Animal models are essential tools for addressing fundamental scientific questions about skeletal diseases and for the development of new therapeutic approaches. Traditionally, mice have been the most common model organism in biomedical research, but their use is hampered by several limitations including complex generation, demanding investigation of early developmental stages, regulatory restrictions on breeding, and high maintenance cost. The zebrafish has been used as an efficient alternative vertebrate model for the study of human skeletal diseases, thanks to its easy genetic manipulation, high fecundity, external fertilization, transparency of rapidly developing embryos, and low maintenance cost. Furthermore, zebrafish share similar skeletal cells and ossification types with mammals. In the last decades, the use of both forward and new reverse genetics techniques has resulted in the generation of many mutant lines carrying skeletal phenotypes associated with human diseases. In addition, transgenic lines expressing fluorescent proteins under bone cell- or pathway- specific promoters enable in vivo imaging of differentiation and signaling at the cellular level. Despite the small size of the zebrafish, many traditional techniques for skeletal phenotyping, such as x-ray and microCT imaging and histological approaches, can be applied using the appropriate equipment and custom protocols. The ability of adult zebrafish to remodel skeletal tissues can be exploited as a unique tool to investigate bone formation and repair. Finally, the permeability of embryos to chemicals dissolved in water, together with the availability of large numbers of small-sized animals makes zebrafish a perfect model for high-throughput bone anabolic drug screening. This review aims to discuss the techniques that make zebrafish a powerful model to investigate the molecular and physiological basis of skeletal disorders.

Over the last years the zebrafish imposed itself as a powerful model to study skeletal diseases, but a limit to its use is the poor characterization of collagen type I, the most abundant protein in bone and skin. In tetrapods collagen type I is a trimer mainly composed of two α1 chains and one α2 chain, encoded by COL1A1 and COL1A2 genes, respectively. In contrast, in zebrafish three type I collagen genes exist, col1a1a, col1a1b and col1a2 coding for α1(I), α3(I) and α2(I) chains. During embryonic and larval development the three collagen type I genes showed a similar spatio-temporal expression pattern, indicating their co-regulation and interdependence at these stages. In both embryonic and adult tissues, the presence of the three α(I) chains was demonstrated, although in embryos α1(I) was present in two distinct glycosylated states, suggesting a developmental-specific collagen composition. Even though in adult bone, skin and scales equal amounts of α1(I), α3(I) and α2(I) chains are present, the presented data suggest a tissue-specific stoichiometry and/or post-translational modification status for collagen type I. In conclusion, this data will be useful to properly interpret results and insights gained from zebrafish models of skeletal diseases.

Classical osteogenesis imperfecta (OI) is a bone disease caused by type I collagen mutations and characterized by bone fragility, frequent fractures in absence of trauma and growth deficiency. No definitive cure is available for OI and to develop novel drug therapies, taking advantage of a repositioning strategy, the small teleost zebrafish (Danio rerio) is a particularly appealing model. Its small size, high proliferative rate, embryo transparency and small amount of drug required make zebrafish the model of choice for drug screening studies, when a valid disease model is available. We performed a deep characterization of the zebrafish mutant Chihuahua, that carries a G574D (p.G736D) substitution in the α1 chain of type I collagen. We successfully validated it as a model for classical OI. Growth of mutants was delayed compared with WT. X-ray, µCT, alizarin red/alcian blue and calcein staining revealed severe skeletal deformity, presence of fractures and delayed mineralization. Type I collagen extracted from different tissues showed abnormal electrophoretic migration and low melting temperature. The presence of endoplasmic reticulum (ER) enlargement due to mutant collagen retention in osteoblasts and fibroblasts of mutant fish was shown by electron and confocal microscopy. Two chemical chaperones, 4PBA and TUDCA, were used to ameliorate the cellular stress and indeed 4PBA ameliorated bone mineralization in larvae and skeletal deformities in adult, mainly acting on reducing ER cisternae size and favoring collagen secretion. In conclusion, our data demonstrated that ER stress is a novel target to ameliorate OI phenotype; chemical chaperones such as 4PBA may be, alone or in combination, a new class of molecules to be further investigated for OI treatment.

Autosomal dominant osteogenesis imperfecta (OI) caused by glycine substitutions in type I collagen is a paradigmatic disorder for stem cell therapy. Bone marrow transplantation in OI children has produced a low engraftment rate, but surprisingly encouraging symptomatic improvements. In utero transplantation (IUT) may hold even more promise. However, systematic studies of both methods have so far been limited to a recessive mouse model. In this study, we evaluated intrauterine transplantation of adult bone marrow into heterozygous BrtlIV mice. Brtl is a knockin mouse with a classical glycine substitution in type I collagen [α1(I)-Gly349Cys], dominant trait transmission, and a phenotype resembling moderately severe and lethal OI. Adult bone marrow donor cells from enhanced green fluorescent protein (eGFP) transgenic mice engrafted in hematopoietic and nonhematopoietic tissues differentiated to trabecular and cortical bone cells and synthesized up to 20% of all type I collagen in the host bone. The transplantation eliminated the perinatal lethality of heterozygous BrtlIV mice. At 2 months of age, femora of treated Brtl mice had significant improvement in geometric parameters (P < .05) versus untreated Brtl mice, and their mechanical properties attained wild-type values. Our results suggest that the engrafted cells form bone with higher efficiency than the endogenous cells, supporting IUT as a promising approach for the treatment of genetic bone diseases.

The clinical phenotype in osteogenesis imperfecta (OI) is attributed to the dominant negative function of mutant type I collagen molecules in the extracellular matrix, by altering its structure and function. Intracellular retention of mutant collagen has also been reported, but its effect on cellular homeostasis is less characterized. Using OI patient fibroblasts carrying mutations in the α1(I) and α2(I) chains we demonstrate that retained collagen molecules are responsible for endoplasmic reticulum (ER) enlargement and activation of the unfolded protein response (UPR) mainly through the eukaryotic translation initiation factor 2 alpha kinase 3 (PERK) branch. Cells carrying α1(I) mutations upregulate autophagy, while cells with α2(I) mutations only occasionally activate the autodegradative response. Despite the autophagy activation to face stress conditions, apoptosis occurs in all mutant fibroblasts. To reduce cellular stress, mutant fibroblasts were treated with the FDA-approved chemical chaperone 4-phenylbutyric acid. The drug rescues cell death by modulating UPR activation thanks to both its chaperone and histone deacetylase inhibitor abilities. As chaperone it increases general cellular protein secretion in all patients' cells as well as collagen secretion in cells with the most C-terminal mutation. As histone deacetylase inhibitor it enhances the expression of the autophagic gene Atg5 with a consequent stimulation of autophagy. These results demonstrate that the cellular response to ER stress can be a relevant target to ameliorate OI cell homeostasis.