Bone architecture and disc degeneration in the lumbar spine of mice lacking GDF-8 (myostatin)

ELSEVIER

Journal of Orthopaedic Research

Journal of Orthopaedic Research 21 (2003) 1025-1032

www.elsevier.com/locate/orthres

Bone architecture and disc degeneration in the lumbar spine of mice lacking GDF-8 (myostatin) Mark W. Hamrick *, Catherine Pennington, Craig D. Byron Department of C e h l u r Biolog,v and Anatomy, Medical College of Georgia, L a n q Wlker Blods CB 2915, Augusta, G A 30912, USA

Received 24 January 2003; accepted 7 April 2003

Abstract

GDF-8, also known as myostatin, is a member of the transforming growth factor+ superfamily of secreted growth and differentiation factors that is expressed in vertebrate skeletal muscle. Myostatin functions as a negative regulator of skeletal muscle growth and myostatin null mice show a doubling of muscle mass compared to normal mice. We describe here morphology of the lumbar spine in myostatin knockout (Mstn-I-) mice using histological and densitometric techniques. The Mstn-1- mice examined in this study weigh approximately IO!h more than controls (p < 0.001) but the iliopsoas muscle is over 50% larger in the knockout mice than in wild-type mice (p < 0.001). Peripheral quantitative computed tomography (pQCT) data from the fifth lumbar vertebra show that mice lacking myostatin have approximately 50% greater trabecular bone mineral density (p = 0.001) and significantly greater cortical bone mineral content than normal mice. Toluidine blue staining of the intervertebral disc between L4-L5 reveals loss of proteoglycan staining in the hyaline end plates and inner annulus fibrosus of the knockout mice. Loss of cartilage staining in the caudal end plate of L4 is due to ossification of the end plate in the myostatin-deficient animals. Results from this study suggest that increased muscle mass in mice lacking myoslatin is associated with increased bone mass as well as degenerative changes in the intervertebral disc. 0 2003 Orthopaedic Research Society. Published by Elsevier Ltd. All rights reserved. K q w o r d x Muscle mass: Bone strength: Mechanical loading; Trabecular bone; Osteoarthritis

Introduction The forces imposed upon bones by muscles are significantly larger than those gravitational forces associated with body mass [6]. Studies of both humans and laboratory animals have documented a strong, positive correlation between muscle mass and bone mass [1,31,47,58]. These studies have, in turn, led many researchers to propose that increasing muscle strength, usually through vigorous physical exercise, is an effective preventative and therapeutic approach for increasing peak bone strength. It has therefore been suggested that increasing back muscle strength may increase the strength of lumbar vertebrae during growth and may in turn reduce the risk of vertebral fractures later in life [53,54]. There is, however, additional evidence to suggest that greater bone mineral density (BMD) is associated

*Corresponding author. Tel.: +I-706-721-1934; fax: +I-706-7216120. E - m d uddress: [email protected] (M.W. Hamrick).

with articular cartilage degeneration in the form of osteoarthritis [9,38,46,57]. GDF-8, hereafter referred to as myostatin, is a negative regulator of skeletal muscle growth and myostatin null mice have approximately twice the skeletal muscle mass of normal mice at both 2 and 10 months of age [35,36]. The effect of myostatin appears be dose-dependent, as mice heterozygous for the disrupted myostatin sequence have muscle weights that are intermediate between those of normal mice and mice homozygous for the myostatin mutation [37]. In situ hybridization data show that myostatin is first expressed in mouse embryos in the myotome compartment of somites and myostatin transcripts can still be detected in adults [29,35]. Myostatin expression during normal growth and development has, so far, only been detected in skeletal muscle although it has recently been observed in bone immediately following fracture [8]. Myostatin knockouts do not differ from normal mice (relative to body weight) in metabolic rate, food consumption, or body temperature [37]. Given that mice lacking myostatin are known to show increased muscle mass relative to normal mice, we

0736-0266/$ - see front matter 0 2003 Orthopaedic Research Society. Published by Elsevier Ltd. All rights reserved doi: 10.l016/S0736-0266(03)00105-0

I026

M. W. Hurnrick r t u1. I Journul of' Ortlioyuedic Ri,.vcwdi 21 (2003) 1025-1032

compare morphology of the lumbar spine between normal and myostatin knockout mice in order to investigate the effects of increased muscle mass on bone architecture and spinal disc morphology in the axial skeleton

Materials and methods &lJl/?/l'

The sample examined in this study includes 18 wild-type CD-I mice (Mstn- ' ) and 18 CD-I mice homozygous for the disrupted myostatin sequence (Mstn-. ). All mice were adult males 6 months of age at the time they were sacrificed by CO? overdose as approved by the Medical College of Georgia. Animal body weights were recorded immediately after sacrifice and the right iliopsoas muscle was then dissected free and weighed to the nearest 0.001 g. The spine was severed above (cranially) the first lumbar vertebrae and then across the sacrum. The articulated Ll-L5 spine was removed along with its musculature. fixed in 10% buffered formalin. and stored in 70% ethanol. The fifth lumbar vertebra (L5) of 10 normal mice and 10 knockout mice was dissected free and cleaned of all soft tissue [or peripheal quantitative computed tomography (pQCT) densitometry. Eight normal and eight knockout specimens did not have the L5 vertebra removed and were saved for histological analysis of the L4-L5 disc. ~ l ~ l l S ~ l ~U l~l dl l ?l i .lS O f ( J~l J~~ t~J l~. ~ ? l ? ( J I ~ ~ l ~ ~ ~ ~

A single scan. 1 mm thick, was taken in the sagittal plane along the cranio-caudal axis of the fifth lumbar vertebra using a Norland Stratec XCT-Research pQCT machine. The pQCT technique is known to be a very useful and reliable approach for measuring trabecular and cortical bone mineral content (BMC) and BMD in small animal bones [13,18]. Cross-sections were scanned at 4 mmls with a voxel size of 0.070 mm and a threshold value of 524.0 mg/cm3 used to distinguish trabecular from (sub)cortical bone. Variables collected include trabecular BILID. trabecular BMC, cortical BMD, and cortical BMC. Following pQCT scanning, fifth lumbar vertebrae were decalcified in 4'%1EDTA. washed, dehydrated, embedded in paraffin and sectioned at 5 pm using a rotary microtome. Sections were cut in the transverse plane through the vertebral body and then stained with hematoxylin and eosin to examine the morphology of individual trabeculae. Section images were captured using a Leica compound DMLS microscope with digital camera attachment and computer interface. A 1 mm2 measurement area was outlined from digital images of the caudal half of each vertebra using SigmaScan@ image analysis software. Trabecular area (Tb.Ar) and total tissue area (Tt.Ar) were measured from digital images and trabecular bone volume (BVITV) calculated as (Tb.Ar/Tt.Ar) * 100. Trabecular number (Tb.N) was measured directly by overlaying a single I mm transect line across the caudal end of each section and then counting the number of trabeculae intersecting the transect line. Trabecular thickness (Tb.Th) was calculated as Tb.Ar/ Tb.N and trabecular spacing = l/(Tb.N-Tb.Th). Standard bone histomorphometry nomenclature follows [40].

sured from toluidine blue stained specimens. A 0.10 mm' region of interest was superimposed on the center of the end plate, the growth plate underlying it, and the inner annulus fibrosus from one side. Staining intensity was quantified from these three regions using the acerage intensity option in Sigmascan". Images are converted to gray scale and the average intensity for any selected region of interest on the image is calculated as the sum of gray level values for all pixels in that region divided by the total number of pixels in that region. A value of 0 is pure black and a value of 255 is pure white. so that lower intensity values are associated with greater proteoglycan staining. Values were logged for statistical analysis and graphic display. MtJ~~?/l~J/lll~tl.J' und .Stflti3ilcU/ N?7U/J'.Si.S

Once the L5 vertebra was removed from 12 normal and knockout mice for densitometry the remaining articulated LI L4 vertebrae were cleaned of soft tissue, prepared for whole-mount clearing and staining using alizarin red, and radiographed at 35 kVP and 2.5 mA for 15 s using a Faxitron X-ray. Whole-mount cleared and stained vertebral columns were viewed in dorsal and lateral aspect using an Olympus stereo light microscope. The following four measurements were collected from each vertebra using the microscope's reticle: dorsoventral height of the spinous process, mediolateral diameter of the transverse processes, crdniocaudal length of the vertebral body, and mediolateral diameter of the vertebral body. Densitometric, histomorphometric. and morphometric variables were compared between normal and knockout mice using ANOVA with genotype as the factor.

Results

Metric dimensions and gross morplzolog,y The myostatin knockout mice are slightly larger than the normal mice but their iliopsoas muscles are approximately 50% larger than the iliopsoas muscles of their normal counterparts (Table 1). The similarity in adult body size is due to the fact that the knockout mice do not accumulate fat like normal mice [37]. The bony attachment sites for muscles of the spine are also expanded in the knockout mice. The spinous processes on L1-4 tend to be larger in the knockout mice compared to the normals and the transverse processes on LlLL2 are significantly broader in the knockout mice to accommodate quadratus lumborum (Table 2). Greater Table 1 Summary statistics [mean (standard deviation)] for body mass, iliopsoas mass, and densitometric properties of fifth lumbar vertebra in control (+/+) and GDF-8 deficient (-/-) mice Parameter

I l i t r o I o ~ unti j ~ iniuge uriu/j~,si.soJ tI7e in rercertchr.al cli.sc

Control GDF-8 defi(+I+) cient (-/-) (n = 10) ( n = 10)

' 4 Dif-

p

ference ~

Eight specimens from each group (wild-types and knockouts) preserving the L4-LS disc were prepared for histological analysis. All of these specimens were decalcified in 4% EDTA and then prepared for sectioning and staining following standard procedures for dehydration. clearing, and pal-afin embedding described by Humason [28]. Speciinens were sectioned at 5 pm using a rotary microtome and alternate sections of these specimens stained with toluidine blue to demonstrate proteoglycans in the articular disc and hyaline end plates. Sections were also stained with hematoxylin and eosin to demonstrate bone structure. Section images were captured using a Leica compound DMLS microscope with digital camera attachment and computer interface. Staining intensity ofthe hyaline end plate on the caudal end of L4. the annulus fibrosus. and the L4 caudal growth plate were mea-

Body mass (g) Iliopsoas mass

(s)

Cortical BMD (mglcm') Cortical BMC (mg/mm) Trabecular BMD (mg/cm3) Trabecular BMC ( mg/mm)

47.1 (3.7) 0.15 (0.02)

52.3 (4.1) II 0.23 (0.05) 53

0.001 0.001

762.8 (2.24)

762.2 (1.4)

0

0.55

41.9 (0.35)

44.4 (1.6)

6

0.001

142.8 (50.8)

218.2 (58.5)

52

0.005

1.8 (0.42)

2.35 (0.77) 30

0.07

M. W. liunirick

ei

rrl. I Jolrrnul

of

Orthopacdic Resiwch 21 (2003) 1025-1032

I027

Table 2 Summary statistics [mean (standard deviation)] for vertebral dimensions (mm) of first through fourth (LILL4)lumbar vertebra in control (+/+) and GDF-8 deficient (-/-) mice. Values for the wild-types are placed above the values for the knockouts in each column

-

p

**

Parameter

L1

L2

L3

L4

Spinous process height

+/+: 1.24(0.23) -I-: 1.49(0.34)

1.54 (0.46) 1.71 (0.38)

1.54(0.47) 1.71 (0.38)

1.96 (0.40) 2.22(0.52)

Vertebral body length

2.81 (0.26) 2.84 (0.40)

2.87(0.26) 3.01 (0.22)

2.87 (0.26)' 3.01 (0.22)

3.14(0.25)*' 3.41 (0.24)

Vertebral body width

2.38 (0.34) 2.28 (0.12)

2.38 (0.27) 2.23(0.07)

2.38 (0.27) 2.23 (0.07)

2.19 (0.1I ) 2.25 (0.19)

Transverse process diameter

3.27 (0.23)" 4.19 (0.37)

3.94 (0.52)' 4.51 (0.31)

4.19 (0.42) 4.39 (0.38)

4.41 (0.37) 4.49(0.19)

< 0.0I . <

1) 0.05

5 mm Fig. I . Radiographs of the lumbar spine (LILL4)in normal (+/+; top row) and myostatin-deficient (-/-; bottom row) mice showing the increased vertebral bone density and decreased vertebral spacing (arrows) of the myostatin knockout mice.

muscling in the knockout mice appears to not only increase the size of bony muscle attachment sites but to also increase size of the vertebral bodies, as the vertebral bodies of L3-4 are significantly longer in the knockouts compared to the normals (Table 2, Fig. 1). This increase in length of the vertebral bodies is also observed in L5, which in part explains the observed increase in cortical and trabecular BMC (see below). The normal mice do not differ from the knockouts in breadth of the vertebral body (Table 2). Bone ii'ensitometrjj and Izistomorphometry

pQCT scans through L5 indicate that the knockout mice have significantly greater cortical BMC and trabe-

cular BMD than the normal mice (Table 1; Fig. 1). In the case of trabecular BMD, the GDF-8 knockouts show an increase of over 50%1compared to the normal mice. Visual inspection of pQCT images indicates that the increased trabecular BMD of the knockout mice is due to increased bone density in the caudal regions of the vertebral body and in the ventral regions of the spinous process. Sections through L5 demonstrate that the knockout mice show an increase in trabecular bone volume of approximately 30%, and this increase is due primarily to an increase in trabecular thickness of approximately 50% (Table 3, Fig. 2). It should be noted that trabecular number in the vertebral bodies of the knockout mice does not differ significantly from that of the normals, revealing that the increased trabecular

M . W. Hamrick rt al. / Journal

1028

of‘ Orlhopaedic

Table 3 Summary statistics [mean (standard deviation)] for trabecular histomorphometry measurements of the fifth lumbar vertebra in control (+I+) and GDF-8 deficient (-/-) mice Parameter

Trabecular bone volume (%I) Trabecular number (mm) Trabecular thickness (mm) Trabecular spacing (mm)

Control

GDF-8 deficient (-/-) (n = 10)

‘YOdiffer-

(+I+) ( n = 10)

24.0 (5.0)

31.0 (5.0)

30

0.007

p

ence

6.0 (1.15)

5.3 (0.94)

11

0.15

0.04 (0.01)

0.06 (0.01)

50

0.001

0.17 (0.03)

0.19 (0.03)

12

0.18

bone of myostatin deficient mice results primarily from an increase in the size of individual trabeculae rather than from formation of additional trabeculae (Table 3). The significant increase in cortical BMC and trabecular BMD observed in the knockout mice shows a strong, positive association with increased muscle mass. Cortical BMC is more highly correlated with ilipsoas mass (Y = 0.66, p < 0.001) than with body mass ( y . = 0.31, p = 0.43) and trabecular BMD is also more highly correlated with iliopsoas mass (Y = 0.84, p < 0.001) than with body mass (Y = 0.40, p = 0.06). Morphology of the intervertebral disc and hyaline end plate

Histological sections through the L4-5 disc stained with toluidine blue show marked reduction in proteoglycan staining of the hyaline end plates in the knockout

Rrsrurch 21 (2003) 1025-1032

mice (Fig. 3). Sections stained with hematoxylin and eosin reveal that the loss of toluidine blue staining in these end plates results from ossification of the cartilage that normally forms the end plate (Fig. 3). Mineralization of the hyaline end plates in the knockout mice is also associated with loss of proteoglycan staining in the inner annulus fibrosus (Fig. 3). The inner annulus is normally rich in proteoglycans, which are hydrophilic and play an important role in binding water within the disc. Quantitative analysis of staining intensity of the Iiyaline end plate and annulus fibrosus show that the knockout mice differ significantly from the normal mice in showing significant loss of toluidine blue staining (“brighter” sections and higher intensity values) compared to normal mice (Fig. 4). Note also that the knockout mice are similar in proteoglycan staining of the caudal growth plate on L4, which is darkly stained in both normals and knockouts, indicating that the differences between the two groups in disc staining are not artifacts of staining preparation between groups.

Discussion

A strong correlation between whole-bone strength and muscle strength was predicted by D’Arcy Thompson [56] and is now supported by a number of more recent studies, primarily analyses of BMD and muscle cross-sectional area (CSA) among athletes [%I. The forces imposed upon bones by muscles are much larger than those gravitational forces associated with body mass, suggesting that muscle strength should be a primary determinant of peak bone strain and therefore a

+ +

I .

I . I

-

. I 0 rnrn

Fig. 2. Morphology of trabecular bone in the fifth lumbar vertebra (L5) of normal (+/+; top row) and myostatin-deficient (-/-; bottom row) mice illustrating the greater trabecular area fraction of the knockout mice. Trabecula = tr and growth cartilage = gc.

M. W. Humrick ef al. I Journal of Orthopedic Reseurch 21 (2003) 1025-1032

1029

+ +

2.

.10 mm

.10 mm

(CI

.10 mm

. I 0 mm

Fig. 3 . Histological sections of the intervertebral disc between L4-L5 illustrating microanatomy of the end plate (ep) and annulus fibrosis ( a n in normal (+I+: left column) and myostatin-deficient (-/-; right column) mice. Sections stained with toluidine blue (a) demonstrate strong proteoglycan staining the end plate of the normal mouse but not in the knockout. Sections stained with hematoxylin and eosin (b) show that the end plate of the normal mouse Is composed of articular cartilage (ac) whereas it is composed of bone (bn) in the knockout mouse. Additional sections stained with toluidine blue show strong proteoglycan staining in the inner annulus fibrosus (af-i) of the normal mouse and comparatively weaker staining of the inner annulus in the knockout mouse. Growth cartilage = gc, Outer annulus fibrosus = af ( 0 ) .

significant factor in determining bone strength [6]. In fact, Schoneau and Frost [48] argue that the largest mechanical loads on bones dominate the postnatal development of bone strength, and the largest loads on bones are in turn proportional to muscle mass [16,17]. This led Schoneau et al. [49] to define the “bone-muscle unit” (BMC/muscle CSA) as an important functional tool for recognizing potential imbalances between bone strength and muscle strength in children and adolescents. Densitometry data from the humerus [23], femur [24], and now from the spine suggest that the increased muscle mass of myostatin-deficient mice is also associated with increased bone mass. The question that arises from data presented here concerns the niechanism(s) underlying the increased cortical BMC and trabecular BMD in the lumbar spine of myostatin-deficient mice. Increased mechanical loading due to increased muscle strength has traditionally

been used to explain the positive correlation between muscle mass and BMD. It is, however, difficult to believe that a 50 g mouse could subject its spine to enough force to stimulate increased bone formation in the vertebral bodies, particularly since mice are quadrupedal and are therefore transmitting most of their weight through their fore- and hindlimbs rather than through the lumbar vertebrae. An alternative explanation is that myostatin has a direct effect on bone metabolism. As noted earlier, myostatin expression during normal growth and development has only been detected in skeletal muscle although it has recently been observed in bone immediately following fracture [8]. The role of myostatin in fracture repair is as yet unknown, but perhaps it has an inhibitory effect on osteoblast or osteoclast differentiation so that loss of myostatin function may either increase bone formation or decrease bone resorption, respectively. It is also possible that endocrine

M . W. Humrick rt ul. / Journril of Orthopaivbc Reseur.ch 21 (2003) 1025-1032

1030

P=.O1

P=.03

T

5.5

Growth Cartilage

End Plate

Annulus fibrosus

Fig. 4. Histogram showing pixel intensities measured from toluidineblue stained sections of the caudal L4 end plate, growth cartilage, and annulus fibrosus. The higher intensity values for the end plate and annulus in the knockout mice (-I-; solid bars) compared to the normal mice (+/+; shaded bars) indicate lighter staining (brighter pixels) in the knockouts versus greater proteoglycan staining (darker pixels) in the normal mice. Error bars equal 1 standard deviation.

changes due to increased muscle mass and decreased fat mass, such as a decline in serum leptin, may better explain the increased vertebral bone mass of myostatin deficient mice (see below). The increased size of the spinous processes in the knockout mice is not surprising given our previous findings showing that bony sites of muscle insertion on both the femur [22,24] and humerus [23] are expanded in mice lacking myostatin. The increased size of these processes is also likely to contribute to the increased cortical and trabecular BMC observed in the fifth lumbar vertebra of the knockout mice. The spinous processes form late in skeletal development and ossification does not begin until two postnatal weeks in mice [30]. The postnatal growth of these bony structures is strongly influenced by functional and epigenetic processes, illustrated by the fact that length of the spinous processes is highly variable both within and among inbred mouse strains [39]. The tendons attaching to the spinous processes are continous with the epimysium surrounding the spinalis muscle, so that as the muscle fibers increase in size and number the epimysium and the tendon must also expand to accommodate the increase in muscle mass. Collagen fibers of the tendon are in turn incorporated into periosteum of the spinous process. Muscle growth produces stretching of collagen fibers and periosteum at the soft tissuelbone interface, which is believed to stimulate local periosteal growth [25]. The increased size of the spinous processes in the myostatin knockout mice therefore reflects the integrated nature of muscle, tendon, and periosteal growth at the tendoosseous junction. The increased trabecular bone density in the lumbar spine of myostatin knockout mice is associated with

degenerative changes in the intervertebral disc. These changes include mineralization of the hyaline end plate (spinal osteoarthritis, or lumbar spondylosis) and loss of proteoglycans in the inner annulus fibrosis. Osteoarthritis (OA) of the lumbar spine is a very common form of OA and increasing evidence suggests that there is a positive association between increased bone mass and osteoarthritis [9, lo]. This association appears to be especially strong in the case of spinal OA. For example, Hordon et al. 1261 found that women with generalized OA had significantly increased BMD compared to normal controls. Peel et al. [41] and El Miedany et al. [I21 also concluded that spinal OA is associated with increased spinal BMD and decreased bone turnover, and Simpson et al. [52] reported that degenerative changes in the lumbar disc were strongly correlated with greater trabecular bone thickness and increased trabecular number. Although the relationship may not hold for other regions such as the ankle [38], a strong correlation between increased BMD and cartilage degeneration appears to exist for both the spine and the hip [4,12,20]. It is unclear what biological process explains this relationship, although multiple mechanisms are sure to exist. Bone remodeling due to increased loading is argued to increase density of the subchondral bone [43]. Increased trabecular and subchondral bone density is, in turn, believed to increase bone stiffness [7,19]. Stiffening of the trabecular and subchondral bone in joint regions increases the level of stress experienced by articular cartilage [44,45], and very large transarticular forces decrease the ability of cartilage to perform its normal metabolic and anabolic functions [21]. Hence, increasing bone stiffness and thereby increasing the stress in articular cartilage may contribute to the onset of articular cartilage degeneration and osteoarthosis [43,44,511. It is also likely that the balance between bone formation and articular cartilage metabolism is somehow disrupted in individuals with osteoarthritis [3,10]. Perhaps the best evidence for such an imbalance comes from mice with impaired TGF-P signaling. Decreased TGF-P signaling in osteoblasts causes an age-dependent increase in trabecular bone mass [ 141 and periarticular cartilage undergoes terminal differentiation and ossification in the absence of TGF-P [50]. In the case of the myostatin knockout mice described here, leptin deficiency may be one endocrine mechanism that explains the cartilage degradation observed in the spines of these animals. Leptin is a hormone secreted by adipocytes and myostatin knockout mice have very low serum leptin levels due to their low fat mass [33,37]. Leptin is now known to act as a growth factor for chondrocytes [ 15,341. Leptin deficiency in myostatin knockout mice would therefore be expected to inhibit chondrocyte proliferation and perhaps cartilage repair.

M . W. HLiiwick et

til.

I Jvurnul of’ Orrhopuedic R c s ~ L ~21 ~ c(2003) I~ 1025-1032

This idea is supported by the fact that the femoral heads of adult myostatin deficient mice are smaller than those of normal mice, and articular cartilage from the proximal femur of young myostatin-deficient animals shows fewer proliferating chondrocytes than cartilage from age-matched controls. Leptin deficiency in mice has also been associated with increased BMD in the spine [ll], providing a possible link between leptin deficiency, increased trabecular BMD, and spinal osteoarthritis. We are currently investigating the growth and aging of bone and articular cartilage in leptin knockout mice to further clarify the role of leptin deficiency in the development of osteoarthritis. Results from this study also shed new light on the pathogenesis of lumbar disc degeneration. The Kirkaldy-Willis degenerative cascade [32] is probably the most well-known model for degenerative changes in the lumbar spine. The cascade involves three phases, termed dysfunctional (phase I), unstable (phase 2), and stable (phase 3) phases. Phase 1 is initiated by microscopic tears or fissures in the annulus fibrosus that result either from a single traumatic event or from repetitive microtrauma. The final phase, phase 3, is characterized by destruction of the hyaline end plates. Pritzker’s [42] study on aging and degeneration in the lumbar intervertebral disc identified focal changes in the cartilaginous end plates preceding any changes in the nucleus and annulus. These changes include mineralization and ossification of the end plate leading to reduction of the disc space and bulging of the annulus fibrosus [42,55]. Calcification and ossification of the end plates contributes directly to degeneration of the nucleus pulposus because end plate mineralization prevents the diffusion of nutrients across the end plate from capillaries of the vertebral body [2,5,27]. Results presented in this paper indicate that ossification of the end plate cartilage is associated with loss of proteoglycans in the inner annulus fibrosus, suggesting that end plate mineralization may indeed be a primary factor in later degenerative changes in the disc tissues. As noted above, these degenerative changes are also associated with increased trabecular bone density, providing further evidence for a relationship between increased spinal BMD and lumbar osteoarthritis.

Acknowledgements We are grateful to Dr. Gary Schneider, NEOUCOM, for access to the pQCT laboratory facility under his supervision and to K. Grecco for helpful assistance in collecting the pQCT data. Drs. Alexandra McPherron and Se-Jin Lee provided helpful advice and assistance with the myostatin knockout mice. Funding for this research was provided by the Medical College of

1031

Georgia and a grant from the National Institutes of Health to MWH (AR47655-01).

References [I] Andreoli A, Monteleone M, Van Loan M. Promenrio L, Tarantino U, De Lorenzo A. Effects of different sports on bone density and muscle mass in highly trained athletes. Med Sci Sports Exerc 2001:33:507-11, Brodin H. Paths of nutrition in articular cartilage and intervertebrdl discs. Acta Orth Scand 19SS:24:177- 83. Bruckner R, Dreier R. Grassel S, Babarina A. Common denominators of osteoarthritis and of endochondral bone formation: is osteoarthritis nothing but an inappropriate attempt at bone formation. Z Rheumatol 2002;61:263-5. Bruno RJ, Sauer PA, Rosenberg AG, Block J , Sumner DR. The pattern of bone mineral density in the proximal femur and radiographic signs of early joint degeneration. J Rheumatol 1999; 26~63640. Buckwalter JA. Aging and degeneration of the human intervertebral disc. Spine 1995;20:1307-14. Burr DB. Muscle strength, bone mass, and age-related bone loss. J Bone Miner Res 1997;12:1547-51. Carter D R , Hayes W. Bone compressive strength: the influence of density and strain rate. Science 1976;194:1174-5. Cho T, Gerstenfeld LC, Einhorn TA. Differential expression of members of the transforming growth factor I3 superfamily during murine fracture healing. J Bone Miner Res 2002; 17:5 13-20, Dequeker J. Inverse relationship of interface between osteoporosis and osteoarthritis. J Rhematol 1997;24:795-8. Dequeker J, Mokassa L, Aerssens J, Boonene S. Bone density and local growth factors in generalized osteoarthritis. Microsc Res Tech 1997:37:358-71. Ducy P, Amling M, Takeda S, Priemel M, Schilling A, Beil F, et al. Leptin inhibits bone formation through a hypothalamic relay: a central control of bone mass. Cell 2000:100:197-207. El Miedany YM, Mehanna AN, El Baddini MA. Altered bone mineral metabolism in patients with osteoarthritis. Joint Bone Spine 2000;67:S2 1-7. Ferretti JL. Perspectives of pQCT technology associated to biomechanical studies in skeletal research employing rat models. Bone 1995;17(Suppl):353S-64S. Filvaroff E, Erlebacher A, Ye J, Gitelman SE, Lotz J, Heillman M, et al. Inhibition of TGF-P receptor signaling in osteoblasts leads to decreased bone remodeling and increased trabecular bone mass. Development 1999;126:4267-79. Figenschau Y, Knutsen G, Shahazeydi S, Johansen 0, Sveinbjornsson B. Human articular chondrocytes express functional leptin receptors. Biochem Biophys Res Comm 2001:287: 190- 7. Frost HM. Perspective: on our age-related bone loss: insights from a new paradigm. J Bone Miner Res 1997; 12: 153946. Frost HM. Changing concepts in skeletal physiology: Wolffs law, the niechanostat, and the “Utah Paradigm”. Am J Hum Biol 1998;10:599-605. Gasser JA. Assessing bone quantity by pQCT. Bone 1995;17: 145s-54s. Gibson L. The mechanical behavior of cancellous bone. J Biom 1985;18:317-28. Goker B, Sumner DR. Hurwitz DE, Block JA. Bone mineral density varies as a function of the rate of joint space narrowing in the hip. J Rheumatol 2000;27:735 -8. Hamrick MW. A chondral modeling theory revisited. J Theor Biol 1999;201:201 8.

1032

M. W. Humrick et ul. I Journal of Orthopuerlir Reseurch 21 (2003) 1025-1032

[22] Hamrick MW, McPherron AC, Lovejoy CO, Hudson J. Femoral morphology and cross-sectional geometry of adult myostatindeficient mice. Bone 2000:27:343-9. 1231 Hamrick MW, McPherron AC, Lovejoy CO. Bone mineral content and density in the humerus of myostatin-deficient mice. Calcif Tissue Intl 2002;71:63-8. [24] Hamrick MW. Increased bone mineral density in the femora of GDF8 knockout mice. Anat Rec 2003:272A:388-91. 1251 Herring SW. Development of functional interactions between skeletal and muscular systems. In: Hall BK, editor. Bone, vol. 9. Boca Raton: CRC Press; 1994. p. 165-91. [26] Hordon LD, Stewart SP, Troughton PR, Wright V, Horsman A, Smith MA. Primary generalized osteoarthritis and bone mass. Br J Rheumatol 1993;32:1059-61. [27] Horner HA, Urban JP. Effect of nutrient supply on the viability of cells from the nucleus pulposus of the intervertebral disc. Spine 2001 ;26:2543-9: [28] Humason GL. Animal tissue techniques. San Francisco: W.H. Freeman: 1972. [29] Ji S, Losinski R, Cornelius S, Frank G , Willis G, Gerrard D, et al. Myostatin expression in porcine tissues: tissue specificity and developmental and postnatal regulation. Am J Physiol 1998; 275:R 1265-73. [30] Johnson ML. The time and order of appearance of ossification centers in the albino mouse. Am J Anat 1937;52:241-71. [31] Kaye M, Kusy RP. Genetic lineage, bone mass, and physical activity in mice. Bone 1995;17:131-5. [32] Kirkaldy-Willis WH. The pathology and pathogenesis of low back pain. In: Kirkaldy-Willis WH, Bernard Jr TR, editors. Managing low back pain. NY: Churchill-Livingstone; 1998. p. 49-60. 1331 Lin J, Arnold HB, Della-Frea M, Azain M , Hartzell D, Baile C. Myostatin knockout in mice increases rnyogenesis and decreases adipogenesis. Biochem Biophys Res Commun 2002;291:701-6. [34] Maor G, Rochwerger M, Segev Y, Phillip M . Leptin acts as a growth factor on the chondrocytes of skeletal growth centers. J Bone Miner Res 2002;17:103443. [35] McPherron AC, Lawler AM, Lee S-J. Regulation of skeletal muscle mass in mice by a new TGF-O superfamily member. Nature 1997;387:83-90. [36] McPherron AC, Lee S-J. Double muscling in cattle due to mutations in the myostatin gene. Proc Natl Acad Sci 1997; 94: 12457-6 1. [37] McPherron AC, Lee S-J. Suppression of body fat accumulation in myostatin-deficient mice. J Clin Invest 2002; 109:595-601, [38] Muehleman C, Berzins A, Koepp H, Eger W, Cole A, Kuettner K, et al. Bone density of the human talus does not increase with the cartilage degeneration score. Anat Rec 2002;266:81-6. I391 OHiggins P, hlilne N, Johnson DR, Runnion CK, Oxnard CE. Adaptation in the vertebral column: a comparative study of patterns of metameric variation in mice and men. J Anat 1997;190:105-1 3. [40] Parfitt AM, Drezner MK, Glorieux FH, Kanis JA, Malluch H , Meunier PJ, et al. Bone histomorphometry: standardization of nomenclature, symbols, and units. J Bone Miner Res 1987;2:595610.

[41] Peel N , Barrington N , Blurnsohn A, Colwell A, Hannon R, Eastell R. Bone mineral density and bone turnover in spinal osteoarthrosis. Ann Rheum Dis 1995;54:867-71. [42] Pritzker KP. Aging and degeneration in the lumbar intervertebral disc. Orthop Clin North Am 1977;8:6&77. [43] Radin EL, Parker H, Pugh J, Steinberg R, Paul I, Rose R. Response of joints to impact loading. 111. Relationship between trabecular microfractures and cartilage degeneration. J Biom 1973:6:51-7. [44] Radin EL, Martin R, Burr D , Caterson B, Boyd R, Goodwin C. Effects of mechanical loading on the tissues of the rabbit knee. J Orthop Res 1984;2:221-34. [45] Radin EL, Burr D. Caterson B, Fyhrie D, Brown T, Boyd R. Mechanical determinants of osteoarthrosis. Sem Arth Rheum I99 1;21: 12-21. [46] Sambrook P, Naganathan V. What is the relationship between osteoarthritis and osteoporosis? Bail1 Clin Rheumatol 1997;1 I : 695-710. [47] Schoenau E, Werhan E, Schniedermaier U, Mokow E, Schiessl H, Schneidhauer K , et al. Influence of muscle strength on bone strength during childhood and adolescence. Horm Res 1996; 45:63-6. [48] Schoenau E, Frost HM. The “muscle-bone unit” in children and adolescents. Calcif Tissue Int 2002;70:405-7. [49] Schoenau E, Neu C, Beck B, Manz F, Rauch F. Bone mineral content per muscle cross-sectional area as an index of the functional muscle-bone unit. J Bone Miner Res 2002:17: 1095101. [50] Serra R, Johnson M, Filvaroff E, LaBorde J, Sheehan D, Derynck R, et al. Expression of a truncated, kinase-defective TGF-P type I1 receptor in mouse skeletal tissue promotes terminal chondrocyte differentiation and osteoarthritis. J Cell Biol 1997;139:541-2. [51] Simon S, Radin E. The response of joints to impact loading. 11. In vivo behavior of subchondral bone. J Bioni 1972;5:267-72. [52] Simpson EK, Parkinson IH, Manthey B, Fazzalari NL. Intervertebral disc disorganization is related to trabecular bone architecture in the lumbar spine. J Bone Miner Res 2001:16:681-7. [53] Sinaki M, McPhee M, Hodgson S, Offord K . Relationship between bone mineral density of spine and strength of back extensors in healthy post-menopausal women. Mayo Clin Proc 1986;6l:116-22. [54] Sinaki M, Wollan P, Scott R, Gelczer R. Can strong back extensors prevent vertebral fractures in women with osteoporosis? Mayo Clin Proc 1996;71:951-6. [55] Taylor JR, Giles L. Lumbar intervertebral discs. In: Taylor Giles L, Giles Singer K , editors. Clinical Anatomy and management of low back pain, vol. I . Oxford: Butterworth Heinemann; 1997. p. 49-71. [56] Thompson DW. On growth and form. London: Cambridge University Press; 1917. [57] Turner CH. Exercise as a therapy for osteoporosis: the drunk and the street lamp revisited. Bone 1998:23:83-5. [58] Witzke KA, Snow CM. Lean body mass and leg power best predict bone mineral density in adolescent girls. Med Sci Sports Exerc 1999;31:1558-63.