https://orcid.org/0000-0003-0810-4173
This editorial refers to ‘Influence of angiotensin II on the gut microbiome: modest effects in comparison to experimental factors’, by R.R. Muralitharan et al., https://doi.org/10.1093/cvr/cvae062.
The collection of microbes inhabiting the gut is collectively termed as the gut microbiota (GM). They comprise of approximately 103 microbes, belonging to at least 160 species,1 and inhabit the entire length of both the small and large intestines. Intrinsic factors like one’s genetic predispositions and sex, as well as extrinsic factors such as the environment, food, and drug intake influence the composition of the GM. In a similar manner, the onset of numerous diseases, including angiotensin II (Ang II)-induced hypertension, is known to influence imbalances in the GM composition,2,3 referred to as dysbiosis.
In human studies, there seems to be a general lack of reproducibility and robustness in identifying specific bacterial species that are differentially abundant. Even though some bacteria such as Klebsiella spp., Enterobacter spp., and Actinomyces spp. often show trends of higher prevalence during hypertension irrespective of the sequencing method,4–6 the bulk of the differentially abundant bacteria varies. This inconsistency could be due to inter-individual variability and to general unconfounded experimental factors. It is therefore difficult to define responsible microbes that are truly associated with the disease. Researchers have therefore been commonly using broader taxonomical classifications such as an increase in Firmicutes/Bacteroidetes ratio to identify hypertension-associated dysbiosis.2 Identifying pathogenic bacteria for therapeutic purposes would be challenging with such a general classification.
In the recent issue of Cardiovascular Research, Muralitharan et al.7 shed light onto the extra-experimental factors, apart from Ang II treatment, influencing the GM that could contribute to the reason behind this lack of reproducibility. They highlighted the GM’s vulnerability to factors such as host genotype, housing facility, age, sex, diet, intestinal region of sample isolation, and sequencing batch (Figure 1), whose modulatory effect on the microbiome was more profound than the modest influence of Ang II treatment itself. This study therefore marks a significant step towards addressing a problem concerning untested variables, as it quantifies their influence through a retrospective analysis of a large sample cohort (538 samples from 303 mice) obtained from independent experiments.
The influence of experimental factors on gut microbiome diversity. Using 538 mice samples originating from independent experiments showed that the region of intestinal compartment, diet, and sequencing batch accounted for a large percentage of overall variance in the gut. However, the actual experimental treatment imparted on the mice, angiotensin II treatment, caused only 0.4% of the total variance, the least contribution compared to all the tested experimental factors. The genotype, housing facility, age, and sex of the mice also contributed to the overall variation. While having significant effects on the gut microbiome, the influence of these experimental factors on the microbial metabolites is yet to be elucidated. The figure was generated with the aid of BioRender.com.
The authors sequenced the V4 region of the 16S rRNA gene to obtain amplicon sequencing variants and found that the above mentioned variables collectively accounted for approximately half (48.8%) of the variations observed in the gut microbial composition. By assessing the α- and β-diversity indices, they found that these variables significantly influenced the diversity and evenness within a sample, as well as the bacterial composition itself. Among these variables, the greatest individual contributions stemmed from the intestinal region of sample isolation (6.8%), dietary fibre content (6%), and sequencing batch effects (4.7%). Surprisingly, treatment with Ang II accounted for only 0.4% of the observed variance, which was notably minimal in comparison to the rest. Furthermore, through multiple regression analysis, the authors found that Ang II was not driving the shift in Shannon index (α-diversity), instead, intestinal compartment of isolation, housing facility, and the age of the mice were. The authors thereby underscored the necessity of multiple controls to account for these confounding factors in microbial research, while at the same time encouraging the readers to critically evaluate the conclusions drawn from studies concerning microbial dysbiosis.
While the study offers valuable insights, it simultaneously provokes consideration regarding the feasibility of controlling these variables in human studies. Some variables such as age and sex can readily be controlled for. However, matching variables such as genetic background, co-housing, and xenobiotic intake8 could pose challenges and would need to rely on in silico corrections. Furthermore, within the context of translatability, the significance of evaluating functional relevance over species relevance should be considered. This can be reflected in the metabolites produced by the GM. It would therefore be insightful to assess the impact of these confounders not only on species diversity but also on the metabolomic profile of the host. Such a comparison could be particularly relevant when extrapolating experimental variables from animal models, such as mice, to the human setting. The authors touched upon this issue by demonstrating that the experimental factors they tested did indeed influence the abundance of short chain fatty acid producing bacteria, thereby suggesting modulations in the metabolite’s amount. Declines of these dietary fibre-obtained metabolites are widely recognized in association with an increase of blood pressure.9,10
Overall, the authors beautifully captured and quantified the complexity of microbial research, particularly in pre-clinical animal settings. Their work serves to emphasize the necessity of implementing rigorous controls, which one might have initially overlooked, as well as interpret microbial results with careful consideration. While the study focused primarily on changes in bacterial species, alterations in the host’s metabolomic profile, an indication of bacterial functional activity, remains to be addressed by future research.
Funding
This work was supported by funding from the Bundesministerium für Bildung und Forschung (project number 01KL2005) to S.S.
Data availability
No new data were generated or analysed in support of this editorial statement.
References
Author notes
The opinions expressed in this article are not necessarily those of the Editors of Cardiovascular Research or of the European Society of Cardiology.
Conflict of interest: none declared.
