless, this highlights the importance of cca-1 in IVM response in C. elegans. The kcc-2 gene encodes a potassium chloride cotransporter, which, in conjunction with the sodium-driven chloride-bicarbonate transporter abts-1, mediates inhibitory GABA signaling. Mediation involves control of cellular chloride gradient to maintain membrane potential in neurons that control locomotion [70]. In addition, kcc-2 and abts-1 mutant of C. elegans were reported to exhibit paralysis after exposure to a GABA receptor agonist, muscimol [70]. Considering that IVM agonistically mediates GABA signaling through the inflow of chloride ions to cause paralysis [9], the role of kcc-2 in IVM response is unknown and warrants exploration. Overall, these findings shed light on additional channels, such as cca-1 and kcc-2, which could be investigated further as potential candidates for involvement in IVM resistance.

Seven of the 11 differentially expressed Transcription factors were downregulated and predominantly enriched for NHRs, which included nhr-17, nhr-21, nhr-55, nhr-57, nhr-99, nhr-115, nhr-119, nhr-121, nhr-137 and nhr-173. Although NHRs have generally been suggested to regulate expression of detoxification genes (phase I, II and III), only a handful have experimental evidence (see review [61]). To our knowledge, none of the NHRs in the current study has been implicated in xenobiotic detoxification regulation. However, a notable example, nhr-8 upregulation has been reported to increase expression of detoxification genes cyp14A2, cyp14A5 and cyp37B1, and Pgp genes (pgp-1,-3,-6,-9 and -13) resulting in the reduced IVM efficacy [71]. Recent work by Guerrero et al. [72] emphasized this phenomenon, demonstrating nhr-8 role in upregulation of Pgp genes (pgp-3, -5, -11 and -13) in the presence of the drug tunicamycin. The GATA-type transcription factor, elt-2 was downregulated in the current study. The proposed genes regulated by elt-2 in C. elegans include those involved in protection against xenobiotic compounds: CYPs, GSTs, and UGTs such as ugt-22 [73]. While it is conceivable that this downregulation of elt-2 might explain the previously described downregulation of ugt-22, it is unclear why elt-2 is downregulated in the presence of a xenobiotic such as IVM.

In this study, an analysis of putative apoptotic genes showed that four out of the five genes identified were downregulated, including C2H2-type zinc finger transcription factor ces-1 and basic region leucine-zipper (bZIP) transcription factor ces-2. These genes have been previously shown to act as repressors of the key apoptotic activator egl-1 under anti-apoptotic conditions. Under pro-apoptotic conditions, egl-1 represses ced-9, thereby activating the pro-apoptotic genes ced-3 and ced-4, hence apoptosis. The downregulation of ces-1 and ces-2 in this study is indicative of egl-1 being liberated and subsequently able to repress ced-9, thereby activating the pro-apoptotic genes ced-3 and ced-4, leading to apoptosis [62]. Additionally, caudal-type homeodomain transcription factor pal-1, a reported activator of ced-3 [74], was found to be upregulated. To further support the pro-apoptotic response hypothesis, genes ces-1, ces-2, and pal-1 were predominantly present in the highest IVM concentration (10⁻⁷ M), in which all worms were previously reported to be immobile [41]. While these findings suggest that IVM at concentrations of 10⁻⁷ M and 10⁻⁸ M partially activates the apoptotic machinery in the N2 C. elegans strain, it is important to note that other mechanisms involving above genes may also be at play as not all relevant apoptotic genes were detected in the data.

Our second objective was to compare C. elegans strains N2 (wild-type) and DA1316 (IVM-resistant) following IVM exposure. Comparing the RNAseq and microarray data from the N2 and DA1316 strains, respectively, revealed that the N2 strain had four times the number of differentially expressed genes with only a 4% (31 genes) overlap between the two strains. Nineteen of these overlapped genes displayed an opposite expression and may have a role in IVM metabolism or resistance. For example, the folate transporter folt-2, whose expression was reduced 6-fold and 2-fold in the resistant DA1316 strain, but increased 4-fold in the sensitive N2 strain, may be involved in the import of IVM. Another potential candidate is the transmembrane transporter T22F3.11, whose expression increased by 3-fold in the resistant strain, but decreased 7-fold in the sensitive N2 strain, suggesting a role in IVM efflux. In addition, genes are that are exclusively expressed in the resistant strain could also be an important group for further exploration. Overall, further studies are needed to confirm the role of these genes in IVM response. Twelve of the 31 overlapping genes were downregulated, including four pud-genes, pud-2.1, pud-3, and pud-4. Pud-genes are unique to Caenorhabditis spp. without known orthologues in other nematodes and their function remains elusive. In spite of that, Cui et al. [75] reported slow growth and hypersensitivity to cadmium of C. elegans after pud-4 (F15E11.12) RNAi knockdown. Another study reported downregulation of pud-1.2 and pud-4 following viral infection of C. elegans [76]. Other overlapping genes such as hsp-17 encoding a heat shock protein, exhibited inverted regulation in either strain. Gene hsp-17, a stress responder, was downregulated 1.7-fold in DA1316 and upregulated 4-fold in N2 strain. This re-emphasizes the inherent sensitivity of the N2 strain to IVM than its resistant or tolerant counterpart. The discrepancy in gene expression between the two studies could be attributable to strain, technique, life-stage, or IVM concentration. Strain differences may account for most of the difference in DEGs, since N2 is more sensitive to IVM than resistant or tolerant strains [34, 77]. Thus, the N2 strain may have more elevated cellular and biological processes, which could account for the increase in DEGs. Overall, we speculate that pud-genes may be universal stress responders exclusively in C. elegans.

As a final objective, we mapped the 615 DEGs derived from the IVM-exposed C. elegans to the recently identified Abamectin-QTL [35]. We identified 45 DEGs that corresponded to the QTL regions and compared them to the proposed list of candidate genes from the QTL [35], but only two genes, putative folate transporter (folt-2) and dod-3, between the two data sets overlapped. Gene folt-2 supposedly participates in the active uptake of folate [78], a B-class vitamin central in the synthesis of nucleotides and amino acids [79]. Like other eukaryotes, C. elegans are intrinsically deficient in folate [80], and therefore rely on dietary sources such as OP50 E. coli. In our data, folt-2 was upregulated, which may be caused by restriction in dietary intake due IVM-induced decreased pharyngeal pumping [9]. This could lead to folate deficiency in the worm. Whether this folate deficiency triggers upregulation of folt-2 remains unclear and therefore entails further study. In contrast, the function dod-3 and its role ML response is unknown. Overall, the role of folt-2, dod-3 and other 43 genes that corresponded to the Abamectin-QTL in ML response is unknown and prompts further investigation.

In conclusion, we have profiled the transcriptomes of adult wild-type (N2) C. elegans after exposure to different IVM concentrations and revealed predominantly downregulated diverse sets of genes. Some genes overlapped with a previous microarray study on IVM-resistant C. elegans while others mapped to recently described abamectin-QTL. Based on this, we have a compiled a list of potential candidate genes for further investigation. Although Abamectin and Ivermectin (IVM) belong to the same subgroup of avermectins, it is important to exercise caution when making comparisons between the two drugs. While Abamectin serves as a precursor drug for IVM, the pharmacokinetics of these drugs have been reported to be considerably different [81, 82]. These differences have been attributed, in part, to variations in the drugs’ lipophilicity [83]. Future studies employing quantitative genetics approaches, such as identifying IVM-QTL and corresponding candidate genes, validated through the use of near-isogenic lines, mutant strains, and transcriptomic profiling, would provide more detailed information.

Supporting information

Acknowledgments

We would like to thank the Science for Life Laboratory (SciLifeLab) for providing sequencing services and Swedish National Infrastructure for Computing (SNIC) at UPPMAX for enabling computations and data handling resources. Shweta Roy’s contributions were invaluable to the success of the experimental setup.