Gut lactate-producing bacteria promote CD4 T cell recovery on Anti-retroviral therapy in HIV-infected patients

Abstract

Anti-retroviral therapy (ART) effectively suppresses viral replication in HIV-infected patients, however CD4 + cell restoration to normal value is not achieved by 15-20% of patients who are called immune non-responders. Gut microbiota composition has been shown to influence host immunity. Herein, to identify intestinal microbial agents that may influence the CD4 recovery in HIV-infected patients, we utilized a "Quasi-paired cohort" method to analyze intestinal metagenome data from immunological responders (IRs) and immunological non-responders (INRs). This method identified significant enrichment for Streptococcus sp. and related lactate-producing bacteria (LAB) in IRs. In a validation cohort, positive correlations between the abundance of these LAB and the post-ART CD4 + recovery was observed, and a prediction model based on these LAB performed well in predicting immune recovery. Finally, experiments using a germ-free mouse model of antibody-induced CD4 + cell depletion showed that supplementation with a lactate-producing commensal Streptococcus thermophilus strongly promoted CD4 recovery. In conclusion, our study identified a group of LAB that was associated with enhanced immune recovery in post-ART HIV-infected patients and promotes CD4 + cell restoration in a mouse model. These findings favour supplementation of LAB commensal as a therapeutic strategy for CD4 + cell count improvement in HIV-infected patients.

Keywords: HIV; Immune recovery; Lactic acid bacteria; Metagenome.

Graphical abstract

Fig. 1

The principle of “quasi-paired cohort”. a. The scheme of the method “quasi-paired cohort” that “paired” patients of similar metagenomic profile. b. Representative species enriched in IR and INR groups identified with the “quasi-paired cohort” approach. Wilcoxon Signed-Rank test, *, p < 0.05; **, p < 0.01; ***, p < 0.001.

Fig. 2

Enrichment of lactate-producing bacteria and lactate-dehydrogenase in IR group. a. Heat map of species enriched in IR and INR groups and their correlations to prognostic immunological markers. Colors indicate the ρ value of Spearman’s coefficient, and asterisk (*) denotes strong correlations of ρ ≥ 0.4 or ρ ≤ -0.4. Species highlighted with a purple background are typical lactic acid bacteria (LAB). Positive prognostic markers include CD4TC (TC, short for T cell count), CD4CD28TC, MCD4TC (Memory CD4 + T cell), HNCD4TC (Homing naïve CD4 + T cell), NCD4TC (Naive CD4 + T cell), and the most informative marker of CD4/CD8 (the ratio of CD4 + T cell count to CD8 + T cell count). Negative prognostic markers include CD8TC, CD8CD38TC, , CD8DRTC (CD8 + DR + T cell). b. Comparison of the abundance of D-LDH and L-LDH in the microbiota of IR and INR patients. LDH, lactate-dehydrogenase; CMP, read count per million reads. c. The relative abundance of the pathway “mixed acid fermentation” in the microbiota of IR and INR patients. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 3

Two clusters of co-occurrent LAB species and their correlations to post-ART immune recovery. a. The top five co-occurrence clusters of species in HIV gut microbiota. The thickness of lines indicates rho value of Spearman’s coefficient. b. Comparison of total abundance of species in cluster I (left), cluster II (right) between IR and INR patients. Wilcoxon rank sum test, *, p < 0.05. c. The correlations of the total abundance of species in cluster I (left) and cluster II (right) to the post-ART growth rate of CD4 + T cells and its subsets of patients in the validation cohort. The growth rate was represented as the After/Before ratio of cell count, and the abundance is averaged between before and after ART. The strength of the correlations was evaluated by the Spearman rank-sum test with the rho value labelled above. d. Change of the total abundance of species in cluster I (left) and cluster II (right) after ART. Wilcoxon signed-rank test, *, p < 0.05; ***, p < 0.001.

Fig. 4

A prediction model based on the abundance of species in Cluster I and II. a. Performance of the prediction model evaluated with ROC. AUC, area under curve. b. Top five of species contributors to the prediction model. Species in blue colour are also identified as species enriched in IR patients. c. The growth rate of CD4 + cell (After/Before ratio of CD4 + cell count) in each quartile of the patients in the validation cohort which was divided according to their immune promotion score. The score of each patient was inferred from the prediction model with the abundance of species in Cluster I and II. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 5

Effects of a LAB species of Streptococcus thermophilus on the recovery of CD4 lymphopenia in a mouse model. a. Scheme of the experimental design for a probiotic-based preventive regimen using the mouse model of transient CD4 + cell depletion. b-c. Representative flow cytometry plots showing the proportion of CD4 + cells and CD8 + cells (b), and Naïve CD4 T cells (c) in blood of the indicated groups. The proportions of each cell subset in all samples were plotted on the right. d. Representative microscopic photos of lymphatic follicles in the lamina propria . Brown cells are immunohistochemistry stained CD4 + T cells. Scale bar = 100 μm. ST, Streptococcus thermophilus gavage group (n = 4); EB, Eubacterium Bakeri gavage group (n = 4); Ctrl, negative control group (n = 4); and Mock, mock control group (n = 4). *, p < 0.05.