Single-Cell Map Reveals Vascular Barrier That Predicts BCG Failure in Bladder Cancer

Bladder cancer that has invaded the muscle layer of the bladder wall is one of the most consequential turning points in urologic oncology, but the stage just before that invasion, known as T1 disease, has long remained a biological black box. Patients diagnosed with T1 tumors face an uncertain future: some respond beautifully to intravesical BCG therapy, the live attenuated tuberculosis vaccine that is instilled directly into the bladder and has been the standard of care for decades, while others progress relentlessly to muscle-invasive disease despite repeated courses of treatment. Clinicians have had no reliable way to predict which path a given tumor will take, and the microenvironmental basis of that divergence has remained poorly defined. Now, a team of researchers at Tianjin Medical University and collaborating institutions has produced a stage-spanning single-cell atlas of human bladder cancer that identifies T1 as a distinct vascular-immune transition state, and in doing so has uncovered a mechanistic explanation for why BCG works for some patients and fails for others.

The study, published in the Journal of Translational Medicine, combined single-cell RNA sequencing of human bladder cancer specimens across disease stages with tissue-level validation and an orthotopic rat model in which bladder tumors were grown in their natural anatomical location. This stage-spanning design was critical. Rather than sampling tumors at a single point in their evolution, the investigators mapped how the cellular composition of the tumor microenvironment changes as disease progresses from earlier stages through T1 and beyond. The approach allowed them to catch, in molecular detail, the precise moment when the tumor begins to remodel its blood supply and reorganize its immune landscape, a transition that their data place squarely at the T1 boundary.

The central discovery concerns a specialized population of blood vessel cells. The researchers found that T1 tumors are characterized by a marked expansion of endothelial cells expressing the chemokine receptor CXCR4, so-called tip endothelial cells that lead the forward advance of growing blood vessels during angiogenesis. These CXCR4-positive tip cells displayed strong angiogenic activity, driving the formation of new vasculature that feeds the expanding tumor. Crucially, however, they showed low expression of adhesion molecules, the molecular tethers that normally allow immune cells traveling in the bloodstream to grab onto vessel walls and squeeze through into surrounding tissue. The result is a vascularly restrictive niche: blood vessels that are metabolically active and proliferative but functionally closed to T-cell traffic, creating a barrier that keeps cytotoxic immune cells out of the tumor bed precisely when the immune system needs access most.

The single-cell data revealed that this vascular barrier does not arise in isolation. Surrounding the CXCR4-positive endothelium, the investigators identified metabolically activated pericytes, the contractile support cells that wrap around blood vessels and regulate their maturation and function. Alongside them, the atlas documented an expansion of macrophages characterized by expression of SPP1, the gene encoding osteopontin, a secreted protein associated with tissue remodeling and immunosuppression. A particularly important subpopulation of these macrophages, marked by expression of MT1X and associated with hypoxic conditions within the tumor, was found to produce vascular endothelial growth factor, or VEGF, the master signaling molecule of angiogenesis. In other words, oxygen-starved regions of the T1 tumor recruit and reprogram macrophages, which in turn secrete VEGF to sustain the very CXCR4-positive tip endothelial cells that wall off the tumor from immune attack. The vascular niche and the immune exclusion it produces are thus two faces of a single, self-reinforcing circuit.

At the same time that this restrictive vasculature was emerging, the atlas captured the beginnings of an organized adaptive immune response within T1 tumors. The researchers observed the appearance of CXCL13-positive T cells, a subset of T helper cells known to orchestrate the recruitment and organization of B cells, together with B-cell gene expression programs characteristic of tertiary lymphoid structures, or TLSs. These are ectopic, lymph-node-like aggregates that form within tissues at sites of chronic inflammation and are widely regarded as favorable prognostic features in many cancers, because they provide local sites where anti-tumor immune responses can be initiated and sustained. Their emergence at the T1 stage suggested that the immune system is attempting to mount a structured response against the tumor even as the vasculature works to exclude it, setting up a competition between two opposing biological programs within the same tissue.

When the investigators compared tumors from patients who responded to BCG therapy with those from patients who did not, the outcome of that competition proved decisive. Mature tertiary lymphoid structures were preferentially enriched in BCG responders, indicating that the ability to build fully organized immune aggregates within the bladder wall correlates with successful response to the vaccine. Conversely, CXCR4-positive tip endothelial cells were enriched in non-responders, consistent with the idea that a vascularly restrictive niche prevents the intravesically administered BCG from generating an effective anti-tumor immune infiltrate. The two cell populations thus function as opposing biomarkers: one signaling an immune microenvironment capable of supporting BCG-induced tumor killing, the other signaling a physical and immunological barrier that the therapy cannot penetrate.

The mechanistic implications of these findings were tested directly in the orthotopic rat model. Because the MT1X-positive, hypoxia-associated macrophages were identified as the source of VEGF driving the CXCR4-positive endothelial program, the researchers reasoned that pharmacologically blocking VEGF-family signaling might dismantle the vascular barrier and restore immune access to the tumor. They treated tumor-bearing animals with sunitinib, a multi-targeted tyrosine kinase inhibitor that blocks both VEGF receptors and PDGF receptors, thereby targeting both the endothelial compartment and the pericyte support cells that stabilize tumor vessels. The results supported the hypothesis: sunitinib enhanced the efficacy of BCG therapy in vivo, providing proof of principle that the vascular-immune niche identified in the human atlas is not merely a correlate of treatment failure but a functional, targetable determinant of response.

The clinical significance of this work is considerable. BCG has been the backbone of treatment for non-muscle-invasive bladder cancer since the late twentieth century, yet a substantial fraction of patients either fail to respond initially or lose response over time, and the only definitive option for BCG-unresponsive disease has historically been radical cystectomy, a major surgical procedure with significant morbidity. A molecular signature capable of stratifying patients at the T1 stage, before months of BCG courses have been completed and precious time has been lost, could allow clinicians to identify likely non-responders early and direct them toward alternative strategies, whether that means combining BCG with vascular-targeting agents, enrolling them in trials of novel intravesical or systemic immunotherapies, or considering earlier surgical intervention. The identification of CXCR4-positive tip endothelial cells as enriched in non-responders and mature TLSs as enriched in responders offers exactly such a stratification framework, grounded in specific, measurable cellular features of the tumor microenvironment.

Beyond its immediate translational promise, the study contributes a conceptual advance to tumor immunology. It reframes the transition from non-invasive to T1 bladder cancer not simply as a matter of deeper tumor penetration but as a coordinated remodeling event in which angiogenic endothelium, pericytes, hypoxic macrophages, and emerging lymphoid structures are locked into a dynamic balance that determines whether immunotherapy can succeed. The vascular-immune crosstalk documented in this atlas, in which VEGF from hypoxia-associated macrophages builds a vessel wall that excludes T cells while CXCL13-positive T cells labor to assemble tertiary lymphoid structures on the other side, provides a mechanistic template that may be relevant well beyond the bladder. The authors note that their atlas defines a T1-stage vascular-immunosuppressive niche linking vascular remodeling, immune accessibility, TLS maturation, and BCG responsiveness, and the demonstration that sunitinib can potentiate BCG in vivo suggests a concrete path toward overcoming BCG resistance. If validated in prospective human studies, the strategy of priming the tumor vasculature before or during BCG instillation could transform the management of one of the most common and treatment-resistant forms of bladder cancer, turning a vascular barrier that has silently sabotaged immunotherapy into a therapeutic target in its own right.

Subject of Research: Single-cell mapping of the T1-stage vascular-immune microenvironment and its role in BCG therapy response in bladder cancer

Article Title: Stage-spanning single-cell mapping reveals a T1 vascular–immune barrier linked to BCG response in bladder cancer

Article References: Song, S., Shao, Y., Wei, Y., Li, P., Yang, Z., Li, S., Zhang, H., Lu, Y., Fang, C., Wang, C., Liu, T., Yang, B., An, H., Liu, L., Tian, J., Hu, H., & Shang, Z. (2026). Stage-spanning single-cell mapping reveals a T1 vascular–immune barrier linked to BCG response in bladder cancer. Journal of Translational Medicine. https://doi.org/10.1186/s12967-026-09023-y

Image Credits: AI Generated

DOI: 10.1186/s12967-026-09023-y

Keywords: bladder cancer, BCG therapy, single-cell RNA sequencing, tumor microenvironment, angiogenesis, CXCR4, tip endothelial cells, tertiary lymphoid structures, VEGF, sunitinib, macrophages, immunotherapy resistance