Gut microbial characteristics in poor appetite and undernutrition: a cohort of older adults and microbiota transfer in germ-free mice

Abstract

Background: Older adults are particularly prone to the development of poor appetite and undernutrition. Possibly, this is partly due to the aged gut microbiota. We aimed to evaluate the gut microbiota in relation to both poor appetite and undernutrition in community-dwelling older adults. Furthermore, we studied the causal effects of the microbiota on body weight and body composition by transferring faecal microbiota from cohort participants into germ-free mice.

Methods: First, we conducted a cross-sectional cohort study of 358 well-phenotyped Dutch community-dwelling older adults from the Longitudinal Aging Study Amsterdam. Data collection included body measurements, a faecal and blood sample, as well as extensive questionnaires on appetite, dietary intake, and nutritional status. Appetite was assessed by the Council of Nutrition Appetite Questionnaire (CNAQ) and undernutrition was defined by either a low body mass index (BMI) (BMI < 20 kg/m2 if <70 years or BMI < 22 kg/m2 if ≥70 years) or >5% body weight loss averaged over the last 2 years. Gut microbiota composition was determined with 16S rRNA sequencing. Next, we transferred faecal microbiota from 12 cohort participants with and without low BMI or recent weight loss into a total of 41 germ-free mice to study the potential causal effects of the gut microbiota on host BMI and body composition.

Results: The mean age (range) of our cohort was 73 (65-93); 58.4% was male. Seventy-seven participants were undernourished and 21 participants had poor appetite (CNAQ < 28). A lower abundance of the genus Blautia was associated with undernutrition (log2 fold change = -0.57, Benjamini-Hochberg-adjusted P = 0.008), whereas higher abundances of taxa from Lachnospiraceae, Ruminococcaceae UCG-002, Parabacteroides merdae, and Dorea formicigenerans were associated with poor appetite. Furthermore, participants with poor appetite or undernutrition had reduced levels of faecal acetate (P = 0.006 and 0.026, respectively). Finally, there was a trend for the mice that received faecal microbiota from older adults with low BMI to weigh 1.26 g less after 3 weeks (P = 0.086) and have 6.13% more lean mass (in % body weight, P = 0.067) than the mice that received faecal microbiota from older adults without low BMI or recent weight loss.

Conclusions: This study demonstrates several associations of the gut microbiota with both poor appetite and undernutrition in older adults. Moreover, it is the first to explore a causal relation between the aged gut microbiota and body weight and body composition in the host. Possibly, microbiota-manipulating strategies will benefit older adults prone to undernutrition.

Keywords: Appetite; Germ-free mice; Gut microbiota; Microbiota transfer experiment; Older adults; Undernutrition.

Conflict of interest statement

The authors declare no competing interests.

© 2022 The Authors. Journal of Cachexia, Sarcopenia and Muscle published by John Wiley & Sons Ltd on behalf of Society on Sarcopenia, Cachexia and Wasting Disorders.

Figures

Figure 1

Variables associated with alpha‐diversity and beta‐diversity in the human cohort. (A) Barplot showing the F‐statistic of the PERMANOVA model for each factor. Bars are coloured based on the type of variable. The R2 is noted in each bar, depicting the variance in Bray–Curtis beta‐diversity explained by each variable. Bold text indicated statistical significance (P < 0.05). (B and C) Boxplots of alpha‐diversity measures species richness, Shannon index, and FPD for participants with or without poor appetite (B) and with or without undernutrition (C). Boxplots show median (centre line), interquartile ranges (boxes), 1.5× the interquartile ranges (whiskers), and outliers. Differences in alpha‐diversity were tested with linear regression models, adjusted for age and Mediterranean Diet Score. ASMMI, appendicular skeletal muscle mass index; CESD, Centre for Epidemiologic Studies Depression; FPD, Faith's phylogenetic diversity; MMSE, mini‐mental state examination.

Figure 2

Log2 fold change of all bacterial taxa significantly associated with either poor appetite or undernutrition. Heatmap depicting the log2 fold change in bacterial abundance of taxa that are significantly associated with either undernutrition or poor appetite. Log2 fold change was calculated with DESeq models, either crude (first and fourth column), adjusted for age (second and fifth column), or adjusted for both age and Mediterranean Diet Score (third and sixth column). Blue cells depict a positive log2 fold change, indicating higher abundance in participants with undernutrition or poor appetite, whereas red cells depict a negative log2 fold change, indicating lower abundance in participants with undernutrition or poor appetite. Numbers behind taxa are arbitrary identifiers for clades within the specified rank. Asterisks indicate Benjamini–Hochberg‐adjusted P‐value < 0.01. Taxa were only considered if the mean abundance was >10. AC, age corrected; ADC, age and diet corrected.

Figure 3

Associations of faecal short‐chain fatty acids with poor appetite and undernutrition. (A) Ternary diagrams showing faecal propionate, acetate, and butyrate as compositional data. Each point represents a participant with or without poor appetite and with or without undernutrition. The closer the participant is plotted to one of the short‐chain fatty acids, the higher the concentration of that short‐chain fatty acid is relative to the concentrations of the other short‐chain fatty acids. (B and C) Boxplots of faecal acetate, butyrate, and propionate concentrations for participants with or without poor appetite (B) and with or without undernutrition (C). Boxplots show median (centre line), interquartile ranges (boxes), 1.5× the interquartile ranges (whiskers), and outliers. Differences were tested with Student's t‐test with Bonferroni correction.

Figure 4

Alpha‐diversity, beta‐diversity, and Blautia of each of the donor groups. (A) Boxplots of species richness, Shannon diversity index, and FPD for each of the donor groups and the remaining cohort. Differences among groups were tested with Kruskal–Wallis tests. Boxplots show median (centre line), interquartile ranges (boxes), 1.5× the interquartile ranges (whiskers), and outliers. (B) Principal coordinate plot based on Bray–Curtis dissimilarity, coloured for each of the donor groups and the remaining cohort. Donors seem equally distributed throughout the cohort; donor group does not explain a significant amount of variance based on PERMANOVA. (C) Boxplot of Blautia abundance. There is no difference between donor groups based on Kruskal–Wallis test. Boxplots indicate same parameters as (A). FPD, Faith's phylogenetic diversity; HC, mice that received faecal microbiota from human donors without low body mass index or substantial weight loss; LBMI, mice that received faecal microbiota from human donors with low body mass index; WL, mice that received faecal microbiota from human donors with substantial weight loss.

Figure 5

Weight, lean mass, and fat mass of mice during experiment. (A and B) Line graphs for average body weight in g (A), and % of baseline weight (B) of mice per group during experiment (error bars indicate standard error). (C and D) Boxplots of lean mass in % of body weight (C) and g (D) per group at baseline and after a 3 week follow‐up. (E and F) Same as (C) and (D), but for fat mass. Boxplot centre line indicates median, boxes indicate interquartile ranges, whiskers indicate 1.5× the interquartile ranges, and paired measurements are connected by grey lines. HC, mice that received faecal microbiota from human donors without low body mass index or substantial weight loss; LBMI, mice that received faecal microbiota from human donors with low body mass index; WL, mice that received faecal microbiota from human donors with substantial weight loss.

Figure 6

Blautia abundance in Week 3 mouse faecal samples. (A) Boxplot for Blautia abundance in Week 3 mouse faecal samples per group. (B) Scatterplot of Blautia abundance in Week 3 mouse faecal samples plotted against the weight difference in the mice from baseline to Week 3. Boxplot centre line indicates median, boxes indicate interquartile ranges, whiskers indicate 1.5× the interquartile ranges, and paired measurements are connected by grey lines. HC, mice that received faecal microbiota from human donors without low body mass index or substantial weight loss; LBMI, mice that received faecal microbiota from human donors with low body mass index; WL, mice that received faecal microbiota from human donors with substantial weight loss.