Several studies demonstrated a clear relationship between aging and bone metabolism in vertebrates, in which early phases associated with changes at molecular and cellular level play a crucial role [2, 17]. Due to the complexity of the regulation systems, the mechanisms are not yet completely understood. Bone markers, systemic metabolic signals, cell behaviour and gene expression could be better investigated in a simple animal model to dissect in detail the early phase of bone aging. Zebrafish is a powerful animal model with many advantages such as small size, low cost of maintenance and rapid development. In addition, thanks to his high conservation of bone structure, it is an excellent model of osteogenesis, bone metabolism and bone remodelling . In particular, the scale represents an optimum model for bone metabolism studies at cellular and molecular level [39, 42]. In the present work, we investigated the early mechanisms of bone aging in zebrafish using the scale as innovative read-out system. Adult fish from 3 to 9 months have been used based on previous studies about age-dependent vertebral microalterations in adult zebrafish [6, 38, 46].
The analysis of the scale ring demonstrated that the osteodeposition rate decreases with age from 3 months to terminate at 9 months with a complete arrest of its growth. These data suggested that the osteodeposition process in fish scale is regulated during the life and coordinated with the body growth, as well as in humans.
In zebrafish scale, the cells responsible for the initial bone matrix deposition are characterized by intense ALP activity  as well as in humans . The histological staining in old-fish scales highlighted an age-dependent downregulation of ALP signal in the matrix-producing cells, suggesting that their activity is progressively switched off during aging, whereas the total ALP activity did not change significantly in the whole scale. In fact, the variation of ALP signal in few matrix-deposing cells (black arrows in Fig. 2b) is clearly too low to be detected with biochemical ALP activity measurement in whole scale.
Since ALP is expressed in the early differentiation process of the human osteoblasts , we hypothesize that the shutdown of the osteodeposition activity is probably due to an impairment of the differentiation processes. To support these data, in humans, age-related bone loss is predominantly due to a reduced osteoblast activity  and associated with a decrease in the number of MSCs and their differentiation potential [31, 56].
The analysis of bone catabolic activity, measured in fish scales through the specific biomarker TRAP, indicates a statistically significant increase related to fish age. In 9-month-old fish, we observed 63% of TRAP-positive scales (Table 1), in which a visible bone resorption process takes place along the scale border.
Osteoclasts and TRAP activity are modulated by human bone aging processes since it has been demonstrated that osteoclast cells become more active on aged bone matrix than on young one [8, 22].
The increase of TRAP activity in aged fish scales could be explained mainly by the upregulation of the tnfrsf1b (homolog of human rank) in 6- and 9-month-old fish. It is known that age exerts multiple effects on expression of gene coding receptors and endogenous bone regulatory factors. In particular, age-related upregulation of RANK may indicate an increase in the number osteoclast progenitors and/or the number of receptors per cell. It has been also demonstrated in mice that such condition could make osteoclast progenitors more responsive to M-CSF and to RANKL . The increase of TRAP activity allows, consequently, to a progressive decline in the ALP/TRAP ratio during aging in fish scales.
The biochemical changes associated to aging in fish involved also the PTH blood level, which increases progressively with age. This condition is found also in humans where sex steroids, calcium and renal function may induce an increase of PTH levels the in blood of aged women and men . The increased serum PTH may act directly on osteoclastic activity increasing bone resorption as demonstrated in the fracture-healing model of orchiectomized mice . The mechanism of PTH-induced osteoclast differentiation and activation could be mediated by parathormone receptor 1 (PTHR1) and Vacuolar-type H+-ATPase (V-ATPase), both expressed in osteoclast precursors . However, we should also consider that environment and diet are different in human and fish.
We conclude that zebrafish possesses a form of senile osteoporosis detectable in the scales and characterized by biochemical, histological and molecular evidences similar to aged human bone. Since we used male animals, the oestrogen hormones were not involved in the regulation of bone metabolism in our model.
The scales, as part of the dermal skeleton, are not subjected to mechanical or gravitational forces, and, in addition, they show great similarity with human bone in terms of cell function and biochemical regulations [40, 47]. For these reasons, the scale represents an excellent model to study the basic mechanisms of metabolic changes in aging bone.
The glucocorticoid PN is a well-known pro-osteoporotic agent in humans  as well as in zebrafish [10, 15, 41].
The treatment of fish with PN inhibits the formation of the new scale ring in fish of all ages and, in addition, induces an important osteoporosis-like phenotype in the scales of older fish. The absence of matrix deposition in young animals indicates that PN has important negative effects on osteoblast activity as confirmed by the reduction of ALP activity after glucocorticoid treatment in the scales of the same age. The bone resorption stimulation of PN can be explained by the enhanced TRAP activity in PN-treated scales with respect to untreated fish of the same age. Indeed, the resorption area is proportional to the TRAP activity level at 6- and 9-month-old fish.
It has been demonstrated that the incidence of vertebral fractures are age-dependent in human patients treated with high-dose of glucocorticoids [33, 49]. Our data support this evidence, suggesting that, in humans as well in fish, these pro-osteoporotic drugs are more effective on old bone than younger one.
The increase of osteoclast activity may be due to a PN-dependent over-stimulation of tnfsf11 (homolog of human rankl) in 6- and 9-month-old fish, which causes the elevation of tnfsf11/tnfrsf11b (rankl/opg) ratio. These data suggest that PN stimulates osteoclast activity, enhancing the catabolic stimuli, which are already elevated in the old bone due to the upregulation of RANK (Table 2).
Although not performed on proteins, the gene expression analysis of OPG, RANKL and OPG/RANKL ratio can be informative in zebrafish as demonstrated in dexamethasone-treated model  and in adult scales after biophysical stimulation .
Talking about molecular mechanisms, it has been demonstrated that glucocorticoid treatment promotes osteoclast differentiation and activity increasing RANKL production  In vitro, RANKL has been found to induce cell-cell fusion during osteoclast differentiation and syncytial organization . Our data indicated that a similar mechanism may be used by PN in zebrafish scale.
In conclusion, the present work demonstrated for the first time that zebrafish develops a form of bone aging detectable in the scales where the osteoblasts become progressively inactive from 6 to 9 months and the osteoclasts are activated at 9 months. These data confirm that early crucial events of bone aging in zebrafish should be studied from 3 to 9 months. In fact, previous works on zebrafish have demonstrated that structural microalterations were detected in spine between 6 and 9 months [6, 38, 46], as well as in male rodent models, where the early events of age-related osteopenia have been detected around few months (., ).
Interestingly, these data indicates that fish scale and male human skeleton share a very similar growth curve in which osteodeposition rate is maximal in juvenile stage (< 6 months/30 years) and slightly declines in early adult stage (6–9 months/30–45 years), generating a negative bone balance (Fig. 6).
Senile male osteoporosis in humans is associate with a reduction in BMD between 0.5 and 1% per year , while in female is reported between 2 and 4% per year in the first 5 to 10 years after menopause . Considering the differences between humans and fish, we can assume that a reduction of 1.02% in fish scale surface could be considered a model of senile male osteoporosis comparable with that of humans.
In aged male fish, the bone resorption activity in the scale of adult fish is accelerated when treated with pro-osteoporotic agents like prednisolone (Fig. 6). In fact, also in humans, osteoclast activity is higher in aged bones .
The whole life span of fish does not have linear timeline with respect to humans because of phylogenetic distance, like rodents [12, 45]. Nevertheless, considering that Danio rerio reproductive phase spans, more or less, between 3 mpf (human 12–14 years, adolescence) and 12 mpf (human 55 years), the curve of bone mass seems to be similar in fish and humans. Further studies will be addressed to verify the persistence and any modulation of bone loss phenotype in the scales of older fish (> 1 year). This study will better elucidate the mechanisms of the clinical complications in the late phase of age-dependent bone loss.
In conclusion, despite the anatomical and phylogenetic distance, the model of senile male osteoporosis in zebrafish scale can help to understand the early mechanisms of the physiological bone aging and screen potential new anti-osteoporotic drugs.