Single cell RNA sequencing reveals emergent notochord-derived cell subpopulations in the postnatal nucleus pulposus

Abstract

Intervertebral disc degeneration is a leading cause of chronic low back pain. Cell-based strategies that seek to treat disc degeneration by regenerating the central nucleus pulposus (NP) hold significant promise, but key challenges remain. One of these is the inability of therapeutic cells to effectively mimic the performance of native NP cells, which are unique amongst skeletal cell types in that they arise from the embryonic notochord. In this study, we use single cell RNA sequencing to demonstrate emergent heterogeneity amongst notochord-derived NP cells in the postnatal mouse disc. Specifically, we established the existence of progenitor and mature NP cells, corresponding to notochordal and chondrocyte-like cells, respectively. Mature NP cells exhibited significantly higher expression levels of extracellular matrix (ECM) genes including aggrecan, and collagens II and VI, along with elevated transforming growth factor-beta and phosphoinositide 3 kinase-protein kinase B signaling. Additionally, we identified Cd9 as a novel surface marker of mature NP cells, and demonstrated that these cells were localized to the NP periphery, increased in numbers with increasing postnatal age, and co-localized with emerging glycosaminoglycan-rich matrix. Finally, we used a goat model to show that Cd9+ NP cell numbers decrease with moderate severity disc degeneration, suggesting that these cells are associated with maintenance of the healthy NP ECM. Improved understanding of the developmental mechanisms underlying regulation of ECM deposition in the postnatal NP may inform improved regenerative strategies for disc degeneration and associated low back pain.

Keywords: Cd9; development; extracellular matrix; goat; intervertebral disc; mouse; single cell transcriptomics.

Figure 1.

A. In situ fluorescence imaging of sections from Shh/tdTomato mouse lumbar spines showing tdTomato+ cells (red) localized to the NP throughout postnatal growth. Krt19 immunofluorescence (green) on the same sections was used to confirm NP cell identity. Blue: cell nuclei (DAPI); mid-sagittal sections; scale = 100μm. B. Representative plots from fluorescence activate cell sorting used to enrich tdTomato+ cells. C. UMAP plot of scRNA-seq results of cells isolated from P30 mice showing clustering of 15 different cell populations. D. Dot plots showing cluster-specific expression of NP cell markers, hematopoietic, endothelial and blood cell markers, and mesenchymal cell markers.

Figure 2.

A. UMAP plot of scRNA-seq results for NP cells only, identified two distinct cell subpopulations. B. Differentiation trajectory analysis demonstrated that NP cells in clusters 1 and 2 aligned along a pseudo timeline; and C. RNA velocity analysis showed that the direction of differentiation was predominantly from cluster 1 to cluster 2. Based on these findings, cluster 1 and 2 cells were denoted “progenitor” and “mature” NP cells, respectively. D. UMAP plots for NP marker genes, showing that Krt8, Krt18, Krt19 and T were expressed across both progenitor and mature NP cells. E. UMAP plots and F. Violin plots showing significantly (*) higher expression of NP-specific ECM genes in mature NP cells (log2 fold change).

Figure 3.

Pathway analysis. A. Top 20 GO terms significantly enriched in mature versus progenitor NP cells. B. Pathways significantly enriched in mature versus progenitor NP cells identified by KEGG analysis. Heatmaps showing differentially expressed genes in the C. ECM-receptor interactions; D. PI3kT-Akt; E. Protein digestion and absorption; F. Focal adhesion; and G. TGF-β signaling pathways respectively, between mature and progenitor NP cells.

Figure 4.

A. Gene network analysis for all 5 pathways enriched in mature versus progenitor NP cells. B. Secreted signaling interactions between NP cell populations identified using CellChat.

Figure 5.

UMAP and violin plots showing significantly (*) higher expression of the surface markers A. Cd109 and B. Cd9 in mature versus progenitor NP cells (log2 fold change). C. In situ fluorescence imaging of tdTomato+ cells (red), and corresponding immunofluorescence imaging of Cd109+ (green) and Cd9+ (white) cells in the NPs of mouse lumbar spines during postnatal growth. Cd9+ cells were localized to the NP periphery, with relative numbers increasing with postnatal age. Blue (DAPI) = cell nuclei; Midsagittal sections; 100μm. D. Flow cytometry analysis showing relative increases in the number of tdTomato/Krt19/Cd9+ positive cells with increasing postnatal age. E. Representative sections showing progressive accumulation of glycosaminoglycan-rich ECM at the NP periphery with increasing postnatal age. Alcian blue and picrosirius red staining; mid-sagittal sections; scale = 100μm.

Figure 6.

A. Representative histological sections of healthy and degenerate goat lumbar discs. Alcian blue and picrosirius red-staining; mid-sagittal sections; scale = 2mm. B. Semi-quantitative grading of healthy and degenerate goat discs. *p<0.05 vs healthy; N=3–4; Students t-test. C. Representative immunostaining of Cd9+ cells in the NPs of healthy and degenerate goat discs. Mid-sagittal sections; scale = 100μm (left) and 20μm (right). D. Quantification of the relative numbers of Cd9+ cells in the NPs of healthy and degenerate goat discs. *p<0.05 vs healthy; N=3–4; Students t-test. E. Flow cytometry analysis showing the relative number of Cd9+ positive cells in the healthy, adult goat NP.

Figure 7.

Schematic representations of A. The transition from progenitor to mature NP cells, which is characterized by increased ECM-receptor interaction, PI3kT-Akt, protein digestion and absorption, focal adhesion and TGF-β signaling, and expression of aggrecan, collagens II and VI, Sox9 and Cd9; and B. Emergent mature NP cells during postnatal growth with concomitant peripheral ECM deposition. Created with BioRender.com, individual license HN262G0UB3.