Single-cell and RNA sequencing identify mechanisms of primary non-response and efficacy biomarkers for vedolizumab-treated ulcerative colitis

To the Editor: The global incidence and prevalence of ulcerative colitis (UC) are steadily increasing.[1] Despite significant therapeutic advancements brought by biologics and small-molecule drugs, challenges such as primary non-response and secondary loss of response remain prevalent. Vedolizumab (VDZ) is currently the only gut-selective biologic approved for the treatment of moderate to severe UC, and demonstrated a favorable efficacy and safety profile. However, approximately 20–40% of patients exhibit primary non-response (PNR) by week 14 of therapy.[2] Previous studies suggest that factors such as pharmacokinetics, serum drug concentrations, and immunogenicity alone are insufficient to explain the mechanisms underlying PNR.[3,4] This study aimed to investigate the cellular mechanisms of PNR to VDZ in UC using single-cell RNA sequencing (scRNA-seq) and identify predictive biomarkers that could guide treatment efficacy. This study was approved by the Ethics Committee of The First Affiliated Hospital of Kunming Medical University (Approval No. 2023L168) and conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants before enrollment. A schematic overview of the study design is provided in Supplementary Figure 1, https://links.lww.com/CM9/C777. We retrospectively analyzed the efficacy of VDZ in 59 patients with UC treated between June 2020 and October 2023, with baseline patient characteristics summarized in Supplementary Table 1, https://links.lww.com/CM9/C778. Among them, 18 subjects were selected for scRNA-seq using sigmoid colon tissue samples. These included five healthy controls (HC, defined as individuals undergoing routine medical examinations with no systemic disease and normal colonoscopy findings), five patients with moderate UC treated only with mesalazine (NP, mayo endoscopic score [MES] = 2), four primary non-responders after 14 weeks of treatment with VDZ (PNR, MES = 3), and four patients who achieved early mucosal healing (EMH, MES = 0). Detailed clinical features of the included patients are provided in Supplementary Table 2, https://links.lww.com/CM9/C779. RNA sequencing (RNA-seq) was performed on formalin-fixed paraffin-embedded (FFPE) intestinal tissues collected before treatment with VDZ, including samples from all PNR and EMH patients used in scRNA-seq analysis. By integrating scRNA-seq and RNA-seq data, we aimed to identify predictive biomarkers of VDZ response in UC. These candidate markers were subsequently validated using an external publicly available dataset (GSE73661). Comprehensive descriptions of experimental procedures, materials, and statistical analyses are provided in the Supplementary Materials [https://links.lww.com/CM9/C777] and Supplementary Table 3–5 [https://links.lww.com/CM9/C778]. The analysis of VDZ efficacy revealed that by week 14, 68% (40/59) of patients achieved a clinical response, 46% (27/59) achieved clinical remission, and 37% (22/59) attained steroid-free clinical remission. Of the total cohort, 30 patients completed 52 weeks of VDZ treatment, with 60% (18/30) achieving clinical remission and steroid-free remission [Supplementary Figure 2A, https://links.lww.com/CM9/C777]. Among the patients who attained clinical remission, five were concurrently receiving glucocorticoid therapy. Partial Mayo scores decreased significantly from a median of 7.0 (6.0–9.0) to 2.0 (2.0–3.0) at week 14, and from 8.0 (6.0–10.0) to 1.0 (0–2.0) at week 52 [Supplementary Figure 2B, https://links.lww.com/CM9/C777]. Among the 43 patients who underwent follow-up endoscopy at week 14, 51% (22/43) achieved endoscopic mucosal remission, while 28% (12/43) achieved mucosal healing. Of the 26 patients who underwent endoscopy at week 52, 58% (15/26) achieved mucosal remission and 39% (10/26) achieved mucosal healing [Supplementary Figure 2C, https://links.lww.com/CM9/C777], accompanied by a notable reduction in MES scores [Supplementary Figure 2D, https://links.lww.com/CM9/C777]. Furthermore, both the clinical response and remission rates at weeks 14 and 52 were significantly higher in biologic-naïve UC patients compared to those previously treated with biologics [Supplementary Figure 2E, F, https://links.lww.com/CM9/C777]. Laboratory analyses showed that patients who achieved clinical response exhibited significantly reduced levels of inflammatory markers, including C-reactive protein and erythrocyte sedimentation rate [Supplementary Figure 2G–J, https://links.lww.com/CM9/C777], and showed correction of hypoproteinemia and anemia [Supplementary Figure 2K, L, https://links.lww.com/CM9/C777]. Finally, multivariable logistic regression analysis of baseline clinical features identified prior immunosuppressive therapy as an independent protective factor, while concurrent glucocorticoid use during VDZ treatment was found to be an independent risk factor for clinical response at week 14 [Supplementary Table 6, https://links.lww.com/CM9/C778]. To investigate the potential mechanisms underlying primary non-response to VDZ in UC, we performed 10× Genomics scRNA-seq on intestinal tissues from a total of 18 individuals, including HC, NP, PNR, and EMH, thereby constructing a comprehensive single-cell atlas of the intestinal microenvironment across different disease states. Following stringent quality control procedures, a total of 108,641 high-quality cells were obtained for the next analysis. Annotation based on canonical marker gene expression identified eight major cell types [Supplementary Figure 3A–C, https://links.lww.com/CM9/C777]. Proportional analysis revealed a marked depletion of epithelial cells in patients with UC, accompanied by predominant infiltration of B cells and plasma cells in individuals with active disease [Supplementary Figure 3D, https://links.lww.com/CM9/C777]. Subclustering analysis of seven major cell types, excluding neurons, identified 49 annotated cell subtypes [Supplementary Figure 4A, B, https://links.lww.com/CM9/C777]. Proportional analysis of T lymphocyte subpopulations—the primary cellular targets of VDZ—revealed significant alterations in T cell composition among patients with UC. Notably, the proportions of CD4+ and CD8+ T lymphocyte subsets in patients who achieved EMH displayed marked differences compared to HC [Supplementary Figure 4C, https://links.lww.com/CM9/C777]. Differential gene expression analysis further demonstrated that CD4_ITGAE_Trm, CD8_IFNG_Cytotoxic, and CD8_KLRB1_MAIT cell subsets were specifically enriched in the EMH group and exhibited high expression levels of cytotoxic effector genes, including IFNG, NFKBIZ, GZMA, GZMB, and GZMK [Supplementary Figure 4D, E, https://links.lww.com/CM9/C777]. Given VDZ’s therapeutic mechanism of selectively targeting integrin α4β7, we evaluated α4β7+ cell counts in the intestinal mucosa of HC and UC patients across different disease states. Interestingly, the numbers of α4β7+ CD4+ T cells, CD8+ T cells, B cells, and plasma cells were significantly reduced in patients from the PNR group compared to those in the NP, EMH, and HC groups [Supplementary Figure 5A, https://links.lww.com/CM9/C777]. Paradoxically, results from a separate cohort using immunofluorescence (IF), Western blotting, and quantitative reverse transcription polymerase chain reaction (RT-qPCR) showed significantly elevated α4β7-positive cell counts, as well as increased α4β7 protein and mRNA expression levels in the intestinal mucosa of both NP and PNR groups compared to the HC and EMH groups [Supplementary Figure 5B–F, https://links.lww.com/CM9/C777]. Previous studies, along with our single-cell transcriptomic data, indicate that the current therapeutic dose of VDZ nearly completely blocks integrin α4β7, effectively inhibiting the α4β7/Madcam-1 pathway responsible for immune cell migration to the gut. However, the persistence of intestinal inflammation despite this blockade suggests the involvement of alternative migration pathways. To investigate this, we performed CellChat analysis to map receptor-ligand interactions between T lymphocytes and endothelial cells, which revealed a significant enrichment of intercellular adhesion molecule (ICAM) and integrin subunit beta 2 (ITGB2) signaling pathways [Figure 1A, B]. Compared to the EMH group, the PNR group demonstrated notably enhanced interactions through the ICAM1-ITGAL_ITGB2, ICAM1-ITGAL, and ITGB2-ICAM1 pathways, promoting strong crosstalk between endothelial cells and both CD4_GZMA_Effector T cells and CD8_TNFRSF9_Cytotoxic T lymphocytes [Supplementary Figure 6A, B, https://links.lww.com/CM9/C777]. Supporting these results, analyses from an independent cohort using IF, Western blotting, and RT-qPCR revealed significantly increased counts of αLβ2 and αMβ2 positive cells, along with elevated protein and mRNA expression levels of ITGAL, ITCAM, ITGB2, and ICAM1 in the intestinal mucosa of the PNR group compared to the EMH group [Supplementary Figure 6C–G, https://links.lww.com/CM9/C777]. Further subpopulation analysis confirmed upregulation of αLβ2 and αMβ2 within CD4_GZMA_Effector T cells and CD8_TNFRSF9_Cytotoxic T cells in the PNR group [Supplementary Figure 6H, https://links.lww.com/CM9/C777], collectively suggesting that β2 integrin/ICAM1-mediated immune cell trafficking may be a key mechanism contributing to sustained intestinal inflammation in UC patients with PNR to VDZ.Figure 1: Investigation of primary non-response mechanisms and predictive biomarkers for VDZ efficacy in UC patients. (A) Chord diagrams illustrating ITGB2-associated signaling network interactions among cell types in intestinal mucosal tissues from HC, NP, PNR, and EMH groups. The thickness of the chords represents the intensity of cellular interactions. (B) Chord diagrams depicting ICAM-associated signaling networks among cell types in the same four groups, with interaction intensity indicated by line thickness. (C) Expression scores of the top 15 marker genes for Vein_iii endothelial cells in the RNA-seq data from the current study. (D) ROC curve analysis evaluating the predictive value of Vein_iii endothelial cells in current study dataset. (E) ssGSEA-based expression scores of nine DEGs identified between Pre_PNR and Pre_EMH groups in the current RNA-seq dataset. (F) ROC curve of all nine genes in the current RNA-seq dataset. AUC: Area under the curve; ROC: Receiver operating characteristic; VDZ: Vedolizumab; UC: Ulcerative colitis; ITGB2: Integrin subunit beta 2; PNR: Primary non-response; EMH: Early mucosal healing; ICAM: Intercellular adhesion molecule; DEG: Differentially expressed gene.To identify predictive biomarkers of VDZ efficacy, FFPE intestinal mucosal tissues from UC patients classified as PNR and EMH were collected prior to VDZ treatment and subjected to bulk RNA-seq and multiplex immunohistochemistry (mIHC). Using the top 15 marker genes from 49 annotated cell subpopulations, single-sample gene set enrichment analysis (ssGSEA) revealed that the Vein_iii endothelial cell cluster score—indicative of cellular abundance—was significantly higher in the Pre_PNR group compared to the Pre_EMH group [Figure 1C, and Supplementary Figure 7A, https://links.lww.com/CM9/C777]. Receiver operating characteristic (ROC) analyses further demonstrated the predictive value of Vein_iii endothelial cells for VDZ treatment response, yielding area under the curve (AUC) values of 0.80 in the internal transcriptomic dataset and 0.65 in a public external dataset [Figure 1D, and Supplementary Figure 7B, https://links.lww.com/CM9/C777]. mIHC validation in an independent cohort confirmed elevated counts of both total Vein_iii cells and ICAM1+ Vein_iii cells in the Pre_PNR group relative to the Pre_EMH group [Supplementary Figure 7C, D, https://links.lww.com/CM9/C777]. Cross-comparison of differentially expressed genes (DEGs) between Pre_PNR and Pre_EMH groups identified nine shared genes and all upregulated in the Pre_PNR group [Supplementary Figure 7E–G, https://links.lww.com/CM9/C777]. ssGSEA and ROC analyses showed that nine-gene signature possessed potential predictive capability for VDZ responsiveness, with AUCs of 0.99 in the internal dataset and 0.67 in the external dataset [Figure 1E, F, and Supplementary Figure 7H, I, https://links.lww.com/CM9/C777] Notably, all nine individual genes exhibited AUCs exceeding 0.7 [Supplementary Figure 7J, https://links.lww.com/CM9/C777]. IF further validated, significantly elevated expression of these nine genes in the intestinal mucosa of Pre_PNR patients compared to Pre_EMH patients [Supplementary Figure 7K, https://links.lww.com/CM9/C777], collectively supporting their potential utility as predictive biomarkers for VDZ treatment outcomes. Our findings highlight substantial inter-individual variability in its expression. Enhanced intestinal trafficking of immune cells mediated by β2-integrin/ICAM1 interactions, as a potential driver of PNR in UC patients receiving VDZ. This study preliminarily evaluates the expression of the ICAM1+ Vein-iii endothelial cell and nine upregulated genes—CCDC3, CD93, CDH6, EBF1, G0S2, MCTP1, MMRN1, SNX10, and TLR2—in the intestinal mucosa of patients with PNR to VDZ, suggesting their potential as biomarkers for predicting PNR and EMH. Their clinical utility warrants further validation in large-scale prospective cohorts. The current study has limitations, primarily stemming from the small sample size, which resulted in notable fluctuations in AUC values. The cross-center validation will be conducted to assess the predictive power and generalizability of the proposed biomarkers, thereby contributing to the advancement of precision medicine in UC. Funding This study was supported by grants from the National Natural Science Foundation of China (Nos. 82170550, 82160107, 82260108, and 82360107), Applied Basic Research Foundation of Yunnan Province (Nos. 202301AT070131, 202201AY070001-048, and 202301AT070091, 202301AS070027), Youth Talents Program of Yunnan Province (No. RLQB20220004), 535 Talent Project of First Affiliated Hospital of Kunming Medical University (Nos. 2023535D06 and 2023533Q03), Special grant for research on the application of single-cell sequencing in inflammatory bowel disease (No. 7012204), and First-Class Discipline Team of Skin & Mucosal Regenerative Medicine of Kunming Medical University (No. 2024XKTDTS10). Conflicts of interest None.