| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1Institute of Physiology and Biochemistry of Nutrition, Federal Research Centre for Nutrition and Food, Kiel; 2Department of Osteopathology, University Medical Center Hamburg-Eppendorf; 3Medical Physics Research Group, Department of Diagnostic Radiology and 4Department of Orthopaedics, University Hospital Schleswig-Holstein, Campus Kiel; and 5Department of Oral and Maxillofacial Surgery, Kiel University, Kiel, Germany
Submitted 20 November 2006 ; accepted in final form 24 April 2007
 |
ABSTRACT |
|---|
 TOP ABSTRACT MATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
bone mineral density; histomorphometry; bone mechanical property; calcium balance; bone marker
25.2 billion in 2002. Most of the total costs per inhabitant were required in the group of subjects more than 65 years old (5). The total medical care expenditures associated with arthritis and other rheumatic conditions in the United States in 1997 amounted to $186.9 billion (68), whereas the actual costs for musculoskeletal disease in the United States were estimated to add up to $215 billion a year, with the prospect of an increase in the future (27). Because of its anti-inflammatory and immunosuppressive action, glucocorticosteroids (GC) are used for therapy of various symptoms and diseases, including rheumatoid arthritis and related diseases, asthma bronchiale, neurodermitis, and, with high doses, after organ transplantation (24, 29). One of the side effects of GC treatment is an accelerated bone loss and an increased risk for bone fracture, especially after a prolonged exposure (2, 8). GC treatment is the third leading cause of osteoporosis, following the loss of sex steroids and old age (67). GC-treated patients are at a twofold higher risk of suffering from a fracture compared with controls, irrespective of their bone mineral density (BMD; see Ref. 66). Limited information is available on its pathophysiology, since the clinical picture of glucocorticosteroid-induced osteoporosis (GIO) mostly reflects a combined effect of the underlying systemic disease and of the secondary effects induced by GC treatment.
To develop strategies for the treatment of GIO, the mechanisms of how GC act on bone have to be known. In GIO, bone loss predominantly affects the trabecular bone, and thus the risk for bone loss is greater at the spine than the hip (8). From regulatory and metabolic effects, it is assumed that the development of GIO, and with it the degree of bone loss, is biphasic. The key action of GC is thought to be the long-term depression of osteoblast number and activity and thus bone formation (8, 67). The initiation step, however, is suggested to be a stimulation of bone resorption as indicated by enhanced RANK-ligand expression and reduced osteoprotegerin production in the presence of GC (8). Both substances are osteoblast-derived cytokines that modulate osteoclastogenesis.
Observations on effects of different GC made in cell culture should be verified in animal models before, finally, applications are tested in clinical trials in human subjects. Animal models, however, have to be tested for suitability. Beyond the testing of new pharmaceuticals in small animal models like the rat or mouse, the U.S. Food and Drug Administration requires that they have to be tested in a large-animal model before a step III clinical study can be performed (38).
In the present experiment, we investigated the effect of GC on calcium metabolism and bone development in adult primiparous Göttingen minipigs. Bone loss, assessed by quantitative computed tomography, changes of markers of bone formation, and resorption and mineral balances were analyzed in a cross-sectional and longitudinal design. Bone morphometry, chemical analyses of bone, and bone mechanical property were analyzed cross-sectionally. The hypothesis was that GC induces the loss of BMD with similar physiological signs as in humans and thus makes the adult minipig a suitable large animal model for research in GC-induced osteoporosis and its treatment.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Animals. We used adult female primiparous Göttingen minipigs of the Institute's own breeding herd because earlier experiments in domestic pigs have shown that osteoporosis induced by ovariectomy was more pronounced in adult primiparous than in growing nulliparous animals (53). At the age of delivery, the sows were 20 mo old. Because pregnancy and lactation may have effects on bone mineral content (BMC) and turnover (4), the lactation productivity was standardized to a litter size of five to six piglets and to a nursing time of six weeks.
Experimental groups, time schedule. The experiment was initiated with 50 primiparous sows. Before intervention, blood, urine, and fecal samples were collected, and BMD measurements were performed. Next, the sows were assigned to the following three groups: animals on GC treatment for 8 mo (GC-short, n = 30), animals on GC treatment for 15 mo (GC-l, n = 10), and animals without treatment for 15 mo (control, n = 10). To longitudinally describe the short-term effect of GC on bone markers in plasma and urine, repeated samples were taken, and samples of all animals on GC treatment (GC-short and GC-l) were pooled (GC-s-all, n = 40) during the first 8 mo. After 8 mo, 10 animals from the GC-short group were killed to assess the short-term effects on bone samples (GC-s), whereas the remaining 20 animals switched to another experiment with therapeutic treatments. The results of the therapeutic treatment have been published in a separate paper (16). To describe the long-term effect of GC, animals of the control group and of GC-l continued for another 7 mo before they were killed.
Diets, housing, intervention. Before starting the intervention, the sows switched from the Institute's standard nutrition regimen for lactating minipigs to the Institute's semipurified diets for adult minipigs for an adaptation time of 4 wk (preexperimental). This diet has been described previously (52). In brief, it consisted of the following components (g/kg diet): 290 corn starch, 240 sucrose, 150 casein, 80 cellulose, 75 lard, 75 margarine, 60 mineral and vitamin premix, and 30 lactulose. The mineral and vitamin premixture provided all nutrients necessary for minipigs in sufficient amounts but were otherwise adapted to an average human diet (rich in fat and protein, poor in fiber). The diet contained 14 MJ metabolizable energy/kg; 6 g/kg calcium and 6.6 g/kg phosphorus from CaCO3 and Na4(H2PO4)2, respectively; and 6,500 IU of vitamin D3. The content of vitamin D3 (Deutsche Vilomix Tierernährung, Neuenkirchen-Vörden, Germany) was higher than in standard pig diets and higher than the recommendations for humans, since this was a long-term experiment on semipurified diets with no other vitamin D sources. Moreover, the pigs were kept indoors since birth with artificial light, i.e., almost no vitamin D production in the skin from precursors.
At an average age of 30 mo, the sows were allocated to the experimental groups on the basis of their body weights, as described above. Feed intake was controlled individually and was restricted to 370 g/day to allow slow, steady growth as it is typical for this age, while avoiding fattening. Feed was completely consumed by all animals. Deionized water was available ad libitum. Animals were housed individually for the whole time. The sows were kept on strawless floor cages.
The glucocorticosteroid prednisolone (Lichtenstein Pharmazeutika, Mühlheim-Kärlich, Germany) was given orally at a dose of 1.0 mg·kg body wt–1·day–1 for 8 wk and thereafter at a dose of 0.5 mg·kg body wt–1·day–1 until the end of the experiment. The tolerability was tested previously in a pilot set of three sows. At the end of the experiment, all animals were killed by exsanguination after anesthesia with a combination of dormicum (Midazolam, 1 mg/kg body wt; Synopharm, Barsbüttel, Germany) and ketamine hydrochloride (5 mg/kg body wt; CuraMed Pharma, Karlsruhe, Germany), and representative bone specimens were taken and stored at –18°C until further procedures and analyses.
BMD in vivo. BMD was assessed in vivo by quantitative computed tomography (QCT) before intervention (t0), after 8 mo (t8), and after 15 mo (t15) with a Siemens Sensation 16 Scanner. Animals were anesthetized with a combination of dormicum and ketamine hydrochloride (see above), and an accessory medication with atropine. At the beginning, three animals of the GC-treated animals died after the second QCT measurement (t8) as a consequence of prolonged anesthesia, which was sometimes necessary if the measurements had to be interrupted because of constrains within the clinic routine. For the remaining QCT measurements, the anesthesia was completed by injections with prednisolone (5 ml of a solution of 10 mg/ml prednisolone-21-acetate) for animals on GC treatment. No more animals died. Trabecular BMD of a central 10-mm-thick slice and an elliptical region in the anterior part of lumbar vertebrae L1–L3 were evaluated. The Siemens OsteoPackage was used to calculate calibrated BMD results (25). Reproducibility of the measurement was assessed by duplicate measurements in four minipigs with interim repositioning. The root-mean-square average of precision errors (15) was 0.83% (3.4 mg/cm3).
Plasma parameters. Blood samples were taken before intervention (t0) and after 2, 5, 7, 10, 13, and 15 mo (t2–t15) after an overnight fast, always at the same time between 08:00 and 10:00 A.M., and stored in ice-cooled water until centrifugation and aliquotting. Calcium was analyzed directly by atomic absorption spectroscopy (PerkinElmer 1100) using an air/acetylene-flame at 2,300°C after dilution with a 0.5% lanthanium chloride solution. Five standards were used for calibration. Phosphorous was estimated in EDTA plasma as inorganic phosphorous with an automatic analyzer (Cobas Bio; Hoffmann La Roche, Basel, Switzerland) using a test kit (Roche Diagnostics, Mannheim, Germany). Parathyroid hormone (PTH) was determined in EDTA plasma using an RIA (Immundiagnostic, Bensheim, Germany). We selected the PTH kit containing a goat antiserum directed against the COOH-terminal end of PTH-(53–84), with a donkey-anti-goat immunoglobulin as precipitation reagent. This kit was the most sensitive to physiological changes like pregnancy and lactation in the minipigs (unpublished observations) compared with the midregional or NH2-terminal kit, or the kit containing antibodies against intact human PTH. This kit showed significantly higher values for ovariectomized pigs compared with intact controls (53). Values for sensitivity, and intra- and interassay variability were 3.4 pg/l, 9.2% coefficient of variation (CV) and 12.5% CV, respectively. 25-Hydroxyvitamin D3 [25(OH)D3] was measured in EDTA plasma by RIA (Immundiagnostik) after acetonitrile precipitation. Values for sensitivity and intra- and interassay variability were 2.5 nmol/l, 9.8% CV, and 14% CV, respectively. 1,25-Dihydroxyvitamin D3 [1,25(OH)2D3] was measured in EDTA plasma by a radioreceptor assay (RRA; Immundiagnostik). For this test, the plasma was cleaned, and the biologically effective vitamin D metabolite was extracted by Extrelut columns followed by a separation on silica columns, where the 1,25-(OH)2D3 metabolite was separated from other vitamin D metabolites. For the quantitative determination of 1,25-(OH)2D3 by RRA, [3H]1,25-(OH)2D3 tracer was used, which competed with the sample for a limited number of 1,25-(OH)2D3 receptors. Values for intra- and interassay variability were 12% CV and 17% CV, respectively. Osteoprotegerin (OPG) was analyzed in serum by sandwich enzyme immunoassay (Immundiagnostik). For technical reasons, only a limited number of samples could be analyzed at baseline and after 8 mo, i.e., in the short-term samples. Total alkaline phosphatase (TAP) and bone alkaline phosphatase (BAP) were analyzed in heparin-plasma directly and as the difference after lectin precipitation, respectively. The kit (Roche Diagnostics) was based on a kinetic color reaction. Values for intra- and interassay variability were 1.6% CV and 2.5% CV, respectively. In our validation studies, we observed that diets with low calcium content raised BAP and those with high calcium content depressed it in ovariectomized minipigs (54).
Urinary parameters. Urine was collected over 24 h before intervention and after 2, 5, 7, 10, 13, and 15 mo and frozen at –18°C until analysis. For technical reasons, urine samples of only 40 animals were collected at t2 (5–6/group). Calcium in urine was measured directly after dilution with a lanthanum chloride solution by atomic absorption spectrometry (PerkinElmer 1100) using an air-acetylene flame at 2,300°C. For lysylpyridinoline, denoted as deoxypyridinoline (DPD), the urine samples were centrifuged, lyophilized, and hydrolyzed with 6 M HCl at 110°C for 24 h and centrifuged again. Each hydrolysate (1 ml) was added to a mixture of 1 ml glacial acetic acid, 2 ml n-butan-1-ol, and 5 ml of a 10% CF-1-slurry (fibrous cellulose powder; Whatman, Maidstone, UK). A column was prepared by adding the mixture of hydrolysate and CF-1-slurry to a polypropylene econocolumn (Bio-Rad, München, Germany). The resin was washed, the pyridinium-containing eluate was eluted, and the lyophilized eluate was redissolved and analyzed by HPLC. The intra- and interassay variability was 2% CV and 4.8% CV, respectively. The hydroxylysyl- and lysylpyridinolines were quantified by external standards with established linearity gained from a commercially available adult bovine bone gelatin (Deutsche Gelatine-Fabriken Stoess, Eberbach/Baden, Germany). They were purified by a preparative reverse-phase HPLC column, and the degree of purity was verified by amino acid analysis (>98% dry wt) according to a method previously described (1). In this paper, only lysylpyridinoline results are reported, since DPD is specific for bone. All parameters in plasma and urine were analyzed in a blind fashion and in duplicate.
Metabolic mineral balance. Four metabolic calcium and phosphorous balances were performed: before the intervention with glucocorticosteroids (i.e., the beginning of the experiment, t0) and after 5 (t5) 10 (t10), and 15 mo (t15). For this, the minipigs were transferred to metabolic cages that allowed the separate collection of feces and urine for 7 days. All samples were collected daily and stored at –18°C until analysis. The calcium and phosphorus absorption and retention was calculated as follows: absorption (mg/7 days) = intake (mg/7 days) – fecal excretion (mg/7 days); retention (mg/7 days) = intake (mg/7 days) – [fecal excretion (mg/7 days) + urinary excretion (mg/7 days)]. Analyses were done in duplicate in a blind fashion.
BMC. BMC (ash, calcium, and phosphorous) was analyzed in vitro in the last thoracic vertebra (5 x 50 mm cylindrical specimens from the center of the vertebrae that were stored at –18°C until analysis) and in the same crushed cylindrical sample of the fourth lumbar vertebra that had been previously used for mechanical testing (see below). Bone specimens contained exclusively trabecular bone and were dried for 4 h at 105°C to determine dry matter (%) by weighing (mg bone after drying/mg bone before drying x 100). Thereafter, the dried bone was ashed in a muffle furnace at 450°C overnight. The ash was weighed (ash content) and dissolved in 20% (vol/vol) hydrochloric acid. Calcium was analyzed by atomic absorption spectroscopy using an air/acetylene flame at 2,300°C (PerkinElmer 1100) after a 1:1,000 dilution with 0.5% lanthanium chloride solution. Inorganic phosphorous was estimated with an automatic analyzer (Cobas Bio; Hoffmann La Roche) using a test kit (Roche Diagnostics). Analyses were done blinded and in duplicate.
Histomorphometry. The second lumbar vertebrae were removed, cleaned of adherent soft tissue, stored in formalin until preparation, and embedded and analyzed using standard methodology. Histomorphometric variables included the following: bone volume/tissue volume, trabecular width, trabecular separation, osteoid volume/tissue volume, osteoid surface/bone surface, osteoblast surface/bone surface, eroded surface/bone surface, osteoclast surface/bone surface, number of osteoclasts/bone surface, and number of osteoclasts/tissue area.
Bone strength. Bone strength was measured as ultimate stress and calculated as failure load divided by the cross-sectional area of the cylinders at the point of failure. Compression testing on the fourth lumbar vertebral body was performed on thawed cylindrical cancellous bone specimens of 7.5 mm diameter and 10 mm length from the center of the vertebral bodies using a core drill and a parallel saw. The bone cylinders were tested by uniaxial compression according to Linde et al. (33). In an electromechanical testing machine, load was measured with a 2 kN load cell (Sensotec, Columbus, OH) and deformation by a static strain gauge extensometer (MTS System, Berlin, Germany) attached to the upper and lower platens. Specimens were conditioned by compression to 0.6% strain at a strain rate of 0.01/s for 10 cycles. Results were calculated as the average of five consecutive cycles. After keeping the specimens at 10°C in isotonic saline for 1 day, the nondestructive testing was repeated, followed immediately by a destructive test. Ultimate strain was also measured at the point of failure. As a measure of elasticity, Young's modulus was calculated as slope of the regression line of the linear portion of the stress-strain curve. Reproducibility was calculated as root-mean-square average of repeated measurements.
Statistics. ANOVA was performed to gain mean and SE of the various diagnostic parameters [of their absolute values for cross-sectional analysis and of their changes with time (delta values) for longitudinal observations]. The statistical package Statgraphics 4.1 (Manugistics, Rockville, MD) was used. Student's t-test was applied for completers to test for significance between groups, and the paired t-test was used to test longitudinal data within groups. The level of significance was defined as follows: tendency with P < 0.1, significance with P < 0.05.
|
RESULTS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
|
|
0.05). The decline in GC-treated animals was more pronounced with –14.96 ± 7.29, –20.51 ± 12.24, and –25.32 ± 12.22 pmol/l at t10, t13 and t15, respectively, but failed to reach significance because of higher variation. The difference between groups was not significant. TAP in plasma was not significantly affected by GC treatment at any time (data not shown). BAP in plasma increased slightly in the control group until 5 mo after the beginning of the intervention and then declined until t13 (P < 0.05 vs. baseline, Fig. 2, bottom). In animals treated with GC, the BAP activity declined from the beginning. This change became significant at t2, t8, and t10. The difference between the control group and GC-s attained significance after 2 and 8 mo. This difference was significant (P < 0.001) when all GC-treated animals were pooled (GC-s-all; Table 2). A nonsignificant depression of the osteoblast-derived osteoclast inhibitory factor OPG was seen after 8 mo in GC-treated animals (Table 2). There was no specific effect of GC treatment on pyridinoline (PD) excretion at any time. We observed slightly decreased urinary DPD in the short term for animals in all groups. Longitudinally, this decline from baseline was significant after 8 mo, when all animals treated with GC (GC-s-all) were pooled (Table 2). Cross-sectionally, no difference was observed between GC-treated animals and control animals.
|
Phosphorus intake (15.54 g/7 days) was the same in each group and animal at any time. Phosphorus absorption (Table 2) and retention decreased nonsignificantly in the short term independent of the treatment. At t15, phosphorus retention was positive in the control group (294 ± 1,158 mg/7 days) but still negative in the GS-l animals (–2,312 ± 2,677 mg/7 days). None of the changes longitudinally or differences cross-sectionally reached significance, except for the change of urinary phosphorus excretion from baseline, which significantly increased in the control group but significantly decreased in animals treated with GC (Table 2).
BMC. The wet weight of the fourth lumbar vertebra specimen (mg) did not differ between groups [595 ± 17 (control), 578 ± 17 (GC-s), and 540 ± 26 (GC-l)]. Compared with the control group, the bone dry weight (mg) tended to be lower in the GC-s group and was significantly lower in the GC-l group (Table 3). The same was true for the content of dry matter (%; Fig. 3, top, P < 0.05). The mineral content (mg ash/bone specimen) tended to be lower in sows after the short-term treatment and more pronounced after the long-term treatment compared with the control group (Table 3). The ash content (mg/g bone wet wt) was slightly lower in sows after 8 mo of GC treatment (–2.7%, P < 0.6) and more pronounced in sows after 15 mo of GC treatment (–7.8%, P < 0.11) compared with the control group. Similar trends were observed for calcium and phosphorus.
|
|
Histomorphometry. Compared with the control group, bone volume/tissue volume was slightly lower in sows after 8 mo of GC treatment but was 15% lower in sows after 15 mo (P < 0.1, Table 3). There was a trend toward higher osteoid volume after 8 mo of GC treatment (borderline significant, P < 0.08). Trabecular separation after 15 mo was significantly higher than after 8 mo (P < 0.05). Trabecular width tended to be lower in animals after 8 and 15 mo of GC treatment compared with controls (P = 0.2; Fig. 3).
Bone strength. Ultimate stress (Fig. 3, bottom) and Young's modulus (data not shown) were slightly reduced after short-term GC treatment. The alteration after long-term intervention was more pronounced (P < 0.09) in the GC group compared with the control group.
|
DISCUSSION |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Animal model. The rat seems to have limitations as model for GIO, since there was no bone loss after intravenous GC administration (28). There are few studies on GC-induced bone loss in large animals, including dogs (35, 40, 49) and sheep (9, 32, 55). The data published on sheep showed detrimental effects of GC only in animals that had additional causes for bone loss, like ovariectomy or calcium/vitamin D deficiency. Moreover, 0.25 mg prednisolone/kg body wt daily was not sufficient on its own to induce loss of BMD or bone strength (11). In fact, these parameters were slightly higher (NS) compared with the untreated group. A significant reduction of histomorphometric variables of bone formation was observed after 3 mo of treatment but without significant changes in markers of bone formation (9).
We used the adult primiparous Göttingen minipig. This large animal model is an omnivore and thus has more similarities with humans than dogs or sheep. Earlier experiments in domestic pigs have shown that effects of ovariectomy-induced estrogen deficiency were more pronounced in adult primiparous than in young nulliparous sows (53), and ovariectomy-induced loss of BMC was seen in adult primiparous minipigs aged 2.5–3 yr (54). Our animals were 30 mo old at the beginning of the experiment. It is difficult to extrapolate this age to humans. With respect to fertility, which starts between 6 and 7 mo, the pigs have multiplied their age about four to five times. In humans, this would come to about 50 yr. If the age is related to epiphyseal closure, which occurs in minipigs at 2–3 yr (unpublished observation), this is comparable to 20-yr-old human subjects (34). Thus we used the model of after epiphyseal closure, which is closer to the target group of adult humans that is mainly affected by osteopathies than studies using animals in the rapid-growing stage (23).
The time of intervention of 15 mo corresponds to 6–7% of the life expectancy of a minipig (up to 20 yr; see Ref. 20). In humans, this would correspond to an intervention time of 
5 yr, which is rather long and allows us the ability to draw conclusions on long-term effects.
Information on GC doses applied to pigs or minipigs are rarely reported. Tolerable doses of 1 mg·kg–1·day–1 orally for 3 mo after three bolus doses of 250 mg on the day of operation were observed in minipigs after organ transplantation (3). In piglets, 0.5 mg/day dexamethasone, equivalent to 3–4 mg prednisolone, was used (65). In humans, maximum doses of 1–3 g/day are given after organ transplantation over a few days (24). Typical initial doses were reported of 20–80 mg/day, chronic doses of 1–10 mg/day for rheumatoid arthritis, and 10–50 mg/day for others (64). The effect of GC on bone is dose dependent; 1–7.5 mg/day, corresponding to up to 0.1 mg/kg body wt prednisolone equivalent, is expected to result in no or small side effects (24), whereas higher doses (60 mg/day) induced osteoporosis (58).
We used prednisolone given orally at a dose of 1.0 mg·kg body wt–1·day–1 for 8 wk, which corresponds to 70–90 mg/day in humans, and 0.5 mg·kg body wt–1·day–1 thereafter. Because the set of pilot sows did not show any signs of intolerability or complications, the dose was not changed during the study.
Body weight. The increase in body weight observed in the short-term and in the long-term interval is also observed in humans, especially after long-term therapy. Weight gain is correlated with the cumulative dose (61). In young minipigs, no effect on body weight was reported, probably because they were still in the rapid growing stage that superimposes on the effect of GC (23).
BMD in vivo.
For practical reasons, the minipig is easy to handle for BMD measurements. The size allowed us to use a standard QCT in the university hospital. The mean BMD was 450 mg/cm3 and thus much higher than that of adult premenopausal women (133 mg/cm3; see Ref. 46). BMD was depressed 45–50 mg/cm3, or –10% from baseline, in the short term. In the long term, only a slight further decline of 1% from t8 values was observed. Bone loss following GC treatment is thus biphasic like in humans (8, 24, 64) with values of 12% during the first few months and 2–3% annually thereafter (36). Bone loss occurs in cortical and cancellous bone, but predominantly in the axial skeleton (36, 45). It is assumed that >5 mg prednisone/day induces bone loss and in 30–50% of patients fractures will occur (63) with strong correlations between cumulative dose and loss of BMD and between daily dose and fracture risk (63). In minipigs after 0.5 mg/kg body wt of 5 days/wk prednisolone subcutaneously for 26 wk, BMD of the femur, but not the lumbar vertebra, was lower compared with control animals (23). In contrast to the sows we used, these pigs were much younger, still growing, and nulliparous. Moreover, the duration was shorter than the short-term observation in our experiment, which may explain their lack of effect in the lumbar spine.
Parameters in plasma and urine. GIO is the consequence of several systemic and local effects on different organs, namely gut, kidney, gonads, parathyroid, and bone (14). It is assumed that GC reduce intestinal calcium absorption, increase urinary calcium excretion, induce hyperparathyroidism, depress gonadal hormone secretion and suppress osteoblast number and activity directly, and cause secondary hypogonadism (14), but the results were not consistent.
PTH. We observed no significant effect of chronic treatment with GC on PTH. This is in contrast to an anticipated hyperparathyroidism after GC treatment in patients (58) but is in agreement with results in healthy men (17) and in men who were treated with 50 mg prednisolone daily for 3.7 mo because of problems with fertility not related to systemic disease (45), and with reports by others (39, 56, 59). Pearce et al. (45) concluded that, rather than causing bone loss, PTH may be suppressed as a consequence of lost bone because of illness and immobility. Recently, it was shown that GC reduced the PTH tonic secretory rate and increased the fractional pulsatile PTH secretion, although the overall PTH concentration was similar among normal and GC-treated subjects (6). This indicated that a significant secondary hyperparathyroidism was not involved.
25(OH)D3.
In our experiment, we observed a significant increase in plasma 25(OH)D3 concentration in all groups within the short term. This was because of the vitamin D content of our semipurified diet (6,500 IU/kg), which was higher compared with our standard pig chow (2,000 IU/kg) the animals consumed before the study. Basal values between 57 and 75 nmol/l in the minipigs are higher than concentrations in young men with very low vitamin D intakes of <200 IU/day, who had mean values of 32 nmol/l before intervention, but were comparable with values of these men after they had been supplemented with 1,800 IU/day. This amount refers to a total of 25 IU vitamin D·kg body wt–1·day–1 compared with the 60–70 IU vitamin D·kg body wt–1·day–1 in the minipigs. In young growing minipigs aged 5 mo, we observed plasma 25(OH)D3 concentrations of 340 nmol/l when animals were on semisynthetic diets containing 8 g calcium/kg diet and 6,500 IU/kg diet vitamin D, whereas their levels were only 40 nmol/l when they consumed the same diet but without vitamin D (51). In intact and ovariectomized adult domestic pigs on conventional feed for breeding sows, we observed 25(OH)D3 concentrations between 180 and 250 nmol/l in multiparous sows and between 180 and 210 nmol /l in nulliparous sows (53).
In the long term, GC treatment tended to transiently slower the raise in plasma 25(OH)D3 compared with the control animals. We speculate that this indicates a lower absorption rate of vitamin D from the gut after GC treatment (12) and/or its reduced hydroxylation in the liver. This observation is in accordance with reports by others (30, 56) showing that GC excess resulted in small but significant decreases of plasma 25(OH)D3 concentrations. A third possibility to explain a lowered bone mass after GC treatment was recently discussed by Pascussi et al. (43) and Zhou et al. (69). They suggest a mechanism via reduced plasma levels of 25(OH)D3 or 1,25(OH)D3 as a consequence of increased catabolism. This is mediated by GC-induced activation of pregnane X receptor (PXR), which is involved in the catabolism and detoxification of xenobiotics and drugs and which is able to cross-talk with the vitamin D-responsive gene that activates the respective hydroxylase that catalyses the 24-hydroxylation of 25(OH)D3 and 1,25(OH)2D3 down to water-soluble and biologically inactive calcitroic acids (42). As a consequence, vitamin D deficiency, lower calcium absorption, and finally bone disease will occur (22). At present, there is still debate whether the hydroxylase responsible for 25(OH)D3 or 1,25(OH)2D3 catabolism activated by PXR is CYP24 (43, 44) or CYP3A4 (69).
To overcome the impeding effects of increased PXR production on vitamin D metabolism following the treatment with GC or other drugs, it is recommended to maintain 25(OH)D3 at levels of 30–50 ng/ml (22). Whether the 25–30% lower levels of 25(OH)D3 in GC-treated animals compared with controls had a marked impact on the loss of BMD is difficult to rate, because the absolute values were still relatively high with 187 ± 43 nmol/l compared with 255 ± 40 nmol/l in controls. These values were lower than those we observed earlier in growing minipigs on a semisynthetic diet with an adequate supply of calcium and vitamin D with 262–342 nmol/l but were much higher than those of vitamin D-depleted growing minipigs with 31–40 nmol/l (51).
1,25(OH)2D3. The effects of GC on plasma 1,25(OH)2D3 are contradictory. Increases (21, 39, 62), decreases (62), and no changes (47, 56) have been reported. From our results, we conclude that plasma 1,25(OH)2D3 seems to play a minor role in the development of GIO, since its plasma level after 8 mo in the GC-treated group was not different from baseline and there were no significant differences between the groups. This effect is in accord with observations in healthy men (17). In the long term, we observed a trend for lower values in the GC-treated animals compared with control animals (e.g., 54.9 ± 7.5 vs. 62.8 ± 14.2 pg/ml after 13 mo). Both values were within the range that we have found earlier in intact domestic pigs of 40 and 100 pg/ml (53) but were significantly higher than values that we have seen before in vitamin D-depleted growing minipigs of 32 ± 6.2 pg/ml (51). The trend for lower 1,25(OH)2D3 concentrations in GC-treated minipigs vs. control animals may be explained by a possibility recently published by Zhou et al. (69). They proposed that activation of a steroid and xenobiotic receptor (PXR or SXR) by GC stimulates the PXR-mediated expression of CYP3A4, the enzyme that degrades 1,25(OH)2D3 by 24-hydroxylation. This mechanism could also contribute to the reduction of the intestinal calcium absorption that occurred in GC-treated animals (Table 2).
BAP.
The most striking physiological effect of GC treatment seen in this study is in lowering BAP activity, which is in agreement with studies in men who were treated with 50 mg prednisolone daily for 4 mo. Their BAP decreased significantly by 24.2 ± 8.6%, and their serum osteocalcin decreased strongly by 28.5 ± 15.5% (P = 0.08; see Ref. 57). The men were treated with GC because of fertility problems and not because of a systemic disease. Similar indications of a reduced bone formation were reported for healthy adult volunteers on a short-term treatment with GC (21), for asthmatics with long-term treatment with doses of 
1.5 mg/day (13), for postmenopausal women on acute high-dose prednisone (60 mg/day), for growing minipigs (23), and for adult minipigs on a diet with a high compared with a low calcium content as an expression of reduced bone turnover (54).
DPD. Ikeda et al. (23) reported significantly reduced urinary type 1 collagen N-telopeptides after 13 wk. We did not see such an indication of depressed bone resorption after chronic administration of GC to healthy animals. This is in agreement with reports in chronically GC-treated young men for reasons of infertility, without a chronic disease (45), in euthyroid patients with Graves' ophthalmopathy treated with GC for one year (47) and in asthmatics with long-term treatment (13). GC treatment may evoke different responses in the growing compared with the mature skeleton. However, in contrast to postmenopausal osteoporosis where bone resorption prevails over bone formation and bone formation is almost unaltered, stimulated bone resorption contributes only slightly to the bone mineral loss in GC-induced osteoporosis, but bone formation is dramatically reduced (8).
Metabolic mineral balance. It is assumed that the bone loss observed after GC treatment has its origin in part in a lower calcium retention and that the reason is a decreased renal hydroxylation of 25(OH)D3, a reduced expression of the duodenal but not the renal calcium-binding protein CaBP-9k (31), an increased catabolism of 25(OH)D3 and 1,25(OH)2D3 (43, 69), and/or a vitamin D-independent defect in transmucosal calcium transport (7, 30). Harrison and Harrison (19) reported on an inhibitory effect on intestinal calcium absorption in rats where dietary cortisol impaired the active transport and passive diffusion.
In some cases, GC treatment increased the urinary excretion of calcium because of a reduced renal tubular calcium reabsorption that is directly mediated by the drug (7, 8).
After acute administration of high-dose GC (40 mg of prednisone) to postmenopausal women, serum calcium and serum phosphate were unchanged (59) like in our minipigs. In contrast to others (39, 58, 59), we did not observe significant changes in urinary calcium. This divergent finding may be explained by the fact that we measured 7-day urinary calcium excretion in healthy adult animals in a long-term intervention with sufficient calcium in the diet. The impact of GC on renal calcium excretion may be different in postmenopausal women (59), in patients that are treated with GC because of a variety of different diseases (58), or in spot urine samples of fasted subjects (39). Moreover, there was no untreated control group and the observation was limited to 5 days (59), or both groups consisted of patients of several different diseases (58). Our results are in accordance with reports by Hahn et al. (18) in which renal calcium excretion of GC-treated patients was not different from normal subjects.
The persistent decline in calcium retention after GC treatment was a result of persistently higher fecal excretion, i.e., lower absorption. Thus the metabolic calcium balance reflected very well the decrease of BMD seen in vivo (Fig. 1), the tendency for a lower BMC (Table 3), and the borderline lower fracture resistance in vitro (Fig. 3).
Our observation is in accordance with a 12% reduction of calcium absorption in normal volunteers within 2 wk of GC treatment (57), and with patients receiving pharmacological doses of prednisone (15–100 mg/day), when calcium absorption correlated inversely with the dose of GC (30). The authors suggested that acute calcium malabsorption in GC-treated patients was the result of an abnormality of vitamin D metabolism because the administration of a physiological dose of synthetic 1,25(OH)2D3 stimulated calcium absorption in all patients (30). Since we did not see large effects on plasma vitamin D metabolites, nor on PTH, the chronic application of GC may deteriorate the vitamin D-independent transmucosal absorption of calcium, whereas the change in vitamin D metabolites that was associated with the reduced calcium absorption in patients may be the result of the underlying disease.
There were no strong effects of GC on phosphorus retention. The trend for a lower phosphorus absorption in minipigs receiving GC was compensated by the significant reduction of urinary phosphorus.
BMC, histomorphometry, bone stability. The positive association between cumulative dose of GC and loss of BMC was indicated by a trend for lower contents of ash, calcium, and phosphorus. The significantly lower dry matter in the lumbar vertebra that paralleled the development of BMD demonstrates a GC-induced change of bone composition with consequences on mechanical properties. Our results are in agreement with lower ultimate load and maximum absorption energy in the lumbar vertebra of growing minipigs with GC treatment (23). The reason for the bone loss might be a reduced GC-associated bone formation in the presence of an unaltered ongoing bone resorption. It was proposed that the inhibition of osteoblastogenesis and a higher rate of apoptosis of osteoblasts and osteocytes are of primary importance (8, 41, 67).
Consequently, a depressed OPG production could be expected. We observed only slightly more depressed plasma OPG concentrations in GC-treated animals. The biphasic course of GC-induced bone loss may be caused by an initial borderline increase in osteoid formation (P = 0.059), a short-term but maintained trend for lower trabecular width, and a significant rise of trabecular separation in the second phase of the study. The resulting trend for ongoing lower trabecular bone volume was in agreement with the direction of histomorphometric parameters of aged women undergoing GC treatment, who had slightly lower bone volume/tissue volume but significantly lower trabecular thickness (50).
Moreover, we could show in this set of minipigs that vascular endothelial growth factor (VEGF) was significantly decreased in the lumbar vertebrae of GC-treated minipigs (48). This observation supports the hypothesis that an impairment of the VEGF-modulated vascularization of the bone remodeling unit by GC contributes to the short-term decrease in osteoblast activity, and thus to the significant decline in plasma BAP (Fig. 2, bottom) and to the trend for a lower number of alkaline phosphatase-positive chondrocytes in articular knee joint cartilage (10).
The controversy over the role of urinary calcium loss, bone resorption, vitamin D metabolites, and hyperparathyroidism to GIO in human studies, and the absence of their contribution to GIO in our study, may be explained by the underlying disease that is treated in patients rather than GC themselves (8, 47), whereas our study was performed in healthy animals.
In summary, the aim was to investigate if the minipig is a suitable animal model for GIO. The main parameter was BMD in vivo as measured by QCT. We have shown a significant and persistent loss of BMD, which was biphasic like in humans. Our study has a limitation in its number of animals compared with some cross-sectional studies in humans. The number of animals, however, was at the limit as to what could be handled with the different long-term samplings in our laboratory. According to the higher variability of other parameters, we have to report that the study just failed to have enough statistical power to ensure the significance of several parameters. However, by looking at the overall picture, i.e., the parallel decline of calcium balance and BMD and the parallel lower bone ash content, we conclude that GIO is the consequence of reduced mineral absorption at the luminal site of the gut. We propose that this may be due more to a direct effect and only partly due to a vitamin D-dependent effect. At the skeletal site, GIO depresses vascularization of bone and bone-forming activity, as indicated by depressed VEGF and BAP in the presence of ongoing bone resorption. This leads to lower tissue mineralization, trabecular area, and ultimately bone strength. The main outcomes are comparable to symptoms of GC-induced osteoporosis in human subjects. Thus the adult Göttingen miniature pig is a suitable animal model for GC-induced osteoporosis.
|
GRANTS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
Present address of B. Stampa and W. Timm: Synarc A/S, Hamburg, Germany.
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
