Exploring the role of intestinal pathogenic bacteria in metronidazole-induced bone loss: focus on Klebsiella variicola

Three-month-old male mice were orally administered drinking water supplemented with MET and age-matched control mice were treated with plain drinking water for 8 weeks. The micro-computed tomography (µCT) analysis showed that MET treatment caused remarkable bone loss and impaired bone microstructures in the femur of mice as indicated by significantly decreased bone mineral density (BMD), trabecular bone volume fraction (Tb. BV/TV), trabecular thickness (Tb. Th), trabecular number (Tb. N), cortical thickness (Ct. Th), endocortical perimeter (Es. Pm), periosteal perimeter (Ps. Pm), and increased trabecular separation (Tb. Sp) (Fig. 1A-I). The three-point bending test indicated that MET treatment significantly reduced the maximum bending load of the tibia (Fig. 1J). Meanwhile, the Hematoxylin and eosin (H&E) staining showed that MET treatment increased inflammatory cell density in mouse bone marrow (Supplementary Figure S1A). Furthermore, the Tartrate-acid-resistant phosphatase (TRAP) staining of mouse femurs and quantitation of the number of osteoclasts (N. OCs) showed that Met treatment induced a significant increase in the number of osteoclasts (K-L). Moreover, Immunohistochemical (IHC) staining showed a significant increase in the released pro-inflammatory mediators including IL-1β and IL-6 in the bone marrow of MET-treated mice (Fig. 1M-P). Thus, oral administration of MET upregulated pro-inflammatory factors and promoted osteoclast differentiation in the bone marrow, ultimately leading to osteoporotic-like changes in mice.

We harvested and analyzed fecal samples from the MET-treated osteoporotic mice and vehicle-treated mice in order to screen for the key pathogenic bacteria contributing to MET-induced bone loss. The 16 S rRNA gene sequencing of fecal samples showed that Enterobacteriaceae increased most significantly at the family level after MET treatment (Fig. 2A-B). Then, the quantitative real-time polymerase chain reaction (qRT-PCR) was used to compare the abundance of the family of Enterobacteriaceae, and significant differences were found in the genus of Klebsiella (Fig. 2C-D). For there are many different species in the genus of Klebsiella, we conducted additional inoculations of fecal samples from MET-treated mice onto both blood agar plates and MacConkey agar plates, which aimed to discern and identify the predominant bacterial species present. The dominate bacteria with the highest number of colonies in these plates were coincident (Fig. 2E) and were identified as Klebsiella variicola (K. variicola) by a matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) with a high credibility score (Fig. 2F). Notably, increased inflammatory responses were not observed in the colonic tissues of mice treated with MET, suggesting that MET stimulation on bone marrow inflammatory responses may act through specific inflammatory mediators (Fig S2).

The abundance of K. variicola was significantly increased in MET-induced osteoporotic mice, but the relationship between K. variicola and skeletal diseases has not been reported. To examine the potential of K. variicola to mediate MET-induced osteoporosis, mice were transplanted with K. variicola by oral gavage once a week for 8 weeks. The µCT results showed that K. variicola transplantation led to a significant decrease of BMD, Tb. BV/TV, Tb. Th, Ct. Th and Tb. N, and increased Tb. Sp, but Es. Pm and Ps. Pm did not have statistically significant difference (Fig. 3A-I). The three-point bending test indicated that K. variicola transplantation reduced the maximum bending load of the tibia (Fig. 3J). Consistent with MET-treated mice, compared to Vehicle-treated mice, the number of osteoclasts and expression of pro-inflammatory factors were significantly increased in the bone marrow of K. variicola treated mice according to TRAP and IHC staining results. (Fig. 3K-P). These results suggested that intestinal K. variicola could independently induce inflammatory bone loss in vivo.

As a part of the communication process between organisms, the gut microbiota produces small bodies, called microbial EVs. EVs act as shuttles for transferring bioactive cargoes (i.e., proteins, mRNAs, miRNAs, DNA, carbohydrates, and lipids) to exert efficient regulatory effects on various organs of the host [16]. Therefore, we next isolated EVs from K. variicola (K. var-EVs), and investigated their direct effects on inflammatory activation and osteoclastogenesis in macrophage lineages in vitro (Fig. 4A-B). As evidenced by the TRAP staining, K. var-EVs dramatically promoted the osteoclastic differentiation in bone macrophages (Fig. 4C-D). The qRT-PCR results showed that K. var-EVs upregulated the expression of IL-1β and IL-6 in a dose-dependent manner (Fig. 4E-F). Therefore, K. var-EVs exhibited the capacity to induce the inflammatory response in bone cells and to promote osteoclast differentiation.

Given the findings in vitro that K. var-EVs promoted inflammatory responses and osteoclastic differentiation, we further evaluated the effects of K. var-EVs on bone mass in mice. Our previous studies have demonstrated that bacterial extracellular vesicles can be translocated to bone tissue and mediate GM regulation of bone metabolism [17]. Similarly, we administered gavage of K. var-EVs to wild-type mice for 8 weeks. Surprisingly, significant bone loss, impaired bone microstructures and decreased bone strength were observed in K. var-EVs-treated mice according to the micro-CT, BMD and biomechanical test results (Fig. 5A-J). The TRAP staining results showed that K. var-EVs increased cell numbers of osteoclasts and the IHC staining results showed that K. var-EVs caused increased expression of inflammatory cytokines comparable to K. variicola transplantation in mice (Fig. 5K-P).

CIP is a long-used antibiotic that is effective in treating a variety of gram-negative bacterial infections. CIP has been reported to be more effective in treating Klebsiella than other common antibiotics [17]. To confirm the key role of K. variicola in contributing to MET-induced osteoporosis, we further intervened with CIP in MET-treated mice. Antibiotic susceptibility testing confirmed the susceptibility of K. variicola to CIP (Supplementary Figure S3A). The 16s rRNA sequencing of mouse feces showed that CIP intervention significantly suppressed the enhanced abundance of Enterobacteriaceae including K. variicola caused by MET (Supplementary Figure S3B). Surprisingly, CIP significantly attenuated bone loss caused by MET as measured by µCT BMD and biomechanical tests (Fig. 6A-J). Moreover, CIP intervention inhibited osteoclast numbers (Fig. 6K-L) and downregulated the expression of pro-inflammatory cytokines in bone marrow in MET-treated mice (Fig. 6M-P). These results further demonstrate the critical role of K.V bacteria in MET-induced bone loss (Fig. 7).

The animal care and experimental procedures were conducted in accordance with the guidelines and regulations of the Ethical Review Board at Xiangya Hospital of Central South University. Male C57BL/6J mice, aged two or three months and weighing between 20 and 24 g, were employed in this study. Mice were randomly assigned to various treatment groups. Exclusion criteria for mice included a body weight of < 18 or > 26 g or poor physical condition prior to treatment initiation. All mice were housed in specific pathogen-free conditions.

Metronidazole (Solarbio) was administered in the drinking water at a concentration of 1 g/L for a duration of 8 weeks, with mice receiving the vehicle under the same treatment regimen serving as the healthy controls. Fecal samples from MET-treated mice were collected at the 4th week of treatment for 16 S rRNA gene sequencing. To assess the impact of K. variicola and K. var-EVs, mice were orally gavaged with K. variicola (10⁶ CFUs in 200 µl of PBS) and K. var-EVs (100 µg in 200 µl of PBS), or an equal volume of PBS, once a week. For therapeutic intervention, ciprofloxacin (Solarbio) was added to the drinking water at a concentration of 0.25 g/L for MET-treated mice. The femurs of ten mice from each group were collected and processed for downstream analyses after 8 weeks of treatment.

After being fixed in 4% paraformaldehyde (PFA) for 2 days, the right femoral samples were analyzed with high-resolution µCT (Skyscan 1176) as described previously [33]. The scanner parameters were set to a voltage of 50 kV, a current of 400 µA, and a resolution of 11.4 μm per pixel, respectively. The analysis of femoral parameters, including Tb. BV/TV, Tb. Th, Tb. N, Tb. Sp, Ct.Th, Es.Pm, Ps.Pm, was conducted using image reconstruction software (NRecon), data analysis software (CTAn v1.11), and three-dimensional model visualization software (µCTVol v2.2).

Following the collection of fecal samples from mice subjected to vehicle, MET, or CIP + MET treatments, 16 S rRNA sequencing was conducted by GeneSky Biological Technology (Shanghai, China), including DNA extraction, PCR amplification, purification, library preparation, sequencing, bioinformatics analysis and statistical analysis.

For histological and immunohistochemical analyses, femurs were embedded in 4% paraformaldehyde (PFA) for 2 days, followed by decalcification in 0.5 M EDTA for 1 week. Subsequently, the specimens underwent dehydration using graded ethanol and immersion in xylene. The specimens were then embedded in paraffin, cut into 5-µm-thick sections, and subjected to Hematoxylin and Eosin (H&E) staining using reagents provided by Servicebio. Immunohistochemical staining for osteogenic marker OCN was performed as described previously [34]. IL-1β and IL-6 antibodies, as well as the secondary antibody (GB23303), were obtained from Servicebio. The sections were photographed with an Olympus CX31 microscope (Tokyo, Japan). The numbers of the IL-1β-, IL-6-, or OCN-positive cells were measured with the Image-Pro Plus 6 software.

Three-point bending tests were conducted on a mechanical testing machine (In-stron 3343, Canton, USA) to assess bone strength. Briefly, tibiae were positioned in the anterior–posterior direction (patella side facing up) on the lower supporting bars, spaced 8 mm apart. A constant vertical compression load (5 mm per minute) was applied to the midpoint of the samples until fracture occurred. The maximum bending load of the femur (N) was calculated from the load-displacement curve.

For isolation of K. variicola-EVs, K. variicola were incubated in fresh LB broth for 3 days and then pelleted by sequential centrifugation at 2000 g for 30 min and 10,000 g for 30 min at 4 °C. The culture supernatant was filtered by a 0.22-µm filter (Millipore, Billerica, USA) and then concentrated 100-fold by centrifugation at 4000 g and 4 °C in Amicon Ultra-15 Centrifugal Filter Units (100 kDa; Millipore). K. variicola-EVs were purified from the concentrated supernatant using OptiPrep density gradient centrifugation. Briefly, K. variicola-EVs (1.33 ml per tube) were added at the bottom of OptiPrep solution [6.67 ml per tube; 60% (w/v) iodixanol; Sigma-Aldrich, St. Louis, USA] in a polyallomer Beckman Coulter tube (38.5 ml), thus producing a 50% OptiPrep layer. Solutions of 40% (8 ml), 20% (8 ml), and 10% (7 ml) OptiPrep and 1 ml of PBS were sequentially and carefully added on the 50% OptiPrep layer to generate a discontinuous gradient. After centrifugation for 18 h at 100,000 g with a SW 32 Ti rotor (k factor, 204), 2 ml each of the OptiPrep density gradient fractions was obtained for nanoparticle tracking analysis to determine the EV-rich fractions (fractions 12 and 13). The EV-rich fractions were diluted with PBS to 30 ml and subjected to centrifugation for 3 h at 100,000 g and 4 °C to pellet K. variicola-EVs. The obtained K. variicola-EVs were resuspended in PBS. A small volume of K. variicola-EVs was subjected to bacterial colony counting assay on LB agar plate to ensure that there is no bacterial contamination. Protein contents of K. variicola-EVs were tested using a Pierce BCA protein assay kit (Thermo Fisher Scientific). K. variicola-EVs were used immediately or stored at − 80 °C until downstream experiments.

RAW264.7 cells were seeded in a 48-well plate at a density of 1.5 × 10⁴ cells per well and treated with 100 ng/ml RANKL, followed by stimulation with K. variicola-EVs or an equivalent volume of PBS. The negative control culture was cultivated in high-glucose DMEM supplemented with 10% FBS. Half of the medium was replaced every 3 days. After 6 days of induction, the cells were washed with PBS and fixed with 4% paraformaldehyde for 10 min. Following a wash with distilled water, the cells were stained for tartrate-resistant acid phosphatase (TRAP) using a commercially available kit (Sigma). TRAP + multinucleated cells (MNCs) with more than three nuclei were identified as osteoclasts. The number of osteoclasts was enumerated using an inverted microscope (Leica).

For fecal samples, total genomic DNA was extracted using the TIANamp stool DNA kit (Tiangen, Beijing, China). In the case of colon and cell samples, total RNA extraction was carried out using TRIzol reagent (Invitrogen, Carlsbad, CA). Following the assessment of total RNA purity and concentration, reverse transcription to cDNA was performed using the PrimeScript RT kit (Takara). Quantitative real-time polymerase chain reaction (qRT-PCR) was conducted using the TB Green Premix Ex Taq II (Takara). The 2^−ΔΔCT method was employed to calculate relative mRNA expression levels. Detailed primer sequences utilized in this study are provided in Supplementary Table 1.

Data are shown as means ± SD. Student’s t test (unpaired, two-tailed) was used for analyzing the differences between two groups. Multiple-group comparisons were performed using one- or two-way analysis of variance (ANOVA) followed by Bonferroni post hoc test. P < 0.05 was considered statistically significant, with *P < 0.05, **P < 0.01, and ***P < 0.001.