MyD88 dimerization inhibitor ST2825 targets the aggressiveness of synovial fibroblasts in rheumatoid arthritis patients

We aimed to elucidate the effect of disrupting MyD88 dimerization by ST2825 on the pathological properties of SFs from RA patients in an unbiased manner using transcriptomic analysis and biological assays. Our working hypothesis is that the synthetic small molecule ST2825 mitigates the aggressive behavior of RA SFs mediated by their proliferative and antiapoptotic capacity, altered expression of inflammatory mediators, and invasive properties.

Experiments on human SFs were performed according to the guidelines and approval of the Emory University Institutional Review Board. OA SFs were prepared from the knee synovium collected from OA patients undergoing total joint replacement as described in our previous publications [33, 34]. RA SFs isolated from the knee synovium of RA patients were collected from cadaveric donors within 48 h of death and were purchased from Articular Engineering, LLC. Dermal fibroblasts from healthy human donors (hDF) were purchased from Lonza. Cells were cultured in DMEM with 10% fetal bovine serum (Corning) and 1% penicillin/streptomycin and utilized between passages 3 and 6. Age and sex of the patients from which the SFs were derived are as described in Fig. S2. Cells were incubated at 37 °C in a humidified 5% CO₂ atmosphere and 95% humidity.

The hDF, OA SFs, and RA SFs were cultured in a 12-well plate after the density adjustment at 1 × 10⁵ cells/mL. The percentages of cells in the G0/G1, S, G2/M phases and apoptotic cells were determined at 24, 48, and 72 h after the incubation with 0, 5, and 10 μM of ST2825. Briefly, cells were permeabilized with 70% ethanol, and their nuclei were stained with DAPI (4 nM) and examined by measuring the amount of DNA per single cell in a multifluorescent channel analysis (blue 377/477) on the Celigo Imaging Cytometer, according to Nexcelom Bioscience protocols (Assay ID: Celigo_02_0001). The percentage of apoptotic cells was determined by gating for intact cells containing less than 2 × amount of DNA calculated at the interface of DAPI integrated intensity on the Celigo Imaging Cytometer (Additional file 1: Fig. S5).

Cell invasion assays were carried out on RA SFs 3D spheroids according to a previously published protocol using the Celigo Imaging Cytometer [35]. RA SFs were plated in Nexcelom3D Ultra-low Attachment Round Bottom 96-well Plates at a density of 4000 cells/well in 200 μL of medium and cultured for a period of 4 days, following which the formation of spheroids was visually confirmed using an Olympus CKX53 Microscope (× 10 magnification). On day 4, the Corning® Matrigel® matrix was thawed on ice according to the manufacturer’s protocols. LPS alone at a concentration of 30 ng/mL or LPS containing 10 μM ST2825 were added to the Matrigel and distributed into the 96-well plates containing RA SF spheroids. For unstimulated RA SFs, pure Matrigel was added. Matrigel containing 10 μM ST2825 was added as a control. Automated image acquisition and quantification of the invaded area in μm² were performed on the Celigo Imaging Cytometer at 0, 48, 72, and 96 h after the addition of Matrigel, according to the Nexcelom Bioscience protocols (Assay ID: Celigo_03_0002) [35].

RA SFs were cultured in 6-well plates after the density of 2.5 × 10⁵ cells/well followed by stimulation with 30 ng/mL of LPS, with or without 10 μM of ST2825 for 24 h. ST2825 was added 45 min prior to LPS stimulation. Unstimulated RA SFs were taken as the control group. Additionally, unstimulated OA SFs were cultured and collected after 24 h and were considered as control cells for unstimulated RA SFs. The total RNA was extracted and purified using Direct-zol RNA MicroPrep (Zymo Research) following the manufacturer’s protocol. RNA quality and quantity were assessed using a 2100 Bioanalyzer (Agilent Technologies). Only samples with an RNA integrity number (RIN) > 8.6 were used for sequencing. Three biological replicates were used per condition. Libraries were generated from 250 ng RNA using TruSeq Stranded Total RNA Sample Prep Kit (Illumina). Sequencing was carried out using the NovaSeq 6000 system (Novogene UC Davis Sequencing Center, Novogene Corporation Inc.). FASTQ files from these samples were uploaded to the DNASTAR Lasergene (version 17.3.0.57) and ArrayStar software for analysis. Paired-end reads were mapped to the GRCh37 human genome assembly. RNA levels were normalized using Log2 RPKM (reads per kilobase of exon model per million mapped sequence) by performing Student’s t-test for genes at 95% confidence.

Differentially expressed genes (DEGs) were identified using the BioJupies web tool [36]. DEGs with p-value < 0.05 and fold change (FC) ≥ 1.5 were used to perform volcano plots, heatmaps, and subsequent pathway analysis. QIAGEN Ingenuity Pathway Analysis (IPA) was used to identify distinct up- and downregulated canonical pathways between different conditions in our study. Data was visualized using GraphPad Prism version 9.4.1, VolcaNoseR, and BioJupies web tools [36, 37]. Protein interaction networks of relevant DEGs were identified and analyzed in STRING version 11.5 [38], the interaction score was set at high confidence = 0.700, and a false discovery rate (FDR) cutoff of 0.05 was used to determine up- or downregulated pathways in the network.

The cDNA was synthesized from 1 μg of RNA using qScript™ cDNA SuperMix 5X (QuantaBio). Quantitative real-time polymerase chain reaction (qRT-PCR) was performed using PerfeCTa SYBR® Green FastMix 2X (QuantaBio) on the QuantStudio™ 6 Flex Real-Time PCR System (Applied Biosystems™). For this analysis, at least two biological replicates and 3 technical replicates were used. Fold change (FC) was determined through the 2^−ΔΔCt method, and TBP mRNA level was used as the reference gene. Primer sequences are detailed in Additional file 1: Table S1.

Cells were harvested and lysed using Pierce™ IP lysis buffer (Thermo Fisher Scientific) containing 1% Halt™ protease and phosphatase inhibitors (Thermo Fisher Scientific). Equal amounts of protein were electrophoresed on 10% SDS–polyacrylamide gels. The primary antibodies used for the western blot are listed in Additional file 1: Table S2. Bands were visualized using ChemiDoc MP Imaging System (Bio-Rad). For luciferase reporter assays plasmids encoding NF-κB (pNL3.2.NF-kB-RE (Promega) and IL-15 [39], RA SFs were transiently transfected with Viromer Red reagent (Origene). The pSV2bgal plasmid was used as a control for transfection efficiency. Cells were treated with LPS (35 ng/mL), ST2825 (10 μM), or both for 24 h. Luciferase and β-galactosidase activities were assayed using the GloMax Discover System (Promega). Reporter activities were normalized for transfection efficiency and reported as fold change in luciferase activity. Assays were performed as triplicates.

Normality tests were performed to determine the data distribution and appropriate descriptive and inferential statistics. The results were analyzed using the software GraphPad Prism version 9.4.1. The specific statistical analysis is described in the corresponding figure caption. One-way ANOVA, two-way ANOVA and Dunnett’s multiple comparisons, Student’s t-test, and one-way ANOVA and Tukey’s multiple comparisons test were used. A p-value less than 0.05 was considered statistically significant.

To identify the effects of inhibiting MyD88 dimerization, the viability of hDF incubated with 0, 5, 10, and 30 μM of ST2825 for 24 h was determined (Fig. 1A). The results indicated that 5 and 10 μM of ST2825 did not compromise cell viability; therefore, those two concentrations were used for subsequent experiments. To evaluate the long-term effect of ST2825 on proliferation, hDF and OA SFs were monitored for 24, 48, and 72 h after incubation with 5 and 10 μM of ST2825 (Fig. 1B, D). The analysis reported no statistically significant differences among conditions, suggesting that cell viability was not compromised. Although not significant, we observed a drop in cell number, particularly in hDF treated with 10 μM of ST2825. Thus, the DNA content of hDF and OA SFs was quantified to investigate the effect of ST2825 on different phases of the cell cycle (Fig. 1C, E and Additional file 1: Fig. S1). Our analysis revealed that 10 μM of ST2825 induced a significant decrease in the percentage of phase S cells in hDF and OA SFs, with a corresponding increase in the percentage of G0/G1 phase OA SFs at 48 h. Interestingly, these changes were accentuated in OA SFs treated with both 5 and 10 μM of ST2825 at 72 h, suggesting that ST2825 may influence cell proliferation through induction of G0/G1 cell cycle arrest.

Since our goal was to determine whether the MyD88 dimerization inhibition mediated by ST2825 modulates pathogenic processes in RA SFs, we confirmed the inflammatory status of the RA SFs samples used in this study by comparing them with OA SFs isolated from OA synovium of patients undergoing total knee replacement (Additional file 1: Fig. S2A). We identified a total of 1879 differentially expressed genes (DEGs) by 1.5-fold change (p < 0.05) between RA and OA SFs. The analysis revealed 826 downregulated and 1053 upregulated genes (Additional file 1: Fig. S2B). Hierarchical cluster analysis of normalized gene expression from RA and OA SFs shows 2129 genes at 95% confidence that clustered together in the appropriate group (Additional file 1: Fig. S2C). Upregulated and downregulated genes were further analyzed to identify canonical pathways underlying pathogenic features in RA SFs (Additional file 1: Fig. S2D). The overexpression of important inflammatory genes, along with the increased expression of MYD88 observed in RA SFs, established the suitability of the RA SFs samples for this study (Additional file 1: Fig. S2E). Importantly, it supports our hypothesis for targeting this protein in SF as a potential pharmacological approach in the arthritic joints.

To determine the biological effects of ST2825 on RA SFs, we monitored cell proliferation for 24, 48, and 72 h after incubation with 5 and 10 μM of ST2825 (Fig. 2A). Although cell numbers remained unchanged, the treatment with 10 μM of ST2825 caused a significant increase in the percentage of G0/G1 phase and significantly decreased the percentage of S and G2/M phase RA SFs (Fig. 2B). The same analysis was performed on RA SFs incubated with ST2825 for 48 and 72 h, even though a similar trend was observed, the changes in cell cycle phases were not significant suggesting that the ST2825 acted within 24 h of the treatment (Additional file 1: Figs. S3 and S4). We, therefore, selected the 24 h time point for transcriptomic analysis by RNA-seq. Genes were selected based on the reactome pathway enrichment (Additional file 1: Fig. S6 and Additional file 2: Supplementary Data 1), and the main downregulated pathways were related to the cell cycle, which matches with the results obtained for the analysis of the cell cycle states. Those downregulated genes with 1.5 FC and p < 0.05 were further analyzed through QIAGEN Ingenuity Pathway Analysis (IPA), and common genes across downregulated pathways were selected for heatmaps. The transcriptomic analysis revealed cell cycle regulation among the top five IPA canonical pathways downregulated by ST2825 (Fig. 2C). We next performed a paired differential gene expression analysis in the 3 independent RA SFs samples; heatmaps show that ST2825 treatment successfully downregulated cell cycle-related genes (Fig. 2D). The predicted protein interaction network showed the main gene cluster involved in the regulation of G1 cell cycle arrest (green), G0 and early G1 (red), and Cyclin D-associated events in G1 (blue) at high interaction confidence = 0.700 and PPI enrichment p-value = 1.43E − 07 (Fig. 2E). The percentage of apoptotic cells was evaluated in all time points among conditions; significant differences in apoptosis were only observed at 72 h with 10 μM of ST2825 (Additional file 1: Fig. S5). Taken together, these observations suggest that MyD88 dimerization inhibition arrests cell cycle progression by inducing G0/G1 cell cycle arrest, which may potentially modulate the exacerbated proliferation and apoptosis resistance observed in RA SFs.

We next investigated the effect of ST2825 under inflammatory conditions by stimulating RA SFs with LPS, a well-known pro-inflammatory inductor known to activate TLR signaling upstream of MyD88 dimerization. RNA sequencing was performed from LPS-stimulated RA SFs after 24 h (Additional file 1: Fig. S7 and Additional file 3: Supplementary Data 2). In order to investigate the possible role of ST2825 in modulating the expression of critical genes involved with RA SF aggressiveness, we manually selected some genes associated with catabolic and inflammatory pathways along with pain mediators (Fig. 3A). The analysis revealed upregulation of critical genes involved in catabolic processes such as CXCL10, TNFSF13B, ADAMTS4 (ADAM metallopeptidase with thrombospondin type 1 motif 4), and IL1B (Interleukin 1 beta). Similarly, the upregulation of several genes related to inflammatory pathways such as TLR1 (Toll-like receptor 1), TLR3 (Toll-like receptor 3), CASP1, and MYD88 was observed. It is important to mention that some of these genes were also upregulated in RA SF vs. OA SF in Additional file 1: Fig. S2, supporting our hypothesis that ST2825 can target distinct aspects of RA SF aggressiveness. One of the most interesting findings was the LPS-mediated upregulation of interleukin 33 (IL33), C–C motif chemokine ligand 2 (CCL2), and prostaglandin-endoperoxide synthase 2 (PTGS2), genes that have been described as pain mediators (Fig. 3A). Differential gene expression analysis also showed that the most downregulated genes by the effect ST2825 on LPS-stimulated RA SFs were those related to cell cycle regulation and DNA mismatch repair (Fig. 3B, E). We successfully validated the ST2825-induced downregulation of the pro-inflammatory genes IL1B and MYD88 and the cell cycle regulators CCNE2 and MYBL2 (Myb-related protein B) genes by qRT-PCR (Fig. 3C, D). The predicted STRING protein interaction network showed that the main gene cluster involved in the regulation of G1 cell cycle arrest (yellow), G0 and early G1 (green), and DNA repair (blue) at high interaction confidence = 0.700 and PPI enrichment p-value = 0.00431 (Fig. 3F). Upregulated genes by the effect ST2825 on LPS-stimulated RA SFs were related to Sirtuin signaling, mitochondrial metabolism, pluripotency, apoptosis, and chondroitin sulfate biosynthesis (Fig. 3G). The predicted STRING protein interaction network also showed that upregulated genes participate in distinct mitochondrial processes; the analysis was performed at high interaction confidence = 0.700, and the PPI enrichment p-value was 4.25E − 08 (Fig. 3H).

Invasiveness represents a key pathogenic feature of RA SFs due to their contribution to pannus formation and joint destruction. In this study, LPS was used to potentiate the invasive properties of RA SFs. In order to evaluate long-term changes, spheroids were performed and monitored for 48, 72, and 96 h after incubation with LPS or LPS plus ST2825 (Fig. 4A). The area of the invasion reached statistical significance at 96 h, an effect that was significantly suppressed by the MyD88 inhibitor ST2825 (Fig. 4B, C). Taken together, the data obtained from this analysis demonstrates that local treatment with ST2825 may suppress the invasiveness of RA SFs contributing to ameliorating joint damage.

In order to elucidate the effect of ST2825 on MyD88 and NF-κB signaling pathways in RA SFs, we treated cells with LPS and ST2825 and performed western blot and luciferase reporter assays. Our results show that the levels of MyD88, IRAK4, and TRAF6 were not significantly changed by ST2825 (Fig. 5A and Additional file 1: Fig. S8). However, this result was expected as ST2825 inhibits MyD88 dimerization and subsequent signaling events by disrupting the formation of the Myddosome complex but does not directly affect the gene expression or protein level of MyD88 or its interacting partners. Nonetheless, we were able to demonstrate that the level of p65, which is the downstream effector of MyD88, was downregulated by ST2825 (Fig. 5A). In addition, our results showed that targeting MyD88 by ST2825 significantly inhibits p65-dependent NF-κB binding site reporter luciferase activity (Fig. 5B) and may have an additional effect by inactivating the IKK-related kinases TAK1 and TBK1, and NF-κB p65 in RA SFs treated with LPS (Fig. 5C and Additional file 1: Fig. S8).