CAR-T cells and allogeneic hematopoietic stem cell transplantation for relapsed/refractory B-cell acute lymphoblastic leukemia

The treatment of relapsed/refractory acute lymphoblastic leukemia (ALL) remains a serious challenge because of low remission rates after chemotherapy, high relapse rates and poor long-term survival even with allogeneic hematopoietic stem cell transplantation (allo-HSCT) [1–3]. The options for the treatment of relapsed/refractory ALL are especially limited [4]. Therefore, alternative therapeutic approaches are necessary to improve overall survival (OS), decrease the risk of relapse and/or minimize the toxicity of relapsed/refractory ALL.

CD19 CARs redirected T cells can improve remission rates for patients with B-lineage hematological malignancies [5–11]. A CAR consists of an antigen-binding region of a monoclonal antibody linked to intracellular T-cell signaling domains [12]. T cells expressing CARs (CAR-T cells) can be redirected to recognize and kill the tumors without the restriction of major histocompatibility complexes. CAR-T cells can expand and effectively induce antitumor responses after transfusion into patients. Compared with therapeutic monoclonal antibodies and related approaches, a greater potency and more durable responses were achieved with CAR-T cells [13]. Although, CAR-T cells showed a favorable outcome in the treatment of relapsed/refractory ALL, the persistence is a noteworthy question and some of the cases quickly relapsed after CAR-T cell treatment [14]. Allo-HSCT is still the best way to treat refractory/relapsed ALL [15]. Thus, treatment with CAR-T cells, followed by allo-HSCT, may be the best treatment for relapsed/refractory ALL. In this review, we discuss CAR-T cells as a bridge for the treatment of relapsed/refractory ALL with allo-HSCT. We first discuss the different types of CAR-T cells. We then discuss the treatment of relapsed/refractory ALL using only CAR-T cells. Finally, we discuss the use of CAR-T cells, followed by allo-HSCT, for the treatment of relapsed/refractory ALL.

CAR-T cells

CARs typically consist of three parts, including an extracellular major histocompatibility antigens-independent antigen-binding domain that is usually derived from a single-chain variable fragment of a monoclonal antibody, a transmembrane-linking domain and an intracellular costimulatory T-cell signaling domain or multiple domains [16]. Some CARs also include an extracellular spacer domain (Figure 1).

There have been four generations of CAR-T cells devised thus far (Figure 1). A CD3ζ intracellular signaling domain (the ‘first signal’) was utilized to activate T cells in the first generation of CARs; however, these CARs did not produce enough IL-2 and had limited activation, expansion and in vivo antitumor effects [17]. The second generation of CARs was then produced through the addition of costimulatory molecules with intracellular costimulatory signaling domains, such as CD28, CD137(4–1BB), CD134(OX40), CD27, ICOS and CD244. The second generation of CARs produced 20-times more IL-2 compared with the first generation and showed continued proliferation and killing of target cells in vivo [18–21]. The third generation of CARs was produced by adding two or more costimulatory molecules with intracellular costimulatory signaling domains. Thus, more cytokines were secreted, and killing ability was enhanced in vitro; however, better outcomes were not achieved in the clinic [22,23]. The fourth generation of CARs, knowns as T-cell redirected for universal cytokine killing, or TRUCK, were produced by adding a sequence of cytokines, such as IL-12, which can kill the tumors through releasing cytokines locally to regulate the microenvironment and recruit immune effector cells. TRUCKs can be used to treat virus infections, auto-immune diseases or metabolic disorders but not cancer [24].

Single CAR-T cells for refractory/relapsed ALL

Treatment of relapsed/refractory CD19⁺ B-cell malignancies in children, adolescents and young adults using CD19-CAR-T cells was performed in Phase I clinical trial. Retrovirally transduced CD19-CAR T cells with CD3ζ and CD28 costimulatory domains (CD19-CAR.z.28) were used. All of the patients received chemotherapy with fludarabine and cyclophosphamide prior to CD19-CAR.z.28 T-cell infusion. In total, 66.7% (14/21) achieved complication remission (CR). For the 20 patients with B-ALL, the CR rate was 70%. Molecular remissions were achieved in a total of 12 out of the 20 patients (60%) with no flow cytometric evidence of minimal residual disease (MRD < 0.01%). Two patients who underwent CAR-T cell treatment relapsed with CD19-negative B-ALL at 3 and 5 months post-treatment [25]. Davila et al. used 19–28z CAR-T cells to treat 16 adults with relapsed/refractory B-cell ALL at the Memorial Sloan-Kettering Cancer Center at a dose of 3 × 10⁶ CAR T cells/kg. The results showed that 88% of patients reached a CR or incomplete count recovery, and 75% of the treated patients achieved MRD-negative disease status [26]. Nearly, all of the patients had a peak number of CAR-T cells within 1–2 weeks after the infusion, and these numbers decreased to low or undetectable levels by 2–3 months. At the National Cancer Institute/National Institutes of Health, CAR-T cells containing the CD28 and CD3z domains were used to treat relapsed/refractory ALL at a dose of 1 × 10⁶/kg [25]. A variably intensive lymphodepletion prior to the treatment with CAR-T cells was used for the 21 pediatric and adult patients with either low- or high-burden disease. Fifteen of these patients reached CR.

Lentivirally-transduced CD19-CAR-T cells with CD3ζ and 4–1BB (CD137) endodomains (CD19-CAR-CD3ζ-41BB) were also used to treat relapsed/refractory B-cell malignancies. Two children with relapsed/refractory ALL were treated with CAR-T cells at a dose of 1.4 × 10⁶ to 1.2 × 10⁷ per kilogram of body weight. Both patients showed CR, but while one patient was still in CR 11 months after treatment, the other relapsed within two months [27]. Maude et al. also treated 30 patients, including 18 relapsed patients after hematopoietic stem cell transplantation (HSCT), with CD19-CAR-CD3ζ-41BB cells. Out of the 30 patients, 27 (90%) achieved morphologic CR, and 22 out of the 27 also achieved molecular CR (MRD < 0.01%) 1 month after CAR-T cell treatment [28]. The 1-year disease-free survival (DFS) rate was 44%, and patients with long-term persistence of CTL019 cells experienced the longest DFS even without subsequent HSCT. However, 20 patients relapsed, and 13 of them had CD19-negative blasts [28,29]. Thirty patients with relapsed/refractory ALL were treated with CD19-CAR-CD3ζ-CD28–41BB cells. All of the patients (100%) survived for more than 21 days and had no morphologically detectable leukemia in the BM. In 27 out of 29 patients (93%), leukemia was undetectable by high-resolution flow cytometry [30].

Allogeneic CD19-CAR-T cells have been used to treat patients with relapsed ALL after allo-HSCT [31]. In a Phase I clinical trial, the feasibility and safety of donor-derived CD19-CAR-T cells were evaluated in four patients relapsing after HSCT, six ALL patients (2 Ph⁺) and two patients at a high risk due to MRD⁺ [32]. The infusion was well-tolerated, and no patient developed graft-versus-host disease (GVHD). Six patients had CR lasting from 2 to 8 months. An 11-year-old girl with relapsed ALL after allo-HSCT was treated with CAR-T cells. After an infusion of 1 × 10⁶/kg CAR-T cells, CAR-T-cell infusions (0.83–1.65) × 10⁶/kg were administered three-times for maintenance. The DFS lasted for 10 months [33].

CD19-CAR-CD3ζ-41BB cells were also used to treat nine adult patients with B-ALL and extramedullary leukemia. Six of the nine patients with definite extramedullary involvement achieved 56% 18-weeks OS. One of the two patients with partial regression of extramedullary lesions who received conditioning chemotherapy achieved a three-month-durable complete response. Four of seven patients without conditioning chemotherapy achieved dramatic regression or a mixed response in the hematopoietic system and extramedullary tissues for 2–9 months [34]. CAR-T cells are also effective against CNS leukemia. Gene marking studies revealed the presence and persistence of CAR-T cells in the cerebrospinal fluid of 46 out of 47 patients without an increased risk of encephalopathy, and four patients with CNS leukemia reached remission after CAR-T-cell treatment [29].

In a Phase II multicenter clinical trial in China, 50 refractory/relapsed ALL patients were treated with the CD19scFv/CD28/CD137/CD27/CD3ζ-iCasp9 (4SCAR-19) lentivector. Out of the patients with blasts < 50% (group 1), 94.3% reached CR and of the patients with blast >50% (group 2), 66.7% (10/15) reached CR [35]. The 120-days LFS for patients in group 1 and group 2 were 86 and 44.4%, respectively. The rates of 10-month OS for patients in group 1 and group 2 were 82 and 36%, respectively.

Recently, a meta-analysis investigated the efficacy of CD19-CAR-T cells in refractory B-cell malignances in Phase I clinical trials. Data from 1991 to 2014 were collected from PubMed and the Web of Science. A response rate evaluation was analyzed from fourteen clinical trials and included data from 119 patients. Progression-free survival (PFS) analysis was performed on data from 62 patients in 12 clinical trials. The overall pooled response rate of CD19-CAR-T cells was 73%. ALL patients had higher response rates than chronic lymphocytic leukemia and lymphoma patients. Lymphodepletion and the absence of IL-2-administrating T cells were two key factors associated with a better clinical response. The 6-month and 1-year PFS rates for the 62 patients were 80.0 and 76.3%, respectively. The median interval of PFS was 7.0 months. Higher numbers of infused CAR-T cells and lymphodepletion were associated with a better prognosis [36].

Altogether, approximately 40–90% of patients with refractory/relapsed ALL can reach complete molecular or cellular remission through treatment with CAR-T cells; however, because CAR-T cells survived in the body for only a short time, some of the patients relapsed quickly, which may be related to the different in the costimulatory molecular or vectors used, or variation in the disease burden across treated individuals [36,37] (Table 1). Thus, other treatments should be performed to improve outcomes in the treatment of refractory/relapsed ALL.

CAR-T cells followed by allo-HSCT for refractory/relapsed ALL

At present, allo-HSCT is still the best way to treat refractory/relapsed ALL [15]. However, allo-HCT is largely ineffective for refractory/relapsed ALL patients at the time of transplantation. Although, the initial outcomes are inspiring for refractory/relapsed ALL treated with CAR-T cells, some of the cases relapsed within 3–6 months, which represents a serious obstacle for CAR-T cells treatment that must be overcome if it is to be widely adopted [38]. We suggest that the best way to treat refractory/relapsed ALL may be to perform CAR-T-cell treatment followed by allo-HSCT to improve survival rates for patients.

Lee et al. reported on 10 patients with refractory/relapsed ALL at the NIH who achieved MRD-negative disease after CAR-T cells infusion and then underwent allo-HSCT while in CAR-induce MRD-negative CR. These patients remain disease-free with no unexpected peritransplant toxicities and no subsequent relapses [25]. However, in the same study, two patients who achieved MRD-negative CR were judged ineligible by their treating physicians for allo-HSCT, and both relapsed with CD19-negative leukemia at 3 and 5 months, respectively.

Brentjens et al. reported four patients that were treated with CAR-T cells, followed by allo-HSCT [7]. One patient died of a suspected pulmonary embolus 2 months after the allo-HSCT while in CR, with no evidence of any disease. The other three patients had no significant complications and remained in MRD-negative CR from 3 to 6 months after allo-HSCT. However, one patient who was ineligible for allo-HSCT relapsed 3 months after CAR-T-cell treatment. Maude et al. reported that three patients in remission were treated with CAR-T cells followed by allo-HSCT, and they remained in remission 7–12 months after CAR-T-cell treatment [30].

Davila et al. reported that no relapses were happened for seven of the 16 (44%) refractory/relapsed ALL patients treated with CAR-T cells followed by allo-HSCT therapy [26]. In another study, adults with refractory/relapsed ALL were treated with CD19-CAR-T cells. After achieving a CR, 13 of 37 patients underwent allo-HSCT. No significant difference was found between patients who underwent allo-HSCT and those who did not (79 vs 80%) in terms 6-month OS [39]. However, it is difficult to estimate whether there was a greater risk of relapse for the patients who underwent consolidative allo-HSCT because it was not a randomized trial. Recently, the absence of post-treatment MRD was used as a predictive marker of survival and the effect of allo-HSCT after CAR-T-cell therapy was studied in refractory/relapsed ALL patients [39]. The OS rate at 6 months was 76% in the MRD⁻ CR cohort and 14% in the MRD⁺ CR cohort. Further, post-treatment MRD status emerged as a strong predictive marker of OS. Further analysis showed that the OS rates at 6 months were 70% in patients who underwent post-CAR allo-HSCT and 64% in patients who did not get allo-HSCT after CAR-T cells therapy; this difference was statistically insignificant, but suggestive. The follow-up time may have been too short and the number of cases too small to detect a statistically significant difference.

Although, other studies have treated refractory/relapsed ALL patients with CAR-T cells followed by allo-HSCT, the number of cases in these studies were too few to draw definitive conclusions, and the authors did not report the outcomes of CAR-T treatment with and without allo-HSCT separately. The question of whether consolidation allo-HSCT is required to maintain remission induced by CAR-T cells in ALL remains largely unanswered, but can and should be tested in clinical trials. Given the exciting clinical outcomes reported for the treatment of refractory/relapsed ALL with CAR-T cells followed by allo-HSCT in the phase I setting, a Phase II trial (ChiCTR-OOC-16008448) already underway at our center will formally assess the efficacy of this experimental treatment in appropriately powered cohorts of patients with relapsed/refractory B-ALL. However, it is very important to identify some key factors in treatment success that will help to prevent recurrence for those patients who really need an allograft as consolidation therapy.

Toxicities

Cytokine-release syndrome (CRS), neurotoxicity and B-cell aplasia are the major toxic effects for CAR-T-cell treatment [40]. The most common and potentially severe toxic effect is CRS [18,32]. The risk of severe CRS may be related to the tumor burden [23]. The level of sIL2Ra, IFN-γ, IL-10 and IL-6 are remarkably elevated in CRS and are accompanied by the mild flu-like symptoms to, in the most severe cases, multisystem organ failure and shock [2,18]. More attention should be paid to the rates of severe CRS associated with CARs with different costimulatory molecules. More serious CRS appears to occur with the 4–1BB CAR-T cells compare with other costimulatory molecules [41–43].

The most common neurologic side effect associated with CAR-T treatment is global encephalopathy, which may be brief and self-limited, but can lead to death [4]. Chronic B-cell aplasia and resultant hypogammaglobulinemia can also occur. Pneumonitis, fever, tumor lysis syndrome, hypotension, hypoxia, tachycardia and fatigue are also associated with CAR-T-cell therapy. However, most of these complications can be controlled.

The additional cost for CAR-T therapy is also high, which may prevent its more widespread use in the near future. Allo-HSCT is good option for some patients because the morbidity and mortality from allo-HSCT is not insignificant, however, we must pay attention to the potential GVHD complications [44].

Conclusion & future perspective

CAR-T-cell therapy represents the beginning of a new era of using engineered T cells for cancer therapy. Although, it has shown its impressive effect in clinical studies, many obstacles preventing its broader application, including the complicated genetic technology required for manufacturing CAR-T cells, tumor escape, inadequate toxicity management, relatively fast relapse of the disease, a lack of persistence and the need to standardize CAR-T-cell treatment [45].

The technology for manufacturing genetically modified T cells is complex and patient-specific. It will take time to build the cell-processing infrastructure to deliver this therapy to large numbers of patients outside major academic centers. However, the US FDA has approved the Novartis biologics license application for CART-19 in pediatric ALL. The Kite's BLA for CART-19 in diffuse large cell B lymphoma is also under consideration. The choice of effector T cells and CAR design must be optimized to improve effectiveness and limit toxicities [46]. It is difficult to compare across clinical trials because of the differences in conditioning chemotherapy, CAR constructs, CAR-T-cell production methods, toxicity management and the size of the tumor burdens.

Another important aspect of CAR-T generation is the means of delivering the foreign gene into human cells. For CAR-T cells, viral system is the most commonly used tool for achieving this. Retroviruses (including lentivirus), adenovirus and adeno-associated virus are the main virus vectors. Among these, engineered retroviruses are the most popular tools for gene delivery [27]. But the insertion mutation used to induce the immune reaction can lead to tumorigenesis and toxicity, the carrier capacity is limited and the titer achieved is not high enough [47,48]. The safest vector that has been developed thus far may be the third-generation minimal lentiviral vector [49]. Thus, perhaps CAR-T cell treatment should be standardized to use this vector [50]. Formulating strategies that can improve cell isolation and culture methods and allow large-scale automation of CAR-modified T-cell production should also be considered, as this could facilitate their utilization in the clinic and reduce the cost of engineered T-cell therapies [24,51]. The accessibility and survival issues of CAR-T cells should also be resolved in the future.

Multispecific CARs could be used to avoid tumor escape. The use of CARs whose activity can be tuned by administration of small molecule pharmaceuticals or incorporation of suicide genes might allow the control of adverse events. T cells that are resistant to hostile microenvironments or have enhanced persistence, such as those from the bone marrow, could be produced [52,53]. The combination of CAR-T cells with other immunotherapeutics, such as oncolytic viruses or checkpoint blockade antibodies, should also be explored, as this can increase efficacy [54].

The management of toxicity is the central question for CAR-T-cell treatment. How toxicities could be solved without interfering with CAR-T-cell expansion, and whether conditioning chemotherapy is necessary should also be determined [13,55]. Many CAR-T-cell therapies used on patients with B-ALL have consisted of CD4⁺ and CD8⁺ T cells at a fixed 1:1 ratio [56]. It is unknown whether the ratio of T-cell subset is important in CAR-T cell therapy, so further study is needed to explore the optimum subset ratio [57]. Another major challenge will be to identify unique tumor antigens that can be targeted with selective T-cell therapy. Preclinical and early-phase clinical studies have shown that targeting multiple antigens in CAR-T-cell therapy can prevent antigen loss relapses. For example, CAR-T cells targeting CD19/CD20, CD19/CD22, or CD19/CD123, can also improve the extremely limited options and poor outcomes of patients with CD19-negative relapses [58–60]. Attention should also be paid to the uncontrolled proliferation of CAR-T-cell therapies. In order to specifically deplete CAR-T cells in a controlled manner, a suicide switch could be incorporated into the CAR construct, although this may also weaken persistence and therefore the ability of CAR-T cells to fight disease long-term [61–64]. Ultimately, cellular therapies should be integrated into the routine practice and standard of care for ALL.

CAR-T cells with prolonged persistence may prevent relapses through long-term immunosurveillance. CAR-T cells generated using a lentiviral vector and 4–1BB costimulation seems to be more persistent than those made using a retroviral vector and CD28, however, there have been no direct comparisons [64]. Some experiments showed that the CAR-T cells with CD28 molecules can cause constitutive stimulation, proliferation and growth [65]. However, the 4–1BB molecules can induce early exhaustion of CAR-T cells, which may limit antitumor efficacy [66,67]. The high short-term potency in some situations may give rise to deep remissions, which may make up for the lack of long-term persistence [26]. Thus, presently the optimal persistence time remains unclear, although it seems that the persistence of CAR-T-cell therapy is critical.

Donor lymphocyte infusion showed limited efficacy in producing a long-lasting graft-versus-leukemia effect or preventing a life-threatening GVHD associated with the infusion of a high number of T cells in ALL for the prevention of relapse or the treatment of relapse after allo-HSCT [68,69]. Patients treated with allogeneic CD19-CAR-T cells used for B-cell malignancies that progressed after allo-HSCT received a single infusion of CAR-T cells obtained from each recipient's allo-HSCT donor. No chemotherapy or other therapies were administered for these patients. eight out of 20 patients obtained remission, with six obtaining CR and two obtaining partial remissions. The response rate was highest for ALL, with four of five patients obtaining MRD-negative CR. No GVHD was observed [70]. Donor-derived allogeneic anti-CD19-CAR-T cells can cause regression of B-cell malignancies resistant to standard donor lymphocyte infusions without causing GVHD [9]. It is therefore a safe and feasible approach for patients to receive donor-derived CAR-T cell therapy for relapse after allo-HSCT [71,72]. Thus, early infusion of CAR-T cells could be used for the prevention of relapse for high-risk cases or when molecular relapse is detected. Presently, our center is performing a clinical trial (ChiCTR-OOC-16008447) in which CAR-T cells are used to prevent the relapse of B-ALL after allo-HSCT when molecular relapse is detected.

Allogeneic CAR-T cells are beneficial for patients with relapsed B-cell malignancies after allo-HSCT, and have low toxicities and complications [68–73]. Recently, coinfusion of haploidentical donor-derived CD19- CAR-T cells and mobilized peripheral blood stem cells following induction chemotherapy was performed in a 71-year-old female with relapsed and refractory ALL. The patient achieved CR with undetectable MRD and full donor cell engraftment was observed with acute GVHD [74]. Therefore, CAR-T cells may be used as a part of conditioning regimen in allo-HSCT for refractory ALL. However, the time and numbers of CAR-T-cell infusion should be researched in this model.

Altogether, CAR-T cells are an extremely promising area of research that holds endless potential for future treatment of ALL. However, we have only seen the tip of the iceberg, and many crucial questions have yet to be answered.