Condensation of hypertrophic chondrocytes at the chondro-osseous junction of growth plate cartilage in Yucatan swine Relationship to long bone growth.
Download Condensation of hypertrophic chondrocytes at the chondro-osseous junction of growth pla...
THE AMERICAN JOURNAL OF ANATOMY 186:346-358 (1989)
Condensation of Hypertrophic Chondrocytes at the Chondro-Osseous Junction of Growth Plate Cartilage in Yucatan Swine: Relationship to Long Bone Growth CORNELIA E. FARNUM AND NORMAN J. WILSMAN College of Veterinary Medicine, Cornell University, Zthaca, New York, 14853 (C.E.F.);School of Veterinary Medicine, University of Wisconisn, Madison, Wisconsin, 53706 (N.J.W .)
ABSTRACT Chondrocytes of the cartilaginous growth plate are found in a spatial gradient of cellular differentiation beginning with cellular proliferation and ending with cellular hypertrophy. Although it is recognized that both proliferation and hypertrophy contribute significantly to overall bone growth, mechanisms acting on the chondrocyte to control the timing, the rate, and the extent of hypertrophy are poorly understood. Similarly, mechanisms acting on the terminal chondrocyte to cause its death at the chondro-osseous junction have not been investigated. In this study we examine the condensation of terminal hypertrophic chondrocytes in proximal and distal radial growth plates of Yucatan swine at 4 weeks of age. The animals were raised in a controlled environment where activity and feeding patterns were synchronized to a given time in the light/ dark cycle. We analyzed cellular condensation both as a function of circadian rhythms in a 24-hr time period, and as a function of overall rate of growth. The data suggest that the magnitude of circadian influences on long bone growth is significantly damped at the level of the hypertrophic chondrocyte compared to that seen by previous investigators studying circadian influences on chondrocytic proliferation. Secondly, the condensation of hypertrophic chondrocytes at the chondro-osseousjunction varies inversely with rate of growth in length of the bone. At any time period, a higher percentage of terminal chondrocytes in the condensed form was found in the slower-growing of the two growth plates. We relate these findings to current hypotheses concerning controls of chondrocytic hypertrophy and possible controls over the timing of hypertrophic cell death. INTRODUCTION
Growth in length of long bones of the appendicular skeleton occurs in the cartilaginous growth plates at the ends of the bone. Each chondrocyte contributes to overall long bone growth by passing through a series of developmental stages characterized proximally by cellular proliferation and ending distally with cellular hypertrophy. During these stages, chondrocytes synthesize and secrete components of the extracellular matrix, and an increase in total matrix production par0 1989 ALAN
R. LISS, INC.
allels the increase in cellular volume during hypertrophy (Buckwalter et al., 1986; Hunziker et al., 1987). Thus, the total amount of growth achieved in a particular growth plate is a function of the size of the proliferative cell pool, the incremental increase in matrix production by each individual chondrocyte, and the increase in size of each chondrocyte in the direction of growth during hypertrophy (Hunziker et al., 1987; Kember, 1978; Luder et al., 1988). It is generally agreed that death of hypertrophic chondrocytes a t the chondro-osseous junction precedes invasion by metaphyseal vasculature and the laying down of osteoid and bone on the previously calcified longitudinal septa of the growth plate cartilage (Schenk et al., 1967, 1968; Brighton et al., 1973, 1982; Hanaoka, 1976; Stambaugh and Brighton, 1980; Farnum and Wilsman, 1987; Miki and Yamamuro, 1987; but see also Shimomura et al., 1975; Shimomura and Suzuki, 1984). Recent studies of porcine growth plates growing a t approximately 140 p d d a y have shown that 5-6 chondrocytes turn over each day, or one every 4 hr (Farnum and Wilsman, 1989). Similarly, studies on developing rat growth plates growing at approximately 300 p m per day have shown that at the level of the chondroosseous junction, vascular invasion may eliminate, for each column of cells, one hypertrophic chondrocyte every 3 hr, or 8 cells a day (Hunziker et al., 1987). In order to achieve a steady state of growth, these losses due to hypertrophic cell turnover must be compensated for by proliferation, matrix production, and hypertrophy of the remaining cells. Of these 3 critical processes, only the first 2 have been studied extensively. The contribution of cellular proliferation to long bone growth has been understood primarily through a well-controlled series of studies by Kember (1960, 1978, 1983) using tritiated thymidine. Numerous investigators have focused on factors associated with matrix production by growth-plate chondrocytes and the cellular controls over matrix calcification (for recent reviews see Boskey, 1981; Wuthier, 1982, 1988). However, mechanisms which control the timing and extent of chrondrocytic hypertrophy currently are poorly understood (Buckwalter et al., 1986; Cowell et al., 1987; Hunziker et al., 1987; Shapiro and Boyde, 1987). Similarly, although there is general
Received February 27, 1989. Accepted June 16,1989.
HYPERTROPHIC CELL CONDENS#ATIONIN THE GROWTH PLATE
agreement that hypertrophic cells die prior to metaphyseal vascular penetration, mechanisms which control the timing and the mode of chondrocytic death at the chondro-osseous junction are not understood (Brighton et al., 1973; Shimomura and Suzuki, 1984; Shimomura et al., 1975; Kakuta et al., 1986; Farnum and Wilsman, 1987, 1989). Nevertheless, it can be hypothesized that mechanisms which control hypertrophy as well as mechanisms which control cellular turnover a t the chondro-osseous junction must be synchronized with cellular proliferation proximally in the growth plate in order to achieve the kind of steadystate kinetics of growth which characterizes controlled endochondral ossification. It is our hypothesis that cellular death at the chondro-osseous junction is manifested as condensation of the terminal hypertrophic chondrocyte. It is hypothesized that chondrocytic condensation is synchronized with cellular division in the proliferative cell zone and that condensation at the chondro-osseousjunction correlates with the rate of overall long bone growth. Many features of growth-plate metabolism in rat and rabbit long bones have been shown to follow a circadian rhythm of activity. At particular times during a 24-hr day, a maximal number of chondrocytes in rat growth plates express their proliferative potential and undergo DNA synthesis, and thus overall proliferative activity may vary by as much as 33% a t different times in the light-dark cycle (Simmons et al., 1979; Oudet and Petrovic, 1981; Kember, 1983; Russell et al., 1983). Similarly, studies of rat and rabbit growth plates have shown that patterns of proteoglycan synthesis (Simmons and Nichols, 1966; Simmons, 1968; Simmons et al., 1979), alkaline phosphatase activity (Simmons et al., 1979, 1983; Oudet and Petrovic, 1981), collagen synthesis (Russell et al., 1984b, 1985), and matrix calcification (Oudet and Petrovic, 1981; Simmons et al., 1983; Russell et al., 1984a) follow discrete circadian rhythms. In this study we examine condensation of hypertrophic chondrocytes both as a function of circadian rhythms and as a function of overall rate of long bone growth.
Yucatan sow was placed in a 75-ft2, environmentally controlled room with a temperature of 22°C and relative humidity of 35%. Temperature and humidity were kept constant and monitored continuously. Standard cool white fluorescent lighting a t 100 fc was turned on at 0700 and off at 1900 (LD 12:12).Each room included a 15-ft2 area which accommodated piglets but from which the sow was excluded. Water was available ad libidum through a nipple watering device and by means of a floor pan for the piglets. Sows were fed a commercial lactation ration and piglets were given a supplemental feeding of commercial starter ration at 3 weeks of age. Piglets used for the study were defined as normal and healthy on the basis of history, physical exam, and normal weight-gain pattern during the first 3 weeks of life. Standards of care exceeded those recommended by AALAC. All manipulations of either the sow or of the piglets were done a t 0800. These manipulations included daily feeding and cleaning, and oxytetracycline injections of piglets 3 days and 1day prior to sacrifice. Collections of growth plates were made from 24 piglets, 6 groups of 4 each, according to the following schedule (HALO refers to hours after lights on): a. b. c. d. e. f.
1 hr after lights on . . . . . . . . . . . . . . 01 HALO Mid-light period . . . . . . . . . . . . . . . . 0 6 HALO 1 hr before lights off . . . . . . . . . . . . . 11 HALO 1 hr after lights off . . . . . . . . . . . . . . 13 HALO Mid-dark period . . . . . . . . . . . . . . . . 18 HALO 1hr before lights o n . . . . . . . . . . . . . . 2 3 HALO
Piglets were individually identified at 3 weeks and assigned randomly to one of the 6 time periods, except that an effort was made t o keep equal sex distribution among pigs within one group. As a result, 5 of the groups had 2 males and 2 females; the sixth group had 3 females and 1male.
Labeling for Bone Growth: Tissue Collection Oxytetracycline hydrochloride (OTC) was injected intraperitoneally at a dosage of 1mg/kg at intervals 48 hr apart (72 and 24 hr prior to euthanasia). This dosage MATERIALS AND METHODS of OTC previously has been shown not to interfere with Animals long bone growth (Hansson, 19671, and IP injections Animals used in this study were Yucatan swine of can be given rapidly and painlessly to piglets with syswild hog ancestry. Yucatan swine achieve adult temic absorption as rapid as by intravenous injection. weights (60-70 kg) that are only 25%of that of rapidly At sacrifice, each piglet was removed from the room growing commercial swine. Because of this lower adult and anesthetized within 1 min of removal with pentobody weight, the term “miniature” has been applied to barbital sodium (50 mg/kg, IP). Growth-plate collections were made from both the these swine. However, the pattern of skeletal maturity in Yucatan swine parallels that of other swine of wild proximal and distal radial growth plates by cutting hog ancestry (Reiland, 19781, and their postnatal l-mm-thick slabs which included articular cartilage, growth is neither chondrodystrophic nor chondrodys- epiphyseal bone of the secondary center of ossification, plastic as is true of “miniature” canine breeds. Their the full thickness of the growth plate, and a few millirelatively small adult size compared to adult swine of meters of metaphyseal bone. Further isolation of the commercial breeds represents only the absence of se- growth plate was achieved by trimming through the lective breeding for rapid growth. These swine are a epiphyseal bone and making sequential l-mm-thick good model for studies of normal mammalian postnatal slices across the remaining tissue. This resulted in tisbone growth, as has been documented previously sue blocks measuring 1 x 1 mm by approximately 6 (Farnum et al., 1984). Swine growth plates more mm including epiphyseal bone on one end and about 2 closely resemble human growth plates than do rodent rnm of metaphyseal bone on the other end. This amount growth plates in the cellular numbers in different of metaphyseal bone was necessary to guarantee visualization of the OTC label. As the slabs were collected growth plate zones (Thurston and Kember, 1985). Four to 6 weeks prior to farrowing, each pregnant they were divided randomly among different fixatives.
C.E. FARNUM AND N.J. WILSMAN
the original plane of section. In this study, all terminal Two methods of chemical fixation which recently lacunae at the chondro-osseous junction containing have been shown to preserve hypertrophic chondro- profiles of cells (either hydrated or condensed), as well cytes in morphology similar to that of living hyper- as lacunae which appeared empty, were analyzed. A lacuna was classified as terminal if i t was separated trophic chondrocytes were used in this study. from the metaphyseal endothelial cell or from any blood vascular cell by the distance of the last transFixation with ruthenium hexamine trichloride (RHT) verse septum. This fixation regime has been shown by previous inAt a magnification of x 400, using a lighted pointer vestigators to preserve hypertrophic chondrocytes in a arrow to maintain orientation, the chondro-osseous fully hydrated state by stabilizing macromolecular junction on the metaphyseal side was identified, and interactions between the plasma membrane and peri- all terminal lacunae across the full width of the section cellular proteoglycans (Hunziker et al., 1982, 1983). were classified as containing a hydrated chondrocyte or Primary fixation in 2% glutaraldehyde-2% paraform- a condensed chondrocyte, or as appearing empty. At aldehyde with 0.7% RHT in 0.1M cacodylate buffer, pH least 4 blocks were chosen from each pig for each of the 7.3, for 2 h r was followed by secondary fixation in 1% 2 fixatives, to give data points on at least 100 profiles. osmium-0.7% RHT bufferred in 0.1M cacodylate, pH Since no significant difference was found between fix7.3, for another 2 hr. atives, these data were combined. The final data represent a n analysis of a t least 200 profiles per pig per Fixation in ferrocyanide-reducedosmium growth plate, greater than 9,600 lacunae for the entire For these slabs, primary fixation in 2% paraformal- study. dehyde-2% glutaraldehyde, 0.1M cacodylate buffer, pH 7.3, for 2 h r was followed by secondary fixation in unTetracyclineLabeling buffered 1% osmium tetroxide-l.5% potassium ferroSlabs to be used for measurements of the rate of cyanide for 2 hr. The fixation effect of osmium-ferrogrowth were fixed directly in 70% ethanol for 4 hr, and cyanide is hypothesized to be related to the ability of reduced osmium to stabilize membrane glycoconju- then dehydrated and embedded as described previgates and their relationship with pericellular macro- ously. Six-micron-thick sections were cut, and fluorescent bands were viewed by using a Nikon Optiphot molecules (Farnum and Wilsman, 1983). Following secondary fixation, all blocks were rinsed epifluorescence microscope with a 405-410-nm exciter in buffer, rapidly dehydrated in graded alcohols, and filter, 430-nm dichroic mirror, and 435-nm barrier filcleared in propylene oxide. Infiltration in 1:l propylene ter. Either bands were measured directly on the section oxide and Epon-Araldite for 2 h r was followed by pure using by a filar measuring device, or else the section Epon-Araldite in a dessicator for 3 days. Polymeriza- was photographed and measurements were made. At least 2 blocks were cut per pig, and when filar meation was for 5 days a t 60°C. surements were made, measurements from at least 3 sections of each block were averaged. Because of rapid Analysis of Serial Sections fading of the fluorescent image, most measurements From each of the 24 pigs, a t least 4 blocks were chodone on photographs. Rate of growth in microns sen a t random and twenty 1-km-thick sections were cut were per day for each growth plate for each pig was calcuand placed on slides which had been washed in 1 : l HC1 lated as the distance between the leading edge of the and coated with gelatin-chrome-alum. Staining of the two 48-hr OTC bands, corrected for magnification facsections was with methylene blueAzure I1 and basic tors, and divided by 2. fuchsin at 50°C (Humphrey and Pittman, 1974). The term lacuna has been used in different ways by Reporting of the Data different investigators. For this study we use the operThe percentage of profiles of chondrocytes in the conational definition of Hunziker et al. (1982) and Eggli e t al. (1985) that the lacuna is the space which is occupied densed form was calculated both a s a function of total by a chondrocyte and its pericellular matrix. In this chondrocytic profiles and as a function of total lacunar study, analysis was of the last hypertrophic chondro- profiles. Corrections were made for the average size cyte in each column of cells on the metaphyseal side. difference between condensed and hydrated cells which The interpretation is complex because terminal chon- relates to the probability of finding a profile of the drocytes may occur in two morphologically distinct chondrocyte within the lacuna in which i t resides (see forms. Terminal chondrocytes are found in a fully hy- Results). In order to compare the cellular data between drated form (consistent with the morphology of living the 2 growth plates as a function of rate of growth of hypertrophic chondrocytes), and a hydrated cell en- the individual plate, a n index of cellular death was tirely fills its lacuna (Farnum and Wilsman, 1989). Al- calculated as the percent of all profiles of terminal hyternatively, terminal chondrocytes may appear in a pertrophic chondrocytes which are in a condensed form condensed form (consistent with the morphology of dy- divided by bone growth in microns in a 24-hr period. ing terminal hypertrophic chondrocytes). A condensed Chronograms of the data were plotted and the statisticell withdraws its attachments to the pericellular ma- cal significance between crest and trough (amplitude of trix and therefore the lacuna in which it resides may the rhythm) was determined by analysis of variance. appear to be devoid of cellular content in some sections Tests of significance of difference between proximal (Farnum and Wilsman, 1987). Thus, terminal lacunae and distal radial growth plates (rate of growth, percent which appear empty on one section either may be truly of cells in the condensed form, index of cellular death) empty or may contain a condensed chondrocyte out of was by Student’s t test for paired data. Fixation Protocol
HYPERTROPHIC CELL CONDENSATION IN THE GROWTH PLATE
Fig. 1. Proximal radial growth plate of a 4-week-old Yucatan pig. The characteristic cellular zones are labeled as reserve chondrocytes (R), proliferative chondrocytes (PI, and hypertrophic chondrocytes (HI. Arrowheads show the chondro-osseousjunetion on the metaphyseal side. Terminal chondrocytesat this bonelcartilageinterface were analyzed. EB, Epiphyseal bone of the secondary center of ossification; MB, metaphyseal bone. Fixation with 0.7% ruthenium hexarnine trichloride. x 85.
RESULTS Morphological Features of Terminal Lacunae
The full thickness of the proximal radial growth plate from a 4-week-old pig (Fig. 1) shows the characteristic growth plate zones. In this study, profiles of chondrocytic lacunae at the chondro-osseousjunction of the metaphyseal side were analyzed. For a lacuna to be
considered as terminal, an intact last transverse septum needed to exist between the lacuna and either the metaphyseal endothelial cell or any metaphyseal blood vascular cell. Figure 2a shows 3 terminal hypertrophic chondrocytes; note that for all 3, the last transverse septum is intact. However, the content of the blood vascular space distal to the last transverse septum is variable. Most commonly, terminal lacunae contain profiles of fully hydrated terminal hypertrophic chondrocytes, similar in morphology in both fixatives. Because the terminal hypertrophic chondrocyte in this form makes direct circumferential attachments of the chondrocytic plasma membrane to the pericellular matrix, the profile of the chondrocyte completely fills the lacuna and the lacuna cannot be visualized as a space (Fig. 2b,c). Terminal hypertrophic chondrocytes in the hydrated form are identical in morphology to hypertrophic chondrocytes positioned more proximally in the growth plate (Fig. 2b). Hydrated terminal cells can be identified even when the profile is of a grazing cut through the periphery of a cell (Fig. 2c). Profiles of terminal hypertrophic chondrocytes in the condensed form vary in size and characteristically do not fill the lacuna in which they reside. For these chondrocytes, the plasma membrane has withdrawn from its attachments to the pericellular matrix, maintaining an attachment only to the last transverse septum (Fig. 3a) (Farnum and Wilsman, 1987,1989).Therefore, profiles of these condensed cells vary in morphology depending upon the plane of the section, but in all cases at least part of the terminal lacuna is visualized as a space (Fig. 3b-d). In deciding whether a lacuna was terminal, a judgment was made concerning the probable sequence of vascular penetration. As an example, in Figure 3b, the hydrated terminal hypertrophic chondrocyte and the condensed terminal chondrocyte share the same territorial matrix and the same last transverse septum. Therefore, both of these chondrocytes are considered to be in the direct path of the penetrating vasculature, and thus both chondrocytes are considered to be terminal cells. In Figure 3c, the condensed chondrocyte is considered to be a terminal cell. The hydrated chondrocyte is considered as a penultimate cell that will not be reached until the lacuna in which the condensed cell resides is penetrated by an endothelial cell. Some terminal lacunae appear totally devoid of cellular contents (Figs. 3a,c 4a). Empty spaces either in direct contract with the blood vasculature (Fig. 4b) or connected by a channel through the last transverse septum to the blood vasculature (Fig. 4c,d) were not considered to be terminal lacunae. Furthermore, empty lacunae were easily distinguishable from sections through the very periphery of a hydrated chondrocyte (compare Fig. 2c to Fig. 3a). Lacunae were considered to be terminal only if the last transverse septum was entirely intact (compare Figs. 3a,c and 4a t o 4b-d). Cellular Condensation as a Function of Circadian Rhphms For each of the 24 pigs, 4 to 6 blocks were cut of each
growth plate for each of the 2 fixatives. Profiles of terminal lacunae on section 10 of the series of 20 serial sections were identified as of a condensed chondrocyte (C), a hydrated chondrocyte (HI, or empty (E). Data
C.E. FARNUM AND N.J. WILSMAN
Fig. 2. Terminal chondrocytes with a fully hydrated morphology. Fixation with 0.7% ruthenium hexamine trichloride. x 925. a: Each chondrocyte indicated by a n asterisk is a terminal cell separated from the blood vasculature by a n intact last transverse septum. Notice that a distinct pericellular matrix surrounds each cell (arrowheads). Only the chondrocyte at the right has a capillary endothelial cell completely filling the loop adjacent to the last transverse septum. Distal radial growth plate. b: Terminal chondrocytes in the hydrated form (asterisk) are identical in morphology to hypertrophic chondrocytes positioned more proximally in the column. Distal radial growth plate. c: A grazing cut has been made through the periphery o f three hypertrophic chondrocytes (arrowheads). No nuclear profiles are seen; it is apparent however, that these profiles are through hydrated chondrocytes. Proximal radial growth plate. Note the size difference between hypertrophic chondrocytes in distal radial growth plates (a,b) compared to the proximal radial growth plate (c). Fig. 3. Terminal hypertrophic chondrocytes in the condensed morphology from the proximal radial growth plate. Fixation with 0.7%
initially were calculated for the 2 fixatives separately, and then, because there were no significant fixative differences, the data were pooled. Data are presented
ruthenium hexamine trichloride. x 925. a: This is the characteristic morphology of a terminal hypertrophic chondrocyte in the condensed form. Notice that the cell has withdrawn the plasma membrane from its attachments to the pericellular matrix and remains attached only at the last transverse septum (arrowheads). The lacuna marked with the asterisk appears devoid of cellular content. b: This condensed terminal chondrocyte does not completely fill its lacunar space. When followed in serial sections, the chondrocyte can be shown to make an asymmetrical attachment to the last transverse septum out of this plane of section. Notice that the hydrated terminal chondrocyte shares the same territorial matrix with the condensed cell. c: In this micrograph, the empty lacunar space (asterisk) as well as the condensed chondrocyte are considered to be terminal. The hydrated chondrocyte is considered to be a penultimate cell that will not be reached until the lacuna occupied by the condensed cell has been vacated and penetrated by the vasculature. d: Condensed chondrocytes vary in size and shape. In this profile of a condensed cell, most of the lacuna appears as a n empty space.
as percent of condensed cells as a function of all cellular profiles %C = [C/(C+H)], or percent of condensed cells as a function of all lacunar profiles
HYPERTROPHIC CELL CONDENSATION IN THE GROWTH PLATE
Fig. 4. Terminal lacunae which appear empty. Fixation with 0.7% ruthenium hexamine trichloride. X 925. a: A lacuna which appears empty. The last transverse septum appears calcified, indicating that this section is through the peripheral part of the lacuna. Proximal radial growth plate. b The asterisks indicate 2 empty lacunar spaces. However, note that the last transverse septum has been eroded so that blood vascular elements make direct contact with these spaces. Therefore, these lacunae were not counted in the present study. Distal radial growth plate. c: The lacuna marked by the asterisk appears as an empty space. However, the last transverse septum is not intact (arrowhead). Therefore, this lacuna would not have been defined as ter-
Fig. 5. The pattern of tetracycline labeling in these growth plates. EB, Epiphyseal bone; MB, metaphyseal bone; GP, growth plate. Tetracycline is incorporated at sites of active calcification and can be identified as a brightly fluorescent band. Rate of growth was measured as the distance between the leading edges of the two 48-hr fluorescent bands in the metaphyseal hone [arrowheads). x 75.
%Cl = [C/(C + H +Ell. These data are presented in Table 1for each of the 6 time periods. The combined data represent a n analysis of more than 9,600 lacunae. Both of these calculations underestimated the true number of condensed chondrocytes because of the average size difference between hydrated chondrocytes and condensed chondrocytes. This relates to the probability of finding a profile of a condensed chondrocyte within the lacuna in which it resides when analysis is
made from only one section. Previously it has been shown that, when followed in serial sections, only 5%of terminal lacunae are truly empty (Farnum and Wilsman, 1989). Most lacunae which appear empty on one section contain a profile of a condensed chondrocyte out of the original plane of section. As a n example, the empty lacunae i n Figures 3a and 4a, if followed in serial sections, can be shown to contain a terminal hypertrophic chondrocyte in the condensed form. Al-
minal for this study. Distal radial grciwth plate. d In this micrograph there is a large direct connection between the blood vasculature and the empty lacunar apace through an eroded last transverse septum. Proximal radial growth plate.
C.E. FARNUM AND N.J. WILSMAN
TABLE 1. Analysis of terminal hypertrophic chondrocytes in the condensed form from proximal and distal radial plrowth plates of twenty-four 4-week-old pigs, collected at 6 time periods1 Proximal radial growth plate 2
Time period 01 HALO 06 HALO 11 HALO 13 HALO 18 HALO 23 HALO Average
136 2 11.2 136 f 14.5 133 f 20.1 130 ? 23.0 153 24.6 104 ? 8.29 132 f 7.2
24.2 f 4.1 22.5 -t 1.7 20.5 .9 20.6 k 4.6 21.6 2 1.9 18.6 -+ 2.2
16.0 f 2.6 22.0 4.5 13.2 f .9 13.9 k 2.9 14.2 f 1.2 12.6 2 1.7
Distal radial growth plate
c,5 34.3 f 5.3 31.0 2 3.4 27.6 f 1.4 29.8 k 5.6 31.1 2.3 27.4 2 2.8 30.2 -t 1.5
Growth (pm) 250 f 20.9 242 -+ 35.4 217 f 34.5 226 k 39.3 226 * 30.5 241 k 18.6 234 2 11.4
C 17.9 -+ 3.4 15.4 k 1.2 15.0 f 1.4 18.4 k .9 15.5 f 2.0 16.5 f 1.5
CC 26.3 2 4.4 22.9 f 1.5 21.4 f 1.4 27.3 2 1.4 25.4 -C 2.3 25.4 2.1 24.6 f 1.0
10.5 f 1.9 9.7 2 .7 9.0 1.1 10.7 k 1.3 10.2 f 1.0 10.4 2 0.6
'Four animals were collected at each time period. At least 100 profiles of terminal lacunae were analyzed for each of 2 fixatives, and the data were pooled. All values are expressed as the mean f S.E. "Growth in microns was measured as the distance between the leading edges of two 48-hr oxytetracycline bands, corrected for magnification factors, and divided by 2. 3%C = [C/(C + H)] where C = number of profiles of terminal chondrocytes in the condensed form and H = the number in the hydrated form. 4%C1 = [C/(C + H + El], where H = number of profiles of terminal chondrocytes in the hydrated form; E = number of profiles of terminal lacunae which do not contain the profile of a terminal chondrocyte. "C, = [C,(C, + Hc)], where C , and H,are the number of chondrocytes in each form corrected for the size difference of terminal chondrocytes in the condensed form versus terminal chondrocytes in the hydrated form.
TABLE 2. Index of cellular condensation as a function of bone growth in 4-week-old pigs: for both growth plates, there was no significant difference for the means of the 6 time periods examined' Hours after N Mean S.E. Individual 95 PCT CI's for mean lights on Analysis of variance: test of difference in means, proximal radial growth plate 01 4 25.1 3.6 t 06 4 t 23.5 2.9 11 4 23.3 5.7 1 13 4 25.8 6.6 t m 18 4 21.8 3.4 I I 23 4 26.7 3.0
F = 0.17 14.0 Analysis of variance: test of difference in means, distal radial growth plate 4 01 10.5 1.5 t 06 4 10.2 1.8 I =
11 13 18 23
11.6 13.4 11.8 10.5
4 4 4 4
3.5 2.6 1.7 1.6
I I I
'The index of cellular condensation (Icois calculated as the corrected percentage o f chondrocytes in the condensed form (%CJ divided by growth (G) in microns in 24 hr: I,, = (%C,/G) x 100.
though the size of chondrocytes in the 2 forms differs, the size of the lacuna in which they reside is the same, whether the lacuna contains a hydrated chondrocyte, a condensed chondrocyte, or is empty (Farnum and Wilsman, 1989). Therefore, counts of lacunar profiles are unbiased. Hydrated cells and condensed cells differ significantly in the extent to which they fill the lacuna. Specifically,profiles of the terminal hypertrophic chondrocyte in the condensed form are visible in only 55% of the profiles through the lacuna in which the chondrocyte resides. By contrast, profiles of terminal hypertrophic chondrocytes in the hydrated form are visible in 92% of the profiles through the lacuna in which they reside (Farnum and Wilsman, 1989).This relationship holds for growth plates growing a t different rates.
Making corrections for these differences and presenting the final corrected data as [C,/(C, + H,)] yields the data in Table 1 where C, is the number of profiles of condensed chondrocytes corrected for the probability of finding a profile of the chondrocyte in a section through the lacuna, and H, is the analogous corrected value for the number of hydrated chondrocytes. This ratio gives %C, or the percent of total cellular profiles which appear in the condensed form, corrected for size differences between condensed and hydrated chondrocytes. %C, was calculated independently for the proximal and the distal radial growth plates from each pig and for each of the 6 collection periods. Thus, the data yielded 3 estimators (%C, %C1, and %C,) of percentage of terminal cells in the condensed form. Of the 3 estimators,
HYPERTROPHIC CELL CONDENSATION IN THE GROWTH PLATE
%C, most accurately reflects the true cell numbers in the condensed form. Only this estimator is corrected for the difference in probability of finding, in one section, a profile of a condensed chondrocyte vs. the profile of a hydrated chondrocyte within the lacuna in which it resides. Additionally, growth (G) in microns per 24 hr was calculated for each time period for each growth plate (Table 1). Figure 5 represents the kind of section from which growth calculations were made, by measuring the distance between the leading edges of the fluorescent OTC bands. The significant comparison between the proximal and distal growth plates is to analyze the percent of chondrocytes in the condensed form as a function of rate of growth. Therefore, cellular and growth data were combined to create a n index of cellular condensation (Ic,) calculated as the percentage of cells in the condensed form divided by growth in microns in 24 hr, Ic0 = (%C,/G) X 100. This index was calculated for each growth plate for each pig for each of the 6 time periods. A test of difference between means of the peak and the trough of the data showed no significant difference, P < .001 (Table 2). The index a t any one time period varied only * 10%from the mean value of the 6 time periods. These data were consistent for both growth plates. The conclusion is that no significant circadian rhythm in the condensation of terminal chondrocytes at the chondro-osseous junction could be demonstrated in these growth plates. Chondrocytic Condensation as a Function of Rate of Growth
For these calculations, comparisons were made between the proximal and distal growth plates by combining data for the 24 pigs. The percentage of condensed terminal chondrocytes as a function of all terminal cells (%Cc) was 30.2% 2 1.5 (S.E.) for the proximal radial growth plate, and 24.6% * 1.0 for the distal radial growth plate (Table 1).This is a significant difference (P < .001). Growth in microns per 24 h r for the proximal radial growth plate (G ) was 132 * 7.2 pm, and for the distal radial growth pfate (Gd), it was 234 ? 11.4 pm. The difference was significant (P < ,001, Table 1). Thus, the distal radial growth plate grows at almost twice the rate of the proximal radial growth plate for swine of this age with a correlation coefficient of + .74when one uses data from all pigs. Combining the growth data and the cellular data into a n index of cellular condensation yielded similar results. The index was 24.3 1.6 for the proximal growth plate and 11.3 4 0.8 for the distal growth plate (P < .0001). These results are shown in Figure 6 and indicate that for the more slowly growing proximal radial growth plate, a higher percentage of terminal cells were in the condensed form compared to the fastergrowing distal growth plate. The correlation coefficient for these data is -.70,indicating the inverse relationship. This contrast between proximal and distal growth plates was consistent for collections at each time period. DISCUSSION
This paper presents 2 new findings which are significant in understanding the role of cellular hypertrophy
-_ . .=$,:.
20 0 I 90
150 1 Growth vrnt24 hours
Flg. 6. The condensation of hypertrophic chondrocytes as a functlon of rate of growth Each value represents the mean of 4 pigs collected at the same time period. Values less than 180 p d 2 4 hr are for the proximal radial growth plate; values greater than 180 p d 2 4 hr are for the distal radial growth plate SC, is the percent of terminal chondrocytes in the condensed form corrected for size differences between condensed and hydrated cells. A=mean of the data for each growth plate, *S.E.
in the mechanism of long bone growth. First, these data suggest th a t chondrocytic condensation at the chondro-osseousjunction varies inversely with the rate of growth in length of the bone. At any given time period, a higher percentage of terminal chondrocytes in the condensed form was found in the slower-growing of the two growth plates. Secondly, the magnitude of circadian influences on long bone growth is significantly damped at the level of the hypertrophic chondrocyte compared to that seen by previous investigators studying circadian influences on chondrocytic proliferation. In this study, no significant circadian rhythm of cellular turnover at the chondro-osseous junction was shown. Mechanims acting on the terminal hypertrophic chondrocyte to cause its condensation at the chondroosseous junction have not been investigated. Similarly, mechanisms which control the onset, rate, and extent of cellular hypertrophy in growth plates growing at different rates are poorly understood (Buckwalter et al., 1986; Cowell et al., 1987; Hunziker et al., 1987; Wroblewski et al., 1987). There is good evidence, however, that the magnitude of hypertrophy is proportional to the overall increase in length contributed by a given growth plate (Buckwalter and Mower, 1987; Hunziker et al., 1987; Wilsman and Farnum, 1989). Therefore, if the rate of chondrocytic condensation does not correlate with the rate of growth in length, it would suggest that the critical transition point which controls hypertrophy is at a level proximal to the terminal chondrocyte. Similarly, controls on the timing of onset of condensation of the terminal chondrocyte may act independently on controls governing the extent of hypertrophy. Although cellular proliferation and cellular hypertrophy both contribute to overall bone growth, controls which act on proliferation are better understood than controls which act on hypertrophy (Kember, 1978, 1983; Seinsheimer and Sledge, 1981; Thorngren and Hansson, 1981). It is known that major controls on chondrocytic proliferation act through circulating
C.E. FARNUM AND N.J. WILSMAN
growth factors, especially the somatomedins (for review see Trippel, 1988), and receptors for somatomedins have been localized on growth-plate chondrocytes (Trippel et al., 1988). Kember (1973, 1978, 1983) has shown, through a very thorough series of wellcontrolled experiments, that a major factor that determines the amount of growth achieved by a particular growth plate is the size of the proliferative cell pool responding to circulating growth factors. There is new evidence that growth hormone may act directly on some chondrocytes of the growth plate (Isaksson et al., 1982;Madsen et al., 1983; Nilsson et al., 1986) and that proliferative cells may be responding to locally produced somatomedins or other local growth factors by an autocrine mechanism (Burch et al., 1986; Nilsson et al., 1986; Schlechter et al., 1986; Makower et al., 1988). Nevertheless, it still is not understood how growth plates of one bone, such as the proximal and distal radial growth plates, grow simultaneously at widely different rates. Mechanisms which control the onset, rate, and final extent of cellular hypertrophy in growth plates growing at different rates currently have not been thoroughly investigated (Buckwalter et al., 1986; Eavey et al., 1988; Giaretti et al., 1988; Luder et al., 1988). In the mandibular condyle, hypertrophy accounts for 4.4 x the total growth compared to a 1.0 x contribution by cellular proliferation (Luder et al., 1988). In the mouse tibia and the rat tibia, there is a 3-fold increase in matrix production by hypertrophic cells compared to proliferative cells, in contrast to a 4- to 10-foldincrease in cellular height during chondrocytic hypertrophy (Buckwalter et al., 1986; Hunziker et al., 1987). Investigators are beginning to examine the role of several newly discovered growth factors such as transforming growth factor p in the control of hypertrophy (Carrington et al., 1988; Kato et al., 1988; O’Keefe et al., 1988;Rosier et al., 1989), as well as membrane changes involving the Na+/K+ pump (Wroblewski and Makower, 1988). However, control of hypertrophy in any tissue, and especially in the growth plate, is poorly understood (Fine et al., 1985; Buckwalter et al., 1986; Campo, 1988; Eavey et al., 1988; Owens et al., 1988; Rizzino, 1988). This is an important area for future investigations since several diseases of postnatal bone development, such as the osteochondroses, are characterized by failure of chondrocytichypertrophy (Farnum et al., 1984). Circadian Influences on Long Bone Growth
Previous investigators have shown that in growth plate cartilage, the rate of proliferative cell division, proteoglycan biosynthesis, and collagen synthesis follow a circadian rhythm of activity (Petrovic et al., 1984; Simmons and Nichols, 1966; Simmons, 1968). Twofold variations in the magnitude of these metabolic activities may occur over a 24-hr period (Simmons et al., 1979; Oudet and Petrovic, 1981). Specific metabolic functions associated with cartilage matrix mineralization, such as alkaline phosphatase activity (Simmons et al., 1979,1983) and deposition of calcium, also can be shown to have daily rhythms (Simmons et al., 1983; Russell et al., 1984a). Interestingly, these different rhythms are not in phase and, as an example, collagen synthetic activity is maximal at mid-light period, while
the laying down of inorganic phosphorus and calcium are maximal during the dark period (Russell et al., 1983, 1984a,b, 1985). This is the same period as peak DNA synthesis (Russell et al., 1984). This kind of pattern is characteristic of circadian rhythms in other body tissues where rhythms for different metabolic functions will be acting simultaneously, but with peaks and troughs occurring a t different times (Pauly, 1983). Presumably the nature of the rhythm in any given species might be dependent upon whether it is a day-active or night-active animal. Additional seasonal influences on the extent of long bone growth also have been documented in lab animals raised totally indoors (Oudet and Petrovic, 1981). This is the first study to examine circadian influences on long bone growth in an experimental animal as large as a pig. In any study involving circadian rhythms, it is critical that environmental conditions be controlled so that the lightldark cycle as well as all activity cycles are maintained constant throughout the period of study (for review see Pauly, 1983). Although development of overt circadian rhythms is thought to be primarily a postnatal event, there is some evidence that prenatal cyclic inputs also may be significant (Hiroshige, 1986). In rodents there is evidence that maternal influences may entrain the initial rhythm of the young, and that the act of birth itself may be timed by circadian influences, as well as serve as a major influence in setting the circadian clock of the young (Takahashi et al., 1984; Davis and Gorski, 1985; Reppert, 1985; Viswanathan and Chandrashekaran, 1985). Although studies of this kind have not been done in swine, it is clear that in many mammals the role of the mother is critical in coordinating the biological clock of the offspring during the time that neuronal pathways of the young are maturing (Kolata, 1985). In this study we were particularly concerned that the farrowing swine be kept in a controlled environment suitable for a circadian rhythm study of this kind. Sows were kept in the controlled environment for a minimum of the last one-third of pregnancy. Different species differ in which environmental clue acts as the dominant synchronizer for development of circadian rhythms. In rodents this is the light/dark (LD) cycle, while in humans the routine of daily activity has a large influence (Pauly, 1983). Similarly, it has been shown that feeding schedules can synchronize behavioral rhythms in many mammals, including humans (Griffiths, 1986; Minors and Waterhouse, 1986; Stephan, 1986). In swine it has been shown that feeding rhythms are a significant component of daily activity rhythms, and thus the timing of feeding (always at 01 HALO) was controlled in this study. There is evidence that swine develop circadian rhythms in response to their particular husbandry conditions, and so swine seem a good choice for these initial studies on the circadian influences on bone growth in a large mammal (Ingram and Legge, 1970; Ingram and Dauncey, 1985). For this study we chose to collect tissue over 6 time intervals, with some clustering of data points around the time of lights on and lights off transitions. There is evidence that for some metabolic variables, the largest circadian differences are found around the transition points between lightldark and dark/light. Specifically,
HYPERTROPHIC CELL CONDENSATION IN THE GROWTH PLATE
(1983) showed that hypertrophic chondrocytes can be preserved with their plasma membranes intact by using ferrocyanide-reduced osmium as the secondary fixative. Additionally, they were able to show by direct observation of living growth-plate explants and timelapse cinematography that all hypertrophic chondrocytes, including the terminal cell, are fully rounded cells which entirely fill their lacunar space (Farnum and Wilsman, 1987; Wilsman and Farnum, 1988). Because the morphology of living terminal hypertrophic chondrocytes has been established, it is now possible to again investigate the mechanism of death of the terminal hypertrophic chondrocyte at the chondroosseous junction (Wilsman and Farnum, 1989). It has been shown that, in optimally preserved growth plates, some terminal hypertrophic chondrocytes can be found in a condensed form, having withdrawn their circumferential attachments to the pericellular matrix, while maintaining their plasma-membrane attachments to the last transverse septum. Since condensed chondrocytes do not entirely fill the lacuna in which they reside, profiles of these chondrocytes differ in size and shape in different sections taken through one lacuna. The condensed morphology is characteristic only of terminal hypertrophic chondrocytes, and not of chondrocytes positioned more proximally in the growth plate. Because of this unique positional relationship, ie., condensed cells appearing only as terminal cells, it has been hypothesized that terminal chondrocytes in the condensed form represent dying cells (Farnum and Wilsman, 1987). The current study does not supply direct evidence that the proportion of condensed terminal chondrocytes is a measure of the rate of chondrocytic death. Before it would be possible to unequivocally equate cellular condensation with cellular death, it would be necessary to verify, by an independent method, that in living growth plates terminal chondrocytes die by cellular condensation. Living terminal chondrocytes in situ have been visualized in short-term CeNular Death at the Chondro-OsseousJunction as a cultures of growth-plate explants, and in this system, Function of Overall Rate of Long Bone Growth although cellular condensation has been seen, the tranWhen growth plates are preserved by using routine sition of hydrated chondrocytes to condensed terminal glutaraldehyde-osmium fixation, the plasma mem- chondrocytes has not been demonstrated (Wilsman and branes of hypertrophic chondrocytes become attenu- Farnum, 1988). Time estimates have been made for the length of ated and lose their attachments to the pericellular matrix. Because with routine fixation hypertrophic time a terminal hypertrophic chondrocyte remains in chondrocytes appear highly vacuolated with discontin- the condensed form before disappearing entirely and uous plasma membranes, it had been hypothesized by leaving a vacated lacuna (Farnum and Wilsman, several investigators that cellular death in the distal 1989). The condensed morphology and the apparently hypertrophic cell zone was by gradual disintegration rapid time in which chondrocytes progress from a hydue to tissue hypoxia (Brighton et al., 1969, 1973, drated cell to a condensed cell to disappearance (40 1982). Hunziker et al. (1982, 1983) were able to show min) suggest that terminal hypertrophic chondrocytes that by incorporating ruthenium hexamine trichloride in this condensed form may be undergoing pro(RHT) into both the primary and the secondary fixa- grammed cellular death by apoptosis (Farnum and tives, chondrocytes at all levels of the growth plate, Wilsman, 1989; Wyllie, 1987). Alternatively, it is posincluding the distal hypertrophic chondrocyte, were sible that the proportion of condensed terminal chonpreserved as fully hydrated cells with an intact plasma drocytes is not related to the rate of cellular death but membrane directly adherent to the surrounding peri- rather to the rate a t which condensed cells are removed cellular matrix. This morphology is identical to that of by macrophages, capillary sprouts, or some other mechhypertrophic chondrocytes preserved by rapid freezing, anism during metaphyseal vascular penetration. Alfreeze substitution, and low-termperature embedment, though it has been demonstrated that the terminal a method which eliminates the artifacts of routine chondrocyte disappears, leaving an empty lacuna prior chemical fixation and high-temperature polymeriza- to invasion of its lacuna by an endothelial cell of any tion (Hunziker et al., 1984; Akisaka et al., 1987; Ar- other blood vascular element (Farnum and Wilsman, senault et al., 1987). Similarly, Farnum and Wilsman 1988), the mechanism of removal of cellular remnants
for variables associated with growth-plate metabolism, alkaline phosphatase activity has been shown to be highest at the beginning of the dark period (Russell et al., 1984a,b; Simmons et al., 1983). Only one previous study has examined circadian rhythms in relation to the overall rate of long bone growth. Hansson and co-workers (1974) gave OTC labels at 6-hr intervals to 5-week-old rabbits and were able to show variations of only f5% in the incremental growth of the rabbit tibia a t 4 different time points in the 24-hr cycle. Because of very low variability in their data, they were able to conclude that there was a subtle circadian rhythm to the rate of long bone growth, but that the magnitude was small. Growth was lowest (95.7%of the mean growth) in the period representing the first 6 hr of darkness. Previous investigators had reported a 33% difference in rate of tritiated thymidine incorporation at different time points, which was significantly greater that the 5%reported by Hansson for overall growth. The magnitude of circadian changes shown by our results is consistent with Hansson’s results. However, in our study there was no statistically significant difference in the mean percentage of condensed chondrocytes at different times in a 24-hr period. The results of both studies suggest that circadian influences on overall bone growth are damped at the level of the hypertrophic chondrocyte compared to that seen in the proliferative cell zone. This suggests that controls over chondrocytic proliferation proximally do not continue to synchronize later steps in chondrocytic differentiation such as chondrocytic hypertrophy and chondrocytic condensation at the chondro-osseousjunction. To test this hypothesis further it will be necessary to measure all differences-chondrocytic proliferation, chondrocytic condensation, the extent of chondrocytic hypertrophy, and overall rate of growth-all in the same growth plate.
C.E. FARNUM AND N.J. WISSMAN
remains unknown. Similarly, mechanisms that control the rate of capillary penetration are poorly understood. Therefore, these alternative hypotheses related to the biological significance of cellular condensation remain untested. In the current study, we have analyzed all terminal chondrocytes in the condensed form. Our study shows that the percent of chondrocytes in the condensed form is significantly greater in the slower-growing of the two growth plates, even after correction for the difference in size of condensed vs. hydrated terminal chondrocytes. Our initial hypothesis was that cellular condensation at the chondro-osseous junction was synchronized with cellular proliferation more proximally, and so this result was unexpected. Such a result could be obtained if the actual time it takes for a chondrocyte to go through all stages of the condensed form were greater in the slower-growing plate Le,, if chondrocytes in slower-growing plates took longer to condense). However, if condensation represents the morphological equivalent of death by apoptosis, it is unlikely that the duration of death would be a variable. Death by apoptosis characteristically is rapid (30-40 min), with little variation in a wide variety of cellular types (Kerr and Searle, 1980; Wyllie et al., 1980; Wyllie, 1987). Alternatively, one could hypothesize that independent mechanisms exist for controlling the timing of cellular condensation at the chondro-osseous junction and the timing of cellular hypertrophy at a point positioned more proximally in the growth plate. The magnitude of the increase in cellular volume varies by a factor of more than x 10 in growth plates growing at different rates (Buckwalter and Mower, 1987; Hunziker et al., 1987; Wilsman and Farnum, 1989). Since the extent of hypertrophy varies directly with rate of growth over a wide range of growth plates, one could hypothesize that the rate at which hypertrophy is achieved also would correlate directly with rate of growth of the bone. The critical variable then becomes the control of the rate a t which chondrocytes hypertrophy and the final size these cells achieve. Although the increase in cell volume which accompanies hypertrophy is continuous, the increase occurs at a variable rate in the column (Buckwalter et al., 1986; Hunziker et al., 1987; Luder et al., 1988). Additionally, maximal hypertrophy is achieved a t some distance proximai to the terminal hypertrophic cell, before the matrix of the longitudinal septa becomes highly calcified (Wilsman and Farnum, 1989). These observations are in agreement with a hypothesis by Kember (1978) that a key control mechanism for overall long bone growth might be at a point that regulates the rate of cell expansion in the hypertrophic cell zone. At a critical control point hypertrophy may be achieved more rapidly in faster-growing growth plates. Thus, the component of overall bone growth which would be contributed by hypertrophic cell size may be proportional not to the rate of turnover of terminal cells but to the rate of growth of cells as they hypertrophy, and maximal hypertrophy is achieved at a point several chondrocytes proximal to the terminal cell. Such a hypothesis would be consistent with our data, but new experiments are required to test this hypothesis directly.
The authors thank Dr. R. Philips and L. Panepinto of Colorado State University for their assistance in acquiring Yucatan swine and in maintaining a breeding colony. We also thank Pam Hutchinson, Gayanna Gilbraith, Heidi Wells, and Jim Dooley for their expert technical help. This research was supported in part by National Institutes of Health Grant #AR-35155. LITERATURE CITED Akisaka, T., G.P. Subita, H. Kawaguchi, and Y. Shigenaga 1987 Improved ultrastructural preservation of epiphyseal chondrocytes by the freeze-substitution method. Anat. Rec., 219:347-355. Arsenault, A.L., F.P. Ottensmeyer, and I.B. Heath 1987 An electron microscopic and spectroscopic study of murine epiphyseal cartilage: Analysis of fine structure and matrix vesicles preserved by slam freezing and freeze substitution. J. Ultrastruct. Res., 98: 1-16. Boskey, A.L. 1981 Current concepts in the physiology and biochemistry of calcification. Clin. Orthop., 157:225-257. Brighton, C.T., R.D. Ray, L.W. Soble, and K.E. Kuettner 1969 In uitro epiphyseal plate growth in various oxygen tensions. J. Bone Joint Surg., 51 -A:1383-1396. Brighton, C.T., Y. Sugioka, and R.M. Hunt 1973 Cytoplasmic structures of epiphyseal plate chondrocytes. Quantitative evaluation using electron micrographs of rat costochondral junctions with special reference to the fate of hypertrophic cells. J. Bone Joint Surg., 55-A:771-784. Brighton, C.T., Y. Sugioka, and R.M. Hunt 1982 Quantitative zonal analysis of cytoplasmic structures of growth-plate chondrocytes in uiuo and in uitro. J. Bone Joint Surg., 64-A:1336-1349. Buckwalter, J.A., D. Mower, R. Ungar, J . Schaeffer, and B. Ginsberg 1986 Morphometric analysis of chondrocytic hypertrophy. J. Bone Joint Surg., 68-A:243-255. Buckwalter, J.A., and D. Mower 1987 Alterations of growth plate associated with slowing of bone growth. Proc. 33rd Orthop. Res. SOC.,San Francisco, CA, p. 490. Burch, W.M., S. Weir, and J.J. VanWyk 1986 Embryonic chick cartilage produces its own somatomedin-like peptide to stimulate cartilage growth in uitro. Endocrinology, 11911370-1376. Campo. R.D. 1988 Effects of cations on cartilage structure: Swelling of growth plate and degradation of proteoglycans induced by chelators of divalent cations. Calcif. Tissue Int., 43:108-121. Carrington, J.L., A.B. Roberts, K.C. Flanders, N.S. Roche, and A.H. Reddi 1988 Accumulation. localization. and comnartmentation of transforming growth factor p during endochondial bone development. J. Cell Biol., 107:1969-1975. Cowell, H.R., E.B. Hunziker, and L. Rosenberg 1987 The role of hypertrophic chondrocytes in endochondral ossification and in the development of secondary centers of ossification. J. Bone Joint Surg., 69-At159-161. Davis, F.C., and R.A. Gorski 1985 Development of hampster circadian rhythms: Prenatal entrainment of the pacemaker. J. Biol. Rhythms, 1:77-89. Eavey, R.D., T.M. Schmid, and T.F. Linsenmayer 1988 Intrinsic and extrinsic controls of the hypertrophic program of chondrocytes in the avian columella. Dev. Biol., 126;57-62. Eggli, P.S., W. Herrmann, E.B. Hunziker, and R.K. Schenk 1985 Matrix compartments in the growth plate of the proximal tibia of rats. Anat. Rec., 211:246-257. Farnum, C.E., and N.J. Wilsman 1983 Pericellular matrix of growth plate chondrocytes: A study using post-fixation with osmium-ferrocyanide. J. Histochem. Cytochem., 31 :765-775. Farnum, C.E., and N.J. Wilsman 1987 Morphologic stages of the terminal hypertrophic chondrocyte of growth plate cartilage. Anat. Rec., 219:221-232. Farnum, C.E., and N.J. Wilsman 1989 Cellular turnover at the chondro-osseous junction of growth plate cartilage: Analysis by serial sections at the light microscopical level. J. Orthop. Res., in press. Farnum, C.E., N.J. Wilsman, and H.D. Hilley 1984 An ultrastructural analysis of osteochondritic growth plate cartilage in growing swine. Vet. Pathol., 21:141-151. Fine, L.G., R.W. Holley, H. Nasri, and B. Badie-Dezfooly 1985 BSC-1 growth inhibitor transforms a mitogenic stimulus into a hypertrophic stimulus for renal proximal tubular cells: relationship to Na+/H+ antiport activity. Proc. Natl. Acad. Sci. USA, 82;61636166. Giaretti, W., G. Moro, R. Quarto, S. Bruno, A. DiVinci, E. Geido, and
HYPERTROPHIC CELL CONDENSATION IN THE GROWTH PLATE R. Cancedda 1988 Flow cytometric evaluation of cell cycle characteristics during in uitro differentiation of chick embryo chondrocytes. Cytometry, 9:281-290. Griffths, R.A. 1986 Natural environmental cues and circadian rhythms of behaviour-a perspective. Chronobiol. Int., 3; 247-253. Hanaoka, H. 1976 The fate of hypertrophic chondrocytes of the epiphyseal plate. J . Bone Joint Surg., 58-At226-229. Hansson, L.I. 1967 Daily growth in length of diaphyses measured by oxytetracycline in rabbits normally and after medullary plugging. Acta Orthop. Scand. LSuppl.1, 10111-178. Hansson, L.I., A. Stenstrom, and K.G. Thorngren 1974 Diurnal variation of longitudinal bone growth in the rabbit. Acta Orthop. Scand., 45499-507. Hiroshige, T. 1986 Maternal phase control of fetal circadian oscillation during pregnancy. Jpn. J . Physiol., 361237-251. Humphrey, C.D., and F.E. Pittman 1974 A simple methylene blueAzure 11-basic fuchsin stain for epoxy tissue sections. Stain Technol., 419:9-10. Hunziker, E.B., W. Herrmann, and R.K. Schenk 1982 Improved cartilage fixation by ruthenium hexamine trichloride (RHTI: A prerequisite for morphometry in growth cartilage. J. Ultrastruct. Res., 81tl-12. Hunziker, E.B., W. Herrmann, and R.K. Schenk 1983 Ruthenium hexamine trichloride (RBT)-mediated interaction between plasmalemmal components and pericellular matrix proteoglycans is responsible for the preservation of chondrocytic plasma membranes in situ during cartilage fixation. J . Histochem. Cytochem., 31:717-727. Hunziker, E.B., W. Herrmann, R.K. Schenk, M. Mueller, and H.Moor 1984 Cartilage ultrastructure after high pressure freezing, freeze substitution, and low temperature embedding. I. Chondrocytic ultrastructure-Implications for theories of mineralization and vascular invasion. J. Cell Biol., 981267-276. Hunziker, E.B., R.K. Schenk, and L.M. Cruz-Orive 1987 Quantitation of chondrocyte performance in growth-plate cartilage during longitudinal bone growth. J. Bone Joint Surg., 69-A:162-173. Ingram, D.L., and M.J. Dauncey 1985 Circadian rhythms in the pig. Comp. Biochem. Physiol., 82A:1-5. Ingram, D.L., and K.F. Legge 1970 Variations in deep body temperature in the young unrestrained pig over the 24 hour period. J . Physiol. (Lond.), 21Ot989-998. Isaksson, O.G.P., J. Jansson, and I.A.M. Gause 1982 Growth hormone stimulates longitudinal bone growth directly. Science, 21 6;12371239. Kakuta, S.,E.E. Golub, J.C. Haselgrove, €3. Chance, P. Frasca, and 1.M. Shapiro 1986 Redox studies of the epiphyseal growth cartilage: Pyridine nucleotide metabolism and the development of mineralization. J. Bone Miner. Res., 1;433-440. Kato, Y., M. Iwamoto, T. Koike, F. Suzuki, and Y. Takano 1988 Terminal differentiation and calcification in rabbit chondrocyte cultures mown in centrifuge tubes: Reeulation bv transforminn growth factor-p and ser;m factors. Pr"c. Natl. Acad. Sci. US& 8.59552-9556. _ _ _ . ...~ .. Kernber, N.F. 1960 Cell division in endochondral ossification. A study of cell proliferation in rat bones by the method of tritiated thymidine autoradiography. J . Bone Joint Surg., 42-B:824-839. Kember, N.F. 1973 Patterns of cell division in the growth plates of the rat pelvis. J . Anat., 116t445-455. Kember, N.F. 1978 Cell kinetics and the control of growth in long bones. Cell Tissue Kinet., 1 Z;477-485. Kember, N.F. 1983 Cell kinetics of cartilage. In: Cartilage, Vol. 1. Structure, Function and Biochemistry. B.K. Hall, gd. Academic Press, New York, pp. 149-180. Kerr, J.F.R., a n d J . Searle 1980 Apoptosis: Its nature and kinetic role. In: Radiation Biology in Cancer Research. R.E. Meyn and H.R. Withers, eds. Raven Press, New York, pp. 367-384. Kolata, G. 1985 Finding biological clocks in fetuses. Science, 2.70; 929-930. Luder, H.U., C.P. Leblond, and K. VonderMark 1988 Cellular stages in cartilage formation as revealed by rnorphometry, radioautography, and type I1 collagen immunostaining of the mandibular condyle from weanling rats. Am. J. Anat., f82r197-214. Madsen, K., U. Friberg, P. Roos, S. Eden, and 0. Isaksson 1983 Growth hormone stimulates the proliferation of cultured chondrocytes from rabbit ear and rat rib growth cartilage. Nature, 304t545-547. Makower, A., J. Wroblewski, and A. Pawlowski 1988 Effects of IGF-I, EGF, and FGF on proteoglycans synthesized by fractionated chondrocytes ofrat rib growth plate. Exp. Cell Res., 179t498-506. Miki, T., and T. Yamamuro 1987 The fate of hypertrophic chondro-
cytes in growth plates transplanted intramuscularly in the rabbit. Clin. Orthop., 218276-282. Minors, D.S., and J.M. Waterhouse 1986 Circadian rhythms and their mechanisms. Experientia, 42;l-13. Nilsson, A., J . Isgaard, A. Lindahl, A. Dahlstrom, A. Skottner, and O.G.P. Isaksson 1986 Regulation by growth hormone of number of chondrocytes containing IGF-I in rat growth plate. Science, 233:571-574. OReefe, R., J.E. Puzas, J.S. Brand, and R.N. Rosier 1988 Effect of transforming growth factor-p on DNA synthesis by growth plate chondrocytes: Modulation by factors present in serum. Calcif. Tissue Int., 43t352-358. Oudet, C., and A. Petrovic 1981 Nyctohemeral and seasonal variations in the number of tritiated labelled cells in the epiphyseal cartilage of the tibia in the growing rat. Effect of lighting duration and temperature. In: Eleventh International Congress of Anatomy, Part C: Biological Rhythms in Structure and Function. H. vonMayersbach, L.E. Scheving, and J.E. Pauly, eds. Alan R. Lisa, Inc., New York, pp. 187-194. Owens, G.K., A.A.T. Geisterfer, Y.W. Yang, and A. Komoriya 1988 Transforming growth factor-induced growth inhibition and cellular hypertrophy in cultured vascular endothelial cells. J. Cell Biol., 1071771-780. Pauly, J.E. 1983 Chronobiology: Anatomy in time. Am. J . Anat., 168: 365-388. Petrovic, A.G., C.L. Oudet, and J.J. Stutzmann 1984 Temporal organization of rat and human skeletal cells. In: Cell Cycle Clocks. L.N. Edmunds, ed. Marcel Dekker, New York, pp. 325-349. Reiland, S. 1978 Growth and skeletal development of the pig. Acta Radiol. Scand. [Suppl.l (Stockh.), 358t15-22. Reppert, S.M. 1985 Maternal entrainment of the developing circadian system. Ann. NY Acad. Sci., 453t162-169. Rizzino, A. 1988 Transforming growth factor-p: Multiple effects on cell differentiation and extracellular matrices. Dev. Biol., 130: 411-422. Rosier, R.N., R.J. O'Keefe, I.D. Crabb, and J.E. Puzas 1989 Transforming growth factor is a n autocrine regulator of chondrocyte growth. Proc. 35th Orthop. Res. SOC.,Las Vegas, p. 522. Russell, J.E., D.J. Simmons, B. Huber, and B.A. Roos 1983 Meal timing as a zeitgeber for skeletal deoxyribonucleic acid and collagen synthesis rhythms. Endocrinology, 113~2035-2042. Russell, J.E., B. Grazman, and D.J. Simmons 1984a Mineralization in rat metaphyseal bone exhibits a circadian stage dependency. Proc. SOC. Exp. Biol. Med., 1761342-345, Russell, J.E., W.V. Walker, and D.J. Simmons 198413 AdrenaVparathyroid regulation of DNA, collagen and protein synthesis in rat epiphysial cartilage and bone. J. Endocrinol., 103t49-57. Russell, J.E., W.V. Walker, R.J. Fenster, and D.J. Simmons 1985 In uitro evaluation of circadian patterns of bone collagen formation. Proc. SOC.Exp. Biol. Med., 180t375-381. Schenk, R.K., D. Spiro, and J. Wiener 1967 Cartilage resorption in the tibial epiphyseal plate of growing rats. J . Cell Biol., 34:275-291. Schenk, R.K., J. Wiener, and D. Spiro 1968 Fine structural aspects of vascular invasion of the tibial epiphyseal plate of growing rats. Acta Anat. (Basel), 69:l-17. Schlechter, N.L., S.M. Russell, E.M. Spencer, and C.S. Nicoll 1986 Evidence suggesting that the direct growth-promoting effect of growth hormone on cartilage in uiuo is mediated by local production of somatomedin. Proc. Natl. Acad. Sci. USA, 8317932-7934. Seinsheimer, F., and C.B. Sledge 1981 Parameters of longitudinal growth rate in rabbit epiphyseal growth plates. J. Bone Joint Surg., 63-Ar627-630. Shapiro, I.M., and A. Boyde 1987 Mineralization of normal and rachitic chick growth cartilage: Vascular canals, cartilage calcification and osteogenesis. Scanning Microsc., 1:599-606. Shimomura, Y., and F. Suzuki 1984 Cultured growth cartilage cells. Clin. Orthop., 184:93-105. Shimomura, Y.. T. Yoneda, and F. Suzuki 1975 Osteogenesis by chondrocytes from growth cartilage of rat rib. Calcif. Tissue Res., 19: 179-187. Simmons, D.J. 1968 Daily rhythms of S3?-incorporation into epiphyseal cartilage in mice. Experientia, 24r363-364. Simmons, D.J., and G. Nichols 1966 Diurnal periodicity in the metabolic activity of bone tissue. Am. J. Physiol., 210t411-418. Simmons, D.J., L.A. Whiteside, and S.W. Whitson 1979 Biorhythmic profiles in the rat skeleton. Metab. Bone Dis. Relat. Res., 2:4964. Simmons, D.J., C. Arsenis, S.W. Whitson, S.E. Kahn, A.L. Boskey, and N. Gollub 1983 Mineralization of rat epiphyseal cartilage: A circadian rhythm. Miner. Electrolyte Metab., 9t28-37. Stambaugh, J.E., and C.T. Brighton 1980 Diffusion in the various
C.E. FARNUM AND N.J. WILSMAN
zones of the normal and the rachitic growth plate. J . Bone Joint Surg., 62-A:740-749. Stephan, F.K. 1986 Coupling between feeding- and light-entrainable circadian pacemakers in the rat. Physiol. Behav., 38537-544. Takahashi, K., N. Murakami, C. Hayafuji, and Y. Sasaki 1984 Further evidence that circadian rhythm of blinded rat pups is entrained by the nursing dam. Am. J. Physiol., 246:R539-R563. Thorngren, K.G., and L.I.Hansson 1981 Cell production of different growth plates in the rabbit. Acta Anat (Basel), 110:121-127. Thurston, M.N., and N.F. Kember 1985 In vitro thymidine labeling in human and porcine growth plates. Cell Tissue Kinet., 18:575582. Trippel, S.B. 1988 Growth plate regulation: The somatomedins. Pathol. Immunopathol. Res., 7:43-47. Trippel, S.B., S.D. Chernausek, J.J. VanWyk, A.C. Moses, and H.J. Mankin 1988 Demonstration of type I and type I1 somatomedin receptors on bovine growth plate chondrocytes. J. Orthop. Res., 61817-826. Viswanathan, N., and M.K. Chandrashekaran 1985 Cycles of presence and absence of mother mouse entrain the circadian clock of pups. Nature, 317.530-531. Wilsman, N.J., and C.E. Farnum 1988 Visualization of living termi-
nal hypertrophic chondrocytes in situ by time lapse cinematography. Proc. 34th Orthop. Res. SOC.,Atlanta, p. 270. Wilsman, N.J., and C.E. Farnurn 1989 Hypertrophic chondrocyte volume correlates positively with rate of longitudinal bone growth. Proc. 35th Orthop. Res. Soe., Las Vegas, p. 173. Wroblewski, J., and A.M. Makower 1988 X-ray microanalysis of cartilage and chondrocytes. Scanning Microsc., 2t1103-1111. Wroblewski, J., A.M. Makower, A. Pawlowski, K. Madsen, and U. Friberg 1987 Cell fractions from rat rib growth cartilage. Morphological and X-ray microanalytical investigation. J. Submicrosc. Cytol., 21t269-274. Wuthier, R.E. 1982 A review of the primary mechanisms of endochondral calcification with special reference on the role of cells, mitochondria, and matrix vesicles. Clin. Orthop., 169.219-242. Wuthier, R.E. 1988 Mechanism of matrix vesicle-mediated mineralization of cartilage. In: IS1 Atlas of Science: Biochemistry. pp. 231-240. Wyllie, A.H., J.F.R. Kerr, and A.R. Currie 1980 Cell death: The significance of apoptosis. h t . Rev. Cytol., 681251-306. Wyllie, A.H. 1987 Apoptosis: Cell death in tissue regulation. J. Pathol., 2531313-316,