Single-cell RNA sequencing reveals different chondrocyte states in femoral cartilage between osteoarthritis and healthy individuals

Abstract

Background: Cartilage injury is the main pathological manifestation of osteoarthritis (OA). Healthy chondrocyte is a prerequisite for cartilage regeneration and repair. Differences between healthy and OA chondrocyte types and the role these types play in cartilage regeneration and OA progression are unclear.

Method: This study conducted single-cell RNA sequencing (scRNA-seq) on the cartilage from normal distal femur of the knee (NC group) and OA femur (OA group) cartilage, the chondrocyte atlas was constructed, and the differences of cell subtypes between the two groups were compared. Pseudo-time and RNA velocity analysis were both performed to verify the possible differentiation sequence of cell subtypes. GO and KEGG pathway enrichment analysis were used to explore the potential functional characteristics of each cell subtype, and to predict the functional changes during cell differentiation. Differences in transcriptional regulation in subtypes were explored by single-cell regulatory network inference and clustering (SCENIC). The distribution of each cell subtype in cartilage tissue was identified by immunohistochemical staining (IHC).

Result: A total of 75,104 cells were included, they were divided into 19 clusters and annotated as 11 chondrocyte subtypes, including two new chondrocyte subtypes: METRNL+ and PRG4+ subtype. METRNL+ is in an early stage during chondrocyte differentiation, and RegC-B is in an intermediate state before chondrocyte dedifferentiation. With cell differentiation, cell subtypes shift from genetic expression to extracellular matrix adhesion and collagen remodeling, and signal pathways shift from HIF-1 to Hippo. The 11 subtypes were finally classified as intrinsic chondrocytes, effector chondrocytes, abnormally differentiated chondrocytes and dedifferentiated chondrocytes. IHC was used to verify the presence and distribution of each chondrocyte subtype.

Conclusion: This study screened two new chondrocyte subtypes, and a novel classification of each subtype was proposed. METRNL+ subtype is in an early stage during chondrocyte differentiation, and its transcriptomic characteristics and specific pathways provide a foundation for cartilage regeneration. EC-B, PRG4+ RegC-B, and FC are typical subtypes in the OA group, and the HippO-Taz pathway enriched by these cell subtypes may play a role in cartilage repair and OA progression. RegC-B is in the intermediate state before chondrocyte dedifferentiation, and its transcriptomic characteristics may provide a theoretical basis for intervening chondrocyte dedifferentiation.

Keywords: cell differentiation; chondrocyte; inflammation; osteoarthritis; single-cell RNA sequencing.

Figure 1

The process of scRNA-seq and the results of cell annotation (A) Flow chart of the experimental strategy. It includes sampling from femur of NC and OA, single cell isolation, cell counting and quality control, sequencing library preparation, single cell RNA-seq, phenotypic classification and statistical analysis. (B) Coefficient for nCount_RNA and nFeature_RNA (C) Parameters for screening nCount_RNA, nFeature_RNA and mitochondrial genes (D) Umap of all cells colored according to cell clusters. The annotation results of 19 clusters are listed on the right. (E) Umap of all cells colored according to cell subtypes (F) Heat map showing scaled differently expressed genes for each cell type defined in (E). Specific representative genes in each cell type are listed along the left of heat map. (G) umap of all cells according to indicated marker genes for each cell subtype.

Figure 2

Enrichment analysis and proportion of each subtype, and RNA velocity analysis of two newly identified subtypes (A) GO enrichment analysis of each subtype. The top 20 biological process terms for each subgroup based on p-value were summarized and visualized by heatmap according to -log10p values. (B) The proportion of each subtype in each sample (C) t-test for the proportion of the number of subtypes in each sample. *p < 0.05, **p <0.01, ***p <0.001 (D) RNA velocity analysis of METRNL+ subtype, HTC-A, HTC-B, and EC-A. Arrows show the direction and rate of cellular movement. (E) RNA velocity analysis of EC-B, RegC-B, and PRG4+ subtypes. (F) The top 10 DEGs of METRNL+ subtype based on log2FC. (G) GO enrichment analysis of METRNL+ subtype. The gene with Log2FC*PCT1/PCT2≥1.5 in this subtype was applied to enrichment analysis. (H) The top 10 DEGs of PRG4+ subtype based on log2FC. (I) GO enrichment analysis of PRG4+ subtype. The gene with Log2FC*PCT1/PCT2≥1.5 in this subtype was applied to enrichment analysis.

Figure 3

Differentiation sequence of chondrocyte subtypes (A, D, G) pseudo-time analysis of all cells, NC group and OA group. The dots in each color represents each state, and each state is named numerically. The direction of the arrow represents the direction of the pseudo-time. (B, E, H) umap visualization of state in (A, D, G). The meanings of the colors, numbers, and arrows correspond to the (A, D, G), respectively. (C, F, I, J) RNA velocity analysis of EC-A, EC-B, and RegC-B in all cells, NC group, and OA group cells. (K) Relative expression levels of COL1/2 subtypes (L) Summary of the differentiation sequence of each subtype. The arrows represent the differentiation direction.

Figure 4

Changes in enrichment analysis during cell differentiation (A, B) GO enrichment analysis of each subtype. The top 20 biological process terms based on p-value for each subtype were summarized and visualized by heatmap according to either the -log10p value (A) or the -log10Rich factor (B). (C–F) KEGG pathway enrichment analysis of METRNL+, EC-A, EC-B and RegC-B. The bottom horizontal axis represents the number of genes enriched into the each pathway. (G) Expression of important downstream target genes of Hippo signaling pathway in each subtype. The size of the circle represents the proportion of cells in the cell subtype expressing the gene, and the color of the circle represents the average gene expression. (H) Expression of WWTR1 in each sample.

Figure 5

Transcriptional regulation of chondrocyte subtypes (A) Heatmap of specific transcriptional factors of EC-A, EC-B, and RegC-B (B) RSS scores for specific transcription factors of EC-A, EC-B and RegC-B. (C–E) umap plot for key transcription factors of EC-A, EC-B, and RegC-B. (F) Gene regulatory network of specific transcription factors. The purple genes are the screened key transcription factor. The green and yellow genes are the highly expressed DEGs in the chondrocyte subtypes, and the yellow genes means that the gene itself is a transcription factor.

Figure 6

IHC results for specific markers for each cell type in different groups and layers. Scale bar, left, 200 μm; right, 50 μm. The scores of the indicated genes in the cartilage of both regions based on the immunohistochemistry assay are shown. *p<0.05, **p<0.01, ***p<0.001; ****p<0.0001; otherwise, not significant NC, normal femur cartilage; OA, OA femur cartilage.