Outcomes of Second Anti-CD19 CAR T-Cell Therapy (CART2) in Acute B Lymphoblastic Leukemia and the Impact of Allo-HSCT on Efficacy

Abstract

For patients exhibiting a suboptimal response to the first chimeric antigen receptor (CAR) T-cell therapy (CART1) or relapse after remission, secondary CAR T-cell therapy (CART2) for the same target may be an option. We retrospectively analyzed patients with acute B-cell lymphoblastic leukemia (B-ALL) receiving CD19 CART1 at our center (n = 84) to report the clinical outcomes of CART2 and to identify the factors that may influence the outcomes. Twenty-six patients received CART2 for suboptimal response or relapse post-CART1. The incidence of cytokine release syndrome (CRS) after CART2 was 65.4% (17/26), with 11 cases classified as grade 1 (42.3%), four cases as grade 2 (15.4%), and two cases as grade 3 (7.7%). Neurotoxicity was observed in one patient (3.8%) after CART2 infusion. Fourteen patients (53.8%) achieved complete remission (CR) after CART2. CART2 exhibited an inferior response rate (CART2: 53.8%, 14/26; CART1: 81.0%, 64/79; P = 0.006) and a lower incidence of severe CRS (CART2: 7.7%, 2/26; CART1: 30.4%, 24/79; P = 0.020) compared with CART1, with a median progression-free survival (PFS) and a median overall survival (OS) of 6.2 months and 11.2 months, respectively. In particular, patients who progressed after consolidative allogeneic hematopoietic stem cell transplantation (allo-HSCT) following CART1 and then received CART2 demonstrated promising outcomes with a response rate of 80.0% (8/10), a median PFS of 7.9 months, and a median OS of 25.1 months. After adjusting for the confounding factors, the response rate (85.7%, 6/7) of CART2 administered to this cohort was better than those who did not bridge to allo-HSCT receiving CART2 (28.6%, 2/7) or non-CART2 treatments (13.3%, 2/15). The median OS after CART2, which was not reached, was significantly better than the median OS after CART2 (3.9 months, P = 0.014) and non-CART2 treatments (6.0 months, P = 0.012) administered in patients who did not undergo consolidative allo-HSCT post-CART1. Our results indicated that, although less effective than CART1, a subset of patients can still benefit from CART2 with mild adverse effects. For patients who relapsed after consolidative allo-HSCT post-CART1, treatment with CART2 is a viable option.

Introduction

Chimeric antigen receptor (CAR) T-cell therapy has demonstrated remarkable results since its development, with particularly notable remission rates ranging from 64% to 93% in hematological malignancies^1–11. However, approximately 50% of patients who achieve remission experience relapse during the follow-up period, with a median event-free survival period of ≤1 year^3,4,7,9–13. These patients subsequently undergo treatments such as salvage chemotherapy, immune checkpoint inhibitor treatment ¹⁴ , or CAR T-cell reinfusion with identical or different targets^15–19. Salvage chemotherapy is a high-risk treatment with poor efficacy and is generally applied to patients without other options. Studies on the administration of immune checkpoint inhibitors after relapse following the first CAR T-cell therapy (CART1) are lacking. Chong et al. ¹⁴ used pembrolizumab to treat 12 patients with B-cell lymphoma, who were either refractory to or relapsed after CD19-targeted CART1, achieving an acceptable safety profile and clinical benefit in 33% (4/12) of the patients. Although CAR T-cell reinfusion is a frequently chosen subsequent treatment, the overall results are unsatisfactory. Gauthier et al. ¹⁵ found that a second CAR T-cell therapy (CART2) targeting CD19 resulted in complete remission (CR) rates of 22%, 19%, and 21% in chronic lymphoblastic leukemia (CLL), non-Hodgkin lymphoma (NHL), and acute B-cell lymphoblastic leukemia (B-ALL) patients, respectively. In another retrospective study on 18 children and young adults with B-ALL who were reinfused with anti-CD19, anti-CD22, or anti-CD19/22 CART2 products, the objective remission rate was 38.9% (7/18) ¹⁷ . However, despite the lack of favorable outcomes in patients treated with anti-CD19 CART2 at our center, we observed that a proportion of patients still achieved remission. This study aims to assess the therapeutic efficacy of CART2 treatment in patients with B-ALL. To achieve this, a retrospective analysis of clinical data was conducted on patients who underwent one or multiple infusions of autologous murine-CD19 CAR T cells at our center. The primary objectives were to elucidate the clinical outcomes of CART2 in comparison with CART1 and to investigate the factors that may influence adverse effects, efficacy, and prognosis of CART2. In addition, we aimed to identify the population characteristics that are most likely to benefit from CART2, which could be valuable in guiding the treatment of patients who have shown poor response or relapses after CART1.

Materials and Methods

Patients and Study Endpoints

Patients diagnosed with B-ALL who received autologous murine-CD19 CAR T cell therapy at our center (registered at https://www.clinicaltrials.gov/ numbered NCT02851589 and NCT02735291) between 2016 and 2021 were included. These patients were subsequently screened for the study, focusing on CART2, based on the following criteria: (1) patients who received either CART2 or non-CART2 treatments after failing to achieve CR or experiencing a relapse after CART1, and (2) patients who did not receive other types of CAR T-cell therapy (such as humanized CD19 CAR or allogeneic CAR) between CART1 and the subsequent treatment. The study protocols were reviewed and approved by the Medical Ethics Committee of the First Affiliated Hospital of USTC (University of Science and Technology of China) and the Second Hospital of Anhui Medical University (SHAMU). Informed consent was obtained from all enrolled patients. The primary endpoints of the study included the response rate, progression-free survival (PFS), and overall survival (OS) of CART2, whereas the safety of CART2 served as the secondary endpoint.

CAR-T Therapy Protocol

Patients received conditioning chemotherapy prior to an infusion of autologous CD19-targeted CAR T cells. The patient-derived T cells were transduced with a lentiviral or retroviral vector that encoded a CAR consisting of a co-stimulatory domain (CD28 alone or CD28/4-1BB), CD3z, and a murine anti-CD19 single-chain fragment variable (scFv) domain. The process of preparing of CAR T cells was previously described in detail in our study ²⁰ . The infusion dose of CAR T cells was generally 1 × 10⁶ cells/kg body weight and adjusted flexibly according to the patient’s clinical condition and disease status. In cases of second infusions, CAR T cells were either recollected from the patients (n = 24) or cryopreserved (n = 2) from CART1.

Definitions and Assessments

Non-CART2 treatments included conventional chemotherapy and symptomatic treatments. Poor prognostic markers included complex karyotypes, BCR-ABL1, MLL-AF4, TP53, and E2A-PBX1. Adverse events and efficacy were evaluated by observing daily symptoms, laboratory indicators, imaging, and bone marrow (BM) examination results after infusion. Cytokine release syndrome (CRS) was graded using the American Society for Transplantation and Cellular Therapy (ASTCT) Consensus Grading System ²¹ ; a CRS grading of ≥3 was considered severe. Neurotoxicity was assessed based on the National Cancer Institute common terminology criteria for adverse events v4.03. BM samples were assessed at approximately 21 days after CAR T-cell infusion; CR was defined as a blast percentage of ≤5%, no blasts in peripheral blood primitive cells, and no extramedullary infiltration, whereas partial remission (PR) was defined as a BM blast ratio of 5% to 25% or a decrease of >50% relative to the pretreatment level. A low tumor burden referred to a <5% BM blast ratio with positive minimal residual disease (MRD), whereas a high tumor burden referred to a ≥5% BM blast ratio. Disease relapse included morphological relapse (>5% of blast cells in the BM) or extramedullary infiltration. Disease progression encompassed MRD positivity, the reappearance of molecular genetic abnormalities, morphological relapse, or extramedullary infiltration. PFS was calculated as the time from CART2 infusion or non-CART2 treatment application to disease progression, death, or the last visit. OS was defined as the time from CART2 infusion or non-CART2 treatment application to death or the last visit.

Statistical Analysis

Descriptive data were presented as follows: Categorical variables were expressed as counts (percentage), continuous variables that followed a normal distribution were presented as mean values ± standard deviations, and continuous variables that did not follow a normal distribution were presented as median values [with the interquartile range (IQR)]. For comparison between groups, categorical variables were analyzed using χ² or Fisher exact tests, normally distributed continuous variables were analyzed using an independent sample t test or one-way analysis of variance (ANOVA), and non-normally distributed continuous variables were analyzed using the Mann–Whitney U test or Kruskal–Wallis test. Binary logistic regression was employed to analyze the factors affecting the CR rate of CART2, whereas Cox regression analysis was used to determine the factors affecting survival of CART2. Kaplan–Meier curves were generated using GraphPad Prism 8 to visualize survival difference. P values of < 0.05 (bilateral) were considered statistically significant. IBM SPSS Statistics 26 was used for all statistical analyses.

Results

Patient Characteristics

A total of 84 patients with B-ALL received murine-derived CD19 CAR T cells; among them, five patients (6.0%) were unable to be evaluated for efficacy due to death related to disease progression or severe CRS (Fig. 1). Seventeen patients exhibited suboptimal response to CART1 [PR: n = 2, non-remission (NR): n = 15]; among these patients, seven received CART2 and 10 received non-CART2 treatments [one patient underwent allogeneic hematopoietic stem cell transplantation (allo-HSCT) after non-CART2 treatment]. Sixty-two patients achieved CR after CART1: 24 patients underwent consolidative allo-HSCT, with seven maintaining CR, three dying from transplantation-related mortality (TRM), one lost to follow-up, and 13 experiencing disease progression (10 received CART2 and three received non-CART2 treatments). Thirty-eight patients did not receive consolidative therapy after CART1-induced CR; among them, one patient sustained CR, one died from infection, five were lost to follow-up, two received non-CART2 treatment followed by allo-HSCT, and 29 subsequently experienced disease progression. Among the 29 patients, one died rapidly after CD19-negative relapse, one developed chronic myeloid leukemia (CML), four received other types of CAR-T cell therapy, nine received CART2, and 14 received non-CART2 treatments. Patients who did not undergo consolidative allo-HSCT after CART1 or did not proceed to CART2 after relapse cited financial reasons or personal preference.

Figure 1.

Patient flow chart. allo-HSCT: allogeneic hematopoietic stem cell transplantation; CART1/2: first/second chimeric antigen receptor T-cell therapy; CML: chronic myeloid leukemia; CR: complete remission; NR: non-remission; PR: partial remission; SCT: stem cell transplantation; TRM: transplantation-related mortality.

Characteristics and Adverse Events of Patients in the CART2 Group

Baseline characteristics were similar between CART1 and CART2 cohorts (Table 1). Twenty-six patients received CART2 for either poor response (26.9%, 7/26) or progression after CR following CART1 (73.1%, 19/26). The median age of these 26 patients was 19.5 years (IQR, 11.0–35.5). Twelve patients (46.2%) had poor prognostic markers. Twenty-three patients (88.5%) experienced ≥2 relapses or sustained non-remission before CART1, whereas three patients (11.5%) experienced <2 relapses. The median blast ratio in BM prior to CART2 infusion was 39.0% (IQR, 6.4%–82.0%). Five patients (19.2%) were in remission status with positive MRD at the time of infusion, whereas 21 patients (80.8%) were not in remission. In addition, five patients had a history of allo-HSCT before CART1, and 10 patients (38.5%) underwent consolidative umbilical cord blood transplantation (UCBT) following CART1-induced CR with 100% chimerism after transplantation.

Table 1.

Characteristics of Patients Who Received CART1 and CART2.

Characteristics	CART1 (n = 84)	CART2 (n = 26)	χ2/Z/t	P
Age (years)	24.5 (16.0~39.0)	19.5 (11.0–35.5)	1.211	0.226
Sex			0.007	0.935
Male	38 (45.2)	12 (46.2)
Female	46 (54.8)	14 (53.8)
Poor prognostic markers a			0.088	0.767
No	48 (57.1)	14 (53.8)
Yes	36 (42.9)	12 (46.2)
Number of relapses before CART1			1.808	0.179
<2	20 (23.8)	3 (11.5)
≥2 or NR	64 (76.2)	23 (88.5)
Allo-HSCT before CART1			–	0.525
No	73 (86.9)	21 (80.8)
Yes	11 (13.1)	5 (19.2)
Consolidative allo-HSCT following CART1			–	–
No	–	16 (61.5)
Yes	–	10 (38.5)
BM blast ratio prior infusion (%)	49.0 (18.0–76.0)	39.0 (6.4–82.0)	0.868	0.386
Tumor burden at the time of infusion				0.525
Low	11 (13.1)	5 (19.2)
High	73 (86.9)	21 (80.8)
Conditioning chemotherapy			–	0.590
Flu-based	81 (96.4)	24 (92.3)
Non-Flu-based	3 (3.6)	2 (7.7)
Infusion dose of CAR-T cells (× 106 cells/kg)	1.1 (0.7–2.2)	1.8 (1.2–2.4)	1.772	0.076
Time interval between two infusions (days)	–	281.5 (83.0–423.3)	–	–
Manufacture of CART2	–		–	–
Recollected		24 (92.3)
Cryopreserved		2 (7.7)
Indication for CART2	–		–	–
No response to CART1		6 (23.1)
Partial response after CART1		1 (3.8)
Relapse after CART1-induced CR		17 (65.4)
Positive MRD after CART1-induced CR		2 (7.7)

Data are presented as numbers/counts (percentages), mean values ± SDs, or median values (with the IQR). allo-HSCT: allogeneic hematopoietic stem cell transplantation; BM: bone marrow; CART1/2: first/second chimeric antigen receptor T-cell therapy; Flu: fludarabine; IQR: interquartile range; MRD: minimal residual disease; NR: non-remission.

^aIncluding complex karyotypes, BCR-ABL1, MLL-AF4, TP53, and E2A-PBX1.

All patients were pretreated with conditioning chemotherapy [fludarabine (30 mg/m² × 3 days) combined with cyclophosphamide (300 mg/m² × 3 days), n = 24; cyclophosphamide monotherapy, n = 2] before the infusion of CART2. The median dose of CART2 was 1.8 (IQR, 1.2–2.4) × 10⁶cells/kg, and the median time interval between two infusions was 281.5 (83.0–423.3) days. Among the CART2 group, five patients received CAR T cells transduced with retroviral vectors for the first infusion and lentiviral vectors for the second infusion, whereas two patients received CAR T cells transduced with retroviral vectors for both infusions and the remaining 19 patients received CAR T cells transduced with lentiviral vectors for both infusions (Supplemental Table S1).

The incidence of CRS after CART2 was 65.4% (17/26), with 11 cases classified as grade 1 (42.3%), four cases as grade 2 (15.4%), and two cases as grade 3 (7.7%). The incidence of severe CRS was significantly lower compared with CART1 (7.7%, 2/26 vs 30.4%, 24/79, P = 0.020; Supplemental Fig. S1A). Thirteen patients had a CRS grade of lower than grade 3 despite an increase in CART2 dosage compared with CART1. Neurotoxicity was observed in one patient (3.8%) after CART2 infusion (Supplemental Fig. S1B). All adverse events were reversible.

Given the low incidence of severe CRS following CART2, correlation analysis was performed after grouping based on CRS grade greater than level 2 (Supplemental Table S2) and no significant relative factors were found.

The Response Rate, PFS, and OS of the CART2 Group

Fourteen patients (53.8%) achieved MRD-negative CR after CART2, despite nine of them (64.3%) presenting poor prognostic markers, 12 (85.7%) experiencing ≥2 relapses or sustained non-remission, and 11 (78.6%) being in a non-remission status at the time of infusion. Notwithstanding, the remission rate of CART2 remained inferior compared with CART1 (81.0%, 64/79; P = 0.006; Supplemental Fig. S1C). Among the patients who responded to CART1, a significant proportion (65.0%, 13/20) maintained their response when treated with CART2. In contrast, only a small percentage of patients (16.7%, 1/6) who did not respond to CART1 achieved CR after CART2 administration. Five patients were in morphologic remission (MRD-positive) status at the time of CART2 infusion, and CART2 resulted in MRD negativity for three patients. Out of the 21 patients in a non-remission status at the time of infusion, 11 of them (52.4%) achieved CR. It is worth mentioning that 57.1% (8/14) of those who achieved remission after CART2 received allo-HSCT following CART1, whereas 83.3% (10/12) of patients who failed to respond to CART2 did not undergo allo-HSCT after CART1 (Table 2).

Table 2.

Patient and Disease Characteristics Stratified by the Response to CART2.

Characteristics	NR (n = 12)	CR (n = 14)	χ2/Z/t	P
Age (years)	26.3 ± 14.2	21.4 ± 13.8	0.877	0.389
Poor prognostic markers a			4.013	0.045*
No	9 (75.0)	5 (35.7)
Yes	3 (25.0)	9 (64.3)
Number of relapses before CART1				1.000
<2	1 (8.3)	2 (14.3)
≥2/NR	11 (91.7)	12 (85.7)
BM blast ratio prior CART1 (%)	55.7 ± 29.7	39.0 ± 30.5	1.387	0.179
Allo-HSCT before CART2			5.793	0.058
No history of allo-HSCT	8 (66.7)	3 (21.4)
Pre-CART1	2 (16.7)	3 (21.4)
Consolidated after CART1	2 (16.7)	8 (57.1)
Response to CART1				0.065
NR	5 (41.7)	1 (7.1)
CR/PR	7 (58.3)	13 (92.9)
Increased dose of CART2				0.099
No	2 (20.0)	7 (58.3)
Yes	8 (80.0)	5 (41.7)
BM blast ratio prior CART2 (%)	54.1 ± 36.7	30.8 ± 36.1	1.479	0.139
Tumor burden prior CART2				1.000
High	10 (83.3)	11 (78.6)
Low	2 (16.7)	3 (21.4)

allo-HSCT: allogeneic hematopoietic stem cell transplantation; BM: bone marrow; CART1/2: first/second chimeric antigen receptor T-cell therapy; CR: complete remission; NR: non-remission; PR: partial remission.

^aIncluding complex karyotypes, BCR-ABL1, MLL-AF4, TP53, and E2A-PBX1.

^*P < 0.05 (bilateral).

At a median follow-up time of 50.8 months [95% confidence interval (CI), 19.0–82.6], the median PFS and OS from CART2 infusion were 6.2 months (95% CI, 3.5–9.0) and 11.2 months (95% CI, 4.7–17.7), respectively (Fig. 2A, B). Among the 14 patients who achieved CR, two (14.3%, 2/14) sustained remission without additional interventions to the end of follow-up, with PFS of 55.0 and 31.7 months, respectively. Three patients (21.4%, 3/14) showed a positive MRD around 2 months after infusion, with one patient dying after failure of salvage chemotherapy, one patient dying from severe myelosuppression, and another patient being lost to follow-up. Eight (57.1%, 8/14) patients experienced a relapse after infusion, with a median time of 6.5 months (IQR, 3.5–19.3). Among them, four died rapidly after relapse, and four received salvage chemotherapy (two died approximately 1 year after relapse, and two are still alive). One (7.1%, 1/14) patient was lost to follow-up 6 months after achieving remission. Among the patients who did not respond to CART2 (n = 12), three were lost to follow-up, and the remaining nine (75.0%) all died within 1 year after infusion.

Figure 2.

Survival curves of patients treated with CART2 and comparison of efficacy between different subgroups. (A) PFS of patients in the CART2 group. (B) OS of patients in the CART2 group. (C) PFS after CART2 in patients who underwent allo-HSCT at different time periods. (D) OS after CART2 in patients who underwent allo-HSCT at different time periods. (E) Comparison of median OS in patients with different responses treated with CART2. (F) Comparison of the CR rate between different subgroups. (G) Comparison of median PFS between different subgroups. (H) Comparison of median OS between different subgroups. Group 1: patients received CART2 after CART1 with consolidative allo-HSCT; group 2: patients received CART2 after CART1 without consolidative allo-HSCT; group 3: patients received non-CART2 treatments after CART1 without consolidative allo-HSCT. allo-HSCT: allogeneic hematopoietic stem cell transplantation; CART1/2: first/second chimeric antigen receptor T-cell therapy; CR: complete remission; NR: non-remission; OS: overall survival; PFS: progression-free survival.

Notably, four patients achieved remission lasting more than 2 years, all of whom received CART2 after consolidative allo-HSCT post-CART1. Survival curves demonstrated relatively better PFS and OS in patients who underwent consolidative allo-HSCT post-CART1 compared with those who did not (Fig. 2C, D). In addition, patients who did not respond to CART2 had a significantly inferior median OS than those who achieved CR [4.6 months vs 18.3 months, P = 0.001, hazard ratio (HR) = 0.186 (95% CI, 0.061–0.570)] (Fig. 2E).

Impact of Allo-HSCT on the Response to and Survival After CART2

Although no significant correlation was observed between factors that may influence prognosis (eg, age, poor prognostic markers, number of relapses before CART1, and tumor burden) and survival of CART2 (Supplemental Tables S3 and S4), the above description indicates that consolidative allo-HSCT after CART1 may have a potential impact on the response rate and survival of CART2. Therefore, we conducted further subgroup analysis after adjusting for confounders.

Considering that the history of allo-HSCT could be confounded with the source of T cells for CAR T-cell manufacture, patients with a history of allo-HSCT before CART1 were excluded. The remaining patients were then divided into four groups to explore the impact of allo-HSCT on the efficacy of CART2: group 1 included patients who received CART2 after CART1 with consolidative allo-HSCT (n = 10); group 2 included patients who received CART2 after CART1 without consolidative allo-HSCT (n = 11); group 3 included patients who received non-CART2 treatments after CART1 without consolidative allo-HSCT (n = 20); and group 4 (n = 3) included patients who received non-CART2 treatments after CART1 with consolidative allo-HSCT.

In group 1, two patients (20.0%, 2/10) did not respond to CART2, whereas eight patients (80.0%, 8/10) obtained CR again. Among the eight patients who responded to CART2, two remained progression-free until the follow-up endpoint (with PFS of 55.0 and 31.7 months, respectively), whereas six patients suffered disease progression at a median time of 7.1 (IQR, 4.2–30.0) months. In group 2, three patients (27.3%, 3/11) achieved CR by CART2, two of them progressed within 3 months and one patient was lost to follow-up at 8 months. In group 3, one patient died of severe myelosuppression and was unable to evaluate efficacy, and only two (10.0%, 2/20) patients achieved CR and relapsed in the short term. In group 4, two patients died due to rapid disease progression and one patient maintained remission after interferon injection due to positive MRD; this cohort was not included in the statistical analysis due to the small number of cases. The median PFS of the second treatment for progression of CART1 in groups 1, 2, and 3 were 7.9, 2.1, and 6.0 months, respectively, and the median OS of was 25.1, 7.6, and 4.4 months, respectively.

Baseline characteristics, including tumor burden before two infusions, poor prognostic cytogenetic markers, and number of relapses before CART1, were comparable among groups, except that patients in group 1 were younger than those in the other two groups (P = 0.003). After adjusting for the confounding factor of age (excluding patients aged younger than 10 years and older than 45 years), the CR rates and survival were compared between groups 1, 2, and 3 (Table 3). The CR rates were 85.7% (6/7), 28.6% (2/7), and 13.3% (2/15) in groups 1, 2, and 3, respectively, showing a significant difference (P = 0.004). Post hoc analysis (Fig. 2F) revealed that the CR rate of CART2 administered to patients who relapsed after CART1 bridging allo-HSCT was higher than that of group 2, although the difference did not reach statistical significance (P = 0.103). However, the advantage of CART2 in group 1 was more pronounced when compared with non-CART2 treatments (P = 0.002) administered to those who did not receive consolidative allo-HSCT after CART1. There was no significant difference in CR rates between groups 2 and 3 (P = 0.565).

Table 3.

Comparison of Efficacy and Survival of CART2/Non-CART2 Treatments Among Different Subgroups.

Characteristic	Group 1 a (n = 7)	Group 2 a (n = 7)	Group 3 a (n = 15)	F/log-rank χ2	P
Age (years)	18.6 ± 7.9	27.0 ± 10.2	25.1 ± 8.3	1.887	0.172
Tumor burden prior CART2				2.739	0.483
Low	0 (0.0)	1 (14.3)	0 (0.0)
High	7 (100.0)	6 (85.7)	15 (100.0)
Poor prognostic markers				3.261	0.208
No	2 (28.6)	5 (71.4)	10 (66.7)
Yes	5 (71.4)	2 (28.6)	5 (33.3)
Number of relapses before CART1				3.460	0.197
<2	3 (42.9)	0 (0.0)	4 (26.7)
≥2/NR	4 (57.1)	7 (100.0)	11 (73.3)
Response				10.404	0.004**
NR	1 (14.3)	5 (71.4)	13 (86.7)
CR	6 (85.7)	2 (28.6)	2 (13.3)
Median PFS (months)	7.9	1.0	6.0	11.390	0.003**
	7.9	1.0		8.333	0.004**
	7.9		6.0	1.824	0.177
		1.0	6.0	2.000	0.157
Median OS (months)	Not reached	3.9	5.6	8.314	0.016*
Post hoc of median OS (months)	Not reached	3.9	–	6.080	0.014*
	Not reached	–	5.6	6.343	0.012*
	–	3.9	5.6	0.147	0.702

Data are presented as numbers/counts (percentages), mean values ± SDs, and median values (with the IQR). allo-HSCT: allogeneic hematopoietic stem cell transplantation; CART1/2: first/second chimeric antigen receptor T-cell therapy; CR: complete remission; IQR: interquartile range; NR: non-remission; OS: overall survival.

^aGroup 1: patients received CART2 after CART1 with consolidative allo-HSCT; group 2: patients received CART2 after CART1 without consolidative allo-HSCT; group 3: patients received non-CART2 treatments after CART1 without consolidative allo-HSCT.

^*P < 0.05 (bilateral); **P < 0.01 (bilateral).

The median PFS were 7.9, 1.0, and 6.0 months in groups 1, 2, and 3, respectively (Table 3), and the median OS were not reached, 3.9, and 5.6 months, respectively. The PFS in group 1 was significantly better than in group 2 (P = 0.004; Fig. 2G), and the OS was significantly better than groups 2 (P = 0.014) and 3 (P = 0.012; Fig. 2H). No difference in median PFS or OS was found between group 2 and group 3.

Discussion

In patients exhibiting an inferior response or experiencing recurrence after anti-CD19 CAR T-cell therapy, reinfusion of the same CAR T cells is considered as a potential therapeutic approach. In a previous retrospective study involving 44 patients with refractory/relapse (R/R) B-cell malignancies, the use of anti-CD19 CART2 resulted in response rates of 33%, 52%, and 21%, and median response durations of 33, 6, and 4 months in patients with CLL (n = 9), NHL (n = 21), and ALL (n = 14), respectively ¹⁵ . In another study, 10 of 40 patients who were unresponsive or relapsed after CART1 achieved CR by CART2 ¹⁶ . In a study of children and young adults with B-ALL, treatment with CART2 led to an objective response rate of 38.9% ¹⁷ . These findings suggest that a second infusion of CAR T cells can be effective for a subset of patients; however, further investigation is needed to identify the groups that benefit from CART2 and to understand the potential factors affecting the efficacy of CART2.

In this study, we retrospectively analyzed the clinical outcomes of B-ALL patients reinfused with the same CAR T cells following an inferior response or recurrence after an initial treatment with anti-CD19 CAR T cells. Accordingly, we identified a population that could potentially benefit from CART2: Patients who relapsed after receiving consolidative allo-HSCT following CART1-induced CR were more likely to benefit from CART2. Although analogous studies are limited, the commonalities and discrepancies between our study and previous studies are discussed in the following.

In this study, CART2 resulted in a lower response rate and less severe CRS compared with CART1, which is consistent with previous studies^15,17. A retrospective review of children and young adults with B-ALL who received reinfusion with an anti-CD19, anti-CD22, or anti-CD19/22 CAR T construct also indicated that the expansion of CAR T cells was more robust in CART1, and a subset of patients resistant to CART2 failed to demonstrate any CAR T-cell expansion, suggesting that limited CAR T-cell expansion might impede robust responses to CART2 and potentially reduce the severity of CRS^15,17. Moreover, the inferior efficacy of CART2 may be attributed to the immunogenicity of CD8⁺ T cells developed after the first infusion, which target the epitope encoded by the CAR transgene, especially in the case of murine-derived CAR T cells^22,23. To address this issue, the use of humanized CAR products in the second infusion, which incorporate humanized scFv, could potentially reduce immunogenicity. Although comparisons are not feasible because the CAR manufactured in our study was of murine origin, other studies with larger sample sizes have found that a second infusion of humanized CAR T cells can achieve better outcomes in patients who showed poor performance following the initial infusion of murine-derived CAR T cells^18,24.

Several variables affected the response to and survival after CART2, among which the history of transplantation was predominant. Based on the observation of cases, most patients who responded to CART2 and who were able to maintain a durable remission after CART2 were those who relapsed after bridging to allo-HSCT post-CART1. We hypothesized that consolidative allo-HSCT could be valuable in predicting outcomes of CART2. Among patients treated with CART2 following relapse after CART1-induced CR, those who received consolidative allo-HSCT after CART1 exhibited higher response rate, longer PFS, and OS than those who did not undergo consolidative therapy, and these improvements were more pronounced when compared with non-CART2 treatments.

Although the results must be confirmed with a larger sample size and rigorous multivariate analysis, patterns can be speculated based on the findings of other studies. Out of the 44 patients included in the retrospective study conducted by Gauthier et al. ¹⁵ , only one received allo-HSCT between CART1 and CART2, the overall objective response rate was 38.6% (CR: n = 9, PR: n = 8). The cohort included 14 patients with B-ALL who received CART2 due to NR after CART1 (n = 3), PR after CART1 (n = 2), and relapse after CR post-CART1 (n = 9). The median OS of the 14 patients was 5 months, and three patients achieved CR with a median PFS of 28 days. In the study of Holland et al. ¹⁷ , 18 children and young adults with B-ALL who received CART1 proceeded to CART2 without interim allo-HSCT, and all were in a non-remission state at the time of CART2 infusion. The objective marrow response rate of CART2 was 38.9% (7/18). The clinical outcomes of both of these cohorts of B-ALL patients is inferior to the outcome of patients who underwent consolidative allo-HSCT after CART1 in our study [response rate, 80.0% (8/10); median PFS, 7.9 months; median OS, 25.1 months], further supporting the therapeutic benefit of CART2 in this specific subset of patients.

Consolidative transplantation after CAR T-induced CR was recommended in previous studies for patients at high risk of recurrence as it can promote immune reconstitution at the stem cell level and lead to achieve deeper remission^20,25,26. This may explain the improved response to subsequent CART2. In addition, studies have found that transgenic T-cell therapy generates an immune response that suppresses T-cell expansion and persistence in vivo^22,27–29 and that the administration of allo-HSCT can mitigate this anti-CAR effect, thereby enhancing the efficacy of CAR T-cell therapy. The observation of the available data in our study revealed that seven patients who achieved CR after CART2 exhibited a significant expansion of CAR T cells, and the copy of CAR remained detectable at 1 month after infusion, with five of them undergoing consolidative allo-HSCT post-CART1. However, due to insufficient data and the absence of antibody monitoring, we could not further explore the underlying mechanisms, pending improvement in future studies. Furthermore, the chimerism ratio of patients with consolidative allo-HSCT in our study was 100%, implying that T cells recollected from recipients were allogeneic during preparation of CART2 ³⁰ , and that allogeneic recipient-derived CAR T cells have additional anti-leukemic effects triggered by T-cell receptor (TCR) recognition, which may therefore result in different clinical outcomes and safety profiles compared with autologous CAR T cells^31,32, but there are no studies comparing the outcomes of these two types of CAR T cells administered post-CART1. Further exploration is necessary to understand the mechanisms by which patients who relapsed after CART1 with consolidative allo-HSCT presented positive clinical outcomes after CART2.

Previous studies have demonstrated that various factors can influence the outcome of CART2. First, disease type and the infusion dose of CART2 have been identified as influential factors. Gauthier et al. ¹⁵ found that patients with ALL receiving CART2 had a lower response rate than those with NHL and a shorter PFS than those with CLL and NHL; they also observed that patients receiving a higher dose exhibited higher cell persistence, although the optimal dose remains unclear. We did not find a correlation between increased CART2 dose and better outcomes, which could be attributed to our small sample size, not strictly controlled infusion dose, or the negative effect of the increased dose. Second, disease status at the time of infusion may have an impact on outcomes of CART2. In our study, approximately half of the patients who were in remission (with positive MRD, n = 5) or not in remission (n = 21) at the time of infusion achieved MRD-negative CR again. Although no significant differences were observed in the proportion of responders, it has been proven that patients with a lower pre-infusion tumor burden are more likely to attain a deep response by CAR T therapy ³³ . Third, the manufacture of CART2 may also be an influencing factor. Cryopreservation can compromise the function of different T cell subsets ³⁴ and may affect the efficacy of CART2. A total of 92.3% of patients in our study received CART2 manufactured from a second leukapheresis procedure instead of cryopreserved cells stored in process for CART1. In addition, the structure of CAR, such as the scFv, vector, and co-stimulatory domain may affect the outcomes. Readministration of CAR T cells with humanized scFv after the failure of murine CAR T therapy has shown promising outcomes; there are two studies that have reported remission rates of 87.5% (7/ 8) ³⁵ and 68.4% (13/19) ³⁶ in R/R B-ALL patients who were refractory to murine-CD19 CAR T and subsequently received humanized-CD19 CAR T. Currently, a study ³⁷ has indicated that products manufactured by lentiviral vectors presented a higher incidence of cytopenia compared with retroviral vectors; however, the difference in efficacy between these two vectors remain unclear. In addition, numerous investigations have also revealed that 4-1BB-based CAR T cells exhibited more durable persistence than CD28-based CAR T cells^38–41. As patients in our study received murine CAR T cells, of whom two received CART2 containing CD28/4-1BB co-stimulatory domain transduced by retroviruses, whereas the remaining 20 received CART2 containing the CD28 co-stimulatory domain transduced by lentiviruses, the impact of different vectors and co-stimulatory domains on the outcome of CART2 needs to be further validated.

Owing to the retrospective nature, multiple confounders, and limited sample size of our study, our results may not provide a comprehensive conclusion, and significant factors may have remained unidentified in our correlation analysis; however, the insights for the selection of subsequent treatment strategies for patients who relapsed after CART1 therapy can be speculated from descriptive statistics. In addition, the type of consolidating transplantation after CART1 in our study was UCBT; it would be more beneficial to expand the scope of study to include patients who underwent BM transplantation or peripheral blood stem cell transplantation (PBSCT). Furthermore, our study focused on patients diagnosed with B-ALL, and the outcomes for other hematologic malignancies require further exploration. In all cases, a larger sample size and more rigorous multivariate analysis are required in subsequent studies, and an in-depth exploration of the underlying mechanisms must be completed using laboratory experiments and animal models.

To summarize, a second infusion of anti-CD19 CAR T-cell is safe for patients who exhibit an inferior response or relapse after an initial infusion of CD19-targeting CAR-T cells. Furthermore, CART2 demonstrated encouraging outcomes in specific populations, especially in patients with B-ALL who have undergone consolidative allo-HSCT after CART1, for whom treatment with CART2 may be a safe and effective option. However, for patients who have not undergone consolidated allo-HSCT or did not respond to CART1, the advantages and disadvantages of CART2 must be carefully evaluated.