High Bacterial Abundances of Dorea and Pediococcus in the Gut Microbiome Linked to Expansion, Immune Checkpoint Expression and Efficacy of CD19-Directed CAR T-Cells in Patients with r/r DLBCL

Introduction: The crosstalk of the gut microbiome and the human host can influence T-cell immune responses and has emerged as a modulator in cancer immunotherapy (Gopalakrishnan et al., 2018). Individual bacteria isolated from human fecal specimen were also shown to shape systemic and gut mucosal T cell repertoires (Geva-Zatorsky et al., 2017), and several commensal members of the human gut microbiome were recently associated with the kinetics of the reconstitution of peripheral immune cells after allo-HCT (Schluter et al., 2020). However, in patients treated with CAR T-cells a link between gut microbiome configurations and T-cell characteristics has not been described yet. Recently, we and others could associate low CAR T-cell expansion and dysfunction with treatment failure. Here, we hypothesize that certain gut microbes or even intestinal monodomination of facultative pathogens may correlate with CAR T-cell expansion and expression profile of immune checkpoint molecules which might impact treatment outcome.

Methods: Patients with relapsed / refractory Diffuse-Large B-Cell Lymphoma (DLBCL) were treated with the CD19 specific CAR T-cell products Axicabtagene-Ciloleucel or Tisagenlecleucel at our institution from April 2019 to May 2021. Within this time-period, peripheral blood and fecal biospecimens from 23 patients were collected sequentially before, during and after CAR T-cell transfusion (specific time points: before lymphodepleting chemotherapy, day of CAR T-cell transfusion, day 7, day 14). CAR T-cells and the expression profile of immune checkpoint molecules (PD-1, TIM-3, LAG-3, 2B4) were studied by multiparameter flow cytometry. CAR T-cell peak expansion was assessed relatively (% CAR ⁺ T-cells of CD3+ T-cells) and absolutely (/ul). In addition, effector : target (E:T) ratios were estimated as absolute peak expansion of CAR T-cells (/ul) per tumor volume (as sum of the product of diameters based on Lugano criteria with up to 6 target lesions in cm ³). 16S rRNA gene sequencing was performed on 83 stool samples. Responder (R, complete or partial remission) were retrospectively assessed compared to Non-Responder (NR, stable or progressive disease) according to response assessment with PET-CT three months after CAR T-cell transfusion.

Results: Higher alpha diversity (i.e., within-sample diversity) prior to CAR T-cell transfusion correlated positively with a favorable E:T ratio (CAR peak expansion/ul : tumor volume in cm ³, p=0.04, r=0.608; Fig. 1). Interestingly, the abundances of certain taxa prior to CAR T-cell transfusion correlated with the expansion of CAR T-cells in vivo. In particular we found a positive association with Pediococcus (p=0.018, r=0.751, FDR corrected) and with Anaerovoracaceae (E:T ratio, p=0.027, r=0.84, FDR corrected). Next, we studied the expression profile of immune checkpoint molecules on CAR T-cells in the context of microbial taxa in R vs NR patients. Notably, we observed a positive association between the frequency of CD8+ CAR T-cells devoid of inhibitory molecules (PD-1, TIM-3, LAG-3 and 2B4) and the relative abundance of the intestinal microbial genus Dorea (p=0.039, r=0.416: see Fig. 1). Furthermore, we found a higher relative abundance of Dorea (p=0.043) prior to transfusion in R compared to NR patients as well as significant positive associations of Dorea (p=0.0.008, r=0.54, not FDR corrected) and Pediococcus (p=0.047, r=0.462, not FDR corrected) with progression free survival .

Conclusion: For the first time, we describe associations between the compositional changes of the gut microbiome and kinetics as well as immune characteristics of CAR T-cell therapy in patients with relapsed / refractory DLBCL. In the future, our findings need to be validated in a larger patient cohort and might serve as a predictive biomarker for microbiome-based patient assessment and microbiome-targeted therapeutic approaches to identify benefiting patients and to improve outcome.

Figure 1

View largeDownload slide

Figure 1

View largeDownload slide

Close modal