Expanding the number of clinically approved choices for osteoanabolic therapy represents the next hurdle on the osteoporosis treatment horizon. The WNT pathway is involved in stimulating new bone formation, and the most recent FDA-approved anabolic therapy-sclerostin neutralizing antibody-works by stimulating the WNT pathway in bone. The challenge with all current osteoanabolic therapies is the short treatment window in which they are efficacious. One strategy to dealing with this limited anabolic window is to further maximize bone formation during the window, using combination therapy. For example, simultaneous pharmacological or genetic inhibition of the WNT antagonists sclerostin and DKK1 potentiate the already strong effects of sclerostin inhibition alone, particularly in cancellous bone. Considerable interest has emerged regarding other candidates that might be co-inhibited along with sclerostin to potentiate its effects in bone, with specific action on cortical bone. In this communication, partnering sclerostin inhibition with the inhibition of another secreted WNT antagonist, NOTUM, is explored. NOTUM is a secreted WNT deacylase that inactivates WNT ligands and has preferential effects on cortical bone. To evaluate combination therapy involving sclerostin/NOTUM inhibition, double knockout mice for Sost and Notum (Sost-/-;Notum-/-) were generated and compared to single knockouts and WT controls. Further experiments were conducted using pharmacologic inhibitors, rather than genomic mutations, for sclerostin (sclerostin neutralizing antibody) and a small molecule inhibitor of NOTUM (LP-922056). Each experiment included evaluation by DXA-derived radiography, μCT, biomechanical testing and bone dynamic histomorphometry. Deletion of Notum alone had mild cortical effects but co-deletion of Sost and Notum improved cortical and some cancellous parameters significantly beyond Sost-/- mice. Co-inhibition of the protein products with antibody/small molecule were less synergistic, with only a small cortical effect, particularly in younger mice. Taken together, the results suggest the potential to improve efficacy of sclerostin inhibition using NOTUM inhibition is promising, but development of additional NOTUM inhibiting tools and optimization of current tools might be necessary to strengthen the partnership.

Genes regulating body fat are shared with high fidelity by mice and humans, indicating that mouse knockout (KO) phenotyping might identify valuable antiobesity drug targets. Male Mrs2 magnesium transporter (Mrs2) KO mice were recently reported as thin when fed a high-fat diet (HFD). They also exhibited increased energy expenditure (EE)/body weight and had beiged adipocytes that, along with isolated hepatocytes, demonstrated increased oxygen consumption, suggesting that increased EE drove the thin phenotype. Here we provide our data on these and additional assays in Mrs2 KO mice. We generated Mrs2 KO mice by homologous recombination. HFD-fed male and female Mrs2 KO mice had significantly less body fat, measured by quantitative magnetic resonance, than wild-type (WT) littermates. HFD-fed Mrs2 KO mice did not demonstrate increased EE by indirect calorimetry and could not maintain body temperature at 4 °C, consistent with their decreased brown adipose tissue stores but despite increased beige white adipose tissue. Instead, when provided a choice between HFD and low-fat diet (LFD), Mrs2 KO mice showed a significant 15% decrease in total energy intake resulting from significantly lower HFD intake that offset numerically increased LFD intake. Food restriction studies performed using WT mice suggested that this decrease in energy intake could explain the loss of body fat. Oral glucose tolerance test studies revealed significantly improved insulin sensitivity in Mrs2 KO mice. We conclude that HFD-fed Mrs2 KO mice are thin with improved insulin sensitivity, and that this favorable metabolic phenotype is driven by hypophagia. Further evaluation is warranted to determine the suitability of MRS2 as a drug target for antiobesity therapeutics.

Mutations in SFRP4 cause Pyle’s bone disease with wide metaphyses and increased skeletal fragility. The WNT signaling pathway plays important roles in determining skeletal architecture and SFRP4 is a secreted Frizzled decoy receptor that inhibits WNT signaling. Seven cohorts of male and female Sfrp4 gene knockout mice, examined through 2 years of age, had a normal lifespan but showed cortical and trabecular bone phenotypes. Mimicking human Erlenmeyer flask deformities, bone cross-sectional areas were elevated 2-fold in the distal femur and proximal tibia but only 30% in femur and tibia shafts. Reduced cortical bone thickness was observed in the vertebral body, midshaft femur and distal tibia. Elevated trabecular bone mass and numbers were observed in the vertebral body, distal femur metaphysis and proximal tibia metaphysis. Midshaft femurs retained extensive trabecular bone through 2 years of age. Vertebral bodies had increased compressive strength, but femur shafts had reduced bending strength. Trabecular, but not cortical, bone parameters in heterozygous Sfrp4 mice were modestly affected. Ovariectomy resulted in similar declines in both cortical and trabecular bone mass in wild-type and Sfrp4 KO mice. SFRP4 is critical for metaphyseal bone modeling involved in determining bone width. Sfrp4 KO mice show similar skeletal architecture and bone fragility deficits observed in patients with Pyle’s disease with SFRP4 mutations.

Purpose Humans with haploinsufficiency of GPR75, an orphan GPCR, are thin. Gpr75 knockout (KO) mice are also thin with improved glucose homeostasis. We wanted to confirm these findings in Gpr75 KO mice and determine whether decreased energy intake and/or increased energy expenditure contributed to the thin phenotype. Methods Gpr75 KO mice were generated by homologous recombination. All studies compared female and male Gpr75 KO mice to their wild type (WT) littermates. Body composition was measured by DXA and QMR technologies. Glucose homeostasis was evaluated by measuring glucose and insulin levels during oral glucose tolerance tests (OGTTs). Food intake was measured in group-housed mice. In singly housed mice, energy expenditure was measured in Oxymax indirect calorimetry chambers, and locomotor activity was measured in Oxymax and Photobeam Activity System chambers. Results In all 12 cohorts of adult female or male mice, Gpr75 KO mice had less body fat; pooled data showed that, compared to WT littermates (n = 103), Gpr75 KO mice (n = 118) had 49% less body fat and 4% less LBM (P < 0.001 for each). KO mice also had 8% less body fat at weaning (P < 0.05), and during the month after weaning as the thin phenotype became more exaggerated, Gpr75 KO mice ate significantly less than, but had energy expenditure and activity levels comparable to, their WT littermates. During OGTTs, Gpr75 KO mice showed improved glucose tolerance (glucose AUC 23% lower in females, P < 0.05, and 26% lower in males, P < 0.001), accompanied by significantly decreased insulin levels and significantly increased insulin sensitivity indices. Conclusion Gpr75 KO mice are thin at weaning, are hypophagic as the thin phenotype becomes more exaggerated, and exhibit improved glucose tolerance and insulin sensitivity as healthy-appearing adults. These results suggest that inhibiting GPR75 in obese humans may safely decrease energy intake and body fat while improving glucose tolerance and insulin sensitivity.

Purpose Obesity is a major public health problem. Understanding which genes contribute to obesity may better predict individual risk and allow development of new therapies. Because obesity of a mouse gene knockout (KO) line predicts an association of the orthologous human gene with obesity, we reviewed data from the Lexicon Genome5000TM high throughput phenotypic screen (HTS) of mouse gene KOs to identify KO lines with high body fat. Materials and Methods KO lines were generated using homologous recombination or gene trapping technologies. HTS body composition analyses were performed on adult wild-type and homozygous KO littermate mice from 3758 druggable mouse genes having a human ortholog. Body composition was measured by either DXA or QMR on chow-fed cohorts from all 3758 KO lines and was measured by QMR on independent high fat diet-fed cohorts from 2488 of these KO lines. Where possible, comparisons were made to HTS data from the International Mouse Phenotyping Consortium (IMPC). Results Body fat data are presented for 75 KO lines. Of 46 KO lines where independent external published and/or IMPC KO lines are reported as obese, 43 had increased body fat. For the remaining 29 novel high body fat KO lines, Ksr2 and G2e3 are supported by data from additional independent KO cohorts, 6 (Asnsd1, Srpk2, Dpp8, Cxxc4, Tenm3 and Kiss1) are supported by data from additional internal cohorts, and the remaining 21 including Tle4, Ak5, Ntm, Tusc3, Ankk1, Mfap3l, Prok2 and Prokr2 were studied with HTS cohorts only. Conclusion These data support the finding of high body fat in 43 independent external published and/or IMPC KO lines. A novel obese phenotype was identified in 29 additional KO lines, with 27 still lacking the external confirmation now provided for Ksr2 and G2e3 KO mice. Undoubtedly, many mammalian obesity genes remain to be identified and characterized.

The 2019 International Skeletal Dysplasia Society nosology update lists 441 genes for which mutations result in rare human skeletal disorders. These genes code for enzymes (33%), scaffolding proteins (18%), signal transduction proteins (16%), transcription factors (14%), cilia proteins (8%), extracellular matrix proteins (5%), and membrane transporters (4%). Skeletal disorders include aggrecanopathies, channelopathies, ciliopathies, cohesinopathies, laminopathies, linkeropathies, lysosomal storage diseases, protein-folding and RNA splicing defects, and ribosomopathies. With the goal of evaluating the ability of mouse models to mimic these human genetic skeletal disorders, a PubMed literature search identified 260 genes for which mutant mice were examined for skeletal phenotypes. These mouse models included spontaneous and ENU-induced mutants, global and conditional gene knockouts, and transgenic mice with gene over-expression or specific base-pair substitutions. The human X-linked gene ARSE and small nuclear RNA U4ATAC, a component of the minor spliceosome, do not have mouse homologs. Mouse skeletal phenotypes mimicking human skeletal disorders were observed in 249 of the 260 genes (96%) for which comparisons are possible. A supplemental table in spreadsheet format provides PubMed weblinks to representative publications of mutant mouse skeletal phenotypes. Mutations in 11 mouse genes (Ccn6, Cyp2r1, Flna, Galns, Gna13, Lemd3, Manba, Mnx1, Nsd1, Plod1, Smarcal1) do not result in similar skeletal phenotypes observed with mutations of the homologous human genes. These discrepancies can result from failure of mouse models to mimic the exact human gene mutations. There are no obvious commonalities among these 11 genes. Body BMD and/or radiologic dysmorphology phenotypes were successfully identified for 28 genes by the International Mouse Phenotyping Consortium (IMPC). Forward genetics using ENU mouse mutagenesis successfully identified 37 nosology gene phenotypes. Since many human genetic disorders involve hypomorphic, gain-of-function, dominant-negative and intronic mutations, future studies will undoubtedly utilize CRISPR/Cas9 technology to examine transgenic mice having genes modified to exactly mimic variant human sequences. Mutant mice will increasingly be employed for drug development studies designed to treat human genetic skeletal disorders.SIGNIFICANCE:Great progress is being made identifying mutant genes responsible for human rare genetic skeletal disorders and mouse models for genes affecting bone mass, architecture, mineralization and strength. This review organizes data for 441 human genetic bone disorders with regard to heredity, gene function, molecular pathways, and fidelity of relevant mouse models to mimic the human skeletal disorders. PubMed weblinks to citations of 249 successful mouse models are provided.

Purpose In humans, single nucleotide polymorphisms (SNPs) near the adjacent protein kinase D1 (PRKD1) and G2/M-phase-specific E3 ubiquitin protein ligase (G2E3) genes on chromosome 14 are associated with obesity. To date, no published evidence links inactivation of either gene to changes in body fat. These two genes are also adjacent on mouse chromosome 12. Because obesity genes are highly conserved between humans and mice, we analyzed body fat in adult G2e3 and Prkd1 knockout (KO) mice to determine whether inactivating either gene leads to obesity in mice and, by inference, probably in humans. Methods The G2e3 and Prkd1 KO lines were generated by gene trapping and by homologous recombination methodologies, respectively. Body fat was measured by DEXA in adult mice fed chow from weaning and by QMR in a separate cohort of mice fed high-fat diet (HFD) from weaning. Glucose homeostasis was evaluated with oral glucose tolerance tests (OGTTs) performed on adult mice fed HFD from weaning. Results Body fat was increased in multiple cohorts of G2e3 KO mice relative to their wild-type (WT) littermates. When data from all G2e3 KO (n=32) and WT (n=31) mice were compared, KO mice showed increases of 11% in body weight (P<0.01), 65% in body fat (P<0.001), 48% in % body fat (P<0.001), and an insignificant 3% decrease in lean body mass. G2e3 KO mice were also glucose intolerant during an OGTT (P<0.05). In contrast, Prkd1 KO and WT mice had comparable body fat levels and glucose tolerance. Conclusion Significant obesity and glucose intolerance were observed in G2e3, but not Prkd1, KO mice. The conservation of obesity genes between mice and humans strongly suggests that the obesity-associated SNPs located near the human G2E3 and PRKD1 genes are linked to variants that decrease the amount of functional human G2E3.

The genetic landscape of diseases associated with changes in bone mineral density (BMD), such as osteoporosis, is only partially understood. Here, we explored data from 3,823 mutant mouse strains for BMD, a measure that is frequently altered in a range of bone pathologies, including osteoporosis. A total of 200 genes were found to significantly affect BMD. This pool of BMD genes comprised 141 genes with previously unknown functions in bone biology and was complementary to pools derived from recent human studies. Nineteen of the 141 genes also caused skeletal abnormalities. Examination of the BMD genes in osteoclasts and osteoblasts underscored BMD pathways, including vesicle transport, in these cells and together with in silico bone turnover studies resulted in the prioritization of candidate genes for further investigation. Overall, the results add novel pathophysiological and molecular insight into bone health and disease.

The 11 existing FDA-approved osteoporosis drug treatments include hormone replacement therapy, 2 SERMs (raloxifene and bazedoxifene), 5 inhibitors of bone-resorbing osteoclasts (4 bisphosphonates and anti-RANKL denosumab), 2 parathyroid hormone analogues (teriparatide and abaloparatide), and 1 WNT signaling enhancer (romosozumab). These therapies are effective and provide multiple options for patients and physicians. As the genomic revolution continues, potential novel targets for future drug development are identified. This review takes a wide perspective to describe potentially rewarding topics to explore, including knowledge of genes and pathways involved in bone cell metabolism, the utility of animal models, targeting drugs to bone, and ongoing advances in drug design and delivery.

Mice with an inactivating mutation in the gene encoding asparagine synthetase domain containing 1 (ASNSD1) develop a progressive degenerative myopathy that results in severe sarcopenia and myosteatosis. ASNSD1 is conserved across many species, and whole body gene expression surveys show maximal expression levels of ASNSD1 in skeletal muscle. However, potential functions of this protein have not been previously reported. Asnsd1−/− mice demonstrated severe muscle weakness, and their normalized body fat percentage on both normal chow and high fat diets was greater than 2 SD above the mean for 3651 chow-fed and 2463 high-fat-diet–fed knockout (KO) lines tested. Histologic lesions were essentially limited to the muscle and were characterized by a progressive degenerative myopathy with extensive transdifferentiation and replacement of muscle by well-differentiated adipose tissue. There was minimal inflammation, fibrosis, and muscle regeneration associated with this myopathy. In addition, the absence of any signs of lipotoxicity in Asnsd1−/− mice despite their extremely elevated body fat percentage and low muscle mass suggests a role for metabolic dysfunctions in the development of this phenotype. Asnsd1−/− mice provide the first insight into the function of this protein, and this mouse model could prove useful in elucidating fundamental metabolic interactions between skeletal muscle and adipose tissue.

The disability, mortality and costs caused by non-vertebral osteoporotic fractures are enormous. Existing osteoporosis therapies are highly effective at reducing vertebral but not non-vertebral fractures. Cortical bone is a major determinant of non-vertebral bone strength. To identify novel osteoporosis drug targets, we phenotyped cortical bone of 3 366 viable mouse strains with global knockouts of druggable genes. Cortical bone thickness was substantially elevated in Notum(-/-) mice. NOTUM is a secreted WNT lipase and we observed high NOTUM expression in cortical bone and osteoblasts but not osteoclasts. Three orally active small molecules and a neutralizing antibody inhibiting NOTUM lipase activity were developed. They increased cortical bone thickness and strength at multiple skeletal sites in both gonadal intact and ovariectomized rodents by stimulating endocortical bone formation. Thus, inhibition of NOTUM activity is a potential novel anabolic therapy for strengthening cortical bone and preventing non-vertebral fractures.

Two large-scale mouse gene knockout phenotyping campaigns have provided extensive data on the functions of thousands of mammalian genes. The ongoing International Mouse Phenotyping Consortium (IMPC), with the goal of examining all ∼20,000 mouse genes, has examined 5115 genes since 2011, and phenotypic data from several analyses are available on the IMPC website (www.mousephenotype.org). Mutant mice having at least one human genetic disease-associated phenotype are available for 185 IMPC genes. Lexicon Pharmaceuticals' Genome5000™ campaign performed similar analyses between 2000 and the end of 2008 focusing on the druggable genome, including enzymes, receptors, transporters, channels and secreted proteins. Mutants (4654 genes, with 3762 viable adult homozygous lines) with therapeutically interesting phenotypes were studied extensively. Importantly, phenotypes for 29 Lexicon mouse gene knockouts were published prior to observations of similar phenotypes resulting from homologous mutations in human genetic disorders. Knockout mouse phenotypes for an additional 30 genes mimicked previously published human genetic disorders. Several of these models have helped develop effective treatments for human diseases. For example, studying Tph1 knockout mice (lacking peripheral serotonin) aided the development of telotristat ethyl, an approved treatment for carcinoid syndrome. Sglt1 (also known as Slc5a1) and Sglt2 (also known as Slc5a2) knockout mice were employed to develop sotagliflozin, a dual SGLT1/SGLT2 inhibitor having success in clinical trials for diabetes. Clinical trials evaluating inhibitors of AAK1 (neuropathic pain) and SGLT1 (diabetes) are underway. The research community can take advantage of these unbiased analyses of gene function in mice, including the minimally studied 'ignorome' genes.

Hip fractures at the femoral neck are a major cause of morbidity and mortality, but aside from biomechanical strength testing, little is known about femoral neck architecture in mice. Procedures were optimized to analyze high-resolution (6 μm voxel size) microCT scans of the mouse femoral neck to provide bone mass and architectural information. Similar to histomorphometric observations in rats, the boundary between cortical and trabecular bone is difficult to identify in the mouse femoral mid-neck and these compartments were not analyzed separately. Analyses included total area, mineralized bone area, and bone volume fraction (BV/TV). Femoral neck architecture varies in C57BL/6J, 129/SvEv and BALB/c mouse strains. Bone cross sectional area and BV/TV were low in Lrp5 but elevated in Sost gene knockout mice. Sfrp4 gene knockout resulted in high total area, normal bone area, low BV/TV and, as indicated by BS/BV values, greater trabecularization. Femoral neck BV/TV declined with age and ovariectomy, but increased with teriparatide treatment. These findings demonstrate that the architecture of the mouse femoral neck mimics phenotypes and treatment effects observed at other skeletal sites and is a relevant bone site for translational studies examining osteoporosis therapies.