Gut microbiota modulation in cardiac cell therapy with immunosuppression in a nonhuman primate ischemia/reperfusion model

Abstract

Gut microbiota affect transplantation outcomes; however, the influence of immunosuppression and cell therapy on the gut microbiota in cardiovascular care remains unexplored. We investigated gut microbiota dynamics in a nonhuman primate (NHP) cardiac ischemia/reperfusion model while under immunosuppression and receiving cell therapy with human induced pluripotent stem cell (hiPSC)-derived endothelial cells (EC) and cardiomyocytes (CM). Both immunosuppression and EC/CM co-treatment increased gut microbiota alpha diversity. Immunosuppression promoted anaerobes, such as Faecalibacterium, Streptococcus, Anaerovibrio and Dialister, and altered amino acid metabolism and nucleosides/nucleotides biosynthesis in host plasma. EC + CM cotreatment favors Phascolarctobacterium, Fusicatenibacter, Erysipelotrichaceae UCG-006, Veillonella and Mailhella. Remarkably, gut microbiota of the EC/CM co-treatment group resembled that of the pre-injury group, and the NHPs exhibited a metabolic shift towards amino acid and fatty acid/lipid biosynthesis in plasma following cell therapy. The interplay between shift in microbial community and host homeostasis during treatment suggests gut microbiome modulation could improve cell therapy outcomes.

Fig. 1. Immunosuppressive treatment increased the gut microbial diversity.

a Schematic illustration of cell transplantation therapeutic procedure in nonhuman primate (NHP) rhesus macaque. b The weighted UniFrac distance of gut microbiota in NHP under immunosuppressive treatment. IRD7, pre-immunosuppressive treatment; IRD28, day 7 post-tacrolimus immunosuppressive treatment. The gut microbial α-diversity determined by (c) observed ASV, (d) Shannon’s index and (e) Pielou’s evenness. Pairwise permutational multivariate analysis of variance (PERMANOVA) was used to analyzed data in (b) and Wilcoxon signed-rank test with p values adjusted was used to analyze data in (c–e). IM intramuscular, IV intravenous, SQ subcutaneous.

Fig. 2. Immunosuppression-induced alteration of gut microbial community.

a LEfSe (Linear discriminant analysis (LDA) Effect Size) analysis for the gut bacterial genus predominant after immunosuppressive treatment. (b) The abundance of Dialister, Anaerovibrio, Faecalibacterium, and Streptococcus under immunosuppression by qPCR. (c) Correlation network of the gut microbiota in the genus level. The correlation was determined with Sparse Estimation of Correlations among Microbiomes (SECOM) (Pearson 2) filter with correlation more than 0.9 and p-value less than 0.01. The Wilcoxon signed-rank test was used to analyze data in (b).

Fig. 3. Metabolic alteration after immunosuppressive treatment.

PICRUSt (Phylogenetic Investigation of Communities by Reconstruction of Unobserved States) analysis of the changes in metabolic pathways influenced by immunosuppressive treatment. Data are presented as mean ± SEM. The error bars refer to the standard error of the mean. The data were analyzed with Welch’s two-sided t-test.

Fig. 4. Correlation between gut bacteria and plasma metabolites under immunosuppression.

a Sparse Partial Least Squares (sPLS) of the metabolite profiling under immunosuppression using LC-MS. b The significant increase of metabolites under immunosuppressive treatment using the linear models with covariate adjustment. c The immunosuppressive-predominant gut bacterial genus and the plasma LC-MS metabolomic profiling were used to calculate the correlation between gut microbiota and metabolites. The prediction was based on the EMBL database. The paired microbiome and metabolomics features that identified as significant by both statistical correlation and GEM-based prediction are marked with a *.

Fig. 5. Cell transplantation influenced gut microbiota of the recipients.

a The schematic illustration of experimental design for cell transplantation. b The weighted UniFrac distance of gut microbiota in NHP after cell transplantation. c The α-diversity of recipient gut microbiota after cell transplantation determined by observed ASV, Shannon’s index and Pielou’s evenness. d LEfSe analysis of the treatment predominant genus of gut microbiota. e Genus-specific qPCR confirmation of Streptococcus, Caproiciproducens, Cerasicoccus, Olsenella, Phascolarctobacterium, and Veillonella. f The microbial dysbiosis (MD) index of end-point (IRD56) gut microbiota in recipients compared with pivotal time points of treatments (pre-IR, IRD7 and IRD28). Data in (c, e) were analyzed with Kruskal-Wallis test followed by Dunn test for multiple comparisons. Data in (e) presented as mean ± SEM. The error bars refer to the standard error of the mean. Ctrl, vehicle; CM, human iPSC-derived cardiomyocytes; EC, human iPSC-derived endothelial cells.

Fig. 6. The alteration in metabolic pathways after cell transplantation.

a PICRUSt predicted metabolic pathway alteration between control and cell transplantion groups. b PICRUst prediction of metabolic pathway alteration between human iPSC-derived cardiomyocyte (CM) and endothelial cell (EC) transplantation. Data are presented as mean ± SEM. The error bars refer to the standard error of the mean. Data were analyzed with Welch’s two-sided t-test.

Fig. 7. Correlation between gut bacteria and the plasma metabolites after cell transplantation.

a Sparse Partial Least Squares (sPLS) of the plasma metabolite profiling after cell transplantation using LC-MS. b The significant increase of metabolites under cell therapy using the linear models with covariate adjustment. c The cell therapy-predominant gut bacterial genus and the plasma LC-MS metabolomic profiling were used to calculate the correlation between gut microbiota and metabolites. The prediction was based on the EMBL database. The paired microbiome and metabolomics features that identified as significant by both statistical correlation and GEM-based prediction are marked with a *.