Currently, there are four different CD19-targeting, and one BCMA-targeting Chimeric antigen receptor (CAR)-T products, approved by American and European Drug Agencies (Food and Drugs Administration and European Medicines Agency, respectively). All these products are used for the treatment of malignancies derived from abnormal B cells at different stages of their differentiation.

The general structure of a CAR construct includes four parts: (1) an extracellular antigen-binding domain, (2) a hinge, (3) a transmembrane helix, and (4) an intracellular signaling domain. Accordingly, decades of research in the area of CAR design have led to the development of five CAR-T cell generations, each of them differing in an intracellular signaling milieu composition. In fact, the first CAR-T generation contains only a single immunoreceptor tyrosine-based activation motif (ITAM) domain derived from a CD3ζ chain of the T cell receptor (TCR). Further approaches have involved adding one (second generation) or two (third generation) co-stimulatory domains (mostly CD28 and/or 4-1BB) that enhance CAR-T cells activation and persistence ^[1][2][3][8,9,10]; however, in order to additionally boost the CAR-T cells activation, proliferation, and killing potential, fourth and fifth CAR-T generations were designed ^[4][11]. Such CAR-T cells are armored with additional intracellular domains responsible for specific cytokines production (e.g., IL-12, IL-2). These CAR-T cells are currently being evaluated in the research phase.

CAR-T cells products used for therapy of hematological patients represent the second CAR-T cell generation, with either a CD28 or 4-1BB (CD137) co-stimulatory domain. Since 2017, four CAR-T cell products aimed to target CD19 were approved based on pivotal phase 1/2 clinical trials, all of which are still active. Initial clinical responses for each trial and appropriate follow-up studies results (with observations lasting longer than 2 years) are summarized in Table 1.

Out of these four products only tisagenlecleucel (Kymriah) is registered for children and young adults (up to 25 years old) with B-ALL, who are refractory to previous treatment regimens, or with more than two relapses. Considering that it was the first CAR-T cell therapy approved, there is the longest follow-up available for this product (median 4.8 years), showing complete response (CR) in more than 60% of the patients and overall survival (OS) equal to 10.5 months ^[6][13]. On the other hand, the access to CD19-targeted CAR-T therapy for adult r/r B-ALL patients was not granted until recently, when in October 2021 the FDA approved brexucabtagene autoleucel (Tecartus) for such indication. The registration was based on the multicenter ZUMA-3 trial ^[13][18][20,25]. The efficacy estimation group in the phase 2 ZUMA-3 trial included 55 adult previously treated patients, of which 27% represented a high-risk group, bearing BCR-ABL1 translocation ^[13][20]. CR was achieved in 56% of the patients, and more importantly, CR or complete remission with incomplete hematological recovery (CRi) was reached in 60% of the patients previously submitted to blinatumomab, another CD19-targeted immunotherapy. The side effects post CAR-T infusions were generally manageable. Taking into consideration the combined results of phases 1 and 2, the median duration of remission in this trial was estimated by the investigators at 13.4 months; however, a longer follow-up is needed to draw stronger conclusions regarding safety profiles and the durability of responses.

In addition to adult B-ALL, brexucabtagene autoleucel is also approved for the treatment of mantle cell lymphoma (MCL). The results gathered during ZUMA-2 trial showed encouraging responses in heavily pre-treated r/r MCL patients, with CR achieved in 67% after single CAR-T administration ^[12][19]. Common adverse events following CAR-T therapy occurred in the majority of treated patients, however, they were mostly manageable. A one-year follow-up of the ZUMA-2 trial presented at the 2020 ASH Annual Meeting showed that with a median of 17.5 month-long observations, the CR was maintained in 67% of the patients ^[19][26].

For the treatment of B cell lymphomas, several CAR-T products are available. Based on JULIET ^[7][14] and ZUMA-1 ^[9][16] trials, tisagenlecleucel (Kymriah) and axicabtagene ciloleucel (Yescarta), respectively, were first to be approved for the treatment of various types of B cell lymphoma, including adult r/r DLBCL, DLBCL arising from follicular lymphoma (FL), or primary mediastinal large B cell lymphoma (PMBCL), for the patients who failed after two or more lines of systemic therapy (see Table 1 for more details). Importantly, long-term follow-up of patients’ status within these trials (median follow-up exceeding 3 years) provided data showing a superior efficacy of Yescarta, with median OS reaching 25.8 months ^[10][17] as compared to 11.1 months for Kymriah ^[8][15]. This discrepancy, however, could arise from the fact that most of the patients in the JULIET study received a bridging therapy (e.g., rituximab, nucleoside analogues, corticosteroids), whereas in the ZUMA-1 trial bridging chemotherapy was not allowed, with the exception of corticosteroids administration. Thus, one can speculate that the JULIET trial included more patients with aggressive lymphomas; however, the direct impact of bridging therapies on CAR-T responses should not be excluded.

The year 2021 was extremely rich in FDA/EMA approvals for CAR-T therapies. In addition to the aforementioned Tecartus approval for r/r adult B-ALL, the additional registrations included axicabtagene ciloleucel (Yescarta) for r/r adult FL ^[11][18], and lisocabtagene maraleucel (Breyanzi) for adult r/r DLBCL, r/r PMBCL, grade 3B FL ^[14][21]. Moreover, in 2021, it wase witnessed that the first approval of a BCMA-directed CAR-T therapy designated for adult r/r MM patients who had failed at least four previous lines of treatment ^[17][24]. Indeed, the efficacy of idecabtagene vicleucel (Abecma) was evaluated in the pivotal KarMMa clinical trial ^[16][23]. Stringent CR (sCR) was detected in only 28%, however, the CAR-T cells were administered to patients with a significant number of prior lines of treatment (median of 6 lines) ^[17][24]. Importantly, 65% of the patients who achieved sCR had a durable response for at least 12 months. All in all, CAR-T cells targeting BCMA seem to be an attractive treatment option for heavily pre-treated, relapsed MM patients. Certainly, a longer follow-up is still required to estimate the safety profiles and survival rates in a long run.

Despite the phenomenal success of CD19 CAR-T therapy, it is not without limitations and side effects. The challenges faced by CAR-T cell therapy are related to both effector and target cells and include: (1) proliferation, persistence, and exhaustion of CAR-T cells, (2) side effects, (3) immunosuppressive tumor microenvironment, and (4) tumor intrinsic factors that account for the development of resistance.

The obstacles related to effector CAR-T cells are elegantly described by others ^{[20][21][22][23][24]}[27,28,29,30,31]. Therefore, below it wase only shortly presented that currently investigated and implemented solutions to modulate insufficient CAR-T cells persistence in vivo, and exhaustion of the effector cells. Considering, however, that more and more data is now gathered on tumor related features contributing to immunotherapy resistance, here itin this review we will mostly focus on limitations and solutions concerning target cells.

The sufficient number of viable and functional CAR-T cells is one of the major contributors to the therapy’s success ^[9][25][16,32]. The poor persistence even after the achievement of complete remission is considered to be the main reason for the development of CD19-positive relapses ^[26][27][33,34]. One of the features affecting CAR-T cell persistence and proneness to exhaustion is the co-stimulatory domain incorporated in the CAR construct. CD28 and 4-1BB are the most frequently exploited co-stimulatory domains in pre-clinically tested and in clinically available CAR products (Table 1). ISeveral st has beenudies have demonstrated that a higher proliferative capacity and longer in vivo persistence of CAR-T cells bearing a 4-1BB domain as compared to CD28 ^[28][29][35,36]. This effect was at least partly attributed to the 4-1BB-mediated activation of non-canonical nuclear factor κB (ncNF-κB) signaling in CAR-T cells ^[30][37]; however, as CARs with a CD28 domain exhibit more rapid tumor elimination ^[31][38], attempts to improve their in vivo performance have led to a generation of mutant CD28 endodomain ^[32][39], or third generation CARs containing both CD28 and 4-1BB sequences ^[33][40]. The latter were demonstrated to retain the beneficial aspects of both domains, showing enhanced antitumor activity and concomitantly increased in vivo persistence.

In the clinical settings, in order to improve CAR-T cells persistence and thus the overall therapy efficacy, lymphodepleting chemotherapy is performed prior to CAR-T cells infusion. The pre-conditioning is most often based on fludarabine and/or cyclophosphamide administration. The reason standing behind this approach was demonstrated in several studies showing higher response rates in patients with r/r B-ALL ^[34][35][41,42], and non-Hodgkin lymphomas, including r/r DLBCL ^[36][37][38][43,44,45], who received lymphodepleting chemotherapy regimens prior to CD19 CAR-T cells. Primarily, lymphodepletion reduces the numbers of endogenous lymphocytes, which normally absorb the cytokines that stimulate T cells proliferation and thus limit their availability for infused CAR-T cells ^[37][39][44,46]. In addition, it partially eliminates T regulatory cells (Tregs) which exhibit immunosuppressive activity towards tumor-specific cytotoxic T cells ^[40][47]. The beneficial effect of lymphodepleting regimens was demonstrated for both CD8-positive and CD4-positive CAR-T cell populations ^[41][48].

However, the intensity of existing pre-conditioning therapy using fludarabine and cyclophosphamide was shown to be correlated with higher toxicity of the CAR-T cell-based therapy, thus it may exclude the application of this approach for numerous patients ^[42][49]. The approach of targeted lymphodepletion with a CD45-directed antibody radioconjugate, which was recently shown to eliminate several subsets of leukocytes while preserving progenitor hematopoietic cells in a murine model, could potentially be proposed as a safer alternative ^[43][50]. However, regardless of the lymphodepleting protocol used, the inevitable side effects of these regimens are lymphopenia and prolonged T cells dysfunction, which may lead not only to overall higher sensitivity to infections and autoimmunity development, but also to tumor relapse ^[44][51].

Some already existing approaches undermine the need for lymphodepletion. The most prominent is based on the construction of CAR-T cells that besides having the ability to recognize specific targets can also produce selected proteins, in particular cytokines ^[45][52]. These 4th generation CAR-T cells are called TRUCKs (T cells redirected for antigen-unrestricted cytokine-initiated killing). Cytokines are crucial for CAR-T cells’ survival, proliferation and persistence, both during in vitro culture and following the administration into the patient ^[46][53]. The additional modification of CD19 CAR-T cells to secrete IL-12 was shown to improve their in vivo cytotoxicity towards murine cells expressing human CD19 and alleviate their suppression by Tregs—the main two aims of pre-conditioning with chemotherapy ^[47][54]. Similar observations were made by Kueberuwa et al., who demonstrated that CD19 CAR-T cells expressing IL-12 effectively eliminated lymphoma cells in fully lymphorepleted mice ^[48][55]. Besides the killing of CD19-positive target cells directly, the TRUCKs also contributed to the induction of an anticancer response elicited by the host immune cells, primarily CD8-positive T cells.

Except for IL-12, other cytokines were also shown to improve the persistence and effector functions of transferred cytotoxic T cells. In particular, IL-2, IL-7, and IL-15 supplementation during the culture of CAR-T cells has beneficial effects ^[49][50][56,57]. Interestingly, CAR-T cells expanded with IL-15 exhibited a higher anticancer efficacy as compared to IL-2-supplemented CD19 CAR-T cells against a lymphoma cell line in vivo ^[51][58]. In addition, IL-15-expressing CAR-T cells appear to be less terminally differentiated and have an increased ability to expand in vitro, but also tend to be more toxic in vivo due to the simultaneously higher production of TNF-α and IL-2, as compared to CAR-T cells without IL-15 ^[52][59]. On the other hand, constitutive cytokine expression poses a risk for the uncontrolled growth of administered CAR-T cells which may lead to overt toxicity. Several research groups have demonstrated that novel CAR constructs may be engineered to induce cytokine signaling only after antigen stimulation (5th CAR-T generation) and preserve their superior antitumor effects while having minimal toxicity ^[53][54][55][60,61,62]. The additional incorporation of an inducible caspase-9-based suicide gene, which mediates the selective depletion of CAR-T cells in case of toxicity, may increase the safety while keeping all the advantages of the therapy based on using anti-CD19 TRUCKs ^[56][57][58][63,64,65].

The effectiveness of CAR-T cells is also impeded by their exhaustion, which is often caused by the enhanced expression of immune checkpoint molecules on cancer cells and CAR-T cells. In particular, PD-1/PD-L1 interaction activates downstream signaling pathways and inhibits T cell activation, which leads to tumor immune escape ^[59][66]. PD-L1 was reported to be overexpressed in classical Hodgkin lymphoma ^[60][61][67,68] and DLBCL ^[62][69]. Moreover, PD-1 expression was observed to be increased on T cells isolated from B-ALL patients’ bone marrow aspirates ^[63][70]. Similarly, CD19 CAR-T cells were also reported to have increased PD-1 expression following the infusion to the patients ^[64][71]. The solution for this issue can be achieved by combining a CAR-based therapy with anti-PD-1 antibodies ^[65][66][72,73], or their administration just after CAR-T failure ^[67][74]. The more advanced and safe approaches include the delivery of CAR-T cells secreting anti-PD-1 scFv ^[68][75], or the transformation of PD-1 in CAR-T cells into a co-stimulatory molecule using the switch-receptor technology ^[69][76].

Recent reports have also shown that CAR-T cells exhaustion occurs due to epigenetic reprogramming ^[70][71][72][77,78,79]. In a clinically-relevant setting, Zebley et al. observed the changes in DNA methylation patterns in CD8-positive CD19 CAR-T cells that were previously administered to B-ALL patients ^[72][79]. Importantly, a deletion of a gene coding DNA methyltransferase 3 alpha (DNMT3A) in T cells bearing CAR constructs augmented a long-term antitumor response and prevented them from developing an exhausted phenotype ^[70][77]. In addition, several other modifications show promise to limit CAR-T cells exhaustion. For instance, CAR-T cells engineered to overexpress a transcription factor c-Jun had an improved ability to resist exhaustion and maintain functionality despite prolonged activation ^[73][80].

There are also other factors contributing to the loss of CAR-T cells effector abilities, including those related to CAR tonic signaling ^[74][81]. One of the consequences of CAR tonic signaling is terminal differentiation to effector cells. The available literature indicates the involvement of sustained phosphoinositide 3-kinase (PI3K)/Akt pathway activation in this process ^[75][76][82,83]. Therefore, pharmacological approaches aimed at PI3K/Akt signaling inhibition have been implemented during the manufacturing phase to maintain CAR-T cells stemness, including the addition of tyrosine kinase inhibitors, e.g., idelalisib ^[77][84], ibrutinib ^[78][85], or duvelisib ^[79][86]. In addition, the modifications of CAR constructs may be performed to address the issue of CAR tonic signaling and involve several aspects, e.g., the substitution of co-stimulatory domains, alterations in the spacer and hinge domain, or the improvement of scFv stability ^[74][81]. An alternative approach is to reduce CAR surface expression and selectively switch-off CAR signaling. This could be achieved by incorporating the ligand-induced degradation/destabilizing domain into the C-terminus of the CAR construct ^[80][81][87,88]. In addition, Weber et al. modulated CAR-T cells exhaustion by the addition of dasatinib, a tyrosine kinase inhibitor that reversed the tonic signaling and led to downregulation of the immune checkpoint markers ^[81][88]. Importantly, dasatinib administration inhibits the secretion of cytokines by CAR-T cells and impairs their in vivo antitumor activity ^[82][89]. This effect, however, is reversible after the drug discontinuation, which implies that short-term dasatinib treatment could be used as a switch to control CAR-T cells’ performance.