Meta-Analysis of Caenorhabditis elegans Transcriptomics Implicates Hedgehog-Like Signaling in Host-Microbe Interactions

Abstract

Controlling nematode-caused diseases that affect cattle and crops world-wide remains a critical economic issue, owing to the lack of effective sustainable interventions. The interdependence of roundworms and their environmental microbes, including their microbiota, offers an opportunity for developing more targeted anthelminthic strategies. However, paucity of information and a currently narrow understanding of nematode-microbe interactions limited to specific infection contexts has precluded us from exploiting it. With the advent of omics approaches to map host-microbe genetic interactions, particularly in the model roundworm Caenorhabditis elegans, large datasets are now available across multiple models, that enable identification of nematode-microbe-specific pathways. In this work we collected 20 transcriptomic datasets documenting gene expression changes of C. elegans exposed to 20 different commensal and pathogenic microbes, performing gene enrichment analyses followed by functional testing using RNA interference directed toward genes of interest, before contrasting results from transcriptomic meta-analyses and phenomics. Differential expression analyses revealed a broad enrichment in signaling, innate immune response and (lipid) metabolism genes. Amongst signaling gene families, the nematode-divergent and expanded Hedgehog-like signaling (HHLS) pathway featured prominently. Indeed, 24/60 C. elegans Hedgehog-like proteins (HRPs) and 15/27 Patched-related receptors (PTRs) were differentially expressed in at least four microbial contexts, while up to 32/60 HRPs could be differentially expressed in a single context. interestingly, differentially expressed genes followed a microbe-specific pattern, suggestive of an adaptive microbe-specific response. To investigate this further, we knocked-down 96 individual HHLS genes by RNAi, using high-throughput assays to assess their impact on three worm-gut infection models (Pseudomonas aeruginosa, Staphylococcus aureus, and Enterococcus faecalis) and two worm-commensal paradigms (Comamonas sp., and Bacillus subtilis). We notably identified new putative infection response genes whose upregulation was required for normal pathogen resistance (i.e., grl-21 and ptr-18 protective against E. faecalis), as well as commensal-specific host-gene expression changes that are required for normal host stress handling. Importantly, interactions appeared more microbe-specific than shared. Our results thus implicate the Hedgehog-like signaling pathway in the modulation and possibly fine-tuning of nematode-microbe interactions and support the idea that interventions targeting this pathway may provide a new avenue for anthelmintic development.

Keywords: stress; C. elegans; Hedgehog; LFASS; RNAi; host-microbe interactions; infection; transcriptomics.

FIGURE 1

Gene expression changes of C. elegans exposed to diverse microbes. (A) Shows a list of 20 different microbes for which gene expression data of exposed C. elegans is available. The table includes: The nature of the interaction with C. elegans, either pathogenic or commensal; the total number of unique significantly differentially expressed (DE) genes combining all different time points and datasets; the top two hits of gene set enrichment analysis performed with the online tool DAVID (https://david.ncifcrf.gov/) using the unique set of DE genes per microbe (both up- and down-regulated together) (the complete results of the enrichment analysis can be found in Supplementary Table 3); and the references from which the datasets were obtained. In (B) a Venn diagram showing the overlaps between the list of unique DE genes of C. elegans exposed to P. aeruginosa, E. faecalis, and S. aureus. The top two hits of the enrichment analysis using DAVID and representative genes for each enrichment term are shown. A full gene ID list and the full results of the enrichment analysis are shown in Supplementary Table 3.

FIGURE 2

C. elegans immune regulation and lipid metabolism genes change their expression in a microbe-specific manner. Gene expression plots (bubble plots) that represent changes in gene expression as log₂Fold Change (Log₂FC) in a color heatmap and the adjusted p-value (FDR) as the diameter of the points. Published transcriptomic datasets were not re-analyzed from raw data, or re-scaled, therefore datasets from different studies cannot be compared quantitatively, only qualitatively. Detailed information about the datasets shown can be found in Supplementary Table 2. (A) Shows gene expression changes in a list of selected immunity genes (Pukkila-Worley and Ausubel, 2012; Martineau et al., 2021). (B) Shows gene expression changes in a list of selected lipid metabolism genes (Watts and Ristow, 2017).

FIGURE 3

In C. elegans Hedgehog-like (HHL) genes and genetic interactors play microbe-specific roles in the defense against infection and microbe-mediated protection from stress. (A) Bubble plots representing gene expression changes in Hedgehog-like (HHL), patched-related receptor (PTR) and hedgehog genetic interactors after exposure to different microbes. Bubble plots represent changes in gene expression as log2Fold Change (Log₂FC) in a color heatmap and the adjusted p-value (FDR) as the diameter of the points. Published transcriptomic datasets were not re-analyzed from raw data, or re-scaled, therefore datasets from different studies cannot be compared quantitatively, only qualitatively. (B) Heatmaps showing results of mean time of death (estimated by LFASS) after infection with P. aeruginosa (Pa), E. faecalis (Ef), and S. aureus (Sa). Bubble plots showing gene expression changes after exposure to the same microbes is repeated for comparison. (C) Heatmaps showing results of mean time of death (estimated by LFASS) after exposure to deadly oxidative stress 7% t-BHP and heat stress (42°C). Before the stress assays, RNAi-treated worms were exposed for 24 h at 25°C to commensal G- (Comamonas sp.) and G + (B. subtilis) stress-protective bacterial isolates. Bubble plots showing gene expression changes after exposure to Comamonas sp. and B. subtilis is repeated for comparison. x represents genes not included in the RNAi screen. YA, Young Adult, GA, Gravid Adult.

FIGURE 4

Schematic representation of approach to screen Hedgehog-like (HHL) genes and genetic interactors by RNAi inactivation. For all assays the RNAi-sensitive C. elegans strain NL2009 was used. Time of death was detected by blue death fluorescence as established in the Label-Free Automated Survival Scoring (LFASS) method (Benedetto et al., 2019). (A) Shows details of the approach for infection with P. aeruginosa, E. faecalis, and S. aureus and (B) shows details of the approach for the oxidative stress assay with 7% t-BHP and the heat stress assay at 42°C. Before the stress assays, RNAi-treated worms were exposed for 24 h at 25°C to commensal G- (Comamonas sp.) and G + (B. subtilis) stress-protective bacterial isolates. YA, Young Adult.

FIGURE 5

Targeting Hedgehog-like signaling to disrupt nematode-microbe interactions. (A) Hedgehog-like signaling (HHLS) involves multiple regulated steps with druggable targets (ligand processing/maturing enzymes, transporters, receptors and co-receptors, regulators of membrane trafficking and recycling) and regulates multiple aspects of nematode physiology (growth, development, immune and structural defenses, nervous system, and gut functions). Based on paralog sequence similarities, overlaps in tissue expression, and transcriptomics data, nematode HHLS may rely on partially redundant and/or combinatorial signaling, to be elucidated. (B) Whether and how nematode Hedgehog-related peptides (HRPs) may directly or indirectly interact with microbes also remain to be clearly established. (C) The main anthelminthic strategies currently in use or being explored* mostly target the nematode neuromuscular system and do not intersect with HHLS (Elfawal et al., 2019; Hahnel et al., 2020; Partridge et al., 2020). (D) Main Hedgehog signaling inhibitors used in oncology either target the less conserved part of the HH pathway (Smoothened and Gli), Wnt signaling, or conserved but low-specificity activators of HH pathway genes (Peer et al., 2019; Jamieson et al., 2020). HH, Hedgehog; HRP, Hedgehog-related peptide; PTRR, Patched-related receptor.

FIGURE 6

Divergence in sequence identity and differential effects of HHLS pathway gene inhibition on C. elegans phenotypes. (A) Percentage identity (bold) and percentage of sequence coverage (parentheses) for the most-likely orthologs of selected C. elegans HRPs and PTRs across nematode clades and mammals. Accession numbers for proteins are provided below percentage identity. Results were obtained using BLASTP or TBLASTN of C. elegans protein against target species databases. Low sequence identities between species suggest that targeting nematode HHLS pathway genes could be a valid strategy for species-specific interventions and may be unlikely to disrupt mammalian host Hedgehog signaling. (B) Scaled Principal Component Analysis (PCA) of individuals applied to survival data from C. elegans exposed to RNAi targeting HHLS genes (median times of death in stress and infection assays from Figures 3B,C). The plot highlights clusters of RNAi conditions that lead to similar outcomes in the phenotypic assays performed. Dimensions 1 and 3 spread variables in a meaningful manner (Dim 1 against Dim 2 is provided in Supplementary Figure 6). Dim 1 recapitulates variation in stress resistance and separates GRL/GRD from WRT ligands, while Dim 3 spreads data mainly according to resistance to infection and seems to separate PTR genes from reported genetic interactors of the C. elegans HHLS pathway (GI). “Other” regroups highly conserved non-PTR genes coding for sterol sensing domain proteins: ptd-2, che-14 (Dispatched orthologs), ncr-1, ncr-2 (NPC1 orthologs), and scp-1 (SCAP ortholog). RNAi clusters may reveal HHLS genes that are co-engaged in a response or a degree of functional redundancy. Overall, effects of individual HRPs and PTRs vary significantly.