Two of the CAR-T products currently available in the market are based on retroviral vectors (Yescarta and Tecartus). Vector-based on murine leukemia virus (MLV) is the most commonly used gamma retroviral vector22 and it has been successfully applied in other therapies such as T-cell immunotherapy for combined immunodeficiency- (SCID) X1 disease.23 Although, immunodeficiency disease has been successfully treated, in some patients, leukemia has been caused by the integration site of retroviral vector.23,24 Gamma retroviral vectors tend to integrate close to promoter regions, which may be the cause of their oncogenic properties. On the other hand, lentiviral vectors, derived from the retroviral family, Lentiviridae, showed safer and better integration properties than their gamma retroviral counterparts.25,26 Lentiviral vectors are able to insert genetic material into non-dividing cells,27 whereas gamma retroviral vectors transduce only dividing cells.28 The advantages of the lentiviral system are reflected in the large number of CAR-T products under development using this system and three of the CAR-T products available on the market use lentiviral vectors (Kymriah, Breyanzi and Abecma). (Table 1).

Since the beginning of CAR-T cell therapy, viral vectors have demonstrated high transduction efficiency. However, this type of vector presents i) a risk of insertional mutagenesis,29 ii) it can be immunogenic and iii) its manufacture is extremely expensive.30

An alternative to viral vectors is the transposon system. A transposon is a sequence of DNA with the ability to change position within a genome via transposase excision and insertion.3 A variety of transposon-based systems have been reported for the production of CAR-T cells. The sleeping beauty (SB) transposon system showed good results in T-cells modified with anti-CD19 CAR, presenting antitumor activity both in vitro and in vivo.18 The main obstacle to using the SB transposon on a large scale is its low integration rate. New systems such as the SB1 and SB100X have much higher transposition rates than the native SB transposon.31,32 SB was the first non-viral vector to be used in clinical trials to generate CD19-specific CAR-T cells for treatment of NHL (Non-Hodgkin lymphoma) and ALL (Acute lymphocytic leukemia) (NCT00968760, NCT01497184) and recently, a new clinical trial with SLAMF7 CAR-T cells prepared with SB100X was initiate to treat multiple myeloma patients.33 Although the integration profile of SB can be considered biologically safe, other transposons such as PiggyBac (PB) transposon demonstrated properties more similar to viral-based vectors.31 PB were successfully used in the generation of CAR T-cells against CD19 to treat hematological malignancies34 and also with CAR anti-CD73,35 MSLN,36 EGFRvIII,37 and PSMA38 to treat solid tumors (Table 1).

Importantly, two cases of malignant lymphoma derived from CAR gene modified T-cells were described in patients that received CAR-T cells anti-CD19 prepared with PB vector,39 emphasizing the need for regular monitoring of new gene transfer methods used in clinical immunotherapy.

The risk of insertional mutagenesis has driven the development of other innovative gene delivery methodologies such as nanocarrier that delivers in vitro-transcribed (IVT) CAR and nanoplasmid. In vitro transcribed (IVT) mRNA has emerged as a new class of disruptive drugs that can be used to delivery CAR into T-cells. Several ongoing clinical trials are testing the efficiency and safety of engineered CAR mRNA T-cells to treat cancer patients (NCT01355965, NCT01897415, NCT02277522 and NCT02624258), and data suggest transient expression of CAR after infusion of cells is already enough to trigger antitumor responses.40

Nanoplasmids are emerging as a safer alternative to integrating vectors. Plasmid vectors based on S/MAR (scaffold and matrix attachment region) allow plasmid retention in the host cell nucleus, through interactions with nuclear matrix proteins ensuring proper plasmid segregation at mitosis.41 Bozza et al. 2021 demonstrates that this type of vector can be used efficiently to manipulate human T lymphocytes, support long-term transgene expression, can be scaled up, and is suitable for clinical use.42

Innovations to optimize vectors that can be used in CAR-T cell therapy will certainly continue to emerge.

• 4)

  Expansion

The conventional protocols for T-cells activation/expansion ex-vivo include a combination of the three steps for optimal quality of the final product: activation strategy, gene delivery and expansion. The phenotypic and functional profiles of CAR-T cells may vary accordingly to the outline used for expansion as well as patient intrinsic features. Some studies have suggested advantages in the development of CAR-T cells presenting phenotypes such as stem cell memory43 and central memory,44 thus, the processes of expansion of CAR-T cells is critical for the quality of the final product. In fact, the choice of culture outline for expansion represents one of the main factors for T-cell expansion. As reviewed by Vormittag and collaborators,3 the most used culture systems reported in clinical trials may be categorized into three main types: 1) plates or T- flasks, present in 22% of the studies; 2) Static culture bags, found in 35% of the reports; and 3) Rocking motion bioreactor, reported by 43% of the clinical trials. Advantages and disadvantages of each of these culture conditions are listed in Table 1.

Considering the variation in workflows for the obtaining of a satisfactory number of cells for infusion, there is an urgent need of standardization of clinical protocols.

• 5)

  Final formulation/product

The final formulation of products involves critical validation steps prior to cryopreservation. These tests aim to check safety, purity and potency, including T cells viability and CAR expression in the cell surface. The cryopreservation formula might contain 5 to 10% clinical grade DMSO, along with an electrolytic solution, human serum albumin and colloid solution in various combinations. Also, freezing rate should be between one and two Celsius degrees per minute, at least until -40°C, when the product could then be transferred to a liquid nitrogen tank. The CAR-T products should also be released under sterility and purity tests in order to ensure the product's safety. The standard protocols consist in testing for bacterial, fungal and mycoplasma contamination as well as for endotoxins.

A series of limitations have been reported in the use of CAR-T cells therapy. One main concern is the uncontrolled magnitude of CAR-T cells’ activation and expansion after infusion.45,46 In addition, CD19-specific CAR-T cells responding patients also shown B-cell aplasia and, consequently, hypogammaglobulinemia, a complication that usually requires periodic intravenous immunoglobulin administration. To overcome these complications, recent studies have described strategies to turn-on and off CAR-T cells.

Among these approaches, some groups engineered regulatory switches to inhibit or promote the infused cells activation. Fedorov and collaborators47 developed an inhibitory-CAR (iCAR) system by combining the antigen recognition domain with the intracellular inhibitory signaling via the receptors CTLA-4 and PD-1, two molecules involved in T-cell suppression. This strategy leads to a limited T lymphocyte capability, reducing T cell proliferation, cytotoxicity and cytokine production. Another series of studies applied the use of small-molecules drugs able to control the magnitude of CAR-T cells activation.48-50 Recently, Jan and colleagues reported the use of lenalidomide in an inducible on- vs off-switch system for CAR-T cells activation and degradation, respectively. For CAR-T receptor degradation, an IKZF3 zinc finger degron tag was inserted in the intracellular domains from 4-1BB and CD3z, which became accessible for lenalidomide binding. This will lead to ubiquitin ligase recruitment, followed by a polyubiquitination and subsequently transient CAR degradation. Conversely, in the ON-switch outline, authors constructed a two-component CAR: one subunit of the CAR containing the CD28 transmembrane and the intracellular domains composed by the lenalidomide-dependent zinc finger from IKZF3; and the second subunit composed by a CD28 domain, a mutated CRBN and a CD3-zeta signaling domain. In this settings, lenalidomide-dependable dimerization induced about fivefold increasing of CAR-T cell activation.

A series of recent studies have shown the emergence of neoplastic B cells lacking CD19 expression in an important proportion of anti-CD19 CAR-T cells treated patients, suggestive of a scaping mechanism.51,52 Recent strategies reported CAR-T cells carrying bispecific or trispecific TCR structures, able to simultaneously target distinct tumor-antigens. The first phase-1 clinical trial was reported recently by Shah and collaborators53 in the use of bispecific anti-CD19/CD20 CAR-T cells to treat patients presenting B cell non-Hodgkin lymphoma or chronic lymphocytic leukemia. Using CliniMACS Prodigy system (Miltenyi), non-cryopreserved bispecific anti-CD19/CD20 CAR-T cells were tested in 22 patients. Results indicate a high response rate (82%) and a low cytotoxicity rate (5%) of cytokine release syndrome.

Another strategy published recently reported promising in vitro and pre-clinical results by using trispecific CAR-T cells able to target CD19, CD20 and CD22 antigens. Fousek et al.54 reported the construction of a trispecific anti-CD19/CD20/CD22 CAR using a single transgene expressing the three molecules with 4–1BB and CD3ζ endodomains. Results obtained from in vitro and in vivo assays showed that trispecific anti-CD19/CD20/CD22 CAR-T cells presented superior killing activity and tumor control when compared to single anti-CD19 CAR-T cells, even in tumor cell lines lacking CD19 molecule. Accordingly, Schneider and colleagues55 reported a newly engineered trispecific anti-CD19/CD20/CD22 CAR-T using lentiviral vectors by testing multiple intracellular T cell signaling motifs. Trispecific anti-CD19/CD20/CD22 CAR-T cells were able to efficiently eliminate and produce cytokines when incubated to a variety of tumor cell lines expressing distinct levels of CD19, CD20 and CD22. In pre-clinical assays, engineered Raji cells expressing heterogeneous levels of CD19, CD20 and CD22 molecules were completely eliminated in NSG-mice in around fourteen days post-infusion.