This study demonstrated that the mouse growth plate calcifies during aging but also that it remains calcified cartilage for a long time without being replaced by bone. Evidence supporting this includes the facts that the growth plate at W55 was strongly stained for hematoxylin and was positive for Alcian blue, whereas bone was stained for eosin and negative for Alcian blue; therefore the W55 growth plate was determined to be cartilage, not bone. In addition, SEM/EDS analysis showed mineral accumulation within the growth plate at W55, indicating that the calcified tissue in the mouse growth plate at W55 was calcified cartilage instead of bone. In humans, invasive histological examinations have limited applicability, and noninvasive radiological analysis is the most common method used to investigate epiphyseal fusion. However, radiological approaches such as CT cannot determine whether calcified regions are comprised of bone or cartilage. Therefore, detailed histological findings on human epiphyseal fusion remain limited. Based on the results of this study, the human growth plate may also contain calcified cartilage for some period of time prior to replacement with bone.

In this study, the core protein of aggrecan was degraded at W55, while GAGs of aggrecan were accumulated within the growth plate at the same time point. Previous studies have shown that proteoglycans inhibit calcification in vitro [14,15,16]; however, GAGs have been reported to promote calcification. Wojtas et al. used an in vitro model based on mouse dental tissues to show that GAGs promote calcification in calcified tissue. In this experiment, the researchers prepared several GAG-removed groups using enzymatic treatment, a group in which non-collagenous protein was removed by trypsin treatment, and an untreated control group. The dental tissues of these groups were first demineralized, then remineralized using a calcium phosphate solution. Their results showed that the rate of mineralization in calcified tissues of GAG-removed groups was lower than the control group, and this effect was much greater than that of the non-collagenous protein-removal group [17]. Moreover, Purnomo et al. used exostosin-like glycosyltranferase 2-deficient (EXTL2 knockout) mice, in which GAGs were expressed at high levels in the aorta. They subsequently induced chronic kidney disease by feeding subject mice a high-phosphate diet, and found that matrix calcification was accelerated in aortic ring and vascular smooth muscle cells. They also showed that the removal of GAGs in vascular smooth muscle cells derived from EXTL2 knockout and wild-type mice effectively suppressed calcium deposition in a high-phosphate environment [18]. Therefore, GAGs that accumulate in the growth plate of aged mice may promote the formation of calcified cartilage.

We initially believed that growth plate cartilage would be replaced by bone due to aging, even in mice. Therefore, we predicted that immunohistochemical analyses would reveal a decrease in aggrecan and type II collagen, two major components of the cartilage matrix, as well as an increase in type I collagen, a major component of bone matrix, at W55. In fact, we observed reductions in aggrecan and type II collagen expression, as predicted, but type I collagen expression did not increase. This finding is one piece of evidence that mouse growth plates are not replaced by bone. Furthermore, an SEM/EDS elemental analysis revealed that the growth plate of aged mice was calcified. Accordingly, an additional immunohistochemical analysis was performed to examine the expression of type X collagen and MMP-13. Type X collagen is a non-fibrous collagen that constitutes the hypertrophic layer of cartilage prior to calcification, and hypertrophic chondrocytes express both type X and type II collagen [19]. In this study, type X collagen was expressed in hypertrophic chondrocytes and the hypertrophic layer of the growth plate cartilage at W10, but was hardly detected in the growth plate at W55. A previous study has reported that the synthesis of type X collagen ceases with the onset of cartilage calcification, followed by a decrease in its expression, and our findings are consistent with this explanation [20]. MMP-13 is an extracellular proteinase that cleaves type II collagen and aggrecan, and is expressed in hypertrophic chondrocytes [19, 21, 22]. It has also been suggested that type X collagen is degraded by MMP-13, and a previous report showed that MMP-13-deficient mice exhibited expansion of the hypertrophic zone of the growth plate and increased type X collagen deposition [23]. Previous studies have reported that MMP-13 may contribute to chondrocyte hypertrophy and cartilage matrix calcification via the degradation of type II and type X collagens and aggrecan [19, 24]. In this study, hypertrophic chondrocytes expressed MMP-13 at W10 but not at W55. Likewise, type X collagen was expressed throughout the hypertrophic layer at W10, but was hardly expressed at W55. Since the growth plate cartilage has already calcified at W55, MMP-13 degradation of type X collagen may be complete, and the synthesis of MMP-13 by chondrocytes may have already ceased.

Growth plate senescence leads to a decrease in the chondrocyte proliferation rate in the growth plate, reduced height (thickness) of the growth plate, and eventually a cessation of chondrocyte proliferation, resulting in epiphyseal fusion [9, 25, 26]. Interestingly, a previous study of growth plate transplantation in rabbits demonstrated that the growth rate of transplanted growth plates depended on the age of the donor, not the recipient [27]. This suggests that any reduction in the growth rate of the growth plate is caused by mechanisms intrinsic to the growth plate itself, rather than by the age of the individual or by factors in the environment external to the growth plate. In contrast, growth plate senescence has been shown to be independent of age, and is regulated chiefly by time—i.e., the cumulative number of chondrocyte divisions [28]. Furthermore, the proliferative capacity of stem cell-like cells within the resting zone has been reported to be finite; thus, growth plate senescence may be caused by the depletion of their proliferative capacity [29].

Estrogen is known to play an important role in longitudinal bone growth within the growth plate and is involved in epiphyseal fusion. For example, early estrogen exposure due to precocious puberty cause premature fusion of the growth plate, and can result in short stature [30]. Estrogen is an essential hormone for normal skeletal growth in both men and women. In a male patient who was estrogen-resistant due to a mutation in an estrogen receptor gene, no fusion of the growth plate was observed, and the patient continued to grow linearly beyond puberty to reach a height of 204 cm [31]. Estrogen can also promote epiphyseal fusion; however, the mechanism by which this occurs is not yet fully understood. A previous report using a female rabbit model showed that estrogen treatment of young, ovariectomized subjects resulted in accelerated growth plate senescence, including in a decreased proliferation rate of chondrocytes, reduced height of the growth plate, and fewer chondrocytes [9]. Thus, estrogen is considered to decrease chondrocyte proliferation and cause accelerated termination of their proliferation. However, the response of chondrocytes present in the growth plate to estrogen may differ between females and males. Moreover, the effect of estrogen on epiphyseal fusion may also differ among animal species. In this study, male mice were used for investigating epiphyseal fusion. This is in contrast to female rabbits, which were used in a previous report. Therefore certain hypotheses related to the mechanisms involved in epiphyseal fusion induced by estrogen may not be directly applicable to the male mouse model. To elucidate the mechanisms responsible for epiphyseal fusion, further investigations should be considered using different sexes and animal species.

Although in this study the growth plate was closed via calcification of growth plate cartilage, epiphyseal fusion generally means that the growth plate cartilage is replaced by bone in the final phase of growth plate senescence. The cartilage matrix of the growth plate is degraded, and chondrocytes disappear during epiphyseal fusion. The cellular mechanism by which chondrocytes disappear has not yet been elucidated, but four main hypotheses have been proposed: (1) apoptosis, (2) autophagy, (3) differentiation and conversion, and (4) hypoxia. A previous review has examined these four hypotheses in detail [6, 7]. The most widely accepted hypothesis is that hypertrophic chondrocytes undergo cell death by apoptosis. However, some previous studies have suggested that autophagy is involved in the death of hypertrophic chondrocytes [32, 33]. Another study by Emons et al. examined the disappearance of hypertrophic chondrocytes in the growth plate during the process of human epiphyseal fusion and reported no evidence of apoptosis or autophagy; in contrast, the authors reported morphological signs of hypoxia and necrosis [34]. The oldest hypothesis is that hypertrophic chondrocytes differentiate into osteoblasts within the growth plate [35], but to date this hypothesis has not been established. Further studies are therefore required to clarify the mechanism by which chondrocytes disappear during epiphyseal fusion.

In a previous study, Fukuda and Matsuoka radiologically studied the process from the appearance of secondary ossification centers to their fusion in the extremities of different animals. They found that patterns of epiphyseal fusion differed among animal species, among different types of long bones in the same animal, and even between the proximal and distal ends of the same long bone. These authors indicated that the incomplete epiphyseal fusion observed in mice and rats may be unique to certain rodent species, since it is not seen in humans, monkeys, or in dogs (beagles) [12]. Mice and rats are common model organisms used in life science and medical research; therefore, it is critical to thoroughly investigate their unique epiphyseal fusion processes. Furthermore, explorations of the involvement of epiphyseal fusion in diseases such as growth retardation and overgrowth may lead to new prevention and treatment strategies.

Epiphyseal fusion is generally considered to be a consequence of replacing growth plate cartilage with bone; however, this study demonstrated that the mouse growth plate calcifies during aging and remains as calcified cartilage, not bone, for a relatively long time. As mentioned above, the mice used in our experiments showed incomplete fusion of the growth plate, whereas complete fusion is observed in humans. Considering this, further studies are required to determine whether the mouse growth plate is replaced by bone after 55 weeks of age or is maintained as calcified cartilage thereafter, as well as whether calcified cartilage also appears both during and after epiphyseal fusion in humans.