Characterizing Environmental Adaption and Virulence of Gut Microbiome Enterococcus Spp. from Patients Receiving CD19-Targeted CAR-T Cell Therapy

Introduction

The gut microbiome plays an important role in modulating the host’s immune system. Multiple clinical studies have linked gut microbiome dysbiosis to poorer clinical outcomes in patients undergoing adoptive immune-cell therapies (Stein-Thoeringer et al. Science 2019, Stein-Thoeringer et al. Nature Medicine 2023). Gut microbiome dysbiosis has been linked in particular to the intestinal expansion of Enterococcus spp., often due to antibiotic pretreatments. This expansion was associated with altered immune responses and pathologies such as GVHD (Stein-Thoeringer et al. Science 2019; Taur et al. Clin Infect Dis 2012). Here, we hypothesized that Enterococcus spp. may also impact the outcomes of CD19-CAR-T cell therapy as these patients are highly susceptible to antibiotic-induced gut dysbiosis.

Methods

59 patients with B cell malignancies who underwent treatment with CD19-CAR-T cells at two university cancer centers in Germany (Munich and Heidelberg) were included, where we collected serial clinical data and stool samples. Enterococcus spp. were isolated from stool specimens and identified via Sanger sequencing. In our downstream analyses, we only focused on samples with E. faecalis or E. faecium and carried out several in vitro functional tests. Antibiotic resistance was determined using disk diffusion assays. Immunogenicity and strain virulence was determined by measuring cytokine responses of enterococcal lysates pulsed on human PBMCs from healthy donors and by quantitative testing for biofilm formation. Metabolic characteristics were assessed by recording growth rates in a minimal medium while adding different carbohydrates. All stool samples were sequenced by shotgun sequencing. In addition, Enterococcus spp. isolated from 42 LMU patients were individually sequenced by whole genome sequencing to determine antimicrobial resistance genes and virulence factors.

Results

In vitro testing showed an increase in ampicillin resistance for E. faecium and E. faecalis from pre- to post-CAR-T time points (Fisher’s exact test p-value = 0.08754 and p-value = 0.06234, respectively). This increase was particularly evident in the Heidelberg patient cohort for E. faecalis (Fisher’s exact test p-value = 0.02105) and in the Munich patient cohort for E. faecium (Fisher’s exact test p-value = 0.07337). In addition, we observed an increased imipenem resistance for E. faecalis (Fisher’s exact test p-value = 0.0534) and, specifically for E. faecium (Fisher’s exact test p-value = 0.05235), in the Heidelberg cohort. These results suggest a general increase in resistance to beta-lactam antibiotics after cell therapy, most likely due to intense antibiotic treatments. When assessing strain virulence in a biofilm formation assay, we detected higher biofilm formation in E. faecium isolated from responder patients (response measured at day 180; Wilcoxon test p-value = 0.03). Notably, interleukin-6 (IL-6) release from stimulated PBMCs was significantly correlated with biofilm formation (p-value ≤ 0.001) and significantly higher in Enterococcus spp. isolates from patients with high-grade ICANS (Wilcoxon test p-value = 0.0054). The association of treatment response and microbial virulence, measured through IL-6 release and biofilm formation, could imply an increased immune activation in patients, possibly resulting in a greater CAR-T cell activity. Strains of E. faecalis that induced high IL-6 release also showed a higher capacity for metabolizing carbohydrates, i.e. glucose, fructose and lactose (Wilcoxon test p-value = 0.0016, p-value = 0.00052 and p-value = 0.019, respectively), which further supports elevated virulence.

Ongoing whole genome sequencing of the isolates aims to characterize the genomic correlates of virulence driving the immunogenic phenotype of the identified strains and will be presented.

Conclusions

Our novel findings indicate that antibiotic resistance, metabolic phenotypes and, notably, virulence of Enterococcus spp. can significantly affect the host on a strain-specific level. These findings suggest the need to perform strain-specific phenotyping in future microbe – host interaction studies in patients undergoing cell-based immunotherapies. Ultimately enabling future microbiome-based patient assessment and monitoring.