Fecal microbiota transplantation mitigates bone loss by improving gut microbiome composition and gut barrier function in aged rats

Experiment animals

The experimental protocol of animal trials was approved by the Ethical Committee of the Ninth People’s Hospital of Shanghai Jiao Tong University School of Medicine (SH9H-2019-A17-1). Thirty-two 18-month-old and four 3-month-old female Sprague Dawley (SD) rats were purchased from SLAC Laboratory Animal Company (SLAC., China, SCXK2012-0002) and raised in the specific pathogen-free Laboratory animal room in the Ninth People’s Hospital. The 18-month-old female SD rats were randomly divided into two groups: the FMT group (FMT, n = 16) and the control group (Con, n = 16). Rats in each group were housed as 2/cage with a 12-h light-dark cycle and free access to water and pelleted rodent diet. As the donors of fecal samples, 3-month-old female SD rats were raised in the same condition. FMT was performed by oral gavage, and rats were sacrificed by rapid cervical dislocation after 12 and 24 weeks following transplantation. The serum, intestine, bone, and feces were collected for subsequent analyses.

Fecal microbiota transplantation

FMT was performed based on an established method (Chang et al., 2015; Di Luccia et al., 2015). Briefly, fresh fecal samples from four 3-month-old donor rats were collected under aerobic and sterile conditions by pressing the end of rectum before sacrificing (Costello et al., 2019). Every 100 mg mixed fecal samples were re-suspended in one mL of sterile saline. The suspension was vigorously vortexed for 10 s before centrifugation at 800G for 3 min. The supernatant was collected and frozen at −80 °C. Diluted supernatant was orally given to recipient rats (one mL/rat) three times per week. Normal saline was given to the control group following the same method.

Bone mineral density detection

To verify the rat model of senile osteoporosis, bone mineral density (BMD) was measured by dual-energy X-ray absorptiometry (DXA, Hologic Discovery A, USA).

Micro-CT analysis

Micro-CT scanning was performed using a high-resolution micro-CT 81 system (Scanco Medical AG, Bruettiselien, Switzerland) (Fu et al., 2020). Before scanning, rats were sacrificed, and tibia samples were collected and fixed in 4% paraformaldehyde. Histomorphometric parameters including bone volume (BV), trabecular bone volume fraction (BV/TV), trabecular number (Tb.N), trabecular thickness (Tb.Th), trabecular spacing (Tb.Sp), and bone surface area ratio (BS/BV) were calculated. Three-dimensional reconstruction of the tibia metaphysis was performed using the built-in Scanco software.

ELISA analysis

The serum was homogenized and used to measure the concentration of osteocalcin (OC), P1NP and CTX using ELISA kits (XLCC, Shanghai YBL Biomedical Technologies, China).

16S rRNA gene sequencing and bioinformatic analysis

Fecal samples were collected from both groups and immediately stored at −80 °C for subsequent 16S rRNA gene sequencing (Ma et al., 2020a, 2020b). For the analysis of the phylogenetic GM composition, the V3–V4 region was amplified using the Illumina MiSeq instrument.

The data of 16S rRNA sequencing were analyzed as previously described (Ma et al., 2020a, 2020b). We grouped the high-quality DNA sequences into operational taxonomic units (OTUs) and compared them with the SILVA reference database (V138) at 97% similarity. The minimum sample size was deemed to be the criteria of data normalization. We performed community richness and diversity analyses (Shannon and Simpson indexes) using Mothur software (V1.44.1). To clarify the similarity or difference of data, we conducted unweighted and weighted UniFrac Principal Co-ordinates Analysis (PCoA) by calculating the ecological distance to obtain the eigenvalues and eigenvectors between samples. ThetaYC algorithm was used to calculate the similarity of the community structure of each sample at the level of OTU = 0.03. Taxonomy was assigned using the online software RDP classifier (80% threshold) based on the Ribosomal Database Project. We performed Redundancy analysis (RDA) using the vegan package in R. We used the KEGG database to predict functional pathways and define significant differences of KEGG ortholog abundances between two groups as P < 0.05, fold change >2 or <−2.

Hematoxylin & eosin staining

The distal ileum was fixed in 4% (v/v) paraformaldehyde and embedded in paraffin, and then cut into four µm sections in thickness. The sections were then stained with HE. Morphological characteristics were examined under the Olympus light microscope (Japan).

Immunohistological staining

Ileum paraffin sections were routinely dewaxed in xylene and dehydrated in gradient alcohol. After blocking with 0.3% triton X-100, in Tris-HCl buffer (0.1 M) containing 5% fetal bovine serum, the slides were incubated with the primary antibody including anti-ZO-1 (1:100, NOVUS), anti-claudin (1:100, SAB signalway antibody) and anti-occludin (1:100, Proteinbtech) at 4 °C over-night. Secondary antibodies used in this study included Cy3-or Alexa Fluor 488-conjugated donkey anti-mouse or rabbit IgG (1:400 dilution; Jackson ImmunoResearch Labs, Inc.). Nuclei were counterstained with 40, 6-diamidino-2-phenylindole (DAPI) (1:30,000 Roche Biochemicals, Monza, IT) and mounted with Vectashield (Vectorlabs, Burlingame, CA). Images were digitally captured by FV1,000 confocal laser-scanning microscope (Olympus, Japan) and processed with FV10-ASW Viewer 1.7 and Adobe Photoshop CS3. The quantitative analysis was performed by ImageJ software (NIH, Bethesda, Maryland, USA).

Statistical analysis

All data were processed by GraphPad Prism 6.0 statistical software. The data were analyzed by Student t-test or multiple t-tests, and corrected by Holm-Sidak method, with α = 0.05. All the measurement data are expressed as mean ± standard error (SE). The statistical significance was as follows: *P < 0.05; ^#P < 0.01.

Successful establishment of the animal model of senile osteoporosis

The experiment flow of the current study was described in the study design (Fig. 1A). DXA was used to measure the BMD in order to verify the animal model of senile osteoporosis. The BMD of the FMT group (0.1503 ± 0.0018 g/cm²) and the control group (0.1494 ± 0.0021 g/cm²) was significantly lower than that of the 3-month-old rats (0.1,647 ± 0.0013 g/cm², P < 0.01) (Fig. 1B), which demonstrated that senile osteoporosis was successfully established in the 18-month-old SD rats used in this study.

GM from young rats reduced bone turnover in aged rats of senile osteoporosis

The bone turnover markers of OC, P1NP, and CTX increased significantly in the control group at the 12th and 24th weeks (Figs. 1C–1E), while these parameters did not change significantly in the FMT group. The OC of the FMT group decreased significantly at the 12th and 24th weeks, compared with the control group (Con12: 127.99 ± 0.81; FMT12: 125.68 ± 0.81; Con24: 129.49 ± 0.99 ng/l; FMT24: 125.43 ± 0.99 ng/l; P < 0.05) (Fig. 1C). Serum P1NP and CTX of the FMT group followed the same trends as OC at the 12th and 24th weeks post-transplantation (P < 0.05) (Figs. 1D, 1E). We concluded from the data that FMT could reverse the high bone turnover status of senile osteoporosis.

GM from young rats alleviated bone loss in aged rats

Three-dimensional images showed that the bone structure of the control group was gradually deteriorated during oral gavage, while little changes of the bone structure were found in the FMT group (Fig. 2A). Quantitative analyses demonstrated that BV, BV/TV, Tb.N, Tb.Th in the control group decreased gradually at the 12th and 24th weeks (Fig. 2B). At the 12th week, these parameters in the FMT group were comparable to the control group. However, at the 24th week post-transplantation, the BV (Con24: 1.20 ± 0.37; FMT24: 2.03 ± 0.37), BV/TV (Con24: 9.05 ± 2.73; FMT24: 16.93 ± 2.73), Tb.N, and Tb.Th of the FMT group were significantly higher than the control group (P < 0.05). These findings together with the serum bone turnover markers suggested that FMT could reverse osteoporosis in aged rats.

GM from young rats improved the gut microbiota dysbiosis in aged rats of senile osteoporosis

Unweighted and weighted UniFrac PCoA was conducted to analyze the beta diversity of GM. The sample distribution of the control group was scattered, which was consistent with our previous findings in senile osteoporosis (Ma et al., 2020b) (Figs. 3A, 3B). However, the GM aggregated at 12 and 24 weeks following FMT, and the ecological distance was similar between the rats in the FMT group and the donor rats (Fig. 3A). These results demonstrated that the diversity of GM changed significantly after FMT, and the intestinal microbiota of the aged rats appeared to be similar to the donor rats.

The total number of OTUs between the two groups was significantly different at the 12th and 24th weeks post-transplantation (Fig. 3C). Shannon and Simpson indexes are commonly used to evaluate the alpha diversity of GM. A low Shannon index and high Simpson index mean low alpha diversity. Therefore, these two indexes showed that the alpha diversity was significantly decreased at 24 weeks after FMT (P < 0.05) (Figs. 3D, 3E). Firmicutes/Bacteroidetes (F/B) ratio is an indicator of GM disorder. High F/B ratio was associated with disease status (Everard & Cani, 2013). The F/B ratio decreased significantly at 24 weeks after FMT, suggesting that the status of GM disorder was alleviated (P < 0.05) (Fig. 3F).

GM from young rats improved GM composition in aged rats at the phylum and family levels

Next, we compared the phylum-level GM contents of the two groups after transplantation (Fig. 4A). The abundance of Firmicutes of the control group was maintained at nearly 70% at the 12th and 24th weeks. However, it decreased from 68.80% at the 12th week to 55.78% at the 24th week in the FMT group, which is quite close to that of the donor rats (54.58%). The abundance of Bacteroidetes was another major phylum of gut microbiota. The abundance of Bacteroidetes in the FMT group at the 24th week (37.92%) was higher than that of the control group (23.25%), which was also similar to the donor rats (38.31%) (Fig. 4A). Furthermore, the abundance of Lachnospiraceae decreased gradually from the 12th week to the 24th week following FMT, which was also similar to that of the donor rats (Fig. 4B). These data suggested that FMT could restore GM composition in rats with osteoporosis at the phylum and family levels.

GM from young rats changed GM composition in aged rats at the genus level

To further investigate the mechanisms of the beneficial effect of FMT on osteoporosis, we constructed a histogram of the main GM abundances of the two groups at different time points (Fig. 5A). Compared with the control group, the contents of unclassified_Lachnospiraceae, Oscillibacter, Lachnospiraceae_incertae_sedis decreased significantly at the 24th week (Figs. 5B, 5D, 5E) (P < 0.05), which was similar to the composition of the donor rats. Blautia increased significantly after FMT (Fig. 5C) (P < 0.05). Meanwhile, the content of Helicobacter and Prevotella decreased significantly following FMT (Figs. 5F, 5G).

RDA analysis of GM and bone parameters following FMT

Redundancy analysis was performed to demonstrate the correlation between GM and bone parameters at 12 weeks and 24 weeks after colonization (Fig. 6). Prevotella, Bacteroides, Clostridium sensu stricto, and Lactobacillus were positively correlated with Tb.Th, BV, BV/TV, and Tb.N. However, Helicobacter, Oscillibacter, and Lachnospiracea_incertae_sedis were negatively correlated with these parameters. Moreover, GM at the genus level in the control group and the FMT group were partially overlapped at the 12th week, while completely separated at the 24th week.

GM from young rats modulated the function of GM in aged rats

To investigate the function of GM following FMT, we predicted pathways based on the KEGG database. Significantly differentially expressed genes were found (Fig. 7) (P < 0.05, fold change >1.2 or <−1.2). Four pathways with fold change >2 after FMT were identified. Steroid hormone biosynthesis, 1, 1, 1-Trichloro-2, 2-bis(4-chlorophenyl) ethane (DDT) degradation, stilbenoid, diarylheptanoid and gingerol biosynthesis, and basal transcription factors were up-regulated in the FMT group.

Correlation network analysis between the FMT group and the control group

To evaluate the correlation between FMT and GM community structure in the recipients, we performed a correlation network analysis of 50 OTUs with the highest abundance to determine the species with key changes. Few correlations and little transitivity between the two groups at the 12th week (Fig. 8A). With the same standard coefficient, a more complex correlation network was constructed 24 weeks after transplantation. A significant correlation and better clustered OTUs were found. OTU06 and OTU18 were selected as the key gene information at the 24th week based on topological properties, which could be annotated to Proteobacteria and Firmicutes, respectively (Fig. 8B). Changes of Firmicutes contributed the most to GM alterations. Proteobacteria had a positive correlation with Firmicutes, while Bacteroidetes had a negative correlation with Firmicutes.

GM from young rats improved the intestinal structure and barrier function in aged rats

Intestinal structure and barrier function were proved to be associated with osteoporosis (Schepper et al., 2020). We have manifested that FMT was beneficial for osteoporosis by restoring GM homeostasis. Next, we investigated the intestinal structure and barrier function following FMT. HE staining showed that the FMT group had less swelling mucosa, sub-epithelial space expansion, and well-arranged villi structure at the 24th week (Fig. 9A). Immunohistological staining showed the upregulation of occludin, claudin, and ZO-1 in the FMT group at the 24th week (Fig. 9B). Furthermore, the quantitative results verified that the expression of occludin (P < 0.01), claudin (P < 0.05), and ZO-1 (P < 0.01) significantly increased at 24 weeks following FMT (Figs. 9C–9E). These results suggested that FMT improved the intestinal structure and barrier function in aged rats.