Cannabidiol-Treated Ovariectomized Mice Show Improved Glucose, Energy, and Bone Metabolism With a Bloom in Lactobacillus

2.2 Metabolic Tests

Body weight and food intake was measured weekly. Body composition was measured using an EchoMRI body composition analyzer (EMR-129, Echo Medical Systems, Houston, TX) at 0, 4, 7, 10, 14, and 18 weeks post-treatment. After 9 weeks of treatment, an oral glucose tolerance test was performed as previously described (Tveter et al., 2020; Mezhibovsky et al., 2021). During the 9th and 10th week post-treatment, indirect calorimetry was performed on 9 mice per group using a Comprehensive Laboratory Animal Monitoring System with temperature and light control enclosure (CLAMS, Columbus Instruments, Columbus, Ohio) as previously described (Mezhibovsky et al., 2021).

2.3 Bone Densitometry

Total body areal bone mineral density (aBMD) and bone mineral content (BMC) was measured using dual-energy X-ray absorptiometry (DEXA, GE-Lunar PIXImus mouse densitometer software version 2.10.41) as previously described (Wang et al., 2016). Frozen carcasses were used for whole body scanning. Immediately after each scan (within 2–3 min), the frozen headless carcass was placed on dry ice. Individual right-side femur, tibia, humerus, and lumbar spine (L1-L5) were dissected to remove surrounding tissues and scanned by DEXA. After scanning the individual right-side bones were wrapped with phosphate buffered saline (PBS) soaked gauze and stored at −20°C. The left femur was dissected, placed in a microfuge tube, and stored frozen at −80°C for RNA extraction.

2.4 Bone Microcomputed Tomography (microCT)

Right femur samples (stored at −20°C) were placed on dry ice and transported to Rutgers Imaging Center for microCT analysis. Tubes containing femurs were placed into room temperature water to thaw, wrapped securely with parafilm to prevent drying, and inserted into a 5 mm polystyrene tube to be scanned using a microcomputed tomography system (SkyScan 1,272; Bruker Corporation, Kontich, Belgium). Settings were 80 kV, 124 μA, image voxel size 6.0 µm isotropically, and camera image settings 1,092 × 1,632 pixels with 0.1 mm Al filter, averaging 3 frames per rotation step of 0.4°. The NRecon was used to reconstruct the scanned images of femoral microstructure and analyzed with SkyScan CT Analyzer (Bruker microCT version 1.18.1.0). Geometric trabecular volumetric BMD (vBMD) and cortical tissue mineral density (TMD) were analyzed and calculated as mg/mL against calibrated phantoms of 250 and 750 mg hydroxyapatite/cc as the standards. Femoral 3D images were generated using CTvol (Bruker microCT version 3.3.0) (Bouxsein et al., 2010).

2.5 Quantitative Polymerase Chain Reaction

RNA from ileum and colon tissues (30–35 mg) were extracted as previously described (Tveter et al., 2020; Mezhibovsky et al., 2021). Each femur bone sample (stored at −80°C) was ground to powder with a mortar and pestle (pre-chilled in −20°C freezer) containing liquid nitrogen. Each powdered femur sample was transferred to a microfuge tube and 900 µL of Qiazol was added for total RNA extraction according to kit instructions (RNeasy Plus universal kit, QIAGEN). RNA concentrations were quantified by Nanodrop (Thermo Fisher Scientific Inc.). cDNA was synthesized from 5 µg RNA and RT-qPCR was performed (with technical duplicates) on a thermocycler (Quantstudio 3, Thermo Fisher Scientific Inc.) using Taqman™ Fast universal PCR master mix using conditions: 20 s at 95°C followed by 40 cycles of 95°C for 1 s (denaturation) and 60°C for 20 s (annealing and extension). TaqMan™ assay primers (Life Technologies) used are summarized in Supplementary Table S2. Hmbs and Gapdh were used as housekeeping genes for intestinal and femoral tissues, respectively.

2.6 Targeted Metabolomics

Previously described (Tveter et al., 2020) methods with modifications described in supplementary methods were used for analysis of serum and ileal content bile acids (BAs). Cecal short chain fatty acids (SCFA) were analyzed by GC-MS using previously described methods (García-Villalba et al., 2012).

2.7 Gut Microbiota Analysis

gDNA was extracted from week 18 fecal samples (n = 28, 7 per group) using DNeasy PowerSoil Pro Kit (QIAGEN). Library preparation and paired-end sequencing of 16S rRNA V3-V4 amplicons (2 × 250 bp configuration) was performed by Azenta LifeSciences (Piscataway, NJ) using an Illumina^® MiSeq instrument.

To confirm the increase in Lactobacillus sp., qPCR was performed using primers (Lac_groEL_F: 5′-GCYGGTGCWAACCCNGTTGG-3′ and Lac_groEL_R: 5′-AANGTNCCVCGVATCTTGTT-3′) targeting the previously described Lactobacillus-specific groEL region (Yasrebi et al., 2017) using conditions: 5 min at 95°C, 40 cycles of 30 s at 95°C, 30 s at 58°C, 60 s at 72°C, ending with 72°C for 7 min. A melt curve stage of 15 s at 95°C, 60 s at 60°C, and 15 s at 95°C ended the run, a single peak confirmed amplification of a single PCR product.

3.2 Cannabidiol Improved Markers of Intestinal Inflammation and Gut Barrier

We investigated the impact of CBD on intestinal markers of gut inflammation, gut barrier integrity, and nutrient metabolism. Compared to VEH-treated OVX mice, CBD-treated OVX mice showed a trend of lower Il1b (p = 0.146) and Il6 (p = 0.124) mRNA levels in ileal tissues (Figure 2A) and significantly lower Il1b, Il6 and Tnf in colon tissues (Figure 2B). CBD-treated SS mice had lower ileal expression of Il1b and Il6 (Figure 2A) and decreased colonic expression of Il1b, Il6 and Tnf (Figure 2B). Compared to the VEH-treated OVX group, CBD-treated OVX mice had increased colonic mRNA levels of Tjp1, encoding tight junction protein 1 (i.e., zonula occludens-1, ZO-1) while Ocln, encoding occludin was unchanged (Figure 2B); expression of Ocln and Tjp1 were unchanged in ileum (Figure 2A). Compared to the VEH-treated SS group, the CBD-treated SS group showed decreased Ocln and Tjp1 in colon tissues (Figure 2B), an unexpected result given the concomitant increase in ileal Tjp1 (Figure 2A) and reduced mRNA levels of Il1b and Il6 in the ileum and colon tissues of CBD-treated SS mice (Figures 2A,B). These data indicated that CBD treatment reduced intestinal inflammation in SS and OVX groups, possibly due to improved gut barrier integrity in the OVX group.

To investigate explanations for the improved glucose tolerance, we further examined ileal gene expression of nutrient signaling peptides, glucose transporters, and bile acid receptors. Gcg encodes preproglucagon, which is cleaved by PC1/3 protease (Pcsk1) to yield the incretin peptide Glp1 and Glp2, which regulates gut barrier, intestinal hexose transport, and bone metabolism (Schiellerup et al., 2019). Gcg expression was unaffected by CBD or OVX (Supplementary Figure S3A). Pcsk1 expression was similar between VEH- and CBD-treated OVX groups; however, reduced Pcsk1 expression was observed in the CBD-treated SS and VEH-treated OVX groups when compared to the VEH-treated SS group (Supplementary Figure S3A) suggesting reduced production of GLP1 and GLP2 peptides. CBD treatment did not alter ileal gene expression of carbohydrate transporters SGLT1 and GLUT2, encoded by Slc5a1 and Slc2a2, respectively, suggesting improved glucose tolerance was not a consequence of altered carbohydrate absorption (Supplementary Figure S3A). Membrane bound Takeda G-protein coupled receptor 5 (TGR5), encoded by Gpbar1, and nuclear transcription factor farnesoid-X receptor (FXR), encoded by Nr1h4, are bile acid receptors regulating glucose metabolism (Pathak et al., 2017). Significant differences were not detected in ileal mRNA levels of Nr1h4, Gpbar1 or Fabp6, a target gene of activated FXR (Supplementary Figure S3A). These data suggest that improved glucose tolerance in CBD-treated OVX mice was independent of changes to intestinal carbohydrate absorption, incretin production, or BA receptor expression.

3.3 Cannabidiol Attenuated Ovariectomized-Induced Osteoporosis and Suppressed Femoral Markers of Bone Resorption and Inflammation

Dual-energy x-ray absorptiometry was performed to evaluate BMC and aBMD after 18-weeks of CBD treatment. Compared to SS groups, CBD treatment partially reversed the OVX-induced decrease in whole-body aBMD (Figure 3A) and completely reversed the OVX-induced decrease in whole body BMC (Figure 3B). CBD did not alter whole-body BMC or aBMD of mice in the SS group (Figures 3A,B). Compared to the SS groups, VEH-treated OVX mice showed a significant decrease in femoral aBMD, which was rescued by CBD treatment (Figure 3C). Femoral BMC (Figure 3D), tibial aBMD and BMC (Figures 3E,F), humeral BMC (Figure 3H), and spinal BMC (Figure 3J) were not altered by OVX or CBD treatment. CBD did not affect OVX-induced reductions in aBMD of humerus (Figure 3G) and spine (Figure 3I).

The resilience of bone to fracturing depends on bone microarchitecture, an indicator of bone strength and quality. Femur segments were therefore subject to microCT to assess the metaphysis area of trabecular (i.e., spongy) bone and the diaphysis area of cortical (i.e., compact) bone. Representative microCT images of femoral trabecular and cortical bone morphology for each treatment group are shown in Figures 4A,B. CBD treatment was able to reverse the OVX-induced reduction in trabecular bone volume fraction (BV/TV) or percentage of spongy bone tissue, raising it to levels comparable with SS groups (Figure 4C). CBD-treated OVX mice had higher trabecular thickness (Tb.Th) than VEH-treated OVX mice and SS groups (Figure 4D). Compared to SS groups, CBD partially reversed the OVX-induced decrease in volumetric bone mineral density (vBMD, Figure 4E). OVX groups had lower trabecular number (Tb.N) indicating lower density of trabeculae and higher trabecular separation (Tb.Sp) indicating more distance between trabeculae; however, these endpoints were not significantly changed by CBD (Figures 4F,G). Compared to the VEH-treated SS group, the mean total cross-sectional tissue area (Tt.Ar) and mean total cross-sectional bone area (Ct.Ar) of cortical bone were unaffected by OVX; however, the CBD-treated SS group showed an increase in these parameters (Figures 4H,I). CBD had no significant effect on the OVX-induced decrease in cortical bone fraction (Ct.Ar/Tt.Ar, Figure 4J) and cortical thickness (Ct.Th, Figure 4K). Polar moment of inertia (PMI), a measure of mechanical rigidity to torsion stress or twisting, was significantly lower in the VEH-treated OVX groups compared to the CBD-treated SS group, but otherwise similar between other groups (Figure 4L). CBD did not alter OVX-induced reduction in cortical tissue mineral density (TMD, Figure 4M).

Femoral gene expression was investigated to uncover possible molecular explanations for the improved bone phenotypes. We assessed tumor necrosis factor receptor superfamily (Tnfrsf) genes encoding members of the RANK/RANKL/OPG signaling pathway as well as tartrate-resistant acid phosphatase type 5 (TRAP), an osteoclastic marker of bone resorption (Udagawa et al., 2021). Compared to VEH-treated OVX mice, CBD-treated OVX mice showed a significant decrease in femoral gene expression of Tnfrsf11, encoding RANKL (Figure 4N), indicating a disruption of the RANKL-RANK interaction and decreased bone resorption. Expression of Tnfrsf11a, Tnfrsf11b, and Acp5 encoding RANK, OPG, and TRAP respectively, was not significantly altered by ovariectomy or CBD (Supplementary Figure S3B). Compared to the VEH-treated OVX group or the CBD-treated SS group, the OPG/RANKL mRNA ratio was significantly higher in CBD-treated OVX mice (Supplementary Figure S3C) suggesting bone formation and inhibition of osteoclast differentiation (Xu et al., 2019). The RANKL/OPG mRNA ratio was not different between the four groups (Supplementary Figure S3C) suggesting similar bone resorption and turnover (Jura-Półtorak et al., 2021). Compared to the VEH-treated SS group, femoral expression of Il6 was increased in VEH-treated OVX mice, and this effect was reversed by CBD-treatment (Figure 4N). Il1b and Tnf levels were similar across groups (Supplementary Figure S3B).

Stimulation of endocannabinoid receptor CB2 was reported to inhibit RANKL release leading to reduced osteoclastogenesis while stimulation of CB1 and TRPV1 activates osteoclastogenesis and bone resorption (Idris et al., 2005; Rossi et al., 2015). CBD is an antagonist of GPR55 that has been shown to inhibit bone resorption in male mice (dosed at 10 mg/kg, 3 times per week), although changes were not significant (Whyte et al., 2009). CBD-treated OVX mice had increased femoral mRNA levels of Cnr2, encoding CB2 (Figure 4N), while mRNA levels of Cnr1 (encoding CB1), Trpv1, and Gpr55 were similar among the four treatment groups (Supplementary Figure S3B).

BA receptors TGR5 and FXR are expressed in osteoblasts and osteoclasts where they regulate bone metabolism (Zhao et al., 2020). Activation of FXR has been shown to suppress osteoclast differentiation, increase osteoblast differentiation (Cho et al., 2013), and promote expression of TGR5 (Pathak et al., 2017). TGR5 inhibited osteoclast differentiation while loss of TGR5 promoted osteoclast differentiation and bone resorption (Li et al., 2019). Compared to VEH-treatment, CBD-treated OVX and SS groups had increased levels of Gpbar1 mRNA encoding TGR5 (Figure 4N) suggesting suppression of osteoclast differentiation. CBD did not alter expression of Nr1h4 or the FXR target gene, Nr0b2 encoding small heterodimer protein (SHP) (Supplementary Figure S3B).

3.4 Cannabidiol Did Not Alter Short Chain Fatty Acids But Induced Bile Acid Changes Consistent With Decreased Inflammation and Improved Glucose and Bone Metabolism

SCFAs levels were not altered by ovariectomy or CBD treatment (Supplementary Figure S4). Bile acid (BA) species have been reported to alter gut bacteria (Ridlon et al., 2014), regulate inflammation (Chen et al., 2019), glucose metabolism (Ahmad and Haeusler, 2019), and bone turnover (Cho et al., 2013). Primary and secondary BAs (i.e., PBAs and SBAs) in ileum content and serum were therefore quantified by LC-MS using external and internal standards.

Serum levels of total BAs, PBAs, SBAs, conjugated BAs, and unconjugated BAs were similar in VEH- and CBD-treated OVX groups (Table 1). Compared to the VEH-treated SS group, total BAs and PBAs were significantly decreased in serum of the CBD-treated SS group (Table 1); specifically individual PBAs that were decreased in the serum of the CBD-treated SS group included tauroursodeoxycholic acid (TUDCA), tauro-α-muricholic acid (TαMCA) and ursodeoxycholic acid (UDCA). Compared to VEH-treated OVX mice, TUDCA showed a trending increase in serum of CBD-treated OVX mice (q = 0.11; Table 1). TUDCA is FDA-approved for treatment of cholestatic liver disease and has also been reported to promote differentiation of osteoblasts, improve markers of bone quality, suppress inflammatory cytokines, and improve glucose tolerance (Yanguas-Casás et al., 2017; Kim et al., 2018; Zangerolamo et al., 2021). Glycocholic acid (GCA), a PBA which has been identified as a biomarker for hepatocellular carcinoma (Li H. et al., 2017) was decreased to non-detectable levels in the serum of CBD-treated OVX mice compared to VEH-treated OVX mice (Table 1). Serum TαMCA and TωMCA, which are FXR antagonists associated with prevention of hepatic cholestasis (Takikawa et al., 1997; Sayin et al., 2013), were decreased in CBD-treated OVX mice, compared to VEH-treated OVX mice (Table 1). Compared to the VEH-treated SS mice, DCA was reduced in the CBD-treated SS group. High deoxycholic acid (DCA) levels have been associated with low BMD in people over 60 (Jovanovich et al., 2018), with colorectal cancer (Bayerdörffer et al., 1995), and with gut dysbiosis, intestinal inflammation, increased bile acid pool, and lower intestinal FXR activity (Wang et al., 2020; Xu et al., 2021a; Xu et al., 2021b).

In ileal content, the concentrations of total BAs, PBAs, and SBAs were similar among treatment groups. SBAs in the CBD-treated OVX group trended lower than the VEH-treated OVX group (q = 0.06, p = 0.02), perhaps due to reduced hyocholic acid (HCA) and ωMCA in CBD-treated OVX mice (Table 1). Compared to SS groups, the VEH-treated OVX group had increased concentration of glycine-conjugated BAs; however, levels were reduced to that of the SS groups in the CBD-treated OVX group (Table 1). Comparison of individual ileal content bile acids in VEH-treated SS and OVX groups revealed that glycoursodeoxycholic acid (GUDCA), glycochenodeoxycholic acid (GCDCA), GCA, HCA, ωMCA, and taurohyodeoxycholic acid (THDCA) were increased in OVX mice; however, in CBD-treated OVX mice the concentrations of these BAs were decreased to levels more consistent with SS groups (Table 1). Increased glycine-conjugated BA in small intestine, specifically GCDCA and GCA, has been correlated with increased intestinal inflammation in indomethacin-treated rats (Lázár et al., 2021). Taurocholic acid (TCA) trended higher in VEH-treated OVX mice compared to the SS groups (q = 0.09) but was significantly lower in the CBD-treated OVX mice (Table 1). Within the SS groups, CBD did not significantly change the profile of ileal BAs except for decreased taurolithocholic acid (TLCA, Table 1).

We thank Jiadong Li for assistance with ovariectomy surgeries and the IFNH GC-MS core facility for analysis of SCFAs. We thank Ali Yasrebi, Joseph Chang, Sofia Lamberto, Kimberly Knowles, and Esther Mezhibovsky for technical support.