Zhitao Ying, Ting He, Shanzhao Jin, Xiaopei Wang, Wen Zheng, Ningjing Lin, Meifeng Tu,Yan Xie, Lingyan Ping, Weiping Liu, Lijuan Deng, Yanping Ding, Xuelian Hu, Bing Bu,Xin’an Lu, Yuqin Song, Jun Zhu

1Key Laboratory of Carcinogenesis and Translational Research (Ministry of Education/Beijing), Department of Lymphoma, Peking University Cancer Hospital & Institute, Beijing 100042, China; 2 Beijing Immunochina Pharmaceuticals Co., Ltd., Beijing 100195, China; 3 Department of Medical Oncology, Shandong Cancer Hospital and Institute, Shandong First Medical University and Shandong Academy of Medical Sciences, Jinan 250117, China

Abstract Objective: Previous studies reported that 4-1BB-based CD19 chimeric antigen receptor (CAR)-T cells were more beneficial for the clinical outcomes than CD28-based CAR-T cells, especially the lower incidence rate of severe adverse events. However, the median progression-free survival (mPFS) of 4-1BB-based product Kymriah was shorter than that of CD28-based Yescarta (2.9 months vs. 5.9 months), suggesting that Kymriah was limited in the long-term efficacy. Thus, a safe and durable 4-1BB-based CD19 CAR-T needs to be developed.Methods: We designed a CD19-targeted CAR-T (named as IM19) which consisted of an FMC63 scFv, 4-1BB and CD3ζ intracellular domain and was manufactured into a memory T-enriched formulation. A phase I/II clinical trial was launched to evaluate the clinical outcomes of IM19 in relapsed or refractory (r/r) B cell non-Hodgkin lymphoma (B-NHL). Dose-escalation investigation (at a dose of 5×105/kg, 1×106/kg and 3×106/kg) was performed in 22 r/r B-NHL patients. All patients received a single infusion of IM19 after 3-day conditional regimen.Results: At month 3, the overall response rate (ORR) was 59.1%, the complete response rate (CRR) was 50.0%.The mPFS was 6 months and the 1-year overall survival rate was 77.8%. Cytokine release syndrome (CRS)occurred in 13 patients (59.1%), with 54.5% of grade 1-2 CRS. Only one patient (4.5%) experienced grade 3 CRS and grade 3 neurotoxicity.Conclusions: These results demonstrated the safety and durable efficacy of a 4-1BB-based CD19 CAR-T,IM19, which is promising for further development and clinical investigation.

Keywords: CD19 CAR-T; 4-1BB; safety; durable efficacy

Introduction

Non-Hodgkin lymphoma (NHL) is a common type of cancer around the world, accounting for about 2.8% of new cancer cases and 2.6% of cancer deaths in 2018 (1). Diffuse large B cell lymphoma (DLBCL) is the most common subtype of NHL (2). Approximately two thirds of DLBCL patients could be cured with standard frontline therapy(2,3). The standard therapy for relapsed or refractory (r/r)DLBCL patients was salvage chemotherapy, followed by autologous stem cell transplant (ASCT) (4,5). Although 50%-60% of DLBCL patients are eligible for ASCT,40%-50% tend to relapse. Thus, more effective new therapies are urgently needed for these patients (6).

CD19-directed chimeric antigen receptor-T (CAR-T)cell therapy has shown remarkable antitumor activity against r/r hematologic malignancies (7). Currently, five CD19 CAR-T products for B cell lymphomas have been approved in the world, including axicabtagene-ciloleucel(axi-cel, Yescarta), tisagenlecleucel (tisa-cel, Kymriah),brexucabtagene autoleucel (brexu-cel, Tecartus),lisocabtagene-maraleucel (liso-cel, Breyanzi), and relmacabtagene autoleucel (relma-cel) (8-13). Although these products used the same single chain variable fragment(scFv) for targeting the extracellular domain of CD19 antigen, their safety and efficacy profiles were different,partially owing to the variations in co-stimulatory domain and manufacturing technologies. The pivotal ZUMA-1 study for axi-cel with CD28 as a co-stimulatory molecule showed that the overall response rate (ORR) was 83%, with complete response (CR) rate of 58% at one month, and the median progression-free survival (mPFS) was 5.9 months(14) . In the JULIET study for 4-1BB-based tisa-cel, the ORR and CR were 52% and 40% at one month,respectively, and the mPFS was 2.9 months (15) . These data suggested that both the initial response and long-term efficacy of the 4-1BB-based product tisa-cel were limited and needed to be improved.

The safety profiles of these products, especially the severe cytokine release syndrome (CRS) and neurotoxicity(NT) rates, were also distinct (14). Grade 3 or worse CRS was 11%, 22% and 2% for axi-cel, tisa-cel and liso-cel,respectively. Regarding ≥grade 3 NT, the incidence rate was 32%, 12% and 10%, respectively. The overall probabilities of severe adverse events were higher for axicel than those of tisa-cel and liso-cel, especially NT.Recently, we compared the efficacy and adverse events of CD19 CAR-T cells with CD28 or 4-1BB co-stimulatory domain in both r/r B-NHL patients and B-cell acute lymphoblastic leukemia patients, and confirmed that 4-1BB signaling was superior to CD28 in safety profiles (16,17).Thus, 4-1BB-based CD19 CAR-T exhibited higher safety profiles for B-NHL therapy.

In this study, we designed a 4-1BB-based CD19 CAR-T(named as IM19) that was composed of FMC63 scFv, hinge and transmembrane domain of CD8α, 4-1BB costimulatory domain, and CD3ζ intracellular domain. CART cells were manufactured into a memory cell-enriched formulation. We investigated the safety and efficacy of IM19 in 22 patients with r/r B-NHL in a phase I/II trial.Our data demonstrated that IM19 had great potential in safety and long-term efficacy in r/r B-NHL.

Materials and methods

Eligibility criteria

Patients with CD19-positive r/r B-NHL who previously received at least one line of treatment were recruited(Supplementary Table S1). The inclusion criteria include: 1)patients with DLBCL, follicular lymphoma (FL), or primary mediastinal B-cell lymphoma (PMBCL) who are refractory or have relapsed after at least two previous lines of treatment; or who have relapsed after transplantation; 2)patients with mantle cell lymphoma who have received at least one line of treatment; 3) patients with evaluable disease lesions; 4) age ≥18 years; 5) life expectancy ≥90 d; 5)Eastern Cooperative Oncology Group (ECOG)performance status of 0-2; 6) females of childbearing potential with a negative blood pregnancy test; and 7) those who voluntarily participate in the trial and sign the informed consent form (ICF). Exclusion criteria are as follows: 1) patients with high-risk organ involvement:tumors invade one of the following organs: central nervous system, gastrointestinal tract, lungs, pericardium and large vessels; 2) those who have a graft-versus-host disease and need to use immunosuppressive agents; or who have an autoimmune system disease; 3) those who received chemotherapy or radiotherapy within 3 d before peripheral blood mononuclear cell (PBMC) collection; 4) those who have used systemic steroids within 5 d prior to PBMC collection (except for recent or current use of inhaled steroids); or 5) patients with active Hepatitis B or Hepatitis C virus, human immunodeficiency virus (HIV) or other untreated active infections.

Study design

This was a single center phase I/II clinical study(ClinicalTrials.gov identifier: NCT03528421) to evaluate the safety and efficacy of IM19 in patients with r/r B-NHL.To assess the dose-limiting toxicity (DLT) of IM19, three dose levels (5×105/kg, 1×106/kg and 3×106/kg) were assessed. Once DLT occurred in more than 2 patients, the recruitment should be terminated at this dose level. This study was conducted in accordance with the Declaration of Helsinki and was approved by the Institutional Review Board of Peking University Cancer Hospital (IRBPUCH No. 2018YJZ03). All participants were informed of the possible risks and side effects of the therapy and provided the signed informed consent.

A schematic design of the clinical trial is shown inFigure 1. Apheresis was performed in all eligible patients.After pretreatment with fludarabine [25 mg/(m2·d)] and cyclophosphamide [250 mg/(m2·d] for 3 consecutive days,patients were infused with IM19 on d 0. The primary endpoints were 3-month ORR and the incidence of DLT.The secondary objectives included CR, partial response(PR) and ORR on d 28 and at month 6. The endpoints such as hematology, blood biochemistry, C reactive protein(CRP), CRS and NT were closely monitored especially within one month. On d 1, 4, 7, 10, 14, 21, 28, 90 and 180,the peripheral blood samples were collected to test the CAR-T and cytokines using flow cytometry.

Vector construction and lentivirus production

IM19 CAR was composed of FMC63-derived scFv, hinge and transmembrane domains of CD8α, and intracellular domains of 4-1BB and CD3ζ. The CAR gene was ligated to an EF1α promoter-based lentiviral transfer plasmid pLenti6.3/V5 (ThermoFisher, Waltham, MA, USA). The transfer plasmid, packaging plasmids (pLP1 and pLP2,ThermoFisher, Waltham, MA, USA) and envelope plasmid(pLP/VSVG, ThermoFisher, Waltham, MA, USA) were co-transfected into suspension 293 cells using polyethyleneimine (Polysciences, Warrington, PA, USA).After purification from the culture medium using Core 700 chromatography (GE Healthcare, Chicago, IL, USA), the IM19 CAR lentiviral vectors were formulated,cryopreserved, and sent for quality control in Immunochina Pharmaceuticals Co., Ltd. (Beijing, China).

Figure 1 Schematic overview of clinical trial. (A) Patient cohorts;(B) Clinical protocol. CAR-T, chimeric antigen receptor-T.

Manufacture of IM19 CAR-T cells

IM19 CAR-T cells were manufactured by Immunochina Pharmaceuticals Co., Ltd. (Beijing, China). Briefly,PBMCs were isolated from the apheresis using a density gradient medium Ficoll (GE Healthcare, Chicago, IL,USA). T cells were purified and activated by CD3/CD28 Dynabeads (ThermoFisher, Waltham, MA, USA) at a T cell/bead ratio of 1:1. One day later, T cells were transduced with IM19 lentiviral vector and cultured in XVIVO 15 medium (Lonza Group, Basel, Switzerland)containing 100 U/mL of IL-2 at a density of 1.5×106/mL for 7-11 d. Then CAR-T cells were formulated,cryopreserved, and sent for quality control in Immunochina Pharmaceuticals Co., Ltd.

Flow cytometry analysis of CAR-T cells

Flow cytometry analysis was performed using NovoCyte 2060R (ACEA Biosciences, San Diego, CA, USA)according to previous report (17). The transduction efficiency of IM19 CAR gene into T cells was detected using a phycoerythrin (PE)-labeled anti-CAR antibody which was developed by Immunochina Pharmaceuticals for recognizing the scFv of IM19. PE-anti-CD4 and FITCanti-CD8 antibodies (BD Biosciences, San Jose, CA, USA)were used to detect the CD4/CD8 ratio. The APC-anti-CD45RA and FITC-anti-CD62L antibodies (Biolegend,San Diego, CA, USA) were used to evaluate the ratio of memory T cells.

Cytometric bead array assay of cytokines

The blood was sampled at the follow-up time points. The protein levels of interleukins (IL-2, IL-4, IL-6, IL-10 and IL-17A), interferon (IFN)-γ and tumor necrosis factor(TNF) were evaluated using the Cytometric Bead Array(CBA) Human Th1/Th2/Th17 Cytokine Kit (BD Biosciences, San Jose, CA, USA) and flow cytometry analysis. All assays were carried out according to the manufacturer’s protocol. Flow cytometry was used to detect the cytokine levels in serum samples. The serum data were expressed in picograms per milliliter.

Evaluation of clinical outcomes

The patients were followed up, with adverse events evaluated according to the clinical symptoms, hematology,blood biochemistry and immune indexes. NT was graded according to National Cancer Institute (NCI) Common Terminology Criteria for Adverse Events (CTCAE) (18).CRS was assessed in line with the Consensus Criteria of American Society for Transplantation and Cellular Therapy (ASTCT) (19). The tumor burden was assessed using positron emission tomography-computed tomography(PET-CT) for response evaluation based on the revised criteria for response assessment (20).

Statistical analysis

Statistical analyses were performed with GraphPad Prism(Version 8.0, GraphPad Software, Inc., La Jolla, CA, USA)and IBM SPSS (Version 22.0, IBM Corp., New York,USA). A two-sided Wilcoxon or Mann-Whitney U test was conducted for a comparison of two groups. D’Agostino-Pearson test and Anderson-Darling test were used for normality tests. For three or more groups, statistical analysis was based on one-way ANOVA and Kruskal Wallis test. Kaplan-Meier curve was used to analyze the survival of patients. A threshold of P＜0.05 was considered statistically significant for all analyses.

Results

Characteristics of patients and IM19 CAR-T cells

From June 2018 to October 2019, a total of 26 patients with r/r B-NHL were recruited, and 22 eligible patients aged 28-69 years were screened. A dose-escalation study was first performed to evaluate DLT and the optimum dose in 15 patients at the dose levels of 5×105/kg (3 patients), 1×106/kg (6 patients) and 3×106/kg (6 patients).Then an extensive investigation of the optimum dose was carried out in 7 patients at the dose level of 1×106/kg(Figure 1A,Table 1). The baseline characteristics of patients are summarized inTable 1, and features of patients are analyzed inSupplementary Table S1. Three lymphoma subtypes were present in this study, including DLBCL(13/22), FL (7/22) and marginal zone B-cell lymphoma(MZL) (2/22). And 90.9% (20/22) of patients received at least 2 lines of therapies before enrollment.

After PBMCs were isolated from the leukapheresis, T cells were purified, activated and transduced with the IM19 lentiviral vectors. The manufacturing time was 6-18 d.When expanding to the dose level, cells were evaluated for the CAR expression ratio, CD4/CD8 ratio and the percentage of memory T cells (Supplementary Table S2).Under our manufacturing process, the percentage of na?ve,stem cell memory, central memory and effector memory T cells in almost all final IM19 products remained at high levels (Supplementary Table S2). On d 0, patients were intravenously infused with IM19 CAR-T cells, and were closely followed up from d 4 to d 28 for safety evaluation.The response was assessed on d 28, 90 and 180 (Figure 1B).

Safety of IM19 CAR-T cells

As of March 1st, 2020, the cut-off date, 22 patients were infused with IM19 CAR-T cells. The adverse events associated with CAR-T cells were monitored. CRS was observed in 59.1% (13/22, 95% CI: 38.7%-76.8%) of patients among whom one patient (4.5%) showed both grade 3 CRS and grade 3 NT at the dose level of 3×106/kg(Figure 2A,Supplementary Table S3). Tocilizumab and dexamethasone were used for the toxicity management. All the adverse events resolved within 2 months. All patients showed tolerability to IM19 at three dose levels, and the DLT was not reached.

To further evaluate the adverse events, we assessed the hematology, blood biochemistry and blood inflammatory cytokines at the follow-up time point. Almost all patients experienced cytopenia and 8 patients showed decreased platelet count after treatment with IM19. Multiple cytokines, especially IL-6, IL-10 and IFN-γ, increased in 12 patients from d 4 to d 28 (Figure 2B-D). The patient who experienced severe CRS and NT showed much higher levels of IL-6, IL-10 and IFN-γ (Figure 2).

Clinical responses of IM19 CAR-T cells

One month and three months after treatment with IM19,patients underwent PET-CT scanning for response evaluation. Decrease in sum of the products of diameters(SPD) was observed in 16 patients, and the ORR was 72.7% (16/22, 95% CI: 51.5%-87.1%) and CR was 40.9%(9/22, 95% CI: 23.2%-61.3%) on d 28 ((Figure 3A,B),Table 2). The 3-month ORR and CR were 59.1% (13/22,95% CI: 38.7%-76.8%) and 50.0% (11/22, 95% CI:30.7%-69.3%), and the 6-month ORR and CR were 50.0% (11/22, 95% CI: 30.7%-69.3%) and 45.5% (10/22,95% CI: 26.9%-65.4%), respectively (Table 2). Nine patients (40.9%) showed ongoing response even after 1year (Figure 3A). Evident eradication of tumors and durable effectiveness were observed in CR patients, and representative PET-CT images are shown inFigure 3C. Six patients were evaluated as progressive disease (PD) on d 28(Figure 3A,B). For patients with aggressive NHL, the ORR was 62.5% (10/16, 95% CI: 38.5%-81.6%) and CR was 31.3% (5/16, 95% CI: 13.9%-55.9%) on d 28 (Table 2).The median PFS was 6 months after IM19 infusion (Figure 3D), and 1-year overall survival (OS) was 77.8% (Figure 3E).

Table 1 Patients’ characteristics

Correlation of IM19 CAR-T characteristics with clinical outcomes

We further evaluated the potential factors which were associated with the safety and clinical responses. The levels of IM19 CAR-T cells in peripheral blood were analyzed at the follow-up time point.In vivoexpansion and high levels of IM19 from d 4 to d 14 were observed in all patients, and then decreased and persisted at a low level. The IM19 CAR-T could be detected in some of the patients even beyond 8 months (Figure 4A). The ratio of CD4/CD8 in final IM19 products was not significantly correlated with the CRS grades. The number of CAR-T cells was proportional to the severity of CRS (Figure 4B).Importantly, the clinical responses (CRvs.PR; PRvs.PD)at month 3 were apparently associated with the peak level of CAR-T cells in peripheral blood, although there were only 2 patients in the PR group (Figure 4C). Particularly,the median area under the curve of IM19 levels within 28 d(AUC-28) in patients of CR and PR groups was more than 10-fold higher than that in PD group (Figure 4D). In addition, the transduction efficiency of CAR and the CD4/CD8 ratio in final IM19 products showed no distinct correlation with clinical responses at month 1 or month 3(Figure 4C), and they were consistently not correlated with the peak level of CAR-T cells in peripheral blood(Figure 4E).

Figure 2 Evaluation of CRS and cytokine expression in patients after CAR-T cell infusion. (A) CRS grade in each patient; (B-D) time course of serum IL-6 (B), IL-10 (C) and IFN-γ (D) in 22 patients after CAR-T cell therapy. The first day of CD19 CAR-T cell infusion was regarded as d 0. CRS, cytokine release syndrome; CAR-T, chimeric antigen receptor-T.

Discussion

Anti-CD19 CAR-T cells have demonstrated impressive clinical outcomes in B cell malignancies, and the functions of CAR-T cells incorporating with different co-stimulatory domains vary in toxicity, safety and persistence. It was reported that 4-1BB-based CAR-T cells tended to induce a central memory phenotype, possibly due to the enhanced fatty acid metabolism (21,22). Although previous studies showed that 4-1BB signaling was beneficial for the safety andin vivopersistence of CD19 CAR-T cells, the durable effectiveness of 4-1BB-based CD19 CAR-T cells should be of concern based on the reported clinical data (9,10).

In this study, we developed anti-CD19 CAR-T cells with 4-1BB domain, and investigated the response and longterm outcomes in patients with r/r B-cell NHL. Compared with US FDA-approved CD19 CAR-T products, our CD19 CAR-T cells demonstrated durable clinical responses and lower NT rates in r/r B-NHL patients. The ORR and CR rate of Kymriah in large B cell lymphoma(LBCL) patients were 52% and 40%, respectively. Twentytwo percent and 12% of the patients experienced grade ≥3 CRS and grade ≥3 NT, respectively (10). The ORR and CR rate of Yescarta in LBCL patients were 83% and 58%,respectively. Yescarta-related severe CRS and NT were observed in 11% and 32% of patients, respectively (14).Recently approved Breyanzi achieved an ORR of 73% and a CR of 53% in LBCL patients. Severe CRS and NT occurred in 2% and 10% of patients who received Breyanzi therapy (11). Moreover, the mPFS was 2.9, 5.9 and 6.8 months for Kymriah, Yescarta and Breyanzi, respectively.In comparison, our IM19 CAR-T cells showed only 4.5%of severe CRS and NT, indicating a favorable safety profile for this 4-1BB-based CD19 CAR-T. Importantly, the mPFS of IM19 was 6 months, similar to that of Yescarta.Our investigation demonstrated that IM19 was a safe and long-term effective CD19 CAR-T.

Figure 3 Clinical response of IM19 CAR-T cells in enrolled patients. (A) Clinical response in each patient after CAR-T cell infusion; (B)evaluation of tumor burden of 22 patients after CAR-T cell infusion; (C) PET-CT images of two representative r/r NHL patients who showed an ongoing CR after IM19 CAR-T therapy; (D) PFS rate of patients after IM19 infusion; (E) OS rate of patients after IM19 infusion. CAR-T, chimeric antigen receptor-T; r/r NHL, relapsed or refractory non-Hodgkin lymphoma; PET-CT, positron emission tomography-computed tomography; CR, complete response; PFS, progression-free survival; OS, overall survival; SD, stable disease; PD,progressive disease; PR, partial response; SPD, sum of the products of diameters.

Table 2 Response rates of 22 patients after IM19 CAR-T cell infusion

Figure 4 Kinetics of IM19 CAR-T cells in patients and correlation with clinical outcomes. (A) In vivo kinetics of IM19 CAR-T cells in different response groups; (B) correlation analysis of CD4/CD8 ratio and CAR-T cell peak in peripheral blood and the CRS grade; (C)correlation analysis of the peak level of CAR-T cells in peripheral blood (left), transduction efficiency (median), and the ratio of CD4:CD8(right) with different clinical response groups at month 1 and month 3; (D) comparison of the AUC-28 between CR+PR groups and PD group. AUC-28, area under the curve of CAR-T cell expansion from d 0 to d 28; (E) correlation of CAR-T cells in peripheral blood with transduction efficiency (left) or with the ratio of CD4:CD8 (right). ＊＊, P＜0.01; ＊＊＊, P＜0.001; ＊＊＊＊, P＜0.0001. CAR-T, chimeric antigen receptor-T; CRS, cytokine release syndrome; CR, complete response; PR, partial response; PD, progressive disease.

The striking clinical outcomes of IM19 were probably attributed to our manufacturing technologies that generated a T cell formulation enriched with na?ve and memory T cell phenotypes (Supplementary Table S2).Gargettet al.reported that the manufacturing conditions for CAR-T cells could affect the phenotype and functions of CAR-T cells (23), supporting the importance of manufacturing technologies and our findings. Moreover,our data showed that the clinical responses were correlated with thein vivoexpansion kinetics and peak level of IM19 cells. The majority of T cells in IM19 products skewed toward to high frequency of central memory, effector memory, na?ve and stem-like memory T cells(Supplementary Table S2). These superior phenotypes of CAR-T cell products were reported to be correlated with sustained remissions in patients (24). Recent studies have revealed that the long-lived CAR-T cells with more na?ve/stem/central memory-like T cells (Tscm)/Tcm phenotype achieved optimal control in tumors (21,25).Thus, an enriched proportion of na?ve and memory T cell phenotypes in IM19 formulation was probably attributed to thein vivopersistence and long-term efficacy.

In addition to the memory phenotype, previous studies had also focused on CD4/CD8 ratio in CAR-T cells. It was demonstrated that a defined ratio of CD4+and CD8+subset of CAR-T cells conferred superior antitumor activity in patients with ALL and B-NHL (25-27).However, our study did not find significant correlation between CD4/CD8 ratio and clinical outcomes including CRS severity, clinical response andin vivoCAR-T kinetics(Figure 4B,C,E).

Conclusions

IM19 with 4-1BB-based co-stimulatory domain showed an effective and durable antitumor activity, as well as very low toxicities for B cell lymphoma. The clinical responses were associated with thein vivoexpansion profile of CAR-T cells. Our study demonstrated the safety and durable efficacy of IM19 with FMC63 scFv, 4-1BB and CD3ζ intracellular domain, which is promising for further development and clinical investigation.

Acknowledgements

This study was supported by the Beijing Natural Science Foundation (No. 7202026) and Capital’s Funds for Health Improvement and Research (No. 2020-2Z-2157).

Footnote

Conflicts of Interest: The authors have no conflicts of interest to declare.

Table S1 Baseline characteristics of all 22 patients

Table S2 Characteristics of IM19 CAR-T cell products

Table S3 AEs after CAR-T cell infusion in 22 patients with B cell lymphoma

Table S3 (continued)

Table S3 (continued)