Another Notch in the Belt of Rheumatoid Arthritis

Abstract

Notch ligands and receptors, including JAG1/2, DLL1/4, and Notch1/3, are enriched on macrophages (MΦs), fibroblast-like synoviocytes (FLS), and/or endothelial cells in rheumatoid arthritis (RA) compared with normal synovial tissues (ST). Power Doppler ultrasound-guided ST studies reveal that the Notch family is highly involved in early active RA, especially during neovascularization. In contrast, the Notch family is not implicated during the erosive stage, evidenced by their lack of correlation with radiographic damage in RA ST. Toll-like receptors and tumor necrosis factor (TNF) are the common inducers of Notch expression in RA MΦs, FLS, and endothelial cells. Among Notch ligands, JAG1 and/or DLL4 are most inducible by inflammatory responses in RA MΦs or endothelial cells and transactivate their receptors on RA FLS. TNF plays a central role on Notch ligands, as anti-TNF good responders display JAG1/2 and DLL1/4 transcriptional downregulation in RA ST myeloid cells. In in vitro studies, TNF increases Notch3 expression in MΦs, which is further amplified by RA FLS addition. Specific disease-modifying antirheumatic drugs reduced JAG1 and Notch3 expression in MΦ and RA FLS cocultures. Organoids containing FLS and endothelial cells have increased expression of JAG1 and Notch3. Nonetheless, Methotrexate, interleukin-6 receptor (IL-6R) antibodies, and B cell blockers are mostly ineffective at decreasing Notch family expression. NF-κB, MAPK, and AKT pathways are involved in Notch signaling, whereas JAK/STATs are not. Although Notch blockade has been effective in RA preclinical studies, its small molecule inhibitors have failed in phase I and II studies, suggesting that alternative strategies may be required to intercept their function.

Figure 1.. Overview of the Notch signaling pathway.

Step 1) Notch receptors are first synthesized in the ER, where they undergo glycosylation at the EGF-like repeats by ER enzymes (ex POFUT). Step 2) In the Golgi apparatus, Notch receptors are cleaved at the S1 site by a furin-like protease, which generates a mature heterodimeric form that is transported to the cell membrane. Also, fringe family glycosyl transferases extend O-fucose residues by adding N-acetylglucosamine. Step 3) Notch receptors are activated by binding to a ligand found on a neighboring cell (trans-activation). Step 4) Endocytosis of Notch ligands produces an essential mechanical pulling force impacting the bound Notch receptor by exposing the S2 site for subsequent cleavage by ADAM metalloproteases. Steps 5 and 6) S2 cleavage by ADAM10, ADAM17, or ADAMTS1 produces the NEXT fragment, which is further cleaved by the γ-secretase complex, releasing the NICD. In the absence of NICD, RBP-J binds corepressors, inhibiting target gene transcription. Step 7A) Once NICD translocates into the nucleus, it binds to RBP-J, facilitating the release of corepressors (Co-R) and the binding of transcriptional co-activators (Co-A), MAMLs and histone acetyltransferase p300 (P300) to activate transcription of target genes. Step 7B) NICD can also stay in the cytoplasm to crosstalk with other signaling pathways.

Figure 2.. Notch ligands and receptors are involved in early active but not established RA.

A-F) Relative expression of A) JAG1, B) JAG2, C) DLL1, D) DLL4, E) Notch1, and F) Notch3 was determined by RNA-seq in early and established RA ST biopsies. G-J) Linear regression of delta biopsy joint power Doppler ultrasonographic measures across 12 representative joints (delta Ultrasound PD 12) following treatment against G) JAG1, H) JAG2, I) DLL1, and J) Notch3 expression. K) Summary of Notch ligand and receptor expression in early active vs. established RA. p values were calculated using linear regression models or the Mann-Whitney test.

Figure 3.. Notch ligands and receptors are not involved in the erosive stage of RA.

Relative expression of A) JAG1, C) JAG2, E) DLL1, G) DLL4, I) Notch1, and K) Notch3 was determined by RNA-seq of RA ST from patients with or without radiographic erosion. Linear regression of biopsy joint synovial thickness measures across 12 representative joints (delta Ultrasound ST 12) against B) JAG1, D) JAG2, F) DLL1, H) DLL4, J) Notch1, and L) Notch3 expression. p values were calculated using linear regression models or the Mann-Whitney test.

Figure 4.. Notch ligands and receptors are differentially expressed in RA MΦs, FLS, and endothelial cells and are upregulated by RA inflammatory mediators.

A) DLL4 is ubiquitously highly expressed across all three cell types. Other Notch ligands are predominantly expressed on RA MΦ and FLS. Notch receptors are highly expressed on RA MΦs but also displayed at intermediate levels by RA FLS and endothelial cells. B) Inflammatory mediators found in the RA joint tend to upregulate several Notch ligands and, in a select few cases, Notch receptors in RA MΦs. In RA FLS, LPS/IFNγ and TNFα upregulate JAG2 and DLL1, but LPS/IFNγ and IL-6 upregulate Notch receptors. In endothelial cells, several inflammatory mediators upregulate Notch ligands, but only flagellin and IL-6 upregulate Notch receptors. lo, low; int, intermediate; hi, high. C) Heatmap displaying log(normalized counts) for Notch signaling components, target genes, and receptor modifiers in M-CSF differentiated CD14⁺ monocytes treated with LPS, IFNγ or primed with IFNγ then activated with LPS compared to control. D) Heatmap displaying log(normalized counts) for Notch signaling components, target genes, and receptor modifiers in TNF-treated RA FLS compared to untreated cells. p values were generated using the Wald test statistic and adjusted for multiple testing. *p < 0.05, ** p < 0.01, *** p < 0.001.

Figure 5.. Impact of RA medications on Notch ligands, receptors, and target genes.

A) Heatmap displaying average log(normalized counts) for Notch ligands, receptors, and target genes in MΦs alone, TNF-activated MΦs, or TNF-activated MΦs exposed to RA synovial fibroblasts. B) Heatmap displaying average log(normalized counts) for Notch ligands, receptors, and target genes for TNF-activated MΦs exposed to RA FLS treated with RA medications compared to TNF-activated MΦs + RA FLS in the absence of treatment. C) Heatmap displaying the scaled mean expression for Notch ligands, receptors, and target genes for a fibroblast plus endothelial cell organoid compared to a fibroblast only organoid. D) Summary table showing the impact of Methotrexate, anti-TNF, anti-IL-6R Ab, and B cell blocker therapies on Notch ligands, receptors, and target genes. p values were calculated using linear regression models. *p < 0.05, ** p < 0.01. Nap, Naproxen; Dex, Dexamethasone; Meth, Methotrexate; Sulfa, sulfasalazine; A77, leflunomide; Hyd, Hydroxychloroquine; Tri, triple DMARD (Hydroxychloroquine + sulfasalazine + Methotrexate); Aur, auranofin; Tofa, Tofacitinib.

Figure 6.. Summary depicting the impact of current RA therapies on Notch ligands, receptors, and target genes.

Naproxen impacts Notch3 and HEY1 expression. Dexamethasone (a steroid) reduces JAG1 and Notch3 expression. DMARD therapies do not impact Notch ligands but reduce Notch3 and HEY1 expression. Anti-TNF therapy broadly impacts expression of all Notch ligands and the downstream target HEY1. Anti-IL-6R Ab therapy only downregulates DLL4 expression. A77, leflunomide; Triple DMARD (Hydroxychloroquine + sulfasalazine + Methotrexate).