A, Dequeker J, Mosekilde L, Einhorn TA, Bouillon R 1996. Effects of recombinant ... Stefansson S, Lawrence DA 1996 The serpin PAI-1 inhibits cell migration by ...
JOURNAL OF BONE AND MINERAL RESEARCH Volume 15, Number 8, 2000 © 2000 American Society for Bone and Mineral Research
Mice Lacking the Plasminogen Activator Inhibitor 1 Are Protected from Trabecular Bone Loss Induced by Estrogen Deficiency E. DACI, A. VERSTUYF, K. MOERMANS, R. BOUILLON, and G. CARMELIET
ABSTRACT Bone turnover requires the interaction of several proteases during the resorption phase. Indirect evidence suggests that the plasminogen activator/plasmin pathway is involved in bone resorption and turnover, and recently we have shown that this cascade plays a role in the degradation of nonmineralized bone matrix in vitro. To elucidate the role of the plasminogen activator inhibitor 1 (PAI-1) in bone turnover in vivo, bone metabolism was analyzed in mice deficient in the expression of PAI-1 gene (PAI-1ⴚ/ⴚ) at baseline (8-week-old mice) and 4 weeks after ovariectomy (OVX) or sham operation (Sham) and compared with wild-type (WT) mice. PAI-1 inactivation was without any effect on bone metabolism at baseline or in Sham mice. However, significant differences were observed in the response of WT and PAI-1ⴚ/ⴚ mice to ovariectomy. The OVX WT mice showed, as expected, decreased trabecular bone volume (BV/TV) and increased osteoid surface (OS/BS) and bone formation rate (BFR), as assessed by histomorphometric analysis of the proximal tibial metaphysis. In contrast, no significant change in any of the histomorphometric variables studied was detected in PAI-1ⴚ/ⴚ mice after ovariectomy. As a result, the OVX PAI-1ⴚ/ⴚ had a significantly higher BV/TV, lower OS/BS, lower mineral apposition rate (MAR) and BFR when compared with the OVX WT mice. However, a comparable decrease in the cortical thickness was observed in OVX PAI-1ⴚ/ⴚ and WT mice. In addition, the cortical mineral content and density assessed in the distal femoral metaphysis by peripheral quantitative computed tomography (pQCT), decreased significantly after ovariectomy, without difference between PAI1ⴚ/ⴚ mice and WT mice. In conclusion, basal bone turnover and bone mass are only minimally affected by PAI-1 inactivation. In conditions of estrogen deficiency, PAI-1 inactivation protects against trabecular bone loss but does not affect cortical bone loss, suggesting a site-specific role for PAI-1 in bone turnover. (J Bone Miner Res 2000;15:1510 –1516) Key words:
plasminogen activator inhibitor 1, mice, ovariectomy, bone turnover, bone histomorphometry
INTRODUCTION ONE REMODELING continues throughout life and involves the resorption of mineralized matrix by the osteoclasts followed by new bone formation by the osteoblasts. This process is controlled by several factors, including systemic hormones, growth factors, cytokines, and proteolytic systems. One of the proteolytic cascades included in the latter
B
group is the plasminogen activator/plasmin system. The plasminogen activator/plasmin pathway results in the generation of plasmin from the zymogen plasminogen. This activation is induced by the two plasminogen activators, tissue-type plasminogen activator and urokinase-type plasminogen activator and is inhibited by the plasminogen activator inhibitor 1 (PAI-1).(1) Several observations have suggested the involvement of this pathway and its inhibitor
Laboratorium voor Experimentele Geneeskunde en Endocrinologie, Katholieke Universiteit Leuven, Leuven, Belgium.
1510
PAI-1 AND OVARIECTOMY-INDUCED BONE LOSS
PAI-1 in bone metabolism. PAI-1 is produced by bone cells,(2,3) and treatment of osteoblast-like cell cultures with bone resorbing hormones down-regulates PAI-1 production.(4) This effect, together with the regulation of the production of the plasminogen activators,(5) results in a marked increase in the plasminogen activator activity.(6) Plasmin itself is a broad spectrum serine protease able to degrade several noncollagenous components of the extracellular matrix including fibronectin, laminin, and the protein core of the proteoglycans.(7) In addition, plasmin is also an important activator of latent matrix metalloproteinases(8,9) and may in this way participate indirectly in collagen breakdown. In agreement with these proposed activities, recently, we have shown that the plasminogen activator/plasmin pathway is involved in the degradation of the nonmineralized bone-like matrix in vitro but not in the resorption of the mineralized matrix.(10) Degradation of the organic matrix of bone is an important component of the resorption process, because it exposes bone mineral to the osteoclasts so that mineral resorption can start, or it removes the remaining organic matrix after the solubilization of the mineral phase. In addition, several growth factors such as transforming growth factor  or insulin-like growth factors are stored in the bone matrix bound to their binding proteins and are released during the resorption process. Plasmin can dissociate these growth factors from their binding proteins and activate them.(11–13) This implicates that plasmin is involved indirectly in bone formation and in the coupling of bone formation to resorption. Taken together, these data suggest a role for the plasminogen activator/PAI-1 system in bone resorption and bone turnover. However, Leloup et al.(14,15) have shown that in organ cultures of bone, the resorption process is largely independent of the plasminogen activator/plasmin function. Noteworthy, PAI-1– deficient explanted fetal mouse metatarsals and calvariae showed a smaller response to bone resorbing hormones.(15) Despite these data from in vitro studies, the role of PAI-1 in bone turnover in vivo is not yet defined. In the present study, the role of PAI-1 in bone metabolism was investigated using mice deficient in the expression of the PAI-1 gene. Because PAI-1 inactivation has little effect on other systems in basal conditions,(16) bone metabolism was studied in a model of increased bone turnover and bone loss resulting from estrogen deficiency. We report here that PAI-1 deficiency has only minimal effects on basal bone turnover and bone mass. Surprisingly, PAI-1⫺/⫺ mice are protected from trabecular bone loss induced by estrogen deficiency, whereas cortical bone loss proceeds independently of the PAI-1 gene function.
MATERIALS AND METHODS Animals PAI-1 knockout mice, generated via homologous recombination in embryonic stem cells,(16) were obtained from P. Carmeliet, Center for Molecular and Vascular Biology, University of Leuven, Belgium, and bred in our animal housing facilities (Proefdierencentrum Leuven, Belgium). Animals
1511
were fed a normal diet and had free access to tap water. Deletion of the genomic sequences was shown by Southern blot analysis of genomic tail-tip DNA. In all experiments age and gender-matched animals of the same genetic background were used as controls.
Study protocol Three protocols using female PAI-1– deficient and wildtype (WT) mice were used: a) death of 8-week-old mice for baseline determinations, b) ovariectomy, or sham operation at 8 weeks of age, and c) ovariectomy or sham operation at 13 weeks of age. Ovariectomy or sham operation was performed under intraperitoneal (ip) pentobarbital anesthesia (100 mg/kg) using the dorsolateral approach and the mice were killed 4 weeks later. All animals received two ip injections of the fluorochrome calcein (16 mg/kg; Sigma Chemical Co., St. Louis, MO, U.S.A.) 4 days and 24 h before death. Twenty-four– hour urine collections were obtained by placing the animals in metabolic cages once before the intervention and the second time before the first calcein injection. At death, the animals were weighed and blood was taken by heart puncture under ether anesthesia. The uterus was isolated and the wet uterine weights were measured to confirm the effect of the intervention. The tibiae, femora, and humeri were dissected and cleaned of surrounding tissue. The left tibia was immersed in Burkhardt’s fixative for 24 h at 4°C and subsequently kept in ethanol 100% until histomorphometric analysis. The left femur was immersed in Burkhardt’s fixative and subsequently was kept in ethanol 70% until bone density and mineral content were measured by peripheral quantitative computed tomography (pQCT). The right femur and humerus were frozen at ⫺20°C and used for chemical analysis (ash weight and calcium content).
Serum concentration of calcium Calcium in serum was determined by microcolorimetric assay (Sigma Chemical Co.).
Calcium and collagen cross-links in urine Urine was collected during 24 h in 10-ml glass tubes and 10-l 6N HCl was added. Calcium was determined using the assay described for measuring serum calcium. Collagen cross-links were quantitated according to an assay previously described.(17) The total daily excretion was corrected for creatinine excretion, which was measured kinetically.
Bone ash weight and calcium content The dry weight of the right femur or humerus was obtained by burning the bones for 24 h at 100°C in a muffle furnace, followed by 24 h at 600°C to obtain the ash weight. Calcium was measured by microcolorimetry in HCldissolved ash dilutions and reported either as total calcium content or corrected for the dry weight of bone.
1512
DACI ET AL.
TABLE 1. GENERAL CHARACTERISTICS OF PAI-1⫺/⫺ MICE AND WT MICE IN WHICH OVARIECTOMY (OVX) OR SHAM-OPERATION (SHAM) WERE PERFORMED AT 8 WEEKS OF AGE
Weight at start (g) Weight gain (g) Uterus weight (mg) Femoral length (mm) Total femoral calcium (mg) Femoral calcium (mg/g dry weight)
WT/Sham (n ⫽ 8)
PAI-1⫺/⫺/Sham (n ⫽ 9)
WT/OVX (n ⫽ 10)
PAI-1⫺/⫺/OVX (n ⫽ 11)
21 ⫾ 0.5 2.31 ⫾ 0.52 109 ⫾ 18 15.2 ⫾ 0.1 8.2 ⫾ 0.3 219 ⫾ 7
20.5 ⫾ 0.3 1.67 ⫾ 0.23 105 ⫾ 11 15.1 ⫾ 0.2 7.7 ⫾ 0.3 213 ⫾ 4
20.2 ⫾ 0.9 3.23 ⫾ 0.38 21 ⫾ 1† 15.0 ⫾ 0.2 7.7 ⫾ 0.2 215 ⫾ 1
21.1 ⫾ 0.6 2.69 ⫾ 0.33 22 ⫾ 1† 15.1 ⫾ 0.2 7.4 ⫾ 0.3 209 ⫾ 4
Data are reported as mean ⫾ SEM for the number of animals given in parentheses. Statistical analysis: († p ⬍ 0.001) vs. Sham of the same genotype.
Bone histomorphometry Bone histomorphometry was performed based on a method previously described(18) and adapted to mouse bones. The left tibia was embedded undecalcified in methylmetacrylate and 4-m-thick longitudinal sections were cut with a rotary microtome (RM 2155; Leica, Heidelberg, Germany) equipped with a tungsten carbide 50° knife. The sections were left unstained or stained according to von Kossa or by a modified Goldner technique. All measurements were performed in the proximal tibial metaphysis with a Kontron Image Analyzing Computer (KS 400 V 3.00; Kontron Electronik, Eching bei Mu¨nchen, Germany). The trabecular bone volume (BV/TV, as a percentage of total volume) was quantitated in the secondary spongiosa of six von Kossa–stained sections, each 60 m apart. In each section, three consecutive fields (tissue area of 1 mm2) were measured along the vertical axis of the central metaphysis, starting 0.25 mm from the distal end of the growth plate. The osteoid surface (OS/BS, as a percentage of the trabecular surface) was assessed on three Goldner-stained sections, 60 m apart, in the secondary spongiosa of the metaphysis, starting 0.25 mm from the distal end of the growth plate and excluding the endocortical surface. The mineral apposition rate (MAR) and bone formation rate (BFR) were assessed by fluorescence microscopy in three unstained sections, at the endosteal side of cortical bone in the proximal metaphysis. MAR was calculated as the mean distance between all double fluorochrome labels, divided by the number of days between the calcein injections. BFR was obtained by the product of MAR with the mineralizing surface (MS), calculated as follows: MS ⫽ (dLS ⫹ sLS/2)/ BS, where dLS and sLS represent, respectively, the double labeled surface and single labeled surface. The cortical thickness was measured in three unstained sections, 1.3, 2.3, and 3.3 mm distal to the growth plate, on the left and right side of each section, and the mean was calculated. Standardized terms and abbreviations used to describe histomorphometric analysis are according to the American Society for Bone and Mineral Research histomorphometry nomenclature.(19)
Bone densitometry The mineral content and density of cortical bone were determined in the distal metaphysis of the left femur by
pQCT. Imaging was carried out with an XCT Research M bone densitometer (Stratec, Pforzheim, Germany) using a voxel size of 0.1 mm. Contour mode 2, cortical mode 1, and a threshold of 480 mg/cm3 were used to analyze cortical bone. First, a low-resolution scout scan was performed, allowing the identification of the condyles and positioning of the reference line. Thereafter, three scans starting 1.3 mm from the reference line and each 0.25 mm apart were obtained. Based on the position of the condyles, one image was selected and used in further analysis. The precision and stability of the measurements were evaluated periodically and the CV was ⬍1.2%. Regression analysis was performed to assess the correlation between total femoral calcium measured biochemically and cortical mineral content assessed by pQCT and showed a good correlation between them (r ⫽ 0.59, p ⬍ 0.001).
Statistical analysis Statistical analysis was performed using a statistical software program (NCSS, Kaysville, UT, U.S.A.). The results are expressed as the mean ⫾ SEM. One-way analysis of variance was carried out to detect intergroup differences, followed by Fisher’s least significant difference (LSD) multiple comparison test. Differences were considered significant at p ⬍ 0.05. Correlation coefficients were calculated to test for an association between two independently measured parameters.
RESULTS General characteristics of OVX and Sham PAI-1⫺/⫺ and WT mice PAI-1⫺/⫺ and WT mice had comparable body weights when operated at 8 weeks (Table 1) or 13 weeks of age (data not shown). No significant difference was observed in the body weight gained during the 4-week experimental period between the different experimental groups, although the OVX mice tended to gain more weight compared with the respective Sham mice (Table 1). To confirm the estrogen status of the OVX or Sham mice, the uterine wet weight was measured at the time of death (Table 1). Ovariectomy performed in 8-week-old mice sig-
PAI-1 AND OVARIECTOMY-INDUCED BONE LOSS
1513
humeri of mice operated on at 13 weeks of age. Four weeks after ovariectomy, calcium content in bone decreased by 5% and 6% in PAI-1⫺/⫺ and WT mice, respectively (Fig. 2A). In addition, when calcium content was corrected for bone dry weight, a 5% and 4% decrease (p ⬍ 0.05) was observed in OVX PAI-1⫺/⫺ and WT mice, respectively, compared with the corresponding Sham mice (Fig. 2B).
Bone histomorphometric analysis
FIG. 1. Urinary excretion of pyridinoline cross-links in PAI-1⫺/⫺ mice and WT mice in which sham operation (Sham) or ovariectomy (OVX) were performed at 8 weeks of age. Pyridinoline cross-links were determined as described in the Materials and Methods section. The results are expressed as the mean ⫾ SEM of three urine samples/ group, each sample obtained by pooling the 24-h urine collected from three mice.
nificantly decreased the uterus weight of both PAI-1⫺/⫺ mice and WT mice by 79% and 81%, respectively. PAI-1 deficiency had no effect on the uterus weight of Sham or OVX mice. Similar data were obtained when ovariectomy or sham operation was performed at 13 weeks of age (data not shown). PAI-1 inactivation did not affect serum calcium concentration and 24-h urinary excretion of calcium measured at baseline (data not shown). When these parameters were measured 4 weeks after the intervention, again, no difference was found among the different experimental groups (data not shown). The urinary excretion of pyridinoline cross-links, a parameter of bone resorption, was increased after ovariectomy similarly in WT and PAI-1⫺/⫺ mice (Fig. 1). However, these effects were statistically not significant, probably because of the low number of samples, because measurements were performed in pooled urine samples. PAI-1 deficiency did not alter urinary pyridinoline excretion in Sham or in OVX mice. The femoral length was similar in PAI-1– deficient and WT mice at baseline (14.5 ⫾ 0.1 mm, n ⫽ 6 vs. 14.2 ⫾ 0.1 mm, n ⫽ 8, respectively) and was not different among the different experimental groups operated at either 8 weeks (Table 1) or 13 weeks of age (data not shown). The PAI-1⫺/⫺ and WT mice of the baseline group showed comparable values for the total calcium content in bone and for calcium corrected for bone dry weight (225 ⫾ 3 mg/g, n ⫽ 6 vs. 224 ⫾ 5 mg/g, n ⫽ 8, respectively). The total femoral calcium was not different between PAI-1⫺/⫺ mice and WT mice in which sham operation was performed at 8 weeks of age. Ovariectomy decreased femoral calcium content in PAI-1⫺/⫺ and WT mice by 4% and 6%, respectively (Table 1). Similar data were obtained by analyzing
The BV/TV was similar in PAI-1– deficient and WT mice at baseline (4.3 ⫾ 0.4%, n ⫽ 6 vs. 3.5 ⫾ 0.7%, n ⫽ 8, respectively). Histomorphometric analysis of mice operated on at 8 weeks of age showed that the BV/TV of Sham mice was not altered by PAI-1 inactivation (Table 2). As expected, ovariectomy induced a significant reduction (by 49%) in the BV/TV of WT mice. In contrast, in PAI-1– deficient mice ovariectomy caused only a 21% reduction in the BV/TV, an effect that was not statistically significant. As a result, OVX PAI-1⫺/⫺ mice had a significantly higher BV/TV compared with OVX WT mice (Table 2). These findings were confirmed in mice in which ovariectomy or sham operation was performed at 13 weeks of age. Four weeks after ovariectomy, the BV/TV of WT mice decreased by 65%, whereas no changes were observed in the BV/TV of PAI-1– deficient mice (Table 2). Again, the BV/TV of OVX PAI-1– deficient mice was significantly higher compared with the OVX WT mice. A comparable picture was observed when the OS/BS of mice in which ovariectomy or sham operation was performed at 8 weeks of age was analyzed. Comparing Sham animals, no significant difference was detected between PAI-1⫺/⫺ mice and WT mice (Table 2). Four weeks after ovariectomy, a 70% increase in OS was observed in WT mice, whereas only a slight and statistically not significant increase (25%) was observed in PAI-1– deficient mice (Table 2). As a result, the OS/BS was significantly smaller in OVX PAI-1⫺/⫺ mice compared with OVX WT mice. No difference was found between PAI-1– deficient mice and WT mice in the endocortical MAR and BFR at baseline (5.7 ⫾ 0.7 m3/m2 per day, n ⫽ 4 vs. 5.9 ⫾ 0.6 m3/m2 per day, n ⫽ 4, respectively) and in Sham mice (Table 2). Both of these parameters increased significantly in WT mice after ovariectomy. In agreement with the other histomorphometric parameters, the MAR and BFR did not change significantly after ovariectomy in PAI-1– deficient mice (Table 2). As a consequence, a significantly smaller MAR and BFR was found in OVX PAI-1⫺/⫺ mice compared with WT mice. The changes observed in MAR and BFR reflect the alterations in the OS/BS, as shown by their strong correlations (r ⫽ 0.69 and p ⬍ 0.0001 and r ⫽ 0.65 and p ⬍ 0.0001) with the latter. The cortical thickness was similar in PAI-1⫺/⫺ mice and WT mice in which sham operation was performed at either 8 weeks or 13 weeks of age (Table 2). This parameter decreased slightly after ovariectomy in both PAI-1– deficient mice and WT mice, without difference between the two genotypes (Table 2).
1514
DACI ET AL.
FIG. 2. (A) The total calcium content in humerus and (B) calcium corrected for bone dry weight in PAI-1⫺/⫺ mice and WT mice in which sham operation (Sham) or ovariectomy (OVX) were performed at 13 weeks of age. Measurements were performed as described in the Materials and Methods section. Both parameters decreased after ovariectomy similarly in PAI-1⫺/⫺ mice and WT mice. Bars represent the mean ⫾ SEM for bones from four to six mice per group. The asterisks indicate a statistically significant difference from the respective Sham mice (* p ⬍ 0.05). TABLE 2. HISTOMORPHOMETRIC ANALYSIS
OF THE PROXIMAL TIBIAL METAPHYSIS IN OVARIECTOMY AND SHAM OPERATION
PAI-1⫺/⫺ MICE
AND
WT MICE AFTER
WT/Sham
PAI-1⫺/⫺/Sham
WT/OVX
PAI-1⫺/⫺/OVX
BV/TV (%) OS/BS (%) MAR (m/day) BFR (m3/m2 per day) Cortical thickness (m)
3.00 ⫾ 0.75 (8) 12.9 ⫾ 2.6 (8) 1.67 ⫾ 0.08 (8) 1.38 ⫾ 0.23 (6) 194 ⫾ 12 (8)
3.17 ⫾ 0.50 (8) 9.2 ⫾ 1.2 (9) 1.55 ⫾ 0.06 (8) 1.13 ⫾ 0.16 (8) 185 ⫾ 4 (9)
1.54 ⫾ 0.33 (10)* 21.9 ⫾ 2.3 (10)** 1.94 ⫾ 0.10 (9)* 2.51 ⫾ 0.36 (9)* 181 ⫾ 6 (10)
2.51 ⫾ 0.36 (10)† 11.5 ⫾ 1.2 (10)††† 1.66 ⫾ 0.07 (9)† 1.67 ⫾ 0.21 (9)† 177 ⫾ 7 (10)
BV/TV (%) Cortical thickness (m)
1.96 ⫾ 0.44 (4) 206 ⫾ 8 (4)
2.76 ⫾ 0.73 (4) 220 ⫾ 9 (4)
0.69 ⫾ 0.14 (6)* 191 ⫾ 9 (6)
2.65 ⫾ 0.50 (5)†† 205 ⫾ 5 (5)
A
B
Data are reported as mean ⫾ SEM for the number of animals given in parentheses; A, histomorphometric analysis of mice ovariectomized or sham-operated at 8 weeks of age; B, histomorphometric analysis of mice ovariectomized or sham-operated at 13 weeks of age. Statistical analysis: * p value vs. Sham of the same genotype (* p ⬍ 0.05, ** p ⬍ 0.01); † p value vs. WT under the same treatment († p ⬍ 0.05, †† p ⬍ 0.01, ††† p ⬍ 0.001).
Quantification of mineral content and density of femoral cortex by pQCT The cortical mineral content and density were similar in PAI-1⫺/⫺ mice and WT mice in which sham operation was performed at 8 weeks of age (Figs. 3A and 3B). Ovariectomy significantly decreased the cortical mineral content in both PAI-1⫺/⫺ mice and WT mice with a similar amount (by 18%; Fig. 3A). The cortical density also decreased significantly after ovariectomy in both PAI-1⫺/⫺ mice and WT mice (by 5% and 9%, respectively), without significant difference between these two groups (Fig. 3B).
DISCUSSION Several in vitro studies have suggested the involvement of the plasminogen activator/PAI-1 system in bone metabolism.(6,10) In the present study we investigated the effect of
PAI-1 inactivation on bone metabolism in vivo in basal conditions and in a model of increased bone turnover induced by estrogen deficiency. We show that in basal conditions PAI-1 inactivation has only minimal consequences on bone turnover and bone mass. Surprisingly, in conditions of estrogen deficiency, PAI-1 deficiency protects against trabecular bone loss but does not affect cortical bone loss, suggesting a complex role for PAI-1 in bone turnover. The present data on the effects of ovariectomy in WT mice are in agreement with other studies(20,21) showing that estrogen deficiency in normal mice increases the rate of bone remodeling resulting in a net bone loss. More precisely, the increase in the urinary excretion of pyridinoline cross-links observed in OVX WT mice is consistent with previous studies(22) and reflects the high resorption rate characteristic for estrogen deficiency. In addition, OS/BS, MAR, and BFR increased significantly in OVX WT mice, indicating that enhanced osteoblast activity is an important
PAI-1 AND OVARIECTOMY-INDUCED BONE LOSS
1515
FIG. 3. (A) Cortical mineral content and (B) cortical density in the distal femoral metaphysis of PAI-1⫺/⫺ mice and WT mice in which sham operation (Sham) or ovariectomy (OVX) were performed at 8 weeks of age. The pQCT analysis was performed as described in the Materials and Methods section. Ovariectomy induced a significant decrease in both cortical mineral content and cortical density of PAI-1⫺/⫺ mice and WT mice. Bars represent the mean ⫾ SEM for bones from 8 –11 mice per group. The asterisks indicate a statistically significant difference from the respective Sham mice (*p ⬍ 0.05; **p ⬍ 0.01; ***p ⬍ 0.001). component of the bone response to estrogen deficiency, as previously reported.(23,24) However, this high bone turnover ultimately results in osteopenia, as confirmed in our study by the decreases in BV/TV, cortical density, and cortical mineral content in OVX WT mice. Evaluation of basal bone turnover and bone mass in PAI-1⫺/⫺ mice and WT mice revealed no differences between these two groups. The normal bone growth and remodeling in PAI-1– deficient mice in basal conditions indicates that PAI-1 does not play a major regulating role in basal bone metabolism and suggests that other alternative pathways may compensate for the lack of PAI-1 function. However, striking differences were observed in the response of PAI-1⫺/⫺ mice and WT mice to ovariectomy. Eight-weekold PAI-1⫺/⫺ mice were partially protected against trabecular bone loss induced by ovariectomy. However, 8-week-old mice are still growing and both bone modeling and bone remodeling contribute to the changes observed in the secondary spongiosa. To confirm the effects of PAI-1 on bone remodeling, ovariectomies were performed in 13-week-old mice. In these very slowly growing mice, ovariectomy caused a 65% decrease in the BV/TV of WT mice, whereas PAI-1– deficient mice were protected completely from the trabecular bone loss. The compensatory bone formation, measured by the OS/BS, MAR, and BFR, was also remarkably low in the OVX PAI-1⫺/⫺ mice, most likely as a result of the low resorption. These data suggest that the turnover of trabecular bone in OVX PAI-1⫺/⫺ mice is low. This finding was unexpected, because PAI-1 deficiency will generate increased levels of plasmin,(25) which is considered to enhance the rate of bone remodeling rather than decrease it. Indeed, the participation of the plasminogen activators and plasmin in the degradation of nonmineralized bone matrix in vitro has been recently described.(10) Moreover, we have found that degradation of nonmineralized bone matrix is enhanced by PAI-1 deficiency (Daci et al., unpublished data). The discrepancy between the in vivo and in vitro data and the understanding of the mechanism responsible for the protection
against trabecular bone loss in OVX PAI-1– deficient mice require further investigation. However, this phenotype resembles the bone pathology described in mice deficient in interleukin-6 (IL-6), in the functional IL-1 receptor, or in tumor necrosis factor (TNF).(26 –28) It is therefore tempting to speculate that down-regulation of cytokine activity by the excessive plasmin is a possible underlying mechanism for the present findings. Alternatively, they may reflect a direct nonproteolytic effect of PAI-1 on bone cells. More precisely, PAI-1 may bind to vitronectin and block the binding of the integrin ␣v3 to this substrate. This interaction, which has been shown to inhibit the migration of smooth muscle cells,(29) may also affect the behavior of bone cells. The binding of the integrin ␣v3 to its ligands may be critical for optimal osteoclast function, as suggested by the protection from the ovariectomy-induced bone loss in mice lacking osteopontin(30) or treated with echistatin.(21) The reduced bone resorption observed in both cases may result from alterations in the binding of ␣v3 to its high-affinity ligands, leading to inefficient signaling through ␣v3 and decreased osteoclast activity. Another intriguing but unresolved question is why trabecular bone of PAI-1– deficient mice responded differently to estrogen deficiency compared with cortical bone. A similar observation has been reported by Poli et al.,(26) showing a different role for IL-6 in cortical and trabecular bone metabolism. An explanation for the fact that certain factors are involved differently in these bone compartments is not available to date. In conclusion, bone growth and modeling proceed normally in mice with inactivation of the PAI-1 gene function, indicating that the role of PAI-1 in basal bone metabolism is not critical but rather redundant. Nevertheless, in conditions of estrogen deficiency, PAI-1 deficiency protects from the increased bone turnover and trabecular bone loss, without affecting cortical bone loss. These observations suggest a specific role for PAI-1 in the regulation of trabecular bone turnover during estrogen deficiency.
1516
DACI ET AL.
ACKNOWLEDGMENTS The authors are grateful to Stratec Medizintechnik, Pforzheim, Germany, for their collaboration in performing the pQCT analysis, and they thank S. Torrekens and R. Van Looveren for technical assistance. This work was supported by a grant from Fonds voor Wetenschappelijk Onderzoek, Belgium (FWO, grant G.0233.97).
17.
18.
REFERENCES 1. Carmeliet P, Collen D 1995 Gene targeting and gene transfer studies of the plasminogen/plasmin system: Implications in thrombosis, hemostasis, neointima formation and atherosclerosis. FASEB J 9:934 –938. 2. Allan EH, Zeheb R, Gelehrter TD, Heaton JH, Fukumoto S, Yee JA, Martin TJ 1991 Transforming growth factor beta inhibits plasminogen activator (PA) activity and stimulates production of urokinase-type PA, PA inhibitor-1 mRNA, and protein in rat osteoblast-like cells. J Cell Physiol 149:34 – 43. 3. Yang JN, Allan EH, Anderson GI, Martin TJ, Minkin C 1997 Plasminogen activator system in osteoclasts. J Bone Miner Res 12:761–768. 4. Fukumoto S, Allan EH, Yee JA, Gelehrter TD, Martin TJ 1992 Plasminogen activator regulation in osteoblasts: parathyroid hormone inhibition of type-1 plasminogen activator inhibitor and its mRNA. J Cell Physiol 152:346 –355. 5. Hoekman K, Lo¨wik CWGM, Van De Ruit M, Bijvoet OLM, Verheijen JH, Papapoulos SE 1991 Regulation of the production of plasminogen activators by bone resorption enhancing and inhibiting factors in three types of osteoblast-like cells. Bone Miner 14:189 –204. 6. Allan EH, Martin TJ 1995 The plasminogen activator inhibitor system in bone cell function. Clin Orthop 313:54 – 63. 7. Mignatti P, Rifkin DB 1993 Biology and biochemistry of proteinases in tumor invasion. Physiol Rev 73:161–195. 8. Murphy G 1995 Matrix metalloproteinases and their inhibitors. Acta Orthop Scand 66(Suppl 266):55– 60. 9. Carmeliet P, Moons L, Lijnen R, Baes M, Lemaitre V, Tipping P, Drew A, Eeckhout Y, Shapiro S, Lupu F, Collen D 1997 Urokinase-generated plasmin activates matrix metalloproteinases during aneurysm formation. Nat Genet 17:439 – 444. 10. Daci E, Udagawa N, Martin TJ, Bouillon R, Carmeliet G 1999 The role of the plasminogen system in bone resorption in vitro. J Bone Miner Res 14:946 –952. 11. Yee JA, Yan L, Dominguez JC, Allan EH, Martin TJ 1993 Plasminogen-dependent activation of latent transforming growth factor beta (TGF) by growing cultures of osteoblastlike cells. J Cell Physiol 157:528 –534. 12. Munger JS, Harpel JG, Gleizes PE, Mazzieri R, Nunes I, Rifkin DB 1997 Latent transforming growth factor-: Structural features and mechanisms of activation. Kidney Int 51: 1376 –1382. 13. Campbell PG, Novak JF, Yanosick TB, McMaster JH 1992 Involvement of the plasmin system in dissociation of the insulin-like growth factor-binding protein complex. Endocrinology 130:1401–1412. 14. Leloup G, Delaisse JM, Vaes G 1994 Relationship of the plasminogen activator/plasmin cascade to osteoclast invasion and mineral resorption in explanted fetal metatarsal bones. J Bone Miner Res 9:891–902. 15. Leloup G, Lemoine P, Carmeliet P, Vaes G 1996 Bone resorption and response to calcium-regulating hormones in the absence of tissue or urokinase plasminogen activator or of their type 1 inhibitor. J Bone Miner Res 11:1146 –1157. 16. Carmeliet P, Kieckens L, Schoonjans L, Ream B, Van Nuffelen A, Prendergast G, Cole M, Bronson R, Collen D,
19.
20.
21.
22.
23. 24.
25.
26.
27.
28. 29. 30.
Mulligan RC 1993 Plasminogen activator inhibitor-1 genedeficient mice. I. Generation by homologous recombination and characterization. J Clin Invest 92:2746 –2755. Mathieu C, Waer M, Casteels K, Laureys J, Bouillon R 1995 Prevention of type I diabetes in NOD mice by nonhypercalcemic doses of a new structural analog of 1,25dihydroxyvitamin D3, KH1060. Endocrinology 136:866 – 872. Verhaeghe J, Van Bree R, Van Herck E, Thomas H, Skottner A, Dequeker J, Mosekilde L, Einhorn TA, Bouillon R 1996 Effects of recombinant human growth hormone and insulinlike growth factor-I, with or without 17-estradiol, on bone and mineral homeostasis of aged ovariectomized rats. J Bone Miner Res 11:1723–1735. Parfitt AM, Drezner MK, Glorieux FH, Kanis JA, Malluche H, Meunier PJ, Ott SM, Recker RR 1987 Bone histomorphometry: Standardization of nomenclature, symbols, and units. J Bone Miner Res 2:595– 610. Ammann P, Rizzoli R, Bonjour JP, Bourrin S, Meyer JM, Vassalli P, Garcia I 1997 Transgenic mice expressing soluble tumor necrosis factor-receptor are protected against bone loss caused by estrogen deficiency. J Clin Invest 99:1699 –1703. Yamamoto M, Fisher JE, Gentile M, Seedor JG, Leu CT, Rodan SB, Rodan GA 1998 The integrin ligand echistatin prevents bone loss in ovariectomized mice and rats. Endocrinology 139:1411–1419. Kitazawa R, Kimble RB, Vannice JL, Kung VT, Pacifici R 1994 Interleukin-1 receptor antagonist and tumor necrosis factor binding protein decrease osteoclast formation and bone resorption in ovariectomized mice. J Clin Invest 94:2397–2406. Manolagas SC, Jilka RL 1995 Bone marrow, cytokines, and bone remodeling. Emerging insights into the pathophysiology of osteoporosis. N Engl J Med 332:305–311. Jilka RL, Takahashi K, Munshi M, Williams DC, Roberson PK, Manolagas SC 1998 Loss of estrogen upregulates osteoblastogenesis in the murine bone marrow. Evidence for autonomy from factors released during bone resorption. J Clin Invest 101:1942–1950. Carmeliet P, Stassen JM, Schoonjans L, Ream B, van den Oord JJ, De Mol M, Mulligan RC, Collen D 1993 Plasminogen activator inhibitor-1 gene-deficient mice. II. Effects on hemostasis, thrombosis, and thrombolysis. J Clin Invest 92:2756 – 2760. Poli V, Balena R, Fattori E, Markatos A, Yamamoto M, Tanaka H, Ciliberto G, Rodan GA, Costantini F 1994 Interleukin-6 deficient mice are protected from bone loss caused by estrogen depletion. EMBO J 13:1189 –1196. Lorenzo JA, Naprta A, Rao Y, Alander C, Glaccum M, Widmer M, Gronowicz G, Kalinowski J, Pilbeam CC 1998 Mice lacking the type I interleukin-1 receptor do not lose bone mass after ovariectomy. Endocrinology 139:3022–3025. Kimble RB, Bain S, Pacifici R 1997 The functional block of TNF but not of IL-6 prevents bone loss in ovariectomized mice. J Bone Miner Res 12:935–941. Stefansson S, Lawrence DA 1996 The serpin PAI-1 inhibits cell migration by blocking integrin ␣v3 binding to vitronectin. Nature 383:441– 443. Yoshitake H, Rittling SR, Denhardt DT, Noda M 1999 Osteopontin-deficient mice are resistant to ovariectomy-induced bone resorption. Proc Natl Acad Sci U S A 96:8156 – 8160.
Address reprint requests to: Geert Carmeliet Legendo, Onderwijs en Navorsing Campus Gasthuisberg Herestraat 49, B-3000, Leuven, Belgium Received in original form July 9, 1999; in revised form February 21, 2000; accepted March 21, 2000.
