A.M. Parfitt

Targeted and nontargeted bone remodeling: relationship to basic multicellular unit origination and progression

Most discussions of bone remodeling focus on how it is influenced by hormones, cytokines, growth factors, and other molecules, and pay no attention to why it occurs. All physiologic processes are aimed at something, however, and it is impossible to understand how they are regulated without some criteria of success or failure. In this mini-review, I examine the purposes and instruments of bone remodeling and how they are related.

Static and dynamic bone histomorphometry in children with osteogenesis imperfecta

Osteogenesis imperfecta (OI) is a genetic disorder characterized by increased bone fragility and low bone mass. Four clinical types are commonly distinguished. Schematically, type I is the mildest phenotype, type II is usually lethal, type III is the most severe form compatible with postnatal survival, and type IV is moderately severe. Although mutations affecting collagen type I are responsible for the disease in most patients, the mechanisms by which the genetic defects cause abnormal bone development have not been well characterized. Therefore, we evaluated quantitative static and dynamic histomorphometric parameters in tetracycline-labeled iliac bone biopsies from 70 children, aged 1.5 to 13.5 years, with OI types I (n = 32), III (n = 11), and IV (n = 27). Results were compared with those of 27 age-matched controls without metabolic bone disease. Biopsy core width, cortical width, and cancellous bone volume were clearly decreased in all OI types. Decreased cancellous bone volume was due to a 41%-57% reduction in trabecular number and a 15%-27% lower trabecular thickness. Regression analyses revealed that trabecular number did not vary with age in either controls or OI patients, indicating that no trabecular loss occurred. The annual increase in trabecular thickness was 5.8 microm in controls and 3.6 microm in type I OI, whereas no trabecular thickening was evident in type III and IV OI. Wall thickness, which reflects the amount of bone formed during a remodeling cycle, was decreased by 14% in a subgroup of 17 type I OI patients, but was not determined in the other OI types. The remodeling balance was less positive in type I OI than in controls, and probably close to zero in types III and IV. Surface-based parameters of bone remodeling were increased in all OI types, indicating increased recruitment of remodeling units. No defect in matrix mineralization was found. In conclusion, there was evidence of defects in all three mechanisms, which normally lead to an increase in bone mass during childhood; that is, modeling of external bone size and shape, production of secondary trabeculae by endochondral ossification, and thickening of secondary trabeculae by remodeling. Thus, OI might be regarded as a disease in which a single genetic defect in the osteoblast interferes with multiple mechanisms that normally ensure adaptation of the skeleton to the increasing mechanical needs during growth.

Normative data for iliac bone histomorphometry in growing children.

Many insights into normal and pathologic bone development can only be gained by bone histomorphometry. However, the use of this technique in pediatrics has so far been hampered by the lack of reference data. Therefore, we obtained transfixing iliac bone samples from 58 individuals between 1.5 and 22.9 years of age (25 male; tetracycline labeling performed in 48 subjects), who underwent surgery for reasons independent of abnormalities in bone development and metabolism. The results of histomorphometric analyses of cancellous parameters and cortical width are presented as means and standard deviations, as well as medians and ranges in five age groups. In addition, the original data are available from the authors. There were significant age-dependent increases in both cortical width and cancellous bone volume, the latter being due to an increase in trabecular thickness. Osteoid thickness did not vary significantly with age. Bone surface-based indicators of bone formation showed an age-dependent decline, reflecting similar changes in activation frequency. Mineral apposition rate decreased continuously with age. Parameters of bone resorption did not vary significantly between age groups. Paired biopsies from adjacent sites, obtained in eight subjects, were used to examine the reproducibility of histomorphometric parameters in children. The lowest coefficients of variation (<10%) were found for structural measures, as well as mineral apposition rate and wall thickness. The highest variability was found for cellular parameters. The availability of reference material will greatly facilitate the use of histomorphometry in pediatrics.

Structural and Cellular Changes During Bone Growth in Healthy Children

Normal postnatal bone growth is essential for the health of adults as well as children but has never been studied histologically in human subjects. Accordingly, we analyzed iliac bone histomorphometric data from 58 healthy white subjects, aged 1.5-23 years, 33 females and 25 males, of whom 48 had undergone double tetracycline labeling. The results were compared with similar data from 109 healthy white women, aged 20-76 years, including both young adult reference ranges and regressions on age. There was a significant increase with age in core width, with corresponding increases in both cortical width and cancellous width. In cancellous bone there were increases in bone volume and trabecular thickness, but not trabecular number, wall thickness, interstitial thickness, and inferred erosion depth. Mineral apposition rates declined on the periosteal envelope and on all subdivisions of the endosteal envelope. Because of the concomitant increase in wall thickness, active osteoblast lifespan increased substantially. Bone formation rate was almost eight times higher on the outer than on the inner periosteum, and more than four times higher on the inner than on the outer endocortical surface. On the cancellous surface, bone formation rate and activation frequency declined in accordance with a fifth order polynomial that matched previously published biochemical indices of bone turnover. The analysis suggested the following conclusions: (1) Between 2 and 20 years the ilium grows in width by periosteal apposition (3.8 mm) and endocortical resorption (3.2 mm) on the outer cortex, and net periosteal resorption (0.4 mm) and net endocortical formation (1.0 mm) on the inner cortex. (2) Cortical width increases from 0.52 mm at age 2 years to 1.14 mm by age 20 years. To attain adult values there must be further endocortical apposition of 0.25 mm by age 30 years, at a time when cancellous bone mass is declining. (3) Lateral modeling drift of the outer cortex enlarges the marrow cavity; the new trabeculae filling this space arise from unresorbed cortical bone and represent cortical cancelization; (4) Lateral modeling drift of the inner cortex encroaches on the marrow cavity; some trabeculae are incorporated into the expanding cortex by compaction. (5) The net addition of 37 microm of new bone on each side of a trabecular plate results from a <5% difference between wall thickness and erosion depth and between bone formation and bone resorption rates; these small differences on the same surface are characteristic of bone remodeling. (6) Because the amount of bone added by each cycle of remodeling is so small, the rate of bone remodeling during growth must be high to accomplish the necessary trabecular hypertrophy.

Misconceptions (2): turnover is always higher in cancellous than in cortical bone

Since it became generally recognized that bone is subject to turnover, it has often been asserted without qualification that cancellous bone has higher turnover than cortical bone. For example: “Much is often made of differences in cortical and metabolically more active cancellous bone . . .”; “. . . trabecular bone . . . is eight times more metabolically active than cortical bone”; “. . . trabecular bone is metabolically more active than cortical bone, with an annual turnover of approximately 20% to 30% for the former and 3% to 10% for the latter”; and “the cancellous bone, which represents about 20% of the skeletal mass, makes up 80% of the turnover, while the cortex, which represents 80% of the bone, makes up only 20% of the turnover.” Herein, I clarify exactly what such a belief implies, trace its origins, define the circumstances in which it is true, and specify the necessary qualifications, which are invariably omitted. But first I must explain precisely what is meant by “bone turnover.” The language of science frequently takes words in common use and redefines them for a different purpose, but the meaning of “turnover” in bone physiology is exactly the same as its everyday meaning, which is replacement. The personnel manager of a business refers to the turnover of employees and the owner of a store to the turnover of items of merchandise. In both cases, the rate of turnover is expressed as a proportion of the total (usually a decimal fraction or a percentage) per unit of time, which could be a year in the former case or a week in the latter. The reciprocal of turnover is always a period of time. For example, if the turnover of checkout clerks is 50%/year, their average length of employment will be 2 years, although their individual lengths may vary over a wide range. In everyday use, turnover is most often of people or things. Bone is a continuous material rather than an assemblage of discrete entities, but otherwise its turnover is defined in exactly the same way—it refers to proportional volume replacement per unit time, usually expressed as percent/year. As before, the reciprocal of turnover is a period of time; if bone turnover is 10%/year, then the mean lifetime of each moiety of bone is 10 years, although their individual lifetimes may vary over a wide range. People are usually replaced in response to resignation, and goods in response to sales, but replacement of bone is a more complex process that occurs in the adult skeleton by remodeling, of which the instrument is the basic multicellular unit, or BMU. Why remodeling occurs at all, a neglected but important question, does not bear on the present topic and has been discussed extensively elsewhere. Most BMUs originate on one of the three subdivisions of the endosteal envelope—cancellous, endocortical, or intracortical. Periosteal bone cell activity in the adult should probably be classified as remodeling rather than as modeling (at least in women), although the evidence remains incomplete. Remodeling on the endosteal envelope is imperfectly efficient after the age of about 40 years, because each transaction replaces slightly less bone than was removed, but within the limits of experimental error, measurement of the tetracycline-based bone formation rate is a reasonable starting point for the histological estimation of bone turnover. This measurement is usually expressed per unit of bone surface (BFR/BS), which must be multiplied by the geometrical quantity surface-to-volume ratio (BS/BV) to obtain a measurement of volume replacement (BFR/BV, expressed as percent/year). The concept of surface-to-volume ratio is commonplace in the physics of such processes as heat transfer, diffusion, and evaporation, but its application to bone raises an infrequently recognized problem that derives from its complex structure— how to decide which volume should be assigned to a particular surface. Rules have been formulated for defining where cancellous bone ends at its junction with the cortex, so that measurement of its surface-to-volume ratio requires only appropriate attention to stereology; however, the situation in cortical bone is less straightforward, because it has three surfaces—periosteal, intracortical (or haversian), and endocortical. For some purposes, it is useful to combine the endocortical and cancellous surfaces because both are in contact with marrow. The bone beneath the endocortical surface forms part of the transitional zone destined for eventual transformation to cancellous bone, and one possibility is to assign this bone to a depth of 75 m (approximately one-half the mean trabecular thickness) to the endocortical surface. Similar calculations have been carried out for the periosteal surface of the ilium using a depth of 50 m, based on the distribution of cement lines. For the present purpose, it seems reasonable to calculate also a single surface-to-volume ratio for the entire cortex including all three surfaces, with a corresponding single value for bone turnover. In whatever way surface-to-volume ratio in bone is defined and measured, it is obvious from simple inspection that this quantity is greater in cancellous than in cortical bone, even if precisely how much greater is not known. The wide appreciation Address for correspondence and reprints: Dr. A. Michael Parfitt, Division of Endocrinology and Center for Osteoporosis and Metabolic Bone Diseases, University of Arkansas for Medical Sciences, 4301 West Markham, Slot 587, Little Rock, AR 72205-7199. E-mail: parfittlwynm@uams.edu Bone Vol. 30, No. 6 June 2002:807–809

Effects of vertebral bone fragility and bone formation rate on the mineralization levels of cancellous bone from white females

Back-scattered electron microscopy was used to study mineralization levels of human iliac cancellous bone of white females (N = 49). Mineralization levels were assessed by converting bone pixel grayscale levels to atomic number (Z) using known calibration standards. The data set consisted of bone biopsies from normal and vertebral fracture subjects that had either high or low values for bone formation rate (BFR(s)) within their respective groups (fracture/low BFR(s), N = 12; fracture/high BFR(s), N = 10; normal/low BFR(s), N = 12; normal/high BFR(s), N = 15). The following three measures of mineralization were quantitatively determined for each specimen: an overall mean mineralization (Z(mean)), the mineralization of trabecular packets deep within the interior of trabeculae (Z(deep)), and the mineralization of superficial exterior packets (Z(superficial)). Two-way analysis of variance revealed that the high BFR(s) group had a significantly lower Z(superficial) than the low BFR(s) group [mean (SD) 10.383 (0.270) vs. 10.563 (0.289)], and there was no significant interaction. BFR(s) had no effect on Z(mean) or Z(deep). For the pooled data, Z(deep) was significantly higher than Z(superficial) [10.866 (0.242) vs. 10.471 (0.291)]. There was no significant difference in Z(mean), Z(deep), or Z(superficial) between normals and those with vertebral fracture, but the standard deviations of the mineralization measures in the fracture group were at least double that of the normal group. Frequency histograms show that the two groups have fundamentally different mineralization distributions. The normal group demonstrates typical Gaussian distributions centered around the mean, and the distributions of the fracture group are bimodal, with peaks occurring at either the high or low tails of the distributions of the normal group. We hypothesize that both low and high patterns of mineralization might detrimentally affect bone material properties, with low mineralization levels causing reduced stiffness and strength and high mineralization resulting in reduced fracture toughness. The degree to which the mineralization differences may affect strength and stiffness of individual elements is estimated. The higher standard deviations of mineralization measures in the fracture group may reflect an inability to properly regulate trabecular level stress and strain. Forward stepwise regression analysis showed significant relationships between Ob.S/OS and both Z(superficial) and Z(mean), suggesting that the osteoblast may play an important role in regulating mineralization.

Age and distance from the surface but not menopause reduce osteocyte density in human cancellous bone

Previous studies of osteocyte density in human cancellous bone have relied mainly on autopsy samples and have demonstrated an age-related decline in men, but there are insufficient data in women. Using previously obtained transiliac bone biopsies from 94 healthy white women, aged 20-73 years, 38 premenopausal and 56 postmenopausal, we measured osteocytes and lacunae in ten randomly selected areas using 5-microm-thick sections stained with Goldner trichrome. For each subject, the number of osteocytes (Ot.N/B.Ar), empty lacunae (EL.N/B.Ar), and total lacunae (Tt.L.N/B.Ar) per bone area, and the proportion of occupied lacunae (Ot.N/Tt.L.N), were calculated. In 92 cases the measurements were made separately in superficial bone (<25 microm from the surface) and in deep bone (>45 microm from the surface). Mean values and differences between extreme values (DEV) for each variable were computed from the ten measured areas. In addition, confocal microscopic examination was performed on 100 microm sections. We found that Ot.N/B.Ar, Tt.L.N/B.Ar, and Ot.N/Tt.L.N decreased, but EL.N/B.Ar increased significantly with age (p < 0.001). The rates of decline were most rapid initially, falling exponentially with increasing age; the linear regressions for all four variables were significant in premenopausal, but not postmenopausal, women. At all ages, there were significantly more osteocytes in superficial than in deep bone; there was no significant decline with age in superficial bone, but a steeper exponential decline in deep bone than in whole trabeculae. DEV did not change with age for any variable. Confocal images revealed that the morphology of the osteocyte network was heterogeneous in different regions and trabeculae. The trabeculae with lower osteocyte density contained acellular areas, especially in interstitial bone. We conclude: (1) osteocyte density declines with age in women as it does in men; (2) the decline occurs exclusively in deep bone, not in superficial bone, suggesting that it is the age of the bone rather than the age of the subject that is important; (3) the rate of age-related decline falls exponentially with age and is not significant in postmenopausal women alone; (4) except for the differences between superficial and deep bone, the pattern of osteocyte distribution within and between trabeculae was not affected by age or menopause; and (5) the data raise the possibility that one function of remodeling in iliac cancellous bone is to maintain osteocyte viability.

Relationships between osteocyte density and bone formation rate in human cancellous bone

Iliac cancellous osteocyte density decreases with age in deep bone but not in superficial bone, most likely because of remodeling. It has been suggested that osteocytes can inhibit bone remodeling. Accordingly, we examined the relationship between osteocyte density and bone formation rate in 92 healthy women. In superficial bone (<25 microm from the surface), we found a weak but significant (p < 0.03) inverse correlation between BFR/BS and Ot. N/B.Ar that was unaffected by menopause and independent of age. A weaker positive relationship with empty lacunar density improved significance. The data appear to suggest a negative feedback loop, but osteocytes explain only 10% of the variance in BFR/BS, and 97% of the variance in osteocyte density is explained by total lacunar density. This measure of initial osteocyte density during bone formation has a high coefficient of variation (20%) indicating large individual differences. We conclude that: (1) our data support the proposal that osteocytes can inhibit bone remodeling; (2) osteocyte density in superficial bone depends mainly on initial osteocyte density during bone formation and is maintained but not regulated by bone remodeling; and (3) the inverse relationship between BFR/BS and osteocyte density may reflect the homeostatic need to maintain calcium exchangeability in the lining cell-osteocyte syncytium.

Osteoclast precursors as leukocytes : Importance of the area code

Osteoclasts are multinucleated cells formed by the fusion of mononuclear precursor cells. It is generally agreed that their immediate precursors, referred to here as preosteoclasts, are derived from the monocyte-macrophage lineage, and so ultimately from the hematopoietic stem cell, and that their fusion product resorbs bone; however, what happens in between these processes is much less clear. Preosteoclast production occurs in hematopoietic (red) marrow, which in the adult is restricted to the central skeleton. Bone resorption occurs at endosteal and, to a much lesser extent, periosteal bone surfaces throughout the skeleton. Preosteoclasts are known to circulate and in the peripheral skeleton the circulation provides their only access route to bone surfaces, via the sinusoid occupying the center of each basic multicellular unit (BMU), which is the instrument of bone remodeling (Figure 1). In the central skeleton, direct migration of preosteoclasts from red marrow to bone surfaces seems possible, and is frequently assumed to occur, but has not been demonstrated directly. The close spatial association between endothelial cells and vascular sinusoids and sites of surface remodeling suggests that, even in central cancellous bone, adjacent to red marrow, the structure of the BMU is essentially the same as in cortical bone, and that many osteoclasts are derived from circulating rather than from locally migrating precursors.

The bone formation defect in idiopathic juvenile osteoporosis is surface-specific.

We have previously shown that idiopathic juvenile osteoporosis (IJO) is characterized by a decreased cancellous bone volume and a very low bone formation rate on cancellous surfaces. Whether IJO similarly affects cortical bone is unknown. We therefore compared tetracycline double-labeled transfixing iliac-crest bone biopsies from eight children with typical clinical features of IJO (six girls; age 10-12 years) and from nine children (four girls; age 9-12 years) without metabolic bone disease. No differences in intracortical remodeling activity were detected. Both structural parameters reflecting intracortical remodeling (cortical porosity, active canal diameter, and quiescent canal diameter) and bone surface-based metabolic parameters (osteoid, osteoblast, mineralizing, osteoclast and eroded surfaces, and bone formation rate) were similar in IJO patients and controls (p > 0.2 each, t-test). Although the internal cortex of the biopsy was thinner in IJO patients than in controls (660 +/- 170 microm vs. 980 +/- 320 microm; p = 0.02), there was no difference in the width of the external cortex (p = 0.36). In growing children, both cortices exhibit an external modeling drift. Therefore, the difference in internal cortical width point to a decreased modeling activity on the endocortical surface of the internal cortex. In fact, bone formation rate on this surface was 48% lower in IJO patients than in controls (82 +/- 45 microm(3)/microm(2) per year vs. 159 +/- 162 microm(3)/microm(2) per year). However, this difference did not achieve statistical significance (p = 0.21) due to the high variability of bone formation rate on modeling surfaces. The disturbance of bone remodeling in IJO is limited to cancellous bone, but there may be a modeling defect affecting the internal cortex. Thus, the process causing IJO appears to mainly affect bone surfaces that are in contact with the bone marrow cavity.