Bone, 12, pp. 429-437, (1991) Printed in the USA. All rights reserved.

Copyright

87%3282/91 $3.00 + .OO 0 1991 Pergamon Press plc

A New Direction for Osteoporosis Research: A Review and Proposal HAROLD M. FROST Departmentof OrthopaedicSurgery, SouthernColorado Clinic, 41 Montebello, Pueblo, Colorado 81001, U.S.A. Address,for correspondence and reprints: Dr. Harold M. Frost, 41 Montebello, Pueblo, CO 81001, U.S.A.

Abstract

volved in controlling bone mass in growing mammals, combines multidisciplinary evidence and ideas.

This article suggests why drugs that only reduce the activity of existing osteoclasts or enhance the activity of existing osteobhtsts probably cannot cure the osteopenias associated with most osteoporoses. Instead they should add only limited amounts of bone, which should begin to disappear after treatment stops. That behavior depends on these facts. Different threshold ranges of mechanical bone strains control gains and losses of bone mass. One threshold controls gains by modeling drifts, and a lower one controls losses by remodeling BMUs. A drug that does not change those thresholds should lhnit gains (or losses) of bone during indefmitely continued treatment. Success in curing those osteopenias should require learning how to change the thresholds. Key Words: Osteoporosis - Bone - Treatment chanics - Physiology - Osteopenia - Modeling eling.

and it

Definitions

Let osteopenia mean a bone or skeleton with less bone than age- and sex-comparable norms. It need not imply disease. Let osteoporosis mean a disease that combines an osteopenia with clinical evidence of increased bone fragility, as M.R. Urist suggested in 1960 (Urist 1960, 1982). Let bone muss mean the amount of bone tissue, and as a volume after removing the marrow space. It does not mean gravimetric mass. This article concerns control of our bone mass in ad&t life, An Analogy

- Biome- Remod-

In many modem homes two thermostats control two machines to keep the temperature within desirable limits. One thermostat controls heating by a furnace, the other controls removal of heat by an air conditioner. The thermostat settings usually leave both machines OFF in a “comfort zone,” say between 68-72” F. If the temperature falls below that zone the furnace goes ON to add heat. If the temperature rises above that zone the air conditioner turns ON to remove heat. Fitting bone mass to mechanical usage (MU) has similar properties. Globally speaking, one biologic mechanism can add but not remove bone, and another can remove but not add it. Something called the mechanostat can turn each ON and OFF, two MES ranges define the settings of its two “thermostats” that control those two functions, and bone mass is the “heat.”

Introduction Here is a blunt statement: Drugs that merely depress osteoclast vigor or enhance osteoblast vigor cannot cure the osteopenias that accompany osteoporoses. This article explains why that is not as foolish as it may seem at first, and some of its nuances. Between 1983-1990 what others call a new paradigm of skeletal biology crystallized out of multidisciplinary evidence (Anderson 1990; Burr 1990; Jee 1990; Martin 1990; Reeker 1990). Its conceptual power combines cell with tissue and organ level information (Burr & Martin 1989; Frost 1973a 199Oc,d; Jee 1989; Martin & Burr 1989; Takahashi 1990). Here and in his post-1973 publications the author serves mostly as a spokesman for legions of others who contributed to the paradigm. Its view of the pathogenesis of osteoporoses supplements more conventional ones. Describing the new direction it suggests for future research on correcting the OSteopenia of these diseases requires agreeing on some definitions, presenting an analogy, and reviewing some bone biology, vital biomechanics, and anatomical evidence that suggest the new direction. For brevity’s sake, references instead of detailed explanations support the statements below. Table I lists the text’s abbreviations and acronyms. The text focuses on adult-acquired osteoporoses, it ignores some physiology in-

The Biologic Mechanisms The modeling and remodeling mechanisms Between 1964-1973 we realized that two operationally different systems can affect bone mass and architecture, yet each depends on osteoclasts and osteoblasts. Modeling drifs can thicken cortices and trabeculae in children, and minimodeling drifts can thicken trabeculae in children and adults. This drift-based modeling becomes inefficient in cortical bone in adult mammals including humans. It can either go ON, or it can stopgo OFF. With respect to bone mass, global modeling can add or not add bone to the bone bank, but apparently cannot reduce the bank (“global” in this text means averaged over the whole skeleton). If heat stands 429

H. M. Frost: Osteoporosis

430

The Vital Biomechanics

Table I. Abbreviations

Materials properties

BMU:

Basic multicellular unit of bone remodeling.

MBSm:

The Minimum Effective Strain that turns previously dormant modeling drifts ON to begin strengthening bone. A range, not a step function. The Minimum Effective Strain for beginning to depress BMU creations and equalize resorption and formation in completed BMUs. A range, not a step function. Mechanical usage. The vigor of MU in this text refers to the size of the mechanical forces on bones, not to their number or frequency.

ME&:

MU:

research

Strain (deformation such as stretching, shortening, twisting) can express bone strength as well as stress can. Special gauges can measure bone strains in vivo (Carter 1987; Cowin 1989; Lanyon 1984). Normal bone fractures at about 25,000 microstrain, which will provide a yardstick to compare with other things below (Burstein & Reilly 1976; Currey 1984). That fracture strain defines a 2.5% stretching or shortening of a bone, It corresponds to stresses of about 130 mega-Pascals or 16,000 psi. The loading history

for bone,

then

the furnace

stands

for

modeling

in the above

analogy (Burr & Martin 1989; Courpron 1981; Frost 1964a,b, 1973a,b, 1985, 1986, 1987c, 1990a; Jee 1989, 1990; Jee et al. in press; Li et al. 1990, in press; Reeker 1983; Uhthoff 1986; Weinmann & Sicher 1955). Bone remodeling BMUs (Basic Multicellular Units) turn bone over in small packets of about 0.05 mm3 of bone per completed BMU. They do it on all bone surfaces throughout life, if at varying rates. However, where bone touches marrow, BMUs normally resorb a bit more bone than they make (Frost 1985). That causes a life-long enlargement of the marrow cavity and losses of spongiosa in all mammals (Johnson 1964). The rate of that bone loss depends on BMU creations and the bone deficit in compkted BMUs. Fig. 1 illustrates the idea. Global remodeling can either remove or conserve bone, but apparently cannot increase bone mass. In that sense remodeling resembles the air conditioner in the above analogy (Anderson & Danylchuk 1978; Baron et al. 1984; Burr & Martin 1989; Courpron 1981; Danylchuk 1978; Darby & Meunier 1981; Eriksen 1986; Frost 1986; High 1988; Jee 1989, 1990; Jee et al. in press; Kimmel & Jee 1982; Li et al. 1990, in press; Marie 1982; Melsen et al. 1989; Parfitt 1983; Reeker 1983; Uhthoff 1986; Weinmann & Sicher 1955).

Between 1964-1973 we realized the biologic responses to mechanical usage depend on some time-averaged value of the mechanical forces on bone, and they assign disproportionately more weight to large than to small loads, no matter how frequent the latter (Carter 1987; Frost 1964b, 1973b, 1986, 1990a; Hart & Davy 1989; Huiskes et al. 1989; Rubin & Lanyon 1984, 1987). Longitudinal bone strains provide a good index of those loads. Investigators still try to find how nature weights and ranks strain rate, magnitude, frequency, and gradients in fitting bone to its mechanical usage. Here mechanical usage (MU) will mean a loading history, a term suggested by Carter (1987), and in the sense of typical peak strains. Qualitative MU (Mechanical Usage) effects on modeling and remodeling In acute disuse there is suddenly more bone than needed and modeling drifts go OFF to stop adding bone. BMU creations increase throughout a bone, and where bone touches marrow completed BMUs form much less than they resorb. These BMU changes remove cortical-endosteal and trabecular bone,

Fig. 1. These are “stair graphs” (after P. J. Meunier). Top row: Given equal BMU creations, on the left a small deficit of bone in completed BMUs would thin a trabecula or cortex as shown here, while the larger deficit on the right would make the same number of BMUs remove more bone (AB). The deficit is signified by Greek rho. Eortom row: Here we see the effect of changing BMU creations (called activation by histomorphometrists and signified by Greek mu). Given similar values of rho, more BMU creations on the right can remove much more bone than on the left. If the left column represents normal, the right column shows the effect both of disuse, and of an increased MESr for BMU-based remodeling.

431

H. M. Frost: Osteoporosis research Table II. Some bone features of steady state MU

BMU creations rho Modeling drifts Cortical bank Trabecular bank Outside diameter Marrow cavity diameter Cortical porosity

Disuse

Normal MU

Hyper MU

N N OFF D D N I N

N N OFF N N N N N

D N ON I I N D N

rho = net loss of bone per typical completed BMU. >neg = greater net loss. I = increased. N = normal. D = decreased. MU = mechanical usage.

Normal MU leaves modeling

drifts still OFF, decreases BMU creations, and tends to equalize their resorption and formation. Those effects tend to conserve existing bone. Hypervigorous MU tends to turn modeling drifts ON to strengthen and increase the bone bank. It may depress BMU creations slightly below normal and does keep resorption and formation equalized. Contrary to what intuition suggests, it apparently does ll~t make BMUs form more than they resorb (although pharmacologic treatment may change that) (Anderson et al. 1984; Hori et al. 1988; Reeve et al. 1980). So, global BMU-based remodeling can adapt bone to underloads but not to overloads, while global modeling drifts can adapt bone to overloads but not to underloads. Increased BMU creations usually increase bone losses next to marrow, and increased MU depresses BMU creations while disuse derepresses them. Table II summarizes that material (Aloia et al. 1986; Amptmann & Oyama 1973; Bachus et al. 1990; Bikle et al. 1987; Frost 1986, 1989a, 199Oa,b; Hansson et al. 1975; Jaworski & Uhthoff 1986; Jaworski et al. 1980; Jee 1989; Jee et al. 1987, 1988, 1989; Jones et al. 1977; Krolner et al. 1983; Li et al. 1989, 1990, in press; Margulies 1986; Martin & Burr 1989; Matsuda et al. 1986; Meunier et al. 1974; Nilsson 1978; Sin&in et al. 1986; Smith & Gilligan 1989; Thomaidis & Lindholm 1976; Turner & Bell 1986; Turner et al. 1991; Uhthoff & Jaworski 1978; Vesterby et al. 1989; Weinreb et al. 1989; Whalen & Carter 1990; Whedon 1984; Wronski & Morey 1982; Wunder et al. 1960). Quantitative MU effects and the “adapted window”

Between 1974-1983 studies of in vivo bone strains revealed that MU thresholds (“thermostat” settings) do control the above responses. Corresponding longitudinal bone strains can express them. The modeling effects. When and where typical peak bone strains exceed about 1500 microstrain (called here the MESm, and 6% of the fracture strain), modeling drifts turn ON to strengthen bone and increase bone mass. Strains below that threshold leave drifts OFF. The remodeling effects. When and where typical peak strains stay below about 50 microstrain (here called the MESr and less than 1% of the fracture strain), BMU creations increase on all bone surfaces, while next to marrow formation in completed BMUs decreases to make remodeling begin removing that bone faster than normal. Strains above that threshold

begin depressing BMU creations and equalizing their resorption and formation, which begins conserving existing bone (Biewener 1990; Bouvier 1985; Burr & Martin 1989; Frost 1983a, 1987b,c, 1989a,b, 199Oa,b; Fukada & Yasuda 1957; Jee 1989; Jee & Li 1990; Jee et al. in press; Keller & Spengler 1982; Lanyon 1984; Li et al. 1989; Li & Jee 1990; Li et al. 1990; MeLeod & Rubin 1990; Nunamaker et al. 1987; O’Connor et al. 1982; ORS 1977-1991; Rubin & Lanyon 1987; Turner et al. 1991). In normal adult bony skeletons, typical peak strains should everywhere stay between the MESr and MESm. Those lower and upper thresholds can define an “adapted window” for bone relative to its typical MU. It follows that in a “disuse window” bone strains would stay below the MESr, and in a “mild overload window” strains would equal or exceed the MESm. Growing mammals, including children, probably live in the mild overload window (Frost 1990 a,b; Martin & Burr 1989; Smith & Gilligan 1989). The adapted window corresponds to the comfort zone in the earlier analogy, where both the furnace and air conditioner tend to stay OFF (understanding that in living bone, OFF can correspond to some minimum level of activity instead of a total lack of it). The mechanostat

Many people realized Nature fits bone architecture and mass to our typical physical activities (Evans 1957; Becker et al. 1964; Cowin 1989; Pauwels 1986; Weinmann & Sicher 1955). Where those activities require more bone, it is added. Where bone is not needed it is removed. We do not know yet what the mechanism that controls these things looks like, but the process is as clear as the relationship between sexual intercourse and pregnancy before the intervention of the microscope. To discuss this mechanism we call it the “mechanostat” (Frost 1987b,c; Martin & Burr 1989; Takahashi 1990). That raises questions about what criteria distinguish adapted from under- or overadapted bones. The in vivo strain studies suggest the strain ranges that correspond to those criteria, the MESm and MESr. In different words: The MESm and ME,!+ values are the mechanostat’s setpoints. Then if something doubled the MESm setpoint, as an example, bone architecture would change until bones became half as strong as before. If something doubled the MESr, any bone previously conserved because it strained between 50 and 100 microstrain, would be removed. Elementary mechanioal considerations indicate most such losses should come from cortical-endosteal and trabecular bone instead of from subperiosteal bone. The mechanostat as u mediator. Bone must generate signals to control how the biologic systems fnt bone mass and architecture to our typical MU (Evans 1957; Becker et al. 1964; Frost 1964b). Streaming potentials and some related effects seem gcod candidates for those signals (Albright & Brand 1987; Chakkalakal 1989; Eriksson 1974; Gross & Williams 1982; Johnson 1984, Kufahl 8z Saha 1989; McLeod & Rubin 1990; Otter et al. 1985; Pienkowski 8~ Pollacks 1983; Pollacks et al. 1984, Takahashi 1990). The mechanostat’s relationship to the other parts of this system could look like this: MU + bone + signals + mechanostat

--, biologic systems

In the earlier analogy the mechanostat provides both thermostat functions, modeling provides the furnace, remodeling the air conditioner, and bone mass the heat. Modeling and re-

H. M. Frost: Osteoporosis

research

Anatomy of Tissue Dynamics of Osteopenias Disuse osteopenia

C :h ild

Fig. 2. These drawings show the configurations that bones should have when they have normal (N), increased (I), and decreased (D) mechanostat setpoints. Children’s bones above, adult bones below, and

steady state effects are assumed. When compared to normal, the lower middle drawing shows the common configuration found in chronic disuse, postmenopausal, and other kinds of osteoporoses, and in osteogenesis imperfecta.

modeling provide the tools the mechanostat uses to change bone mass. MU makes bone generate signals, the mechanostat “hears” them, compares them to its setpoints, and then tells the biologic mechanisms what to do to bone mass and architecture to correct any abnormality. An abnormality would exist when and where typical peak bone strains fell outside the adapted window, meaning below the MESr, or above the MESm. As many have realized, the resulting bone changes usually act in the negative feedback sense, for they tend to change bone mass and architecture in ways that bring typical peak strains falling outside the adapted window back within its MESr-MESm boundaries. By 1990 Burr (1990), Frost (199Oa,b), Jee (1990), Martin (1990), and Reeker (1990) realized the above evidence and ideas imply four things. First, when normal Mu continues, changes in the mechanostat’s setpoints should have the same effect on bone tissue dynamics, mass, and architecture as genuine MU changes. Second, bone mass changes due to MU changes should tend to plateau at a new level instead of continuing indefinitely. Third, increased setpoints should cause disuse responses and an osteopenia that plateaus at a new, lower level. Fourth, partial mechanical disuse should cause smaller increases in BMU creations and smaller deficits of formation than total disuse. Fig. 2 diagrams some effects that changed setpoints should have on bone mass and configuration.

This progresses through two stages: an initial transient one while bone adapts to the disuse, and the finished or steady state adaptation. The transient stage: When disuse begins bone strains would stay below 50 microstrain. Then BMU creations increase throughout a bone, while next to marrow completed BMUs form much less bone than they resorb. Those effects increase cortical porosity, expand the marrow cavity, thin trabeculae, and reduce their number. Outside bone diameter does not change. The right column of drawings in Fig. 1 shows the idea. The steady state: In chronic disuse and after enough bone is lost, BMU creations and cortical porosity decline towards normal, and resorption and formation in completed BMUs tend to equalize. Outside bone diameter stays normal and the previous trabecular losses and thinning persist. The left column of drawings in Fig. 1 shows the idea (Aegerter & Kirkpatrick 1975; Anderson & Kissane 1977; Arnold 1970, 1981; Arnold et al. 1966; Bogomil & Schwamm 1984; Frost 1985, 1986, 1987a; Heaney 1990~; Jaffe 1972; Jaworski & Uhthoff 1986; Jaworski et al. 1980; Jee 1989; Keshawarz & Reeker 1984; Kimmel et al. 1990; Kleerekoper & Krane 1990; Li & Jee 1990; Li et al. 1990, in press; Matsson 1972; Mellish et al. 1989; Meunier 1977; Meunier et al. 1974; Minaire 1973; Reeker 1983, 1989; Reeker et al. 1988; Riggs & Melton 1988; Smith et al. 1984; Smith & Gilligan 1989; Thompson & Rodan 1988; Turner & Bell 1986; Uhthoff 1986; Vedi et al. 1983; Weinreb & Rodan 1989; Wronski & Morey 1983a,b). Postmenopausal osteoporosis We can summarize these findings simply: The tissue-dynamic and structural changes during the perimenopausal (the transient) and late postmenopausal periods (the steady state) tend strongly to copy those of acute and chronic disuse. Table III lists these features for ready comparison (Gong et al. 1964; Keshawarz & Reeker 1984; Kleerekoper & Krane 1990; Lane et al. 1990; Riggs & Melton 1988; Steiniche et al. 1989; Storm et al. 1990; Uhthoff 1986). The same observations apply to many female 1aboratoIy animals (Anderson et al. 1990; Frame & Potts 1983; Jerome et al. 1986; Takahashi 1990; Wronski et al. 1989a,b,c; Wronski et al. in press).

A Synthesis and Proposal Synthesis In the message traffic between bone’s MU, any nonmechanical influences, and any resulting adaptations, the mechanostat should occupy the position shown here: MU + bone + signals + mechanostat

+(modeling-remodeling)

ft Nonmechanical

t agents

We look at two possible situations in that arrangement next. First, a therapeutic agent might change bone resorption or formation by acting directly on the osteoclasts or osteoblasts in the modeling or remodeling mechanisms, indicated by the single upwards arrow. That idea began with Albright (Albright & the thought and Reifenstein 1948), and it permeated

H. M. Frost: Osteoporosis

433

research

Table III. The osteopenias of anatomy

Outside diameter Cortical thickness Cortical porosity Cortical bank Trabecular number Connectivity Trabecular thickness Trabecular bank Ratio C/T loss BMU creations rho

of disuse and menopause and tissue dynamics to normals

Acute disuse*

Chronic disuse

N D

N D N D D D D D

A new direction for osteoporosis research: a review and proposal.

This article suggests why drugs that only reduce the activity of existing osteoclasts or enhance the activity of existing osteoblasts probably cannot ...
1MB Sizes 0 Downloads 0 Views