CD20-specific adoptive immunotherapy for lymphoma using a chimeric antigen receptor with both CD28 and 4-1BB domains: pilot clinical trial results

Abstract

Cellular immune responses have the potential to elicit dramatic and sustained clinical remissions in lymphoma patients. Recent clinical trial data demonstrate that modification of T cells with chimeric antigen receptors (CARs) is a promising strategy. T cells containing CARs with costimulatory domains exhibit improved activity against tumors. We conducted a pilot clinical trial testing a "third-generation" CD20-specific CAR with CD28 and 4-1BB costimulatory domains in patients with relapsed indolent B-cell and mantle cell lymphomas. Four patients were enrolled, and 3 received T-cell infusions after cyclophosphamide lymphodepletion. Treatment was well tolerated, although one patient developed transient infusional symptoms. Two patients without evaluable disease remained progression-free for 12 and 24 months. The third patient had an objective partial remission and relapsed at 12 months after infusions. Modified T cells were detected by quantitative PCR at tumor sites and up to 1 year in peripheral blood, albeit at low levels. No evidence of host immune responses against infused cells was detected. In conclusion, adoptive immunotherapy with CD20-specific T cells was well tolerated and was associated with antitumor activity. We will pursue alternative gene transfer technologies and culture conditions in future studies to improve CAR expression and cell production efficiency.

Trial registration: ClinicalTrials.gov NCT00621452.

Figure 1

Expansion and cytotoxicity of modified T cells. (A) Schematic representation of the αCD20-28-BB-ζ chimeric receptor, not to scale. (B-C) Growth curves of ex vivo expanded T cells for the 3 treated patients (B) and of the 2 attempts to expand cells from patient UPN-01 (C). PBMCs collected by apheresis were stimulated with anti-CD3 Ab and IL-2, transfected by electroporation by the plasmid encoding the CAR shown in panel A, selected with G418, and restimulated every 12 to 15 days in a rapid expansion protocol. Black bars represent periods of G418 selection; and gray bar, the target cell dose range. (D) CD20-specific cytotoxicity of G418-selected autologous patient T cells was assessed with 5-hour chromium-release assays using the following ⁵¹Cr-labeled target cells: Granta cells (MCL), Daudi cells (Burkitt NHL), EL4 cells transfected to express CD20, or the parental EL4 cell line lacking CD20 expression. The cells were tested after G418 selection (top panels) and 54, 27, and 40 days later for the 3 patients at the time of T-cell infusions (bottom panels). Data represent the mean of triplicate values, and error bars represent SEM.

Figure 2

Immunophenotype and relative telomere length of infused T cells. End of production T cells for patients UPN-02, UPN-03, and UPN-04, as well as cells from patient UPN-01 (after 9 stimulation cycles) were analyzed by flow cytometry after being thawed, washed, and stained with the antibodies shown. (A) Geometric mean fluorescence intensity (MFI) after subtracting the isotype control MFI and (B) percent positive cells. (C) Genomic DNA was harvested from CD3-selected apheresis PBMCs and from pre-infusion, ex vivo expanded CAR⁺ T cells for each patient. T cells from both attempts at expansion for UPN-01 were analyzed. Quantitative PCR was used to determine the amounts of telomeric DNA (T) and of a single-copy internal reference gene (S) for each sample. The relative mean telomere length of each cell population is represented by the T/S ratio. Data represent the mean of 2 assays, each performed in triplicate.

Figure 3

Engraftment, persistence, and localization of modified T cells to tumor sites. Quantitative real-time PCR was performed on genomic DNA harvested from study subject PBMCs collected before and at serial time points after T-cell infusions, using primers specific for the transgene. CAR⁺ cells are shown as a percentage of total PBMCs for each of the treated patients over (A) the first month and (B) the first year. The same quantitative PCR assay was used to detect modified cells in genomic DNA harvested from surgically excised lymph nodes (C) or bone marrow aspirates (D) collected 24 to 48 hours after the third T-cell infusion. For patient UPN-03, 2 adjacent lymph nodes were excised and analyzed separately, and an additional bone marrow aspirate was performed 1 month after the third T-cell infusion. Data represent mean values (± SD) of at least 2 assays per time point, with each sample performed in triplicate. Arrows indicate infusions of genetically modified T cells; and black bar, the 14-day period of low-dose IL-2 injections.

Figure 4

Clinical response to T-cell infusions. Patient UPN-04 was treated with CY 1 g/m² 2 days before the first of 3 infusions of αCD20-28-BB-ζ CAR⁺ T cells, which were given 2 to 5 days apart and followed by 14 days of subcutaneous IL-2 injections. (A) Response in a cervical lymph node as imaged by CT scan of the neck at baseline and 1, 6, and 12 months after T-cell infusions. (B) PET-CT scan of the cervical lymph node at baseline and 3 months after T-cell infusions. (C) Change in tumor volume over time. The sum of the products of the diameters of the 7 largest lymph nodes seen on the baseline and 1-, 6-, and 12-month CT scans. (D) Heterogeneity of response in individual lymph nodes. The products of the diameters of the 7 largest lymph nodes from the 6-month CT scan are shown as a percentage of their baseline size. L indicates left; and R, right.

Figure 6

Effects of CY lymphodepletion and IL-2 on lymphocyte subsets and cytokine levels. Multiparameter flow cytometry was performed with patient peripheral blood samples collected at baseline and at serial time points after T-cell infusions to quantify (A) all lymphocytes and CD4⁺ and CD8⁺ T cells, expressed as percentage change from baseline and (B) absolute numbers of Treg cells (CD4⁺/FoxP3⁺) and the ratios of Treg to CD4⁺ T cells and CD8⁺ T cells for UPN-02 (●), UPN-03 (■), and UPN-04 (▴). (C) CD19⁺ and CD20⁺ B cells, expressed as percentage change from baseline. (D) Levels of multiple cytokines were measured in patient serum samples collected at serial time points after CY administration using a Luminex microbead immunoassay and expressed as fold change from baseline (day −15, day −9, and day 0 before CY for UPN-02, UPN-03, and UPN-04, respectively). The range and baseline values (listed in parentheses sequentially for patients UPN-02, UPN-03, and UPN-04) in picograms per milliliter are as follows: IL-2: < 8 to 413 (90.9, 32.7, < 8); IL-4: < 2 to 298 (50.2, 10.6, < 2); IL-6: < 2 to 174 (58.0, 11.7, 2.73); IL-7: 4.53 to 336 (49.7, 14.9, 6.41); IL-10: < 1 to 526 (41.3, 29.4, < 1); IL-12p70: < 8 to 2011 (328, 16.3, < 8); IL-15: 1.61 to 76.4 (21.5, 4.69, 1.74); IL-17: < 15 to 412 (27.3, < 15, < 15); IL-21: < 15 to 4997 (685, 151, < 15); IFN-γ: 1.6 to 23.1 (8.18, 2.52, 1.70); TNF-α: < 3.6 to 164 (28.6, 8.58, < 3.6); and MIP1-α: < 300 to 868 (335, < 300, < 300). Arrows indicate infusions of genetically modified T cells; and black bar, the 14-day period of low-dose IL-2 injections.

Figure 5

Immune responses to modified T cells. Two assays were performed to detect Ab against the transgene in patient serum samples collected before and at serial time points after T-cell infusions. (A) An ELISA was performed testing for the presence of Ab binding to Leu16 mouse anti–human CD20 Ab (from which the αCD20-28-BB-ζ CAR is derived). Baseline patient serum was used as a negative control. Data represent the mean ± SEM of duplicate values. (B) A flow cytometric assay was performed in which serial patient serum samples were incubated with HEK-293 cells genetically modified to express the αCD20-28-BB-ζ CAR, untransfected HEK-293 cells (negative control), or autologous EBV-LCL (positive control), followed by FITC-conjugated goat anti–human F(ab′)₂ Ab. The median fluorescence intensity (MFI) for each sample at various time points at the 1:2 dilution is shown for each cell line. The black bar indicates the 14-day period of low-dose IL-2 injections. (C) The presence of cellular immune responses to infused T cells was assessed by stimulating serially collected patient PBMCs with irradiated autologous EBV-LCL modified to express the αCD20-28-BB-ζ CAR and NeoR gene products (UPN-02 and UPN-03) or irradiated autologous CAR⁺ T cells (UPN-04) at a 2:1 responder/stimulator ratio. After two 1-week stimulations, these PBMCs were used as effectors in ⁵¹Cr-release assays in which target cells were pre-infusion modified T cells, untransfected autologous EBV-LCL, CAR⁺ autologous EBV-LCL, or untransfected autologous PBMCs at E:T ratios of 25:1 (UPN-03 and UPN-04) or 12.5:1 (UPN-02). The lysis of CAR⁺ LCL is expressed as the difference between CAR⁺ and CAR⁻ LCL for patients UPN-02 and UPN-03. The mean ± SEM of triplicate wells is shown.