Preclinical evaluation of human T cell–based cancer immunotherapies requires test systems that capture intravenous delivery, vascular recruitment, and intratumoral T cell activation. Traditional 3D in vitro systems lack perfused vasculature, while rodent models have lower throughput. The chick chorioallantoic membrane (CAM) assay supports rapid growth of vascularized human tumor xenografts in ovo, but its native avian vasculature and species-specific endothelial interactions limit the recruitment and functional study of human T cells.
To address these constraints, the authors established an endothelial cell graft-enhanced CAM model. The core idea is to implant an immortalized murine endothelial cell line into human tumor xenografts growing on the CAM so that these grafted cells can anastomose with the host chick vessels. This produces a perfused, branched mammalian vascular compartment within the tumor microenvironment on the CAM, intended to better model intravenous delivery and immune cell trafficking.
The engineered approach relies on an immortalized murine endothelial cell line capable of forming vascular structures and connecting with the chick vasculature on the CAM. Following grafting into human tumor xenografts, these endothelial cells establish perfused and branched vessel networks that interface with the host circulation, creating a mammalian endothelial layer embedded within tumors grown in ovo.
This perfused mammalian vascular interface aims to replicate key features of tumor vasculature relevant to immune cell extravasation and delivery of systemically administered agents. The authors report that the grafted endothelial cells successfully anastomose with CAM vessels, yielding a functional vascular interface within human tumors on the CAM platform.
To overcome species-specific barriers in leukocyte-endothelial interactions, the authors engineered the grafted endothelial cells to express human ICAM-1. Human ICAM-1 is a key adhesion molecule that facilitates firm adhesion and transendothelial migration of human leukocytes, including activated T cells.
Expression of human ICAM-1 on the murine grafted endothelium was used to enhance the ability of intravenously delivered human T cells to adhere to and infiltrate CAM tumors. The strategy is intended to provide a more physiologically relevant endothelial interface for studying human immune cell recruitment in a high-throughput, in ovo tumor system.
Using the endothelial graft-enhanced platform, the authors demonstrate that intravenously administered human T cells can be recruited into CAM tumors. The presence of a perfused mammalian vascular compartment and engineered human ICAM-1 on grafted endothelial cells improves intratumoral entry of systemically delivered human T cells compared with CAM tumors relying solely on avian vasculature.
This capability addresses a key limitation of conventional CAM experiments and enables investigations of human T cell trafficking, localization, and early activation within tumor xenografts grown in ovo. The model thus supports studies that require systemic cell delivery rather than local cell injection.
The platform was applied to evaluate intratumoral, target-dependent activation of human T cells by bispecific T cell engagers (TCEs). The authors used the system to test TCE activity across different tumor models grown on the CAM.
As an example of translational relevance, the study reports evaluation of the clinically approved DLL3-targeting TCE tarlatamab in small cell lung cancer models on the graft-enhanced CAM. Using this setup, intratumoral T cell activation in response to TCE treatment could be assessed in the context of intravenously delivered human T cells and a perfused mammalian endothelial interface.
This endothelial graft-enhanced CAM tumor model combines the throughput and rapidity of the CAM assay with a perfused mammalian endothelial compartment designed to permit intravenous delivery and recruitment of human immune cells. It offers a platform for testing immune cell–based therapies, antibody formats that require systemic delivery, and early-stage pharmacodynamic readouts of intratumoral T cell activation.
The report is a preprint and has not undergone peer review; therefore, further validation, replication of quantitative results, and detailed methodological review are advisable prior to wide adoption. Specific experimental parameters, quantitative outcomes, and protocol details are reported in the source preprint and should be consulted directly for implementation.
The work is reported as a preprint posted July 17, 2026, on bioRxiv and has not been certified by peer review. Authors disclose a PCT filing related to the approach and report multiple competing interests, including advisory roles and industry relationships noted in the source. Funding sources include several German and European grants and institutional programs listed in the preprint.
Readers seeking to implement or adapt this model should review the original preprint for full experimental detail, validation data, and any supplementary materials provided by the authors.