UCSF and Berkeley Scientists Engineer CAR-T Cells Directly Inside the Body, Clearing Tumors in Mice With a Single Injection
A two-particle gene-editing system published in Nature creates functional CAR-T cells in vivo, eliminating the costly weeks-long manufacturing process and clearing leukemia, myeloma, and solid tumors in preclinical models.
Overview
Researchers at the University of California, San Francisco and the University of California, Berkeley have demonstrated a method to generate chimeric antigen receptor T cells — a potent class of cancer immunotherapy — directly inside the body, bypassing the expensive and time-consuming process of extracting, engineering, and reinfusing a patient’s own immune cells. The findings, published in Nature on March 18, 2026, showed that a single injection of gene-editing particles cleared aggressive cancers in nearly all treated mice within two weeks.
The advance addresses one of the central bottlenecks in CAR-T therapy: the current manufacturing pipeline requires removing a patient’s T cells, genetically modifying them in specialized facilities over several weeks, and infusing them back at a cost that typically ranges from $400,000 to $500,000 per treatment. By performing the genetic engineering in vivo, the researchers aim to make these therapies faster, cheaper, and accessible beyond major academic medical centers.
How It Works
The system relies on two types of nanoparticles administered together. The first particle, an enveloped delivery vehicle, carries CRISPR-Cas9 gene-editing machinery and is coated with antibodies that direct it specifically to T cells circulating in the bloodstream. The second particle, an adeno-associated virus, delivers the DNA encoding a cancer-targeting CAR transgene. Critically, the CAR DNA inserts at the TRAC locus — the site of the T cell receptor alpha gene — placing its expression under the cell’s natural promoter rather than a synthetic one. This targeted integration means only T cells activate the new gene, and expression is regulated physiologically rather than constitutively.
Justin Eyquem, an associate professor of medicine at UCSF and senior author on the study, noted that the dual-particle design can “practically eliminate the chances of accidentally editing the wrong cell” or inserting DNA at an unintended genome location.
Preclinical Results
In mice bearing humanized immune systems, a single injection of the two-particle system produced TRAC-CAR T cells representing approximately 20 percent of splenic T cells, with CAR-T cells comprising up to 40 percent of immune cells in some organs. The engineered cells cleared all detectable cancer in nearly all treated animals within two weeks across three cancer models: acute leukemia, multiple myeloma, and sarcoma.
Notably, T cells programmed in vivo exhibited what the researchers described as greater “stemness” and proliferative capacity compared to cells manufactured using conventional ex vivo methods. The Nature study attributed this to the physiological regulation conferred by TRAC-locus integration, which avoids the exhaustion-promoting effects of constitutive CAR expression seen with standard viral vector approaches.
Path to Clinical Translation
Eyquem and his collaborators, including Nobel laureate Jennifer Doudna of UC Berkeley, have founded Azalea Therapeutics to advance the platform toward human trials. The company announced it is advancing TRAC-targeted in vivo CAR-T programs toward IND-enabling studies, a regulatory prerequisite before testing in patients. Jenny Hamilton serves as president and CEO of the company.
The work was a multi-institutional collaboration involving UCSF, the Innovative Genomics Institute, Gladstone Institutes, Duke University, and Karolinska Institutet, with William A. Nyberg of Karolinska serving as first author on the Nature publication.
What Remains Unknown
While the preclinical data is striking, translating in vivo gene editing from mice to humans introduces substantial unknowns. The human immune system may respond differently to the nanoparticle delivery vehicles, and the long-term safety of site-specific CRISPR editing in circulating T cells remains untested in clinical settings. The timeline from IND-enabling studies to a first-in-human trial has not been disclosed by Azalea Therapeutics. If the approach proves safe and effective in humans, it could fundamentally change the economics and accessibility of CAR-T therapy, potentially bringing a treatment currently limited to specialized centers to community hospitals worldwide.