8756-3282/92 $5 .00 + .00 Copyright © 1992 Pergamon Press Ltd .

Bone, 13, 249-255, (1992)

Printed in the USA . All rights reserved .

Effects of Bisphosphonates and Inorganic Pyrophosphate on Osteogenesis in Vitro H . C . TENENBAUM,'' z M . TORONTALI 2 and B . SUKHU 2' '

Faculty

of Dentistry

and 2Samuel Lunenfeld Research Institute, Mount Sinai Hospital, University

of

Toronto, Canada

Address for correspondence and reprints : Dr . H . C . Tenenbaum, Samuel Lunenfeld Research Institute, Mount Sinai Hospital, 600 University

Ave ., Room 984, Toronto, Ontario, M5G 1X5 Canada . While a majority of studies have focused on the effects of bisphosphonates on resorption (Russell et al . 1973 ; Miller & Jee 1975 ; Fleisch 1980 ; Francis & Martodam 1983), there appears to be less information on osteoblastic activity . Indeed there seems to be some controversial findings in this regard showing, for example, that the bisphosphonate HEBP may stimulate or inhibit bone matrix formation (Francis & Martodam 1983) . This controversy may be a reflection of the problems associated with the use of varied and different model systems . With respect to the effects of bisphosphonates on bone-forming cells, some studies were carried out on whole calvarial cultures (Lerner & Larsson 1987) ; however, whole calvarial explants from chick do not actively produce new bone tissue and seem to remain in a state of stasis when cultured (McCulloch et al . 1989) although mineralization might continue . Other studies were carried out using cell culture models (Felix & Fleisch 1979 ; Felix et al . 1981) . However, unless grown with ascorbate and P-glycerophosphate (GP), most cell culture models do not form mineralized bone (EcarotCharrier et al, 1983 ; Bellows et al . 1986) . In view of the controversial findings, we studied the direct effect of drugs such as the bisphosphonates on osteogenesis in the chick periosteal osteogenesis (CPO) model (Tenenbaum & Heersche 1982 ; Tenenbaum & Heersche 1985 ; Tenenbaum et al . 1986) . The effects of the bisphosphonates HEBP and disodium-lhydroxy-l-aminopropylidine-1,1-diphosphonate (APD), a very potent inhibitor of resorption, as well as inorganic pyrophosphate (PPi), on markers of bone metabolism such as alkaline phosphatase activity, calcium, and phosphate uptake were tested .

Abstract The bisphosphonates, which are chemically related to pyrophosphate, have been studied extensively both in vivo and in vitro to elucidate their effects on bone tissues and cells . However, because these agents have important effects on bone resorption, the majority of investigations have focused on this area . Few studies regarding direct bisphosphonate effects on bone formation have been carried out In the past and, thus, we chose to use the chick periosteal osteogenesis (CPO) in vitro model system to test the direct effects of pyrophosphate and the bisphosphonatesethane-l-hydroxy-1,1diphosphate (HEBP) and disodium-l-hydroxy-1-aminopropylidine (APD) on various parameters of osteogenesis in vitro . The data show that the bisphosphonate HEBP inhibits bone mineralization reversibly while APD, at low doses, may actually enhance mineralization of bone . Similarly, pyrophosphate (PPi) will prevent mineralization in CPO cultures . However, CPO cultures can circumvent PPi-mediated blockage of mineralization with longer-term, continuous (10-day) incubation, whereas this does not occur if cultures are incubated continuously with bisphosphonates . Both drugs appear to be able to reverse j3-glycerophosphate-induced changes in alkaline phosphatase activity, but do not appear on their own to regulate the activity of this enzyme . The findings show that in addition to their well-known effects on resorption, bisphosphonates have significant and direct effects on mineralization in bone-forming cultures . Their direct effects on osteoblastic activity and differentiation remain to be determined. Key Words: Bisphosphonates-Osteogenesis in vitro-Mineralization .

Materials and Methods Culture system

Introduction The CPO model system employs the ectocranial periosteal tissues derived from calvarial halves of 17-day-old embryonic chicks . The periosteal tissues were removed and folded with the osteogenic layer within the fold and in apposition to itself . The osteogenic layer is devoid of differentiated osteoblasts and, thus, bone formation in this model is reliant on the differentiation of osteoprogenitor cells into functional osteoblasts (McCulloch et al . 1989) . The explants were supported at the gas/liquid interface on a Millipore filter (HA 0 .45 Ism) which was resting on a stainless-steel grid over the center well of an organ culture dish (Falcon Plastics, Lincoln Park, NJ, USA) .

Bisphosphonates belong to a class of pharmacologic agents that are chemically related to inorganic pyrophosphate (Katoh et al . 1991) and are used in low doses to inhibit bone resorption (Russell et al . 1973 ; Miller & Jee 1975 ; Fleisch 1980) . However, at higher doses, some bisphosphonates and, in particular, ethane1-hydroxy-1,1-diphosphate (1iEBP) inhibit mineralization (King et al . 1971) . The mechanism behind bisphosphonate-mediated inhibition of mineralization is not known entirely, but it is felt that this is largely due to physicochemical interference with hydroxyapatite formation (Katoh et al . 1991 ; Schenk et al . 1973) . 249

250

H . C . Tenenbaum et al . : Bisphosphonates and osteogenesis

Culture conditions

Alkaline phosphatase assay modulation

All cultures were incubated at 37°C in a humidified atmosphere containing 5% CO, in air for up to 12 days . The culture medium was changed every 48 hours and consisted of BGJ3 medium (Gibco, Grand Island, NY, USA), supplemented with 10% fetal calf serum (Gibco), l0 r M dexamethasone (Sigma, St . Louis, MO, USA) (Tenenbaum & Heersche 1985 ; McCulloch & Tenenbaum 1986) and 300 µg/ml L-ascorbate (Gibco) (Tenenbaum & Heersche 1986) . To induce mineralization 10 mM (GP) (Tenenbaum et al . 1986) was added to the bisphosphonate-treated experimental groups as well as to various controls without bisphosphonate . Controls without GP and bisphosphonate were also done where indicated .

To show that PPi or the alkaline phosphatase substrate GP had direct effects on alkaline phosphatase-mediated hydrolysis of a substrate, the following was done . Pyrophosphate or GP (0, 10, 15, 20, 25, 40, 35, 40 mM) was added directly to the samples before alkaline phosphatase reaction . Alkaline phosphatase was obtained from pooled bicarbonate extracts of 18-20 CPO cultures grown for six days with or without 10 mM GP and the reaction was carried out as described previously (Tenenbaum & Heersche 1985, 1986) . The 18-20 CPO culture extracts were pooled to obtain enough sample for assessment of multiple doses of PPi and GP .

Bisphosphonate treatment HEBP . HEBP was dissolved in GP-containing medium and added to CPO cultures in various dosages (0, 0 .5, 5, 10, 25, 50 µm) from the outset of culture and up to 12 days . Control cultures containing neither GP nor HEBP were carried out as well . To determine whether CPO cultures could recover from HEBP effects, some GP-treated cultures were incubated with the optimal mineralization-inhibiting dose of HEBP (50 µM) over the first six days and then allowed to grow for another six days under HEBP-free conditions . APD . APD was dissolved in medium and added to the GPsupplemented culture media in various concentrations (0, 0 .03, 0.1, 1 .0, 3 .0, 10 KM) from the outset of culture and up to six days . Thus, them was some overlap of dosage between HEBP and APD . However, as APD is used clinically in lower dosage than HEBP, we explored the lower dosage range for APD . PPi . Previous studies showed that a dosage of 10-30 µm PPi effectively inhibited mineralization in six-day CPO cultures (Tenenbaum 1987) . To determine if PPi, being susceptible to hydrolysis by pyrophosphatase and alkaline phosphatase at neutral pH (Xu et al . 1991), could block mineralization in long-term cultures, GP-treated CPO explants were incubated continuously with PPi (10 µm) for 6, 8, and 10 days .

Biochemical measurements Single CPO explants were homogenized in I ml of a bicarbonate buffer (3 mM NaHCO3 in 15 .0 mM NaCl pH 7 .4) using a Polytron homogenizer (Kinematica GMBH, Switzerland) . The homogenate was then transferred to Eppendorf tubes and subsequently centrifuged (high speed) on a microfuge (Eppendorf 5413) for 5 min at 5° C . The supernatant fraction was assayed for soluble protein content (Bradford 1976) and alkaline phosphatase activity (Tenenbaum & Heersche 1985, 1986), while the pellet was used after hydrolysis in 0 .5 N HCI for calcium and phosphate determination- All colorimetric assays were carried out in Titertek 96 well plates, and optical density was measured with a Titertek Multiskan MC Spectrophotometer (Flow Laboratories, Mississauga, Ont., Canada) . The calcium extracted from the pellets was measured by atomic absorption spectrophotometry (Perkin-Elmer, Norwalk, CT, USA) . AB data were compared with Student's t test, and statistical significance was assigned at the p < .05 level .

Histology Some cultures were fixed in neutral buffered formalin and processed for routine paraffin sections and stained with the von Kossa method to demonstrate mineral deposits . Results Long-term PPi effects As shown previously, PPi effectively inhibited mineralization in six-day CPO cultures . This effect was observed in cultures incubated with PPi for up to eight days . However, in cultures incubated for 10 (or more) days with PPi, calcium accumulation occurred and was similar to that which occurred in GP-treated cultures grown without PPi (Fig . 1a) . Alkaline phosphatase activity was suppressed by addition of GP . The effects of GP on alkaline phosphatase activity were reversed by PPi throughout the culture period time course . The PPi effect appeared to be lessened by day 10 although this was not statistically significant (Fig . Ib) . Pyrophosphate effects on alkaline phosphatase reaction The overall level of alkaline phosphatase activity as measured by production of paranitrophenol (PNP) following hydrolysis of paranitrophenylphosphate (PNP-P) was about 50% lower in the pooled bicarbonate extracts obtained from GP-treated cultures than from cultures grown without GP . The addition of increasing concentrations of PPi to the alkaline phosphatase reaction medium led to a progressive inhibition of PNP-P hydrolysis (r = -0 .80, p < .05 + GP, r = -0 .70, p < .05 - GP) . Conversely, the addition of increasing concentrations of the alkaline phosphatase substrate GP to the alkaline phosphatase reaction medium appeared to clearly potentiate hydrolysis of PNP-P only in extracts obtained from -GP cultures (r = +0 .30, p > .05 + GP; r = +0 .88, p < .05 - GP) (Fig . 2) . HEBP A dose-response curve (Fig . 3a) shows that maximal inhibition of mineralization as measured by calcium and phosphate (not shown) uptake occurred at 25-50 µM HEBP . Further experiments with 50 µM HEBP showed that, as with PPi, GP-mediated suppression of alkaline phosphatase activity was reversed in the presence of HEBP (Fig . 3b) . However, in this case enzyme activity was not suppressed as much as usual with GP treatment as observed clearly in other experiments . Unlike that which occurred with PPi, long-term cultures failed to circumvent HEBP-mediated blockage of mineralization



251

H . C . Tenenbaum et al . : Bisphosphonates and osteogenesis

1 .6

0 .076 (a)

1 .4

in J LJ

0

a

0 .070 -

1 .2 0 .065 -

W J

0 .060 -

oa

0 .055 -

I

E 1 0 .4 0 .2 0 .0

10-D PPi

-GP +GP 7-D CONTROL PPI

-5 0 10 16 20 25 30 35

40

45

[PPI] mU

0 .05 (b) h

C 0 .04

w to a a ~ rn = 1 0 .03

a L N c

O \ a. = M Z 0 .02 . 0

I

I

W

z o J

< Y Y

0 .01

J

0 .00

-GP +GP 7-0 8-D 10-D PPi PPi PPi

Fig. 1. The effect of longer-term PPi treatment on calcium accumulation in the GP-heated cultures is shown in (a) . PPi inhibits calcium accumulation in short-term culture up to seven or eight days . However, after 10 days, PPi-treated CPO cultures contained calcium levels significantly above -GP control and those cultures incubated from six to eight days with PPi . All cultures were grown for 10 days and with the exception of -GP group all cultures treated with IO mM GP . The effect of longerterm PPi-treatment on alkaline phosphatase activity in CPO cultures is shown in (b) . GP-treated groups are as described in (a). Alkaline phosphatase activity was suppressed by GP . This effect was reversed with PPi up to eight days . By day 10, alkaline phosphatase levels began to decline to +GP levels . Each bar represents a mean of eight to nine cultures and the vertical line represents the SEM (*p < .05 compared to -GP) . and was confirmed on microscopic evaluation of long-tern cultures showing little mineral staining (Fig . 4a) . HEBP-treated 12-day cultures were capable of reactivated mineralization if the drug was removed on day 6 as demonstrated histologically (Fig . 4b), and these cultures appeared similar but not identical to those incubated only with GP (Fig . 4c). In this regard, HEBP-recovery cultures still contain some regions with excess osteoid ; however, they contain significantly more calcium (Fig . 4d) than the cultures treated throughout with both GP and HEBP . Cultures ,rated continuously for 12 days with GP contained, as would be expected, such large amounts of calcium and phosphate that comparison to the other groups was not considered appropriate .

-5

0 10

16 20 26 30

35 40 46

[GP] Call Fig . 2. The direct effects of PPi on alkaline phosphatase-mediated hydrolysis of paranitrophenylphosphate are shown in (a) . Alkaline phosphatase activity was present in pooled bicarbonate extracts of CPO cultures grown with or without GP (10 mM) . The addition of increasing concentrations of PPi to the reaction media led to dose-dependent decrease in alkaline phosphatase activity (p < .05) . In (b) the effect of GP on alkaline phosphatase activity is shown . Increasing concentrations of GP lead to a dose-dependent increase in hydrolysis of paranitrophenylphosphate primarily in pooled extract derived from cultures not treated previously with GP . A slight but insignificant effect was observed in pooled extract derived from GP-treated cultures . Each point represents the mean of 9-10 measurements and the vertical lines represent the SEM .

APD The APD dose response showed that in the midrange, both calcium and phosphate uptake by CPO cultures was enhanced relative to the APD-free, GP-treated control (almost fourfold at 0 .1 .m APD) (Figs . 5a and b) while this effect was lost at higher p dosage (10 µm) . p-Glycerophosphate-induced suppression of alkaline phosphatase activity in CPO cultures was reversed with



252

H . C . Tenenbaum et al . : Bisphosphonates and osteogenesis 16 (a)

12 -

6-

2-

-I-

0

-GP +GP

5

10

25

50

75

HEBP CONCENTRATIONS (MM)

(b) •

a P Y •

w z J Q Y J Q

0 .06 -

0 .04 -

-c aa z a •

0 .02 -

E Y

0 .00

-CP

GP

5

10

25 50

75

HEBP CONCENTRATIONS (MM) Fig. 3 . HEBP dose-response . Cultures (six-day) incubated without HEBP or GP (-GP), to mM GP only (+GP) or 10 mM GP with various dosages of HEBP as shown . Calcium accumulation (a) was significantly suppressed with increasing concentration of HEBP . Suppression of uptake was observed at 5 µM and beyond . Lower dosages of HEBP similar to those used for APD, had no effect (not shown) . Reversal of GP_ mediated suppression of alkaline phosphatase activity began at 5 .0 sM HEBP and was maintained throughout the dosage range used (b) . Enzyme levels appeared to decrease at 75 µM HEBP, but this could be due to toxicity . Each bar represents the mean of 8-10 cultures and the vertical lines represent the SEM (*p < .05 as compared to +GP only) .

APD (Fig . 5c), although its effects appeared to be somewhat biphasic.

Discussion We have carried out a number of experiments to characterize osteoblastic response to PPi and PPi-related compounds, the bisphosphonates . The ability of PPi and PPi-like compounds (bisphosphonates) to inhibit mineral formation has been described previously

(Miller & Jee 1975 ; Tenenbaum 1987 ; Katoh et al . 1991), and has also been shown here . The data reported in this study also demonstrate that over longer-term incubation, CPO cultures appear to be capable of circumventing PPi-mediated blockage of mineralization . This might indicate that with respect to mineralization, CPO cultures synthesize elevated levels of pyrophosphatase (PPi-ase) in the presence of PPi which then might hydrolyse PPi, thus allowing mineralization to progress in longer-term cultures . While histologic observations of some specimens (not shown) indicated normal distribution of mineral, electron microscopic studies are needed to confirm that biological mineralization is occurring (Tenenbaum et al . 1989) . Alkaline phosphatase does contain come PPi-ase activity (Xu et al . 1991), and in support of this, findings in a preliminary study showed that PPiase activity in PPi-treated CPO cultures increased 67% more than in cultures not treated with PPi (Tenenbaum & Mamujee, 1989) . Interestingly, AP activity did not drop as much as might be expected when the cultures began to mineralize . This may be because there was still a significant mineralization blockage effect of unhydrolyzed PPi that might have masked such a decrease . Alternatively, a different mineralization mechanism may have beenn activated in the presence of PPi that might not be reflected by a decrease in alkaline phosphatase . The reversal of GP-mediated effects on alkaline phosphatase activity by PPi has been shown in earlier studies (Tenenbaum 1987) . However, while GP treatment suppressed AP activity, the suppression was not necessarily statistically significant due to variability . Similarly, calcium uptake levels may vary between experiments and, thus, comparisons were made only within and not between experiments . In this investigation we attempted to determine whether PPi might interfere not only with mineralizalion and mineralization-associated changes in CPO cultures by acting as a crystal poison (Francis 1969), but whether it might also interact directly with alkaline phosphatase . This assumption was based on the contention that PPi might interfere competitively with alkaline phosphatase-mediated metabolism of the substrate GP . When PPi was added directly to alkaline phosphatase assay reaction media, the ability of extracted alkaline phosphatase to hydrolyse PNP-P was decreased in a dosedependent fashion . This might indicate that, in the test tube, PPi acted as a competitor for the alkaline phosphatase substrate PNP-P in much the same way as it could behave in tissue culture with another alkaline phosphatase substrate, GP . To show that PPi was not strictly an alternate substrate for alkaline phosphatase, another experiment was designed whereby GP, a known alkaline phosphatase substrate, was added directly to the reaction media . These studies showed that hydrolysis of PNP-P by alkaline phosphatase was potentiated in the presence of GP, particularly if the enzyme was derived from cultures grown previously without this organic phosphate . Thus, the findings obtained in tissue culture complement those from the enzyme kinetic studies in support of the concept that PPi and perhaps related compounds might regulate bone-cell activity and mineralization in vitro, in part, by modulating alkaline phosphatase-mediated metabolism of organic phosphate . Further studies with HEBP clearly showed that this bisphosphonate inhibited mineralization reproducibly . However, unlike cultures treated with PPi for over six days, those exposed to HEBP do not show a tendency to circumvent HEBP-mediated blockage of mineralization, and this could be related to the fact that HEBP is resistant to enzymatic hydrolysis as are other bisphosphonates (Francis & Martodam 1983) . Since there is evidence that inhibition of mineralization might actually enhance bone matrix formation (Wesselink & Beertsen 1989 ; Harrison et al . 1990), this characteristic of HEBP might be useful if its minemlization-inhibiting effects were reversible since produc-



25 3

H . C. Tenenbaum et al . : Bisphosphonates and osteogenesis

+GP+HEBP +GP+HEBP 0-6 days +GP 0-12 day. 6-12 days

Fig. 4. Light microscopic photomicrograph of CPO cultures grown either continuously with 10 mM GP and 50 µM HEBP (a), first six days with 10 mM GP and 50 .pM HEBP followed by six days with GP only (b), or grown continuously with 10 mM GP only (c) . Calcium levels in a parallel set of cultures are shown in (d) . Culture shown in (a) was not mineralized and showed virtually no von Kossa positive material . The osteoid (0) contains numerous osteocytes and is surrounded by osteoblast-like cells . In (b), them was extensive mineralization as shown by black von Kossa staining (X) . Some of the osteoid had yet to mineralize and was still von Kossa negative (0) . The tissues are difficult to section and artifactual tearing (arrows, star) is noted and may indicate, indirectly, that the newly formed bone is heavily mineralized . The culture in (c) appears to contain fully mineralized bone (B) with a normal osteoid seam (0) . (All magnification x 500 .) Calcium levels in cultures permitted to recover from HEBP-treatment were over twofold more than in cultures grown continuously with HEBP ('p < .05) (d) .

lion of excess osteoid would not provide added bone strength unless it could calcify . Therefore, it was important to show that CPO cultures grown in the presence of HEBP were capable of mineralization once HEBP was removed . This was confirmed both histologically and by demonstrating significant uptake of both calcium and phosphate in HEBP-treated cultures grown subsequently without HEBP . Thus, osteoid formed in the presence of HEBP is capable of mineralization after HEBP is removed and this finding is in agreement with previously reported data (King et al . 1971) . Further studies of this phenomenon with particular focus on collagen synthesis and mineral density are under way and suggest that collagen synthesis is elevated in GPand HEBP-treated cultures as compared to GP-treated cultures without HEBP. While APD in higher doses will inhibit mineral accumulation as does HEBP, at lower dosages which approximate those that might be used in vivo there appeared to be an increase in mineral accumulation relative to APD-free control which was not observed with low dosages of HEBP . This somewhat unexpected result is not related to an antiresorptive effect since there is no evidence for bone resorption in the CPO model . At this time, the actual mechanism behind this phenomenon is not known . However, as discussed above with respect to PPi, the mineralization process in the presence of a bisphosphonate (PPi-related compound) may be entirely different from that which occurs, for

example, in the presence of organic phosphate alone . Thus, decreases in AP which might be expected during mineralization, albeit in the presence of GP, could either have been reversed by, or did not occur in the presence of, APD . Nonetheless, the use of a bisphosphonate which, in low dosage, actually promotes mineralization may be very valuable, particularly when the mineralization blockage effects of HEBP are to be reversed . Moreover, studies carried out in vivo (unpublished observation) suggest that while intermittent HEBP treatment at high dose induces early bone-wound closure, some of the newly formed bone may be subject to resorption . Thus, APD at a low dose could be used not only to promote mineralization of osteoid formed in the presence of HEBP but also to prevent resorption of newly formed bone . The observations reported in this study show that PPi and the bisphosphonates, the latter of which have well-documented effects on bone resorption (Katoh et al . 1991), also have significant effects on various aspects of osteoblastic activity . By using the CPO model it was possible to test the direct effects of such drugs on osteogenesis without the confounding effects of bone resorption . By understanding more fully the effects of these drugs on osteodifferentiation and bone formation, more rational treatment protocols with these drugs might be designed to take advantage, not only of their anti-resorptive activity but possibly their mineralization modulating effects as well .



2 54

H . C . Tenenbaum et al . : Bisphosphonates and osteogenesis 0.7 N 0 .9

N

0 .09 -

i

0 .5

I

n

0 .06 -

d

M

0 .04 -

u

c

n N

a -E

0.

CC0 3 Of 0 z

0.

o.

0

3.0

0 .02 -

0 .00

too

0

APD CONCENTRATIONS (AM)

0 .a

.0

3 .0

io .

APO CONCENTRATIONS (MM)

0 .07 c

a

Effects of bisphosphonates and inorganic pyrophosphate on osteogenesis in vitro.

The bisphosphonates, which are chemically related to pyrophosphate, have been studied extensively both in vivo and in vitro to elucidate their effects...
495KB Sizes 0 Downloads 0 Views