Division of Endocrinology and Metabolism, Center for Osteoporosis and Metabolic Bone Diseases, University of Arkansas for Medical Sciences, Little Rock, Arkansas, USA.
Address correspondence to: A. Michael Parfitt, 4301 West Markham, #587, Little Rock, Arkansas 72205-7199, USA. Phone: 501.686.5130; Fax: 501.686.8148; E-mail: email@example.com.
First published January 2, 2014 - More info
I devised a method for obtaining information on cancellous bone structure from iliac bone histomorphometry that led to the demonstration that architecture is an important component of bone strength and bone fragility. Furthermore, this method contributed to the recognition of the importance of changes in osteoclast and osteocyte apoptosis in response to estrogen deficiency and replacement.
Loss of bone is a universal consequence of aging, from which no subset of the human species is exempt. Age-related bone loss can be demonstrated even in prehistoric populations. The magnitude of age-related cortical bone loss is easily measured on simple radiographs of the hands, but until approximately 25 years ago the magnitude of age-related cancellous bone loss could be measured only by histologic examination, either at autopsy or in iliac biopsy specimens. Cancellous bone has a complex three-dimensional structure that is similar in appearance to a honeycomb or a sponge. Until my 1983 paper (1), the only measurement made on clinical biopsy samples was the proportion of space occupied by bone, which is usually referred to in the clinical literature as cancellous (or trabecular) bone volume and in the biomechanical literature as bone volume fraction. Basing my research on earlier work by Robert Schenk (2, 3), I showed that with minimal expenditure of time, one could obtain additional structural indices of both clinical and scientific interest that only required calibration of an eyepiece graticule with a stage micrometer, and this could be used to determine absolute values of thickness, surface, and separation. The calculations were based on a parallel plate model of cancellous bone, which although extremely oversimplified, was a great improvement compared to ignoring structure altogether. This method enabled a rough description of the age-related transformation of trabeculae from plates to rods.
The most obvious implication of this work was biomechanical — the relationship between bone structure and bone strength. Patients with vertebral fracture had about 30% fewer plates in a cubic millimeter of cancellous bone tissue than expected, based on the values predicted from the linear regression analysis of the relationship between plate number and age in healthy subjects. Furthermore, plate number more clearly distinguished between subjects with and without vertebral fracture than trabecular bone volume. This was true for both sexes, although the deficit in bone thickness was greater in men than in women. In contrast, patients with hip fracture had a relative deficit in plate thickness that was greater than the deficit in plate number. As observed for patients with vertebral fracture, this deficit was greater in men than in women. The notion that there were differences as well as similarities in the pathogenesis of the two fracture types was proposed by Riggs and Melton in the same year, based on quite different reasoning (4).
My colleague Michael Kleerekoper used data from our laboratory to show that patients with nontraumatic vertebral compression fracture had markedly fewer plates than healthy control subjects of similar age, matched for total amount of cancellous bone (5). Due to the method of selection, the patients had thicker trabeculae, which surprisingly had not restored vertebral compression strength. Later we collaborated with Lee Feldkamp, who examined cylindrical biopsy specimens with microcomputed tomography. This technique produced an index of three-dimensional connectivity that revealed greater differences between the groups (6). This led to an explosion of new methods for examining the three-dimensional structure of bone, with progressively greater resolution, that could be applied to patients without biopsy. Together these methodologies confirmed and extended our conclusions. Bone architecture had become a major component of bone quality, a convenient term for all non- mass factors that contribute to fracture risk (7).
Trabecular plate number was significantly lower in healthy postmenopausal women than in healthy premenopausal women; however, the difference in trabecular thickness was much smaller, and in our subjects was not statistically significant (1). In our cross-sectional study, we could not distinguish between the effects of age and the effects of menopause. At this time, it was already known that after oophorectomy, which causes abrupt rather than gradual estrogen deficiency, vertebral cancellous bone mass declines rather rapidly (8). Our data suggested that in the first few years after menopause, some structural elements of cancellous bone are completely removed, and those remaining gradually become thinner, an effect most likely due to a defect in osteoblast function. This sequence could only be initiated if the earliest effect of estrogen deficiency was the perforation of trabecular plates, which could only occur as a result of increased depth of osteoclastic resorption (Figure 1). At first, I postulated the presence of “killer” osteoclasts that removed bone more voraciously as the underlying cause of increased resorption depth. I soon realized that increased depth was most likely due to resorption lasting longer, which led to a focus on increased osteoclast life span. Apoptosis of osteoclasts was first demonstrated by David Hughes (9), and consideration of this process led to formulation of a new model of bone resorption control (10). My current colleagues at the University of Arkansas for Medical Sciences have demonstrated that osteoclast apoptosis is postponed by estrogen deficiency and that osteoclast life span is restored to normal by estrogen replacement via cell-autonomous effects, which are mediated by estrogen receptor α (11).
Estrogens attenuate osteoclast generation and life span via cell-autonomous actions mediated by estrogen receptor α. Loss of these actions is sufficient for loss of bone in the cancellous compartment, in which complete perforation of trabeculae by osteoclastic resorption precludes subsequent refilling of the cavities by the bone-forming osteoblasts. The direction of BMU advance is not shown because of the section plane. Image courtesy of Robert Weinstein and reproduced with permission from Springer (20).
A final implication of this work that is subtler and less well established concerns more recent knowledge of cancellous bone remodeling and the role of osteocytes in this process (12). Osteoclasts work not in isolation but in tandem with osteoblasts in a temporary structure referred to as the bone multicellular unit (BMU), the function of which is replacement of effete bone by new bone (13). Osteoclasts travel not perpendicular to the surface but at the front end of a BMU, which excavates a trench across the surface of cancellous bone (14). Consequently, in order to accomplish plate perforation, the BMU must change direction from a trajectory that is parallel to the surface to one that is at a downward angle. The direction of BMU advance can only be controlled by osteocytes (15). BMUs that repair fatigue microdamage follow signals from dying osteocytes (16). Osteocyte death increases with age in central interstitial bone (17). In a cross-sectional study, we could not find an effect of menopause, but ovariectomy in rats increases osteocyte apoptosis, which is prevented by estrogen replacement (18). BMUs that follow signals from dying osteocytes in interstitial bone would likely lead to plate perforation, particularly if the plate was simultaneously attacked from both sides (Figure 1 and ref. 19).
I thank Stavros Manolagas, Robert Weinstein, and Leah Elrod for their assistance.
Conflict of interest: The author has declared that no conflict of interest exists.
Reference information: J Clin Invest. 2014;124(1):70–71. doi:10.1172/JCI73843.