Beyond CAR-T Cells: exploring CAR-NK, CAR-M, and CAR-γδ T strategies in solid tumor immunotherapy

1. Introduction

Cancer stands as a prominent cause of mortality globally. Traditional modalities such as surgery, radiation therapy, and chemotherapy constitute the cornerstone of treating early to mid-stage cancers (1). However, the substantial adverse effects associated with chemotherapy and radiation have led to the emergence of immunotherapy as a promising alternative for various cancer types. This approach leverages the immune system’s capacity to zero in on cancer cells (2). Tumor cells evade immune surveillance through various mechanisms, leading to uncontrolled proliferation driven by factors such as genetic mutations, chemical carcinogens, physical damage, and viral infections. A major challenge in cancer therapy is minimizing the collateral damage to healthy tissues, which often occurs during treatments such as chemotherapy and radiation therapy. Immune cells play a critical role in cancer defense, with T cells recognizing abnormal cells by detecting antigen peptides presented by major histocompatibility complex I (MHC-I). However, tumor cells can evade immune surveillance by interfering with the immune recognition process, such as downregulating MHC-I expression or altering antigen presentation, which contributes to their uncontrolled proliferation and metastasis (3). Antibody-antigen interactions are fundamental to immune recognition in humans. By grafting antibodies that bind to tumor cell surface receptors onto cell membranes, “smart cells” known as CAR-engineered cells can be created. These cells efficiently recognize and eliminate cancer cells. Introducing these modified cells into patients for cancer treatment is called adoptive cell therapy (4). Engineered immune cells, particularly CAR-T constructs, have emerged as an innovative strategy for tumor eradication. The conventional manufacturing protocol for CAR-T therapies commences with leukapheresis, a procedure to isolate peripheral blood mononuclear cells (PBMCs) from the patient. Following isolation, sequential processing steps are performed: cell washing to remove plasma contaminants, T-cell activation via CD3/CD28 agonists, viral/non-viral gene transfer for CAR integration, ex vivo expansion in cytokine-enriched media, and formulation into a final cryopreserved product for single-dose infusion (5–8). This labor-intensive process spans 10–20 days and necessitates multiple transfers across sterile barriers, introducing logistical complexities. Two critical limitations undermine the druggability of autologous CAR-T therapies. First, patient-specific production contradicts the pharmaceutical industry’s emphasis on standardized production protocols, complicating batch consistency and regulatory quality control. Second, while allogeneic “off-the-shelf” CAR-T strategies aim to circumvent these issues, clinical trials reveal persistent challenges, including immune rejection, limited persistence, and suboptimal efficacy in heterogeneous tumor microenvironments (9, 10). These hurdles highlight the need for innovative manufacturing frameworks to harmonize personalized efficacy with scalable production. A second critical limitation lies in the heterogeneous variable expansion and persistence of CAR-T therapies, which complicate their clinical predictability and safety. Following single-dose administration, CAR-T cells exhibit variable expansion kinetics and tissue distribution, a phenomenon termed “post-administration pharmacokinetic variability.” This unpredictability arises from patient-specific factors (e.g., T-cell fitness, tumor burden) and intrinsic CAR-T behaviors, such as activation-induced exhaustion or uncontrolled clonal proliferation (11, 12). These dynamics may lead to suboptimal therapeutic outcomes, including incomplete tumor clearance or severe adverse events such as CRS and immune effector cell-associated neurotoxicity syndrome (ICANS) (13, 14). CAR-T cell therapy has witnessed remarkable advancements in treating hematologic malignancies, yet it encounters hurdles when addressing solid tumors. The primary challenges include the complexities involved in manufacturing CAR-T cells, the scarcity of specific tumor antigens, inadequate infiltration of CAR-T cells into tumors, immune suppression within the tumor microenvironment, toxicity associated with treatment, and antigen escape mechanisms (15). To address these limitations, alternative CAR-engineered immune cells have emerged. CAR-NK cells leverage innate tumor-homing capacity and MHC-independent cytotoxicity, while CAR-Ms exploit phagocytic clearance and TME remodeling. Preclinical studies demonstrate synergistic antitumor efficacy for both modalities in solid tumors, accompanied by more favorable toxicity profiles compared to CAR-T cells (16, 17). Emerging CAR-γδ T cell therapies combine CAR-mediated antigen specificity with the intrinsic tumoricidal activity of γδ T cells, which recognize stress-induced ligands independently of MHC presentation. This dual-targeting capability enhances tumor selectivity while reducing on-target/off-tumor toxicity.

CAR-NK cell therapy involves the genetic modification of NK cells to express synthetic receptors that recognize tumor-specific antigens. Unlike T cells, NK cells mediate cytotoxicity through MHC-independent mechanisms, enabling them to target malignant cells without prior antigen sensitization—a hallmark of innate immunity (18). This intrinsic capability positions CAR-NK cells as promising candidates for solid tumor immunotherapy, offering advantages such as reduced risk of CRS and compatibility with “off-the-shelf” manufacturing (19). Preclinical studies have validated CAR-NK efficacy across diverse solid malignancies, including breast cancer, ovarian cancer (OC), pancreatic ductal adenocarcinoma, colorectal carcinoma, glioblastoma multiforme, hepatocellular carcinoma (HCC), and head and neck squamous cell carcinoma (HNSCC) (20–23). For instance, CAR-NK cells targeting human epidermal growth factor receptor 2 (HER2) in OC models demonstrated robust tumor regression via dual mechanisms: direct cytolytic activity and interferon-γ (IFN-γ)-mediated TME remodeling (21). CAR-NK cells, engineered to secrete Neo-2/15 (an IL-2Rβγ agonist), demonstrate enhanced cytotoxicity and persistence in immunosuppressive TME. They exhibit significant efficacy in pancreatic ductal adenocarcinoma (PDAC) and ovarian cancer by sustaining mitochondrial fitness via c-Myc/NRF1 activation, leading to reduced exhaustion and superior tumor control (24). Similarly, anti- mucin 1 (MUC1) CAR-NK cells exhibited potent suppression of pancreatic cancer progression in vivo, correlating with enhanced infiltration and reduced stromal barrier density (22). CAR-NK therapies can circumvent HLA compatibility requirements due to their innate MHC-independent recognition mechanisms, which promotes their ‘off-the-shelf’ applicability (16). However, the long-term persistence and efficacy of allogeneic CAR-NK cells can be limited by host immune rejection, mediated through mechanisms such as the development of anti-CAR antibodies or engagement of host NK cell inhibitory receptors. Strategies to mitigate this, including HLA editing and the use of transient immunosuppressive conditioning regimens, are under active investigation to fully realize the ‘off-the-shelf’ potential of CAR-NK products.

Concurrently, CAR-M therapy has emerged as an innovative strategy leveraging the innate phagocytic and immunomodulatory functions of macrophages. These engineered cells are reprogrammed to express CARs, enabling targeted phagocytosis of malignant cells while remodeling the immunosuppressive TME via pro-inflammatory cytokine secretion (e.g., interleukin [IL]-12, tumor necrosis factor-α [TNF-α]) and antigen-presenting functions (25). Unlike CAR-T/NK cells, CAR-Ms uniquely degrade stromal barriers (e.g., fibrotic extracellular matrix) to enhance immune cell infiltration, synergizing with adaptive immunity to sustain antitumor responses (26). Robust preclinical evidence for CAR-M therapy against solid tumors has fueled interest attributed to the inherent abundance of macrophages within the TME and their critical involvement in facilitating tumor progression and metastatic dissemination (27, 28). Furthermore, CAR-M cells enhance phagocytic clearance of malignant cells and improve tumor antigen presentation to T lymphocytes (29). These dual mechanisms highlight their potential to serve as a transformative strategy in immuno-oncology by simultaneously targeting tumor cells and modulating immunosuppressive TME networks. Currently, CAR-M therapy shows promise for solid tumors but faces challenges: macrophage plasticity (M1/M2 polarization), short in vivo persistence, and difficulties in genetic modification. Including breast carcino CAR-NK therapies can circumvent HLA compatibility requirements due to their innate Mma, glioblastoma, and pancreatic adenocarcinoma (30). Early-phase clinical trials (e.g., NCT04660929) are underway to systematically evaluate the safety profile and antitumor activity of CAR-M therapies across multiple cancer types, with a focus on solid malignancies (ClinicalTrials.gov). Building upon the clinical success of CAR-T cell therapies and emerging progress in CAR-NK platforms, the scientific community has garnered significant interest in advancing CAR-M-based approaches for cancer immunotherapy. The development of CAR-M introduces novel therapeutic avenues for solid tumors through genetic engineering of human macrophages. By equipping these cells with tumor-targeting CAR constructs, researchers aim to amplify their innate phagocytic functions while optimizing their capacity for cross-presenting tumor-associated antigens (31).Despite encouraging preclinical/clinical progress, key challenges hinder CAR-NK, CAR-γδ T, and CAR-M efficacy in solid tumors: enhancing in vivo persistence within immunosuppressive microenvironments, improving tumor specificity to mitigate off-target toxicity, and optimizing functional potency. This review analyzes CAR design innovations, mechanistic distinctions and recent advancements to address these barriers. This review will first outline the fundamental barriers limiting CAR-T cell efficacy in solid tumors. It will then comparatively examine the unique mechanisms and advantages of three promising alternative platforms—CAR-NK, CAR-M, and CAR-γδ T cells—in overcoming these challenges, concluding with a perspective on their future clinical translation.

2. Immune cells

Innate immunity constitutes the primary frontline defense system, encompassing physical barriers, chemical mediators, and a diverse array of immune cells expressing pattern-recognition receptors (PRRs). Key cellular components include myeloid lineage cells—dendritic cells (DCs), macrophages, monocytes, neutrophils, eosinophils, basophils, and mast cells—as well as lymphoid subsets such as NK cells, NKT cells, γδ T cells, and mucosa-associated invariant T (MAIT) cells. These innate effectors restrict tumor progression through direct cytolytic activity against malignant cells or by orchestrating adaptive immunity via antigen presentation and cytokine signaling (32). In contrast to the innate system, which provides rapid but nonspecific responses, the adaptive immune system specifically targets and eradicates cancer cells by T and B cells (33). Antitumor immunotherapies, including adoptive cell transfer (34, 35), have undergone extensive validation and received clinical approval for diverse cancer types. Notably, CAR-T therapy has demonstrated remarkable efficacy against hematological malignancies, representing a major advancement in oncological treatment (36, 37). The CAR-T cell manufacturing process involves the genetic modification of autologous T lymphocytes to express engineered receptors (38). These CARs incorporate extracellular antigen-binding domains capable of tumor antigen recognition, coupled with intracellular signaling motifs that replicate T-cell receptor (TCR) activation pathways (39, 40). To date, six CAR-T therapies have received clinical approval for treating B-cell malignancies and multiple myeloma. These include CD19-targeted agents (Yescarta, Kymriah, Tecartus, Breyanzi) and BCMA-directed therapies (Abecma, Carvykti) (41–44). While CAR-T therapy has achieved notable clinical success in hematologic cancers, its efficacy in solid tumors remains limited by three primary obstacles: (1) the immunosuppressive heterogeneity of the tumor microenvironment, (2) antigen loss or downregulation (antigen escape), and (3) insufficient trafficking and infiltration of engineered T cells into tumor sites. To advance therapeutic outcomes, dedicated research efforts must prioritize strategies tailored to overcome the biological and immunological complexities inherent to solid tumors (45–48). Current clinical immunotherapies predominantly target T cells, though their effector mechanisms remain functionally dependent on innate immune interactions to achieve activation, persistence, and immunological memory formation (49). Innate immune components—including NK cells and macrophages—not only prime adaptive immune responses but also exhibit direct tumoricidal activity through cytotoxic signaling and phagocytic clearance. Additionally, innate immune cells expressing Fc receptors (FcRs) amplify adaptive antitumor immunity by facilitating antibody-dependent cellular cytotoxicity (ADCC) and antibody-dependent cellular phagocytosis (ADCP). Given their central involvement in the cancer-immunity cycle, strategically engaging innate immune pathways represents a promising approach to augment therapeutic efficacy and counteract resistance mechanisms. Emerging evidence from preclinical and clinical investigations highlights the potential of CAR-engineered immune cells—spanning conventional T cells, γδ T cells, NK cells, NKT cells, and macrophages—to elicit durable antitumor activity (50–52). Notably, emerging research demonstrates that CAR-modified innate immune cells, mirroring the success of CAR-T therapies, exhibit comparable versatility and therapeutic promise in adoptive cell transfer paradigms (28, 53–55).

3. CAR structure

CARs are synthetic receptors comprising an extracellular single-chain variable fragment (scFv) for antigen recognition (56), a hinge and transmembrane domain for structural stability, and intracellular signaling domains (e.g., CD3ζ) for T-cell activation. Second-generation and later CARs incorporate costimulatory domains (e.g., CD28, 4-1BB) to enhance potency and persistence. “Armored” CARs include additional modifications to improve safety and efficacy against solid tumors (57, 58). We have summarized the updates and iterations of CAR used in immune cell ( Figure 1 ).

Most CAR-NK cell research has employed CAR constructs mirroring those utilized in CAR-T cell studies. These constructs either feature the same CD3ζ intracellular domain found in first-generation CAR-T cells (59) or incorporate the CD3ζ along with the co-stimulatory domain 4-1BB, characteristic of second-generation CAR-T cells (60). The incorporation of the 4-1BB co-stimulatory domain notably enhances NK cell functions, including activation, cytotoxicity, and the secretion of cytokines like interferon-γ and granulocyte-macrophage colony-stimulating factor. Similar to CAR-T cell, each CAR construct is comprised of an antigen recognition module (either immunoreceptor tyrosine-based activation motifs [ITAM] or natural receptors like natural killer group 2 member D [NKG2D]), a transmembrane region (CD8α or CD28), and signaling domains specifical for NK cells (such as 2B4, DNAX activation protein of 10 kDa [DAP10], or DNAX activation protein of 12 kDa [DAP12]). The selection of the targeted antigen is pivotal in the design of CARs, encompassing antigens such as CD19 and HER2. The scope of targeted tumor cells primarily encompasses hematological malignancies (including B-cell malignancies targeted by CD19) and solid tumors (such as epidermal growth factor receptor [EGFR] or programmed cell death ligand 1[PD-L1]-positive tumors). Given the intricate balance of activating and inhibitory receptors that regulate NK cell activation and cytotoxicity, the incorporation of NK-specific intracellular signaling domains—such as the adapter molecule DAP10 or ITAM-containing domains like DAP12 and 2B4, have been proposed to enhance cytotoxicity potentially.

2B4, a surface receptor belonging to the signaling lymphocytic activation molecule (SLAM) family, propagates activating signals in NK cells through its adaptor protein SLAM-associated protein (SAP) (60). The interaction between 2B4 and its ligand CD48 on target cells initiates NK cell activation. This engagement subsequently amplifies cytotoxic activity and stimulates IFN-γ secretion. Critically, CAR-NK cells engineered with 2B4-CD3ζ signaling domains demonstrate superior functionality compared to 4-1BB-CD3ζ-based constructs, exhibiting enhanced cytotoxicity, IFN-γ production, and in vivo tumor control (61). To refine CAR-NK design paradigms, Li et al. systematically compared a CAR-T configuration with nine CAR-NK variants featuring four transmembrane domains and diverse signaling modules targeting mesothelin (62). Among evaluated configurations, CAR-NK cells engineered with an NKG2D transmembrane domain, a 2B4 co-stimulatory domain, and a CD3ζ signaling module demonstrated robust antigen-specific cytotoxicity. Notably, human induced pluripotent stem cell (iPSC)-derived CAR-NK cells harboring this design retained prototypical NK cell surface markers, displayed sustained in vivo survival, and elicited potent antitumor responses in preclinical models. A critical challenge in CAR-NK therapy lies in prolonging their persistence within peripheral circulation and tumor-infiltrating compartments. To address this limitation, researchers have integrated cytokine transgenes—such as IL-21, IL-15, IL-7, or IL-2—into CAR constructs to enhance NK cell proliferation and survival (63). Complementary strategies employ feeder systems, including autologous PBMCs, Epstein-Barr virus (EBV)-transformed lymphoblastoid cell lines (LCLs), or NK-sensitive cell lines (e.g., K562, 721.221), to sustain post-infusion expansion (64, 65). Müller et al. further demonstrated that dual engineering of NK cells with an EGFR vIII-targeted CAR and the CXC chemokine receptor 4 (CXCR4) augmented tumor infiltration capacity, thereby improving therapeutic outcomes in solid tumor models (66).

Similarly, most CAR-M cell research has employed CAR constructs that mirror those used in CAR-T or CAR-NK cell studies. These constructs share either the same CD3ζ intracellular domain as first-generation CAR-T cells (67), or integrate the CD3ζ with the co-stimulatory domain 4-1BB, typical of second-generation CAR-T cells (29). Additionally, they incorporate macrophage-specific intracellular signaling domains, such as Toll/IL-1R (TIR), which promotes the polarization of macrophages towards the M1 phenotype. Notably, TIR has demonstrated the highest efficiency when utilized as the intracellular signaling domain (28).

First-generation CAR-Ms, designed with CD3ζ signaling domains, demonstrated antigen-specific phagocytic activity against tumor cells. Building on this foundation, Lei et al. engineered induced pluripotent stem cell-derived macrophages (iMACs) expressing a CAR incorporating the Toll-like receptor 4 (TLR4)-derived TIR domain. These TIR-integrated CAR-Ms exhibited markedly enhanced antitumor efficacy compared to CAR-M1 constructs (28). Further refinement led to a tandem CD3ζ-TIR dual-signaling CAR, which endowed iMACs with multifunctional capabilities: (1) targeted phagocytosis, (2) antigen-dependent polarization toward pro-inflammatory M1 phenotypes, (3) resistance to immunosuppressive M2 reprogramming, and (4) TME modulation. Beyond core CAR components, emerging strategies propose integrating inflammatory signaling modules (e.g., cytokine receptors or pattern recognition receptors) into CAR architectures. Such innovations aim to convert immunologically inert (“cold”) tumors into inflamed (“hot”) microenvironments, thereby sensitizing malignancies to immune surveillance (68). This genetic modification induced a significant enhancement in phagocytic activity and improved therapeutic efficacy of CAR-Ms. Parallel investigations revealed that integrating TLR4 and/or IFN-γ receptor signaling modules into CAR architectures stimulated macrophages to upregulate M1-polarization markers—including CD86, MHC class II (MHC-II), and TNF-α—while accelerating tumor regression in preclinical models (69). Analogous results emerged in macrophages transfected with sequential intracellular domains of CD3ζ and IFN-γ receptors, or CD3ζ alone, confirming the critical role of pro-inflammatory signaling in optimizing CAR-M functionality (70). Recently, a second-generation CAR-M, derived from iPSCs and engineered with a hybrid CD3ζ-TLR4 intracellular domain, demonstrated superior tumoricidal activity and TME remodeling capabilities compared to earlier iterations (71). Zhang et al. developed HER2-targeted chimeric antigen receptor macrophages (CAR-147) comprising a scFv fused to a hinge region, CD147 transmembrane domain, and intracellular signaling module. These CAR-Ms selectively upregulated matrix metalloproteinases (MMPs), including MMP9, MMP10, and MMP12, facilitating extracellular matrix (ECM) degradation in tumors (72–74). Notably, the CAR-CD147 design achieved ECM disruption without compromising phagocytic capacity, inflammatory cytokine secretion, or reactive oxygen species (ROS) production. Furthermore, this construct significantly inhibited tumor progression and enhanced intratumoral T-cell infiltration in vivo (72, 73). In a parallel approach, Niu et al. engineered CAR-Ms to express C-C motif chemokine ligand 19 (CCL19), the natural ligand for C-C motif chemokine receptor (CCR7), to chemoattract CCR7+ immunosuppressive cells. This strategy promoted CD3+ T-cell recruitment into tumors, amplified pro-inflammatory cytokine levels (e.g., IFN-γ, TNF-α), suppressed tumor growth and metastasis, and prolonged survival in preclinical models (75).

Apart from designing the CAR, the preparation of the cell source for loading the CAR cells, and the functionality of the loaded CAR cells, are both crucial steps in the process.

4. Beyond CAR-T: next-generation CAR-engineered cell therapies

To address the limitations of CAR-T cells in solid tumors, research has expanded to harness other immune effector cells, each offering distinct biological advantages.

5. Clinical trials of cell therapy

NK cells exhibit inherent therapeutic advantages, including MHC-independent target recognition, tumor tissue infiltration capacity, and robust cytolytic activity, while minimizing risks of severe adverse events such as CRS, GvHD, and ICANS. These attributes position CAR-NK cells as a promising modality for addressing solid malignancies. To date, clinical investigations have predominantly focused on CAR-NK products derived from the NK92 cell line, PB-NK, or UCB-K. Current trials registered on ClinicalTrials.gov (summarized in Table 1 ) increasingly target antigens implicated in solid tumors, including NKG2D ligands, mesothelin (MSLN), HER2 and MUC1.

To date, early-phase clinical investigations of CAR-Ms remain limited, with only a small number of registered trials documented on ClinicalTrials.gov (summarized in Table 2 ). The pioneering Phase I trial (NCT04660929), designated CT-0508, evaluates the safety and feasibility of CAR-M therapy in HER2-overexpressing solid tumors. Developed by CARISMA Therapeutics, this candidate utilizes a chimeric adenoviral vector (Ad5f35) for ex vivo genetic modification of autologous PBMC-derived macrophages, enabling HER2-targeted antitumor activity. CT-0508 represents the first-in-human application of CAR-Ms, with preliminary objectives focused on assessing dose-limiting toxicities, pharmacokinetics, and biomarker correlates of response. The trial design incorporates dose escalation followed by expansion cohorts, with planned combinatorial arms exploring synergy with checkpoint inhibitors. Preclinical data underpinning this trial demonstrated that Ad5f35-engineered CAR-Ms not only directly phagocytose HER2+ tumor cells but also remodel immunosuppressive microenvironments via pro-inflammatory cytokine secretion, thereby enhanad5cing T cell infiltration.

The clinical translation of γδ T cell therapy has progressed from early-phase trials utilizing unmodified cells to contemporary studies evaluating advanced, genetically engineered products. Initial clinical studies focused on either activating endogenous Vγ9Vδ2 T cells in vivo with agents like zoledronate and low-dose IL-2, or the adoptive transfer of ex vivo-expanded autologous γδ T cells (159). These foundational trials, conducted in patients with diverse solid and hematological cancers, consistently demonstrated an exceptional safety profile. The therapies were well-tolerated, with a notable absence of GvHD, even in the allogeneic setting. While these studies established safety and feasibility, objective response rates were modest, with disease stabilization being the most common outcome (160).

The crucial safety data from these early efforts provided the impetus for developing the next generation of γδ T cell therapies. A pivotal evolution has been the shift towards allogeneic, “off-the-shelf” products derived from healthy donors, a strategy that leverages the innate safety of γδ T cells to create a standardized, readily available therapeutic. Current clinical trials are increasingly focused on CAR-γδ T cells. Early-phase studies are now underway to evaluate the safety and preliminary efficacy of these constructs in both blood cancers and solid tumors, targeting antigens such as CD19, CD20, and NKG2D ligands. The results from these trials will be instrumental in validating the clinical potential of engineered γδ T cells and will inform the design of future, more sophisticated therapeutic strategies aimed at fully harnessing their potent anti-cancer capabilities. To date, only a few clinical studies evaluating CAR- γδ T cell products have been registered on ClinicalTrials.gov, as summarized in Table 3 .

6. Application of CAR-NK and CAR-M in gynecologic tumors

The therapeutic potential of CAR-NK, CAR-M, and CAR-γδ T cells in solid tumors is driven by their divergent mechanisms. CAR-NK cells exhibit a “dual-modal killing” mechanism. Firstly, through their chimeric antigen receptor, they release perforin and granzymes to mediate direct cytotoxicity and secrete IFN-γ to modulate TME. Secondly, they retain innate ADCC via their CD16 receptor, which enables the elimination of tumor cells that downregulate the CAR-targeted antigen, thereby counteracting tumor heterogeneity (90). CAR-M cells are primarily characterized by their potent phagocytic capacity. A central feature of their function is the ability to remodel the immunosuppressive TME. They secrete pro-inflammatory cytokines including TNF-α and IL-6, recruit cytotoxic T cells, enhance antigen presentation, and can polarize macrophages from the M2 to the M1 phenotype in the TME (161). CAR-γδ T cells combine features of both innate and adaptive immunity. They kill target cells through perforin and granzyme release as well as death receptor-mediated apoptosis such as via the Fas/FasL pathway. A hallmark of these cells is their MHC-independent recognition of stress-induced ligands through receptors like NKG2D, which provides a critical advantage against tumors that evade MHC-dependent recognition (162).

CAR therapies for the above cells are already in the clinical stage, and this paragraph mainly lists CAR-NK and CAR-M in gynecological tumors (e.g., OC). In preclinical studies of OC, CAR-NK targeting HAL-G, CD44, MSLN and αFR has demonstrated superior antitumor effects (163–167). Zhu et al. generated ROBO1-targeted CAR NK cells from PBMCs of OC patients. Efficacy was evaluated using xCELLigence RTCA, CCK-8 and Live/Dead fluorescence assays. Compared with primary NK cells without ROBO1-CAR modification (168), ROBO1-NK cells exhibited higher efficiency in eradicating primary ovarian cancer cells and lysing ovarian tumor organoids. Raftery et al. developed a next-generation CAR targeting CD44v6 that incorporates IL-15 superagonist and checkpoint inhibitor molecules. It could show that CD44v6 CAR-NK cells demonstrated effective cytotoxicity against triple negative breast carcinoma (TNBC) in 3D spheroid models (169). Kutle et al. engineered NK cells with a third-generation CAR targeting MSLN, a tumor-associated antigen overexpressed in primary human cervical carcinomas and established cell lines, as evidenced by their study and prior research. The CAR construct, delivered via self-inactivating (SIN) alpha retroviral vectors, incorporated tandem co-stimulatory domains (CD28 and 4-1BB) to enhance NK cell activation and persistence. In functional assays, anti-MSLN CAR-NK-92 cells exhibited potent cytotoxicity against cervical cancer models, with efficacy validated in both 2D monolayers and 3D spheroid cultures. This antitumor activity correlated with elevated degranulation markers (e.g., CD107a) in CAR-NK-92 cells upon MSLN+ target engagement. To confirm antigen-specific targeting, the authors employed CRISPR-Cas9 gene editing to generate MSLN-knockout cervical cancer cells. Co-culture experiments with these isogenic pairs demonstrated that cytotoxic activity of both CAR-NK-92 cells and primary CAR-NK cells (derived from healthy donors) was strictly dependent on MSLN expression. Furthermore, combinatorial treatment with anti-MSLN CAR-NK-92 cells and conventional chemotherapeutic agents synergistically enhanced tumor cell elimination compared to monotherapy regimens, suggesting potential for integrated therapeutic strategies (170). Poorebrahim et al. engineered a dual-receptor system in NK-92 cells by integrating a clinically validated TCR targeting human papillomavirus 16 (HPV16) E7 oncoprotein with a novel CAR directed against trophoblast cell surface antigen 2 (TROP2). The CAR construct incorporated CD28 and 4-1BB co-stimulatory domains but excluded the CD3ζ signaling module, a design choice aimed at modulating activation thresholds. Flow cytometric analysis revealed significant upregulation of activation markers (e.g., CD69, CD25) and cytolytic mediators (e.g., granzyme B) in NK-92 cells co-expressing CD3, CD8, E7-TCR, and TROP2-CAR following exposure to HPV16+ cervical cancer cells. Notably, dual-receptor NK-92 cells exhibited superior antigen-specific activation and tumoricidal activity compared to counterparts expressing E7-TCR alone. This enhancement was attributed to synergistic signaling between the E7-TCR and TROP2-CAR, wherein the co-stimulatory domains (CD28/4-1BB) amplified intracellular activation cascades. Mechanistically, TROP2-CAR engagement augmented TCR-mediated cytotoxicity without inducing exhaustive phenotypes, suggesting a balanced signaling interplay. This combinatorial strategy holds translational promise for HPV16-associated malignancies, potentially overcoming limitations of single-receptor adoptive therapies. By coupling viral antigen specificity (via TCR) with tumor-associated antigen targeting (via CAR), the approach may broaden targetable epitopes while mitigating antigen escape. Current clinical investigations are evaluating such engineered NK cells in HPV16+ cancers, with implications for improving therapeutic efficacy and durability (171). Klapdor et al. developed a CAR-engineered immunotherapy targeting CD133, a well-established CSC marker, to address therapeutic challenges in OC. The researchers constructed a third-generation anti-CD133 CAR incorporating CD28 and 4-1BB co-stimulatory domains, which was encoded into a lentiviral vector for stable genetic modification. Clinically applicable NK92 cells—a standardized natural killer cell line—were transduced with this construct to generate CD133-specific CAR-NK cells. Functional validation assays demonstrated selective cytotoxicity of engineered CAR-NK92 cells against CD133-positive OC cell lines in vitro. Notably, this activity extended to primary OC cells isolated from sequential ascites samples of patients with advanced disease, underscoring the therapeutic relevance of targeting CD133-expressing tumor populations. The observed antitumor efficacy correlated with CSC depletion, suggesting potential to mitigate relapse driven by chemotherapy-resistant stem-like cells (172).

The HER2 gene is closely associated with tumorigenesis, particularly in breast cancer, and is amplified across multiple solid malignancies. In OC, HER2 overexpression correlates with aggressive phenotypes, including recurrence and metastasis, while maintaining minimal expression in healthy tissues (173). Leveraging this therapeutic window, Chen et al. engineered CAR-Ms dually targeting HER2 and CD47, a “don’t eat me” signal, to enhance tumor-specific phagocytosis. In vitro assays demonstrated that HER2/CD47-targeted CAR-Ms selectively engulfed OC cells and activated CD8+ cytotoxic T lymphocytes (CTLs) via pro-inflammatory cytokine secretion, thereby fostering adaptive antitumor immunity. In humanized mouse models, CAR-M administration induced tumor regression, concomitant with enhanced CD8+ T cell activation and polarization of TAMs toward pro-inflammatory phenotypes (29).

Immune cell therapy has emerged as a transformative strategy to bolster immune responses and counteract tumor-induced immunosuppression. This approach can be utilized either independently or in combination with established conventional therapies, such as surgery, radiotherapy, chemotherapy, and other immunotherapeutic modalities, for treatment in gynecological cancers. By effectively reducing tumor recurrence and metastasis rates while mitigating adverse drug reactions, immune cell therapy offers a promising avenue to enhance patient survival outcomes. Intriguingly, this innovative therapeutic paradigm holds immense potential for extending to women with gynecologic malignancies for improving therapeutic efficacy in future clinical settings.

7. Outlook

While CAR-T cell therapy has demonstrated remarkable clinical efficacy in hematological malignancies, persistent challenges hinder its broader application. Critical limitations include CRS, ICANS, antigen escape mechanisms, poor adaptation to immunosuppressive TMEs, high relapse rates, and prohibitive manufacturing costs. These shortcomings have spurred exploration of alternative immune effector platforms, notably CAR-NK cells and CAR-Ms, which exhibit distinct therapeutic advantages.

CAR-NK cells offer intrinsic benefits such as non-MHC-restricted cytotoxicity, potential for “off-the-shelf” allogeneic use, reduced risk of CRS/ICANS, and lower production costs compared to patient-specific CAR-T products. Furthermore, their limited in vivo persistence may mitigate long-term toxicity risks. CAR-Ms, conversely, demonstrate unique tropic capabilities to penetrate immunosuppressive TMEs and remodel stromal barriers, positioning them as promising agents for solid tumor eradication. These attributes collectively advocate for CAR-NK and CAR-M therapies as viable alternatives to circumvent CAR-T limitations while maintaining robust antitumor activity.

To optimize CAR-based immunotherapies, emerging strategies prioritize:

• Target Antigen Innovation: High-throughput screening for tumor-specific surface markers with minimal on/off-tumor cross-reactivity;

• TME reprogramming: Co-expression of immunomodulatory payloads (e.g., cytokines, checkpoint inhibitors) to counteract immunosuppression (174, 175);

• Combinatorial regimens: Synergistic pairing with radiotherapy, chemotherapy, or targeted therapies to enhance CAR cell infiltration and durability;

• Architectural refinement: Engineering next-generation CAR constructs (e.g., logic-gated receptors, hypoxia-responsive switches) tailored for solid tumor biology. New approaches include hypoxia-sensitive CARs activated under tumor conditions, logic-gated CARs requiring dual-antigen recognition, and CAR constructs with built-in safety switches for toxicity control (176). Together, these designs highlight innovative engineering tactics to enhance the specificity and safety of CAR therapies in solid tumors (177).

Preclinical studies and early-phase trials have validated the feasibility of these approaches, underscoring the necessity for continued translational investment. Key barriers to translating CAR-immunocellular therapies from preclinical to clinical success include: inefficient tumor infiltration, immunosuppressive tumor microenvironments, antigen heterogeneity/escape, on-target/off-target toxicity, and limited survival of CAR-NK/CAR-M/CAR-γδ T cells in solid tumors. In this regard, combining CAR engineering with the intrinsic ability of NK and γδ T cells to recognize stress-induced ligands (e.g., MICA/B, ULBPs, BTN2A1/BTN3A1) offers a dual-targeting strategy that may reduce antigen escape and enhance therapeutic safety and efficacy in solid tumors (178). For successful clinical translation, future work must establish standardized protocols for generating clinical-grade CAR-NK/M/γδ T products from patient-derived monocytes.

The development of allogeneic “off-the-shelf” CAR therapies is complex due to their dual classification as both gene and cell products. Early engagement with regulators is essential to establish appropriate potency assays, quality controls, and long-term safety monitoring. Economically, the high cost of current autologous CAR-T therapies underscores the need for cost-effectiveness analyses of allogeneic alternatives. Demonstrating superior value, not just efficacy, will be vital for reimbursement and patient access.

Future research must overcome these fundamental barriers to fully unlock the clinical potential of emerging platforms such as CAR-NK, CAR-M, and CAR-γδ T cells.

Funding Statement

The author(s) declare financial support was received for the research and/or publication of this article. This work was supported by the Key program of Tianjin Municipal Science and Technology Bureau (22JCZDJC0030) and the Tianjin Hygiene & Health Commission (TJWJ2024ZD012) to SP, Natural Science Foundation of Jiangxi Province Funding Project (20232BAB205086) to CF. The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Author contributions

YH: Writing – original draft, Writing – review & editing. SH: Writing – original draft, Writing – review & editing. CL: Writing – review & editing. XC: Writing – review & editing. YW: Writing – review & editing. YL: Writing – review & editing. ZF: Writing – review & editing. CF: Writing – review & editing. YG: Writing – review & editing. ZL: Writing – review & editing. SP: Writing – review & editing.

Conflict of interest

SH is member of Thinking Biomed Beijing Co., Ltd. that works on potential therapeutic application based on CAR-T products. CF is member of Jiangxi Health-Biotech Stem Cell Technology Co., Ltd. that works on potential cellular therapeutic application.

The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declare that no Generative AI was used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.