Donor-derived CD19-targeted T cells cause regression of malignancy persisting after allogeneic hematopoietic stem cell transplantation

Abstract

New treatments are needed for B-cell malignancies persisting after allogeneic hematopoietic stem cell transplantation (alloHSCT). We conducted a clinical trial of allogeneic T cells genetically modified to express a chimeric antigen receptor (CAR) targeting the B-cell antigen CD19. T cells for genetic modification were obtained from each patient's alloHSCT donor. All patients had malignancy that persisted after alloHSCT and standard donor lymphocyte infusions (DLIs). Patients did not receive chemotherapy prior to the CAR T-cell infusions and were not lymphocyte depleted at the time of the infusions. The 10 treated patients received a single infusion of allogeneic anti-CD19-CAR T cells. Three patients had regressions of their malignancies. One patient with chronic lymphocytic leukemia (CLL) obtained an ongoing complete remission after treatment with allogeneic anti-CD19-CAR T cells, another CLL patient had tumor lysis syndrome as his leukemia dramatically regressed, and a patient with mantle cell lymphoma obtained an ongoing partial remission. None of the 10 patients developed graft-versus-host disease (GVHD). Toxicities included transient hypotension and fever. We detected cells containing the anti-CD19-CAR gene in the blood of 8 of 10 patients. These results show for the first time that donor-derived allogeneic anti-CD19-CAR T cells can cause regression of B-cell malignancies resistant to standard DLIs without causing GVHD.

Trial registration: ClinicalTrials.gov NCT01087294.

Figure 1

An anti-CD19 CAR gives T cells the ability to specifically recognize CD19. (A) The CAR incorporated an anti-CD19 single chain Fv (scFv), part of the CD28 costimulatory molecule, and the cytoplasmic part of the CD3ζ molecule. (B) PBMCs from healthy allogeneic donors were activated with an anti-CD3 monoclonal antibody and transduced with gammaretroviruses encoding an anti-CD19 CAR. The cells were infused after 8 days of culture. (C) Expression of the CAR was detected on the surface of the infusion cells of patient 1 on day 7 of culture by staining with anti–antigen-binding fragment (Fab) antibodies. Staining with isotype-matched control antibodies is also shown. The plots are gated on CD3⁺ lymphocytes. (D) The infusion cells of patient 1 were stained for CD45RA and CC chemokine receptor 7 (CCR7). Staining with isotype-matched control antibodies is also shown. The staining was performed on day 7 of culture, and the plots are gated on CAR-expressing CD3⁺ lymphocytes. (E) On day 8 of culture, infusion cells of patient 1 were cultured for 4 hours with either the CD19⁺ cell line CD19-K562 or the negative-control cell line NGFR-K562 that does not express CD19. CD19-specific upregulation of CD107a occurred. Plots are gated on CD3⁺ lymphocytes. Similar results were obtained with the infusion cells of all patients treated. (F-H) A representative example of CD19-specific cytokine production by anti-CD19-CAR T cells is shown. Infusion cells of patient 4 were stimulated with either CD19-K562 cells or NGFR-K562 cells for 6 hours on day 8 of culture. Intracellular cytokine staining followed by flow cytometry revealed that large fractions of the T cells produced interferon γ (IFN-γ), tumor necrosis factor (TNF), and interleukin (IL) 2 in a CD19-specific manner. Plots are gated on CD3⁺ lymphocytes.

Figure 2

Tumor lysis syndrome and regression of adenopathy occurred after infusion of allogeneic anti-CD19-CAR T cells into patient 1. (A) Serum lactate dehydrogenase (LDH) levels increased after the CAR-transduced T-cell infusion and then decreased as toxicity resolved. (B) The serum uric acid was 5.4 mg/dL (normal range 3.7-8.6 mg/dL) before the CAR T-cell infusion. Eight days after the infusion, the serum uric acid increased sharply before the recombinant urate oxidase drug rasburicase was given to reduce the uric acid. (C) Shortly after infusion of anti-CD19-CAR T cells, the total blood lymphocyte count abruptly decreased and then recovered to normal levels. (D) Serum IFN-γ levels at the indicated days after anti-CD19-CAR T-cell infusion were determined by enzyme-linked immunosorbent assay. A transient increase in serum IFN-γ occurred during the period of clinical toxicity. (E) CT scans were performed before the anti-CD19-CAR T-cell infusion and 26 days after the anti-CD19-CAR T-cell infusion. A decrease in adenopathy occurred at the locations indicated by the arrows.

Figure 3

A profound eradication of CLL cells from the bone marrow and blood occurred after infusion of anti-CD19-CAR T cells into patient 1. (A) A bone marrow biopsy was performed 2 days before anti-CD19-CAR T cells were infused. Bone marrow immunohistochemistry staining showed that 80% to 90% of the bone marrow cells consisted of atypical lymphoid cells that had a morphology consistent with CLL. The atypical lymphoid cells were CD79a⁺ and CD19⁺. Flow cytometry staining for immunoglobulin light chain restriction showed that the cells were clonal (not shown). (B) Twenty-six days after the anti-CD19-CAR T-cell infusion, immunohistochemistry staining of a bone marrow specimen showed no morphologic evidence of CLL. Only rare CD79a⁺ cells were present. (C) Blood B cells, of which 99% were CLL cells, dropped dramatically after infusion of anti-CD19-CAR T cells. (D) Patient 1 was dependent on platelet transfusions before the infusion of anti-CD19-CAR T cells. After the CAR T-cell infusion, the patient no longer required platelet transfusions. His platelet count increased for ∼2 months after the anti-CD19-CAR T-cell infusion; subsequently, his platelet count decreased when his CLL progressed.

Figure 4

Patient 5 obtained a CR after infusion of anti-CD19-CAR T cells. (A) Within 9 days after infusion of anti-CD19-CAR T cells, B cells were eradicated from the blood of patient 5. T cells (B) and NK cells (C) were at normal levels in the blood of patient 5 at the time of anti-CD19-CAR T-cell infusion. T-cell levels remained in the normal range, whereas NK cell levels fluctuated. The dashed line on the plots in panels A-C is the lower limit of the normal range of each cell type. (D) CT scans show that abdominal adenopathy regressed in patient 5 after infusion of anti-CD19-CAR T cells. The red arrows indicate a lymph node mass that regressed to a normal-sized lymph node.

Figure 5

Regression of adenopathy in patient 9 within 1 month after anti-CD19-CAR T-cell infusion. CT scans before the cell infusion (A) and after infusion of anti-CD19-CAR T cells showing that mediastinal adenopathy had regressed in patient 9 (B). The red arrows indicate a lymph node mass that regressed.

Figure 6

Anti-CD19-CAR–expressing T cells were detected in the blood of patients. (A) qPCR was used to determine the percentage of total PBMCs that contained the anti-CD19-CAR gene. The highest percentages of PBMCs containing the CAR gene were generally found between 6 and 14 days after infusion. (B) The absolute number of PBMCs containing the CAR gene per microliter of blood was determined by qPCR. (C) PBMCs from patient 8 were stained with a CAR-specific monoclonal antibody. Pretreatment PBMCs and PBMCs from 9 days after infusion were stained. Most CAR⁺ cells detected 9 days after infusion were CD4⁺. (D) PBMCs from patient 9 were stained with a CAR-specific monoclonal antibody. Pretreatment PBMCs and PBMCs from 7 days after infusion were stained. CAR⁺ cells included both CD4⁺ and CD8⁺ T cells. The plots in panels C and D are gated on live CD3⁺ lymphocytes.

Figure 7

Phenotype of anti-CD19-CAR–expressing T cells. (A) Infusion cells of patient 5 are shown after staining with a CAR-specific antibody. The first plot is gated on CD3⁺CD8⁺ cells. The phenotype of the CD3⁺CD8⁺CAR⁺ cells is shown on the subsequent plots of CD45RA vs CCR7, programmed cell death protein 1 (PD-1) vs CD57, and CD27 vs CD28. (B) CAR expression was assessed on pretreatment lymphocytes of patient 5 by flow cytometry. The plot shows minimal background staining when the cells were stained with a CAR-specific monoclonal antibody. The plot is gated on CD3⁺CD8⁺ lymphocytes. (C) The first plot is gated on CD3⁺CD8⁺ lymphocytes that were obtained from patient 5 12 days after CAR T-cell infusion. CAR⁺ cells made up 2.2% of the CD3⁺CD8⁺ lymphocytes. The subsequent CD45RA vs CCR7, PD-1 vs CD57, and CD27 vs CD28 plots are gated on CD3⁺CD8⁺ lymphocytes that are either CAR⁺ or CAR^–. (D) Infusion cells of patient 5 are shown. The first plot is gated on CD3⁺CD4⁺ cells. Subsequent plots show the phenotype of CD3⁺CD4⁺CAR⁺ cells. (E) The plot shows minimal background staining when pretreatment lymphocytes from patient 5 were stained with a CAR-specific antibody. The plot is gated on CD3⁺CD4⁺ lymphocytes. (F) The first plot is gated on CD3⁺CD4⁺ lymphocytes that were obtained from patient 5 12 days after CAR T-cell infusion. CAR⁺ cells made up 7.4% of the CD3⁺CD4⁺ lymphocytes. Subsequent plots show the phenotype of CD3⁺CD4⁺ cells that were either CAR⁺ or CAR^–. Increased expression of the T-cell inhibitory protein PD-1 was present on CD4⁺CAR⁺ T cells (G) and CD8⁺CAR⁺ T cells (H) relative to CAR^– T cells from the same patient at the time of the peak in the level of CAR⁺ blood T cells in each patient. The percentage of PD-1⁺ T cells is shown for the CAR⁺ and CAR^– T cells of the indicated patients. Note that an insufficient number of CAR⁺CD8⁺ T cells were present in the blood of patient 10 to allow accurate determination of PD-1–expressing CD8⁺ cells in this patient. PD-1 expression was determined as shown in panels C and F. For CD4⁺ T cells, the difference in PD-1 expression between CAR⁺ cells and CAR^– cells was statistically significant (P = .03, Mann-Whitney U test). PD-1 levels are only shown for 4 patients because some patients had very low or undetectable levels of CAR⁺ T cells in their blood and because we lacked sufficient PBMCs from relevant time points to conduct the experiment in some other patients.