T cells to fix a broken heart

Fibrosis that results from excessive extracellular matrix protein deposition by activated cardiac fibroblasts is a hallmark of heart disease and plays a critical role in disease progression to heart failure (1). However, therapies targeting this cardiac fibrosis remain limited. On page 91 of this issue, Rurik et al. (2) describe a new approach to eliminate activated fibroblasts by harnessing the power of engineered T cells. Lipid nanoparticles (LNPs) carrying messenger RNA (mRNA) that encodes a chimeric antigen receptor (CAR) are used to generate CAR T cells in mice, yielding a therapeutic T cell population that is capable of ablating pathogenically activated fibroblasts and attenuating cardiac fibrosis.

CARs are synthetic receptors that allow immune cells—usually T cells—to recognize targeted antigens and initiate antigen-specific immune responses (3). In conventional CAR T cell therapy, T cells are isolated from the blood of patients or healthy donors, genetically modified to express a CAR targeting a disease-associated antigen (e.g., a surface protein found on tumor cells), and expanded ex vivo before infusion into the patient. CAR T cell therapy has shown substantial clinical efficacy against various hematological malignancies, becoming the first genetically modified cell therapy to receive US Food and Drug Administration approval (4). Following these successes, efforts are underway to expand CAR T cell therapy to other indications, such as infectious and autoimmune diseases (5).

Previously, CAR T cells targeting fibroblast activation protein (FAP) were shown to specifically target activated fibroblasts and significantly reduce cardiac fibrosis in a mouse model of hypertensive cardiac injury (6). In this prior work, T cells were stably integrated ex vivo with the CAR-encoding transgene through retroviral transduction, resulting in long-term CAR expression. Extensive clinical experience in the context of cancer therapy has shown that autologous CAR T cells can persist for months or years after adoptive transfer into the patient, and this long-term persistence plays an important role in therapeutic efficacy and durability of response (7). However, cardiac injury has a different temporal profile compared with cancer. Although FAP is generally expressed at low levels in healthy tissue, it is up-regulated during the normal wound-healing process, which involves fibroblast activation (8). As such, long-term persistence of FAP-targeting CAR T cells could present a safety risk if the patient sustains injury after treatment.

To balance the need for robust CAR T cell activity against pathogenic fibroblasts while minimizing long-term safety risks, Rurik et al. developed a method to generate transient CAR T cells in vivo, bypassing the need for ex vivo cell manufacturing (see the figure). In this approach, mRNA that is chemically modified for stability and encodes the FAPCAR is packaged into LNPs decorated with CD5-targeting antibodies to facilitate selective uptake by T cells. Upon LNP uptake, the mRNA is translated to generate CAR T cells. Given the inherent instability of mRNA, CAR expression is transient, and no T cell will remain persistently CAR⁺ by design. Rurik et al. confirmed the emergence of anti-FAP CAR T cells in mice 24 hours after LNP injection, as well as the decline of these CAR T cells to undetectable levels within 7 days. Using a mouse model of hypertensive cardiac injury and fibrosis, the authors demonstrated that transient FAPCAR T cells substantially reduce ventricular fibrosis and improve various cardiac functions, as assessed by echocardiography. These findings provide a rationale for the use of transient CAR T cells and demonstrate the applicability of CAR T cell therapy beyond oncology applications.

Heart disease remains the leading cause of death globally, requiring therapies that can be manufactured at large scale and at reasonable cost. The success of the COVID-19 mRNA-LNP vaccines (9) suggests that LNP-based mRNA delivery could be a viable path to the development of scalable immunotherapies, providing a more economical alternative to the conventional method of ex vivo CAR T cell manufacturing that requires extensive infrastructure and high-cost reagents (10). In vivo CAR T cell generation had previously been reported with the use of adeno-associated viruses instead of mRNA-LNPs (11). A potential advantage of the mRNA-LNP platform is the ability to flexibly decorate the LNP with antibodies to increase target specificity, as well as greater ease of clinical-grade reagent production. Furthermore, compared with conventional ex vivo CAR T cell manufacturing methods, the virus-free, in vivo CAR T cell manufacturing method detailed by Rurik et al. eliminates the risk of unintended host-genome alterations mediated by lentiviral or retroviral integration (12) as well as the need for lymphodepleting chemotherapy before adoptive transfer, thus removing a source of considerable toxicity that is typically associated with CAR T cell therapies.

FAP-targeting CAR T cells were previously evaluated, with some groups reporting minimal side effects (13) and others reporting lethal bone toxicity and cachexia (muscle wasting) in mice due to interactions with bone marrow stem cells (14). No overt toxicity was observed by Rurik et al., and LNP-mediated anti-FAP CAR expression was largely restricted to the spleen. Nevertheless, these discrepancies in toxicity must be carefully evaluated before translation into the clinic, and a more in-depth characterization of the interaction between FAPCAR T cells and activated fibroblasts within the fibrotic tissue will be necessary to demonstrate antigen-specific CAR T cell activity at the disease location. Additionally, although Rurik et al. demonstrated that interstitial fibrosis was largely reduced with mRNA-LNP treatment, it was noted that perivascular fibrosis persisted beyond the treatment duration owing to the presence of activated fibroblasts that do not express FAP. Perivascular fibrosis and associated inflammation can decrease the contact area of neighboring cardiac tissue with blood vessels and reduce oxygen and nutrient availability that may exacerbate pathological cardiac remodeling (1). Further investigation into the biology of perivascular fibrosis could uncover additional therapeutic targets.

The work of Rurik et al. provides a strong rationale for the broadening of immunotherapies into disease areas with unmet needs, and FAPCAR T cells have already been evaluated in the context of cancer (15). Other fibrotic diseases or associated disorders, such as chronic inflammatory diseases with FAP⁺ fibroblasts, may benefit from this approach as well. This work represents an exciting step toward translating personalized immunotherapies into accessible and affordable “off-the-shelf” immunotherapies.