Limosilactobacillus reuteri promotes the expression and secretion of enteroendocrine- and enterocyte-derived hormones

Abstract

Intestinal microbes can beneficially impact host physiology, prompting investigations into the therapeutic usage of such microbes in a range of diseases. For example, human intestinal microbe Limosilactobacillus reuteri strains ATCC PTA 6475 and DSM 17938 are being considered for use for intestinal ailments, including colic, infection, and inflammation, as well as for non-intestinal ailments, including osteoporosis, wound healing, and autism spectrum disorder. While many of their beneficial properties are attributed to suppressing inflammatory responses, we postulated that L. reuteri may also regulate intestinal hormones to affect physiology within and outside of the gut. To determine if L. reuteri secreted factors impact the secretion of enteric hormones, we treated an engineered jejunal organoid line, NGN3-HIO, which can be induced to be enriched in enteroendocrine cells, with L. reuteri 6475 or 17938 conditioned medium and performed transcriptomics. Our data suggest that these L. reuteri strains affect the transcription of many gut hormones, including vasopressin and luteinizing hormone subunit beta, which have not been previously recognized as produced in the gut epithelium. Moreover, we find that these hormones appear to be produced in enterocytes, in contrast to canonical gut hormones produced in enteroendocrine cells. Finally, we show that L. reuteri conditioned media promote the secretion of enteric hormones, including serotonin, GIP, PYY, vasopressin, and luteinizing hormone subunit beta, and identify by metabolomics metabolites potentially mediating these effects on hormones. These results support L. reuteri affecting host physiology through intestinal hormone secretion, thereby expanding our understanding of the mechanistic actions of this microbe.

Keywords: Lactobacillus; GIP; PYY; adipolin; enterocyte; enteroendocrine; hormone; kisspeptin; luteinizing hormone; small intestine; vasopressin.

FIGURE 1

Induced and uninduced NGN3‐HIOs differentially respond to L. reuteri treatment. (A) Overview of RNA‐Seq experiment. (1) L. reuteri conditioned media were prepared by growing L. reuteri 6475 and 17938 in LDM4 media to mid‐log phase. The bacterial cells were spun out. The resulting conditioned media were brought to neutral pH and rendered cell‐free by filtration. The conditioned media were then lyophilized and resuspended in HIO differentiation media. (2) These treatments were then placed onto uninduced or induced NGN3‐HIOs in transwells for 3 h. (3) The organoid cells were harvested, and the isolated RNA was sent for RNA‐Seq. Created with BioRender.com. (B) Multidimensional scaling of transcriptomic data using a Euclidean distance metric from NGN3‐HIOs induced or uninduced and treated with L. reuteri 6475, 17938, or LDM4 media control. Ellipses for illustration purposes are modeled from the data following a t‐distribution. (C) Enriched functional categories of differentially expressed genes in L. reuteri treatments over media alone. U6475 is L. reuteri 6475 versus media control in uninduced NGN3‐HIOs. I6475 is L. reuteri 6475 versus media control in induced NGN3‐HIOs. I17938 is L. reuteri 17938 versus media control in induced NGN3‐HIOs. Some functional groups are listed as belonging to two categories (see Table S3 for further details).

FIGURE 2

Hormone genes differentially expressed by L. reuteri. DEGs annotated as having hormonal function are shown. The genes are annotated with their function, whether they are secreted (s), a receptor (r), or intercellular (i), and the cluster to which they belong, as in Figure S4. The plot shows the log₂ fold change expression of the gene for the indicated comparison. The bars are colored using the log₁₀ scaled mean GeTMM counts to illustrate how abundantly expressed the gene is. Transparent overlays are used for genes not differentially expressed for the given comparison. Comparisons shown: U6475‐ULDM4, L. reuteri 6475 on uninduced HIOs compared to LDM4 media control; I6475‐ILDM4, L. reuteri 6475 on induced HIOs compared to LDM4 media control; I17938‐ILDM4, L. reuteri 17938 on induced HIOs compared to LDM4 media control; I6475‐I17938 L. reuteri 6475 compared to L. reuteri 17938 on induced HIOs; ILDM4‐ULDM4, LDM4 media control on induced versus uninduced HIOs; I6475‐U6475, L. reuteri 6475 on induced versus uninduced HIOs. For each, positive fold changes indicate genes upregulated by the condition listed first.

FIGURE 3

L. reuteri promotes the secretion of known enteroendocrine‐derived intestinal hormones. (A) (1) In order to measure the release of intestinal hormones from human intestinal organoids (HIOs), L. reuteri conditioned media are generated from mid‐log phase cultures of L. reuteri. These cultures are pH neutralized and rendered cell‐free. (2) L. reuteri conditioned media are then placed onto uninduced or induced NGN3‐HIOs. (3) Following incubation on the HIOs, the supernatant is collected and secreted hormones are measured by an ELISA. Created with BioRender.com. Secreted amylin (B), C‐peptide (C), ghrelin (D), GIP (E), PP (F), and PYY (G) measured from uninduced and induced NGN3‐HIOs in response to L. reuteri 6475 or 17938 conditioned media. Hormones in (B–G) were measured on the apical side only of the transwell. In (B–G), batches A and B from the RNA‐Seq experiment were pooled—each plot point represents the result of two organoid batches pooled together. (H) Serotonin released from the apical or basolateral side (as indicated) from uninduced and induced NGN3‐HIOs in response to L. reuteri 6475 or 17938 conditioned media. In H, shape denotes independent batches of organoids. Only p‐values <.1 are shown, with p < .05 being considered significant. Significance was determined using a Dunnett's test.

FIGURE 4

L. reuteri promotes the secretion of novel enterocytic hormones. (A) Gut Cell Atlas annotated UMAP of the adult jejunum (adapted from Danhof et al. ³³ ), highlighting the enteroendocrine marker CHGA (B), the enterocyte marker SI (C), vasopressin (AVP, D), luteinizing hormone subunit beta (LHB, E), and adipolin (C1QTNF12, F). (G) Adipolin visualized in human jejunal tissue. The scale bar represents 50 μm. Secretion of (H) vasopressin, (I) luteinizing hormone subunit beta, and (J) the lack of secretion of adipolin from whole human jejunal tissue using the method shown in Figure 3A except with ex vivo human jejunal intestinal tissue. Shape represents unique human intestinal donors. Significance was determined using a linear mixed model with p < .05 considered significant. Only p‐values <.05 are shown.

FIGURE 5

L. reuteri 6475 and 17938 conditioned media contain differential levels of metabolites. (A) Multidimensional scaling (MDS) using a Bray‐Curtis distance metric of the metabolome of L. reuteri conditioned media and the LDM4 media using normalized imputed data. Ellipses for illustration purposes are modeled from the data following a t‐distribution. (B) MDS as in A but with the metabolites fumarate and malate removed. (C) Heatmap of select metabolites significantly different between L. reuteri strains and from the media using normalized non‐imputed data. The color of the bar on the far left indicates the biochemical pathway to which the metabolites (rows) are associated. Samples (columns) are annotated by the batch of media from which they were prepared. Normalized heatmap values are shown on a red (high) to blue (low) scale, with gray used if the metabolite was not detected.  * Metabolite not confirmed based on a standard.