Laboratory mice engrafted with natural gut microbiota possess a wildling-like phenotype

Abstract

Conventional laboratory mice housed under specific pathogen-free (SPF) conditions are the standard model in biomedical research. However, in recent years, many rodent-based studies have been deemed irreproducible, raising questions about the suitability of mice as model organisms. Emerging evidence indicates that variability in SPF microbiota plays a significant role in data inconsistencies across laboratories. Although efforts have been made to standardize microbiota, existing microbial consortia lack the complexity and resilience necessary to replicate interactions in free-living mammals. We present a robust, feasible and standardizable approach for transplanting natural gut microbiota from wildlings into laboratory mice. Following engraftment, these TXwildlings adopt a structural and functional wildling-like microbiota and host physiology toward a more mature immune system, with characteristics similar to those of adult humans. We anticipate that adopting wild mouse-derived microbiota as standard for laboratory mouse models will improve the reproducibility and generalizability of basic and preclinical biomedical research.

Fig. 1. Natural gut microbiota of wildlings outcompete lab microbiota of the major globally leading vendors.

Fecal material from wildlings was transplanted by oral gavage into a lab mouse, which was housed with four lab mice in the same cage. The same experiment was performed with a wildling mouse transplanted with fecal material from lab mice. Fecal pellets from all mice were collected over 28 days and 16S rRNA gene profiling was performed. a Schematic representation of the experimental design. b PCoA based on beta diversity using Jaccard distance compares the microbiota of wildlings before and at the indicated time points post-transplantation with those of lab donor mice. c Relative abundance of the bacterial phyla of lab mice before and 28 days post-transplantation in comparison to the wildling donor mice. d PCoA based on beta diversity using Jaccard distance compares the microbiota of wildlings before and at the indicated time points post-transplantation with those of lab donor mice. e Relative abundance of the bacterial phyla of wildlings before and 28 days post-transplantation compared to the lab donor mice. f Lab mice were purchased from different vendors and transplanted with wildling fecal material. PCoA based on Jaccard distance shows samples from lab mice of the respective vendors before and 28 days post-transplantation, compared to the samples from the wildling donor mice. g Lab mice were purchased from different vendors and used for transplantation to wildlings. PCoA based on Jaccard distance shows samples from wildlings before and 28 days post-transplantation, compared to the samples from the respective lab donor mice. b–e Data are from three independent experiments, with five donor mice per transplantation and one recipient mouse co-housed together with four cage mates in each experiment. f, g Data are from one experiment. Figure in (a) was created in BioRender. Bruno, P. (2025) https://BioRender.com/60kbw9g. TX: Transplantation.

Fig. 2. The microbiota of TXwildlings closely resemble the functional capability of wildling microbiota and differ significantly from lab microbiota.

Global metabolomics analysis was performed on the cecal content of lab mice, TXwildlings and wildlings. a PCA of all detected metabolites was used to compare lab mice with TXwildlings and wildlings. b The heatmap shows the metabolites significantly different between lab mice and wildlings. Color code depicts the row-wise scaled (z-score) metabolite intensity. PC: Principal component.

Fig. 3. The gastrointestinal tract of TXwildlings phenocopies wildlings and differs significantly from lab mice.

The engraftment of fecal material from wildlings into lab mice resulted in the TXwildlings group, which was then compared to lab mice and wildlings. a Weight of the cecum in lab mice, TXwildlings, and wildlings. b–d RNA sequencing was performed on colon tissue from lab mice, TXwildlings, and wildlings. b PCA of samples from the three groups. The ellipses represent 95% confidence intervals. c Volcano plots compare lab mice and TXwildlings (left panel) or wildlings and TXwildlings (right panel). The number of significantly regulated genes is indicated. The dashed line represents q = 0.05. d GSEA using Gene Ontology (Biological Processes aspect) shows the top ten most enriched gene sets in lab versus TXwildlings according to the normalized enrichment score (NES). The color indicates the significance; only significantly regulated gene sets (q < 0.05) are shown. e Cytokine and chemokine concentrations were analyzed in colon tissue by a multiplex MSD assay. Color code depicts the row-wise scaled (z-score) cytokine concentration. All cytokines that were above the detection limit in >50% of the samples are shown in the heatmap. Data are from two (a–d) or three (e) independent experiments with n = 5 mice per group. Median ± interquartile range (IQR) is shown. Statistical significances in (a) were tested by unpaired, two-tailed Mann–Whitney U tests, p < 0.05 is shown. Source data are provided as a Source Data file. PC Principal component. NES Normalized enrichment score.

Fig. 4. Natural gut microbiota transplantation creates TXwildlings with systemic wildling-defining characteristics.

a Weight of mLN was measured from lab mice, TXwildlings, and wildlings. b mLN from lab mice, TX wildlings, and wildlings were subjected to RNA sequencing. PCA includes the genes significantly regulated between lab mice and wildlings. c, d Livers from mice of the three groups were subjected to RNA sequencing. c PCA was performed with the genes significantly regulated between lab mice and wildlings. d GSEA of lab mice vs. TXwildlings was conducted using Gene Ontology (Biological Processes aspect). The graph shows the top ten most enriched gene sets according to the NES. The color indicates the significance level; only gene sets significantly regulated (q < 0.05) are shown. e The concentrations of the T cell chemoattractants CXCL10 and CCL5 were measured in liver homogenates using an MSD multiplex assay. f The livers of mice from the three groups were subjected to multiplex immunofluorescence staining and CD4- and CD8-positive cells were quantified. g, h The lungs of lab mice, TXwildlings, and wildlings were subjected to RNA sequencing. g PCA using the genes significantly regulated between lab mice and wildlings. h GSEA using Gene Ontology (Biological Processes aspect) of lab mice vs. TXwildlings shows the top ten most enriched gene sets according to the NES. The color indicates the significance, only q < 0.05 is shown. i The heatmap shows the significantly regulated genes (q < 0.05) of the “adaptive immune response” gene set. Color code depicts the row-wise scaled (z-score) RNA normalized expression. j The concentration of cytokines and chemokines was analyzed in lung tissue by a multiplex MSD assay. Color code depicts the row-wise scaled (z-score) cytokine concentration. All cytokines that were above the detection limit in >50% of the samples are shown in the heatmap. Ellipses in (b), (c), and (g) represent the 95% confidence interval. Data are from two (b–d, f–i) or three (a, e, and j) independent experiments with n = 5 per group. Median ± IQR is shown, box plots show the median, IQR, and full data spread via whiskers. Statistical significances in (a) were tested by unpaired, two-tailed Mann–Whitney U tests, p < 0.05 is shown. Source data are provided as a Source Data file. mLN Mesenteric lymph node, PC Principal component, NES Normalized enrichment score, GOBP Gene ontology – biological process.

Fig. 5. The transfer of natural gut microbiota induces systemic immune maturation in TXwildlings.

a, b Blood cells were isolated, analyzed by flow cytometry, and gated for live CD45⁺CD3⁺CD8⁺ cells. a CD8⁺ T cell subsets were identified by their expression of CD44 and CD62L, as illustrated in the left panel. Specifically, T_naive (CD62L⁺CD44^-), T_CM (CD62L⁺CD44⁺), and T_EM (CD62L^-CD44⁺) were analyzed as percentage of total CD8⁺ cells. b KLRG1⁺ cells were gated, as shown in the left panel, and analyzed as percentage of total CD8⁺ cells. c Concentrations of serum cytokines were quantified using a multiplex MSD assay. d Antibody subclasses were measured in the serum of lab mice, TXwildlings, and wildlings. The shown data are from two (c) or three (a, b, d) independent experiments with n = 5 mice per group. Dashed lines in (c) indicate the detection limit. Median ± IQR is shown, box plots show the median, IQR, and full data spread via whiskers. Statistical significances were tested by unpaired, two-tailed Mann–Whitney U tests, p < 0.05 is shown. Source data are provided as a Source Data file.

Fig. 6. TXwildlings and wildlings display comparable physiological responses and possess an adult human-like transcriptome.

Lab mice, TXwildlings, and wildlings were orally infected with C. rodentium, the body weight was analyzed until day 28 post-infection (a), and the fecal bacterial burden was analyzed at days 4 and 7 post-infection (b). a Body weight loss is shown over time and for the greatest significant weight difference on day 15 post-infection. c, d GSEA among the indicated mouse comparisons of uninfected lab mice, TXwildlings and wildlings compared to a ranked gene list derived from healthy adult vs. neonatal PBMCs. Gene signatures include the top 500 significantly differently expressed genes of colon or lung tissues of TXwildlings vs. lab mice (dark blue), wildling vs. lab mice (green) or lab mice vs. TXwildlings (light blue). Data are from two independent experiments with n = 5 per group. Dashed lines in (b) indicate the detection limit. Mean ± SEM is shown. Statistical significances in (a), right panel and (b) were tested by unpaired, two-tailed Mann–Whitney U tests, p < 0.05 is shown. Figure in (c) was created in BioRender. Bruno, P. (2025) https://BioRender.com/2y4432p. Source data are provided as a Source Data file. CFU: Colony-forming units.