Genome-Edited, Donor-derived Allogeneic Anti-CD19 Chimeric Antigen Receptor T-cells in Paediatric and Adult B-Acute Lymphoblastic Leukaemia: Results of Two Phase I Studies

Abstract

Background:

Genome-edited donor-derived allogeneic anti-CD19 chimeric antigen receptor (CAR) T-cells offer an ‘off-the-shelf’ CAR T-cell product, broadening access and applicability. UCART19 is an innovative genome-edited donor-derived allogeneic CAR T-cell product investigated in children and adults with relapsed or refractory B-cell acute lymphoblastic leukaemia (ALL).

Methods:

We enrolled paediatric or adult patients in two on-going phase I clinical trials to evaluate the safety and antileukaemic activity of UCART19. Patients underwent lymphodepletion with fludarabine and cyclophosphamide ± alemtuzumab, then children received UCART19 at 1·1 to 2·3×10⁶ cells/kg and adults received UCART19 doses of 6×10⁶, 6–8×10⁷ or 1·8–2·4×10⁸ cells in a dose-escalation study. These studies were registered at Clinicaltrials.gov (NCT02808442 and NCT02746952).

Results:

7 children and 14 adults were enrolled in the two studies and received UCART19. Cytokine release syndrome (CRS) was the most common adverse event observed in 19 patients (91%); three (14%) had grade 3–4 CRS. Other adverse events were grade 1 or 2 neurotoxicity in eight patients (38%), grade 1 acute skin graft-versus-host disease in two patients (10%) and grade 4 prolonged cytopenia in six patients (29%). Two treatment-related deaths occurred; one caused by neutropenic sepsis in a patient with concurrent CRS and one from pulmonary haemorrhage in a patient with persistent cytopenia. The rate of complete remission or complete remission with incomplete hematologic recovery at day 28 post-infusion was 67%. Patients not receiving alemtuzumab (n=4 patients) showed no UCART19 expansion or antileukaemic activity.

Interpretation:

UCART19 exhibited in vivo expansion and anti-leukaemic activity with a manageable safety profile in heavily pretreated paediatric and adult patients with relapsed or refractory B-cell ALL.

Funding:

Servier

Introduction

Autologous anti-CD19 CAR T-cells have proven to be highly efficacious in relapsed/refractory CD19-positive B-cell acute lymphoblastic leukaemia (ALL), with ~70–90% response rates and durable remissions.^1–5 This led to approval of tisagenlecleucel in the USA and Europe for B-cell ALL in children and young adults (up to age 25 years). However, the process of autologous CAR T-cell manufacturing is time consuming and expensive. Autologous harvests usually have to be collected and transported to distant centralised manufacturing sites, often requiring complex logistics across time zones and international borders. Furthermore, for heavily pretreated lymphopenic patients, infants, and those with highly aggressive disease, it may not always be possible to manufacture an autologous CAR T-cell product. Children under 3 years of age were excluded from the registration study of tisagenlecleucel in B-cell ALL, so there are no data on the use of this strategy in the high-risk group of infants with B-cell ALL and mixed-lineage-leukaemia gene rearrangement. Additionally, in the phase 2 study with tisagenlecleucel, 18·5% of patients who underwent successful apheresis could not receive CAR T-cell therapy due to disease-related complications or manufacturing issues.³

Therefore, an ‘off-the-shelf’ allogeneic CAR T-cell product would be an attractive option for patients with relapsed/refractory B-cell ALL, but several obstacles need to be overcome to allow allogeneic CAR T-cells to safely expand and be active in a non-human leukocyte antigen (HLA)-matched setting. These include the potential for graft-versus-host disease (GVHD) or rejection of HLA-incompatible cells. Strategies to overcome these barriers focus on disrupting genes encoding the T-cell receptor (TCR) αβ in the infused CAR T-cell product to prevent GVHD, and conditioning the recipient’s immune system to minimize CAR T-cell rejection while engineering the infused cells to be resistant to lymphodepleting drugs.⁶

UCART19 is an innovative, genome-edited, allogeneic anti-CD19 CAR T-cell product engineered using transcription activator-like effector nucleases (TALEN^®, Cellectis). For UCART19, TALEN^® mRNA is used to knock out the genes coding for the TCR alpha constant chain (TRAC) to minimise the risk of GVHD by reducing the number of T-cells that are TCRαβ−positive and confers resistance to the anti-CD52 monoclonal antibody alemtuzumab.⁶ Previously, UCART19 was successfully used, in a special access setting, to treat two children with relapsed infant B-ALL, after conditioning with fludarabine, cyclophosphamide, and alemtuzumab.⁷ A single dose of UCART19 resulted in CAR T-cell expansion and a complete molecular remission which was then consolidated with an allogeneic stem cell transplant (SCT).⁷ Both children remain in remission more than 4 years after UCART19 infusion. An adult with advanced B-ALL was also treated before the trials, and after similar lymphodepletion had evidence of UCART19 expansion and CRS but this subject did not survive disease progression.

This informed the design of two phase 1 studies with UCART19 in patients with relapsed/refractory B-cell ALL: the PALL trial in children and the CALM trial in adults. We report here the pooled results as of 27 August 2019 on the safety, anti-leukaemic activity, and cellular kinetics of UCART19 in patients enrolled in the ongoing PALL and CALM trials before October 23^rd 2018.

Methods

The two ongoing phase 1, open-label studies are being conducted at seven centres (three in France, two in the UK, and two in the USA; table S7 in the appendix), with the primary objective of evaluating the safety and tolerability of UCART19. The trials are performed in accordance with the principles of the Declaration of Helsinki. The authors assume responsibility for the completeness of the data and analyses, as well as for the accuracy of the current report. Written informed consent was obtained from all patients or their parents/legal guardians prior to initiating any treatment.

Enrolled patients participated in the study for approximately 12 months. After their inclusion, patients started 7 days of lymphodepletion treatment prior to UCART19 infusion on Day 0. The treatment period (starting at day 0) lasted up to Day 84 post-infusion. All patients then entered into a 9-month follow-up period up to Month 12.

Patients

The PALL trial enrolled patients aged between 6 months and 18 years and the CALM trial enrolled patients aged 16–70 years. To be eligible, patients had to have evidence of CD19-positive B-cell ALL with >5% leukaemic blasts in bone marrow or a minimal residual disease (MRD) level of ≥1×10⁻³ cells assessed by either flow cytometry or quantitative polymerase chain reaction, and to have exhausted other standard treatment options. Key exclusion criteria included the presence of active CNS leukaemia (CNS-3), extramedullary disease or evidence of active infection (table S1 in the appendix).

Treatment

UCART19 was manufactured from healthy donor T-cells by recombinant lentiviral transduction of the CAR construct which consists of the 4G7 anti-CD19 single-chain variable fragment linked to 4–1BB costimulatory and CD3z activation domains. It also incorporates an RQR8 which includes a CD20 target mimotope for rituximab, designed to allow elimination of UCART19 in the event of excessive toxicity. TALEN^® mRNA targeting the TRAC and CD52 genes were electroporated into the cells to simultaneously disrupt cell surface expression of TCRαβ and CD52 (figures S1 and S2 in the appendix).⁶ Batches from 3 different donors were used for the trials and manufactured in advance of patient enrolment. Characteristics of the final product used in the trials are specified in appendix table S2.

Cytoreduction treatment was permitted for patients with high tumour burden (bone marrow blasts >50%, or circulating blasts of ≥15 000 cells/μL, or elevated lactate dehydrogenase suggesting rapidly progressing disease) prior to the start of lymphodepletion. Details are provided in appendix table S3. Lymphodepletion consisted of fludarabine and cyclophosphamide, with or without alemtuzumab. Details are provided in appendix table S4. Institutional antimicrobial prophylaxis guidelines were followed and mostly included acyclovir, itraconazole or posaconazole and co-trimoxazole.

Informed by pre-trial experiences in two infants, children in PALL received UCART19 at a dose of 1·1–2·3×10⁶ cells/kg. However, the CALM trial included a dose escalation phase to establish a maximum tolerated dose in adults. In this study, adults received UCART19 at one of three dose levels: (1) 6×10⁶ cells, (2) 6–8×10⁷ cells, or (3) 1.8–2.4×10⁸ cells (table S4 in the appendix). The modified Toxicity Probability Interval (mTPI)¹⁹ design was used for the dose escalation.

Following cell infusion, patients were monitored for toxicity related to the lymphodepletion and/or UCART19. Disease response was assessed at Day 28 after UCART19 infusion. Patients achieving a complete remission (CR) or CR with incomplete haematological recovery (CRi) could proceed to an allogeneic stem cell transplant (SCT). Redosing with UCART19 was considered if patients had a suboptimal disease response or an early CD19-positive relapse and/or had poor persistence of UCART19 whilst meeting all other eligibility criteria.

Outcomes

The primary outcome measure was adverse events, graded mostly according to the Common Terminology Criteria for Adverse Events version 4.3. Cytokine release syndrome (CRS) was graded according to the approach outlined by Lee and colleagues (2014),⁸ and GVHD according to criteria defined by the Mount Sinai Acute GVHD International Consortium.⁹

Secondary outcomes were the rate of objective response, duration of remission, relapse-free survival, and overall survival. Key exploratory objectives included analysis of expansion, persistence, and phenotype of UCART19, the proportion of patients who underwent a subsequent allogeneic SCT or redosing with UCART19, potential anti-UCART19 immune responses, and potential CD19-negative tumour escape.

Statistical analysis

The incidence of adverse events and all efficacy outcomes were assessed in the full analysis set, which consisted of all patients who received UCART19 (pooled population). Time-dependent survival parameters were estimated with Kaplan-Meier curves. As PALL and CALM are both phase 1 trials with a small sample size, no formal statistical testing was performed.

Role of the funding source

Servier, the sponsor of both studies was responsible for the study design, data collection, data analysis, and data interpretation, and collaborated in the writing of this article. The listed authors had full access to all the data and had final responsibility for the decision to submit for publication. UCART19 is exclusively licensed to Servier from Cellectis and uses Cellectis’ technologies.

RESULTS

As of October 23^rd 2018, 31 patients were screened for inclusion and 21 patients aged between 9 months and 62 years (median 22 years) were enrolled and treated with UCART19 across the two studies: seven children and 14 adults (figure 1). Ten patients did not meet study eligibility criteria. Among adults, the reasons for exclusion were MRD-negative disease (n=2 patients), active infection (n=2), liver dysfunction (n=1), alternative treatment options not yet exhausted (n=1), and temporary study discontinuation (n=1). Three children were excluded because of active infection (n=1), the presence of anti-HLA antibodies (n=1), and insufficient time since previous treatment (n=1).

Figure 1: Screening, enrolment and study completion.

CRS=cytokine release syndrome. FC=fludarabine and cyclophosphamide. FCA=fludarabine, cyclophosphamide and alemtuzumab. SCT=stem cell transplant. Among the 31 patients who were screened, ten patients did not meet eligibility criteria.

In 5 patients who died from infection 3 viral and 3 bacterial infections were recorded, 1 patient experienced both viral and bacterial infections.

Poor risk cytogenetic abnormalities were reported in 43% of patients (table 1). Patients had received a median of four prior lines of treatment, including inotuzumab ozogamicin in eight patients (38%) and blinatumomab in five (24%) (table 1). Thirteen patients (62%) had undergone prior allogeneic SCT, and five of these (38%) had relapsed within 6 months. At screening, six patients (29%) had minimum residual disease ≥10^–3 cells, and the rest had frank relapse including nine (43%) with >25% blasts in bone marrow. Twelve patients (57%) had cytoreduction treatment prior to inclusion.

Table 1:

Patient baseline characteristics

Baseline Characteristics	CALM (N=14)	PALL (N=7)	Pooled (N=21)
Median age ( Q1-Q3), years	29·5 (22–45)	2·7 (2·2–14)	22 (14–39)
(min-max)	(18–62)	(0·8–16·4)	(0·8–62)
Number of previous lines of therapy
1 to 3, n	4	3	7 (33)
≥4, n	10	4	14 (67)
Median ( Q1-Q3)	4 (3–4)	4 (3–6)	4 (3–5)
(min-max)		(2–6)	(1–6)
	(1–5)
High cytogenetic risk*, n (%)	6	3	9 (43)
Previous treatments, n (%)
Inotuzumab ozogamicin	6	2	8 (38)
Blinatumomab	4	1	5 (24)
Allogeneic SCT	10	3	13 (62)
Time to relapse following prior allogeneic SCT, n (%)
<6 months	4	1	5 (38)
≥6 months	6	2	8 (62)
Cytoreduction† before lymphodepletion, n (%) Bone marrow tumour burden prior to lymphodepletion, n (%)	7	5	12 (57)
<5% of blasts	3	3	6 (29)
5% to 25% of blasts	4	2	6 (29)
>25% of blasts	7	2	9 (43)
Bone marrow tumour burden, median (range) % of blasts	27·5 (0–96)	6 (0–80)	8 (0–96)

SCT=stem cell transplant.

^*High cytogenetic risk includes complex karyotypes, MLL rearrangements, Ph+.

^†Cytoreduction regimens are shown in table S4 in the appendix.

Lymphodepletion and UCART19 treatment

Before UCART19 infusion, 17 patients (81%) received lymphodepletion with fludarabine/cyclophosphamide/alemtuzumab and four (19%) received fludarabine/cyclophosphamide because of concerns about the risk of viral infections with alemtuzumab as per Investigator’s discretion. The median time from consenting patients to start of lymphodepletion was 8.5 days. In the CALM trial, six patients received UCART19 at a dose of 6×10⁶ cells, six received 6–8×10⁷ cells, and two received 1.8–2.4×10⁸ cells. Three adult patients received a second dose of UCART19.

The most common adverse events related to UCART19 and/or lymphodepletion were CRS (n=19 patients, 91%), infections (n=13 patients, 62%), neurotoxicity (n=8 patients, 38%), and prolonged cytopenia (n=6 patients, 29%), defined as grade 4 neutropenia and/or thrombocytopenia persisting beyond D42 day from UCART19 infusion, except if >5% bone marrow blasts (table 2). CRS of grade ≥3 severity occurred in three patients (14% - 95% Confidence Interval (CI) [5%;35%]). Details of all-cause grade 4 cytopenia (neutropenia and/or thrombocytopenia) are provided in appendix table S5.

Table 2:

Adverse events related to UCART19 and/or lymphodepletion (n=21)

	Adverse events by worst grade, n (%)
	Any grade	Grade 1	Grade 2	Grade 3	Grade 4	Grade 5
Cytokine release syndrome	19 (91)	4 (19)	12 (57)	2 (10)	1† (5)	0
Neurotoxicity	8 (38)	7 (33)	1 (5)	0	0	0
Acute cutaneous GVHD	2 (10)	2‡ (10)	0	0	0	0
Infections	13 (62)	2 (10)	3 (14)	4 (19)	2 (10)	2 (10)
Prolonged cytopenia§	6 (29)	0	0	0	6¶ (29)	0
B-cell aplasia	2 (10)	1 (5)	1 (5)	0	0	0
Tumour lysis syndrome	2 (10)	0	0	2 (10)	0	0

GVHD=graft-versus-host disease.

^†One dose-limiting toxicity related to UCART19 at the lowest dose level was grade 4 cytokine release syndrome associated with grade 5 neutropenic sepsis and death on day 15.

^‡GVHD was confirmed by biopsy in one patient.

^§Defined as persistent Grade 4 neutropenia and/or thrombocytopenia beyond day 42 post UCART19 infusion, except if > 5% bone marrow blasts

^¶One dose-limiting toxicity related to UCART19 at the second dose level was grade 4 cytopenia associated with infection, pulmonary haemorrhage and death on day 28.

Median time to CRS onset was 8 days (95% CI [7;8] days) following UCART19 infusion and median duration was 6 days (95% CI [5;8] days). Among 19 patients who experienced CRS, eight (42%) received tocilizumab, three (16%) needed steroids in addition to tocilizumab, seven (37%) required intensive care unit admission, and three (16%) needed inotropic support. Eight patients (38%) developed neurotoxicity (grade 1 in seven and grade 2 in one), with onset at a median of 10 days after treatment. Neurotoxicity lasted a median of 3 days and did not require any specific treatment. Neurological events included dyskinesia, hallucination, irritability, disorganised speech, confusion, dysarthria, chorea, myoclonus, and headache. Two patients (10%), one adult and one child, developed grade 1 acute GVHD of the skin (confirmed by biopsy in the child), but no other manifestation of GVHD was seen. There was no expansion of TCRαβ+ cells in either blood or bone marrow in those patients.

As expected with the use of alemtuzumab, grade ≥3 viral infections were encountered in five patients (24%), including cytomegalovirus, adenovirus, human metapneumovirus, and BK virus. Details of all grades of infections are provided in appendix table S6.Two deaths in the CALM trial were considered related to UCART19, and both were reported as dose-limiting toxicity (DLT). One death was caused by neutropenic sepsis, with grade 4 CRS/DLT as a contributing factor, and occurred on day 15 after receiving UCART19 at the lowest dose level. The other death was caused by pulmonary haemorrhage occurring in the context of infection and grade 4 cytopenia/DLT after an allogeneic SCT on Day 82 after receiving UCART19 at the second dose level.

UCART19 cellular kinetics is shown in figures 3A and 3B. Fifteen of the 17 patients treated with fludarabine, cyclophosphamide, and alemtuzumab (88%) demonstrated UCART19 expansion but none of the fludarabine/cyclophosphamide-treated patients showed any discernible expansion, with indications of earlier host lymphocyte recovery in this group (figure S4 in the appendix). Peak expansion was seen around Day 14. The area under the curve for the first 28 days after UCART19 administration was higher in responders than non-responders (figure 3C). UCART19 persisted beyond day 42 in three patients including one who had detectable cells as late as Day 120.

Figure 3: UCART19 cell kinetics.

Kinetics measured by (A) quantitative PCR (log scale) in six patients in the PALL trial, and (B) flow cytometry (log scale) in 14 patients in the CALM trial and two in the PALL trial; (C) comparison of area under the curve for UCART19 cell count to day 28 (AUC_D28) in patients with complete remission (CR) or complete remission with incomplete hematologic recovery (CRi) and patients with residual disease (RD). VCN=vector copy number.

The overall response rate in the pooled cohort was 67% (95% CI [45%;83%]) (n=14 patients) with 71% (95% CI [45%;88%]) being negative for MRD, with <10⁻⁴ cells (figure 2). Among 14 patients who achieved CR/CRi after the first UCART19 infusion, the median duration of follow-up was 7.1 months, the median duration of response was 4.1 months; ten of these patients (71%) proceeded to a subsequent allogeneic SCT.

Figure 2: Clinical course of individual patients after UCART19 infusion at month 0.

# Denotes patients re-dosed with UCART19 in compassionate use after withdrawal from the parent trial, who were then included in the long-term follow-up trial. Allo-SCT=allogeneic stem cell transplant. C=CALM trial. FC=fludarabine and cyclophosphamide. FCA=fludarabine, cyclophosphamide and alemtuzumab. P=PALL trial. *Patient received a single dose (0·3 mg/m²) of inotuzumab ozogamicin 5 months post UCART19 infusion for maintenance of response, prior to Allo-SCT (data reported post cut-off date) **Patients received new anti-leukaemic treatment due to disease relapse. Grey colour is not visible on the figure because the occurrence of relapses and new treatment initiation were concomitant

Notably, the overall response rate in the 17 patients receiving alemtuzumab-containing lymphodepletion was 82% (95% CI [59%;94%])(n=14 patients; figure 2).

Relapse-free survival at 6 months was 33% (95% CI [12%;56%]) (figure 4) with a 6-month overall survival rate of 55% (95% CI [32%;74%]). At the data cut-off of 27^th August 2019, ten of the 14 patients who achieved a CR/CRi had subsequently relapsed or died. All but one of these relapsing patients were CD19 positive. Three adult patients received a second dose of UCART19 following further lymphodepletion, two achieved CR/CRi and underwent allogeneic SCT at days 43 and 64, respectively, after the second infusion of UCART19 (figure S6 in the appendix). The first patient remained in remission for 10 months while the second one remained in remission for 2 months. The third patient progressed with CNS involvement.

Figure 4: Kaplan-Meier curve of relapse-free survival in patients who achieved complete remission or complete remission with incomplete hematologic recovery at day 28 after the first UCART19 infusion.

At data cut-off, 5 patients (24%) in these refractory cohorts were alive. Ten patients had died due to progressive disease and there were five infection-related deaths, including three after allogeneic SCT.

DISCUSSION

These two studies of UCART19 demonstrate the feasibility of administering an allogeneic genome-edited CAR T-cell product in difficult-to-treat leukaemia, in patients who had undergone multiple lines of previous therapy and who were unsuitable for autologous CAR T-cell therapies. After infusion, UCART19 exhibited expansion and anti-leukaemic activity in paediatric and adult patients with relapsed or refractory B-cell ALL. Following TALEN^® editing and processing, residual TCRαβ-expressing cells (<1·1%) were stringently limited below thresholds expected to cause significant GVHD, and around half the infused cells were effectively invisible to alemtuzumab after CD52 disruption. Such multiplexed editing had previously been noted to result in low level translocation events, and, in the products used here, between 3% and 6% of cells exhibited translocated related karyotype abnormalities. Similar findings have recently been reported in trials of triple genome-edited autologous T-cells, without suggestion of adverse effects.¹⁰

The spectrum of adverse effects observed with UCART19 to date seems similar to those reported with autologous anti-CD19 CAR T-cells.^{1, 3–5, 11–13} CRS was encountered in the majority of patients in whom UCART19 expansion was detected and appeared no more severe than with approved autologous products.¹⁴ Standard methods of treating CRS with tocilizumab and steroids were generally effective, except in one patient, who died on Day 15 with concurrent neutropenic sepsis. Notably, UCART19-related neurotoxicity was mild, with no cases of grade >2 severity.

Cytopenia associated with UCART19 arose at a similar frequency to autologous CAR T-cell therapy.³ Prolonged grade 4 cytopenias were likely related to the use of alemtuzumab in combination with both fludarabine and cyclophosphamide in the lymphodepletion regimen. In pre-trial applications, there had been some concern that residual infused TCRαβ+ cells may expand and cause transfusion-associated GVHD, but this was not encountered during the current studies. No significant expansion of donor TCRαβ+ cells was detected and only two cases of grade 1 GVHD limited to skin occurred.

Infectious complications, in particular viral infections were observed, in keeping with the known complications of alemtuzumab use.¹⁵ Patients were closely monitored for cytomegalovirus, adenovirus, and Epstein-Barr viraemia and preemptive treatment strategies were adopted. Alemtuzumab was omitted in four patients who were at high risk of viral infections, but no UCART19 expansion was noted in these patients. The adult dose of alemtuzumab was subsequently reduced from 1 mg/kg to a flat 40 mg dose, while children received 1 mg/kg (capped at 40 mg) to reduce the incidence of severe viral infections.

Encouragingly, UCART19 expansion was observed even at the lowest dose tested in adults. The kinetics of expansion were generally similar to those seen with autologous CAR T-cells,^{3, 5, 12, 13, 16} except that UCART19 persistence was generally limited to 28 days; only 14% of patients had longer persistence. No clear correlation was seen between dose and persistence beyond D28 although the number of patients treated so far is too low to draw any firm conclusion. Alemtuzumab-containing lymphodepletion appears to be required for UCART19 expansion with preliminary analysis showing a correlation between higher serum alemtuzumab levels and superior UCART19 expansion (Figure S3). The four patients treated with only fludarabine and cyclophosphamide exhibited early host T +/− NK cell recovery and no UCART19 expansion, suggesting host cell mediated rejection may have occurred (figure S4 and S5 in the appendix). Unsurprisingly, with humoral immunity impaired during B cell aplasia, donor-specific anti-HLA antibodies were not detected following lymphodepletion and UCART19 infusion.

The overall response rate of 82% (95% CI [59%;94%]) in patients who received fludarabine/cyclophosphamide/alemtuzumab is encouraging and similar to that observed with autologous anti-CD19 CAR T-cells.^{1, 3} Patients in the studies reported here were high risk and heavily pretreated, including infants and older patients with relapsed/refractory B-cell ALL, as well as patients for whom an autologous product could not be generated because of profound lymphopenia. Analysis of UCART19 kinetics revealed a positive correlation between UCART19 expansion and response. Redosing with UCART19 was possible and associated with continued anti-leukaemic activity, as shown by the UCART19 expansion and molecular remission achieved in two of the three patients who underwent a second infusion. These data support the hypothesis that a repeat dosing strategy may deepen and prolong remission, potentially leading to disease eradication. Redosing has been reported with autologous CAR T-cells, albeit with limited efficacy,¹⁷ and with the major hurdle of generating sufficient product for the patients in need.

Most patients who achieved a molecular remission with UCART19 underwent allogeneic SCT. The SCT was performed for bone marrow aplasia in two patients, and the remaining ten had a SCT to consolidate complete remission when UCART19 was no longer detectable. Two of the 12 patients underwent transplantation for the second time using a different SCT donor and reduced intensity conditioning regimens. No impact of prior UCART19 treatment on stem cell engraftment was observed. Park and colleagues, in their series of 19–28z CAR T-treated adult patients with relapsed/refractory B-cell ALL, failed to show any significant difference in outcomes when a subsequent allogeneic SCT was performed.⁴ However, Hay and colleagues showed improved event-free survival with an allogeneic SCT after 19–4-1BBz CAR T-cell therapy compared with CAR T-cells alone.²

Certain differences between these two UCART19 studies in terms of trial design, lymphodepletion regimen and UCART19 cell dose should be noted, and although the studies are ongoing, our pooled analysis provides an early indication of the potential for off-the-shelf therapies. The results so far provide valuable information on GVHD, CRS, infectious complications and genotoxicity risks as well as an indication of efficacy of the first genome edited allogeneic CAR-T cell product. Other similar strategies, without genome editing, include allogeneic anti-CD19 CAR-Natural Killer cells co-expressing interleukin-15, with remissions recently reported in refractory chronic lymphocytic leukaemia and non-Hodgkin’s lymphoma¹⁹. Further studies will help elucidate redosing strategies, factors that may influence UCART19 persistence and the need for allo-SCT.

In conclusion, phase 1 clinical studies in paediatric and adult patients with late-stage relapsed/refractory B-cell ALL have demonstrated the feasibility, safety and the activity of UCART19, an off-the-shelf, pre-manufactured CAR T-cell product. This represents a significant step forward in CAR T-cell development and may herald a new, effective and easily accessible cell therapy for patients with B-cell ALL.

Research in Context.

Evidence before this study

Autologous CD19 targeted CAR T-cells are now well established in the treatment of children and young adult patients with relapsed refractory B-acute lymphoblastic leukaemia (B-ALL) showing 70–90% response rates and long-term remissions of 50–60%. However, the process of autologous CAR T-cell manufacturing is time consuming, logistically challenging, and expensive. Furthermore, for heavily pretreated lymphopenic patients, infants, and those with rapidly proliferative disease, it may not always be possible to manufacture an autologous CAR T-cell product. This has led to considerable interest in developing off-the-shelf allogeneic CAR T-cell products generated from healthy donors that can be given to any patient without matching and whenever required. To overcome the obstacles of graft versus host disease (GVHD) and rejection of HLA incompatible cells, strategies have been developed to disrupt the endogenous T cell receptor (TCR) αβ of the CAR T product as well as lymphodepleting the host immune system to minimize CAR T rejection whilst engineering the CAR T-cells to be resistant to the lymphodepleting drug.

Previously, two infants with relapsed B-ALL were successfully treated in 2015 with UCART19, comprising off-the-shelf allogeneic CD19 targeted CAR T-cells in which TCR αβ and CD52 genes had been simultaneously disrupted by transcription activator-like effector nucleases (TALEN^®). That experience provided important insights which have now been evaluated in the PALL and CALM studies reported here.

We searched PubMed using the terms (“allogeneic CAR*” OR “donor-derived CAR*” OR “allogeneic chimeric antigen receptor*” OR “donor-derived chimeric antigen receptor*”) AND (“acute lymphoblastic leukaemia” OR “acute lymphoblastic leukemia” OR “ALL”) published to April 7^th 2020. We identified 15 articles, of which only one was a clinical study (the others were case reports [3], preclinical studies [2], correspondence/comments [2], and reviews [7]). The clinical trial described the use of unmodified CAR T-cells in six patients with B-ALL who had relapsed after haplo-identical haematopoietic stem cell transplantation; the CAR T-cells came from the same donor as the patient’s stem cells. Five of the six treated patients were able to achieve a remission, but three developed grade 2 or 3 GVHD. Several other genome edited CAR T-cell products are actively being evaluated in ongoing clinical trials in haematological malignancies.

Added value of this study

There is an unmet need for a therapeutic option for patients with B-ALL who are not candidates for autologous CAR T-cell therapy. These 2 phase 1 trials demonstrate the safety and efficacy of UCART19 and represent the first published clinical trials using off-the-shelf CAR T-cells. The studies demonstrate the feasibility of delivering genome-edited allogeneic CAR T-cells to children and adult patients with high risk B-ALL and resulted in manageable side effects with high complete response rates, comparable to autologous CAR T-cells. Encouragingly, UCART19 therapy was found to cause only minimal GVHD and no deleterious expansion of unedited TCR αβ cells from the infused product was seen. The studies also highlight some of the challenges of this field such as the need for an intensive lymphodepletion strategy to prevent host rejection of UCART19, an increased incidence of viral infections and limited persistence of these allogeneic CAR T-cells.

Implications of all the available evidence

Experience with UCART19 from the CALM and PALL trials opens up the possibility of using allogeneic off-the-shelf CAR T-cells to treat children and adults with B-ALL, especially when autologous CAR T-cell therapy may be unavailable or when there is rapidly progressive disease. The study adds to safety profiling of CAR T-cells, now including genome-editing aspects and paves the way for future applications of this technology. Engraftment was limited to patients who received alemtuzumab in addition to standard lymphodepletion and further research is needed to determine optimal lymphodepletion strategies for use with UCART19. Larger studies with longer follow-up will better characterise the safety and efficacy profile of this treatment.