A Novel Method for Molecular Enumeration of Circulating Tisa-Cel and Axi-Cel in Lymphoma Patients

Background: Chimeric antigen receptor (CAR) T-cell therapy is an effective treatment in approximately 40% of relapsed/refractory lymphomas. High rates of complete remission (CR) observed in clinical trials, led to fast-track FDA and EMA approval and commercialization of two CD19-directed CAR T-cell products (tisagenlecleucel/Tisa-cel/Kymriah and axicabtagene ciloleucel/Axi-cel/Yescarta) for refractory/relapsed diffuse large B cell lymphoma (DLBCL) and primary mediastinal large B cell lymphoma (PMBCL). In these studies, CAR T-cell engraftment and in vivo expansion have a crucial impact on disease response and toxicity. It is of paramount importance to develop homogeneous standardized approaches for monitoring expansion and persistence of CAR T-cells also in the case of cells not detectable by flow cytometry (FCM) as these data can guide clinical decision making.

Aims of this study were:

• to assess relevant time-points for FCM analysis across the different products
• to develop a unique quantitative PCR assay to detect both tisa-cel and axi-cel.

Methods:

We prospectively collected samples from 28 patients (pts) (16 DLBCL, 5 transformed follicular lymphoma and 7 PMBCL) treated with axi-cel (n=15) and tisa-cel (n=13) at the Fondazione IRCCS Istituto Nazionale Tumori between Nov 2019 and July 2020. CAR T-cells were monitored in the peripheral blood (PB) on days 0, 4, 7, 9, 14, 21, 28 and monthly post infusion by FCM. Cells were stained with CD19 CAR Detection Reagent (Miltenyi), Anti-Biotin-PE, CD3-FITC, CD4-VioGreen, CD8-APC-Vio770, CD45-VioBlue, CD14-APC and 7-AAD Staining Solution and analyzed by FCM using the MACSQuant® Analyzer (Miltenyi). To evaluate CAR T-cell persistence, we also generated a unique droplet digital PCR (ddPCR) primer-probe assay to target both tisagen-cel and axi-cel. We obtained a partial cDNA sequence of the two CAR constructs by amplification and direct sequencing of fragments generated with primers designed by educated guess to bind to the anti-CD19 murine single-chain variable fragment (scFV) and to the CD8α or CD28 trans-membrane regions of tisagen-cel and axi-cel vectors respectively. The assay was used to quantify CAR+ cells on genomic DNA extracted from PB cells collected serially after infusion. FCM and molecular data were correlated to outcome.

Results:

Peak expansion of CAR T-cells by FCM occurred within the first 14 days post-infusion. At the point of maximal expansion, collectively a mean of 17% of circulating CD3+ T cells were CAR+ and CAR+ cells were more abundant among CD3+ cells in pts receiving axi-cel (26% vs 8.5% for tisagen-cel, p<0.05). Patients achieving CR at 1 month from CAR T-cell therapy (n=10) had significantly more CAR+ cells at day 9 than non-CR pts (mean 146 vs 18 CAR+ cells/ul, p<0.05) and their magnitude of expansion in the first 14 days was higher than the one of non-CR pts (mean area under the curve (AUC)= 1110 vs 177; p<0.05). A cut-off value of CAR+ cells at day 9>62/ul was prognostic with sensitivity of 67% and specificity of 76% (Receiver operating characteristic analysis). At subsequent time points the number of CAR T-cells decreased, but continued to be detectable by FCM up to 6 months post infusion although at very low levels (mean 2 CAR+ cell/ul) irrespective of outcome. To evaluate whether these low counts reliably represent CAR T+ cells or are rather signals due to aspecific binding of the antibodies, a ddPCR assay was set up. The assay displayed an excellent separation between positive and negative droplets on both axi-cel and tisagen-cel products, good reproducibility and specificity with max standard deviation of 0.06 and no CAR positive signal in healthy donors. Additionally, the CAR specific signal was detectable up to the 0,01% dilution. FCM and ddPCR data had a good correlation (r=0.93, p<0.001): in accordance with FCM, ddPCR revealed a peak expansion of CAR+ signals at day 9 post infusion, followed by a progressive reduction of the CAR+ signal over time in all pts that becomes undetectable in case of relapse despite the presence of 1-2 cells/ul by FCM that probably represent aspecific findings.

Conclusion:

The combined use of FCM and the novel molecular assay facilitates the precise enumeration of commercial anti-CD19 CAR T-cells. Higher frequencies of CAR+ T cells at day 9 distinguish responders irrespectively of the product they received in our small prospective series. Updated results will be presented.