JOURNAL OF BONE A N D MINERAL RESEARCH Volume 7, Number 10, 1992 Mary Ann Liebert, Inc., Publishers
Osteogenesis Associated with Bone Gla Protein Gene Expression in Diffusion Chambers by Bone Marrow Cells with Demineralized Bone Matrix YOSHIKO DOHI,’ HAJIME OHGUSHI,’ SHIRO TABATA,3 TAKAFUMI YOSHIKAWA,2 KAZUHIRO DOHI.‘ and TADASHIGE MORIYAMA’
ABSTRACT Diffusion chambers with rat bone marrow cells and demineralized bone matrix (DBM) were implanted subcutaneously to syngeneic 8-week-old rats and were harvested every week 3-7 weeks after implantation, and histochemical examination, determination of alkaline phosphatase activity, total calcium and phosphorus, the bone-specific vitamin K-dependent gla-containing protein (BGP) content, and detection of BGP mRNA relative to mineralization were performed. Alkaline phosphatase in diffusion chamber implants reached the highest activity at 4 weeks and then decreased. Calcium and phosphorus deposits occurred at 4 weeks after implantation and were followed by marked increases until 7 weeks, which was comparable to the accumulation of BGP. The BGP gene within the diffusion chambers began to be expressed at 5 weeks, and its expression increased markedly at 7 weeks after implantation. At 4-5 weeks after implantation, new bone adjacent to the membrane filters and cartilage toward the center of the diffusion chamber were observed histochemically. Light microscopic and immunohistologic examinations of chambers with marrow cells and DBM revealed production of mineralized matrices, typical of bone characterized by the appearance of BGP and mineralized nodules. In contrast, bone marrow cells alone did not show extensive bone formation and yielded very low values for these biochemical parameters. The present experiments demonstrate the potential of bone marrow cells and DBM to produce not only cartilage formation but also membranous bone formation associated with increasing expression of BGP mRNA during the later stages of bone formation, as well as a marked accumulation of BGP.
INTRODUCTION
F
at various stages of development is an active process involving the synthesis of connective tissue macromolecules (collagen, bone-specific proteins, and proteoglycans) by osteoblasts, followed by the deposit of a calcium phosphate mineral phase.“-‘] A number of subsequent studies on assay methods for this highly regulated process of osteogenesis have been reported. On the one hand, a demineralized bone matrix (DBM) implanted into various extraskeletal sites induces the differentiation ORMATION OF BONE
to cartilage and bone by endochondral ossification in host tissues.‘s,61The factor responsible for this induction of osteogenesis was found in demineralized bone matrix to be a protein having an approximate molecular weight of 18,000 and has been termed “bone morphogenetic protein” (BMP) .I n 1 On the other hand, it is known that when cells obtained from the bone marrow are transplanted to heterotopic sites, they differentiate to form bone and cartilage. f7.11 These cells have been designated “determined osteogenic precursor cells” (DOPC). Recently, Ohgushi et al. demon-
‘Department of Public Health, Nara Medical University, Japan. ’Department of Orthopaedics. Nara Medical University, Japan. ’Department of Chemistry, Nara Medical University, Japan. ‘First Department of Internal Medicine, Nara Medical University, Japan.
1173
DOH1 ET AL.
1174
strated heterotopic osteogenesis associated with the developmental appearance of the bone-specific vitamin K-dependent gla-containing protein (BGP) in porous ceramics induced by bone marrow cell^.^^.^^^ Diffusion chamber implantation techniques have also been employed to characterize the bone marrow osteogenic precursor cell population'11,121 and to examine the effect of a partially purified BMP on the ability of bone marrow cells to differentiate to form cartilage and bone.'131 The in vivo diffusion chamber method appears to be reproducible as an assay for various stages of osteogenesis with the content of the chambers being isolated from cells of the recipient, which d o not cross the diffusion chamber membrane filter. Therefore, in this study, we demonstrated the ability of bone marrow cells and DBM on differentiation to form cartilage and bone within a diffusion chamber by monitoring various biochemical parameters, including assay of BGP (osteocalcin) and time-dependent expression of BGP mRNA.
MATERIALS AND METHODS Preparation of demineralized bone matrix Demineralized bone matrix was prepared as described by Nishimoto et al.l1') Briefly, the femora and tibiae from 3-month-old male Wistar rats (about 300 g body weight) were removed and cleaned of soft tissue and marrow, frozen in liquid nitrogen, and pulverized with a Hyscotron homogenizer. The bone powder (74-420 pn) was demineralized with 0.5 N HCl for 1 h at 4°C three times and washed out exhaustively with distilled water up to pH 6 , followed by washing with 95% ethanol three times and then rinsing with ether. The bone powder was air dried overnight at room temperature to obtain the granular demineralized bone matrix and then stored at -25°C.
Marrow cell suspension Marrow cell suspensions from femora of 8-week-old male Wistar rats (about 200 g body weight) were prepared as described by Ohgushi et al.lQ1A cell suspension of the marrow was obtained by pumping the tissue repeatedly through hypodermic needles of decreasing gauge. After these procedures, >90% of the cells were viable judging by dye exclusion methods.
Preparation and implantation of diffusion chambers Diffusion chambers (DC; 9 mm inner diameter x 2 mm thick, 130 ~l capacity, 0.45 pm pore size membrane filters) were assembled from commercially available components (Nihon Millipore, Kogyo K.K., Yonezawa, Japan) and sterilized with ethylene oxide. Marrow cell suspensions (5 x loLcells per 90 p1) and DBM (10 mg) were injected into DC, which were then sealed with Millipore filter cement. Three DCs containing rat marrow cells and DBM were implanted subcutaneously into the backs of 8-week-old male Wistar rats weighing 190-210 g. The DCs were harvested every week from 3 to 7 weeks after implantation. DCs that
contained only marrow cells or 10 mg demineralized bone matrix alone were implanted into rats as control experiments.
Calcium and phosphorus determination Bone plaques in DC or bone samples were weighed and then homogenized in 20% formic acid. An aliquot of the formic acid extract was diluted 1:200 with a strontium solution. Calcium concentration was determined by an atomic absorption spectrometer (AA-810 type; Nippon Jarrel Ash K. K., Kyoto, Japan). Another aliquot of formic acid extract from DC implants was ashed with concentrated sulfuric acid and perchloric acid until white fumes appeared, following which it was subjected to phosphorus analysis. Phosphorus concentration was determined by the method of Chen et al.llsl
Total alkaline phosphatase assay Alkaline phosphatase (ALP) activity in the DC implants was determined as described by Reddi and Sullivan. Briefly, ossicles were removed, weighed, and homogenized with a microhomogenizer in ice-cold 0.2% Nonidet P-40 containing 1 mM MgC12, and the homogenate was centrifuged at 20,000 x g for 15 minutes at 4°C. The activity of the supernatant was submitted to assay for alkaline phosphatase using p-nitrophenylphosphate as a substrate in 0.05 M 2-amino-2-methylpropandiol(AMP) buffer (pH 9.8) containing 1 mM MgCI,.
Preparation of BGP in DC implants BGP in DC was released by demineralization with 20% formic acid and purified by gel filtration over a Sephadex (3-25 (fine) column to avoid the loss of BGP by coprecipitation with Ca phosphate salt. The protein fractions from the Sephadex G-25 column were used for radioimmunoassay (RIA). Rat BGP for the RIA standard was purified from rat cortical bone as described by Otawara et al.1")
Radioimmunoassay of BGP All assays were performed with rabbit antiserum to rat BGP and purified rat BGP for standard and tracer as described by Price and Nishimoto.'lnl The assay system modified by Sugimoto et al.llql was employed. All samples were assayed in duplicate on two separate occasions. A total of seven to nine implants from three rats that were implanted for 3-7 weeks were assayed.
Histochemical study Tissues: The tissues were fixed for 20 h at 4°C by immersion in 10% formalin and 1% CaCl, (pH 7.2) and rinsed in cold 0.88 M gum arabic and 30% sucrose overnight. The tissues were then embedded in Tissue-Tek OCT compound (Miles, Inc., Elkhart, IN) without prior demineralization to avoid the release of BGP and frozen at -80°C.
FIG. 1. Light micrographs of histologic sections of the diffusion chambers containing bone marrow cells and DBM 4 and 5 weeks after implantation. The top of the newly formed bone is the surface in close contact with a Millipore membrane filter (white top area in each photograph). (a and c) Hematoxylin and eosin staining, x 100; (b and d) the same sections as in a and c, respectively, staining by the azo dye for alkaline phosphatase, x 100; (a and b) 4 weeks; (c and d) 5 weeks. Bone together with osteocytic lacunae (B) is clearly seen in close proximity to the membrane filter. Cartilage formation (C) is also seen in the section 5 weeks after implantation (c). M indicates DBM. Arrows indicate the osteoblasts that coincide with the appearance of alkaline phosphatase activity (black area in b).
1176
DOH1 ET AL.
Histochemical Staining: For alkaline phosphatase, 6 gm staining of the ribosomal bands. Tritium-labeled sense thick sections were stained according to the method of RNA was synthesized in vitro using SP6 RNA polymerase Watanabe and FishmanIzo1using naphthol AS-MX phos- and ’H-GTP and BGP cDNA inserted into the EcoRl site phate (disodium salt, Sigma Chemical Co., St. Louis, MO) of plasmid SP6/T-7 containing bacteriophages SP6 and and fast red violet LB (5-chloro-4-benzamide-2-methyl-T-7 promoters. l1w Plasmid SP6/T-7/BGP was linearized benzediazonium chloride, Sigma) for a red end product of by a restriction enzyme, BarnH1. The amount of mRNA was monitored by the use of tritium-labeled sense RNA azo dye. For indirect immunostaining, a fluorescein isothiocya- and scanning densitometry. nate (F1TC)-conjugated IgG method was used. Sections were incubated with rabbit antirat BGP purified by affin- Statistical analysis ity chromatography for 60 minutes at 20°C and then incuStatistical comparisons between groups were made using bated with a FlTC-conjugated goat IgG against rabbit IgG for 60 minutes at 20°C. In control sections the primary Student’s t-test. antibody was replaced by 0.1 M sodium phosphate buffer containing a 1:lOO dilution of normal rabbit serum. RESULTS
Extraction and analysis of R N A Total cellular RNA was extracted from DC implants by the guanidine thiocyanate and LiCl method (RNA extraction kit, Amersham Japan, Tokyo) and then separated in a 1.1 Vo (wtlvol) agarose-formaldehyde gel.121’Total RNA (20-130 pg) was obtained from five to six chambers, and 5 gg was loaded on the gel. RNA was transferred to a nylon membrane filter (Hybond-N’; Arnersham, Japan) and hybridized overnight at 42°C with ”P-labeled probe, followed by autoradiography. The cDNA probe for BGP, kindly provided by Dr. P.A. Price (University of California), was the 363 base pair EcoRI fragment of plasmid constructed from pUC8 and cDNA of rat BGPlZz1and was labeled using a multirandom primer oligonucleotide labeling kit (Amersham Japan) and [a-”P]dCTP”” (specific activity of 1 x lo9 cpm/gg DNA). The quantity of RNA extracted was assessed with reference to ethidium bromide
When freshly isolated bone marrow cells inoculated with the DBM in a diffusion chamber were implanted subcutaneously in syngeneic recipient rats for 4-7 weeks, all DC implants showed bone and cartilage formation. A majority of the cell shapes were nondescript fibroblastic cells 3 weeks after subcutaneous implantation and appeared to be proliferating. This was followed by an increase in occasional islands of cartilage, which were intensely stained for ALP activity in DC implant sections (data not shown). At 4 and 5 weeks after implantation, newly formed intermembranous ossification proceeded adjacent to the Millipore filter, and cartilage was found toward the center of the DC (Fig. 1). The staining for alkaline phosphatase of DC implant sections indicates that the enzyme activity is high in the osteoblast-like cells of newly formed bone and is higher at 4 weeks than at 5 weeks after implantation (Fig. l b and d). None of the DC with bone marrow cells alone showed a
FIG. 2. Section of DC implants containing bone marrow cells and DBM 4 weeks after implantation, stained for BGP with an indirect immunofluorescence technique. Arrowheads indicate immunopositive osteoblasts for BGP. M, demineralized bone matrix ( x 400).
1177
OSTEOGENESIS OF BONE MARROW AND BGP GENE EXPRFSSION mineralized matrix even after 4 weeks, but 5 weeks after implantation small foci of osteogenesis were found in 50% or less of the DC implants with marrow cell suspension alone (data not shown). Furthermore, chambers with DBM alone morphologically did not contain calcified tissue but contained tissue fluid at 4 and 6 weeks after implantation. Figure 2 shows that the active osteoblasts were fluoroimmunostained for the anti-BGP at the region adjacent to the newly formed bone of DC implants 4 weeks after implantation. The sections of DC implants at 3 weeks were hardly immunostained for the anti-BGP IgG. To quantitate osteogenesis by histochemical findings, calcium and phosphorus concentration, ALP activity, and BGP concentration within DC were analyzed every week from 3 to 7 weeks after implantation. As shown in Fig. 3, the accumulation of Ca and Pi and an increase in their contents clearly revealed rapid mineralization in contrast to DC with marrow cells alone. The mean molar ratios between calcium and phosphorus in marrow cells and DBM chambers were 2.11. 1.47, and 1.51 at 5, 6, and 7 weeks after implantation, respectively, values that approximated 1.67 as the molar ratio of these elements in hydroxyapatite. During implantation, ALP in bone plaques in DC changed and its maximum activity appeared at approximately 4 weeks (Fig. 4). Figure 5 shows BGP is rapidly accumulated from 5 weeks after implantation, reinforcing the interpretations from a time course of calcium and phosphorus content.
40.
This interpretation was confirmed by BGP gene expression in the DC implants, as mentioned later. In contrast to the composites of marrow cells and DBM, bone formation was little observed within DC with marrow cells alone, and only traces of A L P activity and BGP in the newly formed specimens were detected, A L P activity and BGP in DC with DBM alone were slight (1-2 ng BGP per mg DBM residue) or not detected at all at 4 and 6 weeks after implantation. Figure 6 shows that the BGP gene began to be expressed at 5 weeks, a time when extensive bone formation could be detected, and that its expression was increasing during the later stages of bone formation. This expression of BGP mRNA at the high level seen in week 7 correlated with rapid BGP accumulation. The probe (3’P-labeled BGP cDNA) recognized a size of 580-660 nucleotides in DC implants, similar to RNA from rat c a l ~ a r i a e , indicating ~~~] that the BGP gene within DC was representative of the BGP message in normal bone. Two bands corresponding to ribosomal RNA in Fig. 6 were regarded as nonspecific because they did not appear when mRNA in DC at 7 weeks was purified by oligo-dT-immobilized latex beads (Oligotex-dT30; Nippon Roche Research Center, Kanagawa. Japan; lane 6 in Fig. 6). From the autoradiograms of nylon filter-immobilized synthetic sense RNA of BGP cDNA hybridized with I’Plabeled rat BGP cDNA, densitometric countings of the hybridized probe were linear within the range of 4.6-460 pg of sense RNA by changing the exposure interval (not shown). The absolute levels of hybridizable BGP mRNA
120
1 .r
:30.
(I (I
*
.d
bl
a
G 20.
/
a a
0
10-
Weeks post-implantation FIG. 3. Calcium and phosphorus accumulation inside diffusion chambers explanted 3-7 weeks after implantation. Calcium (left) and phosphorus (right) concentration (&mg implant tissue) is plotted against the time after implantation. Each data point and bar are the mean and standard error of the mean (SEM) of calcium and phosphorus concentrations from eight or nine DC implants for the composites of marrow cells and DBM (open symbols) and two o r three DC implants for marrow cells alone (filled symbols). Asterisks indicate significant differences compared with values of control DC (bone marrow alone). ***p < 0.005; * + p < 0.05; * p < 0.1.
1178
DOH1 ET AL.
e
,,i Q,
40-
3 La u1
.H
' 30M E
\
M
n I
-4 ..
t
I
A ..
3
4
5
6
7
Weeks post-implantation FIG. 4. Alkaline phosphatase activity within diffusion
20I
2 I
It..-.A 10
0
3
4
5
1
6
'
7
Weeks post-implantation
chamber implants. Alkaline phosphatase activities are plotted against time after implantation. Each data point and bar represent the mean and SEM of the enzyme activities determined from eight or nine DC implants for the composites of marrow cells and DBM (open symbols) and two or three DC implants for marrow cells alone (filled symbols). Asterisks indicate significant differences compared with values of control DC (bone marrow alone). ***p < 0.005; * p < 0.05.
FIG. 5. Appearance of BGP in diffusion chamber implants. Each BGP data point is the average of duplicate radioimmunoassay from eight or nine implants for the composites of marrow cells and DBM (open symbols) and three or four DC implants for marrow cells alone (filled symbols). Asterisks indicate significant differences compared with values of control DC (bone marrow alone). ***p c 0.001; **p < 0.01; * p < 0.1.
from DC implant extracts were calculated from the density of autoradiograms compared to the amount of synthetic sense RNA (lane 7 in Fig. 6). At 5 weeks after implantation, only a few pg BGP mRNA was detected in 5 pg total RNA, although this increased markedly by 400 pg at 7 weeks.
BGP parallels closely the accumulation of phosphorus in DC tissue and that the mean C a / P molar ratio is similar to that of hydroxyapatite. Our data contrast with the low content of BGP found in experimentally induced dystrophic calcification, which may involve deposit of calcium Our reand phosphate through a different sults that DC implants from the composites of marrow cells and DBM mineralize in association with high levels of BGP are comparable to the findings of increases in BGP during the onset of in vivo mineralization. Hydroxyapatite crystal formation may play an important role in the synthesis and secretion of BGP. The Occurrence of the peak of A L P activity. which is considered the earliest demonstrable osteogenic marker, (l7] agrees almost exactly with results previously found by Bab et al.(I2]However, the highest A L P activity was reached early as 15 days in DBM-induced ectopic bone formation.(I41 The delay of expression of A L P activity in this study is probably due to a lack of vasculature under the DC culture conditions, as mentioned later. In this study, there was little or no bone formation in the DC with marrow cells only. Mardon et al.(zalobserved excellent osteogenesis in DC with rat marrow cells alone. They employed 2- to 3-week-old rat bone marrow, younger than that we used. Bab et al.(lzlsuggested that marrow cells from younger animals are more effective in the pro-
DISCUSSION The purpose of our study was to quantitate the degree of developmental osteogenesis by bone marrow cells in an in vivo diffusion chamber by measuring biochemical bonespecific parameters, such as alkaline phosphatase activity and bone-specific vitamin K-dependent protein, BGP content, and BGP gene expression. It has been shown that marrow strornal cells give rise t o bone and mineralized cartilage when the cells were cultured in vivo in DC.(II-I'I W e clearly show that all parameters of osteogenic activity within implanted DC ossicles reflects developmental bone and cartilage formation in vivo induced by composites of the stromal cells and demineralized bone matrix. These tissues in DC exhibited high levels of BGP, 1.3-2 pg BGP per mg Ca, which is comparable to the 1-2 pg BGP per mg Ca in the DBM-induced and the 8-15 pg BGP per mg Ca for rat bone.("] It is interesting that the increase in
1179
OSTEOGENESIS OF BONE MARROW AND BGP GENE EXPRESSION
-
*. BG P-
1
2
u*
Osteogenesis Associated with Bone Gla Protein Gene Expression in Diffusion Chambers by Bone Marrow Cells with Demineralized Bone Matrix YOSHIKO DOHI,’ HAJIME OHGUSHI,’ SHIRO TABATA,3 TAKAFUMI YOSHIKAWA,2 KAZUHIRO DOHI.‘ and TADASHIGE MORIYAMA’
ABSTRACT Diffusion chambers with rat bone marrow cells and demineralized bone matrix (DBM) were implanted subcutaneously to syngeneic 8-week-old rats and were harvested every week 3-7 weeks after implantation, and histochemical examination, determination of alkaline phosphatase activity, total calcium and phosphorus, the bone-specific vitamin K-dependent gla-containing protein (BGP) content, and detection of BGP mRNA relative to mineralization were performed. Alkaline phosphatase in diffusion chamber implants reached the highest activity at 4 weeks and then decreased. Calcium and phosphorus deposits occurred at 4 weeks after implantation and were followed by marked increases until 7 weeks, which was comparable to the accumulation of BGP. The BGP gene within the diffusion chambers began to be expressed at 5 weeks, and its expression increased markedly at 7 weeks after implantation. At 4-5 weeks after implantation, new bone adjacent to the membrane filters and cartilage toward the center of the diffusion chamber were observed histochemically. Light microscopic and immunohistologic examinations of chambers with marrow cells and DBM revealed production of mineralized matrices, typical of bone characterized by the appearance of BGP and mineralized nodules. In contrast, bone marrow cells alone did not show extensive bone formation and yielded very low values for these biochemical parameters. The present experiments demonstrate the potential of bone marrow cells and DBM to produce not only cartilage formation but also membranous bone formation associated with increasing expression of BGP mRNA during the later stages of bone formation, as well as a marked accumulation of BGP.
INTRODUCTION
F
at various stages of development is an active process involving the synthesis of connective tissue macromolecules (collagen, bone-specific proteins, and proteoglycans) by osteoblasts, followed by the deposit of a calcium phosphate mineral phase.“-‘] A number of subsequent studies on assay methods for this highly regulated process of osteogenesis have been reported. On the one hand, a demineralized bone matrix (DBM) implanted into various extraskeletal sites induces the differentiation ORMATION OF BONE
to cartilage and bone by endochondral ossification in host tissues.‘s,61The factor responsible for this induction of osteogenesis was found in demineralized bone matrix to be a protein having an approximate molecular weight of 18,000 and has been termed “bone morphogenetic protein” (BMP) .I n 1 On the other hand, it is known that when cells obtained from the bone marrow are transplanted to heterotopic sites, they differentiate to form bone and cartilage. f7.11 These cells have been designated “determined osteogenic precursor cells” (DOPC). Recently, Ohgushi et al. demon-
‘Department of Public Health, Nara Medical University, Japan. ’Department of Orthopaedics. Nara Medical University, Japan. ’Department of Chemistry, Nara Medical University, Japan. ‘First Department of Internal Medicine, Nara Medical University, Japan.
1173
DOH1 ET AL.
1174
strated heterotopic osteogenesis associated with the developmental appearance of the bone-specific vitamin K-dependent gla-containing protein (BGP) in porous ceramics induced by bone marrow cell^.^^.^^^ Diffusion chamber implantation techniques have also been employed to characterize the bone marrow osteogenic precursor cell population'11,121 and to examine the effect of a partially purified BMP on the ability of bone marrow cells to differentiate to form cartilage and bone.'131 The in vivo diffusion chamber method appears to be reproducible as an assay for various stages of osteogenesis with the content of the chambers being isolated from cells of the recipient, which d o not cross the diffusion chamber membrane filter. Therefore, in this study, we demonstrated the ability of bone marrow cells and DBM on differentiation to form cartilage and bone within a diffusion chamber by monitoring various biochemical parameters, including assay of BGP (osteocalcin) and time-dependent expression of BGP mRNA.
MATERIALS AND METHODS Preparation of demineralized bone matrix Demineralized bone matrix was prepared as described by Nishimoto et al.l1') Briefly, the femora and tibiae from 3-month-old male Wistar rats (about 300 g body weight) were removed and cleaned of soft tissue and marrow, frozen in liquid nitrogen, and pulverized with a Hyscotron homogenizer. The bone powder (74-420 pn) was demineralized with 0.5 N HCl for 1 h at 4°C three times and washed out exhaustively with distilled water up to pH 6 , followed by washing with 95% ethanol three times and then rinsing with ether. The bone powder was air dried overnight at room temperature to obtain the granular demineralized bone matrix and then stored at -25°C.
Marrow cell suspension Marrow cell suspensions from femora of 8-week-old male Wistar rats (about 200 g body weight) were prepared as described by Ohgushi et al.lQ1A cell suspension of the marrow was obtained by pumping the tissue repeatedly through hypodermic needles of decreasing gauge. After these procedures, >90% of the cells were viable judging by dye exclusion methods.
Preparation and implantation of diffusion chambers Diffusion chambers (DC; 9 mm inner diameter x 2 mm thick, 130 ~l capacity, 0.45 pm pore size membrane filters) were assembled from commercially available components (Nihon Millipore, Kogyo K.K., Yonezawa, Japan) and sterilized with ethylene oxide. Marrow cell suspensions (5 x loLcells per 90 p1) and DBM (10 mg) were injected into DC, which were then sealed with Millipore filter cement. Three DCs containing rat marrow cells and DBM were implanted subcutaneously into the backs of 8-week-old male Wistar rats weighing 190-210 g. The DCs were harvested every week from 3 to 7 weeks after implantation. DCs that
contained only marrow cells or 10 mg demineralized bone matrix alone were implanted into rats as control experiments.
Calcium and phosphorus determination Bone plaques in DC or bone samples were weighed and then homogenized in 20% formic acid. An aliquot of the formic acid extract was diluted 1:200 with a strontium solution. Calcium concentration was determined by an atomic absorption spectrometer (AA-810 type; Nippon Jarrel Ash K. K., Kyoto, Japan). Another aliquot of formic acid extract from DC implants was ashed with concentrated sulfuric acid and perchloric acid until white fumes appeared, following which it was subjected to phosphorus analysis. Phosphorus concentration was determined by the method of Chen et al.llsl
Total alkaline phosphatase assay Alkaline phosphatase (ALP) activity in the DC implants was determined as described by Reddi and Sullivan. Briefly, ossicles were removed, weighed, and homogenized with a microhomogenizer in ice-cold 0.2% Nonidet P-40 containing 1 mM MgC12, and the homogenate was centrifuged at 20,000 x g for 15 minutes at 4°C. The activity of the supernatant was submitted to assay for alkaline phosphatase using p-nitrophenylphosphate as a substrate in 0.05 M 2-amino-2-methylpropandiol(AMP) buffer (pH 9.8) containing 1 mM MgCI,.
Preparation of BGP in DC implants BGP in DC was released by demineralization with 20% formic acid and purified by gel filtration over a Sephadex (3-25 (fine) column to avoid the loss of BGP by coprecipitation with Ca phosphate salt. The protein fractions from the Sephadex G-25 column were used for radioimmunoassay (RIA). Rat BGP for the RIA standard was purified from rat cortical bone as described by Otawara et al.1")
Radioimmunoassay of BGP All assays were performed with rabbit antiserum to rat BGP and purified rat BGP for standard and tracer as described by Price and Nishimoto.'lnl The assay system modified by Sugimoto et al.llql was employed. All samples were assayed in duplicate on two separate occasions. A total of seven to nine implants from three rats that were implanted for 3-7 weeks were assayed.
Histochemical study Tissues: The tissues were fixed for 20 h at 4°C by immersion in 10% formalin and 1% CaCl, (pH 7.2) and rinsed in cold 0.88 M gum arabic and 30% sucrose overnight. The tissues were then embedded in Tissue-Tek OCT compound (Miles, Inc., Elkhart, IN) without prior demineralization to avoid the release of BGP and frozen at -80°C.
FIG. 1. Light micrographs of histologic sections of the diffusion chambers containing bone marrow cells and DBM 4 and 5 weeks after implantation. The top of the newly formed bone is the surface in close contact with a Millipore membrane filter (white top area in each photograph). (a and c) Hematoxylin and eosin staining, x 100; (b and d) the same sections as in a and c, respectively, staining by the azo dye for alkaline phosphatase, x 100; (a and b) 4 weeks; (c and d) 5 weeks. Bone together with osteocytic lacunae (B) is clearly seen in close proximity to the membrane filter. Cartilage formation (C) is also seen in the section 5 weeks after implantation (c). M indicates DBM. Arrows indicate the osteoblasts that coincide with the appearance of alkaline phosphatase activity (black area in b).
1176
DOH1 ET AL.
Histochemical Staining: For alkaline phosphatase, 6 gm staining of the ribosomal bands. Tritium-labeled sense thick sections were stained according to the method of RNA was synthesized in vitro using SP6 RNA polymerase Watanabe and FishmanIzo1using naphthol AS-MX phos- and ’H-GTP and BGP cDNA inserted into the EcoRl site phate (disodium salt, Sigma Chemical Co., St. Louis, MO) of plasmid SP6/T-7 containing bacteriophages SP6 and and fast red violet LB (5-chloro-4-benzamide-2-methyl-T-7 promoters. l1w Plasmid SP6/T-7/BGP was linearized benzediazonium chloride, Sigma) for a red end product of by a restriction enzyme, BarnH1. The amount of mRNA was monitored by the use of tritium-labeled sense RNA azo dye. For indirect immunostaining, a fluorescein isothiocya- and scanning densitometry. nate (F1TC)-conjugated IgG method was used. Sections were incubated with rabbit antirat BGP purified by affin- Statistical analysis ity chromatography for 60 minutes at 20°C and then incuStatistical comparisons between groups were made using bated with a FlTC-conjugated goat IgG against rabbit IgG for 60 minutes at 20°C. In control sections the primary Student’s t-test. antibody was replaced by 0.1 M sodium phosphate buffer containing a 1:lOO dilution of normal rabbit serum. RESULTS
Extraction and analysis of R N A Total cellular RNA was extracted from DC implants by the guanidine thiocyanate and LiCl method (RNA extraction kit, Amersham Japan, Tokyo) and then separated in a 1.1 Vo (wtlvol) agarose-formaldehyde gel.121’Total RNA (20-130 pg) was obtained from five to six chambers, and 5 gg was loaded on the gel. RNA was transferred to a nylon membrane filter (Hybond-N’; Arnersham, Japan) and hybridized overnight at 42°C with ”P-labeled probe, followed by autoradiography. The cDNA probe for BGP, kindly provided by Dr. P.A. Price (University of California), was the 363 base pair EcoRI fragment of plasmid constructed from pUC8 and cDNA of rat BGPlZz1and was labeled using a multirandom primer oligonucleotide labeling kit (Amersham Japan) and [a-”P]dCTP”” (specific activity of 1 x lo9 cpm/gg DNA). The quantity of RNA extracted was assessed with reference to ethidium bromide
When freshly isolated bone marrow cells inoculated with the DBM in a diffusion chamber were implanted subcutaneously in syngeneic recipient rats for 4-7 weeks, all DC implants showed bone and cartilage formation. A majority of the cell shapes were nondescript fibroblastic cells 3 weeks after subcutaneous implantation and appeared to be proliferating. This was followed by an increase in occasional islands of cartilage, which were intensely stained for ALP activity in DC implant sections (data not shown). At 4 and 5 weeks after implantation, newly formed intermembranous ossification proceeded adjacent to the Millipore filter, and cartilage was found toward the center of the DC (Fig. 1). The staining for alkaline phosphatase of DC implant sections indicates that the enzyme activity is high in the osteoblast-like cells of newly formed bone and is higher at 4 weeks than at 5 weeks after implantation (Fig. l b and d). None of the DC with bone marrow cells alone showed a
FIG. 2. Section of DC implants containing bone marrow cells and DBM 4 weeks after implantation, stained for BGP with an indirect immunofluorescence technique. Arrowheads indicate immunopositive osteoblasts for BGP. M, demineralized bone matrix ( x 400).
1177
OSTEOGENESIS OF BONE MARROW AND BGP GENE EXPRFSSION mineralized matrix even after 4 weeks, but 5 weeks after implantation small foci of osteogenesis were found in 50% or less of the DC implants with marrow cell suspension alone (data not shown). Furthermore, chambers with DBM alone morphologically did not contain calcified tissue but contained tissue fluid at 4 and 6 weeks after implantation. Figure 2 shows that the active osteoblasts were fluoroimmunostained for the anti-BGP at the region adjacent to the newly formed bone of DC implants 4 weeks after implantation. The sections of DC implants at 3 weeks were hardly immunostained for the anti-BGP IgG. To quantitate osteogenesis by histochemical findings, calcium and phosphorus concentration, ALP activity, and BGP concentration within DC were analyzed every week from 3 to 7 weeks after implantation. As shown in Fig. 3, the accumulation of Ca and Pi and an increase in their contents clearly revealed rapid mineralization in contrast to DC with marrow cells alone. The mean molar ratios between calcium and phosphorus in marrow cells and DBM chambers were 2.11. 1.47, and 1.51 at 5, 6, and 7 weeks after implantation, respectively, values that approximated 1.67 as the molar ratio of these elements in hydroxyapatite. During implantation, ALP in bone plaques in DC changed and its maximum activity appeared at approximately 4 weeks (Fig. 4). Figure 5 shows BGP is rapidly accumulated from 5 weeks after implantation, reinforcing the interpretations from a time course of calcium and phosphorus content.
40.
This interpretation was confirmed by BGP gene expression in the DC implants, as mentioned later. In contrast to the composites of marrow cells and DBM, bone formation was little observed within DC with marrow cells alone, and only traces of A L P activity and BGP in the newly formed specimens were detected, A L P activity and BGP in DC with DBM alone were slight (1-2 ng BGP per mg DBM residue) or not detected at all at 4 and 6 weeks after implantation. Figure 6 shows that the BGP gene began to be expressed at 5 weeks, a time when extensive bone formation could be detected, and that its expression was increasing during the later stages of bone formation. This expression of BGP mRNA at the high level seen in week 7 correlated with rapid BGP accumulation. The probe (3’P-labeled BGP cDNA) recognized a size of 580-660 nucleotides in DC implants, similar to RNA from rat c a l ~ a r i a e , indicating ~~~] that the BGP gene within DC was representative of the BGP message in normal bone. Two bands corresponding to ribosomal RNA in Fig. 6 were regarded as nonspecific because they did not appear when mRNA in DC at 7 weeks was purified by oligo-dT-immobilized latex beads (Oligotex-dT30; Nippon Roche Research Center, Kanagawa. Japan; lane 6 in Fig. 6). From the autoradiograms of nylon filter-immobilized synthetic sense RNA of BGP cDNA hybridized with I’Plabeled rat BGP cDNA, densitometric countings of the hybridized probe were linear within the range of 4.6-460 pg of sense RNA by changing the exposure interval (not shown). The absolute levels of hybridizable BGP mRNA
120
1 .r
:30.
(I (I
*
.d
bl
a
G 20.
/
a a
0
10-
Weeks post-implantation FIG. 3. Calcium and phosphorus accumulation inside diffusion chambers explanted 3-7 weeks after implantation. Calcium (left) and phosphorus (right) concentration (&mg implant tissue) is plotted against the time after implantation. Each data point and bar are the mean and standard error of the mean (SEM) of calcium and phosphorus concentrations from eight or nine DC implants for the composites of marrow cells and DBM (open symbols) and two o r three DC implants for marrow cells alone (filled symbols). Asterisks indicate significant differences compared with values of control DC (bone marrow alone). ***p < 0.005; * + p < 0.05; * p < 0.1.
1178
DOH1 ET AL.
e
,,i Q,
40-
3 La u1
.H
' 30M E
\
M
n I
-4 ..
t
I
A ..
3
4
5
6
7
Weeks post-implantation FIG. 4. Alkaline phosphatase activity within diffusion
20I
2 I
It..-.A 10
0
3
4
5
1
6
'
7
Weeks post-implantation
chamber implants. Alkaline phosphatase activities are plotted against time after implantation. Each data point and bar represent the mean and SEM of the enzyme activities determined from eight or nine DC implants for the composites of marrow cells and DBM (open symbols) and two or three DC implants for marrow cells alone (filled symbols). Asterisks indicate significant differences compared with values of control DC (bone marrow alone). ***p < 0.005; * p < 0.05.
FIG. 5. Appearance of BGP in diffusion chamber implants. Each BGP data point is the average of duplicate radioimmunoassay from eight or nine implants for the composites of marrow cells and DBM (open symbols) and three or four DC implants for marrow cells alone (filled symbols). Asterisks indicate significant differences compared with values of control DC (bone marrow alone). ***p c 0.001; **p < 0.01; * p < 0.1.
from DC implant extracts were calculated from the density of autoradiograms compared to the amount of synthetic sense RNA (lane 7 in Fig. 6). At 5 weeks after implantation, only a few pg BGP mRNA was detected in 5 pg total RNA, although this increased markedly by 400 pg at 7 weeks.
BGP parallels closely the accumulation of phosphorus in DC tissue and that the mean C a / P molar ratio is similar to that of hydroxyapatite. Our data contrast with the low content of BGP found in experimentally induced dystrophic calcification, which may involve deposit of calcium Our reand phosphate through a different sults that DC implants from the composites of marrow cells and DBM mineralize in association with high levels of BGP are comparable to the findings of increases in BGP during the onset of in vivo mineralization. Hydroxyapatite crystal formation may play an important role in the synthesis and secretion of BGP. The Occurrence of the peak of A L P activity. which is considered the earliest demonstrable osteogenic marker, (l7] agrees almost exactly with results previously found by Bab et al.(I2]However, the highest A L P activity was reached early as 15 days in DBM-induced ectopic bone formation.(I41 The delay of expression of A L P activity in this study is probably due to a lack of vasculature under the DC culture conditions, as mentioned later. In this study, there was little or no bone formation in the DC with marrow cells only. Mardon et al.(zalobserved excellent osteogenesis in DC with rat marrow cells alone. They employed 2- to 3-week-old rat bone marrow, younger than that we used. Bab et al.(lzlsuggested that marrow cells from younger animals are more effective in the pro-
DISCUSSION The purpose of our study was to quantitate the degree of developmental osteogenesis by bone marrow cells in an in vivo diffusion chamber by measuring biochemical bonespecific parameters, such as alkaline phosphatase activity and bone-specific vitamin K-dependent protein, BGP content, and BGP gene expression. It has been shown that marrow strornal cells give rise t o bone and mineralized cartilage when the cells were cultured in vivo in DC.(II-I'I W e clearly show that all parameters of osteogenic activity within implanted DC ossicles reflects developmental bone and cartilage formation in vivo induced by composites of the stromal cells and demineralized bone matrix. These tissues in DC exhibited high levels of BGP, 1.3-2 pg BGP per mg Ca, which is comparable to the 1-2 pg BGP per mg Ca in the DBM-induced and the 8-15 pg BGP per mg Ca for rat bone.("] It is interesting that the increase in
1179
OSTEOGENESIS OF BONE MARROW AND BGP GENE EXPRESSION
-
*. BG P-
1
2
u*
