A novel generation 1928zT2 CAR T cells induce remission in extramedullary relapse of acute lymphoblastic leukemia

The trial (ClinicalTrials.​gov number, NCT02822326) was designed to assess the safety and feasibility of infusing 1928zT2 T cells in patients with relapsed or refractory B-ALL and was approved by the Ethics Committee of Guangdong General Hospital. Here, we reported the first three enrolled ALL patients with extramedullary involvement (Table 1).

Patient 1 was a 34-year-old female diagnosed as B-ALL (CD19+, BCR/ABL-) in April, 2015 (Fig. 3a). Although she had no response to chemotherapy regimen of VDLCP at first, the patient achieved CR after Hyper CVAD A therapy. She received four cycles of chemotherapy and underwent allogeneic hematopoietic stem cell transplantation (allo-HSCT) from her 10/10 HLA allele-matched sister in November, 2015. However, 9 months later, she had a relapse in extramedullary tissues including her left breast and multiple lymph nodes identified by Positron emission tomography-computed tomography (PET/CT) (Fig. 3b). The extramedullary leukemia in breast was confirmed histologically (Fig. 3c), and leukemia blast cells were detected as positive for TdT, CD19, CD20, CD79a, CD34, CD99, CD10, PAX5, and Ki67 (15%), and negative for CD3 and Cyclin D1. B-mode ultrasound was used to monitor the tumor mass in the left breast, and about 2.8 × 1.6 cm size of an inhomogeneous hypo-echoic mass was identified (Fig. 3d). No evidence of relapse in BM and CNS was observed with persisted complete donor chimerism or negative minimal residual disease.

Patient 2 was a 15-year-old male also diagnosed as B-ALL (CD19+, BCR/ABL-) in October, 2014 (Fig. 4a). He underwent allo-HSCT from his 10/10 HLA allele-matched sibling in June, 2015, and unfortunately had a relapse in CNS 6 months later. Then, he achieved a second CR after intrathecal chemotherapy (IT), irradiation, and donor lymphocyte infusion (DLI). However, the leukemia recurred again 10 months later with BM and extramedullary involvement. The BM smear showed typical leukemic blasts account for 15% (Fig. 4b). The PET/CT revealed that an abnormal intense high metabolic region beneath the kidney capsule, indicating the extramedullary relapse in kidney (Fig. 4c). Although he received chemotherapy regimen of VDLCP, he had no response at all.

Patient 3 was a 20-year-old male with a third recurrence of ALL with CNS-3 status [14] and extramedullary involvement. He received a diagnosis of ALL (CD19⁺, BCR/ABL-) in 2000 when he was 3 years old (Fig. 5a). During the following 13 years, he had received more than 20 courses of intensive systemic chemotherapy (SC) and IT and remained being in CR. However, he had a relapse in BM and CNS in 2013. Then, he underwent allo-HSCT from a 10/10 HLA allele-matched unrelated female donor in 2014 and the second CR was achieved. Despite he received IT for more than six courses, the patient suffered from second recurrence with CNS-ALL. He had a good response to IT and irradiation and achieved CR for a third time. However, the cancer recurred again 10 months later with CNS and extramedullary involvement in June, 2017. The bone marrow examination still showed signs of complete remission. The whole-body PET/CT showed that extensive extramedullary relapse involving almost the entire pleura, peritoneum, pelvic fascia, pancreas, and multiple lymph nodes (Fig. 5b). A biopsy was performed on right-sided supraclavicular node mass, and characteristic megakaryocytes, erythroblasts and myeloid cells were observed microscopically. The markers were identified as CD3−, CD20−, CD19+, TdT+++, CD10+++, CD34−, and Ki67+ (90%), revealing the leukemic involvement (Fig. 5c). A palpable mass below xiphoid was identified by B-mode ultrasound, and inhomogeneous hypo-echoic mass about 3.0 × 1.9 cm in diameter was seen (Fig. 5d). In addition, aberrant blast cells were detected histologically (Fig. 5e) and examined CD19⁺CD10⁺ phenotype (Fig. 5f) in the CSF obtained by lumbar puncture. X/Y fluorescent in situ hybridization (FISH) showed that XY-type recipient cells accounted for 50% of the cerebrospinal cells, further indicating a relapse in the CSF (Fig. 5g).

Patients underwent an apheresis to obtain sufficient PBMCs for producing 1928zT2 T cells. The manufacture of 1928zT2 was completed in about 14 days. The cells were released for infusion after safety check. All patients received a lymphodepletion conditioning regimen which contained with fludarabine 25 mg/m², cyclophosphamide 300 mg/m², and/or plus other drugs. After 1-day rest, they received a split-dose infusion of 1928zT2 T cells (a total of 5 × 10⁴–1 × 10⁶ cells/kg and 10% on day 1, 30% on day 2, and 60% on day 3). No post-infusion cytokines such as IL-2 were administered. These patients were evaluated for responses and toxic effects. The expansion and persistence of circulating 1928zT2 T cells were also monitored. Adverse events such as CRS after 1928zT2 T cell infusion were graded according to National Institutes of Health criteria (Common terminology Criteria for Adverse Events, version 4) [8], and clinical responses were evaluated and defined as complete remission (CR), CR with incomplete count recovery (CRi), partial response (PR), stable disease (SD), or progressive disease (PD) according with the National Comprehensive Cancer Network (NCCN) guidelines. MRD negative was defined as the absence of leukemia cells in BM determined by flow cytometry, and the absence of fusion gene in BM determined by qPCR.

The 1928zT2 T cells were generated as previously described [14]. Briefly, the 19-28z-TLR2 lentiviral vector was established with 19-28z-TLR2, a CAR-linking FMC63-scFv, CD28 transmembrane and endodomain, the CD3ζ signaling domain, and the TIR domain (amino acid 639-784) of TLR2. Primary human T lymphocytes were isolated and collected, then stimulated with magnetic beads coated with anti-CD3/anti-CD28 antibodies (Miltenyi Biotec) at cell to bead ratio of 1 to 2. Approximately 72 h after activation, T cells were transfected with 19-28z-TLR2 lentiviral vector. These T cells were fed every 2 days with fresh media and used within 14 days of expansion.

The percentages of transduced GFP-positive T cells, CD19 on leukemia cells, T cell phenotypes of the patient were determined by flow cytometry. The antibodies for CD3 (APC), CD4 (Percp-cy5.5), CD8 (PE), CD19 (APC), CD20 (PE), and CD22 (Percp-cy5.5) were used. The flow cytometry was performed with a FACSCalibur flow cytometer (BD Biosciences) and analyzed using FlowJo software.

The target cells K562, K562-CD19, and NALM6 were tagged both GFP and luciferase (GL), specifically incubated with GFP T, 1928z T or 1928zT2 T cells at the indicated ratios in triplicate wells in U-bottomed 96-well plates. Viability of target cells was monitored 18 h later by adding 100 dl/well substrate D-luciferin firefly (Life Sciences) resolved at 150 μg/ml. Background luminescence was negligible (< 1% than the signal from the wells with only target cells). The viability percentage was calculated as experimental signal/maximal signal × 100%, and killing percentage was equal to 100 − viability percentage.

To develop the mouse subcutaneous tumor xenograft model of human extramedullary ALL leukemia [15], NSI mice of ages 6 to 10 weeks were used. All procedures were performed under approval of the Institutional Animal Care and Use Ethics Committee. NALM6-GL (2 × 10⁵ cells in 200 μl PBS) was injected subcutaneously, and 2 × 10⁶ lentiviral transduced human T cells in 200 μl PBS were adoptively transferred to these mice systemically by tail vein injection 2 days later. In vivo whole body imaging of luciferase-labeled cells was performed by cooled CCD camera system (IVIS 100 Series Imaging System, Xenogen, Alameda, CA). D-Luciferin firefly, potassium salt (75 mg/kg) was injected, and the mice were imaged. Quantification of total and average emissions was performed using the Living Image software (Xenogen). The mice were killed at day 34, and their subcutaneous tumor masses were extracted and weighted.

IL-6 levels in plasma were assessed using enzyme-linked immunosorbent assay (ELISA kit, eBioscience, USA) according to the manufacturer’s instructions. The immune-related cytokines were analyzed using Luminex MAGPIX system (Luminex Corp., Austin, TX) according to the manufacturers’ specifications. Samples were detected in triplicate relative to standards supplied by the manufacturer and analyzed for significant differences between different groups.

mRNA was extracted from cells with Trizol (Qiagen) and reverse transcribed into cDNA using the PrimeScript™ RT reagent Kit (Takara). All reactions were performed with TransStart Tip Green qPCR SuperMix (Transgene) on a Bio-Rad CFX96 real-time PCR machine. Delta CT calculations were relative to β-actin and were corrected for PCR efficiencies.

Statistical analysis was performed with SPSS software version 19.0 (Inc., Chicago, IL, USA). Student’s t test (two-tailed) or one-way analysis of variance was used, and data were presented as means ± SEM. A P value < 0.05 was considered significant.

We firstly evaluated the in vitro cytotoxic responses of 1928zT2 T cells exposed to NALM-6 cells (a CD19+ human B-ALL cell line), K562-GL cells with or without CD19 expression. As showed in the Fig. 1, the 1928zT2 T cells generated from PBMC of ALL patient lysed more than 75% CD19-expressing NALM-6 cells at all indicated E:T ratios from 1:1 to 1:16, whereas less than 50% NALM-6 target cells were killed by 1928z T cells at E:T ratio of 1:16, indicating that the cytotoxicity of 1928zT2 T cells against target cells is more potent than that of 1928z T cells. Consistently, the results of killing assay with K562 cells revealed that both 1928z T and 1928zT2 T cells specifically killed the K562-CD19 cells but not the K562 cells without CD19 expression. Notably, the killing percentages of 1928zT2 T cells were higher than those of 1928z T cells at low E:T ratios. Thus, the 1928zT2 T cells exhibited enhanced anti-leukemic effect in vitro.

To further address the potential therapeutic application of 1928zT2 T cells in extramedullary leukemia, we next examined whether these cells harbored cytolytic activity against subcutaneous leukemia ex vivo. The cell line-derived xenografts were established in the immunodeficient NSI mice by subcutaneous injection of 2 × 10⁵ NALM-6 cells that were tagged with GFP and luciferase. Two days later, these xenografts mice were administrated with one intravenous infusion of 2 × 10⁶ GFP T, 1928z T cells or 1928zT2 T cells (Fig. 2a). The BLI results revealed that tumors were significantly suppressed in mice treated with 1928zT2 T cells compared with those received GFP T cells (Fig. 2b), highlighting the potent efficacy of these novel 1928zT2 T cells against extramedullary leukemia. Conversely, there was no significant difference of tumor burden between 1928z T cells and GFP T cells. The tumors from mice that were treated with 1928zT2 T cells were smaller than that from control groups (Fig. 2c). Collectively, 1928zT2 T cells significantly reduced tumor burden in subcutaneous leukemia-bearing xenografts.

The infusion of 1928zT2 T cells was well tolerated without any immediate adverse effects in these three patients. Strikingly, the size and hypometabolism of the first patient’s breast lump was reduced obviously, detected by PET/CT scanning on day 30 after 1928zT2 T cells infusion (Fig. 3b). The patient achieved complete remission on day 46 post infusion. Strikingly, the B ultrasound results showed that her breast lump disappeared compared to the baseline (Fig. 3d). PET/CT and B ultrasound restaging exhibited no evidence of relapse at all during the 12-month follow-up (Fig. 3b, d).

Patient 2 also had a complete remission in BM with no blast cells were observed in the BM smear 10 days after 1928zT2 T cells infusion (Fig. 4b). Six months later, restaging PET/CT revealed the absence of the abnormal intense high metabolic region, suggesting a remission in the kidney (Fig. 4c). This patient remained being in CR and received the second allo-HSCT from an unrelated donor in April 2017.

In patient 3, the visible tumor mass below xiphoid receded gradually beginning on day 3 and disappeared identified by B-mode ultrasound on day 18, indicating the 1928zT2 T cells started to eradicate B-ALL cells rapidly (Fig. 5d). Two months later, restaging PET-CT also showed complete remission in all extramedullary lesions (Fig. 5b). The patient markedly achieved complete clearance of blasts in CSF and relief of symptoms on day 18 post 1928zT2 T cell infusion (Fig. 5e) and obtained recovery of thrombocytopenia and was entitled to receive the lumbar puncture. No CD19⁺CD10⁺ blasts were detected in CSF (Fig. 5f), and X/Y FISH showed that the remaining recipient cells in CSF were 100%XX from the donor (Fig. 5g), suggesting a complete replacement by donor cells and remission in CSF. Collectively, 1928zT2 T cells exhibited potent anti-leukemia effects in these three ALL patients with extramedullary involvement.

We used the flow-cytometry analysis to detect the GFP-positive 1928zT2 T cells in PBMCs in these three patients. In patient 1, the robust 1928zT2 T cells expansion occurred around day 12 and disappeared by day 92 (Fig. 6a). In addition, CD19+ cells were reduced and became absent within 10 days post the 1928zT2 T cells infusion (Fig. 6a). Then, a continuous rise of IgG level was detected (Fig. 6a), suggesting a slow recovery of B cells in IgG level. For patient 2, 1928zT2 T cells in circulation appeared and began to increase on day 10 after infusion and disappeared on day 50 (Fig. 6b). Similarly, 1928zT2 T cells in BM began to expand on day 10 and retreated on day 50 (Fig. 6c). The flow cytometry analysis results of patient 3 revealed that 1928zT2 T cells in circulation became detectable on day 6 after infusion, and continued increasing to a peak that the percentage of 1928zT2 T cells was as high as 57% in PBMC on day 12, then decreased rapidly and became undetectable after day 17 (Fig. 6d). Since the ALL was relapsed in CNS and extramedullary without BM involvement, MRD- statue in BM was persistently observed. In addition, the 1928zT2 T cells were identified in CSF with 1.84%, indicating the potential ability of these cells traffic to blood-brain barrier (Fig. 6e). No detection of blast cells in CSF also suggested that 1928zT2 T cells exhibited robust immunosurveillance effects in CNS (Fig. 6e).

The 1928zT2 T cells infusion was generally well tolerated in all these three patients. No immediate adverse side effects, episodes of tumor lysis syndrome, autoimmune disorder, and Graft-versus-host disease were observed. However, the patient 1 experienced grade 2 CRS on day 7 post infusion, with fever and left breast pain (Fig. 6g). Her clinical presentation was resolved spontaneously on day 13. The level of IL-6, the main contributor of the CRS [16], increased to a peak on day 11 and decreased sharply on day 12, before the CRS remised (Fig. 6g). We also found the level of C reactive protein (CRP), a common inflammation marker [17], followed a similar pattern of IL-6 (Fig. 6g). She experienced a transient neutropenia on day 14 and day 15, and no other changes in hemogram were observed (Fig. 7a). Patient 2 also experienced CRS with grade 2 between day 9 and day 12, when the levels of IL-6, CRP, and PCT were significantly increased (Figs. 6h and 7b). The patient suffered fever and bone pain. This significant but transient toxicity resolved spontaneously. Patient 3 was diagnosed with grade 3 CRS. He started to have a persistent fever on day 7 and emerged refractory hypotension on day 9, which was alleviated after administrating two usages of tocilizumab (4 mg/kg/d), and was resolved on day 15 (Fig. 6i). The level of IL-6 reached a peak on day 9 concurrently (Fig. 6i) and then declined along with the administration of tocilizumab. The variation tendency of CRP was also accompanied. This patient had an increase in absolute neutrophil count, absolute lymphocyte count, and platelets from the baseline due to the lymphodepletion regime (Fig. 7c). These elevations were also coincident with the expansion of 1928zT2 T cells, suggesting that CAR T cells play a role in the recovery of blood picture.

Several cell adhesion and inflammation genes of patient 1 were also monitored by qPCR, and the increased expression pattern is shown in Fig. 7d. The patient was in a relatively stable state without any additional support. The serum cytokines of patient 3 were monitored using luminex assays, and the levels of IL-22, MCP-1, IL-6, and HGF increased prominently from day 6 to day 17 (Fig. 7e), coincident with the expansion of 1928zT2 T cells in the blood. The concentrations of other 34 tested cytokines including IFN-γ, IL-2, TNF-a, and IL-10 did not change much (Fig. 7e). None of these three patients presented neurotoxicity evidence.