henhao
Sign in
ImmunologyNature Immunology
Peer-reviewed journal article

Spatial patterning of fibroblast TGFβ signaling underlies treatment resistance in rheumatoid arthritis

Kartik Bhamidipati, Alexa B. R. McIntyreiD, Shideh Kazerounian, Ce GaoiD, Soon Wei WongiD, Miles TraniD, Sean A. Prell, Rachel LauiD, Vikram Khedgikar, Christopher Altmann, Annabelle SmalliD, Roopa MadhuiD, Sonia PrestiiD, Ksenia S. AnufrievaiD, PHILIP E. BLAZARiD, Jeffrey K. LangeiD, Jennifer A. SeifertiD, Accelerating Medicines Partnership RA/SLE Network, Laura T. DonliniD, Accelerating Medicines Partnership: Autoimmune and Immune-Mediated Diseases Network (AMP-AIM), Melanie H. SmithiD, Ranjeny ThomasiD, Michael J. ZuscikiD, Michael R. ClayiD, Andrew D. Clauw, , Kevin WeiiD

Abstract

Treatment-refractory rheumatoid arthritis (RA) is a major unmet need, and the mechanisms driving treatment resistance are poorly understood. To identify molecular determinants of RA non-remission, we performed spatial transcriptomic profiling on pre- and post-treatment synovial tissue biopsies from treatment naive patients who received conventional DMARDs or adalimumab for 6 months. In the baseline biopsies of non-remission patients, we identified significant expansion of fibrogenic fibroblasts marked by high expression of COMP, a fibrosis-associated extracellular matrix protein. COMPhi fibroblasts localized to perivascular niches that, unexpectedly, served as transcriptional hubs for TGF{beta} activity. We identified endothelial-derived Notch signaling as an upstream regulator of fibroblast TGF{beta} signaling via its dual role in driving TGF{beta} isoform expression and suppressing TGF{beta} receptors, generating a proximal-distal gradient of TGF{beta} activity. Further, disruption of steady-state Notch signaling in vitro enabled fibrogenic fibroblast activation. Analysis of post-treatment biopsies revealed marked expansion of COMPhi fibroblasts in non-remission RA patients, despite evidence of successful immune cell depletion, suggesting a spatiotemporal process of fibrogenic remodeling linked to treatment resistance. Collectively, our data implicates targeting of TGF{beta} signaling to prevent exuberant synovial tissue fibrosis as a potential therapeutic strategy for refractory RA.

PDFDOI
Full-text digest ↓

02 /Specification & Classification

Click any tag to browse matching papers
Citations
32
FWCI
128.59
Field percentile
top 1%
OpenAlex Topics

In-depth summary

✦ Multi-agent verified full-text synthesis

Claude Opus 5.5 · xhigh effort · 5 Nov 2026 · from the full text of version 1 (17 Mar 2025) · CC BY 4.0

Perivascular COMPhi fibrogenic fibroblasts expand in treatment-refractory rheumatoid arthritis synovium, where endothelial Notch signaling patterns fibroblast TGF-beta sensitivity via TGFBR3 regulation and persists despite post-treatment immune cell depletion.

Over 50% of rheumatoid arthritis (RA) patients fail to achieve remission with initial anti-inflammatory therapies, and the tissue mechanisms driving refractory disease despite immune suppression remain unclear. Subcellular spatial transcriptomics of longitudinal synovial biopsies and endothelial-fibroblast co-cultures reveal that endothelial Notch signaling establishes a perivascular TGF-beta sensitivity gradient whose disruption drives COMPhi fibrogenic fibroblast expansion and treatment resistance.

Key results

  1. ShownCOMPhi fibrogenic fibroblasts are enriched for a treatment non-response gene signature and are elevated at baseline in early RA patients who fail to achieve 6-month remission.

    Across 7 spatial niches in 16 treatment-naive early RA biopsies (5,101-gene Xenium panel), a 290-gene non-responder signature from the PEAC bulk RNA-seq cohort (n = 69) enriched in stromal-adipose (padj = 8e-61) and fibroblast-rich (padj = 2e-14) niches and in COMPhi vs POSTNhi fibroblasts in AMP scRNA-seq (n = 82, p < 2.22e-16); baseline COMP was higher in 6-month non-remission vs remission patients (p = 0.0312). · Identification of distinct immune and…; COMP expression defines a fibrogenic program…; Figure legends

  2. ShownEndothelial cells organize a proximal-to-distal perivascular gradient separating Notch-active POSTNhi/TGFB1+/TGFB3+ fibroblasts from distal COMPhi/TGFBR2+/TGFBR3+ fibroblasts.

    In RA synovium (n = 4, 50-gene panel) and HUVEC-synovial fibroblast co-cultures, fibroblasts within ~1 cell layer (~23 um) of endothelial cells expressed Notch targets (JAG1, HES1, NOTCH3; p < 2.22e-308), POSTN (p = 5.27e-33), TGFB1 (p < 2.22e-308), and TGFB3 (p < 2.22e-308), whereas distal fibroblasts expressed COMP (p = 1.36e-111) and TGFBR1/2/3 (p <= 8.83e-205), with pSMAD3 restricted to perivascular niches. · Perivascular compartmentalization of TGFβ…; Endothelial cells generate spatial patterning of…; Figure legends

  3. ShownNotch activation induces fibroblast TGFB1, TGFB3, and POSTN while suppressing TGFBR2, TGFBR3, and COMP, and TGFBR3 overexpression restores COMP production.

    Stimulating synovial fibroblasts (n = 6) with plate-bound DLL4 induced TGFB1 (4.121-fold, p = 0.0006), TGFB3 (5.646-fold, p = 0.0138), pro-collagen I alpha (222.2 vs 30.92 ng/mL, p < 0.0001), and POSTN (56-fold, p = 0.003) while suppressing TGFBR2 (-2.613-fold, p = 0.0075), TGFBR3 (-2.006-fold, p = 0.0530), and COMP (-2-fold, p = 0.0082), reversed by 10 uM DAPT; TGFBR3 OE raised COMP 3.85-fold (p = 0.0001). · Patterning of fibroblast TGFβ responsiveness via…; Figure legends

  4. ShownDisrupting steady-state endothelial DLL4-Notch signaling derepresses fibroblast TGFBR3 and triggers spatial expansion of COMP-expressing fibroblasts.

    In HUVEC-fibroblast co-cultures, siRNA knockdown of endothelial DLL4 decreased POSTN production from day 3 (p = 0.0126) and TGFB3 expression (p = 6e-51), but derepressed TGFBR3 across proximal and distal zones (p < 2e-308) and increased secreted COMP protein from day 4 (p = 0.0109) and single-cell COMP transcript intensity (p = 1.4e-6); in vivo vascular niche Notch scores negatively correlated with TGFBR3. · Disruption of steady-state Notch patterning…; Figure legends

  5. ShownCOMPhi fibroblasts expand in post-treatment biopsies of non-remission patients despite immune cell depletion and associate with systemic serum fibrogenic proteins in long-duration RA.

    Across 17 paired 6-month post-treatment biopsies (triple DMARD or adalimumab), immune infiltrates fell regardless of response, whereas COMP+ fibroblasts expanded in 5 of 6 non-remission patients (vs <1% change in 5 of 10 remission patients) and accumulated in pauci-cellular zones (R = -0.44, p = 0.073); in AMP, CTAP-EFM patients had highest COMP, longer duration (p = 0.014), and higher serum SPARC, OGN, and CCN2. · Spatiotemporal evolution of the COMPhi…; Figure legends

Methods, data and reproducibility

Paired pre- and 6-month post-treatment FFPE synovial biopsies from 17 early RA patients (triple DMARDs or adalimumab) were profiled by 10x Xenium 5,101-gene spatial transcriptomics (Seurat v5, Harmony, Tessera), plus a 50-gene Xenium panel on 4 RA synovia and HUVEC-fibroblast co-cultures. Integrated PEAC RNA-seq (n=69), AMP scRNA-seq (n=82), Olink proteomics, RBPJ ChIP-seq, and DLL4/TGFBR3 perturbations.

  • Paired pre- and 6-month post-treatment FFPE synovial biopsies from treatment-naive early RA patients (396.10 study, triple DMARDs or adalimumab) profiled by 10x Xenium 5,101-gene panel · human · n = 17 patients (16 baseline samples after QC, 17 paired pre/post samples)
  • Additional RA synovial biopsies and HUVEC-synovial fibroblast 2D co-cultures profiled with a 50-gene custom Xenium fibrogenic panel · human · n = 4 RA synovial samples; co-culture of 120,000 fibroblasts and 60,000 HUVECs
  • PEAC cohort treatment-naive RA synovial bulk RNA-sequencing dataset · human · n = 69 patients (53 responders and 16 non-responders at 6 months)
  • AMP RA/SLE synovial single-cell RNA-seq dataset and ~5,400-plex Olink serum proteomics dataset · human · n = 82 patients (scRNA-seq); Olink: 7 CTAP-EFM vs 60 other CTAPs
  • Public HepG2 RBPJ ChIP-seq dataset used to assess promoter binding at TGFBR2 and TGFBR3 · human · ENCSR596FEL
Methods & tools
10X Genomics Xenium Analyzer, Xenium Onboard Analysis, Xenium Explorer, Seurat, Harmony, Tessera, UCell, DESeq2, OlinkAnalyze, limma, Cellpose, MACS2
Reagents & environment
Seurat v5.0.0, Xenium Explorer v3.2, FlowJo version 10, GraphPad Prism version 10.4.1, R version 4.4.1/4.4.2
Reproducibility gaps
Public ChIP-seq data is cited by bare accession ENCSR596FEL without an http(s) URL; no accession numbers or repository URLs are provided for the newly generated Xenium spatial transcriptomics datasets or analysis code.
Limitations (3)
  • The longitudinal spatial transcriptomics cohort was small (17 paired biopsies, with 16 passing baseline quality control), requiring validation in larger longitudinal cohorts. · stated by the authors
  • In baseline synovial tissue vascular niches, Notch activation scores negatively correlated with TGFBR3 expression but did not significantly correlate with COMP expression, potentially due to heterogeneity across vascular subtypes. · stated by the authors
  • Differential serum protein comparisons in the Olink dataset (7 CTAP-EFM vs 60 other patients) reported unadjusted p-values, and statistical comparisons did not formally test for normality. · stated by the authors

Who should read:Read this if you study rheumatoid arthritis treatment resistance, synovial fibroblast heterogeneity, or spatial transcriptomics of perivascular Notch-TGF-beta signaling.

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Related papers

Similarity Map
SelectedSame fieldCross-field
THISPresti '25Županič '26Yoshihara '26Khedgikar '26Mueller '25Diaz '26
72% semantic matchGenomics
Sonia Presti, Ksenia S. Anufrieva, Ce Gao+12 more·bioRxiv·Dec 12, 2025· 2 citations