Intestinal transgene delivery with native E. coli chassis allows persistent physiological changes

Abstract

Live bacterial therapeutics (LBTs) could reverse diseases by engrafting in the gut and providing persistent beneficial functions in the host. However, attempts to functionally manipulate the gut microbiome of conventionally raised (CR) hosts have been unsuccessful because engineered microbial organisms (i.e., chassis) have difficulty in colonizing the hostile luminal environment. In this proof-of-concept study, we use native bacteria as chassis for transgene delivery to impact CR host physiology. Native Escherichia coli bacteria isolated from the stool cultures of CR mice were modified to express functional genes. The reintroduction of these strains induces perpetual engraftment in the intestine. In addition, engineered native E. coli can induce functional changes that affect physiology of and reverse pathology in CR hosts months after administration. Thus, using native bacteria as chassis to "knock in" specific functions allows mechanistic studies of specific microbial activities in the microbiome of CR hosts and enables LBT with curative intent.

Keywords: bile acid metabolism; complex gut microbiome; glucose homeostasis; gut microbiome; metabolism; microbe-host intereractions; non-model gut microbes; precision microbiome modulation; synthetic biology; type 2 diabetes.

Figure 1.. Gut native E. coli are genetically tractable and can serve as a chassis for transgene delivery

(A) Experimental strategy of engineered native bacteria. (B) Optimal characteristics of a chassis for transgene delivery for a potential LBT. (C) Original isolate (EcAZ-1), GFP-producing strain (EcAZ-2), and GFP- and BSH-producing strain (EcAZ-2^BSH+) plated on LB containing TDCA. Precipitate around EcAZ-2^BSH+ is the result of TDCA deconjugated by the bacteria to DCA and qualitatively indicates enzyme functionality. (D) Growth curve of EcAZ-2 compared with EcAZ-1, EcAZ-2^BSH+ compared with EcAZ-1, and EcAZ-2^IL10+ compared with EcAZ-2. The line represents an average of three measurements per strain. (E) Proteomic analysis shows increased BSH protein expression in EcAZ-2^BSH+ compared with EcAZ-1 and EcAZ-2 (n = 4). (F) Log₂ ratio of DCA to TDCA in a timed enzymatic assay (n = 3; see Figure S1G for raw values). (G) IL-10 levels detected with ELISA from cell lysates (n = 6). The marker covers some error bars in (F) and (G). Given the low number of samples for (E) and (F), we used a Student’s t test after the confirmation of normality with Q-Q plot. We used a Mann-Whitney U test for (G). (A) and (B) were created with BioRender.com. See also Figure S1.

Figure 2.. Engineered native E. coli can engraft in the luminal environment

(A) Long-term colonization after gavage with 10¹⁰ CFUs of E. coli MG1655 and E. coli Nissle 1917 (two lab-strain E. coli often used as chassis for function delivery), as well as EcAZ-2 in non-antibiotic-treated CR-WT mice housed in a low-barrier, non-sterile facility after a single gavage. Sample measures in the gray area denote mice where the bacteria were not detectable at a limit of detection of Log₁₀ of 2. (B) Colonization at 110 days (Fisher’s exact test; 4–8 mice/condition). (C) Colonization in co-housed mice housed in a low-barrier, non-sterile facility after a single gavage, where one mouse received 10¹⁰ CFUs of EcAZ-2 and the other received vehicle via oral gavage. Arrow indicates change in stool collection protocol where mice were separated from their cage mates 24 h prior to their weekly stool collection. Mice were returned to their cage mates right after stool collection. Thus, although uncolonized mice are exposed to EcAZ-2 through coprophagia, the level is not sufficient to lead to engraftment. (D) Colonization after gavage of different initial doses of EcAZ-2 in non-antibiotic-treated CR-WT mice housed in a low-barrier, non-sterile facility after a singlegavage (n = 5). (E) Colonization after gavage of 10¹⁰ CFUs of EcAZ-2 in the stool of non-antibiotic-treated CR-WT mice in an SPF facility on a stable diet (10% fat) or on diets with different macronutrient profiles (non-fat diet [NFD] or high-fat diet [HFD]). Diet was changed 4 days post-gavage (4 mice/condition). (F) Colonization after gavage of 10¹⁰ CFUs of EcAZ-2 or EcAZ-2^BSH+ in non-antibiotic-treated CR-WT mice housed in an SPF facility (4–12 mice/condition). (G) Colonization after gavage of 10¹⁰ CFUs of EcAZ-2 or EcAZ-2^IL10+ in non-antibiotic-treated CR-WT mice housed in an SPF facility (4 mice/condition). (H) Colonization of gastrointestinal tract in CR-WT mice 3 months after a single gavage in an SPF facility. Sample measures in the gray area denote mice where the bacteria were not detectable at a limit of detection of Log₁₀ 2. (4 mice/condition.) (I) Colonization of EcAZ-2 and EcAZ-2^IL10+ in the gastrointestinal tract of CR-WT mice housed in an SPF facility 79 days after gavage (same mice depicted in Figure 2E; 5/condition). All error bars indicate standard error of the mean. Samples that were not detectable were excluded from mean and standard error of mean calculations. The marker covers some error bars in (E), (F), and (G). See also Figure S2.

Figure 3.. Native E. coli can be used to change luminal metabolome without measurable effects in the microbiome

(A) Summary diagram showing effects of BSH on fecal bile acid metabolism. The hypothesized effect of increased BSH on the different bile acid pools is shown with red arrows. A question mark denotes uncertainty. (B–F) (B) Total fecal bile acids, (C) primary bile acids, (D) primary conjugated bile acids, (E) primary unconjugated bile acids, and (F) secondary bile acids in fecal samples collected from mice 12 weeks after a single gavage with vehicle, EcAZ-2, and EcAZ-2^BSH+. Significant differences were determined by Kruskal-Wallis test with post hoc Dunn’s multiple comparison test comparing EcAZ-2 and EcAZ-2^BSH+. (G) Faith’s phylogenetic distance (a measure of α-diversity) from 16S rRNA gene sequencing performed on stool samples collected pre-treatment and at 2, 10, and 16 weeks post-gavage (left) and from the terminal ileum samples at the time of euthanasia (right). (H) Relative abundance of E. coli as detected by 16S rRNA gene sequencing performed on stool samples collected pre-treatment and at 2, 10, and 16 weeks post-gavage and from the terminal ileum samples at the time of euthanasia. (I) Principal coordinate analysis of weighted UniFrac β-diversity of the fecal samples collected pre-treatment and at 2, 10, and 16 weeks post-gavage. (J) Principal coordinate analysis of weighted UniFrac β-diversity of the terminal ileum microbiome at the time of euthanasia. (A) was created with BioRender.com. See also Figure S3.

Figure 4.. Native E. coli can be used to physiologically change the host and treat pathophysiological conditions

(A) Total, primary, and secondary serum bile acids from mice treated with a single gavage of vehicle, EcAZ-2, or EcAZ-2^BSH+ 12 weeks prior. (B) Serum levels of CA (top), TCA (middle), and the log₂ ratio of CA to TCA in mice treated with vehicle, EcAZ-2, or EcAZ-2^BSH+. (C) Serum levels of bMCA (top), TbMCA (middle), and the log₂ ratio of bMCA to TbMCA in mice treated with vehicle, EcAZ-2, or EcAZ-2^BSH+. (D) Hepatic gene expression of Fxr, Shp, Cyp7a1, and Cyp27a1, as determined by qRT-PCR. Significant differences were determined by a Student’s t test after normality verified through Q-Q plot. (E) Mouse weights of CR-WT mice treated with a single gavage of vehicle, EcAZ-2, and EcAZ-2^BSH+. (F) Fasting (16 h) and postprandial (30 min) glucose and insulin levels. Significant differences were determined by Kruskal-Wallis test with post hoc Dunn’s multiple comparison test comparing all three conditions done separately for fasted and fed measures. (G) Mouse weights of Ob/Ob mice treated with a single gavage of EcAZ-2 and EcAZ-2^BSH+. (H) Glucose tolerance test in Ob/Ob mice treated with a single gavage of EcAZ-2 and EcAZ-2^BSH+ 15 weeks prior. Significance determined with Mann-Whitney U test corrected for multiple comparisons. Inset shows area under the curve. Significance determined by Mann-Whitney U test. For (A), (B), and (C), significant differences were determined by a Kruskal-Wallis test with post hoc Dunn’s multiple comparison test comparing EcAZ-2 and EcAZ-2^BSH+. See also Figure S4.