Morphological assessment of basic multicellular unit resorption parameters in dogs shows additional mechanisms of bisphosphonate effects on bone.

Bisphosphonates (BPs) slow bone loss by reducing initiation of new basic multicellular units (BMUs). Whether or not BPs simply prevent osteoclasts from initiating new BMUs that resorb bone or also reduce the amount of bone they resorb at the BMU level is not clear. The goal of this study was to determine the effects of BPs on three morphological parameters of individual BMUs, resorption depth (Rs.De), area (Rs.Ar), and width (Rs.Wi). After 1 year of treatment with vehicle (VEH), alendronate (ALN; 0.10, 0.20, or 1.00 mg/kg/day), or risedronate (RIS; 0.05, 0.10, or 0.50 mg/kg/day), resorption cavity morphology was assessed in vertebral trabecular bone of beagle dogs by histology. Animals treated with ALN or RIS at the doses representing those used to treat postmenopausal osteoporosis (0.20 and 0.10 mg/kg/day, respectively) had significantly lower Rs.Ar (-27%) and Rs.Wi (-17%), with no difference in Rs.De, compared to VEH-treated controls. Low doses of ALN and RIS did not affect any parameters, whereas higher doses resulted in similar changes to those of the clinical dose. There were no significant differences in the resorption cavity measures between RIS and ALN at any of the dose equivalents. These results highlight the importance of examining parameters beyond erosion depth for assessment of resorption parameters. Furthermore, these results suggest that in addition to the well-known effects of BPs on reducing the number of active BMUs, these drugs also reduce the activity of osteoclasts at the individual BMU level at doses at and above those used clinically for the treatment of postmenopausal osteoporosis.

Quantitative regional associations between remodeling, modeling, and osteocyte apoptosis and density in rabbit tibial midshafts.

Evidence suggests that osteocyte apoptosis is involved in the adaptive response of bone, although the specific role of osteocytes in the signaling mechanism is unknown. Here, we examined and correlated regional variability in indices of remodeling, modeling, osteocyte apoptosis, and osteocyte density in rabbit tibia midshafts. Histomorphometric analysis indicated that remodeling parameters (BMU activation frequency, osteon density, forming osteon density, and resorption cavity density) were lower in the cranial region compared to other quadrants. In addition, pericortical subregions displayed less remodeling relative to intracortical and endocortical ones. Modeling indices also demonstrated regional variability in that periosteal surfaces exhibited a greater extent of bone forming surface than endosteal ones across all anatomic quadrants. In contrast, endosteal surfaces demonstrated significantly greater surface mineral apposition rates compared to periosteal surfaces in caudal, medial, and lateral but not cranial quadrants. Using TUNEL analysis to detect osteocytes undergoing apoptosis, the density of apoptotic osteocytes was found to be lower in cranial quadrants relative to medial ones. In addition, the densities of osteocyte lacunae, empty lacunae, and total osteocytes were higher in lateral fields relative to caudal quadrants. There was a strong, statistically significant linear correlation between the remodeling indices and apoptotic osteocyte density, supporting the theory that osteocytes undergoing apoptosis produce signals that attract or direct bone remodeling. In contrast, the modeling parameters did not exhibit a correlation with apoptotic osteocytes, although there was a strong correlation between the modeling indices and the density of empty osteocyte lacunae, corroborating previous studies that have found that osteocytes inhibit bone formation. It was found that osteocyte density and osteocyte lacunar density did not significantly correlate with modeling or remodeling parameters, suggesting that cell viability should be examined in studies correlating bone turnover parameters with the functional role of osteocytes in bone adaptation.

Histomorphometric analysis of the effects of osteocyte density on osteonal morphology and remodeling.

Osteocytes, the most abundant cells in the cortical bone matrix, are thought to have mechanosensory and chemosensory regulatory roles. Marotti theorized that osteocytes signal to osteoblasts to recruit them into the osteocyte lineage. Martin extended this theory, assuming that osteocytes display a general inhibitory effect on osteoblast function. The current study provides a quantitative analysis of the relationships between osteonal osteocyte density (Ot.N/BV), wall width (W.Wi), individual osteon porosity (IOP), and formation period (FP) in ulnar cortices from sheep labeled with tetracycline and calcein double labels. We postulated that osteocytes inhibit refilling so that the osteon wall width is thin enough, and the haversian canal is large enough, to allow adequate delivery of nutrients to the osteocytes throughout the forming and completed osteon. Therefore we tested the hypotheses that Ot.N/BV correlates negatively to FP and W.Wi, and positively to IOP, and that FP correlates positively with W.Wi. We found that Ot.N/BV correlated positively with IOP (P < 0.0001) and W.Wi correlated positively with FP (P < 0.0001). Significant negative correlations were observed between Ot.N/BV and both W.Wi (P < 0.0001) and FP (P = 0.006). These data support the general hypothesis that osteocytes contribute to the regulation of osteon morphology via the control of refilling rate and formation period, and the specific hypotheses that, for a given cement line diameter, high osteocyte density (1) reduces the rate of refilling and decreases the formation period and (2) decreases wall width and increases individual osteon porosity.

A theoretical analysis of long-term bisphosphonate effects on trabecular bone volume and microdamage.

Bisphosphonates increase bone mass and reduce fracture risk, but their anti-resorptive action may lead to increases in fatigue microdamage. To investigate how bisphosphonate effects influence changes in bone volume and microdamage in the long term, a strain-adaptive model of bone remodeling and microdamage balance was developed for a continuum-level volume of postmenopausal trabecular bone by invoking Frost's mechanostat hypothesis. Both disuse and fatigue microdamage were assumed to stimulate the activation frequency of basic multicellular units (BMUs) such that bone remodeling served to remove excess bone mass and microdamage. Bisphosphonate effects were simulated as follows: low, intermediate, high, or complete suppression of BMU activation frequency either without a change in resorption by the BMU or with an independent decrease in resorption while the bone formation process was unaffected (i.e., formation initially exceeded resorption). Of the bisphosphonate effects, a reduction in resorption relative to formation dictated the long-term gain in bone volume while the potency of activation frequency suppression controlled the rate of gain. A plateau in the bone mass gain that typically occurs in clinical studies of bisphosphonate treatment was predicted by the model because the resultant reduction in strain forced bone formation by the BMU to decrease over time until it matched the reduction in BMU resorption. A greater suppression of activation frequency proportionally increased microdamage, but the accumulation was limited over the long term as long as remodeling was incompletely suppressed. The results of the model suggest creating bisphosphonates that provide minimal suppression of remodeling and a large decrease in BMU resorption because this would minimize damage accumulation and increase bone mass, respectively.

Trabecular rod thickness by direct measurement from 3D SEM anaglyphs.

This study presents a methodology for measuring the thickness of trabecular rods directly from anaglyphs. Macerated sagittal slices of T12 vertebral bodies from 15 subjects were examined by scanning electron microscopy (SEM). Two digital images (the second image tilted 5 degrees ) were recorded, and a 3D anaglyph was created. The thickness of the trabecular rods (Tb.Th((rods))), and the anatomical orientation of the trabecular rods were measured using an image analyser. Conventional 2D histomorphometry was performed on adjacent bone slices. A total of 1559 rod measurements were made from the 15 vertebral bone slices, with a mean Tb.Th((rods)) of 123 +/- 36 microm. The rod thickness in males (128 +/- 34 microm) was significantly greater than that in females (119 +/- 37 microm, P < 0.001). Tb.Th((rods)) changed significantly with age in the males: the thicker rods in the younger men reduced with age to a thickness similar to that in women. 3D measurements were significantly larger than the 2D estimates, and there was no correlation between the two methods of measurement. An inverse correlation was found between the number of rods and the bone volume fraction (BV/TV), indicating that decreased BV/TV is associated with an increased number of rods. The vertical rods (132 +/- 39 microm) were significantly thicker than the horizontal rods (116 +/- 33 microm, P < 0.001). The determination of rod numbers, and their orientation and individual thicknesses enables a greater understanding of cancellous bone architecture in both individuals and populations, and will allow more reliable finite element modelling. Direct measurements from 3D anaglyphs of intact specimens provide new data that show previously unrecognised age- and sex-related changes.

Long stemmed total knee arthroplasty with interlocking screws: a computational bone adaptation study.

The ability of an interlocking screw fixation technique to minimize bone loss related to stress shielding in the tibia was investigated and compared to the abilities of cement and press-fit fixation. Full bony ingrowth has been associated with greater stress shielding than partial ingrowth; therefore, the effect of intimate bonding of the stem to bone on subsequent bone loss was also studied. A damage- and disuse-based remodeling theory was coupled with a two-dimensional finite element model of the tibia to predict changes in bone remodeling following long stemmed total knee arthroplasty (TKA) for four different fixation techniques (cement, press-fit, interlock with bony ingrowth, and interlock without bony ingrowth). Remodeling changes commenced with the model state variables--bone area fraction, mechanical stimulus, damage, and remodeling activity--at steady-state values predicted by the intact tibia simulation. After TKA and irrespective of fixation technique, the model predicted elevated remodeling due to disuse, in which more bone was removed than replenished. In regions below the tibial tray and along the cortices, the interlocking stem with full bony ingrowth and the cemented stem caused the least amount of bone loss. An interlocking stem with a smooth, matted finish did not reduce the bone loss associated with interlocking fixation.

Tenofovir treatment at 30 mg/kg/day can inhibit cortical bone mineralization in growing rhesus monkeys (Macaca mulatta).

The acyclic nucleoside phosphonate analog, 9-[2-(R)-(phosphonomethoxy)propyl]adenine (PMPA; Tenofovir: Gilead Sciences, Inc., Foster City, CA), has been shown to effectively inhibit simian immunodeficiency virus (SIV) replication in rhesus macaques by blocking reverse transcription. However, chronic long-term tenofovir treatment at 30 mg/kg/day, intended to reduce viral replication and illness, has been shown to result in bone deformities and spontaneous fractures in rhesus monkeys. Based on these findings, we studied the effects of tenofovir treatment and pathogenic SIV infection on cortical bone remodeling in rhesus monkeys. Tibiae from tenofovir-treated or untreated, SIV-infected or uninfected, rhesus macaques were evaluated for bone microdamage and remodeling. We found that tenofovir treatment had a significant effect on osteoid (unmineralized bone) seam width in tibial cross-sections. Regardless of SIV infection status, half of the tenofovir-treated animals had significantly increased osteoid seam widths in tibial cortical bone resulting in an osteomalacia-like condition. Pathogenic SIV infection significantly increased tibial resorption cavity density. and this increase was normalized by tenofovir treatment. These results suggest that tenofovir treatment at 30 mg/kg/day inhibits mineralization of newly formed bone. SIV infection results in increased tibial resorption cavity density, while tenofovir treatment tends to minimize this increase. Both defective mineralization of newly formed bone and increased resorption cavity density may result in greater bone fragility.

Hidden assumptions in a maximum-entropy method stacking analyses.

Two models have been used to describe indefinite self-association (stacking). In the more popular isodesmic model addition of molecules to the growing stack occurs with the same equilibrium constant, while in the attenuated model successive equilibrium constants decrease in value. In an attempt to choose between the two models application was made of the maximum-entropy method. This paper points out that the conclusions drawn from this method are not proven as the application assumed specific limiting values for experimental values of a molecule in the interior of a stack, and these values are not identical in the two models for the two systems considered.

Interactions of Aluminum(III) with Phosphates.

In order to obtain information about aluminum(III)-phosphate interactions, potentiometric measurements were carried out to characterize the complex forming properties of Al(III) with organic phosphates, phosphonates, and nucleoside-5'-monophosphates. The aluminum(III)-orthophosphate system is difficult to study due to AlPO(4) precipitation. To overcome this problem, the stability constant logarithms of the 1:1 Al(III) complexes of ligands with the same donor groups (log K(1:1)) were plotted against the basicities of the ligands (log K(PO)3(H)). The resulting linear free energy relation (LFER) indicates that organic phosphates, phosphonates, and uridine-, thymidine-, and guanosine 5'-monophosphates similarly bind Al(III). Adenosine and cytidine 5'-monophosphate fall above the LFER owing to the presence of a second microform with the nucleic base protonated and a hydroxide bound to the Al(III). From the LFER the log stability constant for Al(III) binding to HPO(4)(2-) is estimated as 6.13 +/- 0.05. From the weakness of any soluble orthophosphate complexes of Al(III) we confirm the importance of citrate as the main small molecule Al(3+) binder in the blood serum. The study includes investigation of Al(III) binding to di- and triphosphates, which bind metal ion differently than monophosphates. Structures of the complexes were supported by (31)P NMR measurements.

An investigation of the interactions between lower-limb bone morphology, limb inertial properties and limb dynamics.

Bone mass and size clearly affect the safety and survival of wild animals as well as human beings, however, little is known about the interactions between bone size and movement dynamics. A modeling approach was used to investigate the hypothesis that increased bone cortical area causes increased limb moments of inertia, decreased lower-limb movement maximum velocities, and increased energy requirements to sustain submaximum lower-limb locomotion movements. Custom software and digital data of a human leg were used to simulate femur, tibia, and fibula cortical bone area increases of 0%, 22%, 50%, and 80%. Limb segment masses, center of mass locations, and moments of inertia in the sagittal plane were calculated for each bone condition. Movement simulations of unloaded running and cycling motions were performed. Linear regression analyses were used to determine the magnitude of the effect cortical area has on limb moment of inertia, velocity, and the internal work required to move the limbs at a given velocity. The thigh and shank moment of inertia increased linearly up to 1.5% and 6.9%, respectively for an 80% increase in cortical area resulting in 1.3% and 2.0% decreases in maximum unloaded cycling and running velocities, respectively, and in 3.0% and 2.9% increases in internal work for the cycling and running motions, respectively. These results support the hypothesis and though small changes in movement speed and energy demands were observed, such changes may have played an important role in animal survival as bones evolved and became less robust.

Theoretical analysis of alendronate and risedronate effects on canine vertebral remodeling and microdamage.

Bisphosphonates suppress bone remodeling activity, increase bone volume, and significantly reduce fracture risk in individuals with osteoporosis and other metabolic bone diseases. The objectives of the current study were to develop a mathematical model that simulates control and 1 year experimental results following bisphosphonate treatment (alendronate or risedronate) in the canine fourth lumbar vertebral body, validate the model by comparing simulation predictions to 3 year experimental results, and then use the model to predict potential long term effects of bisphosphonates on remodeling and microdamage accumulation. To investigate the effects of bisphosphonates on bone volume and microdamage, a mechanistic biological model was modified from previous versions to simulate remodeling in a representative volume of vertebral trabecular bone in dogs treated with various doses of alendronate or risedronate, including doses equivalent to those used for treatment of post-menopausal osteoporosis in humans. Bisphosphonates were assumed to affect remodeling by suppressing basic multicellular unit activation and reducing resorption area. Model simulation results for trabecular bone volume fraction, microdamage, and activation frequency following 1 year of bisphosphonate treatment are consistent with experimental measurements. The model predicts that trabecular bone volume initially increases rapidly with 1 year of bisphosphonate treatment, and continues to slowly rise between 1 and 3 years of treatment. The model also predicts that microdamage initially increases rapidly, 0.5-1.5-fold for alendronate or risedronate during the first year of treatment, and reaches its maximum value by 2.5 years before trending downward for all dosages. The model developed in this study suggests that increasing bone volume fraction with long term bisphosphonate treatment may sufficiently reduce strain and damage formation rate so that microdamage does not accumulate above that which is initiated in the first two years of treatment.

Ethanol and stress activate catecholamine synthesis in the adrenal: effects on bone.

Ethanol consumption and mental stress activate the sympathetic nervous system, which can adversely affect bone. We compared six groups of 10 young adult rats, three with and three without 2 h daily restraint stress. Two groups consumed food and water ad libitum, two received food and 6% (w/v) ethanol as drinking water, and two received the amount of food consumed by ethanol rats the previous day plus water ad libitum (pairfed). After 6 weeks, rats were killed. Plasma, femurs, lumbar vertebrae, and adrenals were harvested. Femoral dimensions were measured and biomechanical properties were tested by three-point bending. Plasma osteocalcin, vertebral osteocalcin mRNA levels, and adrenomedullary tyrosine hydroxylase (TH), dopamine beta-hydroxylase (DBH), and phenylethanolamine N-methyl transferase (PNMT) mRNA levels were quantified. Daily restraint decreased weight gain and femoral length compared to dietary controls, and appeared to partially preserve bone strength, especially in calorie-restricted pairfed rats. Femoral strength was significantly affected by treatment in that bones of pairfed controls were weakest, ethanol drinkers were intermediate, and ad libitum restrained were strongest. Femoral yield load, displacement, and work at yield load were negatively correlated with TH and DBH mRNA levels, but not PNMT, suggesting a negative influence of norepinephrine. Plasma osteocalcin and dry weight of lumbar 3-5 vertebrae were unaffected; however, osteocalcin mRNA in second lumbar vertebrae was positively correlated with TH, DBH, and PNMT levels. Ethanol consumption at this level had little effect on femur morphology or strength. In contrast, the data suggested possible stimulation rather than inhibition of vertebral bone formation.

Predictions on preserving bone mass in knee arthroplasty with bisphosphonates.

Using a computational model of bone adaptation, we investigated the long-term ability of bisphosphonates to minimize proximal bone loss that is associated with stress shielding in the tibia after long-stemmed total knee arthroplasty (TKA). When invoking bisphosphonate effects, the remodeling activity was suppressed, and the resorption size was reduced. Compared with the untreated simulation, bisphosphonate slowed the rate of bone loss after TKA (42% reduction in bone loss at 1 year). Activating the drug 3 months before the surgery reversed bone loss associated with the reduction in such activities as walking, but it did not provide any substantial benefit in the long-term. Late bisphosphonate treatment did not reverse the bone loss that occurred 3.5 years after TKA, although it preserved 3% of bone normally lost without treatment.

Estrogen-dependent actions of bone morphogenetic protein-7 on spine fusion in rats.

STUDY DESIGN: Intertransverse process spinal fusion using recombinant human bone morphogenetic protein-7 (rhBMP-7) was performed in intact and ovariectomized female rats. OBJECTIVES: To examine fusion rates in intact and ovariectomized female rats using rhBMP-7 to determine if spine fusion is dependent on estrogen status. SUMMARY OF BACKGROUND DATA: Rat spinal fusion has been established as a consistent, efficient model for posterolateral intertransverse process fusion. Previous experiments have confirmed the efficacy of pellets containing the carrier, insoluble collagen bone matrix (ICBM), and rhBMP-7 to augment intertransverse process single level fusion in a rat model. Studying these implications in an osteoporosis model is of clinical value because there are many patients undergoing spinal fusion surgery that have osteoporotic bone disease, and there is a steady increase in this group of patients. METHODS: A total of 15 ovariectomized and 15 intact Sprague-Dawley female rats were randomly assigned to groups receiving 25 mg ICBM alone, 25 mg ICBM + 10 microg rhBMP-7, and 25 mg ICBM + 30 microg rhBMP-7. Spinal fusion was evaluated by manual motion testing at each lumbar segment, radiographic evaluation using the Lenke grading system, and histology. RESULTS: Ovariectomized and intact rats receiving 25 mg carrier ICBM alone did not show spinal fusion. With 25 mg ICBM + 10 microg rhBMP-7, there was not a significant difference in fusion rates between intact and ovariectomized rats (P = 0.63). Ovariectomized rats receiving 25 mg ICBM + 30 microg rhBMP-7 showed significantly lower fusion rates than intact rats (P = 0.013). CONCLUSION: These data suggest that spinal fusion using rhBMP-7 is estrogen-dependent in rats. At the dosages used, rhBMP-7 was unable to overcome the inhibitory effects of estrogen deficiency on spinal fusion.

Fatigue damage, remodeling, and the minimization of skeletal weight.

The skeleton has provided many advantages during the course of vertebrate evolution, but it has also contained limitations that have strongly influenced bone biology. These limitations have included weight and the potential for fatigue failure. Calcified bone tissue is approximately twice as heavy as other tissues, so it is important to minimize the size of the skeleton, but this implies increasing bone stresses and strains and the potential for fatigue fracture. This paper first explores the role of fatigue damage removal by remodeling in extending a long bone's fatigue life to match the animal's lifetime. Next, an estimate is obtained for the amount that the cross-sectional area of a bone would have to be increased in lieu of remodeling to achieve the same extension of fatigue life, provided that the associated muscle mass remained constant. The result illustrates how remodeling can provide a gracile bone the same fatigue life as a substantially more robust bone lacking remodeling. Finally, it is shown that if muscle mass increases in linear proportion to bone mass, as experimental data suggest, extending a bone's fatigue life by increasing its cross-sectional dimensions may not be effective because the inertia of bigger bones would result in larger muscles and increased skeletal loads. Thus, bone remodeling to remove fatigue damage may be essential for the existence of relatively large, long-lived vertebrates.

Fatigue microdamage as an essential element of bone mechanics and biology.

The fossil record shows that bone remodeling has existed since the earliest large vertebrates became weight-bearing on land, but the functions of remodeling have long been debated. The principal protagonists in this debate have been those favoring a mechanical function and those asserting that remodeling serves to move calcium in and out of the skeleton. In recent years the arguments of the former school have included not only the adaptation of internal structure to specific kinds of stresses, but the need to remove fatigue damage. It has become clear that (1) physiologic strains continually produce fatigue damage in bone; (2) this damage weakens bone and is associated with both osteocyte apoptosis and the activation of remodeling; and (3) remodeling is the only means by which this damage can be removed. The significance of these observations is increased by the fact that fatigue failure is more likely in larger structures. This "volume effect," along with the advantages of enhanced mobility and metabolic efficiency, may have selected for bone remodeling as a means of controlling fatigue damage as it occurs, allowing larger vertebrates to maintain a relatively light skeleton over an extended lifetime. In this view, bone remodeling is not primarily a mechanism for calcium transport, but is intimately related to other inflammatory repair responses.

Fetal and maternal outcome after administration of tenofovir to gravid rhesus monkeys (Macaca mulatta).

Tenofovir has been shown to cross the placenta in quantities sufficient to sustain reductions in viral load in simian immunodeficiency virus (SIV)-infected fetal monkeys. With chronic exposure (30 mg/kg), however, significant bone-related toxicity has been shown in approximately 25% of infants studied. Further investigations were conducted to determine whether the bone-related toxicity observed was initiated during fetal life. Gravid rhesus monkeys (n = 4) were administered tenofovir subcutaneously once daily from 20 to 150 days of gestation (30 mg/kg; term: 165 +/- 10 days). Fetuses were monitored sonographically, and maternal and fetal blood and urine samples were collected to assess hematologic parameters, clinical chemistry, insulin-like growth factor (IGF) levels, and bone biomarkers. Fetuses were delivered by hysterotomy near term for necropsy and evaluation of bone-related mechanical properties. Results of these studies have shown 1) normal fetal development, although overall body weights and crown-rump lengths were less than those for age-matched controls (p < or = .03); 2) a significant reduction in circulating IGF-I (p <.001); 3) a small reduction in fetal bone porosity (p < or = .03); and 4) transient alterations in maternal body weights and bone-related biomarkers during the treatment period. The results of these studies suggest that chronic fetal exposure to tenofovir at the maternal dose of 30 mg/kg throughout gestation can alter select fetal parameters and transiently affect maternal bone biomarkers.

Biological effects of short-term or prolonged administration of 9-[2-(phosphonomethoxy)propyl]adenine (tenofovir) to newborn and infant rhesus macaques.

The reverse transcriptase inhibitor 9-[2-(phosphonomethoxy)propyl]adenine (PMPA; tenofovir) was previously found to offer strong prophylactic and therapeutic benefits in an infant macaque model of pediatric human immunodeficiency virus (HIV) infection. We now summarize the toxicity and safety of PMPA in these studies. When a range of PMPA doses (4 to 30 mg/kg of body weight administered subcutaneously once daily) was administered to 39 infant macaques for a short period of time (range, 1 day to 12 weeks), no adverse effects on their health or growth were observed; this included a subset of 12 animals which were monitored for more than 2 years. In contrast, daily administration of a high dose of PMPA (30 mg/kg subcutaneously) for prolonged periods of time (>8 to 21 months) to 13 animals resulted in a Fanconi-like syndrome (proximal renal tubular disorder) with glucosuria, aminoaciduria, hypophosphatemia, growth restriction, bone pathology (osteomalacia), and reduced clearance of PMPA. The adverse effects were reversible or were alleviated following either complete withdrawal of PMPA treatment or reduction of the daily regimen from 30 mg/kg to 2.5 to 10 mg/kg subcutaneously. Finally, to evaluate the safety of a prolonged low-dose treatment regimen, two newborn macaques were started on a 10-mg/kg/day subcutaneous regimen; these animals are healthy and have normal bone density and growth after 5 years of daily treatment. In conclusion, our findings suggest that chronic daily administration of a high dose of PMPA results in adverse effects on kidney and bone, while short-term administration of relatively high doses and prolonged low-dose administration are safe.