вход по аккаунту



код для вставкиСкачать
THE ANATOMICAL RECORD 244284-296 (1996)
Increased Cell Diameter Precedes Chondrocyte Terminal
Differentiation, Whereas Cell-Matrix Attachment Complex Proteins
Appear Constant
Department of Anatomy and Neurobiology, Boston University School of Medicine,
Boston, Massachusetts
Background: Chondrocytes in specific areas of chick
sterna have different developmental fates. Cephalic chondrocytes become
hypertrophic and secrete type X collagen into the extracellular matrix,
whereas middle and caudal chondrocytes remain cartilagenous throughout development, continuing to secrete collagen types 11, IX, and XI. In this
report, we ask if the cell size and cytoarchitecture of chondrocytes differ in
cephalic, middle, and caudal portions of whole sterna prior to and during
hypertrophy. In addition, what is the distribution of integrin subunits and
actin associate proteins in differentiating chondrocytes?
Methods: Phalloidin was used to stain filamentous actin, and immunohistochemistry was used to localize the distribution of collagen molecules,
integrin receptor subunits, and actin-associated proteins.
Results: Chondrocytes stained for filamentous actin demonstrated that
on day 14 cephalic chondrocytes had a significantly larger diameter than
middle and caudal chondrocytes. Day 17 chondrocytes in nonhypertrophic
cephalic and middle regions of sterna were significantly smaller than hypertrophic chondrocytes and significantly larger than caudal chondrocytes. In contrast to day 14 chondrocytes, day 17 chondrocytes in the hypertrophic region demonstrated similar diameters at all cartilagenous
depths. The p1integrin subunit appeared punctate and associated with cell
membranes, allowing nonpolarized interactions with extracellular matrix
molecules. The distribution of a integrin subunits was similar to the p1
integrin subunit, although a integrin subunits also appeared cytoplasmic.
Actin-associated proteins, vinculin, and a-actinin, were associated with
F-actin,but vinculin was more specifically localized to the ends of the actin
filaments. Focal adhesion kinase was diffusely distributed throughout the
cytoplasm but also demonstrated areas of colocalization with vinculin.
Zyxin and paxillin demonstrated a punctate distribution, although paxillin
was slightly more diffuse. Using immunohistochemical detection, no difference in integrin subunit or actin associated protein distribution could be
determined between chondrocytes and hypertrophic chondrocytes.
Conclusions: The increased chondrocyte diameter observed in cephalic
regions of sterna on day 14 suggests that intracellular changes may precede
the specific hypertrophic marker, type X collagen, by several days. In addition, the presence of integrin subunits, which are known to interact with
collagen and cytoskeletal proteins, suggests that communication may exist
between chondrocytes and their extracellular matrix via these receptor
molecules. 0 1996 Wiley-Liss, Inc.
Key words: Cartilage, Chondrocyte, Hypertrophy, Type X collagen, Actin,
Throughout fetal develoDment. manv cells change
shape, position, and gene expression. TLe
'YSis embryolodcally derived from mesenchyma1 cells that synthesize and secrete type I collagen
into their extracellular matrix (ECM) prior to differenu
Received june
14,1995;accepted October 3, 1995,
Address reprint requests to K.K.H. Svoboda, Dept. of Anatomy and
Neurobiology, Boston University School of Medicine, 80 E. Concord
St., Boston, MA 02118.
tiation into chondrocytes (Upholt et al., 1979). Chondrocytes, the cells of hyaline cartilage, synthesize and
secrete collagen types 11, IX, and XI (Mendler et al.,
1989).With further fetal development, most of the cartilagenous skeleton is replaced by bone. The process of
bone development includes invasion of blood vessels,
calcification of the cartilage, and deposition of type I
collagen into the ECM. Prior to bone deposition, chondrocytes terminally differentiate into hypertrophic
chondrocytes and change their gene expression to type
X collagen (Schmid and Linsenmayer, 1985; Linsenmayer et al., 1991).
Hypertrophy has been classified as the terminal
stage of chondrocyte differentiation and a marker for
maturation (LuValle et al., 1992). Hypertrophy is
characterized by an increased cell volume of individual
chondrocytes, matrix accommodation for the larger
chondrocytes, production of type X collagen, and a concomitant decrease in collagen types 11, IX, and XI. It
has been suggested that the accumulation of type X
collagen mRNA is the result of a regulated cascade of
events termed the “hypertrophic program” and not the
cause of hypertrophy (Linsenmayer et al., 1991; Reichenberger et al., 1992).
Chondrocytesin specific areas of the embryonic avian
sternum have different developmental fates: Cephalic
chondrocytes become hypertrophic and secrete type X
collagen into the ECM, whereas middle and caudal
chondrocytes remain cartilagenous throughout development (LuValle et al., 1992). Following hypertrophy,
cephalic regions calcify, blood vessels invade, and bone
marrow forms. The mechanisms for chondrocyte terminal differentiation, however, have not been determined.
The high concentration of collagen molecules in the
ECM of chondrocytes has led us to believe that the
differentiation process may involve the interaction of
integrin receptors with collagen molecules extracellularly and the cytoskeleton intracellularly, suggesting
“outside to inside” signal transduction (Sastry and Horwitz, 1993). Before the signaling mechanisms for chondrocyte terminal differentiation can be determined,
however, it is necessary to identify the proteins involved
in the signal transduction pathway.
Until recently the distribution and function of integrin receptors on chondrocytes had not been investigated. Collagen integrin receptors that have been identified in cells other than chondrocytes include: alp1,
cxZp1, a3p1,and a,p3 (Zutter and Santoro, 1990; Hynes,
1992). Some of the OL and p integrin subunits have been
identified and investigated in chondrocytes grown in
culture, in cartilage frozen sections, and in wholemount preparations of embryonic sterna (Durr et al.,
1993; Enomoto et al., 1993; Hirsch and Svoboda, 1994;
Hirsch et al., 1994). The ciZ and p1 integrin subunits
have been individually observed associated with chondrocyte plasma membranes in wholemount sternal cartilage, supporting the hypothesis that the molecules
have a role in cell-substratum interactions (Hirsch et
al., 1994; Hirsch and Svoboda, 1994). Independent investigators have demonstrated that chondrocyte adhesion to type I1 collagen was M 8 + dependent and Ca2+
independent (Durr et al., 1993; Enomoto et al., 1993).
The three amino acid sequence, Arginine-Glycine-Aspartic Acid (RGD),necessary for ligand binding to specific integrin receptors was not necessary for type I1
collagen binding (Durr et al., 1993). The effect on differentiation and gene expression when ligands bind to
chondrocyte integrins has not been determined but has
been demonstrated in other cell types (Schwartz, 1992;
Sastry and Horowitz, 1993).
A focal adhesion model for cultured fibroblast cells
has been proposed that includes the interaction of cytoskeletal proteins (F-actin and actin associated proteins) and integrin molecules (Burridge et al., 1988,
1990; Ezzell, 1993). A similar model, however, has not
been developed for noncultured cells. Brackhan and
Burridge (1993) have shown that focal adhesions form
when cultured chondrocytes dedifferentiate, increasing
their expression of talin, vinculin, and the p1 integrin
subunit. We have previously shown that in freshly isolated whole cartilage, chondrocytes express actin as
bundles of filaments near cell surfaces and closely associated with cell plasma membranes at the optical
plane that includes the nucleus (Hirsch and Svoboda,
1993). The actin-associated protein, vinculin, associated with the ends of the actin filaments in chondrocytes from these wholemount preparations (Hirsch et
al., 1994). Additional actin associated proteins that are
known to interact with integrin cytoplasmic domains,
F-actin, and each other have not been identified in
chondrocytes in situ. Identification of actin associated
proteins are necessary to determine the possible components for signal transduction and differentiation in
this tissue.
Combining studies that involve the hypertrophic
program with the expression and function of collagen
integrin receptors will help delineate the steps in chondrocyte terminal differentiation. Cell diameter and
chondrocyte cytoarchitecture were determined with
phalloidin staining in cephalic, middle, and caudal portions of chick sterna prior to and during hypertrophic
differentiation. An increase in chondrocyte diameter in
cephalic portions of sterna, prior to type X collagen
deposition, confirmed the hypothesis that a hypertrophic program was previously initiated. In addition,
identification of integrin subunits and some actin-associated proteins, previously identified in focal adhesions of cultured fibroblasts (Ezzell, 1993), suggests
that chondrocyte terminal differentiation may involve
a signal cascade that affects cytoskeletal organization,
cellular differentiation, and gene expression.
Cartilage Preparation
White Leghorn chicken eggs were obtained from
SPAFAS (Norwich, CT) at embryonic day 14. Some day
14 eggs were incubated at 39°C to obtain chicks at later
developmental stages. Whole sterna were removed and
dissected free under sterile conditions of all tissue and
perichondrial membranes in 1X Hank‘s balanced saline solution (GIBCO Laboratories, Grand Island Biological Co., Grand Island, NY), or 1X Ham’s F-12 medium (GIBCO) supplemented with 1% penicillin/
streptomycin, 1%nonessential amino acids (Cellgro),
and 1%antibioticlantimyotic. Anatomically, the chick
sternum has a central spine (or keel) that is perpendicular to two lateral plates (Hirsch and Svoboda,
1993). The spine and lateral plates of sterna were separated with a sterile scalpel, maintaining cephalic-caudal and medial-lateral orientations. Cartilage collected
was immediately fixed for either transmission electron
microscopy, protein staining or immunohistochemistry.
Isolated sterna were fixed in freshly prepared 4%
paraformaldehyde/PBS (pH 7.4) or 3.7% formaldehyde/
PBS for 5-15 minutes. After rinsing in PBS, the tisActin Labeling
sues were permeabilized with 0.05% Triton X-100
Filamentous actin (F-actin) was visualized with flu- (Sigma) for 10 minutes. Some tissue was fixed and perorescently tagged phalloidin (Molecular Probes) (Faul- meabilized in cold methanol for 10 minutes. Tissue
stich et al., 1988). Fluorescent phallotoxins were kept used to determine integrin subunit distribution was
in a stock solution containing 300 units/3 ml methanol not treated with Triton X-100. After a brief wash in
(- 3.3 yM). A unit is defined as the amount of material PBS, the tissue was incubated for 10 minutes in 10%
needed to stain one microscope slide of fixed cells (Mo- normal goat serum (NGS) (GIBCO) to block nonspecific
lecular Probes). Best results were obtained when the secondary antibody staining.
Monoclonal antibodies specific for vinculin (mouse
phalloidin was diluted in phosphate-buffered saline
IgG isotype; Sigma); and paxillin (a gift from Dr. Chris(PBS) just prior to use.
To determine the F-actin distribution in chondro- topher Turner) were diluted 1 5 0 in 3% NGSPBS.
cytes, sterna were fixed in 4% paraformaldehyde, Polyclonal antibodies specific for a-actinin (ICN Immurinsed with PBS, extracted with 0.05% Triton X-100 nobiologicals, Lisle, IL), focal adhesion kinase (anti(Sigma, St. Louis, MO) for 10 minutes, rinsed, and in- chicken FAK IgG; Upstate Biolotechnology, New
cubated in phalloidin diluted 1:20 for 60 minutes to York), and zyxin (a gift from Dr. Mary Beckerle) were
overnight a t room temperature. For double-labeled diluted 1:50 in 3% NGS/PBS. Human integrin substudies, immunohistochemistry was performed on the units (&, az,ag,a,) were localized with specific whole
pieces of sterna as described below, prior to incubation IgG mouse monoclonal antibodies (GIBCO) diluted
in rhodamine phalloidin. After staining with phalloi- 150 in 3% NGS/PBS. In some studies a chicken spedin the tissue was rinsed three times in PBS. The cific PI integrin IgG mouse monoclonal was used
stained cartilage was mounted in slowfade mounting (Sigma).
The type X collagen antibody (AC9; Schmid and Linmedia (Molecular Probes) on glass slides with nail polish spacers, coverslipped, and viewed on the confocal senmayer, 1985) was a gift from Dr. Thomas Linsenmayer. The epitope on the type X collagen molecule is
laser scanning microscope (CLSM).
located within the triple helical domain, 19 nm from
the COOH-terminal domain. Prior to fixation, sterna
Determination of Cell Size
were incubated a t 37°C for 1.5-3 hours in 0.1% testicA grid of 16 equal spaces was placed in front of the ular hyluronidase (280 U/mg; Sigma), followed by a
CLSM computer monitor so that nonrandomly selected rinse in PBS. Sterna were subsequently fixed in 4%
chondrocytes could be measured along their long axes. freshly prepared paraformaldehyde for 15 minutes,
Days 14 and 17 lateral sternal plates, stained with rinsed, and blocked with 10% normal goat serum.
phalloidin as described above, were separated from Sterna were incubated overnight in the type X collatheir central spines and mounted onto glass slides with gen antibody a t room temperature on a rotating
nail polish spacers. Optical sections through the carti- shaker, rinsed, and detected with the FITC-conjugated
lage were obtained at the surface (perichondrocytes) AffiniPure Goat Antimouse IgG secondary antibody
and 10,20,30, and 40 ym below the surface at cephalic, (H + L chains; Jackson ImmunoResearch).
middle, and caudal regions of the lateral plates. In adWhole sterna were incubated in a specific primary
dition, chondrocytes of the cephalic region from day 17 antibody overnight at room temperature on a rotator
sterna were measured separately in medial and lateral (Belly Dancer). Excess primary antibody was removed
regions. Only one chondrocyte with identifiable cell pe- with three PBS rinses. All mouse monoclonal primary
ripheries per grid area was used for cell measurements. antibodies were detected with the FITC-conjugated
A minimum of 30 chondrocytes was measured from AffiniPure Goat Antimouse IgG secondary antibody
each area, at each optical plane, from at least two sep- (H + L chains; Jackson ImmunoResearch). The polyarate sterna.
clonal primary antibodies, a-actinin, FAK, and zyxin,
were detected with a FITC-conjugated AffhiPure Goat
antiRabbit IgG secondary antibody (H + L chains;
Statistical Analysis
Jackson ImmunoResearch). Following three 10-minute
An overall comparison of chondrocyte long axis rinses in PBS, the cartilage was mounted in antifade
lengths was conducted with a three-way analysis of mounting media (Molecular Probes) on glass slides
variance with depth, age, and region as “between” sub- with fou; nail polish spacers, coverslipped,and viewed
jects variables. More specific comparisons were then on the CLSM.
obtained by two-way analyses of variance at each age
level with depth and region as “between” subjects variables. These analyses were followed when appropriate Double-labeling procedure
by a posteriori comparisons for simple main effects and
Some chondrocytes were double labeled for transthe Tukey HSD method of pair comparisons among membrane, cytoskeletal, or extracellular matrix promeans (Kirk, 1982; Hayes, 1988). All comparisons were teins. Subsequent to immunohistochemistry, some carassessed according to one- or two-sided methods to pro- tilage was stained with rhodamine phalloidin, as
vide “experimenter wise” control for Type I errors (Mc- described above, to demonstrate the association of F-acDonald and Thompson, 1967). P values were set a t tin and a specific actin associated protein, or F-actin
and type X collagen. To determine the localization of
two cytoskeletal proteins or the localization of a cytoskeletal protein and an integrin subunit, the immunohistochemistry protocol was performed twice on the
same piece of cartilage with one monoclonal and one
polyclonal antibody. The second primary antibody,
however, was detected with a secondary antibody conjugated to Cy3 (Cy3-conjugated AffiniPure Goat Antimouse IgG secondary antibody, H + L chains; Jackson
ImmunoResearch). In addition, it should be noted that
chondrocytes double labeled for integrin subunits and
cytoskeletal proteins were permeabilized with 0.05%
Triton X-100 after detection of the integrin subunits
due to their transmembrane location.
mm camera/50 mm lens on TMAX 100 film or Kodak
Elite daylight film, respectively. Color prints were produced with Adobe Photo Shop, Mt. View, Calif. All
control tissue was collected, analyzed, enhanced, and
photographed with the same conditions as the experimental tissue.
General Morphological Features
The ultrastructural morphological features of chondrocytes and hypertrophic chondrocytes has been established (Farnum and Wilsman, 1987; Hirsch and
Svoboda, 1993). Both chondrocytes and hypertrophic
chondrocytes are surrounded by an abundant collagenous matrix. These cells are characterized as large
spherical or ellipsoid shaped with large central nuclei.
Rough endoplasmic reticulum (RER), Golgi apparati,
and mitochondria are abundant in the cytoplasm of
both chondrocytes and hypertrophic chondrocytes. The
RER distribution also has been observed with confocal
laser scanning microscopy and DiOC6(3) staining
(Terasaki, 1989; Hirsch and Svoboda, 1994). Uneven
cell peripheries appeared to contain fine filaments that
also stain with phalloidin (Hirsch and Svoboda, 1993).
To determine background fluorescent staining,
pieces of cartilage were incubated in the absence of
primary antibody. The tissue was then treated with the
secondary antibody conjugated to FITC and viewed
with the CLSM at the same conditions as the experimental tissue.
In addition to secondary antibody controls, it was
necessary to demonstrate FITC fluorescent probes
were not recorded onto the rhodamine photomultiplier
tube and that the rhodamine fluorescence was not reType X Collagen in ECM of Hypertrophic Cartilage
corded onto the FITC photomultiplier tube. Therefore,
we routinely had single labeled material for each fluChondrocytes in medial cephalic areas of chick
orescent tag scanned at the same confocal microscope sterna become hypertrophic by embryonic day 17 (Fig
settings with the opposite filter set and photomultiplier lA), prior to bone deposition. Hypertrophic chondrotube. These samples were termed “cross-over”controls. cytes produce type X collagen with a concomitant decrease in the synthesis of collagen types 11, IX and XI
Confocal Laser Scanning Microscopy
(Upholt et al., 1979; Linsenmayer et al., 1991). Medial
The specimens were analyzed with the Leica upright cephalic areas of days 14 and 17 sterna were double
confocal laser scanning microscope that is equipped labeled with rhodamine phalloidin and a specific antiwith an argon ion laser with an output power of 2-50 body to type X collagen (Schmid and Linsenmayer,
mW, two photomultiplier tubes, and narrow band fil- 1985; Linsenmayer et al., 1991). At day 14, actin is
ters for double labeling experiments. The continuously observed as bands of filaments (Fig. lB, arrowheads)
variable detection pinhole was set at the minimum size and it also associates with cell membranes (Fig. 1B).
for optimal signal. Smaller apertures (pinhole) allow a No type X collagen is observed in the extracellular manarrower optical section and less background (Pawley, trix (ECM) of day 14 chondrocytes (Fig. 1B). At day 17,
1990;Wilson, 1990).The typical z-series is composed of type X collagen is observed in the ECM of hypertrophic
optical sections in the xy optical plane. These images chondrocytes double labeled for F-actin (Fig. 10. As
are en face optical sections through the vertical axis of previously observed (LuValle et al., 1992) the greatest
the tissue. Each image of the series was taken at 0.3- concentration of type X collagen was observed in the
1.0 pm intervals in the 512 x 512 pixel format. Carti- pericellular matrix (Fig. lC, arrowheads) with less
type X collagen in the interstitial matrix (Fig. 1C, arlage specimens were viewed with either a 2 5 Plan
Fluotar (n.a. = 0.75) water immersion lens, or a 50 x row). It has been suggested that type X collagen in the
Plan Fluotar (n.a. = 1.0) water immersion lens with a interstitial matrix may be associated with collagen
working distance of 100 pm (Leica). Pixel size for the types 11, IX and XI (Schmid and Linsenmayer, 1990).
5 0 objective
lens at zooms 1 and 2 were 0.196 and The appearance of type X collagen on day 17 (Fig. 1C)
confirms that chondrocytes in medial cephalic areas of
0.098, respectively.
Merged images were computer generated from two day 14 sterna have terminally differentiated. By emoptical sections in the same focal plane recorded from bryonic day 20, medial cephalic regions of chick sterna
the FITC or rhodamine photomultiplier tubes. The im- are replaced by bone and osteocytes that stain for F-acages were electronically colored so that the individual tin (data not shown). At the hypertrophic cartilagel
images were green or red. The combined image main- bone interface, type X collagen is still observed in the
tained the red or green color in pixels that do not over- ECM (Fig. 1D). No type X collagen is observed in carlap. Pixels that contain information from both images tilagenous areas lateral or caudal to the hypertrophic
were colored yellow, indicating an overlap in the digi- region or regions that have been replaced by bone (data
not shown).
tal information.
Confocal images were analyzed, enhanced, and
Cell Sizes Determined with Phalloidin Staining
stored on optical discs. Black and white or pseudocolor
As previously stated, the accumulation of type X colimages were computer generated with a minimum of
computer enhancement and photographed with a 35 lagen mRNA is the result of a regulated cascade of
Fig. 1. Type X collagen in the ECM of hypertrophic cartilage. Double-labeled pseudocolored micrographs demonstrate the distribution
of type X collagen in the ECM of hypertrophic ehondrocytes. The
hypertrophic region in the embryonic sternum is located in the medial
cephalic region (A). At day 14 (B), sternal cartilage double labeled for
F-actin (arrowheads) and type X collagen demonstrates that chondrocytes in medial cephalic areas are surrounded by a n ECM that does
not contain type X collagen. At day 17 (C),cartilage double labeled for
F-actin and type X collagen demonstrates that chondrocytes in the
hypertrophic region are surrounded by a n ECM that does contain type
X collagen. A greater concentration of type X collagen is found in the
pericellular matrix (arrowheads) with less type X collagen in the interstitial matrix (arrow). The secretion of type X collagen in the ECM
confirms the fact that chondrocytes in this area have terminally differentiated into hypertrophic chondrocytes. After bone has replaced
the hypertrophic cartilage by day 20 (D), type X collagen can still be
observed a t the hypertrophic cartilage/bone interface. Scale bars =
10 km.
Day 17 Cbondrocytes
Day 14 Chondmytes
Day 12 Cbondrocytes
Cephalic Sterna
Cephalic(htersl) Sterna
f S
Surface 10 pm 20 jun
Depth in Cartilage
Surface 10 pm 20 p
Depth in Cartilage
Depth in Cartilage
Fig. 2. Cell diameters determined with phalloidin staining. A nonrandom selection of chondrocytes stained with phalloidin were chosen
and measured along their long axes. Chondrocytes at the surface (perichondrocytes)and 10, 20,30, and 40 pm below the surface of embryonic day 12 (A), day 14 (B), and day 17 (C) sterna were investigated.
Average chondrocyte diameters are presented in Table I. Analysis of
variance tests determined significant differences in cell diameter (P <
0.05). At day 12 (A), cephalic and middle chondrocytes are significantly larger than caudal chondrocytes. At day 14 (B), chondrocytes in
cephalic regions are significantly larger than chondrocytes in middle
and caudal regions. At day 17 (C), hypertrophic chondrocytes in medial cephalic regions are confirmed by the immunolocalization of type
X collagen (*) (see Fig. 1). Hypertrophic chondrocytes demonstrate the
same diameter 40 km deep into the cartilage and are significantly
larger than all other chondrocytes. In contrast to day 14 (B), middle
chondrocytes are no longer significantly different from the lateral
cephalic chondrocytes at day 17 (C).
events termed the “hypertrophic program” and not thecause of hypertrophy (Linsenmayer et al., 1991; Reichenberger et al., 1992).To confirm if hypertrophy does
in fact precede the appearance of type X collagen, the
cell size (long axis diameter) and cytoarchitecture of
chondrocytes were examined in cephalic, middle, and
caudal portions of whole 12, 14, and 17 day sterna at
the surface and at 10 pm intervals to a depth of 40 pm.
Figure 2 shows the changes in cell diameters observed
at the different regions to a depth of 20 pm. Beyond 20
pm, no significant changes (P < 0.05) in cell size are
observed. Average chondrocyte diameter for each region and depth is presented in Table I.
Chondrocytes were stained for F-actin and measured
along their long axes, demonstrating that at all ages
perichondrocytes are larger than all chondrocytes below the surface except in medial cephalic regions of
day 17 sterna. On day 12 (Fig. 2A, Table I), cephalic
chondrocytes appear slightly smaller than those in
their respective areas and depths on day 14 and are
significantly smaller than hypertrophic chondrocytes
on day 17. Day 12 cephalic and middle chondrocytes
are significantly larger than day 12 caudal chondrocytes. Middle and caudal chondrocytes at day 12, however, are similar or slightly larger in diameter when
compared to days 14 and 17 chondrocytes in their
respective areas and depths. Many perichondrocytes
and chondrocytes appeared to have a fibroblastic morphology a t day 12 in middle and caudal regions, suggesting that some cells had not completely differentiated from their mesenchymal stage into chondrocytes.
This is especially apparent as caudal perichondrocytes
decreased in diameter between days 12 and 14 with
subsequent increased diameters between days 14 and
On day 14 (Fig. 2B, Table I), cephalic chondrocytes
are significantly larger than middle and caudal chondrocytes. In addition these chondrocytes are larger
than cephalic chondrocytes at day 12, suggesting an
onset of cell shape change preceding type X collagen
deposition. Day 17 chondrocytes in nonhypertrophic
(lateral) cephalic and middle regions of sterna are significantly smaller than hypertrophic chondrocytes and
significantly larger than caudal chondrocytes. This
suggests that lateral cephalic chondrocytes continue to
hypertrophy while middle chondrocytes have initiated
the hypertrophic program. In contrast to day 14 chondrocytes, day 17 hypertrophic chondrocytes are significantly larger than any other chondrocytes and remain
the same size 40 p,m deep into the tissue (Fig. 2C, Table
I). The terminal differentiation of these cells into hypertrophic chondrocytes is confirmed by the production
of type X collagen (Fig. 1C).
The increased chondrocyte diameter observed in
cephalic regions of sterna on day 14 and in lateral
cephalic and middle regions on day 17 suggests that
intracellular changes may precede the specific hypertrophic marker, type X collagen, by several days. This
increased cell diameter observed, prior to type X collagen synthesis and deposition, supports the hypertrophic program theory (Linsenmayer et al., 1991; Reichenberger et al., 1992). The increased cell size
observed at all depths in the hypertrophic region of day
17 sterna (Fig. 2C) agrees with the accepted theory
that type X collagen synthesis and secretion (Fig. 1C)
correlate to chondrocyte hypertrophy.
TABLE 1. Average long axis length of 12-, 14-, and 17-day chondrocytes at various cartilage depths and at
various cartilage regions'
12-day sterna
14-day sterna
17-day sterna
10 pm
11.18 +- 0.27
10.94 t 0.24
9.36 t 0.25
10 pm
12.92 t 0.70
9.79 t 0.31
7.94 t 0.23
10 pm
20 pm
9.75 t 0.23
9.82 t 0.22
8.03 t 0.18
20 pm
10.60 t 0.34
8.26 t 0.23
7.62 t 0.28
20 pm
30 pm
9.33 t 0.21
9.46 2 0.20
7.30 t 0.22
30 pm
10.32 t 0.44
7.69 t 0.21
6.89 t 0.26
30 pm
40 pm
8.78 t 0.20
9.33 t 0.19
6.93 rtr 0.21
40 pm
9.95 t 0.37
8.49 t 0.34
6.80 2 0.25
40 pm
14.80 t 0.40
19.33 t 0.77
16.84 t 0.63
15.15 t 0.51
14.44 t 0.38
10.99 t 0.29
10.60 t 0.31
7.26 rtr 0.21
14.19 t 0.40
10.05 t 0.42
9.73 t 0.39
6.95 t 0.19
14.14 t 0.54
9.50 t 0.43
9.19 t 0.28
7.08 t 0.19
13.93 t 0.38
9.23 t 0.32
9.18 t 0.30
6.70 t 0.19
16.79 t 0.45
18.29 t 0.37
16.48 t 0.39
20.40 t 0.75
19.37 t 0.43
13.06 2 0.431
'Values presented as average diameter (pm)
standard error.
meabilized (Fig. 3H) chondrocytes that have been
The high concentration of collagen molecules in the incubated in the human p1 antibody. With permeabilchondrocyte ECM and the change in gene expression isation, the p1 integrin subunit demonstrates a diffuse
accompanying differentiation suggested that molecules pattern that extends from nuclei to plasma membranes
found in cellular focal adhesions also may be found in (Fig. 3G). In contrast, chondrocytes that were not perchondrocytes. Using immunohistochemistry, we have meabilized demonstrate a very discrete, punctate
integrin subunit (Fig. 3H).
demonstrated the localization of (Y and p integrin sub- staining pattern of the
integrin subunit has been
units to chondrocyte plasma membranes from wholemount preparations of cartilage (Fig. 3). The a3,az,a", demonstrated by three-dimensional analysis to associand p1 integrin subunits were investigated because ate with chondrocyte plasma membranes (Hirsch and
these subunits form the heterodimers that act as col- Svoboda, 1994).These results are consistent with other
lagen receptors (Hynes, 1992). Based on morphological cells where the p1 integrin subunits are located and
results, no obvious difference in integrin subunit local- possibly restricted to cell surfaces.
ization was observed between cells from day 14 (Fig. 3) Identification of Actin-Associated Proteins in Chondrocytes
and day 17 sterna (data not shown).
In addition to localizing integrin subunits, the disChondrocytes fixed, permeabilized, and stained for tribution of F-actin and many actin associated proteins
the ci3 integrin subunit demonstrate a punctate distribution associated with plasma membranes and between dividing cells (Fig. 3A,B). Very little staining of
the a3 integrin subunit is observed in the cytoplasm of
Fig. 3.The distribution of integrin subunits in 14-day chondrocytes.
the chondrocytes (Fig. 3B). A similar distribution of the
Lateral sternal plates from days 14 and 17 sterna were incubated in
a2integrin subunit is observed either with (Fig. 3C) or specific
monoclonal primary antibodies to the integrin subunits, a3,
without permeabilisation (Fig. 3D) following fixation. a2,and p1and detected with a secondary antibody conjugated to FITC.
Similar to the a3 integrin subunit (Fig. 3A,B), the a2 Chondrocytes were optically sectioned and viewed with a Leica conintegrin subunit appears to associate with plasma focal laser scanning microscope and photographed. No difference in
distribution is observed between day 14 chondrocytes (A-H)
membranes, however, cytoplasmic staining is also ob- protein
and day 17 chondrocytes or hypertrophic chondrocytes (data not
served. The cytoplasmic staining appears to demon- shown). Nuclei (n) that do not stain can be used to orient the tissue. At
strate a perinuclear Golgi pattern, suggesting that the lower (A) and higher (B) magnifications, chondrocytes fixed and perantibody may be binding to proteins that are being meabilized in cold methanol demonstrate a punctate distribution of
the a3 integrin subunit. This distribution appears to be associated
synthesized andlor transported to the cell surface. In with
chondrocyte plasma membranes. Distribution of the a2 integrin
addition, staining of the a, integrin subunit is less subunit in chondrocytes fixed and permeabilized in cold methanol (C)
abundant; therefore, it was necessary to increase the appears to be associated with chondrocytes plasma membranes, alvoltage of the CLSM to identify these proteins (data not though this protein is also observed in Golgi associated areas. The a2
integrin subunit distribution is more obvious at a higher magnificashown).
tion and does not appear to differ when the chondrocytes were fixed in
The p1 integrin subunit was localized using both an- paraformaldehyde without subsequent permeabilisation (D). Chontichicken (Figs. 3E,F) and antihuman (Fig. 3G,H) drocytes, fixed in paraformaldehyde and incubated in an antichicken
antibody, demonstrate a plasma membrane associated distribution
monoclonal antibodies. Without permeabilising the
the integrin subunit (E,F).
A more punctate distribution of the p1
cells, the chicken p1 antibody appears to associate dif- of
integrin subunit is observed associated with plasma membranes and
fusely with chondrocyte edges and demonstrates very between dividing cells at a higher magnification (F).The distribution
little cytoplasmic staining (Fig. 3E). At higher magni- of the integrin subunit was also determined with the antihuman p1
fication, a more punctate distribution associated with antibody (G,H). In permeabilized chondrocytes the p1 integrin subunit appears cytoplasmic and associated with cell membranes (G).In
plasma membranes and between dividing cells is ob- contrast,
nonpermeabilized chondrocytes demonstrate a punctate disserved (Fig. 3F). An obvious difference in staining pat- tribution of the p1 integrin subunit associated with plasma memtern is observed in permeabilized (Fig. 3G) and unper- branes, and is not obvious in the cytoplasm (H). Scale bars = 10 pm.
Identification of lntegrin Receptor Subunits in Chondrocytes
Fig. 3.
were determined in wholemount preparations of cartilage (Fig. 4). Using immunohistochemistry, no difference is observed in the distribution of these proteins
between chondrocytes of 14 and 17 day sterna and from
hypertrophic chondrocytes of day 17 sterna. The individual distributions of F-actin (Hirsch and Svoboda,
1993) and vinculin (Hirsch et al., 1994) in chondrocytes
has been reported; however, the association of the two
proteins in chondrocytes has not been determined. A
day 14 chondrocyte double labeled for F-actin (Fig. 4A,
arrowheads, red) and vinculin (Fig. 4A, arrow, green)
demonstrates that vinculin colocalizes a t the ends of
actin filaments and adjacent to the cell periphery. This
distribution is expected because vinculin has been
shown t o bind the COOH-terminus of a-actinin, that
cross links F-actin and the p1 integrin receptor subunit
(Otey et al., 1990; Pavalko et al., 1991), and directly to
F-actin (Menkel et al., 1994). A day 17 chondrocyte
double labeled for vinculin (Fig. 4B, arrow, green) and
focal adhesion kinase (FAK; Fig. 4B, arrowheads, red)
demonstrates many punctate areas of colocalization
(Fig. 4B, yellow) on the chondrocyte surface, as expected, since both proteins have been identified in focal
The punctate distribution of FAK resembles that of
the p l integrin subunit distribution (Fig. 3F). Day 14
chondrocytes double labeled for FAK (Fig. 4C, red) and
PI integrin subunits (Fig. 4C, arrowhead, green) demonstrate a more cytoplasmic distribution of FAK a t the
level of chondrocyte nuclei. Many bright, punctate areas of FAK, however, also appear to colocalize with the
p1 integrin subunit near plasma membranes (Fig. 4C,
arrows, yellow) as expected since FAK has been shown
to have a binding site for the cytoplasmic domain of
this subunit (Schaller and Parsons, 1994). Alpha-actinin has been shown to bind the p1 integrin subunit and
to cross link F-actin (Otey et al., 1990; Pavalko et al.,
1991). Chondrocytes from a day 14 sternum demonstrate that a-actinin (Fig. 4D, green) colocalizes (Fig.
4D, arrow, yellow) a t the ends of actin filaments (Fig.
4D, arrowheads, red), although this distribution is less
distinct when compared to the punctate vinculin distribution (Fig. 4A,B).
Zyxin is another actin-associated protein found in
focal adhesions that has been shown to bind the aminoterminus of a-actinin near cell peripheries (Crawford
et al., 1992). The distribution of zyxin (Fig. 4E) in single-labeled day 14 chondrocytes appears punctate
throughout the cytoplasm. A slightly more diffuse but
similar distribution is observed when paxillin is localized in single labeled day 17 chondrocytes (Fig. 4F).
Paxillin can be phosphorylated and has been suggested
to localize with FAk and vinculin in focal adhesions
(Burridge et al., 1988, 1990; Ezzel, 1993). The distribution of zyxin and paxillin observed in chondrocytes appears to be consistent with the focal adhesion model.
Control experiments were performed when chondrocyte protein localization was determined in wholemount preparations of cartilage. Sterna incubated in
the absence of a primary antibody do not demonstrate
significant nonspecific binding of the secondary antibody (Hirsch and Svoboda, 1994; Hirsch et al., 1994).In
addition, single-labeled FITC, Cy3, and rhodamine
samples scanned with the opposite excitation and bar-
rier filter sets demonstrate that crossover between
samples did not occur (data not shown).
The purpose of the present study was to determine
some of the processes that are involved in the development of hyaline cartilage, as chondrocytes terminally
differentiate into hypertrophic chondrocytes. The hypertrophic program theory states that type X collagen
synthesis and secretion are the result rather than the
cause of hypertrophy (Linsenmayer et al., 1991; Reichenberger et al., 1992). We report an increase in
chondrocyte diameter 3 days prior to the appearance of
type X collagen supporting this theory. In addition, we
have identified, in chondrocytes and hypertrophic
chondrocytes, cytoskeletal and transmembrane proteins that have been shown in other cell types to affect
cytoskeleton organization, gene expression, and differentiation (Schwartz, 1992; Hynes, 1992; Sastry and
Horwitz, 1993; Schaller and Parsons, 1995). Although
no functional studies have been performed in our chondrocytes, the presence of these proteins and the high
concentration of collagen molecules in the extracellular matrix (ECM) suggests that chondrocyte terminal
differentiation also may be regulated via signal transduction.
Cellular signaling initiated by integrin receptor ligation results in cytoskeletal organization, modulation
of differentiation, and induction of gene expression
(Sastry and Horwitz, 1993). Subsequent to receptor ligation, molecular events include tyrosine phosphorylation, alkalization, and changes in intracellular calcium
concentrations (Hynes, 1992; Schwartz, 1992). Bonen
and Schmid (1991) have shown that type X collagen
synthesis increases in a time- and dose-dependent
manner in the presence of extracellular calcium. Increased concentrations of extracellular calcium have
been shown to elicit changes in intracellular levels of
calcium in parathyroid cells and osteoclasts by stimulating calcium entry into cells or release from intracel-
Fig. 4. Actin-associated protein distribution in day 14 and day 17
chondrocytes.Proteins were identified in chondrocytes from days 14
and 17 sterna. Filamentous actin was identified with rhodamine phalloidin staining. Actin-associated proteins and p l integrin receptor
subunits were identified with primary antibodies detected with a secondary antibody conjugated to FITC or Cy3. Double-labeled images
were optically sectioned individually with the confocal laser scanning
microscope (CLSM) and superimposed using CLSM software. Superimposed images appear yellow where red and green staining overlaps.
Nuclei (n) do not stain and can be used to orient the tissue. A chondrocyte from a day 14 sternum (A) demonstrates that vinculin (arrow,
green) is associated with the ends of actin filaments (arrowheads,
red). A chondrocyte from a day 17 sternum (B) demonstrates vinculin
(arrow, green) associated near cell membranes. This distribution appears to colocalizewith focal adhesion kinase (FAK;arrowheads,red)
in many areas as expected since both proteins are found in focal adhesions. Day 14 chondrocytes (C) are double labeled for the
integrin subunit (green) and FAK (red). The p l integrin subunit is associated with cell membranes (arrowheads), whereas the FAK
distribution is cytoplasmic, also demonstrating specific punctate areas (arrows). Alpha-actinin (green) in chondrocytes from a day 14
sternum (D) is localized to the ends (arrow)of actin filaments (arrowheads, red). Single-labeled chondrocytes from a day 14 sternum (El
demonstrate a punctate distribution of zyxin. Chondrocytes from a
day 17 sternum (F) demonstrate that the paxillin distribution is also
punctate and cytoplasmic. Scale bars = 10 pm.
Fig. 4.
lular stores (Bonen and Schmid, 1991). It also has been
shown that calcium release from intracellular stores
results when inositol triphosphate (IP,) is produced
from phosphatidylinositol 4,5-biphosphate (PIP2). Integrin receptor ligation activates the phosphorylation of phosphatidylinositol phosphate (PIP) to PIP,
(Schwartz, 1992), making it available for the production of IP, and subsequent increases in calcium concentrations. Combining experiments that examine signaling cascades, hypertrophic program markers, and the
expression and function of collagen integrin receptors
may lead to a better understanding of chondrocyte terminal differentiation and bone development.
We have shown that chondrocyte diameter in cephalic regions on day 14 and in lateral cephalic and middle regions on day 17 increases prior to type X collagen
synthesis and secretion (Fig. 2), confirming the theory
of a previously initiated hypertrophic program. Type X
collagen localization (Linsenmayer et al., 1991) in medial cephalic areas of day 17 sterna confirmed that
chondrocytes had terminally differentiated (Fig. 1C).
Previous studies using flow cytometry determined the
size of chondrocytes from avian tibiotarsi (Schmid et
al., 1991). Nonhypertrophic chondrocytes were 5-8
pm in diameter and did not produce type X collagen.
Chondrocytes from the same tissue that were cultured
for 8 weeks did produce type X collagen and were
10-12 pm in diameter, although some cells were as
large as 30 pm. The cell sizes reported in this study for
nonhypertrophic sternal chondrocytes were comparable to those found in tibiotarsi, although the sternal
chondrocytes were slightly larger (Table I). The average diameter of sternal hypertrophic chondrocytes was
also comparable but slightly larger (Table I) to those in
tibiotarsi, although sternal hypertrophic chondrocytes
were never as large 30 Fm, ranging 10-18 pm. The
larger hypertrophic chondrocytes observed in tibiotarsi
may be a result of the culturing process or the measuring of free-floating cells. Overall, these results demonstrate that in vitro and in vivo studies support the previously initiated hypertrophic program theory.
In contrast, Pacifici et al. (1991) demonstrated that
articular and caudal sternal chondrocytes grown in
suspension for 4-5 weeks did not show a relationship
between cell size or diameter and type X collagen production. This inconsistency may be due to the fact that
articular cartilage and caudal sternal cartilage do not
normally undergo endochondrial ossification as does
cephalic portions of sterna. In addition, it should be
noted that hypertrophic cell size determined in these
studies (Pacifici et al., 1991) were performed in cultured cells, whereas the diameter of chondrocytes and
hypertrophic chondrocytes in the current study were
recorded from wholemount preparations of freshly isolated cartilage. We hypothesize that the changes in
chondrocyte diameter and gene expression that occurs
during cartilage development may be regulated by the
interaction of integrin receptors between the chondrocyte cytoskeleton and the extracellular matrix (ECM).
The inhibition of chondrocyte terminal differentiation
and type X collagen secretion in organ cultured sterna
grown in the presence of p l integrin antibodies further
supports this hypothesis (Hirsch et al., 1995).
As discussed, the binding of integrin receptors to
ECM ligands (outside to inside signaling) may initiate
a signal cascade that affects cytoskeletal organization,
cellular differentiation, and gene expression (Sastry
and Horwitz, 1993). Gene expression in chondrocytes
changes with differentiation, yet specific collagen integrin receptors and associated cytoskeletal proteins
have not been previously identified in chondrocytes in
situ. Using confocal laser scanning microscopy, we
have demonstrated that integrin subunits that have
been determined previously to form collagen receptors
(Hynes, 1992) are distributed on chondrocyte plasma
membranes. Individually, the a3,a,, and p1 integrin
subunits were examined and demonstrated punctate
distributions that associated with chondrocyte plasma
membranes, allowing nonpolarized interactions with
ECM molecules (Fig. 3). This distribution was different
than that observed in cultured fibroblast cells where
integrin receptors are found clustered in membranes
adjacent to the basal substrate in a polarized distribution.
When combined as ap heterodimers, integrin subunits act as transmembrane linkers of the ECM and
the cytoskeleton and can form collagen receptors that
transmit signals into the chondrocyte when these receptors bind ligands. Although specific signaling
caused by the binding of collagen integrin receptors to
ligands has not been demonstrated in chondrocytes in
this study, the downstream effects of integrin binding
has been shown in other cell types (Schwartz, 1992;
Sastry and Horwitz, 1993). Collagen activation of the
integrin receptor, a2p1,in platelets has shown to stimulate second-messengerpathways and protein tyrosine
phosphorylation (Staatz et al., 1989, 1991; Hynes,
1992). The a3p1 integrin receptor that binds fibronectin, laminin, and collagen also has been shown to stimulate tyrosine phosphorylation of proteins when triggered in KB carcinoma cells and NIH 3T3 fibroblasts
(Hynes, 1992). In addition it has been determined that
integrin receptor binding to collagen is highly specific
and dependent on collagen molecule conformation
(Kuhn and Eble, 1994). Future studies of chondrocytes
may demonstrate that collagen integrin receptor binding may have signaling effects similar to that observed
in other cell types. The effects of outside to inside signal transduction may result in cellular differentiation
and changes in gene expression.
The focal adhesion model that suggests an interaction
of extracellular and intracellular proteins via the integrin receptor has been developed from cultured cells
and has been shown to be a site for signal transduction
(Burridge et al., 1988, 1990; Luna and Hitt, 1992; Ezzell, 1993; Sastry and Horwitz, 1993).In addition, other
protein-protein interactions have been demonstrated
with isolated molecules (Luna and Hitt, 1992). Association of a and p integrin subunits form transmembrane
receptors; however, it is the cytoplasmic tail of the p
subunit that binds talin and a-actinin, two actin-associated proteins that also bind F-actin (Otey et al., 1990;
Pavalko et al., 1991).The amino and carboxy terminals
of a-actinin bind zyxin and vinculin, respectively
(Crawford et al., 1992). In addition, vinculin has been
shown to bind F-actin, talin, paxillin, and other vinculin
molecules (Menkel et al., 1994). Focal adhesion kinase
(FAK), a tyrosine kinase localized to focal adhesions,
also has binding domains for the p integrin subunit and
paxillin (Schaller and Parsons, 1994). Recently, it also
has been demonstrated that FAK autophosphorylates
its tyrosine residues and subsequently phosphorylates
paxillin. Phosphorylated paxillin is then able to bind
the SH2 domain of Crk, Csk, and Src (Schaller and
Parsons, 1995).
Our results (Fig. 4) demonstrate that many of the
actin-associated proteins that have been studied in
vitro show a similar distribution in vivo. In cultured
chondrocytes that have adopted a flattened morphology, vinculin, and a-actinin were shown to localize at
the ends and along F-actin, respectively (Marchisio et
al., 1984). In our wholemount cartilage model, chondrocytes demonstrated similar distribution patterns,
even though our chondrocytes maintained their round
morphology. These results suggest that cytoskeletal
proteins found in chondrocytes are necessary for cell
structure, differentiation, and gene expression while
being less involved with the assembly and disassembly
of F-actin necessary for cell movement (Stossel, 1993).
The presence and association of these proteins in our in
vivo model, however, also provides evidence that focal
adhesions in cultured cells do exist and are not an artifact of culturing. Analogous to focal adhesions in
cultured cells, it has been proposed that cytoskeletal
and ECM protein interactions in cells from whole tissues be identified as “cell-matrix attachment complexes (CMAXs)”(Jirawuthiworavong et al., 1994; Wu
et al., 1995).
A role for differentiation and gene expression regulated by cytoskeletal-ECM interactions is further supported by the expression of FAK, paxillin, and zyxin in
chondrocytes (Fig. 4). Tyrosine phosphorylation of
FAK has been shown in epithelial cells, fibroblasts,
and platelets following specific integrin receptor ligation, including the binding of a2p1to collagen (Schaller
and Parsons, 1994). Colocalization of the PI integrin
subunit and FAK and expression of the a2 integrin
subunit on chondrocyte plasma membranes suggest
that tyrosine phosphorylation may occur if the chondrocyte a2Plintegrin receptor binds collagen. Tyrosine
phosphorylation of FAK, resulting from integrin receptor ligation, leads to autophosphorylation, a process
necessary to phosphorylate other proteins such as paxillin (Schaller and Parsons, 1994). Phosphorylation of
paxillin has been shown to recruit other molecules such
as Csk and Crk that allow signals t o be transmitted to
the cytoskeleton and/or nucleus (Schaller and Parsons,
1995). In addition, zyxin has been characterized to exhibit three LIM domains that have been determined
previously in other proteins to have a role in transcriptional regulation and cellular differentiation (Sadler et
al., 1992). Zyxin and paxillin both demonstrated punctate, cytoplasmic distributions in chondrocytes and hypertrophic chondrocytes.
The data presented demonstrate that changes in
chondrocyte diameter precede the appearance of the
specific hypertrophic marker, type X collagen. The
changes observed in cell shape, diameter, and gene expression may be regulated by the interaction of integrin receptors with the chondrocyte cytoskeleton and
ECM. Future studies that combine studying chondrocytes in their natural environment with signaling that
occurs via integrin receptors will allow us to gain a
better understanding of chondrocyte terminal differentiation. Inhibition of “outside t o inside’’ signaling may
prevent subsequent chondrocyte differentiation as
measured by the synthesis and secretion, or lack of
type X collagen. The lack of type X collagen production
by hypertrophic chondrocytes may then further retard
developmental processes, including cartilage calcification and done deposition.
We thank Daniel Orlow for his photographic assistance. We thank Dr. Carl Franzblau, associate dean for
Graduate Biomedical Sciences, for his assistance in establishing the Leica Confocal Facilities at Boston University School of Medicine. We thank Dr. Vickery
Trinkaus-Randall for her advice on the integrin studies. We also thank Drs. Thomas Linsenmayer, Mary
Beckerle, and Christopher Turner for their generous
antibody gifts. This work was supported by a Boston
University Graduate Student Research Award and
NIH EY-08886.
Bonen, D.K., and T.M. Schmid 1991 Elevated extracellular calcium
concentrations induce type X collagen synthesis in chondrocyte
cultures. J. Cell Biol., 115(4):1171-1178.
Brackhan, J., and K. Burridge 1993 Focal adhesions in chondrocyte
de-differentiation. Mol. Biol. Cell, 4t290a.
Burridge, K., K. Fath, T. Kelly, G. Nuckolls, and C. Turner 1988 Focal
adhesions: Transmembrane junctions between the extracellular
matrix and the cytoskeleton. Ann. Rev. Cell Biol., 4t487-525.
Burridge, K., G. Nuckolls, C. Otey, G. Pavalko, K. Simon, and C.
Turner 1990 Actin-membrane interaction in focal adhesions. Cell
Diff. Dev., 32:337-342.
Crawford, A.W., J.W. Michelsen, and M.C. Beckerle 1992 An interaction between zyxin and a-actinin. J. Cell Biol., 116t1381-1393.
Durr, J., S. Goodman, A. Potocnik, H. von der Mark and K. von der
Mark 1993 Localization of pl-integrins in human cartilage and
their role in chondrocyte adhesion to collagen and fibronectin.
Exp. Cell Res., 207t235-244.
Enomoto, M., P.S. Leboy, A.S. Menko and D. Boettiger 1993 p l integrins mediate chondrocyte interaction with type I collagen, type I1
collagen, and fibronectin. Exp. Cell Res., 205:276-285.
Ezzell, C. 1993 Just the FAK(s), Ma’am: Researchers investigate a
new signaling enzyme. J . NIH Res., 5(12):49-54.
Farnum, C.E., and N.J. Wilsman 1987 Morphological stages of the
terminal hypertrophic chondrocyte of growth plate cartilage.
Anat. Rec., 219t221-232.
Faulstich. H.. S. Zobelev. G. Rinnerthaler. and J.V. Small 1988 Fluorescent phallotoxins as probes for filamentous actin. J. Musc.
Res. Cell Motil., 9t370-383.
Hayes, W.L. 1988 Statistics, 4th ed. Holt, Rinehart and Winston, New
Hirsch, M.S., and K.K.H. Svoboda 1993 Confocal microscopy of whole
mount embryonic cartilage: Intracellular localization of F-actin,
chick prolyl hydroxylase and type I1 collagen mRNA. Micron,
Hirsch, M.S., L.Y. Law, V. Trinkaus-Randall, and K.K.H. Svoboda
1994 The intracellular distribution of vinculin and alpha 2 integrin in epithelial cells and chondrocytes. Scanning, 16(5):275-284.
Hirsch, M.S., and K.K.H. Svoboda 1994 Chondrocytes provide a model
for in situ confocal microscopy and 3D reconstructions. In: Three
Dimensional Microscopy: Image Acquisition and Processing, Proceedings of the Society of Photo-Optical Instrumentation Engineers, C.J. Cogswell and K. Carlsson, eds. 2184:159-170.
Hirsch, M.S., M.A. Moses, and K.K.H. Svoboda 1995 Integrin subunit
antibodies (01,a 2 and a3) inhibit type X collagen secretion and
increase metalloproteinase inhibitor expression in serum free
cartilage organ cultures. Molecular Biology of the Cell 6t391a.
Hynes, R.O. 1992 Integrins: Versatility, modulation and signaling in
cell adhesion. Cell, 69t11-25.
Jirawuthiworavong, G., A. Ashrafzadeh, and K.K. H. Svoboda 1994
Zyxin and vinculin are localized in cell-matrix attachment complexes (CMAX) in embryonic corneal epithelia. Invest. Ophth.
Vis. Sci., 36(4):S566.
Kirk, R.E. 1982 Experimental Design, 2nd ed. BrooksXole, Pacific
Grove, CA.
Kuhn, K., and J. Eble 1994 The structural bases of integrin-ligand
interactions. Trends Cell Biol., 4:256-261.
Linsenmayer, T.F., Q. Chen, E. Gibney, M.K. Gordon, J.K. Marchant,
R. Mayne, and T.M. Schmid 1991 Collagen types IX and X in
developing chick tibiotarsus: Analysis of mRNAs and proteins.
Dev., 111:191-196.
Luna, E.J., and A.L. Hitt 1992 Cytoskeletal-plasma membrane interactions. Science, 258:955-964.
Lu Valle, P., K. Daniels, E.D. Hay, and B.R. Olsen 1992 Type X
collagen is transcriptionally activated and specifically localized
during sternal cartilage maturation. Matrix, 12:404-413.
Marchisio, P.C., 0. Capasso, L. Nitsch, R. Cancedda, and E. Gionti
1984 Cytoskeleton and adhesion patterns of cultured chick embryo chondrocytes during cell spreading and rous sarcoma virus
transformation. Exp. Cell Res., 151:332-343.
McDonald, B.J. and W.A. Thompson 1967 Methods for controlling
Type I errors. Biometricka, 54:487-497.
Mendler, M., S.G. Eich-Bender, L. Vaughan, and K.H. Winterhalter
1989 Cartilage contains mixed fibrils of collagen types 11, IX, and
XI. J. Cell Biol., 108:191-197.
MenkeI, A.R., M. Kroemker, P. Bubeck, M. Ronsiek, G. Nikolai, and
B.M. Jockusch 1994 Characterization of a n F-actin-binding domain in the cytoskeletal protein vinculin. J. Cell Biol., 126:12311240.
Otey, C.A., F.M. Pavalko, and K. Burridge 1990 An interaction between a-actinin and the p l integrin subunit in vitro. J . Cell Biol.,
Pacifici, M., E.B. Golden, S.L. Adams, and I.M. Shapiro 1991 Cell
hypertrophy and type X collagen synthesis in cultured articular
chondrocytes. Exp. Cell Res., 192:266-270.
Pavalko, F.M., C.A. Otey, K.O. Simon, and K. Burridge 1991 Alphaactinin: A direct link between actin and integrins. Biochem. SOC.
Trans., 19:1065-1068.
Pawley, J.B. 1990 Fundamental limits in confocal microscopy. In:
Handbook of Biological Confocal Microscopy. J.B. Pawley, ed.
New York, Plenum Press, pp. 15-26.
Reichenberger, E., T. Aigner, K. von der Mark, H. Stob, and W.
Bertling 1992 In situ hybridization studies on the expression of
type X collagen in fetal human cartilage. Dev. Biol., 148562572.
Sadler, I., A.W. Crawford, J.W. Michelsen, and M.C. Beckerle 1992
Zyxin and cCRP: Two interactive LIM domain proteins associated
with the cytoskeleton. J . Cell Biol., 119:1573-1587.
Sastry, S.K., and A.F. Honvitz 1993 Integrin cytoplasmic domains:
Mediators of cytoskeletal linkages and extra- and intracellular
initiated transmembrane signaling. Curr. Opin. Cell Biol.,
Schaller, M.D., and J.T. Parsons 1994 Focal adhesion kinase and associated proteins. Curr. Opin. Cell Biol., 6:705-710.
Schaller, M.D., and J.T. Parsons 1995 pp12FjFAK-dependenttyrosine
phosphorylation of paxillin creates a high-affinity binding site for
Crk. Mol. Cell Biol., 15(5):2635-2645.
Schmid, T.M., and T.F. Linsenmayer 1985 Immunohistochemical localization of short chain cartilage collagen (Type x) in avian skeletal tissues. J. Cell Biol., 100598-605.
Schmid, T.M., and T.F. Linsenmayer 1990 Immunoelectron microscopy of type X collagen: Supramolecular forms within embryonic
chick cartilage. Dev. Biol., 138.53-62.
Schmid, T.M., D.K. Bonen, L. Luchene, and T.F. Linsenmayer 1991
Late events in chondrocyte differentiation: Hypertrophy, type X
collagen synthesis and matrix calcification. In Vivo, 5.533-540.
Schwartz, M.A. 1992 Transmembrane signaling by integrins. Trends
Cell Biol., 2:304-308.
Staatz, W.D., S.M. Rajpara, E.A. Wayner, W.G. Carter, andS.A. Santoro 1989 The membrane elvcoDrotein Ia-IIa (VLA-2) comulex
mediates the Mg+ -depen&&t a'dhesion of platelets to collagen.
J . Cell Biol., 108:1917-1924.
Staatz, W.D., K.F. Fok, M.M. Zutter, S.P. Adams. B.A. Rodrimez. and
S.A. Santoro 1991 Identification of a tetrapeptide recognition sequence for the azPlintegrin in collagen. J. Biol. Chem. 266:73637367.
Stossel, T.P. 1993 On the crawling of cells. Science, 260:1086-1094.
Terasaki, M. 1989 Fluorescent labeling of endoplasmic reticulum. In:
Fluorescence Microscopy of Living Cells in Culture, Part A. Academic Press, San Diego, pp. 125-135.
Upholt, W.B., B.M. Vertel, and A. Dorfman 1979 Translation and
characterization of messenger RNAs in differentiating chicken
cartilage. Proc. Natl. Acad. Sci. USA, 76(10):4847-4851.
Wilson, T. 1990 The role of the pinhole in confocal imaging systems.
In: Handbook of Biological Confocal Microscopy. J.B. Pawley, ed.
New York, Plenum. Press, pp. 113-126.
Wu, X.Y., K.K.H. Svoboda, and V. Trinkaus-Randall 1995 Distribution of F-actin, vinculin and integrin subunits (alpha6 and beta41
in response to corneal substrata. Exp. Eye Res., 60:445-458.
Zutter, M.M., and S.A. Santoro 1990 Widespread histologic distribution of the azpl integrin cell surface collagen receptor. Am. J .
Path., 137(11:113-120.
Без категории
Размер файла
1 750 Кб
Пожаловаться на содержимое документа