Page 84 - SRPSKO DRUŠTVO ISTRAŽIVAČA RAKA

Serbian Association for Cancer Research SDIRSACR

specific CAR T cells using trastuzumab-derived single-chain variable fragments (scFvs), enabling T cells to mimic and
extend the targeting capabilities of the antibody. These engineered T cells exhibit cytotoxicity against HER2-positive
tumor cells, even in settings where trastuzumab is ineffective. Crucially, our studies demonstrate that CAR T cells can
penetrate deep into tumor spheroids and eradicate cells in areas poorly accessible to antibodies, offering a potential
solution to the spatial limitations of passive immunotherapy.
Another critical factor influencing the success of CAR T cell therapy in solid tumors is the composition and functionality
of the infused cell product. Our preclinical work has highlighted the essential role of CD4⁺ T cells in CAR-mediated
tumor eradication. These helper cells contribute not only to the orchestration of immune responses but also possess
direct cytolytic potential when engineered with HER2-specific CARs. The maintenance of a CD4⁺ population within the
CAR T cell product was crucial for sustained antitumor activity, improved in vivo persistence, and reduced exhaustion
in solid tumor models. This finding supports a strategy that deliberately preserves and expands CD4⁺ T cells during CAR
T cell manufacturing to optimize therapeutic outcomes.
The intracellular signaling domain of the CAR construct is another determinant of T cell behavior, impacting activation
kinetics, persistence, and memory formation. We have conducted comparative analyses of CARs incorporating either
CD28 or 4-1BB (CD137) costimulatory domains, as well as combinations of both. CD28-based CARs typically induce rapid
effector responses but can lead to accelerated exhaustion, while 4-1BB domains support longer-term persistence and
mitochondrial fitness. Interestingly, we observed that the choice of costimulatory domain not only alters intracellular
signaling cascades but also affects the spatial organization and mobility of CARs on the T cell surface. These biophysical
differences influence how CAR T cells engage with tumor cells and may contribute to variations in killing efficiency and
durability. These findings provide a rationale for designing hybrid signaling motifs or context-specific CARs based on
tumor type and therapeutic goal.
In addition to engineering individual T cells, we have explored modular CAR T platforms that can flexibly engage different
tumor targets. One such innovation is a universal CAR system that uses a biotin-trastuzumab fusion molecule as an
adaptor between the CAR T cell and the tumor antigen. This system allows for tunable control over antigen recognition,
enabling the same CAR T cells to be redirected toward various biotinylated antibodies targeting different tumors. Using
this strategy, we demonstrated effective infiltration and destruction of large HER2-positive tumor xenografts that were
otherwise resistant to direct antibody therapies. The modularity of this approach offers significant advantages for
treating heterogeneous tumors or for switching targets in response to antigen loss variants.
Parallel to our T cell–based work, we have also investigated the use of natural killer (NK) cells as alternative effector
cells for HER2-targeted therapies. NK cells possess inherent cytotoxic potential and can be used in allogeneic settings
without risk of graft-versus-host disease, making them attractive candidates for off-the-shelf immunotherapy. We
engineered HER2-specific CAR NK cells using the NK-92 cell line and evaluated their efficacy against breast cancer cells.
Our studies revealed a novel resistance mechanism: the expression of CD44, a surface glycoprotein linked to cancer
stemness and epithelial-to-mesenchymal transition, significantly modulates the cytolytic capacity of HER2-CAR NK
cells. Tumor cells with high CD44 expression exhibited reduced susceptibility to NK cell–mediated killing, highlighting
the need to consider CD44 as a functional biomarker when designing CAR NK therapies. These results suggest that
CD44 co-targeting or tumor cell reprogramming may enhance the effectiveness of HER2-CAR NK cell–based strategies.
Taken together, our findings underscore the importance of comprehensive CAR design that considers not only antigen
specificity, but also effector cell type, costimulatory signaling, and modular targeting. The ability of CAR T cells to
penetrate and eliminate trastuzumab-resistant tumor regions, the critical contribution of CD4⁺ cells to antitumor
responses, and the interplay between costimulatory domains and receptor dynamics all represent key variables
that can be manipulated for improved outcomes. Meanwhile, CAR NK cells offer a complementary avenue for HER2
targeting, particularly in settings where T cell–based approaches are impractical or limited by toxicity.
As CAR-based therapies expand beyond hematologic cancers, a deeper understanding of their molecular, cellular, and
physical interactions within the tumor environment will be essential. Our work contributes to this evolving field by
offering mechanistic insights and preclinical validation of HER2-targeted strategies that overcome known resistance
mechanisms. We aim to advance the next generation of adoptive immune cell therapies for solid tumors through
sustained innovation and translational research, improving their efficacy, safety, and accessibility.

Acknowledgments and funding: We acknowledge the financial support from OTKA K143771 and FK132773 (the
National Research, Development, and Innovation Office, Hungary), GINOP-2.3.3-15-2016-00003 (co-financed by the
European Union and the European Regional Development Fund), the HUN-REN Hungarian Research Network, and the
University of Debrecen Program for Scientific Publication.

69