Whole genome shotgun metagenomic sequencing to identify differential abundant microbiome features between dementia and mild cognitive impairment (MCI) in AD subjects

Abstract

AbstractBackgroundPrevious studies demonstrated differential abundances of gut microbiota between Alzheimer’s disease (AD) and normal controls (NCs). However, gut microbiota distribution between dementia and amnestic mild cognitive impairment (aMCI) in AD subjects lacks clear data. With whole‐genome shotgun metagenomic sequencing (WGS), we aimed to identify differentially abundant microbiota as well as metabolic pathways between dementia and aMCI in AD subjects.MethodFecal samples from AD subjects were collected at Shuang‐Ho Hospital, Taipei Medical University, Taiwan and analyzed by WGS technique. 16S rRNA sequencing data from NCs were acquired from Taiwan Microbiome Database. Mini‐Mental State Examination (MMSE) and clinical dementia rating (CDR) scores, food frequency questionnaire, and other principal characteristics were collected along with fecal samplings. Alpha and β diversities were estimated by Chao1 and Shannon indices, and PCoA, respectively. We applied univariate tests (t‐test/Mann‐Whitney) and other modeling approaches (HAllA, MaAsLin, DESeq2, and LEfSe) to identify differential microbial biomarkers.ResultA total of 48 AD (38 dementia and 10 aMCI) and 50 NC samples were analyzed. Alpha diversities at the taxonomy level of Family, Class, and Genus were significantly different between AD and NC. With the results commonly confirmed in the MaAsLin (age and sex‐adjusted) and other models, we found significantly increased abundances of phyla Firmicutes and Actinobacteria; but decreased abundances of Bacteroides and Acidobacteria in AD subjects (Figure 1). There was no significant difference in both α and β diversities between dementia and aMCI in AD subjects. However, lower MMSE (higher CDR) scores in the AD subjects were associated with higher abundances of strains in Blautia hydrogenotrophica, Anaerotruncus colihominis and Gordonibacter pamelaeae. In addition, metabolic pathway analyses by HAllA demonstrated microbial glucose and S‐Adenosyl‐L‐methionine (SAMe) pathways were associated with cognitive function scores (Table 1) and more enriched in aMCI (not significant).ConclusionWe identified significantly higher strains in AD subjects with lower MMSE (higher CDR) scores suggesting that these microbes may have deleterious effects to cognitive function. In addition, the findings of higher enrichment of microbial glucose and SAMe metabolism in aMCI compared to dementia may implicate their beneficial roles in cognitive function in AD subjects.