The microbiota-gut-brain axis is a bidirectional communication system that is poorly understood. Alzheimer's disease (AD), the most common cause of dementia, has long been associated with bacterial infections and inflammation-causing immunosenescence. Recent studies examining the intestinal microbiota of AD patients revealed that their microbiome differs from that of subjects without dementia. In this work, we prospectively enrolled 108 nursing home elders and followed each for up to 5 months, collecting longitudinal stool samples from which we performed metagenomic sequencing and in vitro T84 intestinal epithelial cell functional assays for P-glycoprotein (P-gp) expression, a critical mediator of intestinal homeostasis. Our analysis identified clinical parameters as well as numerous microbial taxa and functional genes that act as predictors of AD dementia in comparison to elders without dementia or with other dementia types. We further demonstrate that stool samples from elders with AD can induce lower P-gp expression levels in vitro those samples from elders without dementia or with other dementia types. We also paired functional studies with machine learning approaches to identify bacterial species differentiating the microbiome of AD elders from that of elders without dementia, which in turn are accurate predictors of the loss of dysregulation of the P-gp pathway. We observed that the microbiome of AD elders shows a lower proportion and prevalence of bacteria with the potential to synthesize butyrate, as well as higher abundances of taxa that are known to cause proinflammatory states. Therefore, a potential nexus between the intestinal microbiome and AD is the modulation of intestinal homeostasis by increases in inflammatory, and decreases in anti-inflammatory, microbial metabolism.IMPORTANCE Studies of the intestinal microbiome and AD have demonstrated associations with microbiome composition at the genus level among matched cohorts. We move this body of literature forward by more deeply investigating microbiome composition via metagenomics and by comparing AD patients against those without dementia and with other dementia types. We also exploit machine learning approaches that combine both metagenomic and clinical data. Finally, our functional studies using stool samples from elders demonstrate how the c microbiome of AD elders can affect intestinal health via dysregulation of the P-glycoprotein pathway. P-glycoprotein dysregulation contributes directly to inflammatory disorders of the intestine. Since AD has been long thought to be linked to chronic bacterial infections as a possible etiology, our findings therefore fill a gap in knowledge in the field of AD research by identifying a nexus between the microbiome, loss of intestinal homeostasis, and inflammation that may underlie this neurodegenerative disorder.

Intestinal homeostasis is tightly regulated by the reciprocal interaction between gut epithelium and adjacent mesenchyme. The mammalian Hippo-YAP pathway is intimately associated with intestinal epithelial homeostasis and re-generation; however, its role in postnatal gut mesenchyme remains poorly defined. We find that, although removal of the core Hippo kinases Lats1/2 or activation of YAP in adult intestinal smooth muscle has largely no effect; Hippo-YAP signaling in Gli1/PDGFR-expressing intestinal stromal cells is critical to maintain the stem cell niche. We show that YAP/TAZ activation drives over-proliferation and suppresses smooth muscle actin expression in the niche-forming Gli1+ mesenchymal progenitors. In addition, mesenchymal YAP/TAZ activation disrupts the epithelial-mesenchymal crosstalk by promoting Wnt ligand production, leading to epithelial Wnt pathway activation. Our data also reveal that YAP/TAZ are upregulated in the stroma during DSS-induced injury and stromal YAP activation promotes intestinal epithelial regeneration. Altogether, our data identify an essential requirement for stromal Hippo-YAP signaling in the stem cell niche during intestinal homeostasis.