JAGGED1 stimulates cranial neural crest cell osteoblast commitment pathways and bone regeneration independent of canonical NOTCH signaling

Abstract

Craniofacial bone loss is a complex clinical problem with limited regenerative solutions. Currently, BMP2 is used as a bone-regenerative therapy in adults, but in pediatric cases of bone loss, it is not FDA-approved due to concerns of life-threatening inflammation and cancer. Development of a bone-regenerative therapy for children will transform our ability to reduce the morbidity associated with current autologous bone grafting techniques. We discovered that JAGGED1 (JAG1) induces cranial neural crest (CNC) cell osteoblast commitment during craniofacial intramembranous ossification, suggesting that exogenous JAG1 delivery is a potential craniofacial bone-regenerative approach. In this study, we found that JAG1 delivery using synthetic hydrogels containing O9-1 cells, a CNC cell line, into critical-sized calvarial defects in C57BL/6 mice provided robust bone-regeneration. Since JAG1 signals through canonical (Hes1/Hey1) and non-canonical (JAK2) NOTCH pathways in CNC cells, we used RNAseq to analyze transcriptional pathways activated in CNC cells treated with JAG1 ± DAPT, a NOTCH-canonical pathway inhibitor. JAG1 upregulated expression of multiple NOTCH canonical pathway genes (Hes1), which were downregulated in the presence of DAPT. JAG1 also induced bone chemokines (Cxcl1), regulators of cytoskeletal organization and cell migration (Rhou), signaling targets (STAT5), promoters of early osteoblast cell proliferation (Prl2c2, Smurf1 and Esrra), and, inhibitors of osteoclasts (Id1). In the presence of DAPT, expression levels of Hes1 and Cxcl1 were decreased, whereas, Prl2c2, Smurf1, Esrra, Rhou and Id1 remain elevated, suggesting that JAG1 induces osteoblast proliferation through these non-canonical genes. Pathway analysis of JAG1 + DAPT-treated CNC cells revealed significant upregulation of multiple non-canonical pathways, including the cell cycle, tubulin pathway, regulators of Runx2 initiation and phosphorylation of STAT5 pathway. In total, our data show that JAG1 upregulates multiple pathways involved in osteogenesis, independent of the NOTCH canonical pathway. Moreover, our findings suggest that JAG1 delivery using a synthetic hydrogel, is a bone-regenerative approach with powerful translational potential.

Keywords: Bone regeneration; Intramembranous ossification; JAGGED1; Non-canonical JAG1-NOTCH pathways; RNAseq.

Figure 1:. JAG1 delivery in a PEG hydrogel stimulates bone regeneration in a critical-sized bone defect mouse model.

As a proof of concept experiment, we (A) incorporated JAG1-dynabeads complex (5 μM, 10 μM or 20 μM), dynabeads alone and BMP2 (2.5 μM) in 4% PEG-MAL hydrogels and implanted them into (B-G) 3.5 mm critical-sized defects in the parietal bones of 6–8-week old C57BL/6 mice and delivered them (J) without or (K) with CNC cells into the defects as 3 separate doses (Initial dose, Week 4, Week 8), as shown in H. After 12 weeks, we quantified differences in regenerated bone volume (J-K) within the defect and compared them between experimental groups by μCT analysis. μCT reconstructions of defects are shown in I. Data were subjected to ANOVA and Tukey’s post-test and are presented as mean (n ≥ 2) ± SD with p values reported.

Figure 2:. JAG1 induces canonical and non-canonical pathways in CNC cells.

(A) A principal component analysis of all treatments revealed that PC1 distinguished JAG1-treated cells to the right from non-JAG1 treated cells to the left along the horizontal axis. PC2 distinguished cells treated with JAG1 alone towards the top and JAG1 + DAPT treated cells towards the bottom along the vertical axis. (B) Clustering analysis using Pearson correlation-coefficient reveals the heatmap showing distinct gene clusters downstream of different treatments (n = 3). Cluster A and B consists of genes downregulated by JAG1 and Cluster C consists of genes upregulated by JAG1.

Figure 3:. JAG1 induces bone-inductive genes.

(A-B) Volcano plots of differentially expressed genes in CNC cells that were treated with JAG1-dynabeads complex (5.7 μM) with or without DAPT (15 μM) compared to Fc-dynabeads complex (5.7 μM). (C) Subsequent comparison of the JAG1 and JAG1 + DAPT treatment groups for genes that were significantly upregulated in either group compared to Fc. Plots indicate genes that were down-regulated in response to DAPT (red) or preserved between groups (grey). (D) Comparison between JAG1 treated-samples in the presence and absence of DAPT revealed various distinct genes, like Cxcl1, Cxcl12, Hes1 and Il6 upregulated downstream of JAG1 alone and Prl2c2 upregulated downstream of JAG1 + DAPT along with some conserved genes like Id1 and Rhou. Data were subjected to ANOVA and Tukey’s post-test and are presented as mean (n ≥ 2) ± SD with p values reported.

Figure 4:. JAG1 enriches canonical NOTCH gene sets.

(A) Gene set variation analysis (GSEA) quantified differences in 2199 gene sets. Clustering analysis (Euclidean + Ward’s minimum variance) revealed 3 clusters: Cluster I involves gene sets that were elevated in the JAG1-treated group, Cluster II shows gene sets that were downregulated in JAG1 + DAPT-treated samples and Cluster III indicates gene sets downregulated in JAG1 alone treated cells. (B) GSVA pathway analysis of selected pathways from CNC cells that were treated with JAG1-dynabeads complex (5.7 μM), with or without DAPT and Fc-dynabeads complex (5.7 μM) revealed various pathways up- and downregulated by JAG1 compared to JAG1 + DAPT. The Hes/Hey pathway (C) and the NOTCH pathway (D) were upregulated canonically and downregulated non-canonically downstream of JAG1. qFDR values are reported from a permutation analysis. Heatmap labels indicate selected genes whose expression levels are significantly different (p < 0.05) (red, blue, green fonts) in JAG1-treated groups compared to the Fc-treated sample. Expression levels of selected genes are significantly upregulated (Red font), downregulated (Blue font) or unchanged (Green font) in JAG1 + DAPT treated cells compared to JAG1 treated cells. See Suppl. Table 1–2 for entire list of genes shown in heatmaps from C & D.

Figure 5:. JAG1 enriches bone development gene sets via the NOTCH non-canonical pathway.

Pathway analysis of data obtained from sequencing of RNA that was isolated from CNC cells that were treated with JAG1-dynabeads complex (5.7 μM), with or without DAPT, a NOTCH canonical pathway inhibitor and Fc-dynabeads complex (5.7 μM) revealed that JAG1 induces (A) the cell cycle pathway, supporting proliferation, (B) the tubulin pathway, involved in cell structure reorganization essential for CNC cell osteoblast commitment, (C) the regulation of Runx2 initiation pathway, essential for osteoblast proliferation and commitment. Heatmap labels indicate selected genes whose expression levels are significantly different (p < 0.05) (red, blue, green fonts) in JAG1-treated groups compared to the Fc-treated sample. Those in black font are not significantly different from Fc-treated groups (p > 0.05). Expression levels of selected genes are significantly upregulated (Red font), downregulated (Blue font) or unchanged (Green font) in JAG1 + DAPT treated cells compared to JAG1 treated cells. See Suppl. Table 3–5 for entire list of genes. qFDR values are reported from a permutation analysis. (D) Luminex analysis revealed that STAT5 phosphorylation was increased by JAG1 treatment. Data were subjected to ANOVA and Tukey’s post-test and are presented as mean (n ≥ 2) ± SD with p values reported.

Figure 6:. Summary & conceptualized pathway.

Cranial neural crest cells treated with JAG1-dynabeads with or without DAPT, a NOTCH canonical pathway inhibitor, showed activation of the JAK2-STAT5 pathway. Red and Green solid arrows indicate non-canonical and canonical NOTCH targets of JAG1, respectively, downstream of JAK2-STAT5 (Prl2c2, Id1, Rhou). Light green arrows indicate non-canonical NOTCH targets and yellow arrows indicate canonical NOTCH targets. Dotted arrows indicate targets that may or may not be activated by STAT5. Navy blue arrows indicate targets that can induce phosphorylation of STAT5. Pathways activated non-canonically downstream of JAG1 revealed by pathway analyses are indicated by light blue arrows. Greater the arrow thickness higher the gene expression levels. All these events lead to JAG1-induced CNC cell osteoblast commitment and thereafter bone formation by intramembranous ossification.