Molecular heterogeneity of C. elegans glia across sexes

Abstract

A comprehensive description of nervous system function, and sex dimorphism within, is incomplete without clear assessment of the diversity of its component cell types, neurons and glia. C. elegans has an invariant nervous system with the first mapped connectome of a multicellular organism and single-cell atlas of component neurons. Here we present single nuclear RNA-seq evaluation of glia across the entire adult C. elegans nervous system, including both sexes. Machine learning models enabled us to identify both sex-shared and sex-specific glia and glial subclasses. We have identified and validated molecular markers in silico and in vivo for these molecular subcategories. Comparative analytics also reveals previously unappreciated molecular heterogeneity in anatomically identical glia between and within sexes, indicating consequent functional heterogeneity. Furthermore, our datasets reveal that while adult C. elegans glia express neuropeptide genes, they lack the canonical unc-31/CAPS-dependent dense core vesicle release machinery. Thus, glia employ alternate neuromodulator processing mechanisms. Overall, this molecular atlas, available at www.wormglia.org, reveals rich insights into heterogeneity and sex dimorphism in glia across the entire nervous system of an adult animal.

Keywords: C. elegans glia; dense core vesicles; glia; glial atlas; sexual dimorphism; snRNA-seq.

Figure 1.. Anatomical and molecular characterization of adult C.elegans glia in hermaphrodites and males using pan-glial transcriptional reporter miR-228.

(A) Schematic example of a C.elegans sense organ or sensilla, consisting of one or more sensory neurons (gray) that associate with one socket (green) and one sheath (blue) glia. Below is a close-up of the nose tip (dotted magenta box) showing interactions between the neuron and two glia. (B) Schematic representation of adult hermaphrodite and male showing sex shared (blue), hermaphrodite-specific (magenta), and male-specific (orange) glial nuclei. Close-up of the head (b), mid-body (b’) and hermaphrodite and male tails (b”) and glia within the region. (C) Z-stack projection and stitched tiles depict adult hermaphrodite (left) and male (right) expressing pan-glial cytoplasmic GFP and nuclear localized RFP. Animals are also expressing co-injection marker coelomocyte RFP. Scale bars = 50μM. (D) UMAP of 51 non-batch corrected clusters from day one adult hermaphrodites and males. (E) UMAP of 51 non-batch corrected clusters from day one adult hermaphrodites and males showing sex contribution to each cluster (male samples = orange; hermaphrodite samples = blue). Genotypes: Figure 2C: him-5; P_miR-228:GFP; P_miR-228:nls:RFP, P_unc-122:RFP.

Figure 2.. Validation of cluster identities.

(A) UMAP projection depicting expression of known AMsh/PHsh genes (F53F4.13, T02B11.3, fig-1, vap-1) within Cluster 14. Heatmap shows expression level. (B) Z-stack projections of adult hermaphrodite and male heads and tails with dotted outlines. Mean gene expression of each gene within cluster 14 shown at the top. Left side panels shows a transcriptional reporter for ZK822.4 depicting expression in the AMsh/PHsh glia in both sexes. Right side shows the far-8 reporter displaying mostly hermaphrodite tail expression. (C) Z-stack projections of L1 animal showing expression of transcriptional reporter for ZK822.4 in AMsh/PHsh. (D) Graph depicting the percentage of adults that retained expression in AMsh/PHsh in each sex. Animals like the representative in (C) were picked as L1 larvae if they had expression in both the head and tail (i.e. AMsh and PHsh) and then were scored again as day one adults. N=18 males and 21 hermaphrodites. (E) Non-batch corrected UMAP depicting 34 glial cell clusters (dark red), 8 neuronal clusters (yellow), and 8 anatomical clusters (dark blue). (F) UMAP of 32 batch corrected glial only clusters and their identifies. (G) UMAP of 32 batch corrected glial only clusters depicting sex specificity: 1 hermaphrodite-specific cluster (blue), 9 male-specific clusters (lavender), and 22 sex-shared clusters (red). (H) Pairwise comparison analysis for all male cluster 24 showing expression of genes (x-axis) per cluster (y-axis). Counts per gene for each sample shown at the top. The higher the number to total number of clusters (i.e. 32), the more unique the expression of the gene. Genes Y67D8C.7 and ttr-59 (red dashes) were chosen due to their high and unique expression. (I) Z-stack projections of the Y67D8C.7 transcriptional reporter. Hermaphrodite tail shown as merged. Male tail shown as individual and merged channels. Arrows point to the four miR-228 RFP+ nuclei that co-localize with the GFP+ cells projecting into the spicule. Genotypes: Figure 2B: him-5; P_ZK822.4:GFP, P_unc-122:GFP; P_miR-228:nls:RFP, P_unc-122:RFP & him-5; P_far-8:GFP, P_unc-122:GFP; P_miR-228:nls:RFP, P_unc-122:RFP. Figure 2C, D: him-5; P_far-8:GFP, P_unc-122:GFP; P_miR-228:nls:RFP, P_unc-122:RFP. Figure 2I: him-5; P_Y67D8C.7:GFP, P_unc-122:GFP; P_miR-228:nls:RFP, P_unc-122:RFP. All scale bars = 10μM.

Figure 3.. Unsupervised clustering identifies markers for populations of glia.

(A) Cosine similarity using highly variable genes shows relationship between clusters. (B) Hierarchical clustering shows separation of clusters into two groups on a dendrogram. (C) Binary classifier machine learning model’s top 10 candidate genes for sheath and socket markers. (D) Z-stack projection of the kcc-3 transcriptional reporter in hermaphrodite head/tail and male tail shows expression in sheath glia. Arrowhead depicts GFP− AMso glial nuclei. (E) Z-stack projection of the zipt-2.2 transcriptional reporter in hermaphrodite head/tail and male tail shows expression in socket glia. Arrowheads depicts GFP+ in PHso2 and PHso1-derived PHD neuron. (F) UMAP of batch-corrected glial only clusters sheath/socket identity based on molecular identity. (G) Binary classifier machine learning model’s top 10 candidate genes for pan-glia markers. Genotypes: Figure 3D: him-5; P_kcc-3:GFP; P_miR-228:nls:RFP, P_unc-122:RFP. Figure 3E: him-5; P_zipt-2.2:GFP, P_unc-122:GFP; P_miR-228:nls:RFP, P_unc-122:RFP. All scale bars = 10μM.

Figure 4.. Transcriptional reporters reveal sexually dimorphic expression.

(A) Z-stack projections of adult hermaphrodite and male heads and tails with dotted outlines showing expression of transcriptional reporter for Y52E8A.3. In hermaphrodites, expression is observed in AMso glia and Phso1 glia (arrows), as well as pharyngeal neurons. In males, expression is observed in Amso glia (arrow), MCM neuron (arrowhead) in the head and PHD neuron in the tail (arrowhead). Arrow depicts potential Phso1 glia still present in male. (B) Top: Z-stack projections of adult hermaphrodite and male tails with dotted outlines showing expression of transcriptional reporter F40H3.2. In both sexes, expression is observed in Phso2 glia (arrows). Bottom: Z-stack projections of adult hermaphrodite and male tails with dotted outlines showing expression of transcriptional reporter F35C5.12. Expression is observed in Phso2 glia (arrow) only in hermaphrodites. (C) Glial only batch corrected UMAP showing lack of ncx-10 gene expression within the glial clusters. (D) Z-stack projections of adult male tail showing expression of col-177 in all ray glia. © Z-stack projections of adult male tail showing expression of T27D12.1 in a subset of ray glia (specifically in RnSt 1, 3, 5, 7). (F) Z-stack projections of adult male tail showing expression of Y71H10B.1 specifically in RnSt 6/ray 6 glia. (G) Glial only batch corrected UMAP showing tra-2 gene expression within Olso clusters only. Genotypes: Figure 4A: him-5; P_Y52E8A.3:GFP, P_unc-122:GFP; P_miR-228:nls:RFP, P_unc-122:RFP. Figure 4B: him-5; P_F40H3.2:GFP, P_unc-122:GFP; P_miR-228:nls:RFP, P_unc-122:RFP & him-5; P_F35C5.12:GFP, P_unc-122:GFP; P_miR-228:nls:RFP, P_unc-122:RFP. Figure 4D: him-5; P_col-177:GFP, P_unc-122:GFP; P_miR-228:nls:RFP, P_unc-122:RFP. Figure 4E: him-5; P_T27D12.1:GFP, P_unc-122:GFP; P_miR-228:nls:RFP, P_unc-122:RFP. Figure 4F: him-5; P_Y71H10B.1:GFP, P_unc-122:GFP; P_miR-228:nls:RFP, P_unc-122:RFP. All scale bars = 10μM.

Figure 5.. Global analysis reveals that glia lack canonical DCV release and neuropeptide processing mechanisms.

(A) Glial only batch corrected UMAP showing lack of unc-31 gene expression in glial clusters. (B) Z-stack projections of adult hermaphrodite and male heads and tails with dotted outlines showing expression of transcriptional reporter for unc-31. In hermaphrodites, no colocalization is observed. In males, expression is observed in MCM and PHD neurons (arrows). (b’) High magnification images and split panels showing expression of unc-31 in the MCM neuron (lower), but not the AMso glia directly above it. (C) Glial only batch corrected UMAP showing lack of egl-3 gene expression in glial clusters. (D) Z-stack projections of adult hermaphrodite and male heads and tails with dotted outlines showing expression of transcriptional reporter for egl-3. In hermaphrodites, no colocalization is observed. In males, expression is observed in MCM and PHD neurons (arrows). (d’) High magnification images and split panels showing expression of egl-3 in the PHD neurons. (E) Glial only batch corrected UMAP showing kpc-1 and bli-4 gene expression in glial clusters. Gene kpc-1 is especially enriched in the IL/OLso clusters. (F) Heatamp showing expression of flp (blue), ins (red), and nlp (green) neuropeptide genes in the batch corrected glial clusters. Clusters are represented on the y-axis and socket genes are on top (black) while sheath genes are on the bottom (brown). Genotypes: Figure 5B: him-5; P_unc-31:GFP; P_miR-228:nls:RFP, P_unc-122:RFP. Figure 5D: him-5; P_egl-3:GFP; P_miR-228:nls:RFP, P_unc-122:RFP. All scale bars = 10μM.