Fig 2: Expansion of PDGFRA+ non-myogenic mesenchymal cells in FSHD muscles.A Immunofluorescence staining for laminin (red), non-myogenic mesenchymal cells with PDGFRA (green) and nuclei with DAPI (blue). ×20 magnification. On the right, there are two enlarged representative areas showing PDGFRA+ cells for each experimental group. B The number of PDGFRA+ cells/field was reported as mean ± SD for each group. The ratio between the number of PDGFRA+ cells and the number of myofibers per field was assessed considering only the MRI parameter on STIR sequences (C) or also the T1 MRI value (D). The results were reported as mean ± SD. For all the analysis p-values were assessed by one-way ANOVA followed by Kruskal–Wallis multiple comparisons post hoc test; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Fig 3: Correlation of non-myogenic mesenchymal cell expansion with progression of fatty degeneration.A Patient 1 (FSHD#25): left gastrocnemius lateralis (green arrows) appears hyperintense on STIR sequence (STIR+) and normal on T1w sequence (no fat substitution). After 1-year follow-up T1w sequence showed partial fat substitution. Patient 2 (FSHD#8): right gastrocnemius lateralis (orange arrows) appears normal on STIR sequence (STIR-) and normal in T1w sequence (no fat substitution). After 1-year follow-up T1w sequence showed no fat substitution. Ratio of CD201+ cells/number of myofibers/field (B) and number of CD201+ cells/field (C), in muscles that did not worsen and that had a worsened on T1w sequences at follow-up at MRI, respectively. Ratio of PDGFRA+ cells/number of myofibers/field (D) and number of PDGFRA+ cells/field (E), in muscles that did not worsen and that worsened on T1w sequences at follow-up at MRI, respectively. For all the graphs, results are presented as mean ± SD and p-values were assessed by one-way ANOVA followed by Kruskal–Wallis multiple comparisons post hoc test; **p < 0.01, ***p < 0.001.

Fig 4: Evidence for cell lineage of the cortical tRG in the human brain. A Schematic diagram of the experimental design for this study (also see Materials and Methods). B Re-analysis of scRNA-Seq data from 3,355 frontal cortical EGFR+ cells at GW21-GW26 [35] revealed the molecular identity of cortical tRG and their progeny (PyN-IPCs, bMIPCs, OPCs, APCs, OBiN-IPCs and EP-ependymal cells). Note that pericytes and endothelial (Endo) cells also expressed EGFR. C tRGs and APCs expressed RGC general markers FABP7, HES1, SLC1A3, SOX2, and VIM. D, E EGFR+ASCL1+PAX6+ cortical tRGs generated EGFR+ASCL1+PAX6+EOMES+ PyN-IPCs. Note that very few NEUROD6+ immature PyNs expressed EGFR. F–I Cortical tRGs also generated EGFR+ASCL1+OLIG2+OLIG1+DLL3+ bMIPCs, which in turn gave rise to OPCs (H), APCs (C, G) and OBiN-IPCs (I) (Slingshot unsupervised pseudo-time analysis in F). Note that PDGFRA was also expressed in a subpopulation of bMIPCs and APCs (H). Also note that OBiN-IPCs expressed GAD2 earlier than SP8 and Sp9 (I).

Fig 5: Isolation and characterisation of non-myogenic mesenchymal cells.A Schematic representation of non-myogenic mesenchymal cell isolation protocol (panel created with Biorender.com). B Gating strategy for non-myogenic mesenchymal cell sorting from the CD56- cell fraction. C Representative histograms of specific markers in the CD56- cell fraction. D Heatmap showing the relative expression levels of positive markers. Original values are ln(x)-transformed. Rows are centred and scaled. Rows are clustered using Euclidean distance and Ward linkage. 9 rows, 13 columns. E PCA analysis by singular value decomposition of the same dataset showed in D. X and Y axis show principal component 1 (PC1) and principal component 2 (PC2). The percentage of the total variance explained by each PC is indicated. F Loadings (i.e. contributions) of each variable to PC2. G–J Relative expression, expressed as median fluorescence intensity (MFI) of CD201, CD140a, CD73 and CD90.