A Novel Autologous CAR-T Therapy, YTB323, with Preserved T-cell Stemness Shows Enhanced CAR T-cell Efficacy in Preclinical and Early Clinical Development

Abstract

CAR T-cell product quality and stemness (Tstem) are major determinants of in vivo expansion, efficacy, and clinical response. Prolonged ex vivo culturing is known to deplete Tstem, affecting clinical outcome. YTB323, a novel autologous CD19-directed CAR T-cell therapy expressing the same validated CAR as tisagenlecleucel, is manufactured using a next-generation platform in <2 days. Here, we report the preclinical development and preliminary clinical data of YTB323 in adults with relapsed/refractory diffuse large B-cell lymphoma (r/r DLBCL; NCT03960840). In preclinical mouse models, YTB323 exhibited enhanced in vivo expansion and antitumor activity at lower doses than traditionally manufactured CAR T cells. Clinically, at doses 25-fold lower than tisagenlecleucel, YTB323 showed (i) promising overall safety [cytokine release syndrome (any grade, 35%; grade ≥3, 6%), neurotoxicity (any grade, 25%; grade ≥3, 6%)]; (ii) overall response rates of 75% and 80% for DL1 and DL2, respectively; (iii) comparable CAR T-cell expansion; and (iv) preservation of T-cell phenotype. Current data support the continued development of YTB323 for r/r DLBCL.

Significance: Traditional CAR T-cell manufacturing requires extended ex vivo cell culture, reducing naive and stem cell memory T-cell populations and diminishing antitumor activity. YTB323, which expresses the same validated CAR as tisagenlecleucel, can be manufactured in <2 days while retaining T-cell stemness and enhancing clinical activity at a 25-fold lower dose. See related commentary by Wang, p. 1961. This article is featured in Selected Articles from This Issue, p. 1949.

Figure 1.

YTB323 vs. traditional CAR T-cell manufacturing. A, Using the novel T-Charge manufacturing platform, T cells are enriched from the patient's leukapheresis material, activated, and transduced with a lentiviral vector encoding the CD19-targeting CAR construct. After a short culture period, CAR-positive T cells are harvested, washed, and formulated in <2 days. Following successful clinical manufacturing, quality assurance assays (release testing) to assess cell viability, dose, and vector copy number are performed prior to final product release. Once manufactured in a commercial facility, YTB323 door-to-door time in the United States is anticipated to be 10 days or less. B, Using traditional manufacturing, T cells are enriched from the patient's leukapheresis material, activated, and transduced with a lentiviral vector encoding the CD19-targeting CAR construct. Following transduction, CAR T cells are expanded for 9 to 10 days before being washed and formulated. Following successful clinical manufacturing, quality assurance assays (release testing) to assess cell viability, dose, and vector copy number are performed prior to product release. Median door-to-door time for tisagenlecleucel has been reported at 20 or more days (25, 26). Door-to-door time is defined as the time from pickup of leukapheresis material to delivery of the final product back to the treatment site.

Figure 2.

Preclinical evaluation and characterization of YTB323 and CTL*019. A, YTB323 and CTL*019 as well as untransduced T cells (UTD) were manufactured according to the traditional manufacturing and T-Charge processes, respectively. Data are representative of 3 full-scale runs with 3 different healthy donors (n = 3 healthy donors) performed in the preclinical setting. Phenotypic analysis of leukapheresis (input) and CAR-T products after thaw of the cryopreserved product. Gating on live CD3⁺ events used to determine T-cell subsets (CCR7 vs. CD45RO). Numbers in quadrants indicate the percentage of the parental CD3⁺ population. Extended culture of T cells ex vivo, such as in traditional manufacturing, leads to a loss of T_SCM and naive phenotypes (red boxes). In contrast, limited culture ex vivo allows for the maintenance of naive-like cells by the T-Charge process (green boxes). B, Composition of YTB323 and CTL*019 is shown for the CD3⁺ population of donor 1 (donor 2 and 3 compositions can be found in Supplementary Fig. S1). The percentage (%) of the respective populations (colors) refers to the percentage of CD3⁺ T cells. 4 and 8, CD4⁺ and CD8⁺, respectively; cm, central memory; eff, effector; em, effector memory; n, naive-like. C, Single-cell gene set enrichment for input (leukapheresis), YTB323, and CTL*019 cells. Gene signature sets for T effector memory compared with T stem cell memory (left; ref. 7) cells and stemness (right; ref. 36) were assessed for gene set enrichment in each cell in the 3 samples. Violin plots compare the gene set enrichment scores between the 3 samples, with lines indicating the 25%, 50%, and 75% quantiles. Data shown are derived from 1 healthy donor. D, Top: NSG mice were injected with the pre–B-acute lymphoblastic leukemia line NALM6, expressing the luciferase reporter gene; the tumor burden is expressed as total body luminescence (BLI, p/s), depicted as mean tumor burden with 95% confidence interval (CI). On day 7 after tumor inoculation, mice were treated with YTB323 or CTL*019 at the respective doses (number of viable CAR⁺ T cells). The high-dose YTB323 group was terminated on day 33 due to the onset of xenogeneic graft-versus-host disease. UTD served as negative control. n = 5 mice for all groups, except n = 4 for T-Charge UTD, 1 × 10⁶ dose, and all CTL*019 dose groups. Five xenograft studies were run with CAR T cells generated from 5 different healthy donors, 3 of which included a comparison with CTL*019. Bottom: Time course of CAR⁺ T-cell concentration in NALM6 tumor–bearing mice treated with UTD, CTL*019, or YTB323 at 3 respective doses. Blood samples were taken at 4, 7, 14, 21, and 28 days after CAR-T injection. CAR⁺ T-cell (CD3⁺CAR⁺) concentrations were analyzed by flow cytometry at designed time points, depicted as mean cells with 95% CI. Reprinted from Blood, 138 (suppl 1), Engels B, et al. Preservation of T-cell stemness with a novel expansionless CAR-T manufacturing process, which reduces manufacturing time to less than two days, drives enhanced CAR-T cell efficacy, Abstract 2848, Copyright © 2021, with permission from American Society of Hematology. Published by Elsevier Inc. All rights reserved (13).

Figure 3.

Efficacy of YTB323 in patients with r/r DLBCL. A, Swimmer plot of YTB323 efficacy in patients with r/r DLBCL. Nineteen patients with r/r DLBCL were infused with YTB323 and evaluable for efficacy. Three patients at DL1 and 12 at DL2 responded to YTB323. Five patients at DL2 had high lactate dehydrogenase (LDH) at baseline (rows 4, 6, 8, 13, and 15 from the top), 7 patients at DL2 had high LDH at screening (rows 2, 4, 6–8, 14, and 15 from the top), and 4 patients had elevated LDH at both baseline and screening (rows 4, 6, 8, and 15 from the top). At DL1, 2 patients had high LDH at baseline (rows 18 and 19 from the top), 2 patients had high LDH at screening (rows 17 and 19 from the top), and 1 patient had elevated LDH at both baseline and screening (row 19 from the top). Arrow denotes ongoing efficacy assessments. Gray bars represent screening prior to infusion; purple bars represent screening after infusion. The symbols at time 0 represent each patient's disease status at the time of YTB323 infusion. B, Complete response rate following YTB323 treatment. ^aPatients who were in CR prior to receiving YTB323 due to either a late effect of prior therapies or bridging chemotherapy. cCR, continuous CR indicates patients in CR before infusion of YTB323; DL, dose level; PD, progressive disease; SD, stable disease; UNK, unknown. Panel A reprinted from Blood, 138 (suppl 1), Flinn I, et al. A first-in-human study of YTB323, a novel, autologous CD19-directed CAR-T cell therapy manufactured using the novel T-charge TM platform, for the treatment of patients (pts) with relapsed/refractory (r/r) diffuse large B-cell lymphoma (DLBCL), 2021; Oral 740. Copyright © 2021 American Society of Hematology. Published by Elsevier Inc. All rights reserved (20).

Figure 4.

Cellular kinetics of YTB323 and cytokine profiles. A, Cellular kinetics profile of YTB323 assessed by flow cytometry (left) and qPCR (right) up to 90 days overlaid with tisagenlecleucel data (JULIET) in patients with r/r DLBCL. Left: CAR expression by flow cytometry, as a percentage of CD3⁺CAR19⁺ cells in CD3⁺ cells, after YTB323 infusion in patients who received full DL1 (2.5 × 10⁶ cells) are presented in blue, below target DL2 (>2.5 × 10⁶ and <12.5 × 10⁶ cells) in green, and full DL2 (12.5 × 10⁶ cells) in red; CAR expression by flow cytometry after tisagenlecleucel infusion (0.1–6.0 × 10⁶ cells; median, 3.0 × 10⁶ cells) is presented in gray. Right: CAR transgene levels by qPCR, as copies/μg DNA, after YTB323 infusion in patients who received full DL1, are presented in blue, partial DL2 in green, and full DL2 in red; CAR transgene levels by qPCR after tisagenlecleucel infusion are presented in gray. B, Cytokine profiles following YTB323 infusion. IL6, IL10, and IFNγ were measured by mean square displacement analysis up to 30 days for patients with cCR (long dashed line), CR (solid line), and non-CR (short dashed line), and with no CRS (black), grade (Gr) 1–2 CRS (blue), and Gr >3 CRS (red). Pink bar indicates the median time of peak expansion measured by qPCR as in A. cCR, continuous CR indicates patients in CR before infusion of YTB323; CD, cluster of differentiation; DL, dose level; IFNγ, interferon gamma; IL, interleukin; LOQ, limit of quantification; qPCR, quantitative polymerase chain reaction.

Figure 5.

T-cell phenotype of CD4 and CD8 subsets in leukapheresis and YTB323 and tisagenlecleucel final products. A, Left: T-cell memory subsets in the leukapheresis (n = 17) preserved in the YTB323 final product (n = 16) using flow cytometry. Phenotype percentages of CD4 or CD8 for paired patient samples: naive/T_SCM = CCR7⁺CD45RA⁺; central memory = CCR7⁺CD45RO⁺, effector memory = CCR7⁻CD45RO⁺, and senescent/exhausted = CD27^–CD28^–. Right: Absolute changes in phenotype frequencies [mean difference and 95% confidence interval (CI)] show overall preservation of memory T-cell subsets. B, Left: Significant changes in T-cell memory subsets between the leukapheresis (n = 122) and tisagenlecleucel final product (n = 140), including lower frequency of naive/T_SCM in the final product using flow cytometry. Phenotype percentages of CD4 or CD8 for paired patient samples: naive/T_SCM = CCR7⁺CD45RA⁺, central memory = CCR7⁺CD45RA^–, effector memory = CCR7^–CD45RA^–, and effector = CCR7^–CD45RA⁺. Right: Absolute changes in phenotype frequencies (mean difference and 95% CI) show significant shifts in memory T-cell subsets between the leukapheresis and final product. Leukapheresis material and tisagenlecleucel final product were from patients enrolled in the JULIET clinical trial (NCT02445248). C, YTB323 retains a naive stem-like gene signature. Single-sample gene set enrichment for leukapheresis and cell product T cells from bulk RNA-seq. YTB323: <2-day manufacturing with T-Charge platform. Tisagenlecleucel: standard 9-day tisagenlecleucel manufacturing. Gene signature sets for stemness (left; ref. 22) and naive T cells (Tn) vs. effector memory T cells (TEM; right; ref. 7) were assessed for gene set enrichment scores in the leukapheresis and YTB323 and tisagenlecleucel cell products. P values are calculated for the comparison of YTB323 leukapheresis (n = 12) vs. cell products (n = 11) and tisagenlecleucel leukapheresis (n = 9) vs. cell products (n = 23).