The first-generation CAR T model comprises a CD3ζ chain as a key transmitter of signals from endogenous TCRs. Following a successful outcome in pre-clinical trials, this type of drug entered into phase I clinical trials for leukemia, lymphoma and various other types of cancer, including ovarian cancer and neuroblastoma (17). Despite the inadequate antitumor action owing to the lack of activation, persistent exposure to the tumor environment has resulted in continued therapeutic effects in patients with B-cell lymphoma infused with α-CD20-CD3ζ CAR T cells and a number of patients with neuroblastoma treated with scFv-CD3ζ CAR T cells (18). The heavy and light chains constitute structural parts of the B cell receptor or antibodies, called scFv, which are fused to the CD3 domain or T cell-activating ζ chain of the TCR to create non-MHC-restricted activating receptor molecules. These modified molecules are capable of enhancing T cell antigen detection and cytotoxicity by specifically targeting tumor cells (19). The original ‘true’ CAR was designed by integrating an scFv antibody receptor directly with the CD3ζ domain. This model was subsequently named the ‘T body approach’ and these synthetic signaling receptors are now known as CARs or chimeric immune receptors (20).

The success of first-generation CARs in phase 1 clinical trials paved the way for second-generation CAR T cell therapy. This model of CAR T cells was established to elicit a more effective anti-leukemic response in phase I clinical trials with complete remission rates of up to 90% in patients with recurrent B-cell ALL (B-ALL). Here, the second-generation anti-CD19 T cells were integrated into a 4-1BB or CD28 co-stimulatory domain attached to the CD3 domain (21). However, significant concerns remain regarding its efficacy and safety and making it more robust. Anticipating these concerns, second-generation CARs aimed at integrating intracellular signaling domains from different co-stimulatory molecules, such as CD28, 4-1BB or CD137, inducible T cell costimulator (ICOS) or CD278, OX40 or CD134 fused to the cytoplasmic tail of the CAR, thus amplifying the signal (22). First-generation CARs contain a CD3ζ chain as a key transmitter of signals from endogenous TCRs, whereas second-generation CARs contain a CD3ζ chain and a single costimulatory molecule, which is why the receptor is known as a second-generation CAR. For example, anti-CD19 CARs consisting of 4-1BB or CD28 signaling domains produced notable complete response (CR) rates in patients with relapsed and refractory B-cell malignancies (23). CR represents complete response or disappearance of all signs of cancer in response to treatment [The United States Food and Drug Administration (FDA) 2007]. Consequently, CD28-based CARs have a rapid proliferative reaction, thus enhancing T effector cell functions, whereas 4-1BB-based CARs lead to improved T cell accumulation (24).

To extend the antitumor efficacy, third-generation CARs comprise two signaling domains and the CD3ζ chain, such as CD3ζ-CD28-OX40 and CD3ζ-CD28-4-1BB, to achieve improved activation signal, prolonged proliferation, elevated cytokine production and effective function (25). For example, a third-generation CAR consisting of α-CD19-CD3ζ-CD28-4-1BB reported complete remission rates by infiltrating and lysing cancer tissue in patients with chronic lymphocyte leukemia (26). In addition, certain CAR T cells function as memory cells, thus preventing tumor relapse. Despite the significant curative effect, the irrepressible activity of CARs and their increased antitumor efficacy is associated with life-threatening and unfavorable outcomes, with increased secretion of pro-inflammatory cytokines, multi-organ dysfunction, pulmonary failure and death (27).

All earlier CARs were based on a precise stratagem and helped in mediating the T cell anticancer response. However, these suffered from limitations, including a lack of antitumor activity against solid tumors owing to large phenotypic heterogeneity and deterioration attributed to antigen-negative cancer cells. These shortcomings led to the development of a novel CAR stratagem (28). The fourth-generation CAR was introduced to establish the tumor background via the inducible expression of transgenic immune modifiers, such as interleukin (IL)-12, which activate innate immune cells and enhance T cell activation to reduce antigen-negative cancer cells in the marked lesion (29,30) (Fig. 2).

Viral vectors, such as γ-retrovirus, lentivirus and adenovirus vectors are generally used in gene therapy. Retrovirus transduction is one of the commonly used delivery methods in gene therapy. The method involves a reverse transcriptase that promotes the stable integration of artificial genes into the host genome (31). To create a vector, γ-retroviral coding sequences are substituted by a gene of interest. The retroviral vectors possess an innate capability to disturb the genomic section and results in neoplastic transformation. Thus, γ-retroviral vectors have been used in gene therapy applications (32). Lentiviral vectors are retroviruses derived from human immunodeficiency virus-1, which serve as a major tool for the delivery of transgenes into mammalian cells. The advantage of using lentivirus vectors is their efficient transduction and their stable integration and expression into non-dividing and dividing cells both in vitro and in vivo (33).

The mechanism of transposon transfection is different from that of viral transfection. Delivery via transposons is a non-viral process that uses transposon DNA and a transposase enzyme for stable gene transfer. Two important vectors, namely piggyBac (PB) and sleeping beauty (SB), are frequently used (34). Transposons shift from one gene position to another position via a cut-and-paste mechanism. The mechanism of the transposon system using SB and PB involves four steps: i) The transposase enzyme helps in recognition and binding to the transposon; ii) a synaptic complex is produced by the coupling and binding of the repeat elements at both ends of the transposon; iii) the transposon excises the genetic element to be transposed, and iv) the excised element is reintegrated into the target location (35). The transposase for both vectors (SB and PB) consists of the DNA-binding domain, transposable catalytic domain and nuclear localization signal. The SB vector is a safer substitute for viral vectors owing to its innately low enhancer activity, non-pathogenic source and minimum epigenetic alterations at the incorporation site (36). By contrast, the PB vector has been shown to provide a large load capacity and additional competent transposition action in in vivo studies. PB vector-mediated CAR T therapy is considered safer than that based on the SB vector. Therefore, use of the PB vector results in greater CAR production, as evident from the production of CAR targeting CD19 using a PB vector (37).

Electroporation has evolved as a useful technique for modifying genes of diverse cell types. The target cells are exposed to electric fields to temporarily disrupt their cell membranes. This allows the charged molecules to enter the cells. Square-wave and pulse-based systems are new electroporation devices (38). The Lonza Nucleofector II Electroporator performs effective genetic alteration of T cells using certain electric parameters and electroporation buffers (39). The electroporation of human T cells has been reported to be associated with ~40–60% of gene expression and 80% of cell viability (40). One of the limitations of electroporation is the low transfection efficiency and redundant cell damage (41).

The most critical step in CAR T cell therapy is T cell activation, for which cells need to be incubated with the viral vector including a CAR gene. However, stable insertional mutagenesis can be dangerous as its possible effects on humans remain to be fully elucidated (42). Retroviral or lentiviral vectors offer permanent gene integration into host cells, with the potential integration in close proximity with a proto-oncogene which can result in an oncogene. The various disadvantages associated with these transfer methods suggest that a safer alternative is needed (43).

The study by Smith et al elucidated the use of nanotechnology to resolve the tumor-targeting problem and make the therapy cost-effective. They reported that polymeric nanocarriers can efficiently deliver leukemia-specific CAR genes targeted to specific ligands on the host T cells in situ (44). A number of nanoparticles used for gene delivery have been investigated preclinically and reported. For example, magnetic nanoparticles, such as Fe₃O₄, integrate into the cell-penetrating peptide complexes. These oligonucleotides maintain stable cell transfection and plasmid transfection in addition to gene silencing and splice correction (45). A recently published study by Smith et al at the Fred Hutchinson Cancer Research Center, Seattle, confirmed that polymeric nanoparticles carrying DNA effectively introduced CAR genes into T cell nuclei and recognized leukemia cells distinctively (46). The use of nanoparticles to replace viral vectors for gene delivery has proved advantageous, as evident from the restriction of unspecific pro- or anti-inflammatory effects or pro- or antiproliferation effects (47).

A bispecific receptor is one that contains two distinct antigen recognition domains attached and placed with two distinct intracellular signaling domains that are expressed as two different CARs on a single cell surface. At present, bispecific CAR CD19/CD20 has been introduced as a novel synthetic molecule that can recognize and bind to more than one targeted tumor antigen on the cancer cell surface. Therefore, it can create a synergistic cascade of effector molecules when it encounters two tumor antigens (53). Additionally, the bispecific CAR conserves the cytolytic capacity of T cells, i.e., if one of the objective molecules is not accessible to CAR T cells due to a cellular hindrance such as mutation of a target antigen or loss of the target antigen commonly found in malignant cells, a bispecific CAR can counterbalance the tumor evasion (54). Investigated therapeutics include CD3 of T cells and tumor antigens, such as CD19, on malignant cells. The curative ability is evident from blinatumomab, an approved bispecific T cell-engager against relapsed/refractory B-ALL (55). Positive results have been reported in patients <15 years of age with relapsed/refractory ALL, with a heightened response in 26/36 (72%) patients in 9 months. Blinatumomab was discovered while studying a patient who was negative for minimal residual disease (MRD) and who was found to be MRD-positive following consolidated chemotherapy (56). Numerous other bispecific CARs have been investigated preclinically by several research organizations, including CD20/CD19 and CD20/CD3 (57,58). Targeting T cells with bispecific CARs was shown to eliminate transplanted pediatric ALL in a study, whereas T cells targeted by CD20 CAR did not control the disease (59).

Conventional CARs cannot meet higher expectations under certain circumstances, such as the downregulation or alteration of targeted antigens that can occur in cancerous cells. These conditions subsequently lead to antigenic loss or escape variations (60). To overcome this shortcoming, scientists are attempting to develop advanced technology, in which two particular antigen recognition sites are joined by a linker, placed in tandem on a single intracellular domain and expressed as a single CAR on a cell surface, termed a Tan-CAR. This model enables the synchronized targeting of both antigens on a single cancer cell or tumor microenvironment. In this manner, it enhances the activation and stimulation of T cells by increasing their avidity and expanding their therapeutic properties (61). Tan-CARs have several significant curative implications as they are as potent as standard disease models with single antigen-specific CARs. Additionally, they are more efficient and less noxious in a higher disease load setting (62). This may be attributed to optimized cytokine production and limited cell killing, as evident from preclinical studies on Tan-CAR T cells in a mouse tumor model, which confirmed its possibility for remedial implication in human disease consisting of B-cell antigen CD19 and human epidermal growth factor receptor 2 (HER2) (63). Recently, a trivalent CAR T cell was designed by a group of scientists by simultaneously co-targeting multiple antigens, such as HER2, IL-13 receptor α2 and ephrin A2, to overcome interpatient changeability with a propensity to target almost 100% of tumor cells (64).

It has been reported that certain novel immunoinhibitory receptors are involved in T cell activation and the attenuation or termination of T cell responses, such as programmed death-1 (PD-1), programmed death-ligand 1 (PD-L1) and cytotoxic T-lymphocyte-associated antigen 4 (CTLA-4). These pathways are regarded as cancer immunotherapy breakthroughs (65). Recently, two immunologists won the Novel Prize in Physiology/Medicine, Dr James Allison and Dr Tasuku Honjo, for their profound discovery in the field of cancer immunotherapy. Dr Allison's work focused on the T cell surface protein CTLA-4; he found that the inhibition of immune cells and antibodies against CTL-4 eliminated malignant growth and prevented new tumor formation (66). This finding was tested on 14 patients with metastatic melanoma, with relapse observed in three patients. In 2011, the FDA approved an anti-CTLA-4 (ipilimumab) antibody as a treatment for high-grade melanoma (67).

Similarly, Dr Honjo's team investigated a novel T cell protein, PD-1, which was discovered in 1992 (68). They observed that PD-L1 presenting on healthy cells and malignant cells bind to PD-1. They reported another similar molecule, PD-L2, that also binds to PD-1. Based on these findings, they published a report that malignant cells produced PD-L1 and blocked PD-L1 using a counteracting antigen to inhibit tumor growth (69). The first clinical trials to target malignancy were launched in 2006, which indicated significant viability in a number of patients in 2012. The FDA approved the main PD-1 checkpoint inhibitors, nivolumab and pembrolizumab (70,71), for treating melanoma in 2014 (72). However, a single treatment may not be sufficient, which is the reason that a consolidated treatment targeting CTLA-4 and PD-1 is currently being investigated. Both Dr Allison and Dr Honjo encouraged the merging of various strategies to generate more advanced forms of the drug for the immune system to inhibit tumor cells more effectively (73).

Certain tumor cells contain a high level of PD-L1, which assists in evading immune attack. It has already been investigated that PD-1- or CTLA-4-based I-CARs can efficiently control the cytotoxicity, secretion of cytokines, and proliferation and cytotoxicity stimulated by endogenous TCR or an activating chimeric receptor (74). I-CARs are designed to control the actions of CAR T cells through inhibitory receptors. This advanced feature unites the action of two chimeric receptors; one of these generates a dominant-negative signal that restricts the responses of activated CAR T cells by the activating receptor. I-CARs can inhibit the activator CAR response to antigens expressed only by normal cells, thus differentiating between cancer and normal cells (75). Therefore, using genetic engineering to inhibit T cell inhibition physiology and regulate T cell response can be harnessed in an antigen-selective manner. This has been experimentally confirmed preclinically in a mouse model by designing an I-CAR using surface antigen recognition domains CTLA-4 and PD-1. In mice lacking CTLA-4 receptor, substantial T cell activation and proliferation were observed, ultimately leading to rigorous systemic autoimmune disease (76). Similarly, PD-1 is another the inhibitory receptor, which was found to be particularly expressed by activated T cells causing glomerulonephritis and arthritis in C57BL/6 mice and certain non-obese diabetic mice affected by insulitis (77).

Initially, several CAR constructions contained scFv of murine origin. This is associated with the risk of an immune response to the modified cells, and the resulting anaphylaxis of CAR T cell transfer cannot be avoided. These disadvantages limit the persistence of the infused cells. Therefore, besides conventional CAR, a physiological CAR has been developed, also known as a receptor-ligand CAR, which can recognize and bind to tumor antigens, such as HER3 and HER4 (78). The physiological CAR consists of an antigen receptor and a CD3ζ intracellular signaling domain with or without a transmembrane region, which can also be engineered into immune cells to target the ligands expressed on tumor cells. This approach increases the capability of T lymphocytes to distinguish tumor-related targets and eliminate cancer cells (79). This physiological CAR is an emerging field of CAR T cell therapy with limited published reports. Experimental trials may have been initiated but the outcomes have not yet been published.

Although scFv is specifically directed against tumor-associated antigens, the recognition specificity potential of CAR T cells is inadequate. Therefore, uCARs were developed to overcome this limitation. To construct a universal CAR, biotin or anti-fluorescein isothiocyanate (FITC) scFv is used as a targeting region, which is fused to a transmembrane domain with one or two endodomains (80). The uCAR-expressing T cells can efficiently recognize and remove cancer cells through the binding of FITC-labeled or biotinylated antigen-specific mAbs, which, in turn, activate the T cells and stimulate their proliferation and production of cytokines (81). The mSA2 CAR T cell is a uCAR, which can be fused to a biotinylated tumor-specific antibody to specifically target various types of tumor. As with interferon γ, it is well equipped for interceding malignant cell lysis and cytokine production (82). Human clinical trials involving uCAR T cells are ongoing. Two children with relapsed and refractory B-ALL achieved molecular and cytogenic remission following uCAR T therapy. Clinical trials have been initiated on uCAR T cell therapy specific to CD19-positive cells (NCT03166878 and NCT03229876). However, detailed information and conclusions are not yet available (83).

NK cells are a type of cytotoxic T cell that are essential for natural immunity. The function of NK cells is similar to that of cytotoxic T cells in the adaptive immune response of vertebrates. Immune cells generally identify MHC complexes expressed on infectious cell surfaces to elicit cytokine production, thus generating an immune response, culminating in the apoptosis or death of infectious cells (84). NK cells are the only immune cells that identify infected cells in the absence of MHC and antibodies, thus eliciting a rapid immune response. These are known as ‘natural killers’ as they do not require activation to destroy cells that are devoid of ‘self’ MHC class I molecule markers (85), making them important for destructive cells with missing MHC I markers.

Cancer cells that do not cause any inflammation are treated as self by the immune system and do not stimulate a T cell response. NK cells produce a number of cytokines, including tumor necrosis factor α, interferon γ and IL-10, which act as immune suppressors (86). The activation of NK cells leads to the gradual formation of cytolytic effectors cells, such as dendritic cells, macrophages and neutrophils, which consequently facilitate antigen-specific T and B cell responses. NK cell-mediated tumor cell lysis involves several receptors, including NKp44, NKp46, NKG2D, NKp30 and DNAM. Malignant cells usually express NKG2D in addition to ULBP and MICA (87,88). The clinical effectiveness of CAR T cells has been shown for ALL; however, this therapeutic approach has not been confirmed for acute myeloid leukemia (AML), suggesting the need for other therapeutic options (89). Hypothetically, automotive NK cells have an additional favorable toxic effect in comparison to CAR T cells, particularly for avoiding unfavorable effects such as CRS. Additionally, in contrast to T cells, donor NK cells do not target non-hematopoietic cells, indicating that NK cell-mediated antitumor activity may be activated in the absence of graft-vs.-host disease (90) (Fig. 4).

CRISPR is a gene modifying tool that uses a guide gRNA to modify a DNA sequence. As this technology is an integration-free gene incorporation system, it offers a foolproof and competent gene knock-in process. Following the significant progress associated with CRISPR technology, it has the potential to emerge as a promising immunotherapy (91). The CRISPR system can be directly applied to mammalian cells via transfection using a plasmid that contains both nuclease and sgRNA. Cas9 encodes a large molecule with a multifunctional DNA endonuclease and is known to excise dsDNA from 3-bp upstream of the protospacer adjacent motif (PAM) (92). Once the nuclease binds to its gRNA, the compound scans for an integral target DNA sequence (93). The PAM sequence has a significant role in recognizing self and non-self sequences. The local PAM sequence is usually used as a ‘spy’ on the nuclease sequence 5-NGG-3, in which N is any of the four deoxyribonucleic acid bases (94).

Recently published reports suggest that CRISPR technology can deliver the CAR gene to the TRAC locus of T cells. In this regard, CRISPR-edited universal-CAR T cell therapy has been used in humans (NCT03166878 and NCT03229876). This technology is rapidly developing with the potential for gene correction (82). It is reported that anti-CD19 CAR to the TCR α constant locus (TRAC locus) not only results in increased T cell potency but also in the consistent expression of CAR in peripheral blood T cells (95). CRISPR-modified cells have been shown to perform well in generally constructed CAR T cells in a mouse model of ALL (96). It has also been demonstrated that targeting the TRAC locus turns on CAR signaling, thereby initiating successful internalization and stimulation of the CAR for single and repeated exposure to a tumor antigen. It also delays effector T cell differentiation and exhaustion (97). Multiplex genome editing is another attractive application of the CRISPR/Cas9 tool. Proficient genomic disturbance of multiple gene loci to create a universal donor cell and a potent effector T lymphocyte targeted to different inhibitory pathways, such as PD-1 and CTLA4, is established by incorporating several gRNAs into a CAR vector. Furthermore, two-fold knockout of the TCR gene and HLA class I can effectively permit the generation of an allogeneic uCAR T cell, in addition to CAR T cells that are universally Fas-resistant via three-fold gene disruption (98).