JOURNAL OF BONE AND MINERAL RESEARCH Volume 7, Number 12, 1992 Mary Ann Liebert, Inc., Publishers
Bone Resorption and Formation on the Periosteal Envelope of the Ilium: A Histomorphometric Study in Healthy Women RAFFAELLA BALENA,* MEI-SHU SHIH, and A. MICHAEL PARFITT
ABSTRACT Continuation of net periosteal bone gain after cessation of longitudinal growth has been inferred from sequential radiographic morphometry. Accordingly, we performed histomorphometry of the periosteal surfaces of transilial bone biopsies from 57 healthy women aged 24-74 years, 29 premenopausal and 28 postmenopausal. Compared to the endocortical surface, the extents of eroded and osteoid surfaces were very similar, but the extents of osteoclast- and osteoblast-covered surfaces were 80-9OV0 smaller, and both wall thickness and osteoid thickness were about 30% lower. Double tetracycline labels were present in only 11 cases. The second (demethylchlortetracycline) label was almost four times as long as the first (oxytetracycline) label, a much greater difference than on the endocortical surface, so that the extent of mineralizing surface was based only on the second label. Even so, adjusted apposition rates and bone formation rates were only about 20% of the endocortical values, and unlike the endocortical surface, formation rates were not higher in the postmenopausal than in the premenopausal women. Resorption, reversal, and formation periods were each much longer than on the endocortical surface. There was no correlation between periosteal and endocortical values for any variable. At least 54% of total cement line length was scalloped, implying reversal of remodeling direction from resorption to formation, and at least 18% of total cement line length was smooth, implying temporary arrest of bone formation. Convincing evidence of modeling, related to growth or mechanical stimulation, was not observed. We conclude (1) periosteal bone turnover occurs by remodeling rather than modeling; it is much slower than endosteal turnover and extends less far from the surface, (2) periosteal bone formation is probably discontinuous with long periods of interruption, and (3) periosteal bone cell recruitment and activity appear to be relatively uninfluenced by the hormonal changes of aging and menopause; presumably local factors, possibly biomechanical, are more important.
INTRODUCTION the periosteum is defined as an envelope that demarcates the bone from the surrounding soft tissue.") It is composed of an outer fibrous layer and an inner cellular cambium layer; these are separated by an intermediate zone rich in blood vessels, mostly capillaries, ( m ) that gradually involutes and finally disappears with skeletal rnat~rity."~The human periosteum is devoid of lymphatics.(5JIn the growing skeleton the peri-
A
NATOMICALLY,
osteum is highly osteogenic, and it is possible to isolate from it two distinctive cell populations: epithelial-like and fibroblast-like cells; the former appear to be the cells from the cambium layer and the latter those from the outer fibrous layer. ( w By contrast, it is traditionally believed that the adult periosteum is minimally cellular and mainly fibrous,(') although it shows high osteogenic potential in response to pathologic stimuli. w) The hallmark of the aging skeleton is bone loss, but agerelated widening of bone contour has been documented in
Bone and Mineral Research Laboratory, Bone and Joint Center, and Department of Medicine, Henry Ford Hospital, Detroit, Michigan. *Present affiliation: Merck, Sharp and Dohme Research Laboratories, West Point, Pennsylvania.
1475
1476
BALENA ET AL.
the s k ~ l l , ( ~femur,(L1.L’) .~~) metacarpal,(”) and other bones. (I4) This phenomenon has been attributed to periosteal bone apposition continuing in adulthood, for which there is direct evidence in the rib.(1sJHowever, the nature and mechanism of the periosteal changes remain incompletely understood. (14-17J Periosteal bone formation in the adult skeleton could represent remodeling, if it occurred in relation to previous resorption, or modeling, if it occurred directly without preceding resorption.‘”’ In the latter case, there could be continuation, at a much slower rate, of the processes characteristic of growth(17)or a response to changes in the mechanical demands on the bone.(19)We examined the process of periosteal bone apposition at the intermediary level of organization of bone, ( l o ) applying histomorphometry to the periosteum of iliac biopsies obtained from healthy women who received two time-spaced tetracycline labels.
MATERIALS AND METHODS A group of 57 healthy women were studied, recruited as previously described and j u ~ t i f i e d . ( ~ ’Of -~~ these, ) 29 were premenopausal (mean age 35.2 years; range 24-48) and 28 postmenopausal (mean age 59.7 years, range 50-74); of these 11 (8 premenopausal and 3 postmenopausal) were black and the remainder white. Double tetracycline of bone in vivo was accomplished by giving oxytetracycline (green fluorescence) for 3 days, and demethylchlortetracycline (yellow fluorescence) for 3 days, separated by an 11 day interval. ( 2 1 , 1 3 ) Transiliac bone biopsies were obtained 4 days after the end of the second labeling period.(24’The only inclusion criterion was the presence in the specimen of periosteum on both outer and inner cortices, together with good quality of the sections. The bone specimens were fixed, dehydrated, sectioned, stained, and measured as previously described. (11-23) A special template was derived for use with the Bioquant IV digitizing system. The periosteal bone surface was defined as the border between cortical bone and the fibrous tissue above it. The mean length of the available periosteum was approximately 12.5 mm (6.0 mm on the external cortex and 6.5 mm on the internal cortex). In each case, six sections were measured, three sections with bright-field illumination and three adjacent sections under ultraviolet light, and the average for each variable was calculated. Total magnifications used were x 100 for the lengths of total periosteal and eroded surface, x400 for osteoid thickness, and x200 for all other measurements. Osteoid was measured only when it exceeded 2 pm in thickness. Wall thickness was measured as the mean distance between the resting periosteal bone surface and the nearest reversal line in sections stained with either gallocyanine or toluidine blue. cZs) Equidistant multiple measurements were taken, ensuring that the micrometer remained perpendicular to the cement line; the same was done for osteoid thickness. Because of substantial inequality of label length, the mineralizing surface was calculated using only the length of the second label.‘23)Derived dynamic indices in biopsies void of double labels were
handled as previously suggested. (”) Mineral apposition rate was assumed to be the minimum value of 0.3 pg/day if only a single label was present and to be a missing value if no label was present.(’” Histomorphometric data are reported in three-dimensional terms with use of standard abThe results are compared with data previbreviations. ously obtained by the same methods from the endocortical surfaces of the same biopsies using the same sections. To establish the occurrence or not of bone remodeling in the periosteum, two different investigators independently evaluated the morphology of the cement lines under the periosteal surface. Cement lines were counted and assigned to three different categories: reversal (irregular), arrest (regular), and mixed, with both irregular and regular segments in the same line; lines of indeterminate morphology were not included. This was possible only in 21 cases; in the remaining 25 cases too few cement lines were stained with sufficient clarity to be identifiable. The length of each line was measured by the grid method., ( 2 6 . 1 e ) Additionally, we determined the distance of each line from the periosteal surface for allocation to zone A ( 5 35 pm from surface) or zone B ( > 35 pm from surface). For each variable a two-way ANOVA was performed,(29) with classification according to menopausal status (before versus after) and surface (periosteal versus endocortical). In addition, differences between the pre-and postmenopausal groups were analyzed by unpaired Student’s t-test, and differences between surfaces and between zones by paired t-test; all t-tests were two-sided. Linear regression analysis was used to determine the relation between age and selected histologic measurements. In each zone, data for the three types of cement lines were compared by one-way ANOVA. All statistical computations were done using the software package Stat-view (Maclntosh).
RESULTS Approximately 30% of the periosteal surface was either covered with osteoid or eroded, and approximately 70% was quiescent; the latter proportion was significantly lower in mean value in the postmenopausal subjects, although the ranges were similar (Table 1). For both osteoid and eroded surface, the mean values were very similar to those on the endocortical surface, which also showed a significantly lower value for quiescent surface in the postmenopausal subjects. Osteoclasts were found in only 14 cases, although some eroded surface was present in every case. The mean value for osteoclast-covered surface did not differ significantly between pre- and postmenopausal subjects but was significantly lower than on the endocortical surface. Osteoblasts were found in 45 cases; all were classified‘”) as intermediate (type 111) in shape and degree of basophilia between classic cuboidal osteoblasts (type 11) and flattened terminal osteoblasts (type IV). The osteoblast surface was significantly greater in the postmenopausal than in the premenopausal subjects but in both groups was significantly lower than on the endocortical surface. Osteoid thickness was significantly higher in the postmeno-
1477
PERIOSTEAL BONE REMODELING IN NORMAL WOMEN
TABLE1. BONEREMODELING IN THE PERIOSTEUM: STATIC
Premenopausal (n = 29) Periosteal OS/BS, To ES/BS, To QS/BS, To Oc.S/BS,
0'9
Oc.S/Md.S, Ob.S/BS, To Ob.S/OS, To Rv.S/BS, Vo O.Th, pm W.Th, pm
070
18.6 f 1.99 (5.10-48.3) 9.19 f 1.07 (0.70-26.1) 72.2 f 2.12 (44.7-89.6) 0.12 f 0.06 (0.00-1.5O)f 0.26f 0.15 (0.00-4.00)f 0.61 f 0.19 (0.00-4.14)' 3.49 f 0.99 (0.00-22.8)h 9.07 f 1.07 (0.70-26.1) 5.93 f 0.26 (3.65-8.79) 25.3 f 0.981 ( 1 9.6-34.1)
Endocortical
19.7 f 2.30 (3.50-46.3) 9.16 f 1.11 (1.26-19.0) 71.1 f 2.52 (43.7-93.7) 0.63 f 0.17 (0.00-3.20)" 0.86 f 0.22 (0.00-3.91)d 5.59 f 0.91 (0.00- 16.5)i 28.9 f 3.74 (0.00-79.3)i 8.56 f 1.03 ( 1 .17-18.5) 8.70 f 0.43 (4.48-13 .O) 37.1 f 0.98 (25.7-49.O)
VARIABLESa
Postmenopausal (n = 28) Periosteal
23.5 f 1.96 (4.80-44.2) 11.2 f 1.29 (1.60-29.6) 66.1 f 2.14b (47.2-83.4) 0.11 f 0.04 (0.00-0.80)e 0.14 f 0.05 (0.00-1.W)e 2.34 f 0.36e (0.00-6.71)J 9.74 f 1 . 1 9 (0.00-21.99 11.1 f 1.28 (1.60-29.6) 6.83 f 0.36b (4.57-10.8) 26.2 f 1.55m (16.7-41.9)
Endocortical
26.4 f 3.06 (1.44-80.3) 9.78 f 0.94 (1.97-20.5) 63.8 f 2.86b (10.9-87.9) 0.88 f 0.12 (0.00-2.54)f 1.45 f 0.36 (0.00-10.8)f 6.12 f 1.51 (0.00-35.8)k 18.8 f 2.96 (0.00-47.2)k 9.04 f 0.94 (1.97-19.7) 8.95 f 0.47 (4.08-15.2) 36.9 f 1.16 (20.5-49.7)
ANOVA
M
S
I
< 0.02
NS
NS
NS
NS
NS
< 0.01
NS
NS
NS
< 0.001
NS
NS
Bone Resorption and Formation on the Periosteal Envelope of the Ilium: A Histomorphometric Study in Healthy Women RAFFAELLA BALENA,* MEI-SHU SHIH, and A. MICHAEL PARFITT
ABSTRACT Continuation of net periosteal bone gain after cessation of longitudinal growth has been inferred from sequential radiographic morphometry. Accordingly, we performed histomorphometry of the periosteal surfaces of transilial bone biopsies from 57 healthy women aged 24-74 years, 29 premenopausal and 28 postmenopausal. Compared to the endocortical surface, the extents of eroded and osteoid surfaces were very similar, but the extents of osteoclast- and osteoblast-covered surfaces were 80-9OV0 smaller, and both wall thickness and osteoid thickness were about 30% lower. Double tetracycline labels were present in only 11 cases. The second (demethylchlortetracycline) label was almost four times as long as the first (oxytetracycline) label, a much greater difference than on the endocortical surface, so that the extent of mineralizing surface was based only on the second label. Even so, adjusted apposition rates and bone formation rates were only about 20% of the endocortical values, and unlike the endocortical surface, formation rates were not higher in the postmenopausal than in the premenopausal women. Resorption, reversal, and formation periods were each much longer than on the endocortical surface. There was no correlation between periosteal and endocortical values for any variable. At least 54% of total cement line length was scalloped, implying reversal of remodeling direction from resorption to formation, and at least 18% of total cement line length was smooth, implying temporary arrest of bone formation. Convincing evidence of modeling, related to growth or mechanical stimulation, was not observed. We conclude (1) periosteal bone turnover occurs by remodeling rather than modeling; it is much slower than endosteal turnover and extends less far from the surface, (2) periosteal bone formation is probably discontinuous with long periods of interruption, and (3) periosteal bone cell recruitment and activity appear to be relatively uninfluenced by the hormonal changes of aging and menopause; presumably local factors, possibly biomechanical, are more important.
INTRODUCTION the periosteum is defined as an envelope that demarcates the bone from the surrounding soft tissue.") It is composed of an outer fibrous layer and an inner cellular cambium layer; these are separated by an intermediate zone rich in blood vessels, mostly capillaries, ( m ) that gradually involutes and finally disappears with skeletal rnat~rity."~The human periosteum is devoid of lymphatics.(5JIn the growing skeleton the peri-
A
NATOMICALLY,
osteum is highly osteogenic, and it is possible to isolate from it two distinctive cell populations: epithelial-like and fibroblast-like cells; the former appear to be the cells from the cambium layer and the latter those from the outer fibrous layer. ( w By contrast, it is traditionally believed that the adult periosteum is minimally cellular and mainly fibrous,(') although it shows high osteogenic potential in response to pathologic stimuli. w) The hallmark of the aging skeleton is bone loss, but agerelated widening of bone contour has been documented in
Bone and Mineral Research Laboratory, Bone and Joint Center, and Department of Medicine, Henry Ford Hospital, Detroit, Michigan. *Present affiliation: Merck, Sharp and Dohme Research Laboratories, West Point, Pennsylvania.
1475
1476
BALENA ET AL.
the s k ~ l l , ( ~femur,(L1.L’) .~~) metacarpal,(”) and other bones. (I4) This phenomenon has been attributed to periosteal bone apposition continuing in adulthood, for which there is direct evidence in the rib.(1sJHowever, the nature and mechanism of the periosteal changes remain incompletely understood. (14-17J Periosteal bone formation in the adult skeleton could represent remodeling, if it occurred in relation to previous resorption, or modeling, if it occurred directly without preceding resorption.‘”’ In the latter case, there could be continuation, at a much slower rate, of the processes characteristic of growth(17)or a response to changes in the mechanical demands on the bone.(19)We examined the process of periosteal bone apposition at the intermediary level of organization of bone, ( l o ) applying histomorphometry to the periosteum of iliac biopsies obtained from healthy women who received two time-spaced tetracycline labels.
MATERIALS AND METHODS A group of 57 healthy women were studied, recruited as previously described and j u ~ t i f i e d . ( ~ ’Of -~~ these, ) 29 were premenopausal (mean age 35.2 years; range 24-48) and 28 postmenopausal (mean age 59.7 years, range 50-74); of these 11 (8 premenopausal and 3 postmenopausal) were black and the remainder white. Double tetracycline of bone in vivo was accomplished by giving oxytetracycline (green fluorescence) for 3 days, and demethylchlortetracycline (yellow fluorescence) for 3 days, separated by an 11 day interval. ( 2 1 , 1 3 ) Transiliac bone biopsies were obtained 4 days after the end of the second labeling period.(24’The only inclusion criterion was the presence in the specimen of periosteum on both outer and inner cortices, together with good quality of the sections. The bone specimens were fixed, dehydrated, sectioned, stained, and measured as previously described. (11-23) A special template was derived for use with the Bioquant IV digitizing system. The periosteal bone surface was defined as the border between cortical bone and the fibrous tissue above it. The mean length of the available periosteum was approximately 12.5 mm (6.0 mm on the external cortex and 6.5 mm on the internal cortex). In each case, six sections were measured, three sections with bright-field illumination and three adjacent sections under ultraviolet light, and the average for each variable was calculated. Total magnifications used were x 100 for the lengths of total periosteal and eroded surface, x400 for osteoid thickness, and x200 for all other measurements. Osteoid was measured only when it exceeded 2 pm in thickness. Wall thickness was measured as the mean distance between the resting periosteal bone surface and the nearest reversal line in sections stained with either gallocyanine or toluidine blue. cZs) Equidistant multiple measurements were taken, ensuring that the micrometer remained perpendicular to the cement line; the same was done for osteoid thickness. Because of substantial inequality of label length, the mineralizing surface was calculated using only the length of the second label.‘23)Derived dynamic indices in biopsies void of double labels were
handled as previously suggested. (”) Mineral apposition rate was assumed to be the minimum value of 0.3 pg/day if only a single label was present and to be a missing value if no label was present.(’” Histomorphometric data are reported in three-dimensional terms with use of standard abThe results are compared with data previbreviations. ously obtained by the same methods from the endocortical surfaces of the same biopsies using the same sections. To establish the occurrence or not of bone remodeling in the periosteum, two different investigators independently evaluated the morphology of the cement lines under the periosteal surface. Cement lines were counted and assigned to three different categories: reversal (irregular), arrest (regular), and mixed, with both irregular and regular segments in the same line; lines of indeterminate morphology were not included. This was possible only in 21 cases; in the remaining 25 cases too few cement lines were stained with sufficient clarity to be identifiable. The length of each line was measured by the grid method., ( 2 6 . 1 e ) Additionally, we determined the distance of each line from the periosteal surface for allocation to zone A ( 5 35 pm from surface) or zone B ( > 35 pm from surface). For each variable a two-way ANOVA was performed,(29) with classification according to menopausal status (before versus after) and surface (periosteal versus endocortical). In addition, differences between the pre-and postmenopausal groups were analyzed by unpaired Student’s t-test, and differences between surfaces and between zones by paired t-test; all t-tests were two-sided. Linear regression analysis was used to determine the relation between age and selected histologic measurements. In each zone, data for the three types of cement lines were compared by one-way ANOVA. All statistical computations were done using the software package Stat-view (Maclntosh).
RESULTS Approximately 30% of the periosteal surface was either covered with osteoid or eroded, and approximately 70% was quiescent; the latter proportion was significantly lower in mean value in the postmenopausal subjects, although the ranges were similar (Table 1). For both osteoid and eroded surface, the mean values were very similar to those on the endocortical surface, which also showed a significantly lower value for quiescent surface in the postmenopausal subjects. Osteoclasts were found in only 14 cases, although some eroded surface was present in every case. The mean value for osteoclast-covered surface did not differ significantly between pre- and postmenopausal subjects but was significantly lower than on the endocortical surface. Osteoblasts were found in 45 cases; all were classified‘”) as intermediate (type 111) in shape and degree of basophilia between classic cuboidal osteoblasts (type 11) and flattened terminal osteoblasts (type IV). The osteoblast surface was significantly greater in the postmenopausal than in the premenopausal subjects but in both groups was significantly lower than on the endocortical surface. Osteoid thickness was significantly higher in the postmeno-
1477
PERIOSTEAL BONE REMODELING IN NORMAL WOMEN
TABLE1. BONEREMODELING IN THE PERIOSTEUM: STATIC
Premenopausal (n = 29) Periosteal OS/BS, To ES/BS, To QS/BS, To Oc.S/BS,
0'9
Oc.S/Md.S, Ob.S/BS, To Ob.S/OS, To Rv.S/BS, Vo O.Th, pm W.Th, pm
070
18.6 f 1.99 (5.10-48.3) 9.19 f 1.07 (0.70-26.1) 72.2 f 2.12 (44.7-89.6) 0.12 f 0.06 (0.00-1.5O)f 0.26f 0.15 (0.00-4.00)f 0.61 f 0.19 (0.00-4.14)' 3.49 f 0.99 (0.00-22.8)h 9.07 f 1.07 (0.70-26.1) 5.93 f 0.26 (3.65-8.79) 25.3 f 0.981 ( 1 9.6-34.1)
Endocortical
19.7 f 2.30 (3.50-46.3) 9.16 f 1.11 (1.26-19.0) 71.1 f 2.52 (43.7-93.7) 0.63 f 0.17 (0.00-3.20)" 0.86 f 0.22 (0.00-3.91)d 5.59 f 0.91 (0.00- 16.5)i 28.9 f 3.74 (0.00-79.3)i 8.56 f 1.03 ( 1 .17-18.5) 8.70 f 0.43 (4.48-13 .O) 37.1 f 0.98 (25.7-49.O)
VARIABLESa
Postmenopausal (n = 28) Periosteal
23.5 f 1.96 (4.80-44.2) 11.2 f 1.29 (1.60-29.6) 66.1 f 2.14b (47.2-83.4) 0.11 f 0.04 (0.00-0.80)e 0.14 f 0.05 (0.00-1.W)e 2.34 f 0.36e (0.00-6.71)J 9.74 f 1 . 1 9 (0.00-21.99 11.1 f 1.28 (1.60-29.6) 6.83 f 0.36b (4.57-10.8) 26.2 f 1.55m (16.7-41.9)
Endocortical
26.4 f 3.06 (1.44-80.3) 9.78 f 0.94 (1.97-20.5) 63.8 f 2.86b (10.9-87.9) 0.88 f 0.12 (0.00-2.54)f 1.45 f 0.36 (0.00-10.8)f 6.12 f 1.51 (0.00-35.8)k 18.8 f 2.96 (0.00-47.2)k 9.04 f 0.94 (1.97-19.7) 8.95 f 0.47 (4.08-15.2) 36.9 f 1.16 (20.5-49.7)
ANOVA
M
S
I
< 0.02
NS
NS
NS
NS
NS
< 0.01
NS
NS
NS
< 0.001
NS
NS
