Emerging CAR immunotherapies: broadening therapeutic horizons beyond cancer

Abstract

Chimeric antigen receptor-based immunotherapy has transformed cancer treatment, especially for hematologic malignancies like acute lymphoblastic leukemia and diffuse large B-cell lymphoma. Innovations in CAR design from first-generation constructs relying on CD3ζ signaling to next-generation CARs with co-stimulatory domains have enhanced T cell persistence and antitumor efficacy. Despite these successes, translating CAR-T therapy to solid tumors faces significant challenges, including antigen heterogeneity, immunosuppressive tumor microenvironments, and toxicities such as cytokine release syndrome and neurotoxicity. To overcome these hurdles, CAR therapies involving alternative immune cells are currently being developed, such as CAR-natural killer, CAR-T regulatory (Treg), CAR-macrophages (Ms), and others, each offering distinct biological advantages and potential for broader applications. Beyond oncology, CAR approaches are being explored for autoimmune diseases, infectious diseases, and fibrosis, expanding their therapeutic scope. Manufacturing complexities and safety concerns related to gene modification also highlight the need for scalable, safe production methods, including non-viral gene delivery systems. This review summarizes the evolution, current applications, and future prospects of CAR-based therapies, emphasizing the importance of ongoing innovation to enhance specificity, safety, and clinical efficacy across diverse disease contexts.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10238-025-01820-x.

Background

The development of chimeric antigen receptor (CAR) T cell therapy was a landmark event in the field of cancer treatment, as it integrates the adaptive immune system’s functionality with precise molecular targeting. Since the first CAR-based therapies were engineered in the late 1980s [1, 2], this pioneering innovation has undergone extensive refinement and has become an integral component of contemporary oncology. Early studies of the efficacy of CAR T cell therapy in treating hematologic malignancies, such as acute lymphoblastic leukemia (ALL) and diffuse large B cell lymphoma, demonstrated its promise, with remission rates surpassing those achieved with conventional treatments [3]. Developing CAR T cells that are effective against solid tumors is more challenging because of several substantial barriers, including the immunosuppressive tumor microenvironment (TME), antigen escape, lack of physical infiltration, and systemic cytotoxic effects, such as the induction of cytokine release syndrome (CRS) and neurotoxicity [4–6]. Nevertheless, new CAR T cells are now being evaluated in numerous clinical trials, with a focus on improving their efficacy against both hematologic malignancies and solid tumors [7]. The process used to generate CAR T cells underlies both the therapeutic promise and the challenges associated with the implementation of this technology. For example, autologous CAR T cells, which are derived from a patient’s T cells, have exhibited therapeutic efficacy but are constrained by variability in the starting material, lengthy manufacturing timelines, and patient-specific factors [8]. Efforts to reduce production time and enhance consistency, including the use of automated cell culture systems and standardized protocols, have had a positive impact; however, production limitations continue to delay the provision of timely treatment for high-risk patients. While these limitations and other logistic and economic obstacles emphasize the need for new strategies to boost the efficacy and applicability of CAR-based therapies [9], ongoing advances in CAR-based therapies suggest a future of broader clinical applicability and affordability in personalized cancer immunotherapy. In this review, we present a comprehensive overview of the current state of CAR T cell research and therapy, encompassing technological developments, clinical applications, limitations, and prospects.

Development of CAR-based systems

CAR T cell therapy was initially conceived as a technology that would enable target-specific T cell mediated cytotoxicity independent of major histocompatibility complex (MHC) restriction via a single-chain variable fragment (scFv) specific for a tumor-associated antigen [1, 2]. A schematic illustration of the structural evolution of CAR designs across successive generations is presented in Fig. 1. The first generation of CAR T cells featured an extracellular scFv derived from an antibody to confer antigen specificity. This scFv was fused to a transmembrane domain and an intracellular signaling domain derived from CD3ζ, which mediates T cell activation through the endogenous T cell receptor (TCR) [10–12]. However, these CAR T cells, which depended solely on the CD3ζ chain to mimic TCR signaling, exhibited limited proliferation, engraftment, cytotoxicity, and efficacy in clinical trials [13]. Subsequently, modifications were made to enhance their cytotoxicity, cytokine production, and proliferation in response to target stimulation. This limitation spurred the development of second-generation CARs.

Fig. 1.

Diagrammatic representation of the structural evolution of CAR generations

In second-generation CAR T cells, a co-stimulatory domain derived from CD28 or 4-1BB was inserted between the transmembrane and CD3ζ signaling domains. CD28 confers rapid and robust activation, it is often associated with limited persistence, whereas 4-1BB promotes long-term survival and memory formation. The dichotomy between these two signaling motifs has driven differential CAR design tailored to tumor biology and treatment goals. Notably, the first successful clinical trials involved second-generation CD19-targeted CAR T cells, with sustained therapeutic responses observed in patients with B cell leukemia [14, 15]. Consequently, in 2017, the US Food and Drug Administration (FDA) approved the first CD19-targeted CAR T cells for the treatment of hematologic malignancies. Currently, all FDA-approved CAR-based products employ second-generation technology [16].

Third-generation CAR T cells were developed to optimize T cell activation through the incorporation of additional co-stimulatory domains. Preclinical studies have shown that third-generation CAR T cells exhibit enhanced proliferative capacity and prolonged persistence compared to second-generation CAR T cells in CD19-targeted therapies [17], and clinical trials have shown that third-generation CAR T cells demonstrate promising efficacy and reduced procedure-specific toxicity [18, 19]. However, third-generation CAR T cells have been linked to higher rates of severe adverse effects (AEs) and more rapid T cell exhaustion, which is likely due to the overactivation of co-stimulatory signaling pathways [20, 21]. In addition, this type of therapy does not result in favorable outcomes in all patients [17, 18]. In some cases, the overactivation of T cells has led to accelerated exhaustion and increased adverse events. This discrepancy underscores a critical insight in CAR engineering: more co-stimulatory signals do not necessarily equate to better efficacy, and signal tuning, rather than stacking, may be essential to optimize therapeutic outcomes.

Fourth-generation CAR T cell-products include T cells redirected for universal cytokine-mediated killing (TRUCKs) [22], universal CAR T cells [23], and “armored” CAR T cells [24]. These constructs are often based on second-generation CAR T cells and feature modified intracellular signaling domains that enable cytokine secretion, which enhances T cell activation and antitumor activity [25, 26]. However, fourth-generation CAR T cells exhibit markedly reduced efficacy against solid tumors and have been associated with issues such as “on-target off-tumor” activation (TRUCKs) and unintended cytokine release in healthy tissues, resulting in systemic toxicity [27].

Fifth-generation (or next-generation) CAR T cells are now under development to address the limitations of earlier constructions. Constructs are being designed to enhance the survival and antitumor efficacy of CAR T cells within the TME and to target low-antigen-density tumors and multiple antigens [28, 29].

Fourth- and fifth-generation CAR T cells represent a paradigm shift toward the functional reprogramming of T cells, moving beyond conventional co-stimulatory enhancements. While these designs expand the functional scope of CAR T cells, such as through autocrine cytokine loops, inducible modules, or resistance to immunosuppressive signaling, they introduce new layers of complexity. These include risks of off-tumor cytotoxicity, uncontrolled cytokine release, and increased immunotoxicity, especially in the immunosuppressive microenvironment of solid tumors where trafficking, infiltration, and persistence remain challenging. As a result, recent efforts in CAR design have shifted from simply stacking intracellular domains to developing context-responsive, tightly regulated constructs that are more adaptable to the tumor milieu. This transition from “more signals” to “better signals” reflects a conceptual evolution in CAR engineering, one that prioritizes precision, safety, and tunability over brute-force activation.

Nanobody (Nb)-based CAR T cells have recently emerged as a promising alternative to conventional scFv-based constructs. Nbs are single-domain antibodies derived from camelid heavy-chain-only antibodies and are composed solely of a variable heavy chain. Their small size (approximately 12–15 kDa), lack of light chains, and structural simplicity confer advantages, such as improved solubility, thermal stability, and ease of engineering, as well as the ability to target sterically hindered epitopes inaccessible to traditional scFvs [30]. The efficacy of Nbs against myeloid leukemia [31] and solid tumors, such as glioblastoma, [32] cervical cancer [33], and large solid tumors [34], has been reported in preclinical and clinical studies.

CAR gene delivery systems

In parallel with advances in the structure of CAR T cells, novel delivery platforms are being developed to boost the efficacy and application of CAR T cells. CAR gene transfer into T cells relies on two major types of delivery systems: virus-based and nonviral methods. The respective advantages and limitations of these approaches are outlined in Table 1, offering a concise comparison across critical parameters such as transduction efficiency, safety, and clinical feasibility. Complementing this, Fig. 2 provides a more detailed overview of the gene delivery landscape, including schematic illustrations of typical delivery strategies and a direct comparison of retroviral and lentiviral vectors, commonly used in clinical settings.

Table 1.

Advantages and disadvantages associated with the different delivery platforms

Delivery platform	Method	Advantages	Disadvantages	References
Viral	Lentivirus	- High transduction efficiency - Stable gene integration - Large transgene capacity (~ 10 kb) - Prefer both dividing and non-dividing cells - Long-term safety	- High production costs - Risk of insertional mutagenesis	[37, 281, 282]
	Gamma-retrovirus	- High transduction efficiency - Stable gene integration - Simple vector design	- Genomic integration risk remains - Prefers dividing cells	[36, 282, 283]
	Adeno-associated virus (AAV)	- Low immunogenicity - Long-term expression	- Limited packaging capacity - Transient expression in dividing cells - Lower transduction efficiency	[39, 42]
Non-viral	Sleeping beauty transposon	- Low production costs - Simplified manufacturing - Long-term expression	- Random genomic insertion - Low transduction Efficiency	[45, 284]
	Piggybac transposon	- Large gene insertion - Stable integration	- Insertional mutagenesis risk	[55, 285]
	CRISPR/Cas9	- Specific gene editing - High efficiency - Safe harbor integration	- Off-target effects - Electroporation toxicity	[69, 286]

Fig. 2.

Overview of gene delivery methods and comparative analysis of retroviral and lentiviral vectors in CAR-T cell manufacturing

Viral vectors

Currently, most CAR T cells are produced by transferring CAR genetic material into T cells via virus-based vectors. Gamma retroviruses and lentiviruses have been adapted for this purpose. The efficacy of gamma-retroviral vectors was first demonstrated in clinical gene therapy trials for the treatment of severe combined immunodeficiency-X1 [35], and they have proven to be effective in the generation of CAR T cells [36]. However, gamma-retroviral vectors carry a risk of oncogenesis because of their tendency to integrate into gene promoters. LVs are appealing because of their stable integration, established in vitro technology, and growing adoption in clinical applications. Compared to early-generation LVs, third-generation LVs are more efficient, have a higher genetic payload capacity, and have improved safety profiles (e.g., they do not replicate and have reduced pathogenicity) [37]. Hence, gamma-retroviral vectors and LVs are commonly used for CAR T cell generation because of their high transduction rates and long-term stable transgene expression. In addition, LVs are not restricted by cell division, enabling efficient transduction of a broad range of cell types, including noncycling, terminally differentiated cells [38].

Adeno-associated virus (AAV) is a small, nonpathogenic, single-stranded DNA virus that has gained significant attention as a therapeutic gene delivery vector in recent years. It can efficiently transduce both dividing and nondividing cells [39], and it is widely employed as a vector for introducing genetic material into target cells to treat or prevent various diseases in clinical settings, across academia, and in industry [40]. AAV has been shown to mediate persistent transgene expression and long-term correction of disease phenotypes with minimal toxicity in both clinical and preclinical studies [41]. In a humanized mouse model of T cell leukemia, a sufficient volume of human CAR T cells capable of inducing tumor regression was generated in vivo via the injection of an AAV vector containing a CD4-specific CAR gene [42]. To date, more than 300 clinical trials involving AAV have either been completed or are ongoing [43].

Nonviral delivery systems

Nonviral delivery systems are a promising alternative to viral vectors in the generation of CAR T cells because of their higher cargo capacity and lower immunogenicity, insertional mutagenicity, and production costs. These systems facilitate either transient gene expression or stable transgene integration. While transient gene expression offers flexibility, its short-lived potency limits its use in CAR T cell therapies that require sustained in vivo expansion, memory generation, and immunosurveillance. Conversely, stable transgene integration supports longer-term therapeutic outcomes.

The widely adopted Sleeping Beauty (SB) transposon system offers enhanced safety, scalability, and faster production compared to viral vectors [44–46]. It enables long-term gene expression across various applications, including cancer immunotherapy and genetic disease treatment [47]. For instance, SB has been used to generate anti-CD19 CAR T cells from peripheral and cord blood through plasmid DNA or mRNA delivery [48, 49]. Clinical trials with SB-engineered CAR T cells have demonstrated sustained responses in treating B-ALL and non-Hodgkin’s lymphoma [50, 51]. Advanced techniques, such as shortening the ex vivo culture time to two weeks and incorporating “UltraCAR-T cells” that co-express membrane-bound interleukin (IL)-15, have further improved the safety and response rates in relapsed acute myeloid leukemia (AML) and ovarian cancer patients [52, 53]. Additionally, trials such as the CARAMBA trial, which focuses on multiple myeloma, continue to explore the feasibility and safety of SB-based CAR T cell therapies, reinforcing their potential in cancer treatment [54]. Nonetheless, despite its lower risk of insertional mutagenesis compared to retroviral vectors, the SB system still integrates semi-randomly within the genome, raising concerns about long-term genotoxicity. Moreover, the efficiency of SB-mediated transposition remains relatively lower than that of viral vectors, which may limit its application in rapidly progressing diseases requiring quick T cell manufacturing.

The precision, efficiency, and flexibility of the PiggyBac (PB) transposon system make it a prime candidate for inclusion in CAR T cell production. Derived from Trichoplusia ni, it delivers PB transposase DNA or mRNA and a transfer plasmid carrying the genetic cargo [55–57]. Recent advances, such as optimizing CAR constructs with smaller cassettes and integrating survival-promoting cytokines, have enhanced PB-based CAR T cell therapy [57–59]. The use of the PB system in clinical trials has gained momentum, partly because of preclinical data confirming its safety and efficacy [60–63]. However, PB’s propensity to integrate into transcriptionally active regions, while advantageous for sustained expression, also poses a potential risk of dysregulating endogenous genes. Ongoing refinements such as the development of "safe harbor" targeting and hyperactive transposase variants aim to mitigate these risks.

Nanocarriers are another promising delivery system for introducing CAR genetic material into T cells. They exhibit minimal off-target toxicity and immunogenicity, can be manufactured at scale, and can be tailored via conjugation to various biomaterials to achieve specific targeting [64]. Nevertheless, challenges such as endosomal escape, nucleic acid degradation, and limited nuclear delivery remain key hurdles that must be addressed to realize their full translational potential in CAR T cell production.

Advances in gene-editing technologies (e.g., transcription activator-like effector nucleases and zinc finger nucleases) have significantly broadened their clinical applications, which now include transgene insertion to enhance therapeutic potential and precision. The CRISPR/Cas9 system stands out because of its unparalleled efficiency and versatility, enabling precise single-gene modifications and complex multiplexed editing [65, 66]. It has become a cornerstone technology in genome editing and regulation and is extensively used in genomic silencing, knock-in integration, transcriptional modulation, and other applications [67]. Various delivery methods have been developed for the transfer of DNA templates in CRISPR knock-in applications, with the use of viral vectors among the most prominent. AAV has proven to be one of the most efficient delivery systems to date [68, 69]. In terms of the CRISPR/Cas9-based approaches that utilize nonviral delivery systems, cytotoxicity remains a significant challenge. This toxicity is primarily attributed to the electroporation of double-stranded DNA and ribonucleoprotein aggregates, as well as the activation of the innate immune response via DNA sensor pathways. To address these limitations, several innovations have been introduced. For instance, incorporating truncated Cas9 target sequences into DNA templates in combination with ribonucleoproteins has been shown to improve nuclear translocation and enhance knock-in efficiency. Additionally, single-stranded DNA templates, although generally less efficient than double-stranded DNA templates, are a less toxic alternative, exhibiting up to 40% efficiency in CAR T cell production in clinical studies [70]. Further improvements have been achieved by combining these approaches with the use of transposon-based systems to enhance RNA-guided integration [71, 72]. It must also be noted that the identification and mitigation of the risks associated with genomic rearrangement and translocation, including collateral damage, are crucial and will be instrumental in driving the clinical translation of gene-editing technologies in immunotherapy. Nonetheless, unintended genomic rearrangements, off-target edits, and chromosomal translocations remain major safety concerns. Rigorous characterization of genome integrity post-editing is thus essential to ensure translational readiness of CRISPR-engineered CAR T cells.

Taken together, nonviral delivery systems are not merely technical alternatives to viral vectors but represent a rapidly evolving frontier in CAR T cell engineering. As safety, regulatory control, and cost become increasingly important for scalable clinical deployment, the integration of gene editing, transposon systems, and advanced biomaterials holds the promise of ushering in the next generation of customizable, efficient, and safe CAR T cell platforms.

Applications of CAR T cells

Hematologic malignancies

All CAR T cell products currently approved by the FDA are second-generation CAR T cells, as summarized in Table 2 [16]. The first anti-CD19 CAR T cell therapy, tisagenlecleucel (Kymriah™), was approved by the FDA in 2017 after demonstrating significant effectiveness against pediatric B-ALL in clinical trials [73]. Since then, four additional anti-CD19 CAR T cell therapies have been approved: axicabtagene ciloleucel (Yescarta™), brexucabtagene autoleucel (Tecartus™), lisocabtagene maraleucel (Breyanzi™), and obecabtagene autoleucel (Aucatzyl™). Two anti-B cell maturation antigen (BCMA) CAR T cell therapies, idecabtagene vicleucel (Abecma™) and ciltacabtagene autoleucel (Carvykti™), have been approved for use in patients with multiple myeloma [74, 75]. Carvykti™ was the first Nb-based CAR T cell therapy to be approved by the FDA (February 2022) [76]. Among these FDA-approved anti-CD19 and anti-BCMA CAR T cell therapies, two utilize gamma-retroviral vectors (Yescarta™ and Tecartus™), while the remaining five use LVs (Kymriah™, Breyanzi™, Abecma™, Carvykti™, and Aucatzyl™). Notably, all CAR T cell therapies currently approved by the FDA are limited to the treatment of specific hematologic malignancies. Given that most cancers are nonhematologic (solid) in nature [77], there are increasing efforts to develop CAR T cell therapies for solid tumors and even nonmalignant conditions.

Table 2.

FDA-Approved CAR-T Cell Therapies (as of 2025)

Product name	Generic name	Target antigen	Vector type	Approved indications	Manufacturer	FDA approval year
Kymriah™	Tisagenlecleucel	CD19	Lentiviral	Pediatric and young adult B-ALL (≤ 25 years); r/r DLBCL	Novartis	2017
Yescarta™	Axicabtagene Ciloleucel	CD19	γ-Retroviral	r/r large B-cell lymphoma; r/r follicular lymphoma	Kite pharma (gilead)	2017
Tecartus™	Brexucabtagene Autoleucel	CD19	γ-Retroviral	r/r mantle cell lymphoma; r/r adult B-ALL	Kite pharma (gilead)	2020
Breyanzi™	Lisocabtagene Maraleucel	CD19	Lentiviral	r/r large B-cell lymphoma; r/r CLL/SLL (2024 label expansion)	Bristol myers squibb	2021
Aucatzyl™	Obecabtagene Autoleucel	CD19	Lentiviral	r/r B-cell malignancies (e.g., DLBCL; recently approved in 2023–2024)	Nanjing bioheng /	2023–2024
Abecma™	Idecabtagene Vicleucel	BCMA	Lentiviral	r/r multiple myeloma (after ≥ 4 prior lines of therapy)	Bristol myers squibb / 2seventy bio	2021
Carvykti™	Ciltacabtagene Autoleucel	BCMA (dual nanobody)	Lentiviral	r/r multiple myeloma (after ≥ 4 prior lines of therapy); first nanobody-based CAR-T therapy	Janssen (J&J) / legend biotech	2022

Nonhematologic malignancies (solid tumors)

The efficacy of CAR T cells against solid tumors remains limited. Figure 3 highlights the contrasting challenges associated with targeting hematologic and nonhematologic malignancies and compares the clinical efficacies of CAR T cell therapies developed to treat these two types of conditions. The limited efficacy of CAR T cells against solid tumors is partly a result of the restricted availability of targetable antigens, as CARs typically recognize cell-surface antigens. Additional factors include immunosuppressive TME and T cell exhaustion. Furthermore, CAR T cell therapies that target solid tumors have been shown to result in AEs because of target antigens being expressed in both tumor and healthy tissues. Nevertheless, an increasing number of phase I/II clinical trials are investigating the effects of CAR T cells specific for novel solid tumor-associated antigens [78]. Many of these trials focus on treating glioma, pancreatic cancer, and lung cancer, with the treatment of breast cancer and prostate cancer also under investigation [16].

Fig. 3.

The difference of clinical applications and outcomes for CAR therapy in hematologic malignancies and solid tumors

Among the clinical trials registered on the ClinicalTrials.gov database (accessed May 16, 2025) that focus on CAR T cell therapies for the treatment of solid tumors, most of the trials focus on two types of tumor-associated antigens: protein antigens and carbohydrate antigens. The following paragraphs highlight selected clinical trials that exemplify these two antigen categories.

Proteins as targets

The protein antigens discussed here are representative of the range of protein antigens being examined as potential targets of CAR T cell therapy in solid tumors. Human epidermal growth factor receptor 2 (HER2, also known as ERBB2) is expressed on the surface of some cancer cells. It plays a key role in cancer cell growth and is overexpressed in several types of solid tumors. HER2-targeted CAR T cell therapies are currently being evaluated in six clinical trials. In trial NCT02442297, a combination of HER2-targeted antibodies and CAR T cells is being tested in patients with HER2-positive brain tumors [79]. In trial NCT02792114, the safety of HER2-targeted CAR T cells in patients with metastatic HER2-negative breast cancer is being assessed to determine appropriate dosing levels In trial NCT00902044, the safety and preliminary efficacy of HER2-CD28 CAR T cells administered after chemotherapy are being evaluated in patients with sarcoma, aiming to determine whether this therapeutic approach can enhance tumor-specific targeting and effectively eliminate cancer cells without causing unacceptable toxicity [80]. In trial NCT03696030, the direct delivery of HER2 CAR T cells into the brain is being investigated as a method for treating cancers that have spread to the brain or leptomeninges [81]. In trial NCT03500991, which involves children and young adults with HER2-positive brain tumors, a new approach in which CAR T cells are delivered locally and directly to the tumor site to increase effectiveness and reduce AEs is being evaluated [82]. In trial NCT04511871, the safety and tolerability of a new CAR T cell product (CCT303-406) are being tested in patients with advanced, treatment-resistant HER2-positive solid tumors [83]. In addition, the safety, optimal dose, and preliminary efficacy of a combination of intratumoral CAdVEC (an oncolytic adenovirus) and HER2-specific CAR T cells are being evaluated in patients with HER2-positive cancers in trial NCT03740256, with the aim of enhancing the antitumor immune response via a two-pronged approach [84].

Epidermal growth factor receptor (EGFR)-targeted CAR T cell therapies for solid tumors are being examined in six clinical trials: five phase I trials and one phase I/II trial. In trial NCT05341492, a single-arm, open, exploratory clinical study, the safety and efficacy of EGFR/B7H3 CAR T cells are being assessed in patients with EGFR/B7H3-positive advanced solid tumors, including lung cancer and triple-negative breast cancer [85]. In the phase I trial NCT03198052, the safety and early efficacy of multi-targeted third-generation CAR T cells are being evaluated in patients with cancers expressing specific tumor antigens, including glypican (GPC) 3, mesothelin (MSLN), claudin18.2, GUCY2C, B7H3, PSCA, PSMA, mucin-1 (MUC1), TGFβ, HER2, Lewis-Y, AXL, and EGFR [86]. In trial NCT05060796, a single-arm, open-label study, the intravenous infusion of anti-EGFR CAR T cells modified by CXC chemokine receptor type 5 is being tested in patients with advanced adult non-small cell lung cancer [87]. In trial NCT06186401, the safety, AEs, and optimal dose of E-SYNC CAR T cells following lymphodepleting chemotherapy are being assessed in patients with EGFRvIII-positive glioblastoma [88]. In the phase I trial NCT03710421, the safe dose and AEs of CS1 CAR T cells after chemotherapy are being evaluated in patients with relapsed/refractory CS1‑positive multiple myeloma [89]. The aim of trial NCT06682793 is to evaluate A2B395, an allogeneic logic-gated Tmod™ CAR T cell product, in patients with EGFR-positive, HLA-A*02-negative solid tumors. In phase 1, the safe dose will be determined, and the tumor-killing efficacy and selectivity will be determined in phase 2. Participants will undergo lymphodepleting chemotherapy before receiving A2B395 [90].

MSLN is the most frequently targeted antigen in solid tumors. It has been assessed in 51 clinical trials to date, including three completed phase I trials. The safety and feasibility of MSLN-targeted CAR T cell therapies across various cancer types have been analyzed. For instance, in the phase I trial NCT02159716, CAR T cells modified to target MSLN were tested in patients with MSLN-expressing cancers [91]. Similarly, in trial NCT01897415, autologous MSLN-directed CAR T cells were assessed in patients with chemotherapy-refractory metastatic pancreatic cancer [92]. In another phase I trial, NCT03054298, both intravenous and local delivery of lentivirally transduced huCART-meso cells, with or without prior lymphodepleting chemotherapy, were tested to establish safety and feasibility [93].

Glycoforms as targets

Cancer cells frequently undergo aberrant glycosylation as a result of altered expression or activity of glycosyltransferases, leading to the production of tumor-associated glycoforms that differ markedly from those found in normal tissues [94]. These abnormal glycosylation patterns are closely linked to cancer initiation, progression, and metastasis, and can generate distinct neoantigenic epitopes that are attractive targets for immunotherapy [95]. Among the most well-characterized examples are the Tn (GalNAcα1-O-Ser/Thr) and sialyl-Tn (STn; NeuAcα2-6GalNAcα1-O-Ser/Thr) antigens, which are rarely expressed in healthy adult tissues but are commonly found on the surface of epithelial tumors [96]. These tumor-specific glycoforms are not only useful as diagnostic biomarkers but also play a pivotal role in modulating immunogenicity. Glycosylation can significantly influence antigen recognition and CAR-T cell functionality. For example, it has been shown that the glycosylation status of CD19 can impact the efficacy of CAR-T cells, highlighting the need to consider glycoform variation even in canonical targets [97]. This has spurred the development of glycoform-specific CAR-T cell therapies aimed at improving tumor selectivity and minimizing off-tumor toxicity. One notable advancement in this area is the development of CARs that recognize the Tn glycoform of MUC1, a mucin protein aberrantly glycosylated in many carcinomas. The 5E5 CAR, specific for a Tn-MUC1 glycopeptide epitope, has demonstrated potent antitumor activity in preclinical models of leukemia and pancreatic cancer, successfully eliminating Tn-MUC1-positive tumor cells without affecting normal tissues [98]. Targeting such glycoform-specific epitopes not only enhances therapeutic precision but also reduces the likelihood of off-tumor effects [99]. The identification and validation of suitable tumor-specific glycoforms are thus critical to expanding the therapeutic window of CAR-T cells. The Tn-MUC1 epitope, for example, is overexpressed across various tumor types and represents a compelling target for precision immunotherapy [98]. By engineering CARs that discriminate between tumor-associated and normal glycosylation patterns, it may be possible to develop safer and more effective CAR-T cell therapies [98, 100].

Carbohydrates as targets

A distinctive feature of malignant cells is the abnormal expression of cell-surface carbohydrates, a significant trait that often receives less attention than other well-established hallmarks of cancer [101]. These aberrant carbohydrates, which differ from those in healthy cells in type, structure, or abundance, are collectively referred to as tumor-associated carbohydrate antigens [102].

GPCs are heparan sulfate proteoglycans found in the extracellular matrix and on the surface of many cell types. GPC3 is expressed in a variety of solid tumors, including hepatocellular carcinoma (HCC) [103], urothelial carcinoma, thyroid cancer, ovarian clear cell carcinoma, melanoma, lung squamous cell carcinoma, and salivary gland tumors [104]. Among the 36 clinical trials involving GPC3-targeted CAR T cells, 5 have been completed. In trials NCT03884751 [105] and NCT03980288 [106], both phase I clinical studies, GPC3-targeted CAR T cells were tested in patients with advanced HCC. The latter trial focused on a fourth-generation CAR T cell construct, with the aim of improving antitumor activity and persistence. Similarly, trial NCT03146234 was designed to assess the safety and therapeutic potential of single and multiple doses of GPC3-targeted CAR T cells in relapsed or refractory HCC patients for whom there is a lack of effective treatment options [107]. In trial NCT02395250, autologous T cells containing a GPC3-specific CAR were evaluated, with a focus on their safety and efficacy in patients with relapsed or refractory HCC [108]. Finally, trial NCT02905188 was designed to assess GLYCAR T cells: GPC3-targeted CAR T cells containing a CAR derived from the GC33 antibody. This trial involved patients with advanced HCC that was recurrent, metastatic, or resistant to standard therapies, and its objectives included identifying the maximum tolerated dose, tracking the persistence of GLYCAR T cells in the body, monitoring AEs, and evaluating antitumor responses [109].

MUC1 is a glycoprotein overexpressed in various adenocarcinomas, such as lung, liver, colon, breast, pancreatic, and epithelial ovarian cancers [110]. CAR T cells targeting MUC1 are currently being assessed in 13 trials, and 1 trial has been completed. In trial NCT05812326, the safety and preliminary efficacy of an immunotherapy using PD-1-knockout anti-MUC1 CAR T cells were assessed in patients with advanced MUC1-positive breast cancer [111].

Disialoganglioside GD2 is highly expressed in neuroblastoma and other tumors of neuroectodermal origin [112]. Currently, there are 31 phase I and II clinical trials in progress investigating GD2-targeted CAR T cells, with 3 trials already completed. In trial NCT02765243, the safety, AEs, and appropriate dosing of a fourth-generation GD2-targeted CAR T cell product were determined [113]. In trial NCT02761915, a first-in-human study that assessed the safety and feasibility of 1RG-CAR T cell therapy, patient-derived T cells were modified to express a GD2-specific CAR and then reinfused to combat tumor cells [114]. Meanwhile, trial NCT02107963 evaluated the production feasibility and safety of anti-GD2 CAR T cells in children and young adults with GD2-positive solid tumors by utilizing escalating doses following lymphodepletion with cyclophosphamide [115]. Collectively, these studies highlight the significant progress being made in applying CAR T cell therapy to improve outcomes for patients with GD2-expressing neuroblastoma.

Nonmalignant conditions

CAR T cells are also being investigated as therapeutic agents for various nonmalignant conditions, including infectious, inflammatory, fibrotic, and autoimmune diseases. For example, CAR T cell therapies developed to treat human immunodeficiency virus infection primarily target gp120, an envelope protein, and have shown promise in suppressing viral replication and targeting reservoirs of latent virions [116–118]. CAR T cells have also been engineered to treat hepatitis B virus [119, 120], chronic hepatitis C virus [121], human cytomegalovirus [122, 123], and influenza A [124] infections, as well as Epstein–Barr virus-associated malignancies [125]. They demonstrated targeted cytotoxicity or immunomodulatory effects in preclinical models. In terms of fungal infections, CAR T cells directed toward Aspergillus fumigatus were found to reduce fungal burden [126]; however, specificity remains a challenge [127].

The effects of CAR T cells on allergies and asthma have also been analyzed. Immunoglobulin (Ig)E-targeted CAR T cells and carcinoembryonic antigen (CEA)-directed CAR regulatory T cells (Tregs) have been shown to modulate allergic inflammation and airway hyperresponsiveness [128, 129]. Furthermore, CAR T cells targeting fibroblast activation protein have been investigated for treating cardiac fibrosis [130, 131] and found to reduce fibrosis in mouse models [132]. Further refinement is needed to mitigate adverse inflammatory effects.

Using CAR T cells to treat autoimmune diseases is a rapidly evolving and promising area. The clinical translation of research findings now spans a wide array of autoimmune conditions, including systemic lupus erythematosus (SLE), multiple sclerosis (MS), lupus nephritis, systemic sclerosis, Sjögren’s syndrome, antisynthetase syndrome, myasthenia gravis, and idiopathic inflammatory myopathies. While the precise pathogenesis of many autoimmune diseases remains unclear, it is well established that the failure of T cell tolerance plays a central role, and that this leads to the activation of autoreactive B cells and the suppression of Tregs. Hence, most approaches focus on depleting autoreactive B cells. Chimeric autoantibody receptors, specialized CAR variants, have been designed to target autoreactive B cells by binding to their autoantigen receptors, thereby facilitating precise cytotoxicity while avoiding broad immunosuppression. This approach has been adopted to develop a treatment for pemphigus vulgaris, and after selective elimination of pathogenic B cells in vitro and in vivo was observed in preclinical studies, clinical trials are now in progress to evaluate the safety and efficacy of this approach [133]. Another approach involves converting CAR T cells into CAR Tregs through the introduction of FOXP3, a transcription factor crucial for Treg development and function, to restore immune tolerance. CAR Tregs suppresses autoimmune responses via mechanisms such as inducing anergy and clonal deletion, and they have demonstrated efficacy in preclinical models of diseases such as type 1 diabetes (T1D), MS, and ulcerative colitis. In T1D models, CAR T cells targeting insulin-reactive immune cells delayed disease onset [134], whereas CAR Tregs targeting myelin oligodendrocyte glycoprotein effectively reduced central nervous system (CNS) inflammation in MS models [135]. CEA-targeted CAR Tregs were found to alleviate inflammation and improve survival in a mouse ulcerative colitis model, highlighting their therapeutic potential in human inflammatory bowel diseases [135]. Several lines of research focus on depleting autoreactive B cells via CD19 or BCMA targeting. For example, BCMA-specific CAR T cells are being analyzed for their efficacy against myasthenia gravis [136], a condition characterized by muscle weakness. CD19-specific CAR T cells are being evaluated for their effects on SLE [137], an autoimmune disorder that causes systemic inflammation, and CAR T cells specific for both CD19 and CD20 are being assessed for their efficacy against neuromyelitis optica spectrum disorder [138]. A notable trial is also investigating CD7-targeted CAR T cells for their capacity to eliminate autoreactive T cells in Crohn’s disease, ulcerative colitis, dermatomyositis, and Still’s disease [139, 140].

Although none of the therapies directed toward autoimmune diseases have yet gained FDA approval, the emerging clinical data are encouraging. Patients have experienced sustained remission and marked symptom improvement, which are commonly associated with the depletion of autoreactive B cells and autoantibodies. Importantly, these therapies have favorable safety profiles, with most AEs being low grade and no severe CAR T cell related toxicity reported to date [140]. Collectively, these developments underscore the potential of CAR T cells to transform the treatment landscape for autoimmune diseases. They also provide valuable insights into therapy optimization, patient response dynamics, and risk mitigation strategies, thereby informing future clinical and translational advances in this rapidly evolving field.

Limitations hindering the safety and effectiveness of CAR T cells

Despite the progress made in developing CAR T cell treatments for solid tumors, several challenges and obstacles remain. The efficacy and safety of CAR T cells as a treatment modality for solid tumors are limited by both intrinsic and tumor-driven mechanisms, as well as treatment-related toxicity. The major factors limiting their efficacy against solid tumors are outlined below.

The immunogenicity of the targeting domains and counter strategies

CAR T cell therapy, while revolutionary in treating hematologic malignancies, continues to face limitations stemming from the immunogenicity of its synthetic construct. This immunogenicity can provoke host immune responses leading to the development of anti-CAR antibodies, which may neutralize CAR T cells, accelerate their clearance, and ultimately compromise therapeutic efficacy and persistence [141, 142]. Other structural elements of the CAR, including the hinge, transmembrane, and intracellular signaling domains, can also contribute to this immunogenic profile [143]. These immune-mediated effects are clinically significant, as reduced CAR T cell persistence is associated with diminished tumor clearance and increased risk of relapse [141]. In response, several strategies have been developed to mitigate immunogenicity. These include designing fully human or humanized CAR constructs to reduce xenogeneic responses [144, 145], optimizing hinge and transmembrane domains to minimize off-target immune interactions [143, 146] and combining CAR T cell therapy with immune checkpoint inhibitors like PD-1 blockade to counteract immune suppression [147, 148]. Additional approaches include dual-targeting CARs to prevent antigen escape [149], lymphodepleting chemotherapy to reduce host anti-CAR responses [150], and the incorporation of suicide genes for controllable CAR T cell elimination in the event of adverse events [151, 152]. The choice of T cell subsets, such as memory T cells, also plays a role in enhancing persistence and reducing immunogenicity [142, 153]. Moreover, novel constructs based on nanobodies are being explored for their lower immunogenic potential compared to scFv-based CARs [149]. Ongoing research continues to focus on identifying more specific tumor antigens [154], refining CAR architecture, and integrating combination strategies, such as pairing CAR T cells with oncolytic viruses or additional immunotherapies, to further enhance clinical outcomes [141, 155]. Notably, beyond immunogenicity, another intrinsic limitation of CAR design lies in the phenomenon of tonic signaling, defined as antigen-independent CAR activation, which represents a distinct but equally critical barrier to therapeutic efficacy and persistence.

Tonic signaling in CAR T cell therapy refers to the spontaneous, antigen-independent activation of CAR T cells, which can result in premature exhaustion and reduced therapeutic efficacy, particularly in the context of solid tumors [156]. This phenomenon occurs in the absence of the specific antigen targeted by the CAR T cells and is often attributed to structural characteristics of the chimeric antigen receptor itself. A primary driver of tonic signaling is the instability of the single-chain variable fragment (scFv), which can induce unintended receptor activation through self-aggregation [157]. Tonic signaling undermines CAR T cell persistence and functional longevity, thereby compromising their capacity for sustained immune surveillance and long-term remission [158]. To address this challenge, ongoing research focuses on refining CAR constructs to suppress tonic signaling while preserving cytotoxic function. One major strategy involves redesigning the targeting domain. For instance, D domains—engineered receptor-ligand binding regions—can replace traditional scFvs, offering improved stability and controlled binding kinetics that reduce off-target or constitutive activation [159]. When using scFvs, optimizing their physicochemical properties is critical. Parameters such as solubility, polarity, surface charge, and electrostatic patch size on the complementary determining regions (CDRs) significantly influence the likelihood of tonic signaling [156, 159]. Affinity tuning is also essential: while low-affinity CARs may show inadequate target recognition, overly high-affinity CARs are more prone to tonic signaling and loss of specificity [160]. In addition to the targeting domain, the hinge or spacer region plays a crucial role. Modulating its length and biochemical composition has been shown to influence CAR signal strength and reduce unwanted activation [161]. Another approach involves regulating CAR expression levels using weaker or inducible promoters to prevent overexpression and receptor clustering, though this must be balanced against the need for sufficient CAR density to ensure therapeutic efficacy [162]. Furthermore, precise genomic engineering strategies such as site-specific CAR knock-in at the T cell receptor alpha constant (TRAC) locus have demonstrated improved receptor regulation and signaling homeostasis, thereby limiting tonic activation [163]

Altogether, rational CAR design, including the careful selection and engineering of targeting domains, signaling components, and expression systems, is essential for minimizing tonic signaling. These innovations are pivotal not only for enhancing the durability and safety of CAR T cell-therapy but also for extending its application beyond hematologic malignancies to solid tumors and other challenging disease contexts [164].

The human microbiome

The human microbiome is the complex ecosystem of microorganisms found in the human body, including in the gastrointestinal tract and on the skin and mucosal surfaces. This microbial community exists in symbiosis with the host and plays a vital role in the maintenance of homeostasis and the regulation of immune functions. Dysbiosis, an imbalance in the composition or function of the human microbiome, can disrupt immune responses and contribute to the development and progression of various diseases, including cardiovascular conditions, respiratory illnesses, autoimmune disorders, and cancers.

More than 70% of lymphocytes transit through the gut, where they are modulated by the gut microbiota [165]. This modulation can influence immune responses and, thus, lead to enhanced cancer surveillance or the development of autoimmune diseases [166–168]. Recently, gut microbiota have been recognized as a critical factor that influences both the efficacy and toxicity of CAR T cell therapies. Notably, a highly diverse baseline gut microbial population is associated with improved CAR T cell treatment outcomes [169], and pretreatment antibiotic exposure has been linked to poor outcomes, with specific bacterial taxa (e.g., Bacteroides and Ruminococcus) identified as key predictors of the treatment response [170]. In addition, studies have indicated that modulating gut microbiota could enhance the efficacy of CAR T cells [171, 172]. These findings collectively highlight the potential of the gut microbiota as a noninvasive prognostic marker for the efficacy of CAR T cell therapy, although further research is necessary to fully elucidate its role in the efficacy of CAR T cell therapy.

Cellular factors

The selection of the target antigen(s) during the initial stages of CAR T cell therapy can be particularly challenging. Several studies have shown that many patients experience disease relapses because of antigen escape, which occurs when cancer cells downregulate or stop expressing the target antigen [173–175]. Tumor heterogeneity is another critical factor that contributes to variations in treatment efficacy, resistance, and failure [176]. Given that the antigens expressed in tumor cells can change over time and that potential target antigens are frequently expressed in normal tissues at varying levels, it is vital to develop innovative antigen selection strategies that will prevent antigen escape, enhance antitumor efficacy, and limit on-target off-tumor toxicity.

The efficacy of CAR T cells as a treatment option for solid tumors relies on their capacity to reach and infiltrate solid tumors. Chemokines, chemokine receptors, adhesion molecules, vasculature leakage, and other factors all influence T cell infiltration [177, 178].

In addition, infiltrating cells of various types secrete factors that contribute to the creation of an immunosuppressive TME, which allows tumor cells to escape antitumor immune processes. The immunosuppressive nature of the TME hinders the effector function of CAR T cells and thus impedes their clinical efficacy [179–181].

The success of CAR T cell therapy is further hindered by the development of CAR T cell exhaustion. Persistent antigen stimulation within the TME results in increased expression of inhibitory receptors and the presence of suppressive immune cells and cytokines [182], which significantly limit the effectiveness of CAR T cells by impairing their function. Notably, ex vivo manufacturing processes have been shown to induce CAR T cell exhaustion prior to infusion. In a recent study on the generation of CAR T cells against multiple myeloma, the initial PBMCs exhibited low expression of exhaustion markers; however, TIM-3 and LAG-3 were upregulated in the final product. The majority of the CAR T cells acquired an effector memory phenotype that, despite its high cytotoxic potential, is more susceptible to exhaustion. Additionally, CAR T cells with CD28 co-stimulatory domains were found to promote strong early cytolytic responses and exhibit more rapid exhaustion than those with 4-1BB domains. These findings indicate that exhaustion is influenced not only by the TME but also by manufacturing conditions, underscoring the need to optimize CAR constructs and cell culture protocols [153].

The efficacy of CAR T cell therapies is also influenced by the composition and functional quality of the CAR T cell product. Notably, the distribution of T cell subsets in the initial apheresis material plays a critical role in determining the fitness and therapeutic effectiveness of the manufactured CAR T cells. In addition, the relative frequencies of CD4⁺ and CD8⁺ T cells within CAR T cell products have been shown to significantly impact their immunological behavior and clinical performance. CD4⁺ CAR T cells are characterized by the enhanced production of Th1-type cytokines, such as interferon-gamma (IFN-γ), tumor necrosis factor-alpha (TNF-α), and IL-2, and robust proliferative responses upon antigen engagement. In contrast, CD8⁺ CAR T cells demonstrate superior cytolytic activity against target cells. Preclinical studies have shown that a 1:1 ratio of CD4⁺ to CD8⁺ CAR T cells can elicit synergistic antitumor effects in vivo [183]. Furthermore, clinical data indicate that patients receiving CAR T cell products with a lower ratio of CD4⁺ to CD8⁺ CAR T cells exhibit improved response rates at three months post-infusion and prolonged progression-free survival. A CD4⁺:CD8⁺ CAR T cell ratio of approximately 1.12 has been proposed as a threshold predictive biomarker for clinical outcomes, indicating a slightly higher proportion of CD4⁺ cells relative to CD8⁺ cells. Interestingly, patients with lower ratios of CD4⁺ to CD8⁺ CAR T cells also had a higher incidence of neurotoxicity, suggesting a complex interplay between T cell subset composition and toxicity profiles [184].

Acute AEs

CAR T cell therapy demonstrates significant efficacy against certain refractory cancers but may cause severe or life-threatening AEs that vary individually. The two prevalent acute AEs are CRS and neurologic events, also termed CAR T cell–related encephalopathy syndrome (CRES). These AEs are generally reversible and treatable with timely interventions; however, inadequate management can lead to severe or life-threatening outcomes [185, 186]. CRS is a form of systemic inflammatory response syndrome that arises because of the proliferation of activated T cells following CAR T cell-infusion, and it occurs in about half of all patients who receive CAR T cells [187], usually within a week of cell administration [185]. When CAR T cells engage with their target antigen, they are activated and secrete large amounts of pro-inflammatory cytokines, notably IL-6, IL-1, IFN-γ, and TNF-α [186]. These cytokines further stimulate macrophages (Ms) and endothelial cells, resulting in a self-perpetuating cycle of inflammation. In CRS, they also affect multiple organ systems. A hallmark of CRS is increased vascular permeability, and “third spacing” of fluid has been observed following the “cytokine storm,” leading to vasodilation, decreased cardiac output, and intravascular volume depletion [188]. CRS symptoms range from mild fever to life-threatening multi-organ failure. Acute adverse events, including CRS and CRES, were graded according to the American Society for Transplantation and Cellular Therapy (ASTCT) consensus criteria, which is widely adopted as the standard for CAR-T cell related toxicities. This system provides a clinically relevant and standardized framework, emphasizing objective clinical parameters such as fever, hypotension, and hypoxia for CRS, and neurologic assessment for CRES/ICANS. Unlike older grading scales, the ASTCT criteria are specifically tailored for immune effector cell therapies and offer more practical guidance for clinical intervention [189]. The grading system and recommended clinical management strategies, which emphasize early intervention with immunomodulatory therapies, are summarized in Fig. 4. Tocilizumab, an FDA-approved IL-6 receptor antagonist, is the first-line treatment and is often combined with corticosteroids. These agents are most effective when administered early in the course of CRS [190] and may also be used prophylactically in high-risk patients [191]. Multiple studies have shown that neither tocilizumab nor corticosteroids significantly impair the therapeutic function of CAR T cells [189].

Fig. 4.

Grading criteria and clinical management strategies for acute adverse events following CAR-T cell therapy adapted from ASTCT consensus guidelines

CRES is the second most frequently observed AE following CAR T cell therapy and may occur concurrently with or subsequent to CRS. It is hypothesized to result from cytokine-mediated disruption of the blood brain barrier and endothelial activation, which facilitates the infiltration of leukocytes and cytokines into the CNS. Clinically, CRES typically presents toxic encephalopathy. Early manifestations often include diminished attention, language disturbances, and impaired handwriting. As the condition progresses, patients may exhibit confusion, disorientation, agitation, aphasia, somnolence, and tremors [185].

Long-term AEs

Secondary malignancies have been reported in patients who received CAR T cells. In November 2023, the FDA announced that T cell malignancies, including CAR-positive lymphomas, had been reported in patients who received BCMA- or CD19-targeted autologous CAR T cells [192]. Subsequently, several studies have explored the risk of secondary malignancies in patients treated with CAR T cell therapies [193–195]. The FDA now requires a boxed warning for T cell malignancies following treatment with BCMA- or CD19-targeted autologous CAR T cell immunotherapies [196].

There is some evidence of a link between such secondary malignancies and insertional mutagenesis. This implies that the viral vector used to integrate the CAR transgene into the host T cells may inadvertently activate oncogenes or disrupt tumor suppressor genes and thus induce malignant transformation [60, 193, 197]. Although the incidence of secondary malignancies is low, the risk warrants attention, especially as CAR T cell therapies expand into earlier lines of treatment and broader patient populations.

Additionally, there is emerging evidence of an association between the development of secondary lymphomas and nonviral gene delivery methods. For example, in a clinical trial that utilized PB to engineer CAR T cells for the treatment of relapsed/refractory B cell malignancies, 2 out of 10 patients developed lymphoma [60]. This raised significant safety concerns about the use of PB and other transposon/transposase-based gene-modification platforms. No insertion of the transgene into known oncogenes was detected, and the precise mechanism underlying the malignant transformation remains undetermined. It has been proposed that the occurrence of lymphoma might have been associated with the manufacturing process or an increase in the global copy number variations noted in the therapeutic products [198]. These findings have highlighted the lack of standardized methods for evaluating genotoxicity when nonviral gene delivery systems are used. Currently, techniques for detecting small insertions and deletions, potential off-target DNA breaks, and endonuclease-mediated translocations are under development [199].

To address the risk of secondary T cell malignancies occurring after CAR T cell therapy, as well as other long-term AEs, comprehensive long-term clinical management and surveillance protocols have been implemented. Current guidelines, particularly those recommended by the FDA for gene-modified cellular products, recommend implementing extended follow-up periods of up to 15 years in both clinical trials and post-marketing surveillance programs. Molecular monitoring (e.g., insertion site analyses and clonality assays) is used to detect early-stage neoplastic changes.

The clinical management of patients with confirmed secondary T cell malignancies typically involves standard anti-cancer treatments, such as chemotherapy or radiotherapy, with investigational therapies considered in refractory cases [200]. The re-administration of autologous CAR T cells is contraindicated because of the risk of reintroducing malignant clones.

Patient registries and real-world data are crucial for identifying risk factors, informing clinical decision-making, and improving the safety of CAR T cell therapy. The risk benefit profile of CAR T cell therapy has been evaluated by regulatory agencies such as the FDA and the European Medicines Agency, as well as through multiple clinical trials and post-marketing surveillance studies. These evaluations consistently indicate that the therapeutic benefits—particularly in patients with relapsed or refractory hematologic malignancies—outweigh the associated risks, including cytokine release syndrome and neurotoxicity. Nevertheless, clinical caution remains essential, and continued investigation is needed to optimize patient outcomes and minimize adverse events.

Other CAR-based cellular immunotherapies

The expression of CARs in cells other than conventional polyclonal αβ T cells has been explored with the dual aims of generating more effective antitumor agents and broadening the application of CAR-based therapies. The range of cell types engineered to express CARs is illustrated in Fig. 5. Collectively, these alternative CAR-based cellular therapies provide novel platforms to overcome the limitations associated with CAR T cell therapy, particularly in the context of solid tumors, safety concerns (e.g., GvHD and CRS), and manufacturing challenges. Recent years have seen the launch of numerous clinical trials evaluating these approaches, many of which have yielded encouraging early-phase results. While most are still in early clinical stages (Phase I/II), their feasibility, safety, and initial efficacy data support the clinical potential of non αβ CAR-engineered immune cells as next-generation immunotherapies.

Fig. 5.

CAR-based cellular

CAR natural killer cells

Natural killer (NK) cells are a subset of innate lymphoid cells that play a critical role in the body’s defense against tumors and viral infections [201]. Because of their intrinsic antitumor properties (i.e., they do not need to undergo prior sensitization), they are promising candidates for CAR engineering [202]. Another major advantage of NK cells is that their activity is not restricted by HLA recognition, which not only contributes to their potent antitumor activity but also eliminates the risk of graft-versus-host disease (GvHD) when used in allogeneic settings [203]. By comparison, CAR T cells are often required to undergo autologous harvesting or genetic modification to avoid GvHD. Furthermore, CAR NK cells can be manufactured from pre-existing cell lines or allogeneic sources because of the lack of MHC restriction, which can result in simpler and more affordable “off-the-shelf” treatments. In contrast, CAR T cell therapy requires a patient-specific approach that is labor-intensive, costly [204], and susceptible to delays because of prior chemotherapy impairing the quality of T cells [205].

CAR NK cells have been shown to effectively kill cancer cells via both CAR-dependent and CAR-independent pathways while exhibiting low toxicity, particularly in terms of CRS and neurotoxicity [206, 207]. Despite their potential, there are limitations hindering the use of CAR NK cells. First, NK cells are more resistant to viral transduction than T cells [208]. Second, NK cells have short half-lives and exhibit poor persistence post-infusion [209]. Hence, there is a potential need for repeated dosing. The findings of preliminary trials suggest that it is safe to administer multiple CAR NK-cell infusions [210]; however, data are limited, and such regimens could affect the cost-effectiveness. The results of over 50 clinical trials conducted in the past five years with CAR NK-cell therapies provide further evidence of the potential of these therapies for treating various cancers, including solid tumors. The field is characterized by cautious optimism, and further research is required to fully understand and optimize CAR NK cells as a critical component of future cancer immunotherapy strategies [211].

Clinical evidence supporting CAR NK cell therapy is rapidly expanding. A landmark Phase I/II trial (NCT03056339) utilizing CD19-targeted CAR NK cells derived from cord blood reported a 73% overall response rate in patients with relapsed or refractory CD19-positive lymphoid malignancies, notably without CRS, neurotoxicity, or GvHD [212]. Building on these promising results, ongoing clinical trials are investigating CAR NK cells targeting HER2 (NCT03692637) [213] and NKG2D ligands (NCT05247957) [214] in solid tumors, highlighting the broad therapeutic potential of this approach across hematologic and solid malignancies.

CAR natural killer T cells

Natural killer T (NKT) cells are a small subset of αβ T cells. These cells are restricted by CD1d recognition and characterized by their invariant TCR [215]. NKT cells are particularly attractive as CAR carriers because of their capacity to infiltrate primary tumors, a feature correlated with improved outcomes across various cancer types. Like CAR NK cells, CAR NKT cells are being explored as off-the-shelf allogeneic therapeutic agents, addressing the need for readily available treatment options [216]. Additionally, their high specificity and low toxicity in both autologous and allogeneic settings are attributed to the monomorphic nature of the CD1d gene, which is expressed in limited cell types [217–220]. Notably, CAR NKT cells have exhibited stronger in vivo antitumor activity than CAR T cells, which is attributed to their CD1d-dependent elimination of immunosuppressive Ms, the promotion of spreading epitope, and the activation of endogenous T cell responses within the TME. Despite showing signs of exhaustion under a high tumor burden, CAR NKT cells can be functionally restored through PD-1 blockade and therapeutic vaccination [216, 221]. Although CAR NKT cells are still in the early stages of development, the inherent pro-inflammatory properties of NKT cells (e.g., robust cytokine production and rapid cytotoxic response) position these cells as a promising treatment option in adoptive cell therapy. A landmark first-in-human Phase I/II clinical trial (NCT03294954) [222] evaluated GD2-targeted CAR NKT cells in children with relapsed or refractory neuroblastoma. The study demonstrated a favorable safety profile and early signs of clinical benefit, with no cases of GvHD report [223]. Expanding beyond neuroblastoma, CD70-targeted CAR NKT cells are currently being investigated in a Phase I trial (NCT06182735) for the treatment of advanced malignant solid tumors, including renal cell carcinoma [224]. Ongoing research continues to explore the potential of CAR NKT cells across a broader spectrum of solid tumors.

CAR cytokine-induced killer cells

Cytokine-induced killer (CIK) cells are a subset of cytotoxic T cells that exhibit both T- and NK-cell phenotypes. These cells are capable of MHC-unrestricted tumor killing and are manufactured ex vivo [225]. In contrast to CAR T cells, which are generated from autologous T cells and act via MHC-restricted antigen recognition, CAR CIK cells are generated from a heterogeneous population that primarily consists of T cells with NK cell-like features. Hence, CAR CIK cells exert MHC-unrestricted cytotoxicity, making them highly suitable for off-the-shelf allogeneic applications with a low risk of GvHD [226]. Additionally, CIK cells produce high levels of IFN-γ, which has been shown to reduce GvHD in mouse models [227]. Comparative studies have demonstrated that CAR CIK cells possess superior cytotoxic activity against HER2-positive rhabdomyosarcoma (RMS) cells in vitro and ex vivo compared to conventional CAR T cells derived from the same donors [228]. Moreover, compared to CAR T cells, CAR CIK cells have a greater proliferative capacity, a more favorable cytokine secretion profile (characterized by elevated levels of IFN-γ, perforin, and granulysin), and an enhanced capacity to activate chemokine signaling and NK cell–mediated cytotoxic pathways. While both CAR T and CAR CIK cells were found to extend survival in RMS mouse models, CAR CIK cells exhibited superior persistence and a sustained effector function in vivo, reinforcing their potential as robust alternatives to CAR T cells, particularly for targeting solid tumors [228]. In another study, anti-CD19 CAR CIK cells engineered using the nonviral SB transposon system achieved high complete response rates in patients with B-ALL without inducing GvHD or neurotoxicity [229]. Furthermore, CD33-targeted CAR CIK cells have shown efficacy against AML in preclinical studies [230]. Similarly, CAR CIK cells engineered to target HER2 were effective in treating RMS [231], while CSPG4-specific CAR CIK cells demonstrated robust antitumor activity in soft tissue sarcoma models [232]. The distinct NK cell-like phenotype of CAR CIK cells, combined with their safety in allogeneic applications and MHC-unrestricted cytotoxicity, underscores their potential as cancer immunotherapy. Anti-CD19 CAR CIK cells engineered using the Sleeping Beauty (SB) transposon system were tested in a Phase I clinical trial (NCT03389035) in B-ALL, demonstrating high complete response rates with minimal toxicity, notably without neurotoxicity and GvHD [233]. Building on these results, a Phase I/II trial (NCT05869279) is currently underway to assess the safety and efficacy of Anti-CD19 CAR CIK cells in both adult and pediatric patients with relapsed/refractory B cell non-Hodgkin lymphoma [234].

CAR γδ T cells

Gamma delta (γδ) T cells, a subset of T cells characterized by variable chains in the TCR complex, exhibit features associated with both the innate and adaptive immune systems [235]. Unlike αβ T cells, which rely on MHC-dependent antigen presentation for target recognition, γδ T cells recognize a broad range of tumor-associated and stress-induced ligands through an MHC-independent mechanism. Given that γδ T cells can eliminate cancer cells in a manner that does not rely on MHC recognition [236], they can be used to generate allogeneic off-the-shelf CAR γδ T cells with a low risk of GvHD that can target tumor cells that evade immune surveillance by downregulating MHC molecules [237, 238]. Importantly, CAR γδ T cells possess a dual-recognition capability because of the presence of engineered CARs and their endogenous γδ TCRs, which enhance their specificity and minimize off-target effects [238].

Preclinical studies have demonstrated that CAR γδ T cells are effective against both hematologic malignancies, such as leukemia and lymphoma, and solid tumors, which are typically challenging to treat with conventional CAR T cells because of poor tumor infiltration and antigen heterogeneity [237, 238]. Their capacity to present antigens to αβ T cells further amplifies their antitumor response [238]. The adoptive transfer of CAR γδ T cells is a promising therapeutic strategy for cancers such as GD2-expressing neuroblastoma [239] and CD19-positive malignancies [240].

Nevertheless, the clinical translation of CAR γδ T cells faces challenges, including limited in vivo persistence, inefficient expansion protocols, and potential functional exhaustion upon chronic antigen exposure. To address these limitations, multiple strategies are being explored, including developing CAR γδ T cells with co-stimulatory domains, enhancing cell survival via gene editing, using cytokines and zoledronic acid for ex vivo expansion, and using different delivery systems (e.g., garlic-derived nanoparticles or γδ T cell engagers) to improve therapeutic sustainability and functionality [238]. The results of ongoing clinical trials in which CAR γδ T cell therapies are being evaluated as treatments for various cancers, including relapsed or refractory B cell malignancies and solid tumors, indicating that CAR γδ T cell therapy has potential as a next-generation cancer immunotherapy [237, 238].

CAR γδ T cells are under evaluation in Phase I trials, such as NCT04735471, for treating relapsed/refractory B-cell malignancies and neuroblastoma. Preliminary findings suggest that these therapies are well tolerated, with early signs of antitumor activity and a lower incidence of immune-related adverse events [241]. Additionally, a Phase 1/2 trial (NCT06480565) is evaluating ADI-270, a CAR γδ T cell therapy in patients with relapsed/refractory clear cell renal cell carcinoma who have previously received immune checkpoint inhibitors and VEGF-targeted therapies [242]. This study aims to assess safety, cellular kinetics, immunogenicity, and antitumor efficacy.

CAR Tregs

Tregs are a subset of αβ T cells that play a central role in maintaining immune homeostasis and preventing autoimmunity [243]. The lack of detailed understanding about the antigen recognition mechanisms employed by Tregs, combined with the nonspecific antigen recognition exhibited by polyclonal Tregs, limits their therapeutic potency. CAR technology provides an MHC-independent approach for directing Tregs toward pathogenic cells or affected tissues [244]. CAR Tregs are designed to enhance the suppressive function and stability of Tregs and prevent CAR Treg exhaustion. Hence, CAR Tregs and CAR T cells represent two distinct therapeutic strategies: CAR T cells are designed to activate immune responses to target cancer, while CAR Tregs are designed to suppress immune activity to treat autoimmune, inflammatory, and fibrotic diseases and to promote tolerance in transplantation and gene therapy [244, 245]. CAR Tregs are MHC-independent, are less dependent on IL-2 than CAR T cells, and exhibit bystander suppression and infectious tolerance, allowing for broader application and sustainable immune regulation [244, 245].

CAR Tregs have been reported to suppress B cell pathology without inducing GvHD [246], to be associated with a lower incidence of GvHD [247], and to be effective against hemophilia A [248], vitiligo [249], autoimmune encephalomyelitis [135], colitis [250, 251], asthma [252, 253], cardiovascular diseases [130], and senescence-associated pathologies [254]. There are several obstacles limiting the use of CAR Tregs, such as potential CRS, development delays due to antigen specificity, and high costs [245, 255]. Engineering solutions, such as the introduction of co-stimulatory domains, IL-2Rβ signaling, and telomerase expression, are being explored to enhance CAR Treg stability and persistence [245]. Off-the-shelf allogeneic CAR Tregs and universal CAR systems are being tested to increase accessibility [245, 255]. Data from preclinical and clinical studies support the use of CAR Tregs; however, further optimization is needed for broader clinical adoption [244].

CAR Tregs are currently undergoing early-phase trials (NCT04817774) aimed at addressing immune-mediated graft rejection and promoting immunologic tolerance post renal transplantation [256], as well as in NCT05234190 for preventing liver transplant rejection after withdrawal of immunosuppression [257]. These trials, although in early stages, demonstrate the potential safety of CAR Tregs infusion. Initial biomarker data also suggest immune modulation and successful engraftment in targeted tissues.

CAR macrophages

Macrophages (Ms) are a key component of the innate immune system. They play essential roles in immune defense, tissue homeostasis, and the regulation of various pathological processes [258]. The development of CAR Ms is appealing because Ms can infiltrate solid tumors, modulate the TME, and sustain prolonged antitumor activity [259, 260]. CAR Ms have been shown to recognize tumor-associated antigens, initiate intracellular signaling cascades, induce tumor cell phagocytosis, secrete pro-inflammatory cytokines, upregulate MHC class I and II molecules, maintain the M1 phenotype, and stimulate adaptive immune responses [261, 262]. Collectively, these functions support a targeted and sustained antitumor response [263, 264], and CAR Ms clearly utilize a range of mechanisms to mediate their antitumor effect, unlike CAR T cells, which exert their effects mainly via direct cytotoxicity [265, 266].

In head-to-head in vitro studies, CAR Ms engineered with different intracellular signaling domains (FcRγ, Megf10, or PI3K) demonstrated varying phagocytic and killing capacities, with FcRγ-based CAR Ms showing superior performance [267]. Importantly, CAR Ms can synergize with CAR T cells, as inflammatory cytokines secreted by CAR T cells induce M1 polarization and upregulate CD80/CD86 on CAR Ms, thereby enhancing the overall cytotoxic effect [267]. Despite their tumor-killing efficacy, CAR Ms have displayed more limited tumor lysis in vitro compared to CAR T cells; however, their natural effector functions (e.g., cytokine secretion) can effectively modulate the immune landscape [266]. CAR Ms also shows enhanced tumor infiltration compared to CAR T cells [266, 268]. The efficacy of these cell types is limited by similar challenges, including antigen escape and the limited availability of tumor-specific antigens [268].

Compared to CAR T cells, which have been extensively studied at the preclinical and clinical levels, the use of CAR Ms is still in the early stages but gaining momentum because of promising preclinical and early clinical results [265, 268]. CAR Ms are being evaluated in ongoing clinical trials for their capacity to target HER2-, MSLN-, and GPC3-expressing tumors. Early data show that the use of CAR Ms is safe and feasible; however, further validation is required [265]. Other studies are assessing the capacity of CAR Ms to reduce tumor burden and their safety profile in patients with refractory or advanced-stage malignancies who have exhausted conventional therapeutic options, and preliminary results indicate that they effectively localize to tumor sites and exhibit sustained persistence [268, 269]. While CAR T cells exhibit superior tumor homing activity and are more amenable to systemic administration, CAR Ms have a better safety profile, with a reduced risk of CRS and potential for allogeneic off-the-shelf use [266]. Combining CAR Ms with CAR T cells and designing multifunctional CAR constructs are promising avenues for overcoming the limitations of the different cell types and improving therapeutic outcomes in the treatment of solid tumors [266, 267].

The first-in-human Phase I trial of HER2-targeted CAR macrophages (NCT04660929) in patients with advanced solid tumors demonstrated a favorable safety profile, with no CRS observed, and confirmed successful tumor infiltration [270]. Additionally, another study (NCT05007379) is evaluating the antitumor activity of novel CAR-macrophages using patient-derived breast cancer organoids [271]. Interesting, CAR-M therapy is being explored for hepatocellular carcinoma, targeting Glypican-3, a tumor-associated antigen highly expressed in liver cancer but absent in most normal adult tissues, making it an ideal target for HCC treatment [272].

CAR mucosal-associated invariant T cells

Mucosal-associated invariant T (MAIT) cells, an evolutionarily conserved, innate-like subset of T cells that bridge the innate and adaptive immune responses, play a crucial role in the defense against infectious diseases [273, 274]. They are abundant in the human body and enriched at mucosal sites, such as the lungs and gut [275–277]. MAIT cells reportedly promote inflammation and mediate antitumor responses in cases of human cancer [278]. It has also been reported that MAIT cells are not alloreactive because they are not activated by classical MHC/peptide complexes and express semi-invariant TCRs. Thus, MAIT cells have potential as a source cell type for the development of an off-the-shelf CAR-based therapy. CAR MAIT cells are a viable alternative to autologous CAR T cells, as they overcome the limitation of HLA disparity and can be produced in a faster and more cost-effective manner [276].

Compared to conventional CAR CD8⁺ T cells, CAR MAIT cells demonstrate robust cytotoxicity and lower cytokine secretion (e.g., IFN-γ, GzB, and GM-CSF secretion). This means that CAR MAIT cells may be safer and more effective than other CAR-based therapies because of a lower risk of CRS [279]. However, MAIT cells constitute only 1–10% of the T cell population in the peripheral circulation [274]. Consequently, large-scale production of CAR MAIT cells is currently limited by the availability of PBMC-derived MAIT cells, and this is hindering extensive research being conducted on CAR MAIT cells. To overcome this production barrier, it may be viable to utilize genetically engineered hematopoietic stem cells or induced pluripotent stem cells to generate CAR MAIT cells.

Although CAR MAIT cell clinical trials are not yet registered, preclinical models demonstrate robust efficacy against various cancer types. The potential for allogeneic application, coupled with reduced cytokine release, has generated enthusiasm for advancing CAR MAIT cells into forthcoming early-phase clinical trials [278, 280].

Conclusion

CAR-based immunotherapy has revolutionized the treatment of hematologic malignancies, particularly B cell-associated and relapsed/refractory cancers. It is resulting in transformative outcomes in patients who are unresponsive to conventional therapies. Building on this success, the field has rapidly advanced to now include fifth-generation CAR constructs and potential therapies based on a growing variety of cell types, such as CAR NK cells, CAR Tregs, and CAR Ms. These developments are broadening the therapeutic scope of CAR constructs beyond cancer treatment to the treatment of autoimmune, infectious, and fibrotic diseases. For instance, CAR-T cells targeting autoimmune diseases such as lupus and rheumatoid arthritis have demonstrated promising results in preclinical models by selectively eliminating pathogenic immune cells, leading to disease remission. In infectious diseases, CAR-T and CAR-NK cells engineered to target viral antigens like HIV and HBV have shown potential in reducing viral reservoirs and controlling infection in animal studies and early-phase clinical trials. Despite these achievements, significant challenges remain. Therapeutic efficacy against solid tumors is hindered by antigen heterogeneity, immune escape mechanisms, and immunosuppressive TME. Additionally, severe AEs, such as CRS, neurotoxicity, and secondary malignancies, currently limit wider application. These issues highlight the need for safer, more targeted CAR constructs and toxicity-mitigation strategies. Efforts to improve CAR T cell trafficking, persistence, and antigen specificity while maintaining safety are crucial to enhancing the therapeutic sustainability of this type of therapy. Addressing manufacturing complexity and developing safe and scalable nonviral gene delivery systems are also essential to achieving broader clinical accessibility. In this review, we have summarized the evolution and current status of CAR-based immunotherapies, emphasizing their remarkable impact on the treatment and outcomes of hematologic malignancies and their promise as therapeutic agents for nonhematologic malignancies and nonmalignant conditions. Continued innovation, rigorous clinical evaluation, and technological progress will be key to unlocking the full potential of CAR-based therapies across a diverse range of medical conditions.

Competing Interests

The authors declare no competing interests.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

AAV

Adeno-associated virus

AE

Adverse event

ALL

Acute lymphoblastic leukemia

αβ T cells

Alpha beta T cells

AML

Acute myeloid leukemia

B-ALL

B-cell acute lymphoblastic leukemia

BCMA

B-cell maturation antigen

CAR

Chimeric antigen receptor

CAR CIK

Chimeric antigen receptor-cytokine-induced killer cell

CAR Ms

Chimeric antigen receptor-macrophage

CAR MAIT

Chimeric antigen receptor-mucosal-associated invariant T cell

CAR NK

Chimeric antigen receptor-natural killer cell

CAR NKT

Chimeric antigen receptor-natural killer T cell

CAR T

Chimeric antigen receptor T cell

CAR Treg

Chimeric antigen receptor-regulatory T cell

CAR γδ T

Chimeric antigen receptor-gamma delta T cell

CD

Cluster of differentiation

CD1d

Cluster of differentiation 1d

CEA

Carcinoembryonic antigen

CIK

Cytokine-induced killer cell

CRISPR

Clustered regularly interspaced short palindromic repeats

CRES

CAR-T cell-related encephalopathy syndrome

CRS

Cytokine release syndrome

DC

Dendritic cell

dsDNA

Double-stranded DNA

EGFR

Epidermal growth factor receptor

Env

Envelope glycoprotein

FAP

Fibroblast activation protein

FDA

U.S. food and drug administration

FOXP3

Forkhead box P3

γδ T cells

Gamma delta T cells

gB

Glycoprotein B

GM-CSF

Granulocyte-macrophage colony-stimulating factor

GPC

Glypicans

gp120

Glycoprotein 120

GvHD

Graft-versus-host disease

GzB

Granzyme B

HBV

Hepatitis B virus

HBsAg

Hepatitis B surface antigen

HCC

Hepatocellular carcinoma

HCMV

Human cytomegalovirus

HCV

Hepatitis C virus

HER2

Human epidermal growth factor receptor 2

HIV

Human immunodeficiency virus

IFN-γ

Interferon gamma

IgE

Immunoglobulin E

IL

Interleukin

iPSC

Induced pluripotent stem cell

LMP1

Latent membrane protein 1

LV

Lentiviral vector

MAIT

Mucosal-associated invariant T cell

mbIL15

Membrane-bound interleukin-15

MHC

Major histocompatibility complex

mIgE

Membrane-bound immunoglobulin E

MOG

Myelin oligodendrocyte glycoprotein

MUC1

Mucin-1

MS

Multiple sclerosis

MSLN Ms

Mesothelin macrophages

Ms

Macrophages

Nbs

Nanobodies

NK

Natural killer cell

NK cells

Natural killer cells

NKT cells

Natural killer T cells

NMOSD

Neuromyelitis optica spectrum disorder

PB

PiggyBac

PBMC

Peripheral blood mononuclear cell

PreS

Pre-surface region of (HBV envelope protein)

SB

Sleeping beauty

scFv

Single-chain variable fragment

SLE

Systemic lupus erythematosus

T1D

Type 1 diabetes

TALEN

Transcription activator-like effector nuclease

TCR

T cell receptor

TME

Tumor microenvironment

TRUCK

T cell redirected for universal cytokine-mediated killing

Treg

Regulatory T cell

VHH

Variable heavy chain

ZFN

Zinc finger nuclease

Author Contribution

Natthaporn Sueangoen: Original draft, Writing, Visualization, reviewed Somsak Prasongtanakij: Writing, review & editing, Supervision, Conceptualization.

Funding

Open access funding provided by Mahidol University.

Data Availability

No datasets were generated or analysed during the current study.