Handling stress may confound murine gut microbiota studies

Abstract

Background: Accumulating evidence indicates interactions between human milk composition, particularly sugars (human milk oligosaccharides or HMO), the gut microbiota of human infants, and behavioral effects. Some HMO secreted in human milk are unable to be endogenously digested by the human infant but are able to be metabolized by certain species of gut microbiota, including Bifidobacterium longum subsp. infantis (B. infantis), a species sensitive to host stress (Bailey & Coe, 2004). Exposure to gut bacteria like B. infantisduring critical neurodevelopment windows in early life appears to have behavioral consequences; however, environmental, physical, and social stress during this period can also have behavioral and microbial consequences. While rodent models are a useful method for determining causal relationships between HMO, gut microbiota, and behavior, murine studies of gut microbiota usually employ oral gavage, a technique stressful to the mouse. Our aim was to develop a less-invasive technique for HMO administration to remove the potential confound of gavage stress. Under the hypothesis that stress affects gut microbiota, particularly B. infantis, we predicted the pups receiving a prebiotic solution in a less-invasive manner would have the highest amount of Bifidobacteria in their gut.

Methods: This study was designed to test two methods, active and passive, of solution administration to mice and the effects on their gut microbiome. Neonatal C57BL/6J mice housed in a specific-pathogen free facility received increasing doses of fructooligosaccharide (FOS) solution or deionized, distilled water. Gastrointestinal (GI) tracts were collected from five dams, six sires, and 41 pups over four time points. Seven fecal pellets from unhandled pups and two pellets from unhandled dams were also collected. Qualitative real-time polymerase chain reaction (qRT-PCR) was used to quantify and compare the amount of Bifidobacterium, Bacteroides, Bacteroidetes, and Firmicutes.

Results: Our results demonstrate a significant difference between the amount of Firmicutes in pups receiving water passively and those receiving FOS actively (p-value = 0.009). Additionally, we found significant differences between the fecal microbiota from handled and non-handled mouse pups.

Discussion: From our results, we conclude even handling pups for experimental purposes, without gavage, may induce enough stress to alter the murine gut microbiota profile. We suggest further studies to examine potential stress effects on gut microbiota caused by experimental techniques. Stress from experimental techniques may need to be accounted for in future gut microbiota studies.

Keywords: Bifidobacteria; Methods; Microbiota; Milk oligosaccharides; Mouse study.

Figure 1. Timeline for each treatment group.

Samples were collected at time points post-natal day (PND)0, PND7, PND14, and PND21.

Figure 2. Daily dosage of FOS or water in microliters.

Dosage of water or FOS increased from 2 microliters (µL) to 12 microliters (µL) over the course of the experiment.

Figure 3. log10 Colony-forming unit (CFU) equivalents/ng of sample DNA for control and treatment groups.

Wilcoxon rank sum test demonstrated significant differences between the control and treatment samples for all Bifidobacterium (p < 0.001), Bacteroides (p < 0.001), and Bacteroidetes (p-value = 0.008).There was also a significant difference for Firmicutes between the WP and FA groups (p-value = 0.009). W-P, water passive, W-A, water active; F-A, FOS active; F-P, FOS passive.

Figure 4. log10 colony-forming unit (CFU) equivalents/ng of sample DNA for feces collected from the GI tracts of treatment litter sires.

Samples from water sires contained Bifidobacteria, while samples from FOS sires did not.

Figure 5. The log10 colony-forming units (CFU) equivalents/ng of sample DNA for control and experimental dams.

There was no significant difference between the samples, despite the treatment dams having no bifidobacteria. The F3 dam was the buccal water dam that did not deliver a litter. W-P, passive water; W-A, active water; F-A, active FOS; F-P, passive FOS.

Figure 6. The log10 colony-forming unit (CFU) equivalents/ng of sample DNA from control and treatment samples collected on post-natal day 14.

The median Bifidobacterium and Bacteroides counts in the control samples were significantly different from the treatment samples (p-value = 0.03 for both comparisons) as tested by Wilcoxon rank sum test. The medians of Bacteroidetes counts were not significantly different between the treatment and control groups (p-value = 0.057). Firmicutes could be compared between control and treatment groups, because there were significant differences between treatment groups for this taxa. W-P, passive water; W-A, active water; F-A, active FOS; F-P, passive FOS.