Identifying core microbiotas in the human donors
To compare the establishment of human fecal bacterial communities in HMA mice and piglets, we inoculated GF mice and piglets maintained in gnotobiotic isolators with fecal matter from four separate human donors. The donors selected had diverse microbial communities (Fig. 1) and represented different stages of human development (see “Methods” for donor information). All animals in a given isolator (for both mice and piglets) were inoculated with the inocula obtained from a single donor. Both recipient species of animals were inoculated twice during the study—the initial round of inoculations were performed after weaning and the second round of inoculations occurred two weeks after the first round of inoculations. All inocula were prepared at the same time under the same conditions and both mice and piglets were fed the exact same sterile solid diet.
Statistical comparisons were performed using the Wilcoxon rank-sum test. Boxes with different letters indicate statistically significant differences (p < 0.05). The box represents the 25th and 75th percentiles as the interquartile range (IQR) and the short black line represents the median. The whiskers represent the minimum and maximum values. Outliers are shown as black dots. n = 3 inoculum aliquots for Donor_1 and n = 4 inoculum aliquots or Donor_2, Donor_3, and Donor_4.
To compare the engraftment of donor microbiota in the two HMA animal models, the establishment of amplicon sequence variants (ASVs) identified in the donor inocula was evaluated. To reduce variation in composition due to sequencing depth, a ‘core’ bacterial community was defined for each donor by identifying ASVs that were found in all aliquots of the pooled donor inoculum used to inoculate the corresponding HMA animals (see “Methods” for details on donor inoculum preparation). There were a total of 4 inoculum aliquots per donor – two used for inoculating the GF mice and two used for inoculating the GF piglets. All 4 inoculum aliquots were sequenced separately. For Donor_1, one of the inoculum aliquots used to inoculate the corresponding GF piglets did not have sufficient material for DNA extraction, which resulted in only three inoculum aliquots from this donor for subsequent analysis. Twenty-six, 76, 140, and 134 core ASVs that fit the above criteria were identified from Donor_1, Donor_2, Donor_3, and Donor_4, respectively. As a percentage of the total number of reads, the core ASVs identified accounted for 99%, 96%, 98%, and 96% of the total reads in the 4 donor inocula used in the study. The taxonomy and relative abundance information of the core ASVs identified for each donor across the animal models is provided in Supplementary Data 1. Since core ASVs identified are found in all inoculum aliquots of a given donor and represent the major portion of the reads generated (96–99%), we reasoned that these taxa can be considered as the major taxa within the human donors having the highest probability of colonizing each animal model. With the exception of alpha diversity analysis, the remainder of the results presented are focused on the core ASVs identified.
Global diversity patterns in the HMA animal models
To evaluate global bacterial community differences, we compared the establishment of the core microbiotas of the corresponding human donors in the two HMA animal models using Principal coordinate analysis (PCoA) plots based on unweighted UniFrac and Bray-Curtis distances (Fig. 2). For Donor_1, the Bray-Curtis distance-based analysis revealed the donor inoculum aliquots to cluster more closely to the HMA mouse fecal samples than the HMA piglet fecal samples (Fig. 2a). However, for Donor_2, Donor_3, and Donor_4, the donor inoculum aliquots were more similar to the HMA piglet fecal samples than the HMA mouse fecal samples (Fig. 2b–d). As previously reported25, we also compared microbial community engraftment using UniFrac distances and Bray–Curtis distances (Fig. 3), which further corroborated the results from the beta diversity analysis.
(a) Donor_1, (b) Donor_2, (c) Donor_3, (d) Donor_4. n = 13 (3 piglets/donor, with the exception n = 4 for Donor_1, 2 days post inoculation) and n = 37 mice (Donor_1 = 7, Donor_2 = 10, Donor_3 = 10, Donor_4 = 10). Red circles, human donor inocula; green triangles, HMA mouse fecal samples; blue squares, HMA piglet fecal samples.
(a) Unweighted UniFrac distances and (b) Bray–Curtis distances. The box represents the 25th and 75th percentiles as the interquartile range (IQR) and the short black line represents the median. The whiskers represent the minimum and maximum values. Outliers are shown as black dots. Statistical comparisons were performed using the Wilcoxon rank-sum test. ***p < 1 × 10−5; ****p < 1 × 10−10; ns–not significant. n = 13 (3 piglets/donor, with the exception n = 4 for Donor_1, 2 days post inoculation) and n = 37 mice (Donor_1 = 7, Donor_2 = 10, Donor_3 = 10, Donor_4 = 10).
To evaluate the weekly variation in alpha diversity of recipient microbiotas, we compared the alpha diversity of the HMA animals with that of their respective human donors at each time point (Fig. 4). Since alpha diversity analyses are sensitive to the presence/absence of low prevalence taxa, we considered all ASVs (not only core ASVs) of a given donor for this analysis. Donor_1-HMA mice maintained a stable alpha diversity over time, with only the fecal samples from week 4 showing a Shannon index significantly different (p = 0.018) from that of Donor_1 which represents an infant microbiota. However, greater variation in alpha diversity was observed in Donor_1-HMA piglet fecal samples with days 2, 7, 14, 21, and 40 post-inoculation yielding significantly different Shannon indexes compared to the Donor_1 inoculum aliquots (Fig. 4a). Diversity estimates for Donor_2-HMA and Donor_3-HMA piglet fecal samples were maintained similar to that of the estimates for the human donors throughout the study (Fig. 4b, c). However, significantly lower Shannon index values for Donor_2-HMA and Donor_3-HMA mouse fecal samples were observed compared to the donor inoculum aliquots. Donor_4-HMA piglet fecal microbiota displayed notable variations in alpha diversity throughout the study period (Fig. 4d). The Donor_4-HMA mouse fecal samples did not show much variation in alpha diversity between sampling time points although the Shannon indexes were significantly lower relative to the donor inoculum aliquots.
(a) Donor_1, (b) Donor_2, (c) Donor_3, and (d) Donor_4. The box represents the 25th and 75th percentiles as the interquartile range (IQR) and the short black line represents the median. The whiskers represent the minimum and maximum values. Outliers are shown as black dots. Statistical comparisons are based on the Wilcoxon rank-sum test. *p < 0.05; **p < 0.01; ns–not significant. n = 13 (3 piglets/donor, with the exception n = 4 for Donor_1, 2 days post inoculation) and n = 37 mice (Donor_1 = 7, Donor_2 = 10, Donor_3 = 10, Donor_4 = 10).
Establishment of core donor ASVs in the two HMA animal models
Heatmaps (Supplementary Figs. 1–4) and chord diagrams (Fig. 5) were generated to visualize the establishment of the core donor ASVs in HMA mice and piglets, which revealed notable differences in their success of engraftment in the two species. Notably, certain ASVs only established in one of the two animal models, while some others did not establish in either. With the exception of Donor_1, more core ASVs found in the human donors were identified in the HMA piglets compared to the HMA mice.
(a) Donor_1, (b) Donor_2, (c) Donor_3, (d) Donor_4. Each sector represents an ASV and the size of the sector corresponds to the mean relative abundance of that ASV in the host species. Links indicate which core donor ASVs established in each animal model and the widths at each end of a link are proportional to the abundance of that ASV in the respective host species. Core donor ASVs which failed to colonize both HMA animal models are not shown. ASVs have been designated according to the following example: ASV_1 is ASVH1 in human donors, ASVM1 in HMA mice, and ASVP1 in HMA piglets. n = 13 (3 piglets/donor, with the exception n = 4 for Donor_1, 2 days post inoculation) and n = 37 mice (Donor_1 = 7, Donor_2 = 10, Donor_3 = 10, Donor_4 = 10).
To determine how successfully the core ASVs found in human donors established and persisted in the corresponding HMA animal models, we defined ‘colonization criteria’, which were established based on the number of animals and sampling time points in which a donor ASV was detected. If a core donor ASV was detected in at least one fecal sample of a given HMA animal model, then this ASV was considered as a ‘colonizer.’ However, if a donor ASV was detected in fecal samples of >50% of the animals in at least 4 of the 7 sampling time points, then this ASV was considered as a ‘persistent colonizer’ for that animal model (Table 1 and Supplementary Data 2). Relative abundance and taxonomic information related to persistent colonizers identified following these criteria for each donor and animal model are provided in Supplementary Data 3.
For Donor_1, a single ASV out of the five core ASVs that were not detected in the HMA mice was detected in the HMA piglets. Similarly, 2 ASVs of the 6 core Donor_1 ASVs that were not detected in the HMA-piglets were detected in the HMA mice. Four ASVs (15%) failed to colonize either animal model.
For Donor_2 and Donor_3, a higher proportion of core donor ASVs established in the piglets compared to the mice (Table 1). Twenty-eight of the 39 Donor_2 core ASVs that failed to colonize the mice colonized the piglets. The 11 ASVs that were not detected in the piglets were also absent in the mice. Only 3 out of the 34 Donor_3 core ASVs that failed to colonize the HMA piglets colonized the HMA mice. However, 43 out of the 74 core donor ASVs that failed to colonize the HMA mice colonized the HMA piglets. Eleven (14%) and 31 (22%) core ASVs failed to colonize either animal model for Donor_2 and Donor_3, respectively.
For Donor_4, Forty-one out of the 79 core ASVs that did not colonize the HMA mice colonized the HMA piglets. Only 4 out of the 42 ASVs that failed to colonize the HMA piglets colonized the HMA mice. Thirty-eight (28%) core Donor_4 ASVs failed to establish in either animal model. The taxonomic classifications of the core ASVs in each donor that failed to colonize both HMA animal models are provided in Supplementary Data 4.
Establishment of common ASVs found among donors in the two HMA animal models
With the assumption that ASVs common to all donors may represent part of the autochthonous microbes in the human gut, we evaluated the establishment of such ASVs in the two HMA models. However, Donor_1 was not considered for this analysis as the Donor_1 microbiome is from an infant, and infants are known to have unstable and highly dynamic gut microbiota compositions26,27,28. We identified 27 core donor ASVs (Supplementary Data 5) that were shared by all 12 inoculum aliquots originating from Donor_2, Donor_3, and Donor_4 (4 aliquots for each donor). Twenty-four of these shared ASVs were classified as members of the phylum Firmicutes and 2 were classified as members of phylum Bacteroidetes. The remaining ASV belonged to the phylum Desulfobacterota (Supplementary Data 5). The 27 common ASVs identified were compared between the two HMA models in terms of how many ‘colonizers’ and ‘persistent colonizers’ were present. The results of this analysis (Supplementary Data 6) revealed that, for all three donors, the 27 shared core ASVs established more efficiently in the HMA piglets compared to the HMA mice. Specifically, 24 of these 27 ASVs (89%) had established in the Donor_2-HMA piglets, while 25 (93%) and 19 (70%) had colonized the HMA piglets inoculated with Donor_3 and Donor_4, respectively. In contrast, only 9 (33%), 15 (55%), and 10 (37%) of these ASVs had colonized the HMA mice for Donor_2, Donor_3, and Donor_4, respectively. In terms of persistent colonizers, 16 (59%), 20 (74%), and 6 (22%) of the 27 common ASVs were identified as persistent colonizers of the piglets inoculated with Donor_2, Donor_3, and Donor_4, respectively. For the same donors, 5 (18%), 6 (22%), and 5 (18%) ASVs were persistent colonizers of the mice (Supplementary Data 6).
Members of the phylum Firmicutes establish in the HMA piglets with greater success compared to the HMA mice while Bacteroidetes successfully colonize both animal models
We compared core donor taxa establishment at phylum, family, and genus levels between the two HMA animal models for each donor (Supplementary Data 7). At the phylum level, the most notable differences between the two animal models were observed for the phylum Firmicutes. Donor_2 had 58 core ASVs belonging to Firmicutes, and only 21 (36%) of these ASVs established in the HMA mice; however, 48 (83%) colonized the HMA_piglets. Out of the 48 core Firmicutes ASVs that established in the HMA piglets, 31 (65%) were persistent colonizers, while 10 of the 21 (48%) Firmicutes ASVs persistently colonized the HMA mice (Supplementary Data 2). The Donor_3 inocula contained 97 core Firmicutes ASVs, 71 (73%) of which established in HMA piglets and 31 (32%) established in the HMA mice. Only 6 out of the 31 (19%) core Firmicutes ASVs colonizing the HMA mice were persistent colonizers, whereas 44 of the 71 (62%) Firmicutes ASVs persistently colonized the HMA piglets. Similarly, out of the 95 core Firmicutes ASVs found in the Donor_4 inocula, 21 (22%) colonized the HMA mice and 57 (60%) colonized the HMA piglets. Eight of the 21 (38%) Firmicutes core ASVs that established in the HMA mice were persistent colonizers, while 15 of the 57 (26%) Firmicutes ASVs persistently colonized the HMA piglets. Phylum Firmicutes was represented in 18 ASVs and was the predominant phylum in the Donor_1 core with a mean abundance of 62.77% (SD = 5.18%). Fourteen and thirteen of these Firmicutes ASVs colonized the HMA mice and HMA piglets, respectively, with mean relative abundances of 40.5% (SD = 11.15%) and 10.72% (SD = 7.6%), respectively (Supplementary Data 7).
The differential establishment of Firmicutes in the two HMA animal models was observed at the family level by members of Lachnospiraceae and Ruminococcaceae (Supplementary Data 7). Of the 34 Donor_2 core Lachnospiraceae ASVs, only 11 (32%) were detected in the HMA mice while 28 (82%) were detected in the HMA piglets. Sixteen of the 28 (57%) Lachnospiraceae ASVs detected in the piglets were persistent colonizers, while 6 of the 11 (54%) Lachnospiraceae ASVs that established in the mice were persistent colonizers (Supplementary Data 2). Similarly, the Donor_3 core consisted of 40 Lachnospiraceae ASVs, 15 (37%) of which colonized the HMA mice and 31 (77%) which colonized the HMA piglets. Only 3 out of the 15 (20%) Lachnospiraceae ASVs colonizing the HMA mice were persistent colonizers. In contrast, 20 out of the 31 Lachnospiraceae ASVs (64%) were persistent colonizers in HMA piglets. The Donor_4 core contained 40 Lachnospiraceae ASVs; only 7 of these ASVs (17%) established in the HMA mice, whereas 22 (55%) established in the HMA piglets. Two core Lachnospiraceae ASVs (29%) persistently colonized the HMA mice, while 4 out of the 22 (18%) core Lachnospiraceae ASVs colonizing the HMA piglets were persistent colonizers (Supplementary Data 2). Similar to Lachnospiraceae, core donor ASVs belonging to the family Ruminococcaceae were poorly established in the HMA mice relative to the HMA piglets (Supplementary Data 7). Donor_2 contained 6 core ASVs belonging to Ruminococcaceae. All 6 of these ASVs (100%) successfully colonized the HMA piglets, while only 1 (17%) colonized the HMA mice. Four of the 6 (67%) core Ruminococcaceae ASVs that successfully colonized the HMA piglets were identified as persistent colonizers (Supplementary Data 2). Out of the 17 core ASVs belonging to Ruminococcaceae found in Donor_3, 12 (71%) colonized the HMA piglets, while only 6 (35%) colonized the HMA mice. One of these 6 core Ruminococcaceae ASVs (17%) persistently colonized the mice. However, 9 out of the 12 (75%) Donor_3 core Ruminococcaceae ASVs that established in the piglets were persistent colonizers. Donor_4 consisted of 14 core Ruminococcaceae ASVs but only 4 (29%) colonized the HMA mice. In contrast, 12 (86%) core Ruminococcaceae ASVs colonized the piglets. Three of the 4 (75%) core Ruminococcaceae ASVs persistently colonized the mice, whereas 5 out of the 12 (42%) core Ruminococcaceae ASVs persistently colonized the piglets (Supplementary Data 2).
The Firmicutes family Christensenellaceae, which has recently been identified as the most heritable bacterial family in the human microbiota and also related to health benefits29,30, successfully colonized the HMA piglets but was not detected in the HMA mice (Supplementary Data 7). Specifically, Christensenellaceae ASVs were identified as core members only in Donor_3 and Donor_4 inocula, with 6 core ASVs in Donor_3 and 7 core ASVs in Donor_4. Out of the 6 core Christensenellaceae ASVs in Donor_3, 4 established in the HMA piglets–3 of these were persistent colonizers. In addition, 4 out of the 7 core Christensenellaceae ASVs in Donor_4 established in the HMA piglets. A single core Christensenellaceae ASV persisted in these HMA piglets (Supplementary Data 2).
In contrast to the notable differences in colonization efficiency of members of the phylum Firmicutes between the HMA mice and the piglets, ASVs belonging to Bacteroidetes successfully colonized both animal models (Supplementary Data 7). All 3 of the core Bacteroidetes ASVs identified in Donor_1 persistently colonized both animal models (Supplementary Data 2), although their abundances were much higher in the HMA mice (mean = 35.56%, SD = 14.42%) compared to the HMA piglets (mean = 9.74%, SD = 6.98%). Similarly, all 8 core Bacteroidetes ASVs of Donor_2 persistently colonized both HMA animals (Supplementary Data 2) and at higher relative abundances compared to the donor (Supplementary Data 7). Twenty-one of the 25 core Bacteroidetes ASVs of Donor_3 established in both animal models, with 18 (85.7%) and 16 (76.2%) of these ASVs persistently colonizing the HMA mice and HMA piglets, respectively. Higher relative abundances for these Bacteroidetes ASVs were observed in the HMA mice compared to both the donor and the HMA piglets (Supplementary Data 7). Finally, of the 25 core Bacteroidetes ASVs detected in Donor_4, 21 (84%) and 23 (92%) colonized the HMA mice and piglets, respectively, although their relative abundances were much lower compared to the donor (Supplementary Data 7). Fourteen of the 21 (67%) Bacteroidetes ASVs colonizing the mice were persistent colonizers, while 13 of the 23 (56%) Bacteroidetes ASVs persistently colonized the piglets (Supplementary Data 2).
Together, our results demonstrate that a greater number of ASVs belonging to the phylum Firmicutes, especially those of the families Lachnospiraceae, Ruminococcaceae, and Christensenellaceae, from the human donors established and persisted in the HMA piglets compared to the HMA mice. Our results also indicate that members of the phylum Bacteroidetes successfully colonized both animal models, although their relative abundances were higher in the HMA mice compared to the HMA piglets.
Bacteroides and Bifidobacterium members successfully colonize both animal models
The genera Bacteroides and Bifidobacterium are key members of the human gut microbiota and were identified as part of the core microbiotas of all four donors in this study (Supplementary Data 1). The Donor_1 core included a single Bifidobacterium ASV classified as Bifidobacterium bifidum (mean abundance = 4.5%, SD = 3.47%), which colonized both animal models at much lower relative abundances compared to the donor with means of 0.81% (SD = 0.71%) and 0.84% (SD = 1.17%) for the HMA mice and HMA piglets, respectively. This Bifidobacterium bifidum ASV also persistently colonized both HMA animal models (Supplementary Data 3). The Donor_2 core contained 3 Bifidobacterium ASVs, with all 3 ASVs colonizing the HMA piglets and 2 out of the 3 ASVs colonizing the HMA mice. The Donor_3 and Donor_4 cores contained 1 and 3 Bifidobacterium ASVs, respectively, all of which established in both HMA animal models. The single core Bifidobacterium ASV identified for Donor_3 did not persist in either animal model. Only 1 of the 3 core Bifidobacterium ASVs of Donor_4 persistently colonized the mice while all 3 ASVs persistently colonized the piglets (Supplementary Data 2).
All core Bacteroides ASVs found in the donors successfully colonized both animal models (Supplementary Data 3). Donor_1 core contained 3 Bacteroides ASVs at a mean abundance of 18.05% (SD = 0.34%). Two of these Bacteroides ASVs were assigned to the species B. vulgatus while the remaining ASV was classified as B. uniformis (Supplementary Data 1). These Bacteroides ASVs bloomed in the HMA mice with a mean abundance of 35.56% (SD = 14.42%). In contrast, these same ASVs reached a mean relative abundance of only 9.74% (SD = 6.98%) in the HMA piglets. All 3 of these Bacteroides ASVs persistently colonized both animal models (Supplementary Data 2). Donor_2 contained 6 core Bacteroides ASVs (mean abundance = 11.37%, SD = 1.42%), which established at higher abundances in both HMA animal models—26.86% (SD = 7.23%) and 22.80% (SD = 6.09%) for the HMA mice and piglets, respectively (Supplementary Data 7). In addition, all 6 of these ASVs persisted in both groups of animals (Supplementary Data 2). The 7 core Bacteroides ASVs in Donor_3 (mean abundance = 6.51%, SD = 0.67%) established in the HMA piglets at similar levels but with greater variation (mean abundance = 7.01%, SD = 8.67%). However, in the HMA mice, the abundances were much greater (mean abundance = 18.9%, SD = 6.34%) compared to both the donor and the HMA piglets. All 7 ASVs persistently colonized the HMA mice, while 6 out of the 7 persisted in the HMA piglets. The Donor_4 core contained 7 Bacteroides ASVs (mean abundance = 7.13%, SD = 0.62%) and these ASVs colonized the HMA mice at comparable relative abundances (mean = 9.52%, SD = 4.53%). The relative abundances for the same Bacteroides ASVs were much lower in the HMA piglets (mean = 3.01%, SD = 4.16%). Six and 4 of these 7 core Bacteroides ASVs persistently colonized the HMA mice and piglets, respectively (Supplementary Data 2).
Establishment of ASVs classified as Faecalibacterium prausnitzii, Roseburia and Ruminococcus bromii in the HMA animal models
Previous studies have demonstrated Faecalibacterium prausnitzii and Roseburia species to be functionally important members of the human gut microbiota and potential biomarkers of gut health31,32. We therefore evaluated the establishment and persistence of these ASVs in the two HMA animal models. Donor_2, Donor_3, and Donor_4 had core ASVs which were classified as Faecalibacterium prausnitzii (Supplementary Data 1). The Faecalibacterium prausnitzii (ASV_3233) of the Donor_2 core was detected only in the HMA piglets but at a lower relative abundance. Two ASVs classified as Faecalibacterium prausnitzii (ASV_3233 and ASV_3238) were part of the Donor_3 core. While ASV_3238 had colonized both HMA animal models, ASV_3233 only colonized the HMA piglets (Supplementary Data 1). Donor_4 contained three ASVs classified as Faecalibacterium prausnitzii (ASV_3233, ASV_3238, and ASV_3239). All three ASVs were detected in the HMA piglets, while none were detected in the HMA mice (Supplementary Data 1). Thus, all the core Faecalibacterium prausnitzii ASVs from the different donors colonized the HMA piglets, albeit at lower relative abundances. However, only ASV_3233 found in the Donor_3-HMA piglets was a persistent colonizer (Supplementary Data 3).
The butyrate-producing human gut commensal genus Roseburia was represented among the core ASVs of Donor_2, Donor_3, and Donor_4 inocula (Supplementary Data 1). All core Roseburia ASVs found in each donor (2, 3, and 2 ASVs for Donor_2, Donor_3, and Donor_4, respectively) colonized the corresponding HMA piglets. The relative abundances of these ASVs were comparable or higher to that of the donor, with the exception of ASV_6057 found in the Donor_2-HMA and Donor_4-HMA piglets (Supplementary Data 1). Furthermore, apart from the 2 core Roseburia ASVs of Donor_4, the remaining core Roseburia ASVs that established in the HMA piglets were persistent colonizers (Supplementary Data 2). In contrast, the only core Roseburia ASVs that established in the HMA mice were 2 core Roseburia ASVs found in Donor_3. These ASVs colonized at very low abundances (mean abundance = <0.01%, SD = 0.01%) in the HMA mice. Across donors, the main species of Roseburia identified in the core microbiotas were R. inulinivorans and R. intestinalis (Supplementary Data 1).
The prominent resistant starch degrading gut symbiont Ruminococcus bromii has also been proposed as a keystone species within the human gut microbiota33. R. bromii (ASV_3350) was identified in the core microbiotas of Donor_2 and Donor_4 (Supplementary Data 1). R. bromii had only engrafted successfully in the corresponding HMA piglets and at higher relative abundances than in the human donors (Supplementary Data 1). Moreover, this ASV had persistently colonized both Donor_2-HMA and Donor_4-HMA piglets (Supplementary Data 3).
The ability of persistent colonizers to maintain donor-like abundances in the HMA animal models over time
To determine if the persistent colonizers identified in the HMA animals were capable of maintaining abundances similar to the human donors over time, we compared the relative abundances of the persistent colonizer ASVs established in the two HMA animal models to their abundances in the respective human donors at each sampling time point. For persistent colonizer ASVs colonizing each HMA animal model, we determined the number of time points in which the relative abundance of each of the ASVs were significantly different from the corresponding human donor inocula. We considered a persistent colonizer ASV as having maintained a ‘donor-like’ relative abundance if that ASV’s relative abundance in a given animal model was not significantly different from the corresponding human donor in at least 4 out of the 7 sampling time points. Supplementary Data 8 and 9 provide the relative abundance and taxonomic information for all persistent colonizer ASVs with donor-like relative abundances with respect to each HMA animal model.
Based on the above criterion, we observed 7 out of 16 (44%) and 8 out of 15 (53%) of the persistent colonizer ASVs to maintain donor-like relative abundances for Donor_1 in the HMA mice and HMA piglets, respectively. Bifidobacterium bifidum (ASV_3778) was able to maintain a donor-like relative abundance in the HMA mice but not in the HMA piglets (Supplementary Data 9). For Donor_2, the HMA mice contained only 2 out of the 21 (9%) persistent colonizer ASVs with donor-like abundances. The HMA piglets contained 15 out of 43 (35%) persistent colonizer ASVs with donor-like relative abundances. For Donor_2, members of the genera Bacteroides, Blautia, and Roseburia were among the ASVs with donor-like abundances in the HMA piglets (Supplementary Data 9). For Donor_3, 40 out of the 69 (58%) persistent colonizers maintained donor-like abundances in the HMA piglets while only 6 out of 31 (19%) persistent colonizer ASVs identified in the HMA mice were able to do so. Five out of the 6 persistent colonizers with donor-like abundances identified for Donor_3 in HMA mice were Bacteroidetes members. Bacteroidetes and Firmicutes were widely represented among the persistent colonizers with donor-like abundances for Donor_3 in HMA piglets (Supplementary Data 9). For Donor_4, 7 out of 28 (25%) and 16 out of 38 (42%) persistent colonizers were maintained at donor-like relative abundances in the HMA mice and HMA piglets, respectively.
Altogether, these findings suggest that only a small proportion of persistent colonizers are able to consistently maintain relative abundances similar to those of the respective donors within the two HMA animal models. The HMA piglets, however, had higher proportions of persistent colonizers with donor-like relative abundances compared to the HMA mice.
Phylogenetic analysis of Firmicutes colonizing the two HMA animal models
As noted previously, ASVs of the phylum Firmicutes were the most conspicuous in terms of their differential establishment between the two HMA animal models. To further identify phylogenetic lineages that may affect colonization of core ASVs of the phylum Firmicutes among the two animal models, we analyzed the phylogenetic relationships among the core Firmicutes ASVs identified in the 4 donors (Fig. 6).
Red squares indicate the presence of core Firmicutes ASVs in the human donors while the green squares and blue squares indicate which of those Firmicutes ASVs were able to colonize the HMA mice and piglets, respectively. Green stars indicate persistent colonizers of HMA mice and blue stars indicate persistent colonizers of HMA piglets. n = 13 (3 piglets/donor, with the exception n = 4 for Donor_1, 2 days post inoculation) and n = 37 mice (Donor_1 = 7, Donor_2 = 10, Donor_3 = 10, Donor_4 = 10).
In total, we identified 178 core ASVs across the 4 donors which belonged to the phylum Firmicutes. Out of these 178 ASVs, 131 (74%) were identified as colonizers of the HMA piglets while only 69 (39%) colonized the HMA mice. Twenty-four of the 69 (39%) Firmicutes ASVs colonizing the mice were persistent colonizers while 74 out of the 131 (56%) were persistent colonizers of the piglets. As shown in Fig. 6, we identified 3 clades as preferentially colonizing the HMA pigs. This included a clade consisting of members of the family Christensenellaceae and two other clades consisting of members of the order Oscillospirales. We also observed five closely related Lachnospiraceae ASVs which persistently colonized both animal models. In addition, we identified a clade consisting of Clostridia UCG-014 that failed to colonize either animal model, with the exception of a single ASV which colonized the HMA piglets (Fig. 6).
Core microbiota establishment in HMA animals using a different set of donors
As described in Table 1, markedly low percentages of colonizers, as well as persistent colonizers were identified for HMA mice compared to HMA piglets for Donor_2, Donor_3, and Donor_4. In order to further verify these observations and investigate the potential donor-to-donor variation in colonization efficiency, we performed a second study where we inoculated mice with fecal inocula from a different set of donors (Donor_5, Donor_6, and Donor_7; see “Methods” for donor information) and provided the same solid diet. Core ASVs (i. e., present in all inoculum aliquots), “colonizers” and “persistent colonizers” were identified using the same approach. These donors were also selected to represent the variation in the human microbiome (Supplementary Fig. 5) and belonged to similar age groups as our first set of donors. In addition, since Donor_1, an infant donor, was the only donor from our first set of donors which had a bacterial community that established more successfully in HMA mice compared to HMA piglets, we investigated the repeatability of this observation by performing another inter-species transplantation experiment using fecal matter from a different infant donor (Donor_8) to GF mice (C3H/HeN) and GF piglets (Sus scrofa domesticus), similar to our first study.
The results of this subsequent study (Table 2) demonstrated that the Donor_7-HMA mice had comparable percentages of colonizers and persistent colonizers relative to Donor_4 HMA mice (Both Donor_4 and Donor_7 were seniors; see “Methods” for details). A similar comparison of core ASV establishment among the child donors (Donor_2 and Donor_5) revealed that these two groups of mice had very similar levels of persistent colonizers while Donor_5-HMA mice had a greater percentage of colonizers (Tables 1, 2). Donor_6-HMA mice had notably higher percentages of colonizers, as well as persistent colonizers, compared to the similarly-aged donor (Donor_3) of the first study (Table 2). These results demonstrate the possibility of having considerable donor-to-donor variation in colonization efficiency when transplanting human fecal bacterial communities into GF mice. The taxonomic distribution of the colonizing core ASVs in the HMA mice (Supplementary Data 10) was consistent with our previous observations – Bacteroidetes members colonized with high efficiency while Firmicutes ASVs established poorly (Supplementary Data 10). In agreement with our previous results, ASVs belonging to the phyla Lachnospiraceae and Ruminococcaceae displayed low colonization success, with the exception of the mice inoculated with Donor_6, which were efficiently colonized with Lachnospiraceae members (Supplementary Data 10). Additionally, the mice inoculated with Donor_6 and Donor_7 were poorly colonized by members of the family Oscillospiraceae. In contrast to the observation with the previous set of donors, some Christensenellaceae ASVs from Donor_6 (1 out of 1) and Donor_7 (1 out of 6) colonized the GF mice of this second study.
Similar to the results for our first infant donor (Donor_1), the core bacterial community of the donor inocula of our second infant donor (Donor_8) was more similar to the donor in the HMA mice than HMA piglets (Supplementary Fig. 6). Once again, the major taxonomic groups were established at much lower relative abundances in the Donor_8 HMA piglets compared to both the donor and the HMA mice (Supplementary Data 10).
Comparison of predicted function between the human donors and the HMA animals
To determine the potential functional consequences of the differential establishment of fecal bacterial communities between the two HMA animal models, we used PICRUSt34 to predict the functional capacity of the human donors and the corresponding HMA animal models. Using the PICRUSt-predicted functional annotations, we identified the KEGG Orthology (KO) features that were detected in all the inoculum aliquots of a given donor and defined them as ‘core predicted KO features’ for each donor microbiome. Subsequently, this information was used to compare potential functions of the microbiomes established in the HMA models against the donors. The results, represented as PCoA plots (Supplementary Fig. 7), showed a similar clustering pattern to what we observed with the core ASVs (Fig. 2) where the Donor_1 inocula clustered closer to the HMA mouse samples while for the remaining donors, the donor inocula clustered closer to the HMA piglet samples. Thus, these results indicate that the differences in establishment of human donor microbiotas between the two animal models also results in potential differences in their functional capacities. As such, function may have favored the established taxa. Therefore, our belief is if similar taxa to the donor are established, similar function would also be present in the corresponding animal model.
Source: Ecology - nature.com
