Chronic kidney disease–mineral and bone disorder (CKD–MBD) is a common complication of chronic kidney disease (CKD), with an incidence in developing countries from 33.3% to 81% of CKD patients (Seyedzadeh et al., 2022). CKD–MBD is accompanied by systemic dysfunction of mineral and bone metabolism, such as abnormalities in serum calcium (Ca), phosphate (Pi), parathyroid hormone (PTH), vitamin D, and fibroblast growth factor 23 (FGF23) metabolism, as well as in bone turnover, mineralization, and volume (TMV). CKD–MBD usually induces skeletal growth defects, fragility, and calcification resulting from extravasation, particularly vascular calcification (Elias et al., 2018; Reiss et al., 2018). Because CKD–MBD contributes to cardiovascular disease, which is the main cause of death in patients with CKD (Reiss et al., 2018; Tani et al., 2020), it is necessary to prevent and treat CKD–MBD as early as possible. However, the present treatments aiming to control serum Ca, Pi, and PTH levels are not effective, and the mortality rate among patients with CKD remains extremely high (Elias et al., 2018). Therefore, an appropriate animal model of CKD–MBD is urgently needed to explore effective strategies to prevent and treat CKD–MBD.

In recent years, 5/6 nephrectomy (5/6 Nx) (Shalhoub et al., 2012; Watanabe et al., 2018; Liu et al., 2021a), adenine with high-phosphorus diet consumption (Tani et al., 2020; Tsuboi et al., 2020), unilateral ureteral obstruction (UUO) (Nordholm et al., 2018; Jiang et al., 2021; Jin et al., 2022), and Adriamycin nephropathy (AN) (Zhang et al., 2019; Li et al., 2020) have been utilized to model CKD–MBD. However, there is a lack of systematic studies comparing these models. To identify a model that accurately mimics the characteristics of CKD–MBD observed in the clinic, we need to determine the differences in CKD–MBD-related manifestations among the models. Therefore, we induced CKD–MBD in rats fed with a low-calcium and high-phosphorus diet (0.6% Ca and 1.2% Pi) by 5/6 Nx, AN, and UUO and then assessed the renal function by measuring serum creatinine (Scr) and blood urea nitrogen (BUN) levels; determined the bone metabolism index by measuring Ca, Pi, and FGF23 content; monitored the levels of key markers associated with bone metabolism, such as intact parathyroid hormone (i-PTH), FGF23, osteocalcin (OC), and cross-linked C-telopeptide of type 1 collagen (CTX-1) in serum; examined histopathological changes in the kidney and femur by hematoxylin–eosin (HE), Masson’s trichrome (Masson), Goldner’s trichrome (Goldner), and tartrate-resistant acid phosphatase (TRAP) staining; performed micro-CT imaging of the femur; and evaluated the expression of osterix and cathepsin K by immunohistochemical (IHC) staining. This systematic study may offer some useful information for choosing an experimental animal model of CKD–MBD.

Patients with CKD experience mineral and bone metabolism disturbance, which results in an increased risk of fracture, especially for those on dialysis (D'Arrigo et al., 2022). It is commonly believed that bone turnover patterns are altered with progression of chronic kidney disease. Increasing evidence indicates that a high prevalence of low bone turnover first occurs in the early stages of CKD, with adynamic bone disease being the predominant form, and that renal function progressively deteriorates. Persistent PTH secretion may lead to secondary hyperparathyroidism, which results in osteitis fibrosa; a high-turnover bone disease resulting mainly from secondary hyperparathyroidism; and finally deterioration in bone quantity with prominent bone resorption in end-stage renal disease (Drüeke and Massy, 2016; Massy and Drueke, 2017; Liu et al., 2018; El-Husseini et al., 2022).

In the present study, we generated rat models of bone disease associated with renal insufficiency by 5/6 nephrectomy, Adriamycin injection, and UUO accompanied by consumption of a low-calcium and high-phosphorus diet to provide information for the study of CKD–MBD. The 5/6 Nx model animals, which were fed a high-phosphorus diet to induce hyperphosphatemia, suffered acute injury with creatinine clearance that was approximately equivalent to late stage 4 or 5 CKD in humans (Wanna-Udom et al., 2022). Therefore, 5/6 Nx model rats develop severe hyperphosphatemia and hyperparathyroidism after acute kidney injury, which promote the progression of CKD–MBD. Because Adriamycin interferes with membrane proteins and podocyte cytoskeletal proteins and can thus impair podocyte function and albuminuria filtration and ultimately lead to glomerular sclerosis and renal failure, the AN model is the classic model of focal segmental glomerulosclerosis (Dou et al., 2020). Moreover, the UUO model is a well-established animal model used to explore tubulointerstitial injury and progressive fibrosis (Wang et al., 2020). Complete ureteral obstruction initiates a sequence of events in the obstructed kidney, including activation of fibroblasts and EMT, leading to a massive increase in the number of myofibroblasts during the early stage, ECM accumulation in tubulo-interstitium, collagen formation, and eventually end-stage obstructive nephropathy after approximately 2 weeks (Humphreys, 2018; Martínez-Salgado et al., 2022). Thus, UUO results in renal fibrosis, especially tubular interstitial fibrosis. Therefore, the different renal damage models generated in this study are beneficial for comprehensively studying bone turnover accompanied by disturbances in mineral and bone metabolism.

In CKD diagnosis, renal function is always evaluated by measuring the levels of Scr and BUN in serum, and then the estimated glomerular filtration rate (eGFR) is calculated based on the level of Scr. Furthermore, the stages of CKD are classified according to the eGFR level (Wan et al., 2022; Zhang et al., 2022). With the progression of CKD, the levels of Scr and BUN in serum increase and the eGFR decreases. As a complication of CKD, CKD–MBD often silently begins in the early stages of CKD and is usually diagnosed in G3b–G4 stages of CKD (Elias et al., 2018; Ketteler et al., 2018). Therefore, CKD–MBD patients always have extremely high levels of Scr and BUN in serum. In this study, renal injury in the different models was evaluated by measuring Scr and BUN levels in serum. The data showed that 5/6 Nx and AN led to severe renal injury at the fifth week after the first operation; furthermore, renal failure progressed within the experimental period, especially in the 5/6 Nx model rats. Comparatively, Scr and BUN levels increased slowly in the UUO group; a reasonable explanation for this is that the normal contralateral kidney has a compensatory effect on renal excretion function, which is consistent with a previous report (Martinez-Klimova et al., 2019). Similarly, the difference in the results of kidney histopathology can be explained when whole or focal damage is taken into full consideration. Thus, the aforementioned results indicate that all the models caused renal injury; the 5/6 Nx model rats exhibited obvious damage, while the UUO model rats exhibited slight damage.

According to the CKD–MBD guidelines (Ketteler et al., 2017; GBD Chronic Kidney Disease Collaboration, 2020), the levels of Ca, Pi, and i-PTH in serum should be monitored in adult patients at the beginning of stage 3 CKD, and then the severity of CKD–MBD should be assessed based on the level of i-PTH. Extremely high levels of i-PTH indicate high-turnover bone disease (Laowalert et al., 2020) and vice versa. In this study, accumulation of Pi, exhaustion of Ca, and excessive secretion of i-PTH in serum were observed in the 5/6 Nx model rats at the eighth week of the experiment, and the variation trend was consistent with that observed in CKD–MBD patients; elevation of serum FGF23 levels, which corresponded with increases in Pi and i-PTH levels, was also observed in the 5/6 Nx model rats. Therefore, it is clear that 5/6 Nx induced high-turnover bone disease in rats under our experimental conditions. However, UUO model rats exhibited less marked changes in the levels of Pi, i-PTH, and FGF23 during the experiment. Because both serum i-PTH and FGF23 levels can be used to predict the occurrence of high-turnover bone disease, but not low-turnover bone disease (Massy and Drueke, 2017), it is difficult to distinguish the types of bone turnover in UUO model rats according to the serum levels of Pi, i-PTH, and FGF23. A previous study reported that serum sclerostin is a good predictor of low bone turnover and low bone volume (Massy and Drueke, 2017). Unfortunately, the serum levels of sclerostin do not correlate with changes in bone mineral density in patients with osteoporosis and are of limited diagnostic value in this disease (Figurek et al., 2020; Dincel and Jørgensen, 2022; Vavanikunnel et al., 2022) because sclerostin acts locally in the bone microenvironment, and the serum level of sclerostin does not reflect changes in sclerostin expression by osteocytes (Dreyer et al., 2021). Furthermore, it has been reported that romosozumab, a humanized anti-sclerostin antibody, is associated with an increased incidence of cardiovascular events, limiting the diagnostic application of serum sclerostin. Therefore, serum sclerostin was not used as a monitoring index in the present study, although sclerostin may aggravate PTH resistance in CKD (Canalis, 2018). In the AN model rats, accumulation of Pi, exhaustion of Ca, and excessive secretion of FGF23 in serum were observed at the eighth week of the experiment; however, excessive elevation of i-PTH levels was not observed, making it difficult to distinguish the different forms of bone turnover. However, the levels of OC and CTX-1 in serum were abnormally increased in the 5/6 Nx model rats and the AN models rats, which confirmed that the 5/6 Nx model rats and the AN model rats suffered from high-turnover bone disease. On the other hand, the levels of OC and CTX-1 in the UUO model rats showed no obvious change, so it was difficult to determine the forms of bone turnover induced by UUO according to the serum indices only.

In addition to hyperphosphatemia, hypocalcemia, and secondary hyperparathyroidism, active proliferation of osteoblasts and osteoclasts is typical of high-turnover bone disease and leads to excessive bone formation and resorption and eventually bone destruction because of the lack of Ca and overload of Pi (Baptista et al., 2020; Ferreira et al., 2021; Jorgensen et al., 2022). Therefore, bone destruction is not commonly observed in the early stage of CKD–MBD and progresses in the later stages of CKD–MBD. Bone biopsy is the gold standard for diagnosing CKD–MBD; however, it is not strongly recommended because of the invasiveness of the procedure and the difficulty of clinical implementation (Chiang, 2017; Salam et al., 2021). In this study, histopathological analysis of the femur showed that all the experimental methods caused some degree of bone damage, such as a reduction in the number of trabeculae, a decrease in bone mineralization, and overactivation of osteoclasts. In particular, osteoclasts were overactive in the 5/6 Nx model rats, whereas bone mineralization was decreased and trabeculae were degraded in the AN model rats and UUO model rats, meaning that the 5/6 Nx model rats all had high-turnover bone disease and that the AN model rats and UUO model rats had later-stage CKD–MBD. Micro-CT images and measurement of femur parameters, such as BMD, BV/TV, Tb.Th, and Tb.N, all demonstrated that AN resulted in the most severe femur destruction and that 5/6 Nx caused minimal damage relative to the other two techniques, which is consistent with the histopathological analysis of the femur; these results indicate that the 5/6 Nx model rats experienced high bone turnover accompanied by increased bone formation and resorption and that the AN model rats and UUO model rats had later-stage CKD–MBD and suffered severe bone destruction. The expression of osterix and cathepsin K was assessed by IHC, and the results verified that bone formation and resorption were increased in the 5/6 Nx model rats and that bone resorption was increased in the AN model rats and UUO model rats, which is consistent with the histopathology results and micro-CT images.

In UUO model rats, the serum levels of creatinine, Pi, i-PTH, FGF23, OC, and CTX-1 were similar to those in the control group; however, the histopathological analyses micro-CT imaging and IHC all showed that UUO model rats exhibited increased bone resorption. This inconsistency might be attributed to the compensatory effect of the normal contralateral kidney on renal excretion function, which makes it difficult to determine the metabolic status of bone according to serum biochemical indices. According to a previous report, obstructive nephropathy usually occurs within approximately 2 weeks in end-stage CKD (Humphreys, 2018; Martínez-Salgado et al., 2022). Therefore, it is also possible that we fail to detect an increase in serum creatinine, Pi, i-PTH, FGF23, OC, and CTX-1 levels due to the timing of our blood draws. In brief, the UUO model suffered drastic damage with subtle biochemical changes and should be used carefully in the study of CKD–MBD, and bone histomorphometry should be the primary method for determining the form of bone turnover, with serum biochemical indices being less informative.

AN resulted in rapid progression of renal injury during the early period of the experiment; eventually, obvious bone destruction was observed at the end of this experiment. Unfortunately, we did not observe the transformation of bone formation and absorption in the femur, and it is possible that we did not collect femur samples at the most appropriate time. However, utilizing AN to model bone disease and study the course of progressive CKD in rats and humans may provide insight into both normal and abnormal physiology if the timeline of experiments is selected prudently as AN can rapidly induce severe kidney injury. It is worth noting that we observed focal swelling of hepatocytes, multifocal focal necrosis, inflammatory cell infiltration around the portal area, and fibrous tissue hyperplasia in the AN model rats, suggesting that Adriamycin may lead to liver injury (Supplementary Figure S2); thus, the scope of the application of the model should be evaluated.

5/6 Nx induced remarkable variations in Ca, Pi, and i-PTH levels in serum and progression of renal injury. It also led to high turnover of bone formation and absorption according to the levels of OC and CTX-1. These changes were consistent with the findings of histopathological staining and micro-CT imaging. Thus, the 5/6 Nx model can be used to easily investigate the progression of high-turnover bone disease on this experimental schedule.

Overall, 5/6 Nx, AN, and UUO accompanied by a low-calcium and high-phosphorus diet successfully induced CKD–MBD in rats. The AN and UUO models exhibited deterioration in bone quantity with rapid and severe bone resorption at the end of the experiment; however, the changes in the biochemical indices were subtle in the UUO model, and liver injury was obvious in the AN model. The 5/6 Nx model presented high-turnover bone disease at the end of the experiment, with consistency between serum biochemical indices and femur histomorphometric parameters.

In conclusion, 5/6 Nx, AN, and UUO accompanied by a low-calcium and high-phosphorus diet can result in CKD–MBD. On our experimental schedule, the 5/6 Nx model exhibited high-turnover bone disease, which is easy to evaluate according to serum biochemical indices and histomorphometry. The AN and UUO models presented severe bone resorption and deterioration in bone quantity at the end of the experiment, according to histomorphometric analysis of the femur.