Different Immune Cells Modified With Chimeric Antigen Receptors Are Being Applied to Ovarian Cancer: Which Is the Most Effective?

ABSTRACT

Ovarian Cancer (OC), the deadliest gynecological malignancy, poses a major therapeutic challenge in advanced stages owing to its high recurrence rate and metastatic potential. In this regard, it is noteworthy that immunotherapy has recently gained significant attention in OC treatment, a phenomenon attributable to notable advances in over‐the‐counter Chimeric Antigen Receptor (CAR)‐based cell therapy. At the heart of CAR‐T Cell (CAR‐T) immunotherapy is genetically modified CAR molecules that enable immune cells to target and recognize tumor antigens. Based on such strategies, CAR‐T therapies have developed rapidly in hematological oncology and are gradually being extended to solid tumors. Despite their potential in OC treatment, several factors, including off‐target effects attributable to the lack of Tumor‐Specific Antigens (TSAs), as well as severe side effects such as tumor immune barriers, Cytokine Release Syndrome (CRS), and neurotoxicity, have been established to limit the clinical use of CAR‐T therapies. Moreover, compared to CAR‐T, CAR‐Natural Killer (NK) and CAR‐Macrophage (M) therapies have distinct advantages. The killing mechanism of NK cells integrates both CAR‐dependent and non‐dependent pathways, avoiding severe CRS and neurotoxicity. Furthermore, besides directly phagocytosing tumors due to its strong ability to infiltrate tumors, CAR‐M therapy could also effectively improve the Immunosuppressive Microenvironment (IME) via immunomodulatory factor secretion to remodel M2‐type Tumor‐Associated Macrophages (TAMs) into the M1 phenotype with anti‐tumor function. In this review, we systematically describe the research progress in CAR‐T therapy for OC and compare the similarities and differences of three types of cellular therapies (CAR‐T, CAR‐NK, and CAR‐M) regarding their mechanisms of action, clinical advantages, and technological bottlenecks. We hope that our findings will provide a theoretical basis for optimizing immunotherapeutic strategies for OC.

Trial Registration: ClinicalTrials.gov identifier: NCT03585764

1. Introduction

Ovarian Cancer (OC), the eighth most prevalent and deadliest gynecological malignancy [1], characteristically presents with insidious early‐stage lesions and non‐specific clinical symptoms, often leading to delayed diagnosis. Consequently, patients are often diagnosed at advanced stages, with limited therapeutic options and a poor prognosis [2]. Owing to this “high lethality‐low detection rate” clinical profile, OC is presently recognized as a “silent killer”.

Multiple immune cells exert antitumor effects in the Tumor Immune Microenvironment (TIME) through synergistic interactions based on specific molecular recognition mechanisms and divergent effector pathways. For instance, Cytotoxic T Lymphocytes (CTLs) recognize tumor antigens via T Cell Receptor (TCR)‐mediated binding to MHC class I/peptide complexes [3], thus activating the perforin/granzyme B‐dependent apoptotic pathway [4]. Meanwhile, CTLs also secrete Th1‐type cytokines such as Interferon‐γ (IFN‐γ), which trigger the JAK–STAT signaling cascade, thus enhancing the functionality of Antigen‐Presenting Cells (APCs) and upregulating MHC class I molecules on tumor cells [5]. Additionally, Natural Killer (NK) cells engage tumor cells via Killer Activation Receptors (KARs) that bind to stress‐induced ligands (e.g., MICA/MICB) [6] or via Killer Immunoglobulin‐like Receptors (KIRs) that recognize MHC class I deficiency [7], thereby initiating Antibody‐Dependent Cellular Cytotoxicity (ADCC) [8]. Moreover, the FcγRIII (CD16) receptor binds to the Fc region of tumor‐specific IgG antibodies, inducing cytolytic granule polarization and release via the PI3K‐Akt pathway [9]. Following Pattern Recognition Receptor (PRR)‐guided immune sensing, Tumor‐Associated Macrophages (TAMs) could also engage in Fc receptor‐dependent phagocytosis [10] and generate cytotoxic Reactive Oxygen Species (ROS) and Reactive Nitrogen Intermediates (RNI) through the NADPH oxidase system and inducible Nitric Oxide Synthase (iNOS), respectively [11]. These mechanisms, along with paracrine pro‐inflammatory cytokines (e.g., TNF‐α, IL‐1β, and so on), could drive TAM repolarization from an immunosuppressive M2 phenotype to an antitumor M1 phenotype, thus remodeling the immunosuppressive TIME [12].

Recent advances in cellular therapeutics, particularly Chimeric Antigen Receptor T‐cell (CAR‐T) therapy, have markedly transformed cancer immunotherapy [13]. This approach mainly involves genetically engineering T cells to express surface Chimeric Antigen Receptors (CARs), enabling tumor antigen‐specific recognition, thus bypassing Major Histocompatibility Complex (MHC)‐restricted antigen presentation mechanisms and redirecting T cell cytotoxic activity. This novel strategy could be leveraged to address clinical challenges in OC treatment, including difficulties in targeted therapy and insufficient immunogenicity due to high tumor antigen heterogeneity and low mutational burden.

Despite its remarkable efficacy, CAR‐T therapy has been associated with unique toxicities such as Cytokine Release Syndrome (CRS) and Immune Effector Cell‐Associated Neurotoxicity Syndrome (ICANS) [14]. Upon identifying their target, CAR‐T cells release initial cytokines, including IFN‐γ and TNF‐α, which subsequently trigger the extensive activation of macrophages. These macrophages then secrete significant quantities of IL‐6, leading to CRS toxicity. This highly inflammatory cascade further results in endothelial cell activation and capillary leakage [15], which clinically manifests as fever, hypotension (HoTN), and multi‐organ dysfunction [16]. On the other hand, ICANS characteristically presents with an altered mental status, speech deficits, or seizures [17].

To address the challenges of CAR‐T therapy in solid tumors, such as limited tumor infiltration and toxicity, the use of tocilizumab, a chimeric anti‐IL‐6 monoclonal antibody, or siltuximab to inhibit the CRS cascade has emerged as the primary treatment option for severe cases of CRS [18]. Moreover, researchers have leveraged the core attributes of CAR technology—the genetic reprogramming of immune cells for precise targeting—to develop novel interventions such as CAR‐NK (Chimeric Antigen Receptor Natural Killer Cells) and CAR‐M (Chimeric Antigen Receptor Macrophages) (Figure 1). Due to their “off‐the‐shelf” feasibility, inherent anti‐tumor activity, and reduced CRS risk [19], CAR‐NK cells have shown great promise in OC treatment [20]. On the other hand, CAR‐M could phagocytose tumor cells and reprogram the Immunosuppressive Microenvironment (IME), thus exerting therapeutic effects. However, while reprogramming the immunosuppressive microenvironment, the pro‐inflammatory toxicity of CAR‐M cells necessitates further evaluation. The sustained activation of macrophages may trigger the NF‐κB and NLRP3 inflammasome pathways. This hyperinflammatory state results in the excessive production of reactive oxygen species (ROS) and nitric oxide (NO), leading to a detrimental cascade of IL‐1β, IL‐6, and TNF‐α. The systemic dissemination of these mediators poses the risk of inducing fatal hemophagocytic lymphohistiocytosis (HLH) or a macrophage activation syndrome (MAS)‐like state, which is clinically characterized by severe coagulation disorders, cytopenia, and associated endothelial cell necrosis [21].

FIGURE 1.

Schematic diagram of CAR cell therapy | The in vitro model for CAR cell therapy (A). Mechanisms of tumor cell killing by CAR‐NK cells (B). Mechanisms of tumor cell killing by CAR‐T cells (C). Mechanisms of tumor cell killing by CAR‐Ms cells (D). The pattern was created with BioGDP.com.

This article systematically reviews the advancements in CAR‐T therapy for OC, examines the similarities and differences among CAR‐T, CAR‐NK, and CAR‐M strategies regarding the mechanisms of action, clinical benefits, and technical limitations, and discusses the optimal cell sources for CAR‐T, CAR‐NK, and CAR‐M therapies for OC treatment. Finally, it offers future directions for achieving curative outcomes in OC via CAR‐engineered cellular therapies.

2. Structure and Generational Evolution of CARs

As earlier mentioned, CAR is a genetically engineered synthetic receptor that enables immune cells (e.g., T cells, NK cells, or macrophages) to recognize and target tumor cells precisely. Mechanistically, it primarily bypasses natural immune cell‐activated restrictive pathways and directly targets tumor surface antigens. Notably, CAR comprises three key functional domains. First, the extracellular structural domain contains the antigen‐binding domain and the hinge region. The core component of the antigen‐binding domain is the Single Chain Fragment Variable (SCFV), which comprises the heavy chain variable regions of tumor‐associated antigen‐specific antibodies (VH) and the light chain variable regions of the light chain (VL). The light chain (VL) contains a flexible linker peptide that recognizes antigens on target cell surfaces; the hinge region that increases spatial plasticity and maintains CAR T cell activity and stability, thus promoting antigen binding. Second, the transmembrane domains connect the extracellular and intracellular domains involved in antigen recognition and signal recruitment and are commonly used in transmembrane regions of natural proteins such as CD8α or CD28. Third, the intracellular structural domain contains co‐stimulatory domains and signal transduction units, often comprising the CD3ζ chain and co‐stimulatory molecules (e.g., CD28, 4‐1BB). The selection of costimulatory domains plays a crucial role in shaping the cytotoxic characteristics and dynamics of CAR‐T cells. Incorporating CD28 significantly enhances activation of the PI3K/AKT/mTOR signaling pathway, leading to a metabolic shift in T cells towards aerobic glycolysis. This shift results in rapid and robust proliferation along with a substantial release of cytokines. However, clinically, this heightened activity is associated with an increased risk of severe cytokine release syndrome (CRS) and swift cellular exhaustion [22]. In contrast, the 4‐1BB (CD137) domain transmits signals via the TRAF2/TRAF1 cascade, upregulating AMPK and promoting mitochondrial biogenesis, fatty acid oxidation, and oxidative phosphorylation (OXPHOS) [23]. This metabolic network reduces the acute peak of cytokine release, lowering the risk of severe CRS while promoting the development of central memory T cells (Tcm) for long‐term residency in the body. Notably, upon antigen binding, this structural domain activates the killing effect of T cells [24, 25, 26].

3. Generational CAR Optimization

Due to iterative upgrades in co‐stimulatory signaling domains, CARs are classified into four key generations: First‐generation CARs (with only the CD3ζ signaling unit); second‐generation CARs (with a single co‐stimulatory molecule, e.g., CD28 or 4‐1BB); third‐generation CARs (combining two co‐stimulatory molecules, such as CD28 and 4‐1BB); and fourth‐generation CARs (carrying additional cytokine secretion modules, such as IL‐12, which activate nuclear factors, thus enhancing the anti‐tumor immune response) [27]. Systemic IL‐12 activates STAT4 signaling in immune cells, creating a feedback loop that produces IFN‐γ and TNF‐α, leading to hepatotoxicity, cytopenia, and mucosal damage [28]. To address these toxicities, modern fourth‐generation CARs, specifically TRUCKs cells, employ the nuclear factor of activated T cells (NFAT) response promoter. Upon binding to a tumor antigen, an increase in intracellular calcium ion concentration activates calmodulin‐dependent protein kinase, leading to the dephosphorylation of NFAT. This process facilitates the translocation of NFAT into the nucleus, thereby driving IL‐12 transcription. Once the tumor antigen is eliminated and calcium levels normalize, IL‐12 production is rapidly downregulated, ensuring that the release of this highly toxic cytokine is transient and strictly limited to the tumor microenvironment [29].

Furthermore, the traditional classification of CAR generations is primarily T‐cell‐centric, relying on T‐cell receptor (TCR) signaling cascades (CD3ζ, CD28, and 4‐1BB). Directly applying these conventional T‐cell CARs to NK cells or macrophages often results in poor activation, necessitating structural customization specific to the cell type. For CAR‐NK cells, replacing T‐cell costimulatory domains with NK‐specific signaling adaptors is a more effective alternative. For example, integrating the transmembrane and intracellular domains of NKG2D or 2B4, or utilizing DAP10/DAP12, aligns more closely with the natural activation pathways of NK cells, thereby enhancing cytotoxicity and cytokine secretion [30]. Similarly, in CAR‐M therapy, the CD3ζ chain often encounters challenges in effectively initiating phagocytosis. Currently, CAR‐M constructs incorporate the intracellular domains of phagocytic receptors, including Megf10 and the Fc receptor gamma (FcRγ) chain. These domains facilitate actin cytoskeletal reorganization and enhance the phagocytosis of target cells by activating the Syk‐PI3K signaling axis [31], highlighting that optimal CAR design must fundamentally align with the biological characteristics of the engineered immune cells.

4. IME Composition and Function

The TIME is a dynamic ecosystem comprising tumor tissues and their surrounding immune cells (e.g., cytotoxic T cells and NK cells), non‐immune cells (e.g., Tumor‐Associated Fibroblasts [TAFs]), cytokines (e.g., TGF‐β and IL‐10), chemokines (e.g., CXCL12), and the Extracellular Matrix (ECM), among others [32]. Anti‐tumor immune components, such as CD8⁺ T cells, NK cells, M1‐type macrophages, and Dendritic Cells (DCs), exert anti‐tumor effects through the direct killing or activation of the immune response [33]; and pro‐tumor immune components, such as Regulatory T cells (Treg), M2 macrophages, and Myeloid‐Derived Suppressor cells (MDSCs), promote tumor progression through immunosuppression or pro‐angiogenesis [34]. On the other hand, chemokines and immunosuppressive factors (e.g., Programmed Death Ligand 1 [PD‐L1]) regulate immune cell migration and maintain a state of local immunosuppression, respectively.

After migrating from Hematopoietic Stem Cells (HSCs) in the bone marrow to the thymus, mature T cells differentiate into self‐tolerant CD4⁺ or CD8⁺ T cells through TCR gene rearrangement, positive selection (to ensure MHC recognition), and negative selection (to remove self‐reactive clones). In response to the CXCL9/10/11 chemotactic gradient, CD8⁺ T cells (CTLs)—enriched in the tumor core—migrate towards the tumor via the CXCR3 receptor, making them the core effector cells of anti‐tumor immunity [35]. On the other hand, CD4⁺ T cells are distributed at tumor margins and comprise T helper cells (Th) and Regulatory T cells. While Th enhance CTL function via IFN‐γ secretion [36], Tregs suppress the immune response through the IL‐10/TGF‐β axis [37].

5. Anti‐Tumor Killing Mechanism of CTL

Mechanistically, CTLs specifically recognize Tumor‐Associated Antigens (TAAs) or Tumor‐Specific Antigens (TSA) presented by MHC class I molecules on the surface of tumor cells via the TCR to activate the killing program. According to research, APCs (e.g., DCs, macrophages) could capture tumor antigens, which, in turn, could activate Th cells via MHC class II molecules or CTLs via MHC class I molecules through cross‐presentation, causing the dual signaling (TCR‐MHC/antigen peptide recognition + co‐stimulatory signals)‐driven T cell activation and expansion [38].

Besides overexpressing MHC class I molecules and costimulatory molecules (e.g., B7‐CD28), cells in tumors with high immunogenicity also act directly as non‐classical APCs to activate CTLs. According to research, CTLs eliminate tumor cells via granule exocytosis (perforin/granzyme B) and death receptor‐mediated apoptosis (e.g., Fas/FasL) [39]. Furthermore, in solid tumors, tumor cells often downregulate MHC class I molecules, necessitating professional APCs to perform antigen cross‐presentation. Activated Th1 cells could promote CTL expansion by enhancing APC functionality via IL‐2, IFN‐γ, and CD40L‐CD40 interactions, while secreting CXCL9/10 to recruit CTLs to tumor sites and maintain their survival and effector functions [40, 41, 42](Figure 2).

FIGURE 2.

T cell function | MHC class I molecules and co‐stimulatory molecules are upregulated on the surface of certain tumor cells and bind to the TCRs of CD8⁺ T cells via the MHC I‐TAA complex, while B7‐CD28 interactions provide co‐stimulatory signals that stimulate the activation and secretion of autocrine IL‐2 by CD8⁺ T cells and promote their proliferation and differentiation into effector CTLs, inducing tumor cell apoptosis via the release of various cytokines (such as IFN‐γ, TNF‐α, etc.) and perforin granzymes to induce tumor cell apoptosis, while B7‐CTLA4 interactions provide synergistic inhibitory signals to prevent T‐cell activation and proliferation (A). In contrast, in most solid tumors, the expression of MHC molecules and co‐stimulatory molecules on tumor cells is decreased. In this case, APC can bind to the TCR of CD8⁺ T cells via the MHC I‐TAA complex, whereas the B7‐CD28 interactions provide a second signal that activates CD8⁺ T cells to exert specific cytotoxicity against tumor cells; APC can also activate CD4⁺ T cells through the delivery of MHC class II molecules, and activated CD4⁺ T cells secrete IL‐2, IFN‐γ, and other cytokines to induce diverse immune effects, and bind to CD40 on the surface of APC. In addition, activated CD4⁺ cells could also secrete chemokines to recruit CTL to the tumor site (B). The pattern was created with BioGDP.com.

6. Immunotherapy Dilemma

The lack of TSAs and solid Tumor Microenvironment (TME) heterogeneity are currently the key dilemmas in immunotherapy, collectively leading to immune escape. The other key challenge is the physical barrier, primarily based on the tumor stromal architecture, which comprises tumor cells, Cancer‐Associated Fibroblasts (CAFs), and a dense ECM. In OC, aberrant deposition of ECM components (e.g., collagen and fibronectin) forms rigid networks that physically impede drug penetration and immune cell infiltration [43]. Besides the physical barrier, there is also the immunosuppressive barrier, which comprises various cellular components. Examples include Tregs and Myeloid‐Derived Suppressor Cells (MDSCs), which suppress effector T cell function via immunosuppressive cytokine (e.g., TGF‐β, IL‐10) secretion and direct cell contact inhibition [44]. From a molecular standpoint, the immunosuppressive milieu is further reinforced by two key pathways: (1) Aberrant PD‐1/PD‐L1 overexpression, which drives T‐cell exhaustion [45]; and (2) Soluble factors such as IL‐10 and VEGF, which amplify immune suppression via paracrine signaling [44].

7. Tumor Immune Escape

Gene inactivation occurs when the gene encoding tumor‐specific TSAs is silenced due to various processes, such as point mutations, shift mutations, or Loss of Heterozygosity (LOH). Epigenetic silencing has also been implicated in gene inactivation, especially given that abnormal hypermethylation in the promoter region of the TSA gene could inhibit transcription. Furthermore, immunoediting plays a critical role, especially when the immune system effectively removes high TSA‐expressing tumor clones. Consequently, low or no TSA‐expressing sub‐clones could proliferate, benefiting from the “survival of the fittest” phenomenon. This selective pressure underscores the dynamic immune system–tumor evolution interaction.

Multiple key mechanisms define the Infiltrated‐Immune Phenotype (IIP). First, tumor‐secreted CXCL9/10/11 chemokines, which effectively draw CD8⁺ T cells into the TME, facilitate immune cell recruitment [46]. Once recruited, activated T cells secrete IFN‐γ, which triggers STAT1 signaling to upregulate tumor MHC‐I expression [47]. Conversely, IFN‐γ induces PD‐L1 overexpression on tumor cells and CAFs, suppressing TCR signaling via the PD‐1/SHP‐2 pathway [48]. On the other hand, the Immune‐Desert Phenotype (IDP) presents a different set of challenges. First, the metabolic suppression caused by Indoleamine 2,3‐Dioxygenase (IDO)‐mediated tryptophan catabolism depletes local tryptophan and generates kynurenine, which ultimately impairs CD8⁺ T‐cell function [49]. Second, an immunosuppressive network emerges where Tregs inhibit DC activity through CTLA‐4 engagement. Furthermore, Tregs could secrete IL‐10 and TGF‐β, which synergistically drive CD8⁺ T‐cell exhaustion [50].

8. Immune Infiltration Heterogeneity and Therapeutic Strategies for Solid Tumors

Various solid tumors (e.g., Breast Cancer (BC), Renal Cell Carcinoma (RCC), melanoma, OC, and gastrointestinal mesenchymal stromal tumors) exhibit significant T‐cell infiltration heterogeneity [51, 52, 53, 54, 55], highlighting the need for specific intervention strategies based on tumor type. To address the complex TME, current clinical interventions mainly include Immune Checkpoint Blockades (ICBs), such as PD‐1/PD‐L1 inhibitors [56]; and Adoptive Cell Therapy (ACT), such as CAR‐T [57].

As earlier mentioned, CAR‐T therapy genetically engineers T cells to target tumor antigens. Notably, its killing mechanism does not rely on MHC‐mediated antigen presentation, allowing for the direct recognition of tumor surface antigens and triggering cytotoxic effects. Furthermore, in highly T‐cell‐infiltrated tumors, an overactivated IME could inhibit CTL function, leading to immune escape. Conversely, CAR‐T could bypass the traditional antigen presentation pathway, precisely identifying tumor antigens and avoiding excessive immune activation. However, because the majority of target proteins in ovarian cancer are tumor‐associated antigens (TAAs) rather than strictly tumor‐specific mutations, this non‐selective killing mechanism has the potential to cause significant off‐target effects. This risk arises when the CAR binds to antigens that are consistently expressed on healthy epithelial tissue [58]. Therefore, optimizing CAR affinity or utilizing logic‐gated synthetic biology circuits is essential for expanding the therapeutic index and minimizing collateral tissue damage [59]. In OC, immune infiltration characteristically presents with a high ratio of CD8⁺ T cells to Foxp3⁺ Tregs in tumor tissues, a phenomenon that correlates with a favorable prognosis for patients [60]. Regarding CAR‐T potential, OC cells express high levels of highly immunogenic antigens such as mesothelin and Folate Receptor alpha (FRα) on their surfaces, and preclinical studies have demonstrated the efficacy of CAR‐T‐targeted therapy [61]. Furthermore, in hematological oncology, the US Food and Drug Administration (FDA) approved anti‐CD19 CAR‐T therapy for B‐cell malignancies, highlighting a therapeutic breakthrough [62]. Finally, in solid tumors, Steven Rosenberg et al. developed an over‐the‐counter T‐cell therapy with ≥ 50% durable remission rates in melanoma [63].

9. CAR‐T Application in OC

9.1. Tumor‐Associated Glycoprotein Targets

9.1.1. MUC16 (CA125)

Mucin 16 (MUC16) is a transmembrane protein whose extracellular structural domain (MUC16‐ecto) forms the core target region for CAR design. Serum CA125 (MUC16 extracellular region) levels were found to be significantly elevated in > 80% of advanced OC patients [64]. Preclinical studies also showed that CAR‐T therapies targeting MUC16 exerted anti‐tumor effects in Pancreatic Cancer (PCa) [65], Lung Cancer (LC) [66], and OC [64]. Furthermore, a phase I clinical trial involving second‐generation MUC16‐CAR‐T with a 4‐1BB co‐stimulatory domain for recurrent OC revealed the following: (i) CA125 levels in ascites decreased by > 50% in some patients; and (ii) median Overall Survival (OS) was prolonged to 12 months (compared to the control group's 8 months). Among participants in trial cohorts 2 and 3 who received CAR‐T infusions, the incidence of CRS was 33.33% and 60%, respectively. Although the therapeutic efficacy of CAR‐T is promising, these results indicate that its clinical safety warrants further investigation.

9.1.2. Mesothelin (MSLN)

Previous research revealed MSLN overexpression in > 70% of ovarian plasmacytoid carcinoma patients [67]. Furthermore, Liang Zhiqing et al. [68] developed a tandem CAR construct targeting two secreted antigens to enhance the efficacy of CAR T cells via a dual antigen recognition strategy. The Tandem‐CAR T cells secreted IL‐12, which significantly reduced the risk of tumor antigen escape while increasing T cell proliferation and killing activity, thus enhancing the anti‐tumor immune response. Zhao Xudong et al. [69] further developed CAR‐T cells targeting both mesothelin and MUC16, a highly expressed antigen in OC. This dual‐targeting strategy establishes a nuanced equilibrium between efficacy and safety. On one hand, the simultaneous targeting of MSLN and MUC16 diminishes the likelihood of tumor immune evasion, as tumors must downregulate multiple surface antigens concurrently to evade T‐cell recognition. Conversely, this broader antigen coverage inherently raises the risk of off‐target effects. Given that both MSLN and MUC16 are expressed at baseline levels in healthy mesothelial tissues, including the pleura, pericardium, and peritoneum, tandem CARs may demonstrate heightened cross‐reactivity with normal tissues [70]. This expanded targeting spectrum suggests that the integration of a suicide gene switch may be necessary for clinical application to ensure patient safety. Moreover, in vitro experiments in a hormone‐immunodeficient mouse model showed that MSLN/MUC16 CAR T cells could effectively eliminate OC cells and their stem cells, with the intervention significantly inhibiting tumor growth, thus confirming its anti‐tumor potential in vivo.

9.2. Growth Factor (GF) Receptor Family Targets

9.2.1. FRα

High FRα expression was previously reported in 80% of OC patients, a phenomenon that correlated significantly with the degree of tumor malignancy and a poor prognosis [71]. Consequently, the scholars developed a second‐generation FRα CAR, designed based on the MOv19 single‐chain antibody (SCFV) and containing the 4‐1BB co‐stimulatory domain, which induced the secretion of inflammatory factors such as IFN‐γ and IL‐2 in vitro [72]. Furthermore, Georgios Coukos et al. conducted a Phase I clinical trial and found that FRα‐CAR‐T therapy, along with lymphocyte clearance pretreatment, significantly prolonged Progression‐Free Survival (PFS) and OS in patients [73].

9.2.2. Human Epidermal Growth Factor Receptor 2 (HER2)

Chang et al. [74] found a specifically high and low HER2 expression in OC and normal tissues, respectively, providing a theoretical basis for developing OC‐specific HER2‐targeting therapies. Although anti‐HER2 monoclonal antibodies (e.g., trastuzumab) significantly improved the prognosis of patients with HER2⁺ BC, drug resistance still limits their clinical application. Consequently, due to its potentially efficient anti‐tumor effect, HER2‐targeted CAR‐T cell therapy has recently become a key research hotspot. Sun et al. [75] successfully constructed HER2 CAR‐T cells that specifically recognized and killed HER2⁺ BC and OC cells in vitro, demonstrating the feasibility of this strategy.

9.3. TME‐Related Targets

9.3.1. Urokinase‐Type Plasminogen Activator Receptor (uPAR)

Monoclonal antibodies targeting uPAR (e.g., ATN‐658, 2G10) are currently on preclinical or clinical trials [76, 77]. Furthermore, uPAR‐specific CAR T‐cell therapy represents a new direction for tumor immunotherapy. In 2019, Wang et al. [78] constructed the first uPAR‐targeted CAR‐T cells, whose antigen recognition domain used a uPA Amino‐Terminal Fragment (ATF) instead of a conventional single‐chain antibody (SCFV) to mimic the uPA‐uPAR natural ligand‐receptor binding mode. The ATF‐CAR T cells showed specific recognition and significant killing activity against uPAR‐positive OC cells. Furthermore, in 2024, Amor et al. [79] found that uPAR‐CAR T therapy could be extended to fibrotic illnesses and oncology treatment, significantly removing senescent cells and ameliorating liver fibrosis in animal models, while also prolonging survival in a mouse model of Lung Adenocarcinoma (LUAD). Furthermore, given that uPAR is physiologically upregulated on activated endothelial cells and tissue‐resident macrophages during inflammation, this significantly heightens the risk of off‐target effects from uPAR‐CAR T cells. Clinically, this could result in severe capillary leak syndrome, while targeting macrophages might provoke intense systemic inflammation. Consequently, future clinical applications may require the incorporation of a suicide gene switch to mitigate the risk of these potential off‐target effects.

10. CAR‐NK Application in OC Treatment

Natural Killer (NK) cells, the innate immune system's core effector cells, belong to a subpopulation of innate lymphocytes that can directly kill malignant or virally infected cells. Based on surface markers and functional differences, NK cells could be divided into two main subpopulations: NK1 cells [CD56bright CD16dim/neg; which mainly secrete cytokines (e.g., IFN‐γ, TNF‐α) to mediate immunomodulation]; and NK2 cells (CD56dim CD16⁺; which account for 90% of NK cells in peripheral blood and are known for their immunomodulatory effects through the release of perforin and granzymes) [80]. Kärre et al. [81] proposed the ‘missing self’ hypothesis, positing that the presence of specific signals could influence NK cell activation. In their model, the interaction of Killer Immunoglobulin‐like Receptors (KIRs) with Major Histocompatibility Complex class I (MHC‐I) molecules on the surface of target cells regulated NK cell activation. Furthermore, normal cells expressed high levels of MHC‐I molecules, which bind to KIRs and transmit inhibitory signals to NK cells, thus preventing attacks on healthy tissues. Conversely, virally infected or cancerous cells downregulated MHC‐I, releasing KIR‐mediated inhibitory signals, thus activating the killing function of NK cells (Figure 3).

FIGURE 3.

Function of NK Cells | Normal cells express MHC class I molecules on their surface. The binding of inhibitory receptors on NK cells to these molecules results in suppression of the NK cell function, but does not affect their cytotoxicity (A). In abnormal cells (e.g., tumor cells), MHC class I molecules are typically downregulated, while ligands for NK cell‐activating receptors (such as NKG2DL) are often upregulated. The interaction between activating ligands and activating receptors (e.g., NKG2D) leads to the absence of inhibitory signals and the enhancement of activating signals, thereby inducing NK cell activation and cytotoxic targeting of abnormal cells. Furthermore, NK cells induce tumor cell apoptosis via the death receptor pathway. The release of cytokines and recruitment of other immune cells (e.g., macrophages) amplify the immune response (B). The NK cells may be activated via CD16‐mediated recognition of antibodies bound to target cells, exerting cytotoxicity through antibody‐dependent cellular cytotoxicity (ADCC) (C). The pattern was created with BioGDP.com.

Perforin, a key component of the perforin‐granzyme pathway [82], forms the transmembrane pores in the target cell membrane and promotes the inward flow of granzymes. Subsequently, granzymes, such as granzyme B, enter the target cell through these pores and activate the caspase cascade reaction, cleaving key proteins such as Poly ADP‐Ribose Polymerase (PARP) and inducing programmed apoptosis. The other mechanism involved in apoptosis is the death receptor pathway [83], which relies on key ligands such as Fas Ligand (FasL) and Tumor Necrosis Factor‐related Apoptosis‐inducing Ligand (TRAIL). These ligands bind to death receptors on the surface of target cells, including Fas and DR4/DR5, activating caspase‐8/10 and triggering apoptosis signaling. Antibody‐Dependent Cytotoxicity (ADCC) [84] also occurs when the NK cell surface CD16 (FcγRIII) recognizes the Fc segment of IgG antibodies bound to target cell antigens. This interaction could trigger granule cytotoxicity or activate the death receptor signaling pathway, thus synergistically enhancing the killing efficiency of NK cells.

In addition to their cytolytic functions, NK cells could also secrete various cytokines [85]. Among them, Interferon‐gamma (IFN‐γ) serves as the primary immune regulatory factor that drives TH1 polarization and activates CD8⁺ T cells to coordinate adaptive immune responses [86]. Furthermore, NK cells could secrete GFs such as FMS‐like Tyrosine Kinase 3 Ligand (FLT‐3L) and Granulocyte‐Macrophage Colony‐Stimulating Factor (GM‐CSF), which crucially regulate the differentiation of hematopoietic and immune cells. Meanwhile, NK cell‐produced chemokines such as XCL1, CCL3, and CCL4 could effectively recruit DCs and T cells to tumor sites, enhancing the anti‐tumor immune response cascade [87]. It is through these mechanisms that NK cells crucially mediate the immune response.

Unlike T cells, to properly function, NK cells do not rely on Human Leukocyte Antigen (HLA) compatibility, a phenomenon that might endow them with a lower risk of immune rejection in xenotransplantation. Conversely, the realm of allogeneic cell therapy, particularly over‐the‐counter immunotherapies such as CAR T‐cell therapy, still faces major challenges. Specifically, due to the restrictive nature of HLAs, CAR‐T cell therapy could trigger life‐threatening complications such as CRS, ICANS, and Graft‐versus‐Host Disease (GvHD). However, with NK cells, CRS risk is extremely low due to differences in their activation and signaling pathways [88]. These characteristics signify considerable advantages for the clinical application of CAR‐NK therapy. Mechanistically, excessive inflammatory cascades lead to severe CRS. In this process, effector cells induce the substantial activation of monocytes and macrophages, which are highly dependent on elevated levels of IL‐6 and GM‐CSF [15].

In contrast to CAR‐T cells, which significantly secrete factors that exacerbate the pathological cycle, activated NK cells display a distinctive cytokine secretion profile. They primarily produce IFN‐γ and a limited quantity of GM‐CSF, while exhibiting almost no secretion of IL‐6, the key molecule associated with the induction of CRS. Furthermore, when compared to CAR‐T cells, NK cells release lower absolute levels of pro‐inflammatory factors, including TNF‐α and GM‐CSF. This comparatively restrained secretion pattern effectively inhibits the myeloid‐cell‐driven inflammatory cycle, thereby mitigating the risk of severe CRS [88]. Furthermore, compared to the vigorous expansion and long‐term persistence of memory T cells in vivo, transplanted NK cells have a shorter physiological half‐life (typically surviving only a few days to a few weeks) [89, 90]. This transient nature guarantees that the effector response, although powerful, is self‐limiting. This characteristic helps to avoid side effects typically linked to CAR‐T cell therapy, such as enduring cytokine accumulation, extended systemic inflammation, and the risk of significant off‐target effects.

Notably, CAR engineering is not limited to T cells, highlighting its potentially unique biological properties, especially concerning NK cells. Compared to CAR‐T cells, the intracellular signaling module of CAR‐NK cells is not limited to the traditional CD3ζ chain‐dependent pathway, but integrates multiple Immunoreceptor Tyrosine‐based Activation Motif (ITAM) signaling elements. According to reports, the NK cell‐specific articulatory molecule DAP12 (DNAX activation protein 12) can specifically recruit Syk and ZAP70 kinases (but not T cell‐associated CD3ζ kinases) through its ITAM structural domain, yielding a differentiated signaling network that significantly enhances the activation threshold and cytokine secretion capacity of NK cells [91]. Particularly, NKG2D, a natural activation receptor for NK cells, could be used as the extracellular antigen recognition domain of CAR and chimerically integrated with the DAP12 intracellular signaling module, yielding an optimized CAR‐NK with synergistic effects of endogenous signaling. Such a design could multiplex the DAP12‐NKG2D signaling axis intrinsic to NK cells, markedly enhancing the receptor activation efficiency and targeted killing specificity [92]. However, it is crucial to recognize that prolonged activation of the DAP12‐NKG2D signaling pathway introduces significant safety concerns. This powerful pathway can lead to early exhaustion of NK cells and may also cause unintended off‐target effects. This risk is particularly increased because NKG2D ligands (NKG2DL) are often upregulated in healthy tissues during physiological stress or inflammatory states [93]. Furthermore, significant evolutionary differences between mouse models and humans in the DAP12 signaling network and NKG2D ligand expression patterns greatly complicate preclinical toxicity assessments [94]. These species‐specific differences limit the predictive value of standard in vivo safety pharmacology studies, creating an urgent need to develop humanized models or conduct studies with non‐human primates to validate safety for clinical translation.

In solid tumor therapy, the heterogeneity of tumor antigens and the IME are some of the key factors that could limit the efficacy of CAR therapy. However, with advances in cancer molecular typing, engineered immunotherapy based on NK cells demonstrated a unique potential for application. For instance, in HER2⁺ BC, researchers constructed primary CAR NK cells targeting HER2 antigens using gene editing technology, demonstrating significant tumor eradication ability in mouse models. This approach could overcome traditional monoclonal antibody resistance [95]. Furthermore, anti‐HER2 NK92 cell therapy has entered translational research, and preclinical data showed that engineered NK92 cells could cross the Blood–Brain Barrier (BBB) while depleting tumor stem cells [96]. Moreover, a phase I/II clinical trial is currently underway to assess its safety and efficacy [97].

As earlier mentioned, MSLN‐targeted CAR‐T therapy represents a breakthrough in OC treatment. Rizwan Romee et al. [98] recently proposed a pertinent innovative solution. First, combining Cytokine‐Induced Memory‐Like (CIML) NK cells with MSLN‐specific CAR yields MSLN‐CAR‐CIML‐NK cells with enhanced NK cell proliferation and persistence properties. Additionally, the CAR‐MSLN design precisely targets mesothelin antigens, which are highly expressed in Epithelial Ovarian Cancer (EOC).

To evaluate the potential of FRα‐targeted CAR‐NK cells in treating OC, Jianxin Guo et al. [99] systematically constructed three FRα‐specific CARs with different signaling domains: FRα‐ζ (first‐generation CAR, which contains only the CD3ζ intracellular signaling module); FRα‐28ζ (second‐generation CAR, which integrates the CD28 co‐stimulatory domain with the CD3ζ signaling chain); and FRα‐28BBζ (third‐generation CAR, which combines the CD28 and 4‐1BB (CD137) dual co‐stimulatory domains with the CD3ζ signaling chain). Notably, NK‐92 cells, a human‐derived NK cell line, when transformed with the above CARs, exhibited specific killing ability against FRα‐overexpressing OC cells in an in vitro model. The EC₅₀ of first‐generation FRα‐ζ CAR‐NK cells against SKOV3 cells was 12.3:1, with a maximum killing rate of 42.6% at an effector‐to‐target ratio of 20:1. In contrast, the EC₅₀ of second‐generation FRα‐28ζ CAR‐NK cells was 4.7:1, with a maximum killing rate of 71.8% at an effector‐to‐target ratio of 20:1. The third‐generation FRα‐28BBζ CAR‐NK cells had an EC₅₀ of 3.9:1, with a maximum cytotoxicity of 78.2% at the same effector‐to‐target ratio. Furthermore, compared to first‐generation CARs that rely solely on CD3ζ signaling, the FRα‐28ζ and FRα‐28BBζ CAR‐NK cells, which incorporate co‐stimulatory domains, exhibited significantly enhanced cytotoxicity at the same effector‐to‐target ratio. These findings suggest that co‐stimulatory signaling plays a key role in the sustained activation and maintenance of NK cell function [100].

Despite these promising characteristics, the clinical translation of CAR‐NK therapy encounters several challenges. First, in contrast to T cells, NK cells exhibit a relatively short lifespan in vivo and possess limited expansion capacity without ongoing support from exogenous cytokines. Second, variations in the quality of donor cells can affect batch‐to‐batch consistency, thereby impacting therapeutic efficacy. Lastly, although CAR‐NK cells are marketed as an “off‐the‐shelf” product, the processes of cryopreservation and thawing significantly compromise their viability, cytotoxicity, and in vivo persistence [101]. To enhance in vivo persistence, a new generation of CAR‐NK cells is being engineered to co‐express autocrine cytokines (such as IL‐15) exogenously, thereby providing sustained support for their autonomous survival and proliferation [102]. To address the challenge of donor heterogeneity, standardizing the cell source through the use of induced pluripotent stem cells (iPSCs) offers a scalable and uniform approach, ensuring rigorous batch‐to‐batch consistency.

11. CAR‐M Application in OC Treatment

Among the immune cells that infiltrate the solid TME, TAMs are the most prominent. In addition to regulating tumor invasion and metastasis, these cells also mediate treatment resistance among other mechanisms, thus driving tumor progression. Furthermore, their function is highly dependent on the polarization status, specifically the pro‐inflammatory M1 phenotype and the inflammatory M2 phenotype, which correlate closely with the pathological evolution of tumors [103]. Besides inducing tumor cell apoptosis via the TNF‐α‐TNFR1‐caspase8 pathway, M1‐type macrophages, characterized by a pro‐inflammatory phenotype, also secrete IL‐12, thus activating the STAT4‐T‐bet axis, which, in turn, enhances CD8⁺ T‐cell cytotoxicity [104, 105]. Conversely, M2‐type macrophages exhibit an anti‐inflammatory phenotype and could upregulate PD‐L1 through IL‐10‐STAT3 signaling, leading to T‐cell depletion. Moreover, the TGF‐β‐SMAD3 pathway drives Epithelial‐Mesenchymal Transition (EMT) and VEGF‐mediated neovascularization, ultimately remodeling the IME [106, 107] (Figure 4).

FIGURE 4.

Functions of Macrophages | Macrophages exist in two polarization states: M1 and M2 subtypes. In the tumor microenvironment (TME), the IFN‐γ signaling can inhibits M2 polarization and promotes M1 polarization. M1 macrophages activate CD8⁺ T cells and mediate tumor cell apoptosis via the FasL–Fas death receptor pathway and the TNF‐α‐TNFR1 apoptotic pathway. Moreover, M1 macrophages regulate matrix metalloproteinase (MMP) activity to remodel the immunosuppressive microenvironment. In contrast, M2 macrophages induce T cell exhaustion, drive epithelial‐mesenchymal transition (EMT), and enhance VEGF‐mediated angiogenesis, creating an immunosuppressive TME. M2 macrophages also secrete MMP9 and LOX (lysyl oxidase) to remodel the extracellular matrix (A). Following tumor cell clearance via Fcγ receptor‐mediated antibody‐dependent cellular phagocytosis (ADCP), macrophages leverage MHC‐II molecules and co‐stimulatory markers (e.g., CD80/CD86) to present tumor‐associated antigens (TAAs) to CD4⁺ T cells. Alternatively, they activate tumor‐specific cytotoxic CD8⁺ T cells through cross‐presentation mechanisms (B). The pattern was designed using the BioGDP.com.

According to research, TAMs are significantly enriched in solid tumors such as breast, colorectal, lung, and ovarian cancers [108, 109, 110, 111], and their M2‐like phenotype correlates with tumor growth, metastasis, and a poor prognosis. In the OC microenvironment, M2‐polarized TAMs drive ECM transformation and secrete MMP9/LOX, remodeling the ECM through the TGF‐β/SMAD signaling axis and significantly promoting tumor invasion and metastasis [112, 113]. On the other hand, M1‐type TAMs activate tumor‐specific CD8⁺ T cells via the IL‐12/STAT4 pathway and induce the apoptosis of FasL–Fas death receptor pathway‐dependent OC cells [114, 115]. Furthermore, preclinical studies confirmed that TAM polarization modulation prolonged the survival of ID8 hormone‐treated mice by 2.3‐fold [116].

Macrophages have shown significant phenotypic plasticity in TIME. According to reports [117], M1 macrophages could reverse the immunosuppressive phenotype of M2 macrophages by secreting cytokines such as Tumor Necrosis Factor‐α (TNF‐α) and IFN‐γ and systematically reverse the IME by modulating Matrix Metalloproteinase (MMP) activity, which mediates tumor mesenchymal remodeling and systematically reverses the IME. Furthermore, the sustained release of CCL5 (C‐C Motif Chemokine Ligand 5), CXCL9 (C‐X‐C Motif Chemokine Ligand 9), and other chemokines in the TIME forms a positive feedback loop that induces CD8⁺ T cells and NK cells to specifically infiltrate the tumor site [118]. Additionally, the formation of immune synapses could enhance the effector cell‐mediated tumor killing efficiency while inhibiting tumor killing via IFN‐γ signaling, thus suppressing M2‐type macrophage polarization [119]. Moreover, pro‐inflammatory microenvironment‐induced vascular normalization and ECM remodeling could significantly enhance the tumor penetration capabilities of drugs such as Immune Checkpoint Inhibitors (ICIs) [120].

After macrophages have cleared tumor cells via Fcγ Receptor (FcγR)‐mediated Antibody‐Dependent Cellular Phagocytosis (ADCP), their surface MHC‐II molecules synergize with co‐stimulatory molecules (e.g., CD80/CD86) to efficiently deliver TAAs to CD4⁺ T cells and activate CD8⁺ CTLs via cross‐presentation [121].

Notable breakthroughs have been achieved with CAR‐M. For instance, CAR‐engineered M1 Macrophages (CAR‐M1) can specifically recognize tumor antigens (e.g., HER2 or MSLN), and enhance their phagocytosis and antigen‐presenting ability to OC cells, while synergistically fighting tumors by secreting chemokines to recruit T cells. In this regard, it is noteworthy that chimeric adenoviral infections and cytokines could differentiate M2 macrophages into the M1 phenotype. Furthermore, CAR macrophages (CAR‐MS) demonstrated antigen‐specific phagocytosis and tumor clearance in vitro [122]. Reprogramming of macrophages with CAR significantly improved immune cell infiltration into the tumor area, with HER2‐targeted CAR‐147 macrophages significantly increasing the density of CD8⁺ T cell tumor infiltration through the formation of a chemotactic gradient via the secretion of CXCL9/CCL5 [114]. Additionally, CAR‐147 upregulated MMP‐12 at a rate of 63.4%, with a 4.7‐fold increase in T‐cell infiltration efficiency, thus mediating tumor interstitial collagen (Col I/III) degradation. After infusion, peripheral blood TNF‐α and IL‐6 levels in mice decreased by 71.3% and 69.8%, respectively, significantly reducing CRS risk [116]. Abdin et al. further assessed the scalability of generating functional CAR macrophages from various stem cell sources, including induced Pluripotent Stem Cells (iPSCs). They particularly focused on the efficacy of anti‐CD19 CAR macrophages in eradicating CD19⁺ leukemia, elucidating the versatility and applicability of CAR macrophages in targeting specific cancer types. According to the results, CAR macrophages exhibited enhanced antigen‐dependent phagocytosis of CD19‐targeted cancer cells and increased pro‐inflammatory responses, regardless of the stem cell source. These findings illuminate the potential challenges in obtaining cells for CAR cell therapy [123]. From a clinical translation perspective, utilizing iPSC‐derived CAR‐M cells introduces several challenges, particularly in regulatory and manufacturing domains. A primary concern is the consistency across batches; variations in differentiation paths can result in functional heterogeneity, complicating predictions of in vivo pharmacokinetics and leading to inconsistent therapeutic outcomes. Additionally, insufficient maturation into terminally differentiated macrophages diminishes the overall phagocytic capacity and increases the risk of teratoma formation due to the presence of residual pluripotent stem cells [124]. Ensuring rigorous quality control and achieving uniform differentiation can mitigate the risk of unpredictable toxicity and ensure the safety of clinical translation for off‐the‐shelf iPSC‐CAR‐M therapies.

The exploration of macrophage presence in cancer, particularly in OC, is an emerging field with significant limitations in clinical studies. While TAMs have been implicated in various malignancies, including BC, their role in OC remains underexplored. For instance, a study reported that the presence of TAMs in cancer correlates with a worse prognosis, although this phenomenon is yet to be explored in vivo [125].

However, the clinical translation of CAR‐M therapy encounters numerous challenges. Firstly, unlike lymphocytes, terminally differentiated macrophages do not possess the capacity to proliferate in vivo, necessitating the infusion of significantly high doses of cells. Most importantly, while CAR‐M cells are infused in the antitumor M1 state, the cytokine‐rich TME of ovarian cancer easily repolarizes them into a pro‐tumor, immunosuppressive M2 phenotype. This shift could potentially accelerate disease progression if not meticulously regulated [126]. To prevent M2 repolarization, next‐generation CAR‐M constructs are engineered to co‐deliver potent pro‐inflammatory cytokines such as IL‐12 or to employ CRISPR/Cas9 technology to knockout immune checkpoint molecules, such as SIRPα. This approach aims to permanently lock macrophages in the anti‐tumor M1 state [26].

12. Conclusion

The field of CAR‐mediated immunotherapy is rapidly evolving, particularly in OC treatment. This review summarizes the latest findings on CAR‐T, CAR‐NK, and CAR‐M, as well as their advantages and disadvantages (Table 1), highlighting their unique benefits and limitations.

TABLE 1.

Comparison of CAR‐T, CAR‐NK, and CAR‐M.

Parameter	CAR‐T	CAR‐NK	CAR‐M
Cell source	Autologous T cells MHC‐matched allogeneic T cells (under clinical exploration)	Peripheral blood mononuclear cells (PBMCs) Umbilical cord blood (UCB) Hematopoietic stem cells (HSCs) iPSC‐derived NK cells NK cell lines	Autologous monocyte‐derived macrophages iPSCs‐derived macrophages immortalized lines
Cytotoxicity mechanisms	CAR‐dependent cytotoxicity: T cell receptor activation induces apoptosis	Dual mechanisms: CAR‐dependent targeting Innate receptor‐mediated killing	CAR‐dependent phagocytosis Tumor microenvironment (TME) remodeling Antigen presentation to activate adaptive immunity
Cytokine release syndrome and neurotoxicity	All‐grade CRS incidence: 70%–90% Severity: Mostly Grade 3–4 High risk of neurotoxicity (e.g., ICANS)	All‐grade CRS incidence: < 10% Severity: Mostly Grade 1–2 Neurotoxicity rarely reported	Limited clinical data Preclinical models suggest localized inflammation and potential pro‐tumor effects if M2 polarization occurs
Dose‐limiting toxicities observed in clinical trials	On‐target, off‐tumor toxicity Target antigen heterogeneity CRS	Generally low risk of DLTs Potential for on‐target, off‐tumor toxicity	Risk of unregulated localized hyper‐inflammation or collateral tissue necrosis Risk of systemic macrophage hyperactivation Paradoxical risk of pro‐tumor M2 phenotype repolarization driven by the immunosuppressive TME
Intra‐tumoral infiltration	Low: Restricted by TME immunosuppressive factors (e.g., PD‐L1, TGF‐β)	Moderate: NK cells infiltrate low MHC‐I tumors but are inhibited by Tregs	High: Macrophages inherently infiltrate TME and penetrate dense stroma
Clinical application and star targets	Maturation phase Angioma (target:CD19) [112] Solid tumor Breast & Lung Cancer Targets: HER2 [113], MUC1 [114], ROR1 [115] Neuroblastoma & Osteosarcoma Target: GD2 [116] Stomach Targets: Mesothelin, MUC16, CLDN18.2 [117]	Early clinical Angioma (target:CD19) [118] Solid tumor Glioblastoma Targets: HER2 [119], NKG2DL [120] Stomach Target: TRII/21R [121]	Start‐up: Less clinical data available Solid tumor Glioblastoma Target: NKG2DL [120]
Ovarian cancer targets	MUC16, FRα, UPAR, Her2, Mesothelin, TAG72 [122]	NKG2DL, Mesothelin, CD24, FRα, Claudin‐6 [123], CD133 [124]	/

Note: CLDN18.2: Claudin18.2. TRII/21R: Consists of TGF‐β receptor II and IL‐21 receptor.

Although CAR‐T cell therapy has revolutionized cancer treatment, particularly for hematological malignancies, its application in solid tumors, including OC, faces significant challenges. The TME of solid tumors is often immunosuppressive, causing a poor response to CAR‐T therapy [127]. Limitations such as CRS and neurotoxicity further complicate the use of CAR‐T cells in OC treatment. Therefore, current research efforts should prioritize achieving a balance between improving therapeutic efficacy and maintaining safety in clinical applications [128, 129, 130, 131]. In summary, there are three key areas for enhancing therapeutic efficacy: (1) Development of novel antigen‐binding structural domains with enhanced specificity; (2) Designing bispecific CAR‐T cells for dual‐target engagement; and (3) Incorporation of intracellular signaling structural domains for enhanced co‐stimulatory signals. While the three approaches outlined above focus on addressing tumor immune evasion, it is equally important to consider the necessity of minimizing toxicity during clinical translation. In contemporary CAR design, considerable attention has been directed towards strategies that aim to reduce adverse effects, including: (1) Precise calibration of CAR activation thresholds for optimal signaling while maintaining a balance between agonistic and antagonistic signals; (2) Co‐expression of dominant negative receptors to regulate effector function; and (3) Strategic modification of cytokine receptor pathways to attenuate or enhance specific cytokine‐mediated signaling cascades. These optimization strategies could improve treatment efficacy while mitigating potential Adverse Effects (AEs). These optimization strategies enhance treatment efficacy while reducing the risk of adverse events. The three safety‐oriented strategies outlined above are essential for minimizing potential dose‐limiting adverse effects and expanding the therapeutic index of CARs in the clinical management of solid tumors.

On the other hand, due to their innate ability to target tumor cells without prior sensitization, CAR NK cells represent a promising alternative. These cells could overcome some of the limitations associated with CAR‐T cells, such as the need for extensive in vitro expansion and the risk of allogeneic rejection. However, the efficacy of CAR‐NK cells in solid tumors remains unclear, necessitating further studies, especially regarding their interactions in the TME.

Additionally, CAR‐macrophages, an emerging paradigm for CAR‐engineered immunotherapy, harness the innate immune system's ability to phagocytose tumor cells. The potential of CAR‐M in OC treatment is still being explored, with studies suggesting that it may be a complementary strategy to CAR‐T and CAR‐NK cells. However, to optimize the therapeutic potential of macrophages, the issues of macrophage polarization and the impact of the TME on their function should be addressed.

Overall, although CAR‐T cells have gained a firm foothold in cancer immunotherapy, their application in OC is subject to several limitations. Moreover, while CAR‐NK cells and CAR‐M cells present promising alternatives with notable advantages, their potential to enhance treatment outcomes is tempered by specific challenges. In the case of CAR‐NK therapy, there are intrinsic limitations regarding in vivo persistence without ongoing cytokine support, and variations in donor cell quality can significantly affect batch‐to‐batch consistency. On the other hand, CAR‐M therapy raises primary safety and efficacy concerns related to the risk of uncontrolled local phagocytosis overactivation and the instability of the polarized state. Furthermore, the proliferative stroma and immunosuppressive microenvironment of ovarian cancer complicate the application of all three therapies. CAR‐T cells often struggle to penetrate the tumor core effectively and are particularly vulnerable to rapid exhaustion caused by PD‐L1 and Tregs. Although CAR‐NK cells can leverage innate receptors to enhance their CAR targeting capabilities, they also encounter challenges in physical penetration and are susceptible to suppression by TGF‐β within the TME. In contrast, CAR‐M cells have an inherent ability to infiltrate and colonize solid tumors, which sets them apart. Their distinctive advantage lies in their ability to secrete matrix metalloproteinases (MMPs), enabling them to degrade the dense extracellular matrix and cross‐present tumor antigens to endogenous T cells. This dual functionality allows them to act not only as direct effector cells but also as modulators of the TME. These specific challenges underscore the need for ongoing, mechanism‐oriented research to fully harness the safe and effective clinical application of these CAR‐engineered cells in the context of ovarian cancer.

Author Contributions

Yisen Cao: conceptualization, investigation, writing – original draft, methodology, validation, visualization, writing – review and editing, software, formal analysis, data curation. Liying Wang: conceptualization, investigation, writing – original draft, methodology, validation, visualization, writing – review and editing, software, formal analysis, data curation. Yuhang Zhang: conceptualization, investigation, methodology, writing – original draft, writing – review and editing, visualization. Yuan Ren: conceptualization, investigation, writing – original draft, writing – review and editing, visualization. Liang Wang: conceptualization, investigation, funding acquisition, writing – original draft, writing – review and editing, validation, methodology, visualization, project administration, formal analysis, software, supervision, resources, data curation.

Funding

The study was supported by the National Natural Science Foundation of China (NSFC) (82303734), the Science and Technology Innovation Joint Fund Project of the Fujian Provincial Health Commission (2024Y9535), and the Fujian Provincial Natural Science Foundation (2025J01201).

Ethics Statement

The authors have nothing to report.

Consent

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.