Advances in CAR T cell therapy: antigen selection, modifications, and current trials for solid tumors

First-generation CARs carry a single CD3ζ-chain or FcϵRIγ intracellular domain (Figure 2) without any additional costimulatory domains. Though these complexes are like endogenous TCRs, they are unable to produce sufficient interleukin-2 (IL-2). As a result, first-generation CARs need to be supplemented with exogenous IL-2 for an efficient response. This proved to be a major drawback in the first-generation CAR Ts (11). First generation CARs have also demonstrated low cell proliferation and a short in vivo lifespan; both these points contributed to their poor performance in a clinical setting (12, 13). These drawbacks and poor clinical performance prompted the development of the second generation of CAR Ts.

To improve on the first generation of CARs, co-stimulatory domains such as CD28, 4-1BB, or OX-40 were added to the CAR structure (Figure 2). CARs carrying these cytoplasmic costimulatory domains proved much more effective in delivering a secondary signal upon encountering a tumor antigen. A comparison of T cells carrying only one signaling domain to those carrying an internal CD28 costimulatory domain showed a significant increase in proliferation and persistence (23). In general, addition of co-stimulatory domains improved proliferation and cytotoxicity, while also leading to a prolonged response and an increased in vivo life span. Each of these costimulatory domains imparted specific characteristics to the CAR T. 4-1BBζ-CAR T cells could persist longer in circulation than CD28ζ-CAR T. It was also observed that while the former caused early exhaustion of CAR T cells, the latter led to constitutive stimulation even in the absence of the antigen. Thus, subsequent CAR design focused on the engineering of more effective costimulatory constructs (23).

Third-generation CARs utilize a combination of multiple costimulatory signaling domains within the endodomain (Figure 3). Some of the common constructs carry CD3ζ-CD28-OX40 or CD3ζ-CD28-4-1BB (14, 24). Amongst these third generation CAR constructs, CD3ζ-CD28-4-1BB seems to be the most promising given the CD28 costimulatory domains contribute to rapid tumor elimination while the 4-1BB endodomains increase CAR survival (25, 26). Third generation CARs have been successful in the treatment of cancers as they demonstrate high proliferation, maintain longer survival periods, and display good safety profiles. However, their efficiency in tumor elimination does not always remain superior compared to second-generation CAR T cells (20). Hence, efforts to improve efficiency and safety resulted in the development of fourth generation CARs.

Subtle modifications of third generation CAR Ts gave rise to fourth-generation CARs. These CARs feature expression of transgenic proteins, often cytokines, that can be inducibly or constitutively expressed. In some cases, these CAR Ts are also designed with a regulatory switch for a suicide gene or an element that enhances T cell function (27, 28). In one instance, inclusion of Caspase 9 enhanced the safety of CAR Ts (29). An advanced version of this CAR was developed by Labanieh et al. (30). The group designed protease-based CARs that could be regulated with drug doses to enhance the safety and efficacy of the CAR Ts. Also belonging to the fourth generation of CAR Ts are T cells redirected for universal cytokine-mediated killing (TRUCKs). These CAR Ts are engineered to carry a nuclear factor of the activated T cell (NFAT)-responsive cassette (containing a transgenic cytokine such as IL-12) and deliver the transgenic product to the tumor site (Figure 2). The presence of a cytokine transgene enhances the efficacy of CAR T cell therapies in preclinical models with low systemic toxicity (31, 32). Thus, in TRUCKs, the expression of the transgene is induced when CD3ζ-containing CARs encounter their specific target. TRUCK CAR T cells thus carry two transgenic cassettes – one for the CAR structure and another for the inducible cytokine.

Fifth generation CARs integrate an additional membrane receptor into their structure. Of the several different approaches tested, one of the most promising is the addition of IL-2 receptors (Figure 2) that allow for antigen-dependent JAK/STAT pathway activation (33).

Another promising approach is the inclusion of switch receptors. Incorporation of a drug-dependent OFF-switch that leads to CAR depletion or the incorporation of an ON-switch that leads to activation remains a promising option. In fact, lenalidomide-gated CARs are designed on this principle (34). Even though these CARs exhibit less efficiency in vitro, they remain favored from a research perspective given their increased regulation and enhanced safety profile. Li et al. (2022) further developed this concept by developing a multiantigen targeted collection of CARs controllable with FDA-approved antiviral drugs (35). The VIPER (VersatIle ProtEase Regulatable) CARs are inducible with ON and OFF switch CAR circuits engineered using a viral protease domain. The group demonstrated limited toxicity of their system in a cytokine storm animal model. In addition, they also developed several complex CAR circuits with logic and multiplex control functions.

Dual-signaling CARs make use of two separate CARs on the same T cell with each CAR cell having its own intracellular domains (Figure 3A). Such CARs are much more effective since they can target different antigens at the same time. Dual targeting, split costimulatory signaling and shared CD3ζ chain tailored CARs have been used to target two clinically relevant antigens (15). These dual-signaling CARs were targeted against GD2 and B7-H3 in a neuroblastoma disease model (36–38). In fact, this approach was further validated with other pairs of targets, namely mesothelin (MSLN) and chondroitin sulfate proteoglycan 4 (CSPG4) (15).

Similarly, a dual-signaling CAR T targeting CD19 and CD123 was created and proved more efficient in tumor eradication in preclinical trials than using CAR Ts with a single signaling moiety (CD19-CAR T cells and CD123-CAR T cells) (18). Additionally, this dual-signaling CAR T prevented disease recurrence caused by downregulation or loss of target antigens on the tumor cell surface much more efficiently. Thus, it has been shown that the dual-signaling CAR strategy allows for rapid and sustained antitumor effects while preventing tumor escape in heterogeneous antigen expression tumor cell populations (15).

Refinement of antigen binding domains in the form of scFvs has also led to the production of CARs that target more than one antigen in the form of multiple antigen-binding CARs (Figure 3B). These CARs exist in the form of dual CARs, consistent of two separate binding domains, and in the form of TandemCARs, which contain a modified CAR ectodomain that encodes for multiple antigen-binding domains in a CAR module using scFvs from monoclonal antibodies (39–41).

To further enhance the efficiency of CARs, antigen binding domains have been designed to carry three scFvs in tandem. As a result, these CARs can target three different tumor associated antigens (TAAs), allowing for their utility in treating tumors with heterogeneous cell population (42). Constructed CARs with a single universal tricistronic transgene were designed to recognize three different TAAs, namely, human epidermal growth factor receptor 2 (HER2), interleukin-13 receptor subunit alpha-2 (IL13Rα2), and ephrin-A2 (EphA2) to treat glioblastoma. The outcomes of this CAR T therapy revealed that the trivalent-tandem CAR had strong anti-tumor activity and could overcome tumor antigen heterogeneity at a level greater than nonspecific or bispecific CAR T cells. On similar lines, Balakrishnan et al. (2019) designed a trivalent-tandem CAR by using three scFvs in tandem with ankyrin repeat proteins (DARPins) (43). These CAR T cells displayed potent anti-tumor effects and performed well despite extensive tumor antigen heterogeneity and immune escape.

Another modification of CAR involves a design concept in which the CAR T can only be activated when two scFvs bind to corresponding antigens simultaneously (Figure 3C). Boolean logic gates are the recent developments to CAR T cells as safety switches. Integrated signals from multiple receptors at once can regulate CAR T cells activity based on their environment. The “AND”- gate logic utilizes two receptors that recognize different tumor antigens. Binding to both antigens of interest is needed to trigger CAR T cell activation (44). This combinatorial antigen sensing circuit is dependent on the synNotch receptor that forms the core design of “AND”-gate CARs. The synNotch binds to the first selective tumor antigen, inducing the expression of the second CAR, which then binds to specific second tumor antigen leading to the activation of CAR T.

The utilization of two scFvs binding to corresponding antigens can also be used to limit on-target off-tumor toxicity. Inhibitory CARs (iCARs) contain a T cell inhibitory signaling domain. An AND-NOT gate CAR T strategy allows for the CAR on the T cell to recognize the tumor antigen and activate the T cell while the iCAR recognizes a normal tissue antigen and inhibits T cell activity (45). This allows the CAR T cell to limit its on-target off-tumor toxicity by distinguishing between a tumor cell and a normal cell that both have the same CAR target antigen.

While CARs exhibiting a TCR-like antibody form utilize scFvs from monoclonal antibodies, there exist other mechanisms for antigen binding (46). Another approach for CAR binding relies on variable heavy domains of heavy chains (VHHs), solely the variable domains of heavy-chain-only antibodies (HCAbs) (47). Given advancements in the complementarity-determining regions (CDRs) of VHHs, these VHHs, or nanobodies, can bind antigens with affinities comparable to those of antibodies (Figure 3D). These binding affinities are enhanced by further mutations in the VHH framework which allow for these molecules to remain stable in aqueous solutions while containing soluble protein domains (48). The applicability of these nanobody-driven CARs, or nanoCARs, is derived from investigations showing the drawbacks associated with scFv framework regions in CARs - such regions have shown to cause CAR aggregation and tonic signaling with low cell activation (49). Although nanoCARs are still subject to this mentioned aggregation issue, nanoCARs remain more predictable in their binding ability given their monomeric structure, allowing for the use of immune libraries (50).

T cell receptor fusion constructs (TRuCs) utilize the biology of TCRs and fuse them with an antibody-based binding domain (44). The fusion of the antibody-based binding domain to the TCR to create TRuCs allows these modified TCR complexes to effectively recognize surface antigens (Figure 3E). These modified TCR complexes, specifically referred to as TRuC-T cells, have shown to kill tumor cells with the same level of potency as second-generation CAR-T cells while displaying a significantly lower cytokine release (51). Additionally, TRuCs have been used in targeting treatment-refractory mesothelin expressing solid tumors through the fusion of an anti-mesothelin antibody to the endogenous TCR complex. This TRuC-T therapy, gavocabtagene autoleucel, has shown positive results in a Phase I/II clinical trial (NCT03907852↗) has shown promising results with a disease control rate of 77% in thirty patients (52).

After leukapheresis separates leukocytes from the blood of patients, enrichment of the product allows for collection of T cells (53). Rather than completing the intensive process of purification of autologous antigen-presenting cells (APCs) from patients to activate T cells, beads coated with anti-CD3 and anti-CD28 monoclonal bodies have been employed to activate T cells in an efficient way (54, 55). Traditionally, autologous T cells, cells purified from the patient to be treated, have been used to produce CARs (56). However, a considerable drawback associated with the usage of autologous T cells remains the retrieval of patient specific T cells for CAR development for effective CAR T therapy. In specific, drawbacks include the time-consumption and cost associated with autologous T cell production, the functional quality of autologous T cells, and the potential manufacturing failures associated with autologous T cell preparation (57, 58). Unlike autologous T cells, allogeneic T cells serve as universal T cells from healthy donors.

To circumvent graft-versus-host diseases, allogeneic T cells are produced through genetic disruption of HLA-A, HLA-B, class II major histocompatibility complex transactivator (CIITA), T-cell receptor alpha constant (TRAC), and Programmed Cell Death Protein 1 (PD-1), leading to their inability in recognizing allogeneic antigens (57, 59, 60). Allogeneic T cells can thus be used to produce nearly universal CAR T cells for antigens of interest (Figure 4). Though these genetic disruptions to the HLA loci, CIITA, TRAC, and PD-1 present on T cells serve to mitigate the effects of graft-versus-host disease, there remains a life-threatening risk of graft-versus-host disease. In addition, universal allogeneic generated CAR T cells are eliminated by the patient’s immune system, whereas autologous CAR T cells last on the timescale of months to years (4, 58).

Within activation, T cells are subject to incubation with a produced lentiviral vector encoding for the CAR. Viral machinery can be utilized to inject RNA within patient cells such that the RNA is reverse transcribed into patient T cell DNA. This reverse transcription encodes for the CAR and allows for maintained expression, via transcription and translation, of the CAR within the patient T cells as they are expanded (61, 62). While lentiviral vectors are commonly employed to express CARs on patient T cells, manufacturing of the CAR T cell can also be completed by utilizing mRNA transfection and the Sleeping Beauty transposon system.

mRNA based transfection utilizes electroporation to modify T cells with transient CAR expression (63). This approach yields significant antitumor effects in preclinical models of solid tumors like mesothelioma, while remaining a cost-efficient manner to manufacture CAR T cells (64). One drawback, however, is that the mRNA transfection approach requires multiple rounds of CAR T cell infusion (65). Additionally, in vivo delivery of mRNA can be mediated by lipid nanoparticles in a non-viral vector system (66). These lipid nanoparticles-mRNA formulations protect the mRNA from nuclease degradation in cellular fluids while avoiding clearance by renal glomerular filtration as mRNA is internalized by target cells (67). Lipid nanoparticles have been administered in the clinical setting for small molecule and mRNA delivery (68).

The Sleeping Beauty transposon-based system utilizes electro-transfer of non-viral plasmids for DNA transduction - this clinical grade production occurs at a cost that is nearly 10% of the cost associated to produced good manufacturing grade (GMP) virus. Additionally, with proven potential for human gene therapy, the Sleeping Beauty system can generate significant amounts of modified T cells within 28 days (69, 70). Further positives associated with the Sleeping Beauty transposon system include its relative inexpensiveness in early-stage clinical trials. However, there remain lingering questions regarding the efficacy of this system in comparison to lentiviral vectors and the potential remobilization of transposons (71).

CAR T cells can be produced in vivo using various techniques. The first of these techniques employs the utilization of lentiviral vectors. Though the lentiviral vector approach can be used to manufacture CAR T cells in vivo, the tropism of CAR-encoding lentiviral vectors to T cells must be confined. This can be done by replacing natural receptor usage of the vectors and fusing binders which recognize solely T lymphocyte binders, such as CD8 or CD4 (72–74). Lentiviral vectors have been used to generate human anti-CD-19 CAR T cells by targeting human CD8+ cells in a mouse model (75). While such an approach did lead to B-cell depletion and signs of cytokine release syndrome (CRS), the utilization of CD4 lentiviral vectors led to the production of anti-CD-19 CD4+ CAR T cells that led to more efficient tumor cell killing activity (76).

Adeno-associated virus (AAV) vectors can also be used to manufacture CAR T cells in vivo. The injection of an AAV vector containing the CD4 targeting CAR gene allowing for the generation of potent CAR T cells in vivo that ultimately led to tumor regression in a mouse model (77).

Non-viral approaches have also been established in the in vivo manufacturing of CAR T cells. Lipid nanoparticles (LNPs) can be used to transfect T cells with mRNA in vivo through conjugation to antibodies or ligands that target various receptors on these T cells, including CD3, CD8, CD4, CD7, and CD5 (78–83). Specifically, LNPs using antibody conjugation platforms have been used to selectively and efficiently produce functional CAR T cells in vivo through the use of CD3 and CD7-LNPs (84).

While the methods to manufacture the engineered CAR T cell therapy are further expanded and refined, there remains the manufacturing bottleneck of growth of large enough numbers of cells for clinical use with the use of bioreactor culture systems (85). There are significant costs associated with both the aforementioned leukapheresis and expansion steps. While methods and devices, like the G-Rex (Wilson-Wolf) exist to expand cells from low seeding densities, propagation of such devices pose an obstacle to the widespread manufacturing of CAR T cells (86). However, devices like the CliniMACS Prodigy, produced by Miltenyi Biotec, provide hope for the future given its ability to prepare, enrich, activate, transduce, expand, and sample cells (87). While expensive, this device has been used to produce autologous CD19 CAR T cells safely and efficiently for treatment of lymphomas, making it plausible for use in clinical trials (88, 89).

Rapid manufacturing of CAR T cells also remains a prominent area of research and development. The typical ex vivo manufacturing of T cells involves activation, viral transduction, and ex vivo expansion for at least six days – this time in activation and expansion can negatively affect CAR T cell therapies due to the progressive differentiation and loss of anti-leukemic activity of these CAR T cells over that time (90). Utilizing lentiviral vectors to transduce non activated quiescent T cells allows for the stable expression of a CAR in these non-activated T cells. This process saves time and resources while also promoting higher efficacy at lower doses as compared to those seen in activated T cells (90, 91).

The Epidermal Growth Factor Receptor (EGFR) is considered an important target for anti-tumor therapy in many solid tumors such as non-small cell lung cancer, gastrointestinal cancer, and colorectal cancer. Ongoing clinical studies are evaluating CAR T cell therapy targeting EGFR. In the case of NSCLC, a phase I trial for patients with EGFR mutation (NCT01869166↗) enrolled fourteen patients. Four of fourteen patients showed partial response for two to four months, and eight patients showed stable disease for two to four months as well. All Grade 3 events were manageable. The progression-free survival (PFS) was three months, and the overall survival (OS) was 4.9 months (134). Another phase I trial for patients with refractory/relapsed NSCLC showed therapeutic response (NCT03182816↗). One of nine patients maintained partial response for more than thirteen months, six patients showed stable disease, otherwise, and two patients showed progression in disease. All Grade 1-3 adverse events were manageable with proper treatment. The PFS was 7.13 months and OS was 15.63 months (135). An open label phase I clinical trial for patients with glioblastomas with EGFRvlll mutation (NCT02209376↗) assessed the efficacy of autologous CAR T cells targeting EGFRvlll, however, it was terminated prior to completion due to futility (136).

Human Epithelial Growth Factor Receptor 2 (HER2) plays a key role in targeted therapy, especially in breast cancer, and gastrointestinal malignancies. In addition, other types of solid tumors are applicable to Her2 targeted therapy and researchers highlighted the potential of HER2-targeted-CAR T cell therapy. A phase I trial of HER2-specific dual-switch CAR T cells in previously treated HER2 positive solid tumors (NCT04650451↗) showed the therapeutic and safety profile of CAR T cell products in combination with rimiducid. Five of nine patients achieved a prostate specific antigen (PSA) 50 response, and four of nine patients also got a PSA90 response. However, the last patients with metastatic castration-resistant prostate cancer experienced Grade 4 cytokine release syndrome (CRS) and reached second dose limiting toxicities (DLT) in the dose-escalation cohort. Given this safety issue, the trial was terminated. A phase I trial (NCT0195843) revealed that HER2 directed CAR T had the activity in advanced biliary tract cancer and pancreatic cancer. One patient achieved partial response for 4.5 months and five obtained stable diseases. Regarding safety profiles, most adverse events (AEs) were Grade 1 or 2, except one Grade 3 febrile syndrome and one abnormal liver function test (LFT) elevation that were both manageable (137). An ongoing phase I trial (NCT04511871↗) was designed to evaluate the efficacy of CCT 303-406, a HER2 targeting CAR T in refractory HER2 positive tumors. The primary objective was safety, tolerability, and optimization of the dose for phase II trial. In 2023, a new phase I trial (NCT05745454↗) began to recruit patients with HER2 positive solid tumors.

Carcinoembryonic Antigen (CEA) is an established tumor marker in epithelial solid tumors including colorectal cancer (CRC). As a result, CEA is regarded as a promising target for CAR T therapy. A phase I/II trial, which began to recruit patients in August 2023. is underway to assess safety and tolerability of CEA-targeted CAR therapy for patients with CEA-positive advanced solid tumors such as gastrointestinal (GI) malignancy, lung cancer, and breast cancer (NCT06006390↗). Another ongoing phase I trial (NCT06043466↗) is enrolling patients with various solid tumors, including GI malignancies and NSCLC. The primary objectives are determining the DLT and the maximum tolerated dose (MTD), while secondary objectives include disease control rate (DCR) within three months, maximum concentration of dosage, maximum time to reach the highest concentration, and the content of CEA in peripheral blood after infusion of the CAR T cells. Multiple other trials are evaluating the safety and efficacy of CEA-targeted CAR T cell therapy for patients with CEA- positive advanced solid malignancies (NCT05538195↗, NCT06126406↗). Lastly, a logic-gated Tmod CAR T therapy of A2B530 showed its safety and efficacy in patients with solid tumors associated with CEA expression and loss of heterozygosity (LOH) of HLA-A*02 (NCT05736731↗, EVEREST-1). The primary objectives of this phase I study are to assess the safety and tolerability of this therapy in patients with NSLC, CRC, and pancreatic cancer, while determining the maximum tolerated dose and the recommended phase II dose (RP2D).

Mesothelin (MSLN) is usually expressed in normal mesothelial cells; however, it is overexpressed in pancreatic cancers, ovarian cancers, and mesothelioma as well. MSLN is considered as a potential target for CAR T cell therapy and multiple trials tested its clinical activity (138). A phase I of MSLN-targeting CAR T cells demonstrated anti-tumor activity in pancreatic ductal carcinoma, ovarian cancer, and malignant mesothelioma (NCT03054298↗ and NCT03323944↗). In these studies, eleven of fifteen patients obtained stable disease (the best response), and one patient experienced dose-limiting toxicity of Grade 4 sepsis. Short persistence after the intravenous infusion of the CAR T cells appears to be associated with limited clinical activity. Ongoing studies are evaluating the regional delivery (intrapleural administration, intramural administration, etc.) of CAR T cells to enhance their proliferation, persistence, and function (139). MSLN-targeting CAR T cells (mesocarp-T) that express CTLA-1 and PD-1 antibodies are under evaluation for patients with various MSLN positive solid malignancies (NCT03182803↗). Other early phase trials (NCT03941626↗) aimed to evaluate MSLN-targeted CAR T therapy in hepatocellular carcinomas, glioblastomas, esophageal carcinomas, and gastric carcinomas, but their results remained veiled so far (140–142).

Receptor tyrosine kinase-like orphan receptor (ROR) is a subfamily of RTK and serves as a nuclear receptor interacting with intracellular transcription factor. It plays a significant role in embryonic development, and numerous studies have shown its aberrant expression in many diseases including advanced lung and triple negative breast cancers, making it a candidate for CAR T therapy (143, 144). A ROR1 targeted CAR T cell therapy, LYL797, is being tested in patients with advanced lung and triple negative breast cancers (NCT05274451↗). This phase I trial aimed to assess safety and determine the RP2D. The study plans to enroll fifty-four patients in dose-escalation and expansion cohorts. Another phase I trial of autologous T cell therapy for patients with ROR1+ triple negative breast cancer and ROR1+ non-small cell lung cancer (NCT02706392↗) assessed the safety profile in six patients, and no DLT was observed. However, the trial was terminated prematurely due to slow accruals (145). A trial of PRGN-3007 UltraCAR-T cells (NCT05694364↗) targeting ROR1 is currently under clinical development for patients with hematologic malignancies as well as ROR1+ triple negative breast cancers. The UltraCAR-T platform is designed to address various limitations of traditional CAR T cell manufacturing by using a nonviral gene delivery system and an overnight manufacturing process (146). In addition, PRGN-3007 is designed by incorporating Intrinsic down-regulation of PD-1 expression into a CAR T cell. The first patient was enrolled in April 2023 and a total of eighty-eight patients are expected to participate in this study. ROR2 has been considerably less studied than ROR1, however, various studies showed that it is highly expressed in solid tumors including melanoma, osteosarcoma, renal cell carcinoma, and head and neck squamous cell carcinoma (147). Hence, ROR2 has emerged as a candidate for CAR T cell therapy. A phase I trial of CCT301-59 (NCT03960060↗) is investigating the efficacy and safety of autologous ROR2-targeting CAR T cell therapy in adults with relapsed and refractory stage IV ROR2 metastatic solid tumors. The results have not yet been reported.

Claudin-18.2 (CLDN 18.2) is a member of the claudin family and is expressed in gastric and pancreatic adenocarcinoma (148). Thus, CLDN is a promising target for CAR T cell therapy. A multicenter phase IB trial of salvage CT041 CLDN 18.2 specific CAR T therapy (NCT03874897↗) against CLDN18.2-pisitve advanced gastrointestinal cancers indicated that the investigational drug had tolerable safety profile; there was a Grade 3 or higher hematologic toxicity and Grade 1/2 CRS in most patients. The overall response rate (ORR) and disease control rate (DCR) were 48.6% and 73.0%, respectively. Patients with gastric cancer showed a better result in ORR and DCR (57.1% and 75.0% respectively). Encouraged by this result, a phase II trial was initiated in heavily pretreated patients with gastric cancer (149). An ongoing autologous T cell therapy of CT048 targets CLDN 18.2 in gastroesophageal (GE) junction cancer, gastric cancer, and pancreatic cancer (NCT05393986↗). Another phase I trial (NCT05472857↗) is evaluating the safety and efficacy of IMC002, an autologous CAR T cell therapy, in patients with CLDN 18.2-positive solid tumors. An exploratory trial is currently enrolling patients to assess the safety profile of XKDCT086 (iPD-1-claudin 18.2 CAR T) (NCT05952375↗) in patients with CLDN 18.2 positive gastric cancer.

Glycoprotein tumor antigen (GD2) is a membrane protein overexpressed in tumors from neuroectodermal origin (150). Neuroblastoma, retinoblastoma, and melanoma are examples of highly GD2 expressing tumors. Given its limited overall expression in normal tissue and high expression in the aforementioned tumor types, GD2 is considered as a viable candidate for targeted therapy. Since 2018, an ongoing phase I/II trial of autologous GD2-targeting CAR T cells (NCT03373097↗) is assessing the safety and efficacy of the therapy in pediatric patients with high-risk neuroblastoma (109). Twenty-seven patients received GD2-CART01 and no DLTs were observed. CRS was observed in twenty out of twenty-seven patients, however, nineteen out of twenty experienced a mild form of CRS. ORR was 63% (seventeen of thirty patients); nine patients presented with a complete response (CR), and eight patients presented with partial response (PR). Additionally, a GD2-targeting CAR T therapy (NCT04196413↗) indicated clinical efficacy in young adults with H3K27M-mutated diffuse intrinsic pontine glioma and spinal diffuse midline glioma. Ten of twelve patients showed improvement in clinical and radiologic courses without severe toxicity. Third generation GD2 CAR T cells were tested in patients with GD2 positive neuroblastoma and osteosarcoma (NCT02107963↗). Thirteen patients received the drug. These patients maintained stable disease by day 28 and, ultimately, progressed in their disease (151). Conversion between monocyte and myeloid cells may have played a role in limited expansion and activity of the CAR T cells in this trial. Andras et al., conducted a phase I trial to evaluate the clinical benefit of GD2 CAR and iCaspase suicide containing T cells (NCT01822652↗) in patients with neuroblastomas. The efficacy of C7R-GD2 CART cells has been under investigation in targeting neuroblastoma and other GD+ solid tumors (NCT03635632↗). A fourth generation GD2 specific CAR T cell therapy (4SCAR-GD2) is also one of investigational drugs for GD2-positive solid tumors (NCT02992210↗). In the case of cervical cancer, multiple antigens are being examined for patients presenting with GD2, PSMA, Muc1, mesothelin, or other markers (NCT03356795↗).

Glycoprotein Mucin 1 (MUC1) is a member of the mucin family, which consists of mucus for lubricating and protecting normal epithelium. However, MUC1 promotes cancer invasion, metastasis, and neovascularization when it is expressed in cancer cells (152). A study of CAR T-TnMUC1 (NCT04025216↗) tried to assess the safety, feasibility, and efficacy of autologous T cells in TuMUC1 mutated tumors. Initially, six patients were enrolled, and two patients reached the planned dose. No significant toxicity profile was reported at the initial analysis. The study was terminated in 2023 due to an unfavorable risk-benefit analysis. A phase I/II study of autologous T cells co-expressing immune checkpoint antibodies (CTLA-4 and PD-1) and MUC1-targeting CARs was conducted in patients with MUC1 positive advanced solid tumors (NCT03179007↗). Another phase I/II trial (NCT02617134↗) evaluated MUC1-tageting CAR T cells for patients with MUC1 positive solid tumors. The enrollment status of the study is unknown, and the report is not available.

B7-H3(CD276) is a regulatory protein involved in the immune checkpoint pathway. It facilitates immune evasion of cancer cells and leads to proliferation, metastasis, and drug resistance. Given these findings, B7-H3 is a promising target for immunotherapy (153, 154). A TAA06 injection (B7-H3 targeting CAR T cell) trial (NCT05190185↗) was conducted for patients with malignant melanoma, lung cancer, and colorectal cancer. The enrollment status for this study is unknown. Another B7-H3 targeted autologous CAR T cell trial is planned (KT095 CAR T injection, NCT05515185↗). For pediatric patients with adrenocortical carcinoma, a trial of B7-H3-targeting autologous CAR T cell therapy is now recruiting patients (NCT04897321↗).

The glypican family is defined by the heparan sulfate proteoglycans consisting of six members (GPC1 to 6). It is related to regulation of Wnt hedgehogs, and fibroblast growth factors and bone morphogenetic proteins signal (155). GPC3-targeting CAR T cells are being tested in children with relapsed or refractor liver tumors (NCT02932956↗, GAP trial). An Interleukin -15 and -21 armored GPC3 targeted CART therapy will be ready to open (NCT04715191↗). For other pediatric GPC3-positive solid tumors, interleukin-15 armored GPC3 specific CAR T therapy is recruiting patients. (NCT04377932↗). An ongoing trial is also assessing T cells engineered with GPC3-CAR and Interleukin-15 (CATCH T cells) in patients with GPC3-positive solid tumors (NCT05103631↗).

Prostate stem cell antigen (PSCA) is a glycosylphosphatidylinositol (GPI) anchored cell surface antigen. It is expressed in epithelial cells of the genitourinary tract organ, skin, esophagus, stomach, and placenta. Overexpression of this antigen is observed in prostate cancer, bladder cancer, renal cell cancer, and hydatidiform mole (156). The results of a phase I trial of BPX 601 in prostate cancer were reported (NCT02744287↗). A total of eight patients were enrolled, and one patient gained sustained response (SD) after nine months. All patients experienced CRS (6 G1, 2 G2), and one patient reached DLT of neutropenic sepsis (G5). Due to the DLT of neutropenic sepsis, investigators suspended the trial. Additionally, a phase I trial of PSCA-CAR T cells was performed to assess the safety and tolerability (dose limiting toxicity) and determine the RP2D in patients with PSCA positive metastatic castration resistant prostate cancer (NCT03873805↗). A total of fourteen patients received CAR T cell infusion with varying treatments – the three cohorts were treated with the starting dose level (DL1, 100 million CAR T cells), the starting dose level that included the incorporation of lymphodepletion (DL2), and the starting dose level using a reduced lymphodepletion regimen (156). No DLTs were observed at DL1, while a DLT of grade 3 cystitis was observed at DL2, which led to the addition of DL3 which had no observed DLTs. CRS of grade 1 or 2 observed in five patients; prostate-specific antigen declines occurred in four patients with observed radiographic improvements. While limited persistence of CAR T cells was seen beyond twenty-eight days post-infusion, there were indicative changes as a result of therapy. Such changes included activation of peripheral blood endogenous and CAR T cell subsets, change in the tumor microenvironment, and TCR repertoire diversity in a subset of patients (156).

Prostate specific membrane antigen is expressed at high levels in metastatic castration-resistant prostate cancer and has been established as a tumor-associated antigen for immune therapy (157). A phase I clinical trial of autologous anti-PSMA CAR T therapy in metastatic castration-resistant prostate cancer looked to determine the efficacy and safety of the treatment (NCT04249947↗). In this study, thirteen patients were treated with the P-PSMA-101 autologous CAR-T therapy. Declines in prostate specific antigens were seen in seven patients. CRS was seen in six of patients (one Gr 3 or greater). DLT was seen in one patient with macrophage activation syndrome (Gr 3 or greater CRS event) (158). Additionally, a first-in-human phase I trial of castration-resistant prostate cancer directed CAR T cells, targeting PSMA, and equipped with a dominant negative TGFβ receptor looked to determine the safety, bioactivity, and disease response. A total of eighteen patients were enrolled, and thirteen received varying doses of therapy (NCT03089203↗). Five of thirteen patients developed greater than grade 2 CRS – one of these patients experienced clonal CAR T cell expansion and death following grade 4 CRS with sepsis (157).

Guanylyl cyclase 2C (GCC) is a transmembrane receptor for an enterotoxin produced by diarrheagenic enteric bacteria. The receptor has a unique expression pattern in colorectal cancer (159). Therefore, a GCC-targeting CAR T therapy was tested in patients with metastatic colorectal cancer (NCT05287165↗). Nine patients received the treatment, and five of nine patients developed GR 1 - 2 cytokine release syndrome. The DCR was 66.7% and ORR was 11.1%. No DLT was observed. These results will open a dose-expansion cohort. In addition, the investigational drug XKDCT080 is recruiting candidates with GCC positive advanced solid tumors (NCT05875402↗).

Tyrosinase Related Protein 1 (TYRP1) is a key protein in melanine synthesis, and it is present in the membranes of normal and malignant melanocytes (160). Therefore, targeting TYRP1 was considered as a potential therapeutic approach in patients with malignant melanoma, and pre-clinical studies were performed to prove its efficacy. A recent report by Ribas and Puig-Saus et al., showed that TYRP1 targeting CAR T cell therapy demonstrated significant antitumor in tumor cells with high TYRP1 expression in murine and patient-derived cutaneous, acral, and uveal melanoma models. The study did not observe significant toxicities and clinical translation is underway based on the efficacy and safety profile (161).

Six transmembrane epithelial antigen of the prostate 1 (STEAP1) is a transmembrane protein that showed enriched expression in metastatic castration-resistant prostate cancer (mCRPC) and other solid tumors (162). An antibody-drug conjugate (ADC) targeting STEAP1 was tested in a phase I trial (NCT01283373↗) and showed an acceptable safety profile in patients with mCRPC (163). A phase I trial of STEAP1 and CD3 targeting bispecific-antibody is ongoing as well in patients with mCRPC (164). Bhatia V. et al. reported the study evaluating STEAP1-targeting CAR T cell therapy for mCRPC using in vitro and in vivo models. The study revealed that STEAP1 is highly expressed among mCRPC samples. Additionally, the in vitro cytotoxic assay was specific and strong even with minimal expression of STEAP1 on cancer cell surfaces. An in vivo study with NSG mice and a syngeneic mouse model demonstrated that STEAP1 CAR T cell therapy is efficacious and safe. Based on these promising results, clinical translation is underway (162).

NKG2D is a cell surface receptor usually expressed on cytotoxic immune cells. NKG2D is a natural receptor on the surface of natural killer cells that can recognize multiple ligands on the tumor cells. NKG2D therefore could be an efficient chimeric antigen receptor for these tumors. NKG2D ligands are present in tumor cells of various origins including colorectal cancer, ovarian carcinoma, pancreatic cancer, prostate cancer, acute lymphoblastic leukemia, lymphomas, and in a range of cancer cell lines indicating their use for both hematological and solid tumors (165, 166). NKG2D-targeting CAR T therapy was evaluated in patients with advanced solid tumors (NCT04107142↗ and NCT05302037↗).

Ephrin type-A receptor 2 (EphA2) is a transmembrane glycoprotein observed in epithelial cells during proliferation. EphA2 is upregulated in malignant lymphomas and EphA2-positive metastatic solid tumors, and it is regarded as a target for CAR T cell therapy (NCT05631886↗ and NCT05631899↗). A CAR T cell trial targeting epithelial cell adhesion molecule (EpCAM) is ongoing now (NCT02915445↗) (167). CD22 and CD79, B cell membrane antigens, are targets for CAR T cell therapy in patients with CD22-positive and CD79-positive B cell malignancies given their in vivo results (NCT04556669↗ and NCT02830724↗) (168). Lewis Y is a subtype of blood group related antigens and is associated with poor prognosis in breast cancer (169). Lewis Y specific CAR T cell therapy showed efficacy in an in vivo trial, and a phase I trial was conducted (NCT03851146↗) in patients with Lewis Y antigen-expressing solid tumors. This infused CAR-T therapy displayed poor persistence in patients; however, further tuning through utilization of features of stem-like T cells in designing the CAR T cell therapy may lead to better clinical performance (170).

While there are positive results associated with the presented clinical trials, there remain significant therapeutic limitations. As evidenced by the presented CAR T therapy clinical trials for solid tumor treatment, a recurring theme was adverse event reporting in patients due to dose limiting toxicities (DLT). Namely, clinical trials for CAR T therapy in HER2 (NCT04650451↗) and PSCA(NCT02744287↗) were suspended due to DLTs. Another recurring challenge in trials was cytokine release syndrome (CRS), which is the most common and one of the most critical adverse events associated with CAR T therapy. Many trials, notably CAR T trials targeting HER2 (NCT04650451↗), Claudin 18.2 (NCT03874897↗), GD2 (NCT03373097↗), PSCA (NCT02744287↗), and PSMA (NCT04249947↗, (NCT03089203↗), reported that patients experienced CRS (regardless of severity). As such, the management of DLTs and CRS are pivotal issues in successful treatment and advancement of CAR T therapies for solid tumors.

CRS usually occurs within the first week of treatment, and its clinical manifestation varies from fever to severe hypotension requiring vasopressors. The mechanism of development is related to the release of inflammatory cytokines such as IFN-γ, IL-6, IM-10, macrophage inflammatory protein (MIP)1β, monocyte chemoattractant protein (MCP)-1, and GM-CSF (171, 172). Several factors may predict a higher possibility of CRS – high tumor burden, high CAR T dose, prior lymphodepletion conditioning regimen, and high peak CAR T cell counts (173). Ironically, CRS may also be associated with a positive response to CAR T cell therapy given its immune-stimulatory mechanism. Secondary hemophagocytic lymphohistocytosis/macrophage activation syndrome (HLH/MAS) originates from immune hyperactivation, which shares common mechanisms with CRS. Thus, HLH/MAS may overlap with CRS. Furthermore, serum markers in severe CRS meet the criteria of HLH/MAS.

Neurotoxicity, known as immune effector cell-associated neurotoxicity syndrome (ICANS), is also a common adverse event. It occurs in approximately 65% of patients with hematologic malignancy (174). ICANS presents with tremors, dysgraphia, aphagia, delirium, and altered mentality in early phases; ICANS can deteriorate to seizures, stupor, obtundation, and coma, sometimes associated with cerebral edema. The exact mechanism of ICANS has remained unclear, but immune stimulating cytokines such as IL-1, IL-6, and GM-CSF play roles in its development (175, 176). The inflammation caused by myeloid cells can induce capillary leaks, endothelial activation, and dysfunction of the blood-brain barrier.

Optimization of the efficacy of CAR T therapy aims to maximize immune responses with acceptable toxicity. IL-6 inhibition with tocilizumab is a widely used method to control excessive reactions; similarly, corticosteroids suppress overall immune cells (177). These two drugs are the backbone of CRS management. ITK inhibition by ibrutinib may also reduce the activity of inflammatory cytokines (178, 179).

Zarei et al. showed that the immunomodulatory drug lenalidomide enhances the effect of NKG2D CAR T cells to kill colorectal cancers in vitro (180). The NKG2D CAR T cells, generated from transduction of second generation NKG2D-CAR, showed significant cytotoxic activity against colorectal cancer cell lines, HCT116 and SW480, as compared to non-transduced parental T cells. Addition of lenalidomide showed dose dependent increase in the cytotoxicity and cytokine secretion by NKG2D CAR T cells. This study outcome suggests that combinational therapy utilizing NKG2D-based CAR T cells and lenalidomide has a high potential for effectively eliminating tumor cells in vitro and that immunomodulatory medication lenalidomide (LEN) may increase the effectiveness of CAR T cells in the treatment of solid tumors.

Additionally, the effect of natural products on the function of CAR T cells has been investigated. Huang et al. (2023) assessed the effect of gastrodin (GAS) on CAR T cells targeting interleukin-13 receptor α2 antigen (IL-13Rα2 CAR T) in the brain against glioblastoma multiforme in both in vitro and in situ glioblastoma models (181). In a transwell assay, GAS increased the migratory and destructive capacity of IL-13Rα2 CAR T cells with no effect on cytokine release. There was elevated expression of S1P1 with GAS treatment which encouraged the entry of CAR T cells into the brain and bone marrow. In the transcriptomic analysis, there was upregulation of genes such as add2 and gng8 related to skeletal migration. GAS treatment improved the mobility of IL-13Rα2 CAR T, enhancing their ability to recognize the tumor antigen of glioblastoma, indicating its potential application of CAR T for the treatment of solid tumors.

These studies show that the utilization of drugs and natural products can further improve CAR T cell function in vitro.

Some gene mutations were identified as enhancers to autologous CAR T cell therapy through CRISPR screening. For example, the deletion of TCR encoding genes and inactivation of PD-1 by CRISPR/Cas9 methods caused an increase in immune activity (19). In addition, the knock-down of Tet methylcytosine dioxygenase 2 (TET2) encoding genes led to the improved efficacy of CD19-based therapy (20). CRISPR knock-out screens of RASA2, which is a RAS GTPase-activating protein, in T cells led to increased T cell activation and led to more persistent cancer cell killing (201). Similarly, an in vivo pooled CRISPR-Cas9 screening approach targeting REGNASE-1 showed that T cells can be reprogrammed to long-lived effector cells with better persistence in tumors as compared to control T cells. Further CRISPR-Cas9 screens showed that BATF, the key target of REGNASE-1, loss led to the greater accumulation of REGNASE-1 deficient T cells (202).

Additionally, deletion of DNA methyltransferase 3 alpha (DNMT3A) indicated the enhanced immunogenicity and inhibition of T cell exhaustion (203). The recent study by Garcia J. et al. demonstrated that the other fusion mutation in caspase recruitment domain-containing protein (CARD11) and Phosphoinositide-3-Kinase Regulatory Subunit 3 (PIK3R3) improved tumor killing effect by enhanced signaling of CARD11–BCL10–MALT1 complex (204). Therefore, the exploration of various genetic mutations in T cell biology can be a promising solution for improvement and overcoming resistance. In the context of engineering CAR T cells, the overexpression of canonical AP-1 factor c-Jun has enhanced T cell expansion potential, increased functional capacity, and improved anti-tumor response in multiple in vivo mouse tumor models (205).

In November 2023, the FDA announced an investigation regarding the risk of T cell malignancies, including CAR positive lymphomas, for those who received BCMA or CD19 directed CAR T therapies. The recent report by Ghilardi et al. indicated that three cases of T cell lymphoma were observed after anti-CD19 CAR T therapy for non-Hodgkin B cell lymphoma was administered (206). According to this report, further studies among patients treated at the University of Pennsylvania showed that 3.6% (16 of 449 patients) reported a secondary malignancy – one case was identified as T cell lymphoma with a very low incidence. In addition, Garcia et al. discussed the concerns associated with T cell malignancies in those administered these CAR T cell therapies (204). Their study, and namely their experimental data derived from a longer term follow up in murine studies, did not suggest that there was an increased risk of developing secondary T cell malignancies when fusion genes were inserted into CAR T cells. Typically, these insertions are one of the naturally occurring mutations in human T cell cancers. Instead, Garcia et al. concluded that pre-existing genetic features could contribute to secondary sporadic malignancies. Moving forward, the field will need to continue address these concerns and adhere to a more cautious approach until more evidence is gathered.

CAR T therapy is proving to be a viable option in cancer treatment, especially with the advancements in CAR engineering. The areas of concern, namely antigen heterogeneity, the immunosuppressive tumor microenvironment, and tumor infiltration, have been identified. Focused research in these areas, as well as in expanded areas of CAR engineering and antigen discovery, has progressed. One of the prevailing problems that remains is that often CAR Ts do not reach target antigens in solid tumors in sufficient numbers because of their destruction by lymphatic system or by the tumor microenvironment. To overcome these problems, a nanoparticle-based packaging and delivery system is being examined. A variety of nanotechnologies, including hydrogel, nanoparticle conjugation, transient CAR expression in T cells through RNA delivery, and others are being explored (207). These advances and further engineering of CARs looks to further the ability of CAR T cell therapy to be used as a therapeutic approach for solid tumors.