BoneKEy-Osteovision | Commentary
Glucocorticoid osteoporosis - relations to BMU theory and to bone cell birth and death
DOI:10.1138/2001026
Dalle Carbonare et al. () recently reported interesting differences in iliac cancellous bone remodeling and microstructure between glucocorticoid (GC) and postmenopausal (PM) osteoporosis. In accordance with my conviction that the most important role for bone histomorphometry is as a tool of in vivo bone cell biology (), I will examine the differences in relation to a conceptual model of bone remodeling and consider to what extent they can be accounted for by the effects of GC on bone cell recruitment and lifespan.
By relating their findings to the total GC dose they were able to estimate the order in which the changes occurred. In the low dose group, in whom the (unstated) duration of treatment was probably from 0.5 to about 2 years, osteoclast number and eroded surface were higher, adjusted apposition rate and bone formation rate were lower, and the only structural difference was a reduction in wall thickness. All of these findings have previously been reported, with differing severity and consistency (), and inhibited bone formation was confirmed in two animal models (). In the high dose group, in whom the probable duration of treatment was from about two to more than ten years, the indices of bone remodeling were not significantly more abnormal, but there was a further reduction in wall thickness, a significant reduction in trabecular thickness, as observed by others (), and confirmation of microstructural differences indicating removal of more trabeculae and poorer connectivity ().
Bone remodeling is executed, not by isolated osteoclasts and osteoblasts, but by unique temporary anatomic structures known as basic multicellular units (BMU; 12). The individuality of the BMU is less clear in cancellous than in cortical bone and it is rarely captured in randomly oriented histologic sections, but its existence rests on compelling evidence (). Each BMU has a target-a region of effete bone that for some reason needs to be replaced (), originates close to the nearest blood vessel () and excavates a surface trench toward and for some distance beyond its target (). Cortical BMU are directed to their target by osteocyte apoptosis (); whether the same holds for cancellous bone is not known. Progression of the BMU across the surface depends on the continued arrival of new pre-osteoclasts from the circulation (). The depth of the trench depends on the timing of osteoclast apoptosis (), and the extent to which the trench is refilled depends on the number of new osteoblasts at each cross sectional location (), the proportion that persist as osteocytes or lining cells, and the timing of apoptosis in the remainder ().
The increase in surface histologic indices of bone resorption was interpreted by the authors (and will be similarly interpreted by the great majority of readers) as indicating an increased rate of bone resorption, but this cannot be correct. From the mean values for BFR/BS and Tb.Th it is possible to estimate mean values for BFR/BV of 24.6 and 10.5%/y in the PM & GC groups respectively. Assuming a mean value for BV/TV at age 50 of 24.5% () the mean value of 16.4% at age 66 in the PM group represents an exponential loss of 2.5%/y and a mean value for BRs.R/BV of 27.1%/y. If this value was increased in the GC group in proportion to the increase in surface resorption indices it would be 54.2%/y, and the net rate of bone loss in this group would be 43.7%/y. Such a rate of loss could not be sustained for more than a few years without virtually complete disappearance of the bone, which evidently did not occur. In fact, with similar calculations, BRs.R/BV in the GC group was 15.1%/y, substantially lower, not higher, than in the PM group. The increase in eroded surface is the result of a long delay between completion of resorption and onset of formation in each BMU (). The increase in osteoclast number despite the reduction in the rate of resorption could reflect smaller cells with fewer nuclei, or indicate that individual osteoclasts were working more slowly than normal, so that they took longer to resorb the same amount of bone, as was also demonstrated many years ago in chronic renal failure (), a condition of high rather than low bone turnover.
From the mean values for BFR/BS and wall thickness, it is possible to estimate mean values for activation frequency (Ac.f) of 0.45 and 0.21 events/y in the PM and GC groups respectively. Ac.f is the best index of the overall intensity of bone remodeling, and represents the probability that a new cycle of remodeling will begin at any point on the bone surfaces (). Such cycles are oriented perpendicular to the direction of advance of the BMU, and their frequency depends not only on how often new BMU's are originated but also on the average distance traveled by each BMU across the surface (). Consequently, in the GC group there were either fewer new BMU's, or a shorter distance traveled by the average BMU, or some combination of these changes. It is possible to define a minimum rate of bone remodeling consistent with the preservation of bone health, based on the need for one new BMU for the repair of each new region of microdamage, and the mean distance from the site of origination to the location of the microdamage. Normally there is a wide margin of safety (), so that even a 50% reduction would not necessarily lead to microdamage accumulation, provided the reduction was mainly in the distance traveled beyond the target rather than in the number of new BMU's. Consequently the very low remodeling rate induced by GC would not necessarily make an independent contribution to increased bone fragility.
The adjusted apposition rate (Aj.AR, not Aj.AR/BS as in table 1) is, in the absence of osteomalacia, the most satisfactory histomorphometric index of the mean rate of matrix apposition by teams of osteoblasts (). The teams are recruited sequentially as the BMU advances across the bone surface (), and they partly refill the trench in a direction perpendicular to the surface. The rate of apposition is most rapid at the onset and progressively falls as the osteoid surface moves upward to approach its final location; what is measured as Aj.AR represents the mean value averaged throughout the lifespan of the osteoblast team (). Although the mean rate will be influenced by the rate of matrix synthesis by individual osteoblasts, the main source of variation in Aj.AR is the number of osteoblasts currently present per unit of osteoid surface ().
At any time during the lifespan of an osteoid seam the number of current osteoblasts depends both on the number initially assembled on the cement surface, and on the number that have so far escaped entrapment as osteocytes or death by apoptosis. As will subsequently be further discussed, the low Aj.AR in the GC group probably reflects contributions from both mechanisms. Whatever the explanation, the low Aj.AR is the reason for the increase in osteoid surface despite the reduction in bone formation rate (), simply because the lifespan of the osteoid seam is prolonged (). The low osteoblast number also accounts for the reduction in wall thickness, which is an index of the total volume of bone matrix synthesized by an osteoblast team during its lifespan (). The extent to which wall thickness falls short of erosion depth is the main determinant of the rate of bone loss in a state of low bone turnover, as is found during long term GC administration. Even with fewer cycles of remodeling, a significant loss after each transaction leads to progressively thinner trabeculae, not only with long term GC administration (), but also after intestinal bypass surgery ().
There are several aspects of the three-dimensional organization of the BMU that are inaccessible to measurements on standard two-dimensional histologic sections. Nevertheless, the BMU model provides an instructive means of relating such measurements to the underlying cell biology (). Since BMU progression depends on continued recruitment of new circulating preosteoclasts, a GC induced reduction in the number of osteoclast precursors formed in the bone marrow () provides a logical explanation for the inferred reduction in the mean distance traveled by each BMU. Whether reduced pre-osteoclast supply would decrease the rate of longitudinal advance as well as the extent of progression is unclear. A reduced supply is not inconsistent with the observed increase in osteoclast prevalence (). This has been an inconstant finding () and depends on the opposing effects of decreased precursor birth () and impaired cell function (). In vitro effects of GC on various indices of bone resorption have been conflicting () but in at least one such study there was inhibition (). Based on indirect biochemical and radiokinetic methods, there have been many reports that “bone resorption” is increased in GC treated patients (), but this is true only during the first 6 or12 months, when osteoclast production may be increased by effects of GC on osteoprotegerin and its ligand (), and osteoclast apoptosis is postponed (), contributing to the most rapid bone loss. In some patients secondary hyperparathyroidism increases bone turnover and expands the remodeling space (), but it does not usually persist ().
As the BMU progresses across the surface, from two to three new teams of osteoblasts are recruited daily, depending on the rate of advance (). New osteoblasts assemble only in a narrow zone corresponding to the end of the reversal phase and cannot join older teams that have already begun to make bone (). Consequently, the GC induced reduction in the number of osteoblast precursors formed in the bone marrow () would reduce the number of osteoblasts in each new team, with consequent reductions in both Aj.AR and wall thickness. Unlike osteoclasts, all of which eventually undergo the same fate of apoptosis (), some osteoblasts (at least temporarily) escape apoptosis and become osteocytes or lining cells. Consequently the GC induced increase in the prevalence of cells undergoing apoptosis () could reflect the earlier occurrence of apoptosis in the same number of cells, or apoptosis in a larger number of cells, of which some would otherwise have become osteocytes or lining cells. In either case, the average matrix synthesizing lifespan of the cells would be shortened, contributing further to reductions in Aj.AR (because the mean osteoblast density during the osteoid seam lifespan would further decline), and in wall thickness (because the aggregate total amount of matrix made by the osteoblast team would further decline). The effect of GC to promote apoptosis in osteocytes as well as in osteoblasts () could account for the unexpectedly rapid increase in fracture susceptibility (), but this was not addressed by the histomorphometric findings being discussed (). These findings can to a large extent be accounted for by reduced birth rate of osteoclasts and osteoblasts, and earlier death of osteoblasts ().
The changes in cancellous bone architecture found in patients after prolonged GC administration () are less easy to relate to the disorders of bone cell function that have been identified in such patients. The architectural changes are the result of complete removal of some structural elements, beginning with either perforation of a plate () or transection of a rod (). In PM osteoporosis, perforation and transection are the result of increased depth of resorption (), most likely due to delayed osteoclast apoptosis (). This occurs mainly during the first few years after menopause and is followed by slow thinning of the structural elements that remain (). The situation in GC osteoporosis is paradoxical. In the early stages bone turnover is high and osteoclast apoptosis delayed () but the architectural changes are no worse than in PM osteoporosis (). In the later stages, when the cellular abnormalities that cause perforation or transection have subsided, the architectural changes have progressed! One possible explanation is that because the structural elements are thinner, they are more susceptible to perforation or transection even though resorption depth is no longer increased (). A more interesting, but entirely speculative, possibility is that because of increased osteocyte death there is loss of directional control of the BMU, so that it wanders further from the bone surface ().
This work is licensed under a
Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 United States License.