A small molecule first isolated from a medicinal orchid may have cracked one of cancer immunotherapy’s most stubborn problems: getting engineered T cells through the barricaded blood vessels that shield glioblastoma, the deadliest cancer originating in the brain. In a study published in the journal Angiogenesis, researchers report that erianin, a bibenzyl compound derived from Dendrobium, normalizes the structurally deranged vasculature that glioblastoma builds around itself, transforming an impenetrable vascular fortress into an open gateway for chimeric antigen receptor (CAR) T cells. When the compound was combined with CAR-T cells engineered to recognize the EGFRvIII mutation, a tumor-specific genetic alteration that has been pursued in multiple clinical trials, glioblastoma mouse models responded far better than to either intervention alone. The work, led by Fan Yang of Shanghai Jiao Tong University School of Medicine together with Yanqing Gong of the University of Pennsylvania, also pinpoints the precise molecular target of erianin and maps the signaling chain it disables, offering drug developers a blueprint for vascular reprogramming that could extend well beyond brain cancer.

CAR-T cell therapy has produced remarkable, sometimes curative remissions in leukemia and lymphoma, cancers whose cells circulate freely and are physically accessible to infused immune cells. Solid tumors are another matter entirely. To destroy a solid tumor, CAR-T cells must survive in the bloodstream, latch onto the vessel wall, squeeze through the endothelial barrier, migrate through hostile stromal tissue and then remain functional inside a microenvironment that the tumor has engineered to suppress them. Every step is an obstacle. Glioblastoma, the most common and aggressive primary brain tumor in adults, has seen median survival barely improve over decades despite maximal surgery, radiation and temozolomide chemotherapy, and for this disease the obstacles are exceptionally high. The brain adds further complications, including the specialized endothelial barriers of the central nervous system and an organ-level immune privilege that blunts conventional T cell responses. Clinical attempts to treat glioblastoma with CAR-T cells directed against antigens such as EGFRvIII, IL13Rα2 and HER2 have produced encouraging anecdotes but no durable breakthroughs, largely because the engineered cells fail to reach, enter and expand within the tumor in sufficient numbers. The failed traffic, a growing body of evidence suggests, begins at the tumor’s own blood vessels.

The new study starts from the question of why that infiltration fails, and the answer lies in vascular architecture. Tumors do not simply grow a blood supply; they grow a corrupted version of one. Unlike the orderly, hierarchically branched vessels of healthy tissue, tumor vessels are dilated, tortuous, hyperpermeable and unevenly perfused, a chaos fueled by overshooting vascular endothelial growth factor signaling and chronic hypoxia. The consequences cut both ways: poorly oxygenated tumor regions resist drugs and radiation, while a disorganized, anergic endothelial lining secretes too few of the adhesion molecules that circulating T cells need to exit the bloodstream and actively suppresses their transit. Analyzing human glioblastoma samples with single-cell transcriptomics, the researchers focused on a process called endothelial-to-mesenchymal transformation, or Endo-MT, in which vessel-lining endothelial cells abandon their normal identity. Driven by transcription factors such as SNAIL and SLUG, these cells lose VE-cadherin, the adhesive protein that welds neighboring endothelial cells into a continuous and selective barrier, and instead acquire motile, matrix-producing, mesenchymal traits. The analysis indicated that Endo-MT is a key mechanism behind the vascular abnormalities that keep glioblastoma profoundly immune-cold, and that reversing it could reopen the route for immunotherapy.

To find a way to reverse the process, the team ran a functional screen through a curated chemical library, searching for compounds capable of blocking Endo-MT. The molecule that stood out was erianin, a natural bibenzyl isolated from Dendrobium, a genus of orchids long prized in traditional Chinese medicine. Erianin was not a newcomer to these laboratories. Work published two decades ago by members of the same group had documented its anti-angiogenic activity in human umbilical vein endothelial cells, and later studies tied the compound to blockade of ERK1/2-regulated HIF-1α/VEGF signaling in retinal angiogenesis and to calcium/calmodulin-dependent ferroptosis in lung cancer cells. What the new study contributes is specificity and therapeutic intent. Rather than simply poisoning endothelial cells, erianin was found to push them back toward a normal, quiescent, barrier-forming state, inhibiting the Endo-MT program that glioblastoma exploits and thereby normalizing the very vessels the tumor had weaponized. That distinction matters, because indiscriminate vessel destruction with anti-angiogenic drugs has repeatedly disappointed in brain tumors, sometimes even tightening the barrier that immunotherapy needs to cross.

Identifying how erianin accomplishes this required chemoproteomic and biophysical analyses, and the answer proved to be a protein not previously associated with vascular normalization: P4HA1, the alpha subunit of collagen prolyl 4-hydroxylase 1. P4HA1 is an α-ketoglutarate-dependent dioxygenase that hydroxylates proline residues in nascent collagen chains, a chemical modification essential for collagen’s triple helix to mature and for the extracellular matrix to be properly assembled. The enzyme also plays a second, less obvious role: it stabilizes HIF1α, the master transcriptional regulator of the cellular hypoxia response, feeding forward into angiogenesis, glycolysis and invasive behavior. The researchers showed that erianin binds P4HA1 at the Arg379 residue, located inside the pocket that normally accommodates the cofactor α-ketoglutarate. By occupying that pocket, erianin disrupts the enzyme’s catalytic cycle, an interaction the authors verified through chemoproteomic target mapping and biophysical binding assays. In doing so, the study converts an enzyme better known for collagen biochemistry into a druggable switch controlling the state of the tumor endothelium.

Blocking P4HA1 triggers a cascade that runs straight through the core of the Endo-MT program. With the enzyme inhibited, HIF1α levels fall, and with them the expression of SNAIL and SLUG, the transcription factors that orchestrate the endothelial transition under hypoxic stress. Freed from that repression, endothelial cells re-establish their VE-cadherin-mediated junctions, restoring vessel integrity and converting leaky, chaotic plumbing into structured, better-perfused conduits. Simultaneously, the treated endothelium upregulates ICAM1, intercellular adhesion molecule 1, the surface ligand engaged by the integrin LFA-1 on T cells. That molecular handshake is far from decorative: the mechanical forces transmitted through LFA-1/ICAM-1 bonds are known to fine-tune T cell receptor signaling, and firm adhesion to the endothelium is the non-negotiable first step for a T cell to crawl out of a vessel and into tissue. In effect, erianin does not merely open the vascular door for incoming immune cells; it installs the handle and the welcome mat, while the re-oxygenated, better-drained tumor interior becomes a more navigable and less hostile terrain.

The functional payoff was demonstrated in glioblastoma mouse models. Animals treated with erianin showed restored endothelial architecture and markedly increased T cell infiltration into tumor tissue, historically one of the hardest outcomes to achieve in this disease. When erianin was combined with CAR-T cells engineered against EGFRvIII, the tumor-specific epidermal growth factor receptor variant that has anchored several clinical trials, the combination proved markedly more effective than either treatment alone, a result the authors describe as sensitizing glioblastoma to the engineered cells. The vascular effects also paid dividends for conventional treatment: erianin enhanced the efficacy of chemotherapy, consistent with the principle that normalized, efficiently perfused vessels deliver drugs more predictably than the leaky, interstitially pressurized vessels of untreated tumors. The strategy is consistent with earlier work from the same laboratories, which showed that targeting the kinase PAK4 could reprogram the vascular microenvironment to improve CAR-T immunotherapy for glioblastoma, and that the small molecule toosendanin could reverse macrophage-mediated immunosuppression in the disease. Together, these studies sketch a coherent doctrine: before immune cells can be supercharged, the ground they must cross has to be rebuilt.

The findings arrive as the field converges, from several directions, on the tumor vasculature as a master regulator of immunotherapy success. The concept of vascular normalization, which steers tumor vessels toward function rather than destroying them outright, was articulated by Rakesh Jain and Peter Carmeliet more than a decade ago, and clinical imaging of glioblastoma patients treated with the pan-VEGF receptor inhibitor AZD2171 demonstrated years ago that vessel normalization is achievable in the human brain, though transient. What the new study adds is a druggable entry point upstream of that process. P4HA1 had previously been implicated in hypoxic adaptation and chemoresistance in triple-negative breast cancer, in a feedback loop driving glycolysis in pancreatic cancer and in HIF1α-mediated Wnt signaling in colorectal cancer, and a recent study in Cancer Cell reported that inhibiting P4HA1 expands progenitor-like CD8-positive T cells and strengthens systemic anti-tumor immunity. The new results position the enzyme inside the endothelium’s decision machinery, linking hypoxia signaling, junctional integrity and immune-cell adhesion, and nominate the Arg379 pocket as a specific vulnerability amenable to medicinal chemistry.

The work remains preclinical, and glioblastoma research is painfully familiar with the distance between mouse models and human benefit. CAR-T cells have yet to deliver durable responses in large glioblastoma trials; EGFRvIII is expressed in only a fraction of patients and is frequently lost under therapeutic pressure; and vascular normalization is a moving target, a transient window that must be timed so immune cells arrive while the vessels, but not the tumor, have been tamed. Erianin’s pharmacokinetics, safety profile and optimal dosing in humans remain unknown, and its effects on normal vasculature will require careful scrutiny before any clinical translation. Still, the study delivers something the field has lacked: a structurally defined natural product with a validated intracellular target that converts the tumor’s vascular shield into a portal for engineered immune cells while simultaneously improving drug delivery. If the strategy can be carried into patients (whether with erianin itself or with next-generation P4HA1 inhibitors designed around the Arg379 pocket), the implications would extend far beyond glioblastoma, to the many solid tumors whose vessels stand between CAR-T cells and their prey.