Neurotrophin signaling is a central mechanism of salivary dysfunction after irradiation that disrupts myoepithelial cells

Abstract

The mechanisms that prevent regeneration of irradiated (IR) salivary glands remain elusive. Bulk RNAseq of IR versus non-IR human salivary glands showed that neurotrophin signaling is highly disrupted post-radiation. Neurotrophin receptors (NTRs) were significantly upregulated in myoepithelial cells (MECs) post-IR, and single cell RNAseq revealed that MECs pericytes, and duct cells are the main sources of neurotrophin ligands. Using two ex vivo models, we show that nerve growth factor (NGF) induces expression of MEC genes during development, and upregulation of NTRs in adult MECs is associated with stress-induced plasticity and morphological abnormalities in IR human glands. As MECs are epithelial progenitors after gland damage and are required for proper acinar cell contraction and secretion, we propose that MEC-specific upregulation of NTRs post-IR disrupts MEC differentiation and potentially impedes the ability of the gland to regenerate.

Fig. 1. Neurotrophin signaling is a key pathway dysregulated post-IR.

A Representative H&E staining of control and IR human SMG and PG. Areas in dotted boxes of IR SMG and IR PG are enlarged in right two panels to highlight increased adiposity (a), immune infiltration (i), and fibrosis (f). B PCA plot of log-transformed CPM counts from bulk RNAseq analysis of human biopsies. C Heatmaps of the top 50 DEGs by fold change in irradiated glands compared to controls. The color scale represents scaled gene expression values. *Long non-coding RNAs, small nuclear RNAs, SNORAs, and other pseudogenes were overrepresented among downregulated genes and were removed from heatmap (see Supplementary Data 1 for complete list); a representative expression profile is shown at the bottom of the heatmaps. D Bar graph with fold change gene expression of selected genes in irradiated glands (n = 6 IR-SMG; n = 5 IR-PG) compared to controls from RNAseq analysis. Stars denote statistical significance (adjusted p < 0.05, statistical analysis with EdgeR-DeSeq2 (Wald test p-value and Benjamini-Hochberg adjusted p-value)). E Results from Ingenuity Pathway Analysis (IPA) Software showing common pathways dysregulated in PG and SMG. A combined score was calculated by adding the –Log(p-val) for a given pathway in both glands. The Venn diagram highlights the overlap of genes associated with the top dysregulated pathways. * NGF and NTRK1 were only significant in IR-SMG and NTRK3 in IR-PG. F Heatmap showing results from Upstream Regulator analysis with IPA. Only genes that met the criteria for significance in our dataset (p-value < 0.05 and fold change > 2) and that were predicted to be upstream of DEGs in both glands are shown. Analysis was adjusted to only predict genes annotated to function as growth factors, receptors, and transcription factors, based on IPA’s database.

Fig. 2. NTRs are upregulated in MECs post-IR.

A Box plots showing the median expression of neurotrophin signaling genes in bulk-RNAseq data from human salivary glands (n = 7 SMG, n = 6 SMG-IR, n = 6 PG, n = 5 PG-IR). The box represents the interquartile range and the bars span the minimum and maximum values. Star denotes statistical significance (p-value < 0.05 and fold change > 2, EdgeR-DeSeq2-Limma pipeline). B qPCR analysis of selected genes in irradiated human SMG and PG samples (n = 4 per group). Gene expression was normalized to GAPDH and non-irradiated controls. Statistical significance (p-value < 0.05) was determined by two-tailed t-test with log-transformed fold changes and is shown with a star above bars. Error bars represent standard error of the mean (SEM). C UMAP of scRNAseq from human SG from the Tabula Sapiens. D Balloon plot of expression of IPA’s upstream regulator and neurotrophin signaling genes in Tabula Sapiens. Color scale represents average scaled expression and size reflects percentage of cells expressing a given gene. E Number of IPA’s upstream regulator genes expressed in each cell type. F, G Immunofluorescence staining for NGFR and neurotrophin receptors (green) in control (A) and IR (B) human SMG. Smooth muscle actin (SMA, red) labels myoepithelial cells, and nerve-specific tubulin beta 3 (TUBB3, white) labels peripheral nerves. Scale bar = 50 μm.

Fig. 3. Murine SMG is an ideal model to investigate neurotrophin signaling in healthy salivary glands and confirms that MECs are a major neurotrophin signaling hub.

A Immunostaining of SMA in mouse SMG at embryonic days E16 and E17, and postnatal days P1 and P8. B scRNAseq from mouse SMG (GSE150327). C, D Balloon plot of expression of neurotrophin signaling genes in mouse SMG at selected developmental stages. E Immunostaining for SMA (red), TrkA, TrkB and TrkC (green) and nuclei (DAPI, blue) in mouse SMG at E16, P1 and P8. White boxes (a-c) are shown enlarged. Scale bar, 50 µm except c) Scale bar, 20 µm. F Representation of putative ligand-receptor interactions between MECs and other cell types via neurotrophin signaling genes. Arrows point from the source of ligands in the direction of the receptor and the thickness of the arrow is relative to the number of potential interactions between two cells. Analysis and plots generated with the LigandReceptor script available on GitHub (https://github.com/chiblyaa/LigandReceptor). A list of curated pairs was obtained from Ramilowski et al..

Fig. 4. NGF correlates with late MEC differentiation whereas NTRK2 and NRTK3 mark early MEC development.

A PCA plots of MECs and endbud cells from scRNAseq data. Left plot is colored by developmental stage and right plot by pseudotime score. B Expression of neurotrophin receptors across pseudotime in endbud cells and MECs. C Heatmap showing MEC genes that are differentially expressed between developmental stages.

Fig. 5. NGF promotes MEC differentiation in SMG organ cultures.

A Experimental setup. Organ culture of E16 mouse SMG treated with NGF or GNF5837 for 24 h. B, C PCR results showing fold change gene expression of selected acinar, MEC, and duct genes 24 h post-treatment with NGF or GNF5837 at multiple doses. Statistical significance is represented by stars (n = 4 per group, p < 0.05; Two-way ANOVA with Dunnet’s correction for multiple comparisons). Error bars represent standard error of the mean. D) Immunofluorescence staining for MECS (CNN1, Green), developing endbuds/ducts (KRT19, Red) in E16 SMG collected 48 h after treatment with NGF (100 nM) or GNF (1 µM). Panels on the left show composite images while panels on the right show single-channel images for CNN1 staining. Scale bar = 20 µm. E Quantification of CNN1 + area and KRT19 + area from D. Dots represent individual areas analyzed from each gland where at least 5 random images per gland were analyzed. Stars denote statistical significance (students t-test, p < 0.001, n = 3 glands per group). The horizonal line and error bars represent the mean +/− SEM. F Heatmap of expression of neurotrophin signaling genes in scRNAseq from E16 SMG. G) Proposed mechanism for NGF induction of MEC and acinar differentiation via NTRK1 in nerves and glial cells.

Fig. 6. Dysregulation of neurotrophin receptors in MECs is associated with abnormal MEC differentiation and morphology.

A Experimental setup of murine MEC culture treated with exogenous NGF or RO for 48 h. PCR results showing fold change gene expression of selected MEC and duct genes 48 h post-treatment with NGF or RO (n = 4 per group). Statistical significance (*p < 0.05; **p < 0.01; ***p < 0.001; Unpaired t-test compared to each control). B Immunostaining of MEC culture treated with RO for 48 h, with antibodies to NGF, neurotrophin receptors, KRT14 and KRT19 (all Green). All cultures were stained with SMA (red) and DAPI (blue). Scale bar = 150 µm. C Relative quantitation of MEC immunostaining from B, normalized to nuclei staining. Data is shown as mean +/− SEM. Two-tailed unpaired t-test was used. D Summary of NGF signaling in MEC culture.

Fig. 7. Upregulation of NTRs is directly associated with abnormal morphology and co-expression of MIST1 and KRT19 in MECs post-IR in humans.

A Control and IR SMG stained for KRT14 (red), KRT19 (green), and DAPI (blue). White arrows point at cells with overlap between KRT14 and KRT19. Scale bars, 50 µm. B Control and IR SMG glands immunostained for KRT19 (green), SMA (red), and E-cadherin (blue). C Control and IR SMG and PG glands stained for MIST1 (green), SMA (red), and NKCC1 (blue). White arrows point at cells with overlap between SMA and MIST1. Scale bars, 50 µm. D Quantification of double-positive MIST1+/SMA+ cells normalized to number of MIST1+, and %SMA/K19 co-localization from IF staining (n = 3 PG IR, 2 PG non-IR, 4 SMG IR, 1 SMG non-IR). E Immunostaining of human SMG for KRT14 (Red), KRT19 (white), and TRKA (green). Delineated areas are expanded and denoted by labels ‘a’, ‘b’ and ‘c’. White arrows point at representative MECs. All scale bars, 50 µm. F Representation of morphological differences between control and IR MECs.