Changes in the distribution of extracellular matrix vesicles during healing of rat tibial bone (computerized morphometry and electron microscopy)

8756-3282187 $3.00 + .OO Copyright 0 1987 Pergamon Journals Ltd.

Bone, 8, 245-250 (1987) Printed in the USA. All rights reserved

Changes in the Distribution of Extracellular Matrix Vesicles during Healing of Rat Tibia1 Bone (Computerized Morphometry and Electron Microscopy) J. SELA,’ D. AMIR,2 Z. SCHWARTZ3 and H. WEINBERG2 1 Departments of Oral Pathology, 3 Periodontology and * Orthopedic Surgery, Mt. Scopus.

The Hebrew University and Hadassah

Address for correspondence and reprints: Professor J. Sela, Department School of Dental Medicine, P O.B. 1.172, Jerusalem 91010, Israel.

Schools of Dental Medicine, of Oral Pathology,

and Medione

The Hebrew

University

and Hadassah

Abstract

Introduction

A study of the distribution of extracellular matrix vesicles during the first 3 weeks of healing of adult rat tibia1 bone was performed by transmission electron microscopy in combination with computerized morphometry. Bone injury comprised removal of the marrow followed by regeneration of the tissue via a phase of primary mineralization. A total number of 39,498 vesicles were traced on electron micrographs and sorted according to their diameters, distance from the calcified front and types. The different vesicular types were defined as follows: (a) vesicles with electron lucent contents, i.e. “empty”, (b) vesicles with amorphous electron opaque contents, i.e. “amorphic”, (c) vesicles containinhg crystalline depositions, i.e. “crystal”, and (d) vesicles containing crystalline structures with ruptured membranes i.e. “rupture”. The vesicles were studied on the days 3, 6, 14 and 21 after bone injury. Most of the vesicles were concentrated between diameters of 0.07 and 0.17 pm. Most of the vesicles were found in a distance less than 3 km from the calcified front. The sequence of changes of distances from the calcified front and of the vesicular diameters were recorded as follows: “rupture”, “crystal”, “amorphic” and “empty”, the “rupture” type being the closest to the front and of the largest diameter in each day. The results of the present study confirm the accepted hypothesis on calcification via extracellular matrix vesicles. It is thought that the cell secretes “empty” vesicles that accumulate Ca and Pi forming amorphous calcium phosphate that is then converted to hydroxyapatite. This is followed by rupture of the vesicular membrane. The maturation of the process herein on each day is accompanied by increase in the vesicular diameter and its approximation to the calcified front. With the propagation of healing the vesicles of all types decreased in diameter and distance from the calcified front. This process is accompanied by increase in the number of the mature vesicles and decrease in the number of the young vesicles.

It has been repeatedly reported in both normal and pathological conditions that extracellular matrix vesicles have an important role in primary mineralization in cartilage, bone and dentin. It must be stated herein that calcification via matrix vesicles is not the exclusive mechanism of crystal formation in vertebrate tissues (Anderson, 1967; Bonucci, 1967; Anderson, 1976; Boskey, 1981; Wuthier, 1982; Ali, 1983). The occurrence and activity of matrix vesicles have been studied by ultrastructural and biochemical methods. It is accepted that chondroblasts, osteoblasts and odontoblasts may be directly involved in the initial phase of mineralization via matrix vesicles. The understanding of vesicular calcification is supported by data based on studies of the in vitro enzymatic activity. Alkaline phosphatase (ALPase) activity was the first to be detected. Hydrolysis of different phosphate compounds by ALPase is followed by an increase in the concentration of orthophosphate which is essential for crystal formation. Since the activity of ALPase is enriched in this extracellular organelle it has been accepted as the vesicular marker enzyme (Ali and Evans, 1973; Ali, 1976; Ali et al., 1976; Felix and Fleisch, 1976; Hsu and Anderson, 1978; Boskey, 1981; Wuthier, 1982; Ali, 1983; Anderson, 1985). In addition, high levels of ATPase activity were found in isolated vesicles. The action of ATPase could be of a two-fold nature, increase of intravesicular Ca and hydrolysis of ATP to remove its inhibitory effect on the transformation of amorphous calcium phosphate to hydroxyapatite (HA) crystals. Different vesicular phosphatases hydrolyse pyrophosphates which are also potent inhibitors of calcification. Finally, the participation of a calcium binding protein in the transport of Ca2+ through the vesicle membrane has been suggested. A protein of such potency could act to increase intravesicular concentration of ionic calcium (Ali, 1980; Ali and Evans, 1981). The current “working-hypothesis” states that extracellular matrix vesicles have biochemical properties that serve the process of formation of the initial HA crystal (Ali, 1983). There is a wide consensus with regard to the interpretation of ultrastructural findings during the first stages of calcification. It is generally accepted that the secretion of an electron lucent vesicle by the cell is followed by ve-

Key Words: Matrix Vesicle Distribution-Bone TEM-Morphometry.

Healing 245

246

J Sela et al. Distnbution

sicular loading with calcium and phosphate. lntravesicular formation of calcium-phosphate complexes is responsible for an electron opaque phase. The appearance of a nidus of calcification on the internal membranal surface of the opaque vesicle is believed to be followed by the formation of an HA crystal. Crystalline structures could be found in numerous electron micrographs of matrix vesicles in different positions in the vesicular wall (Ali, 1983). It is thought that crystal growth results with the rupture of the vesicle membrane. HA crystals adhere to each other to form calcospheritic structures. Secondary nucleation is thought to be a mechanism guided by the collagen. Electrically charged crystallites align alongside collagen fibrils to incorporate in “hole sites” (Wuthier, 1982). it must be stated that direct evidence to the stages agreed upon as the accepted sequence of calcification via extracellular matrrx vesicles, was still lacking. Quantitative evaluation of the vesicular content of different isolated fractions has been recently performed by morphometric analysis of electron micrographs. A high correlation was found between the percentage of area occupied by vesicles with electron-dense content and the enzymatic activity. Highest enzymatic specific activities and electron-dense vesicle fractional area were recorded in a ‘light” vesicle enriched fraction. These parameters revealed lowest values in a “heavy” vesicle enriched fraction (Bab et al., 1983). In addition, differences in the vesicular behaviour during healing were detected by biochemical studies. Increased ALPase activity was found on the 9th day of socket healing and a recovery to normal levels was recorded on the 22nd day after exodontia (Muhlrad et al., 1981). The vesicular size distribution in growth cartilage was recently studied by morphometric methods. It was demonstrated that in rickets, the size distribution of matrix vesicles in the lower half of the growth plate manifested an increased number of smaller vesicles. The conclusion was that matnx vesicle degradation is not sufficient to start mineralization and that a possible subpopulation of larger vesicles have a role in the process of mineralization (Englfeldt et al, 1985). The

Fig. 1. Electron micrograph 0.2 pm).

of a typical field of primary calcification.

of matrrx vesrcles in healrng bone

purpose of the present study was to examrne the working hypothesis of the sequence of matrix vesicle calcrfrcation during bone healing by computerized tissue morphometnc electron microscopy. The model of bone healing comprised removal of marrow followed by regeneration of the tissue preceded by endosteal ossification (Amsel et al., 1969; Bab et al., 1985; Patt and Malloney, 1985).

Materials and Methods Thrrty-two male rats of the Hebrew Unrversrty (Sabra) strarn, weighing 4009 each, were used for the experiment The rats were divided Into 4 groups 8 anrmals each marked a, b, c and d Animals were operated under ether anesthesra An infrapatellar rncrsion was performed to achieve access to the proximal aspect of the tibia1 bone. This was followed by penetratron of the frontal aspect of the bone with a safrne-cooled round dental burr (No 4) using a 20.000 RPM motor The bone marrow was evacuated by repeated washtngs with saline that was Introduced to the intrabony space by a cannule. The skrn operation wounds were then sutured and the 4 groups a, b, c and d were kept for 3, 6. 14 and 21 days, respectrvely. This was followed by removal of the trbial bone under ether anesthesra and evacuation of the contents of the marrow space by qection and curretage under a constant flow of 2% glutaraldehyde in phosphate buffer solutron The tissue was collected and each subgroup of two rats In the drfferent groups was pooled for further idem fixation and processed for transmisson electron microscopy as previously described (Bab et al., 1983). In short, post fixation was performed by 1% osmrum tetraoxide in 0.1 M cacodylate buffer followed by stainrng with 1% uranyl acetate in 0 1 M veronal buffer and lead crtrate Ultrathrn sections were studted with a Philrps EM 300 electron mrcroscope Ten randomly chosen electron mrcroscopic fields of primary calcrfication were selected from each of the pooled samples (Fig. 1). Ten consecutrve transmission electron micrographs were obtarned from each field This was based on a prelrmrnary study that has shown that 100 electron micrographs contain more than 500 vesicles in this experimental setup At the completion of thrs step 400 electron mrcroscope fields for each group (a total of 1600) were performed on 9.3 x 7.8 cm cut down films at magnrfrcation x25.700. Thrs was followed by printing of electron micrographs on Kodabromide photographic paper (Eastman Kodak Co. Roch-

C = cell, M = matrix with calcosphentes,

F = calcified front (Bar =

J. Sela et al.: Distribution

of matrix vesicles

Fig. 2. Higher magnification (Bar = 0.i km).

of collagen

rich extracellular

matrix with different types of vesicles E = empty, C = crystal and I3=i ‘Ulsture

ester, N.Y. 14650, USA). With further x2.5 magnification to achieve a blow-up of x 64.250. Preparation of two identical prints from each electron micrograph completed this step. Extracellular matrix vesicles of diameters ranging between 0.02-0.35 pm were detected on each electron micrograph. In instances of disagreement of 5% or more between the two examiners a third examiner was engaged to decide on the structures in doubt. The examined vesicles were divided into 4 types: A. “empty” vesicles, namely electron lucent organelles. B. “amorphic” vesicles, I.e. vesicular structures containing electron densities other than crystalline. C. “crystal” vesicles, namely vesicles containing crystalline forms, and D. “rupture” vesicles, i.e. vesicles with ruptured membranes containing crystalline forms (Figs. 2 and 3). The outlines of matrix vesicles and the distance between the vesicles and the calcified front were traced. This was performed by a digitizer connected to

Flg. 3. Extracellular

247

in healing bone

an a-Micro computer (a-Micro System Inc. Irvine, Calif. USA). The parameters for each vesicle as recorded on the electron micrograph included the diameter in microns as calculated from the area, the distance in microns between the vesicles and the calcified front and the vesicular type. The results are based on a total pool of data obtained from four pooled samples for each group following a variance test that revealed no significant difference between the four pooled samples. These were processed to calculate the following relationships: distributions of vesicles according to diameter, distance from calcified fronts and type. The distribution of vesicles according to diameters was calculated as based upon 6 subgroups, i e. CO.02 pm, 0.02-O 07 pm, 0.07-0.12 pm, 0.12-O 17 pm. 0 17-0.22 km and >0.22 km. The distribution of vesicles according to their distance from the calcrfied front was based on 6 subgroups, i.e ~0 5 km, 0.5-l pm,

matrix with vesicles A = amorphic,

R = rupture (Bar = 0.1 Fm)

248

J Sela et al

1-l 5 km, 1 S-2 km, 2-3 pm and >3 km. In addition, the dlstributions of types of vesicles according to number, diameter and distance from calcified fronts were calculated, respectively. Calculation of the correlations between diameters and distance from the calcified front were performed. Statistical evaluation was performed in each instance according to the following mean and standard error of the mean (SEM) The differences between groups and days were tested by Student’s t-test with level of slgnificance of P < 0.05. In addition, the correlation between the different parameters was tested between the groups by regression analysis The statistical studies were processed automatically by the computer after Input of the raw data through the digitizer

Distribution

of matrix vesicles

in healing bone

1.1 -

l-

.9-

.8 -

Results .7 -

A total number of 39.498 vesicles were studled in 1600 electron micrographs representing different fields. The vesicles in each one of the four groups manifested a normal distribution with regard to their diameters. A study of the different groups revealed that the diameters of more than 95% of the vesicles ranged between 0.07 to 0.17 pm. The mean vesicular diameter according to the day of treatment on the day% 3, 14, and 21 was not significantly different and ranged between 0.09 and 0.1 pm on the 6th day the diameter was significantly higher (0.1 14 rJ,m) (Fig. 4). The largest percentage of vesicles in each group was detected near the calcified front. More than 95% of the vesicles were found up to a distance of 3 pm from the front. The distribution of the vesicular mean distance from the calcified fronts according to the days after treatment is presented in Fig. 5. The largest mean vesicular distance from the front was found on the sixth day after treatment. The vesicles were found significantly nearer to the front on the 3rd day and significantly closer on the 14th and 21st days with no difference between the last two groups. Distribution of vesicular diameters according to their types and day after treatment are presented in Figure 6. The

.6 -

.5

_

.4’ Ll

I 6

I 3

I 9

I 15

I 12

0PY.S QFTER

1 18

24

TREFlTMENT

Fig. 5. Mean vesicular &stance from the calcified front accordrng to day after treatment. Each point represents mean i SE of all vesicles measured

patterns of the distrlbutlons of the diameters according to days among the different types showed a similar pattern to the one presented in Figure 4. With regard to vesicular diameter on the third day the sequence was recorded as follows “empty”, “amorphoic”, “crystal” and “rupture ‘, the last group being the largest. On the sixth day there was no

H

.I21

I

I 21

Er4PTT

l -..~RMORPHIC

.I4

X--X A-

CRTSTAL -A

RUPTURE

.12

.I

.OE I-

.Ob’

0

I

I

I

1

,

I

L

3

6

9

12

15

IB

21

DFIYS RFTER

.06 24

TREFITMENT

Fig. 4. Mean vesicular diameter according to day after treatment Each point represents mean 5 SE of all vesicles measured

O-

I

I

/

I

I

I

I

3

6

9

12

15

18

21

DFlYS

FiFTER

TREDTMENT

Fig. 6. Vesicular diameter according to type and day atter treatment Each point represents the mean -c SE

J. Sela et al.: Distribution

of matrix vesicles

249

in healing

difference between the “empty” and “amorphic” groups or between the “crystal” and “rupture” groups, however the crystal and rupture were significantly larger than the “empty” and “amorphic” groups. On the 14th day the rupture group was significantly larger than all other type groups which in turn were of a similar diameter. On the 21st day the rupture group manifested a significantly larger diameter when compared to the crystal. Both rupture and crystal showed a significantly larger diameter when compared to the empty and amorphic groups than in turn showed no significant difference. Figure 7 represents the distribution of the vesicular distance from the calcified front according to the day after treatment and vesicular type. The patterns of vesicular distance from the front according to the different types showed a high similarity to the one found in Figure 5. On the third day the empty, amorphic and crystal groups were of a similar distance from the front, the rupture being significantly closer. On the sixth day the sequence of distances from the calcified fronts was recorded as follows: empty, amorphic, crystal and rupture, the rupture group being the closest to the front. On the 14th day the rupture group was significantly closer to the front than the crystal group. The two groups being significantly closer than the empty and amorphic groups, no significant difference could be found between the two last groups. The empty and amorphic groups on days 14 and 21 with regard to their distance from the front showed no differences. On the 21st day the crystal and rupture groups were significantly closer to the front than the amorphic and empty groups the rupture group being significantly the closest. Figure 8 represents the numbers of vesicles according to types and day after treatment. In this plot two main patterns of vesicular distribution are shown. One pattern comprises the empty and amorphic types showing a continuous decrease of the numbers (in

1.6 ~EMPTT

HEtiPTT .---*AMORPHIC X--KCRTSTRL &--&RUPTURE

DFlYS

RFTER

TREFlTMENT

Fig. 8. Number of vesicles (In percent) according after treatment.

to type and day

percent) with the day after treatment. The amorphic group was larger than the empty all through the experiment. The second pattern comprises the crystal and rupture groups. These show an increase in numbers throughout the experiment. The crystal being larger than the rupture except for the third day after treatment. After the 14th day a decrease in the number of the crystal and an increase in the number of rupture, yet, the crystal showing the higher value between these two.

l ..-*AnORPnIC

1.2

X--K

CRTSTRL

A--A

RUPTURE

Discussion

I

.8

.6

c 0

I

/

I

I

I

I

I

3

6

9

12

15

18

21

DRYS

FlFTER

J

TREFlTMENT

Fig. 7. Vesicular distance from the calcified front according to type and day after treatment. Each point represents the mean k SE

The distribution of the total number of vesicles studied according to their diameters in each group manifested a normal pattern. This was characterized by concentration of more than 95% of the vesicles in the range of 0.07-0.17 km in diameter. Our biological model was based on primary mineralization in endosteal bone formation devoid of cartilaginous phase (Amsel et al., 1969; Bab et al., 1985; Patt and Malloney, 1975). The present findings are in agreement with a previous study which demonstrated that in young rat jaw bones the diameter range of over 95% of the vesicles was below 0.2 km. Although vesicles obtained from jaw bones were studied in isolated fractions and the study of vesicles in this model in the tibia1 bones was based on tissue morphometry, the observations corroborate each other. Therefore it could be suggested that isolation of the vesicles in our previous study did not result in morphological changes (Bab et al., 1983). Studies on epiphyseal growth plate cartilage stated that their diameters ranged between 0.05-0.45 km (Anderson, 1969; Ali et al., 1970). However tissue morphometry revealed diameters between 0.05 and 0.1 km (Reinholt et al., 1983). The mean vesicular diameter in the present study was larger than 0.09 Km and smaller than 0.12 km. The differences

250

between our data and those found in cartilage could be attributed to either that the vesicles herein are of bone tissue or that the vesicles are typical to the healing process. The differences between the vesicular diameters observed in cartilage and bone must be subjected to further studies. The mean vesicular distance from the calcified fronts was less than 1.2 pm, yet, on the 14th and 21st days it manifested a value smaller than 0.7 pm. A similar pattern could be detected with regard to the vesicular diameter. The 6th day manifested largest and most remote vesicles with regard to the front. With progress of healing, on days 14 and 21 these values decreased. Day 3 manifested smallest diameters and a distance similar to the one found on the 6th day. This increase can be correlated to changes in ALPase enzymatic activity in the different stages of alveolar socket healing (Muhlrad et al, 1981). It was clearly demonstrated herein that the different types of vesicles manifested similar patterns of development with regard to their diameter and distance from the calcified front represented in the mean value patterns. Moreover, two different pairs of types could be distinguished i.e. empty and amorphic as the first pair and crystal and rupture as the second pair. The first pair manifested smaller diameters, larger distances from the calcified front and a decrease in numbers throughout the healing process when compared to the second pair. The second pair showed an increase in numbers throughout healing except for the 14th day after which the crystal showed a relative decrease and the rupture a relative increase. In summary, with the propagation of healing the vesicles of all types decreased in diameter and distance from the calcified front. This process is accompanied by increase in the number of the mature vesicles and decrease in the number of the young vesicles. No correlation could be recorded between the diameters and the distance of the matrix vesicles from the calcified front in each group. It must be pointed out that due to the obliquity of sections one could apply a correction factor of cosin 45” (0.707) to the present results (Frost, 1976). A general pattern of vesicular enlargement was observed during each day throughout the maturation of the process of crystallization. This was characterized by a tendency of increased diameters of the rupture and crystal as compared to empty and amorphic. It is suggested that this process is dependent upon osmotic changes. The study of the distances of the different types of vesicles from the calcified front revealed that the vesicles with ruptured membranes were closest to the front followed by vesicles containing HA crystals, amorphic and the empty vesicles being most distant from the front. This confirms the accepted hypothesis on calcification via matrix vesicles (Wuthier, 1982; Ali, 1983). It could be concluded thereupon that the cell secretes the empty vesicles, once in the extracellular matrix, these accumulate amorphous Ca and Pi to form HA crystals which in turn rupture the vesicular membranes. The application of tissue morphometry combined with electron microscopy in the present study was performed herein on healing bone for the first time. The results show a better understanding of the mechanism of primary mineralization via extracellular matrix vesicles.

References Ali S Y Analysis of matrrx vesrcles and therr role in the calcrfrcatlon of epiphyseal cartilage Fed. Proc. Am Sot Exp Biol 35 135-142, 1976

J

Sela et al.. Distributton

of matrix

vesicles

in healing

bone

All S Y Mechanism of calcrftcatron. In. Swent,hc Foundations of Offhopaedics and Traumatoology R Owen, J Goodfellow and P Bollough. eds., Helnmann, London, 1980, pp 175-184 Ali S Y Calcrficatron of cartrlage In Cartrlage, Structure, Funct/on and B/ochemWy B K Hall, ed , Vol 1 Acad Press. New York, 1983, pp 343-378 All SY and Evans L The uptake of Ca *+ Calcrum Ions by matrix vesrcles isolated from calctfyrng cartilage Brochem J 134 647.-650, 1973 All S Y and Evans L Mechanrsms of mineral formatron by matrix vesicles In. Ma&/x Vesrcles Proceed/ngs of the 3rd /nfemaf!ona/ Conference of Mafrlx Vewles A Ascenzi, E Bonuccl and B de Bernard, eds Wichtig, Milan, 1981. pp 67-72 Ali S Y., Satdera S.W and Anderson H C lsolatron and characterrzatron of calcifyrng matrix vesrcles from eprphyseal cartilage Proc Naf/ Acad SC/. USA 67 1513-1520. 1970 Air S Y , Wrsby A and Crarg Gray J The sequence of calcium and phos phorous accumulation by matrix vesicles Caiof TISS Res 22 490493. 1976 Amsel S Manratls A, Tavassolr M and Crosby W H The sgnrfrcance of rntramedullary cancellous bone formation rn the reparr of bone marrow tissue Anaf. Rec. 164 101-l 11, 1969 Anderson H.C. Electron mrcroscoprc studres of Induced cartilage develops ment and calclfrcatron J Cell. Biol. 35 81 -101, 1967 Anderson H C Vesicles assocrated with calcrfrcatron rn the matrrx of eprphyseal cartilage J Celi B/o/ 41 59-72, 1969 Anderson H C Matrrx vesrcles of cartrlage and bone In The B/ochem!sfry and Physiology of Bone, Vol IV, H G Bourne, ed Academrc Press, New York, pp 1976, 135-157 Anderson H C Matrrx vesicle calciftcatron Review and update In Bone and Mneral Research, WA Peck, ed Vol 3. Elsevrer Science Publisher, Amsterdam, 1985, pp 1099149 Bab I Gazrt D Massarawa A and Sela J Removal of trbral marrow Induced Increased formatlon of bone and cartrlage in rat mandrbular condyl Calof TISS Inf 37 551-555, 1985 Bab I Schwartz 2, Deutsch D Muhlrad A and Sela J Correlatrve marphometrrc and brochemrcal analysrs of purrfied extracellular matrix vesrcles from rat alveolar bone Calof. TLSS.Int 35 320-326, 1983 Bonuccr. E Fine structure of early cartilage calcifrcatron J Ultrasfr Res 20 33-50. 1967 Boskey, A L Current concepts of physrology and brochemrstry of calclflcation, Cbn Orfhop. Rel Res 157 225-257, 1981 Englfeldt B Hterpe A, Rernholt F P and Wernerson A Srze dlstrrbutron of matrix vesrcles dunng mrnerallzahon of eprphyseal growth cartrlage Proc. Confer of Ceil Mediated CalohcaCon and Matrix Vesrc/es. Elsevier Publrshrng, Amsterdam. 1985 Felrx R and Fleisch H Role of matrrx vesicles in calcrfrcatron fed Proc Am Sot B/o/ 35 1699171, 1976 Frost, H M Hrstomorphometry of cortrcal bone errors due to malorrentatron of sectron plane In Bone H~sfomorphomefry, P J Meunler, ed 2nd lnternatronal Workshop, Lyon, 1976, pp 69974 Hsu H H T and Anderson H C Calctfrcation of Isolated matrix vesrcles from fetal cartriage Proc Nat Acad SC/ U S A 75 3805-3808. 1978 Muhlrad A Bab I and Sela J Dynamic changes In bone cells and extracellular matrrx vescles durrng healing of alveolar bone In rats Metaboi Bone 0s and Rel Res. 2 3477356. 1981 Patt H M and Malloney M A Bone marrow regeneration after local rntury a revrew Exp Haemafoi 3 135- 148. 1975 Rernholt F P.. Hterpe A Jansson K and Englfeldt B Stereologrcal studtes on matrrx vesicle dlstributlon In the epiphyseal growth plate durrng healing of low phosphate, vrtamin D deftcrency rrckets Vrrchows Arm chw (Cell Pafbol) i3 44.257-266, 1983 Wuthrer, R F A revrew of the primary mechanisms Qf endochondral calcrftcation with special emphases on the role of cells, mltochondna and matrix vesrcles Chn Orfhop. Rel Res. 171 219-242. 1982

Recewed. June 16. 1986 Revised November 25. 1986 Accepfed February 18. 1987