Functional Integrin Subunits Regulating
Cell–Matrix Interactions in the Intervertebral Disc
Christopher L. Gilchrist,1 Jun Chen,1 William J. Richardson,2 Richard F. Loeser,3 Lori A. Setton1,2
1
Department of Biomedical Engineering, Duke University, 136 Hudson Hall, Box 90281, Durham,
North Carolina 27708-0281
2
Division of Orthopaedic Surgery, Department of Surgery, Duke University Medical Center, Durham, North Carolina
3
Section of Molecular Medicine, Wake Forest University School of Medicine, Winston-Salem, North Carolina
Received 10 July 2006; accepted 19 October 2006
Published online 22 February 2007 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/jor.20343
ABSTRACT: Cellular interactions with the extracellular matrix are key factors regulating cell
survival, differentiation, and response to environmental stimuli in cartilagenous tissues. Much is
known about the extracellular matrix proteins in the intervertebral disc (IVD) and their variations
with region, age, or degenerative state of the tissue. In contrast, little is known of the integrin cell
surface receptors that directly bind to and interact with these matrix proteins in the IVD. In almost
all tissues, these integrin-mediated cell–matrix interactions are important for transducing
environmental cues arising from mechanical stimuli, matrix degradation fragments, and cytokines
into intracellular signals. In this study, cells from the nucleus pulposus and anulus fibrosus regions
of porcine IVDs were analyzed via flow cytometry to quantify integrin expression levels upon
isolation and after monolayer culture. Assays of cell attachment to collagens, fibronectin, and
laminin were performed after functional blocking of select integrin subunits to evaluate the role of
specific integrins in cell attachment. In situ distribution and co-localization of integrins and laminin
were also characterized. Results identify integrin receptors critical for IVD cell interactions with
collagens (a1b1) and fibronectin (a5b1). Additionally, dramatic differences in cell–laminin
interactions were observed between cells of the nucleus and anulus regions, including differences
in a6 integrin expression, cell adhesion to laminin, and in situ pericellular environments. These
findings suggest laminin–cell interactions may be important and unique to the nucleus pulposus
region of the IVD. The results of this study provide new information on functional cell–matrix
interactions in tissues of the IVD. ß 2007 Orthopaedic Research Society. Published by Wiley
Periodicals, Inc. J Orthop Res 25:829–840, 2007
Keywords:
integrin; intervertebral disc; collagen; fibronectin; laminin
INTRODUCTION
The intervertebral disc (IVD) is a heterogeneous
tissue with regional variations in composition,
structure, and cellular morphology and phenotypes.1,2 Cells of the IVD are responsible for
maintaining tissue homeostasis, and alterations
in cellular function or viability may play a key role
in the progression of IVD degeneration. Important
regulators of cell function and survival in many
tissues are cellular interactions with their surrounding extracellular matrix (ECM). The ECM
provides physical and biochemical cues that
regulate processes from cell differentiation in
Correspondence to: Lori A. Setton (Telephone: 919-660-5131;
Fax: 919-681-5490; E-mail: [email protected])
ß 2007 Orthopaedic Research Society. Published by Wiley Periodicals,
Inc.
development to cell-mediated repair or breakdown
in mature or aging tissues.3–6 The structure and
composition of the ECM in the IVD has been
relatively well-characterized2,7–10; however, considerably less is known about the specific interactions of IVD cells with matrix components.
Characterization of cell–matrix interactions in
the nondegenerate IVD, and their subsequent
changes in the aged or pathologic IVD, may yield
insight into the aging and degenerative process
and provide new directions for intervention or
repair.
The integrin family of cell surface receptors11
link cells to their extracellular matrix and have
been shown to regulate cellular signals affecting
cell survival, differentiation, proliferation, biosynthetic activity, and responses to environmental
stimuli in many tissues12 including cartilaginous
tissues.13 Integrin receptors function as heterodimers
JOURNAL OF ORTHOPAEDIC RESEARCH JUNE 2007
829
830
GILCHRIST ET AL.
consisting of one alpha and one beta subunit, with
specific heterodimer combinations determining
extracellular matrix-binding specificity. Integrinmatrix binding events may transduce specific
downstream effects that are both matrix and
integrin subunit specific.14–16 In the IVD, several
studies have identified specific integrin receptors
present during IVD development,17–20 including
the fibronectin-binding integrin receptor a5b1 and
the collagen-binding integrins a10b1 and a11b1.
Additionally, the collagen-binding integrin a2b1
has been identified in cultured cells isolated from
the rat nucleus pulposus region of the IVD and
shown to play a role in adhesion of these cells to
Type II collagen.21
In previous work from our laboratory,22 tissue
immunostaining revealed the presence of a set of
alpha and beta integrin subunits in immature
porcine IVD tissues as well as the mature human
IVD. Integrin subunits that classically bind collagens (a1, b1) and fibronectin (a5, av, b3, b1, b5)
were identified in both the anulus fibrosus (AF) and
nucleus pulposus (NP) regions of the porcine IVD,
as well as in the fibrous IVD tissues obtained from
human adults. Several differences between regions
were noted, with findings for higher expression
levels of a6 and b4 integrin subunits in the
immature porcine NP. The a6 integrin subunit (in
combination with b1 or b4) is a primary receptor for
the matrix protein laminin in other cell types, and
the presence of a6 in the NP may indicate an
important phenotypic difference in cellular interactions with laminin between IVD regions or cell
types. The presence of laminin has been documented in the developing rat IVD,10 but there is
currently no information regarding its presence
and distribution in the post-embryonic IVD.
Although our previous tissue immunostaining
work has qualitatively identified the presence of
individual integrin subunits in the IVD, it does not
provide quantitation of integrin expression nor
indicate their potential role in mediating cell–
matrix interactions in the IVD. Understanding the
functional interactions of individual integrins and
extracellular matrix constituents in the IVD is a
first step towards the goal of determining downstream cellular responses of particular cell–matrix
interactions, as well as identifying changes that
may occur with aging or degeneration.
The objective of this study was to determine the
roles of individual integrin subunits in mediating
IVD cell interactions with specific extracellular
matrix proteins. Cell–matrix interactions were
studied in immature porcine IVD tissues that
contain both a gelatinous NP region (which may
JOURNAL OF ORTHOPAEDIC RESEARCH JUNE 2007
contain both ‘‘notochordal-like’’ and ‘‘chondrocytelike’’ cell populations) and a fibrous AF region, in
order to test for differences between distinct IVD
regions. Cells were isolated from porcine IVDs and
integrin subunit expression was quantified via flow
cytometry to investigate integrin expression levels
upon isolation and over the time in culture required
to perform in vitro experiments, in order to confirm
that integrin subunit expression patterns observed
in IVD tissues previously22 were maintained
in vitro. Integrin subunits mediating interactions
with specific matrix proteins were identified via cell
attachment and gel contraction assays, with function-blocking antibodies used to inhibit individual
integrin subunits. Additionally, immunohistochemical staining was used to evaluate the
region-specific variations in laminin protein
expression. The results of this study identify IVD
cell integrin receptors that mediate interactions
with collagens and fibronectin, and elucidate
distinct differences in laminin interactions and
integrin expression between cells from different
IVD regions. These studies lay the foundation for
future work examining integrin- and matrixmediated biological responses of IVD cells and
their roles in IVD tissue homeostasis, degeneration, and pathology.
MATERIALS AND METHODS
Cell Isolation and Culture
Lumbar spines were obtained from pigs immediately
following sacrifice (L1–L5, 9–16-week-old, Duke University Vivarium). Cells were isolated from the NP and
AF regions of IVDs using sequential pronase-collagenase digestion as described previously.23 Isolated cells
were either used immediately for some experiments, or
cultured in subconfluent monolayers (50,000 cells/cm2)
on 0.1% gelatin-coated (Sigma, St. Louis, MO) tissue
culture flasks for 2–6 days in culture media (F-12 media
supplemented with 10% FBS, 10 mM HEPES, 100 U/mL
penicillin, and 100 mg/mL streptomycin).
Flow Cytometric Analysis
AF and NP cells were analyzed via flow cytometry to
quantify expression levels of several integrin subunits.
Integrin subunits selected for analysis were previously
identified in porcine IVD tissues22 as being highly
expressed (a1, a5, b1) or regionally varying (a6). Freshly
isolated (Day 0) and monolayer cultured (Days 3 and 7)
cells (n ¼ 3 separate cell isolations from AF and NP
tissues; each isolation pooled from two spines) were
resuspended in serum-free F12 media, with freshly
isolated cells allowed to recover in suspension for 2 h at
378C. Cells were then incubated for 30 min at 48C with
DOI 10.1002/jor
INTEGRIN-MEDIATED CELL–MATRIX INTERACTIONS IN THE IVD
one of the following integrin antibodies (10 mg/mL): a1
(MAB1973Z, Chemicon, Temecula, CA), a5 (555651, BD
Pharmingen, San Diego, CA), a6 (555734, BD Pharmingen), b1 (4B4, Beckman Coulter, Fullerton, CA), or
appropriate IgG isotype negative controls (mouse IgG1,
CBL600; rat IgG2a, CBL605, Chemicon). All flow
cytometry integrin antibodies were confirmed to crossreact with pig. Cells were washed twice with PBS
and incubated with appropriate secondary antibody
(10 mg/mL Alexa 488, Molecular Probes, Eugene, OR)
for 30 min at 48C. Cells were analyzed on a FACScan
flow cytometer (Becton Dickinson, Franklin Lakes,
New Jersey) to measure the (geometric) mean fluorescence intensity (MFI) and percentage of cells with
positive surface proteins (%). IgG control values were
recorded and subtracted from experimental values of
fluorescence.
Integrin Blocking Assays
Ninety-six–well culture plates (Corning Costar, Acton,
Massachusetts) were coated with one of the following matrix proteins (Sigma) diluted in PBS: bovine
Type I collagen (40 mg/mL), chicken Type II collagen
(40 mg/mL), human plasma fibronectin (1 mg/mL), by
overnight incubation at 48C; or mouse laminin-1 (mouse
EHS laminin, 20 mg/mL, Chemicon) incubated 2 h at
378C. Coated wells were then blocked with 3.75% bovine
serum albumin (BSA, Gibco, Gaithersburg, MD) for 3 h
at 378C. Wells without matrix protein (BSA blocked
only) were used as a substrate negative control to
identify nonspecific cell attachment.
Cells in monolayer culture (n ¼ 3 separate isolation
pools each for NP and AF cells) were detached from the
culture surface using 0.025% trypsin/EDTA (Cambrex,
East Rutherford, NJ) and immediately washed in
soybean trypsin inhibitor (Sigma). Cells were resuspended in serum-free F12 media containing one of the
following integrin function-blocking antibodies specific
for a particular integrin subunit or dimer (from Chemicon, 10 mg/mL, unless otherwise noted): a1 (clone FB12),
a2b1 (BHA2.1), a5 (P1D6), a6 (NKI-GoH3, 40 mg/mL), av
(AV1), avb3 (LM609), avb5 (P1F6), b1 (4B4, Beckman
Coulter). Cross-reactivity with pig was confirmed by the
manufacturer or via flow cytometry and an ability to
functionally inhibit attachment; confirmation was
obtained for all antibodies except those against the a5
subunit, for which no cross-reacting antibody could be
identified. Control cells were resuspended without
blocking antibody or with nonblocking mouse IgG antibody. Additionally, small peptides containing the RGD
sequence (1 mM GRGDSP, Anaspec, San Jose, CA) or
control RGE sequence (GRGESP, Anaspec) were used to
investigate whether cell attachment was dependent on
this common integrin-binding peptide sequence contained in matrix proteins such as fibronectin.
Cells were preincubated in suspension with antibodies or peptides for 30 min, seeded on matrix-coated
culture plates, and incubated at 378C for 1 h to allow for
attachment (20,000 cells per well, 4 replicates (wells)
DOI 10.1002/jor
831
per blocking condition). Wells were rinsed twice with
serum-free media to remove unattached cells. Remaining
cells were fixed with 4% formaldehyde, cell nuclei were
labeled (Hoechst 33342, Sigma), and cells in the central
region of each well were imaged via fluorescence
microscopy [Zeiss Axiovert S100 (Carl Zeiss, Thornwood,
NY) equipped with Nikon, Melville, New York CoolPix
990 digital camera]. Total number of cells per image were
counted and normalized to unblocked controls for each
substrate.
Cell Size Measurements
To further elucidate findings on laminin substrates, NP
cell attachment to laminin (in the presence/absence of
integrin blocking antibodies) was evaluated by light
microscopy to quantify the distribution of attached cell
diameters. DIC images of attached cells in each well
(1 image field/well, for all wells of control, a6-, and
b1-inhibited cells attaching to laminin) were recorded
and cell diameters were measured using ImageJ software (NIH, Bethesda, Maryland). Attached NP cells
were spherical and had not spread out on substrates at
time of fixation.
Collagen Gel Contraction
Contraction assays of a Type I collagen gel24 were
utilized to determine the functionality of specific
integrin subunits in mediating interactions with fibrillar collagens, as cells may utilize different integrin
receptors in interactions with fibrillar (vs. monomeric)
collagens. Ninety-six–well plates were coated with
3.75% BSA at 48C overnight and washed twice with
PBS. Porcine AF cells were trypsinized from plates as
described above and resuspended in culture media with
or without integrin function-blocking antibodies (a1, b1)
and incubated at 378C for 30 min. The cell suspension
was combined with collagen solution (Cultrex bovine
Type I collagen solution, Trevigen, Gaithersburg, MD;
HEPES, and NaOH) to yield a final concentration of
300,000 cells/mL, 1.5 mg/mL collagen, and 10 mg/mL of
blocking antibody or IgG control. One hundred microliters of cell/collagen solution was dispensed per well
(3–4 wells per condition) and the plate was immediately
incubated for 45 min at 378C to allow for collagen
gelation. After collagen gels had formed, 100 ml per well
of culture media was injected to release gels from the
plate, resulting in free-floating gels. The plate was then
incubated at 378C and gel diameters were measured via
inverted microscope at 0 and 18 h.
Immunohistochemical Detection of
Laminins and Integrins in IVD Tissues
AF and NP tissues from immature porcine spines (L1–
L5, n ¼ 3 spines) were dissected, separated, and flashfrozen immediately in liquid nitrogen for cryosectioning
(7 mm-thick sections). Sections to be labeled for laminin
and integrins were either left unfixed (laminin) or fixed
JOURNAL OF ORTHOPAEDIC RESEARCH JUNE 2007
832
GILCHRIST ET AL.
in acetone at 208C for 10 min (co-labeled for integrins
and laminin). Sections were blocked for 45 min in
blocking solution (3.75% BSA (Gibco)/5% goat serum
(Zymed, Carlsbad, CA)) and those to be labeled for
laminin were incubated for 60 min with a mouse
antibody (clone D18, DHSB, no dilution) specific for
the g1-chain of laminin. Sections co-labeled for integrins
were incubated for 60 min with either rat anti-a6 (NKIGoH3, 1:50) or rabbit anti-b4 (ab1922, Chemicon, 1:200)
diluted in blocking solution, followed by incubation with
anti-laminin antibody. Sections were washed twice in
PBS, incubated with appropriate secondary antibodies
(AlexaFluor 488 or 633 secondary antibodies, Molecular
Probes) for 45 min in blocking buffer (1:200 dilution).
Cell nuclei in all sections were counterstained with
propidium iodide (1 mg/mL, Sigma) and mounted.
Control sections were incubated with appropriate IgG
controls (10 mg/mL mouse IgG1 or rat IgG2a, Chemicon)
instead of primary antibody. Specimens were imaged
using a Zeiss LSM 510 confocal laser scanning microscope with 20 (NA 0.5) and 63 (water immersion, NA
1.2) objectives with consistent settings.
Statistical Analyses
Flow cytometry results were evaluated using a threefactor ANOVA (integrin subunit, tissue region, culture
time). For integrin blocking experiments, differences in
cell attachment and collagen gel contraction between
blocking conditions were detected via one-factor (integrin blocking condition) ANOVA with repeated measures.
Fisher’s post-hoc test was used for all ANOVAs, with a
significance level of p 0.05. Differences in cell size
distributions between integrin blocking conditions for
NP cells attaching to laminin substrates were detected
using a chi-squared analysis (p 0.001), with cells
grouped into bins based on cell diameter (bin size ¼
5 mm).
RESULTS
Integrin Expression in Isolated IVD Cells
Immediately upon isolation from immature porcine IVD tissues (Day 0), expression of all integrin
subunits examined was detected [mean fluorescence intensities (MFI) and % cells positive >
negative controls] for both AF and NP cells
(Fig. 1), with the exception of the a6 subunit,
which was not detectable in AF cells. Integrin
expression (MFI, % positive) increased significantly (ANOVA, main effect of culture time,
p < 0.05) after 3 and 7 days in monolayer culture
across all integrins examined and both cell types
(Fig. 1). For the a1, a5, and b1 subunits, MFI levels
increased two- to eight-fold and the percentage of
positive cells increased to between 70%–95% after
3 days of culture for cells from both regions.
Differences between cell type (IVD region) were
JOURNAL OF ORTHOPAEDIC RESEARCH JUNE 2007
detected for two integrin subunits, regardless of
culture time (ANOVA, cell type by integrin
subunit interaction). For the a5 subunit, cells from
the AF region were found to exhibit significantly
higher MFI levels than NP cells (Fig. 1A, p < 0.01),
although no differences were detected for the
percentage of cells that were positive for this
subunit (Fig. 1B). For the a6 subunit, a prominent
difference in integrin expression between cell type
was detected, with NP cells expressing significantly higher MFI levels and percent cells positive
(Fig. 1A and B, both p < 0.01) as compared to AF
cells. Upon isolation (Day 0), 29% of NP cells had
detectable levels of a6, while just 3% of AF cells
were positive. After 3 and 7 days in culture, a6
expression for NP cells was higher than other
alpha subunits (89%–97% of cells positive,
MFI ¼ 155–170) while expression in AF cells was
increased but remained comparatively low (%
positive ¼ 45%–57%, MFI ¼ 35–43). These results
suggest that regional differences in a6 integrin
expression identified in situ22 are maintained
upon cell isolation, and over 7 days of monolayer
culture.
IVD Cell Attachment to Collagens
and Fibronectin
Cells from both regions of the porcine IVD were
found to attach readily to types I and II collagen
substrates, as well as fibronectin substrates.
Attachment to surfaces without protein (wells
blocked with BSA, Fig. 2) was negligible (always
<10% of control), indicating cell attachment was
due to cellular interactions with the matrix
protein coated on the well surface. On Type I
collagen substrates, blocking the a1 integrin
subunit partially inhibited cell attachment (49%
inhibition for AF, 64% NP), whereas blocking the
b1 subunit resulted in almost complete inhibition
of attachment (AF: 97%, NP: 82%), as shown in
Fig. 2A. These results suggest that the a1b1
integrin partially mediates attachment of IVD
cells to Type I collagen, but other b1-containing
integrins are also involved. However, blocking
another known collagen-binding heterodimer,
a2b1, did not inhibit attachment of AF cells and
showed slight effects (19% inhibition) for NP cells.
Incubation in the presence of RGD peptide or
blocking the avb3 heterodimer showed no inhibitory effect of binding to Type I collagen, indicating
a secondary interaction via cell-secreted fibronectin was not involved in attachment. For IVD cells
seeded on Type II collagen substrates, cell attachment was found to be mediated entirely by the a1
DOI 10.1002/jor
INTEGRIN-MEDIATED CELL–MATRIX INTERACTIONS IN THE IVD
833
Figure 1. Integrin expression for IVD cells isolated from AF and NP tissue regions and
analyzed via flow cytometry. Cells were analyzed for mean fluorescence intensity (A) and
percent cells positive (B) immediately following isolation (Day 0) and after monolayer
culture for 3 and 7 days. Data shown (mean SEM) represent results from three separate
cell isolations (n ¼ 3). Differences were analyzed via three-way ANOVA (tissue, culture
time) for each integrin subunit. Significant main effects between tissue regions
(*, p < 0.01) and integrin subunit (**, p < 0.05) are shown. Additionally, a significant
effect of culture time was detected (p < 0.05) for all integrins for both MFI and % positive.
(AF: 97%, NP: 96%) and b1 (AF: 97%, NP: 83%)
integrin subunits, as shown in Fig. 2B, while other
blocking conditions showed no inhibitory effects.
Enhancement of NP cell attachment to collagens
was detected in the presence of both RGD and RGE
peptides (p > 0.05, for all except RGE on Type II
collagen). Although the specific mechanism
is unclear, this may indicate an effect of high
concentrations of free peptide in the cell solution,
since both functional (RGD) and nonfunctional
(RGE) peptides resulted in similar enhancement.
It is not known why NP cells were uniquely
affected by these peptides, although NP cells
may express receptors involved in cell–cell interactions which could be inhibited by these concenDOI 10.1002/jor
trations of free peptide, allowing more cells to
interact with the culture surface.
On fibronectin substrates (Fig. 2C), functionblocking antibodies to the known fibronectin-binding subunit av and heterodimers avb3 and avb5
showed no inhibitory effect for either cell type.
Blocking with the b1 antibody or with RGD peptide
completely inhibited porcine AF and NP cell
attachment to fibronectin, while blocking the a1
subunit showed no effects. We were unable to
identify a function-blocking antibody for the a5
subunit that cross-reacted with porcine cells. This
corroborates the findings of another study where
another anti-human a5 function-blocking antibody
(clone SAM-1) was found to be ineffective in
JOURNAL OF ORTHOPAEDIC RESEARCH JUNE 2007
834
GILCHRIST ET AL.
Figure 2. Effects of integrin function-blocking antibodies on IVD cell attachment
to substrates coated with (A) Type I collagen, (B) Type II collagen, (C) fibronectin, and
(D) laminin. Attachment numbers (mean SEM) were normalized to uninhibited control
cells for each cell type. Nonspecific cell attachment to wells without matrix protein was
low (data not shown, always <10% of control for all cells and substrates). n ¼ 3 cell
isolations for each substrate, 4 replicates per condition for each experiment. *ANOVA,
p < 0.05 from control and IgG; **, p < 0.05 between conditions [shown for (D) only].
blocking porcine monocyte attachment to human
fibronectin.25 Since the a5b1 heterodimer is a
common fibronectin receptor in numerous cell
types and the primary fibronectin receptor involving the b1 subunit, and since blocking antibodies
to the av, avb3, and avb5 integrins were unable to
inhibit attachment, it is likely that the a5b1
integrin heterodimer is the critical receptor in
IVD cell attachment to fibronectin. When porcine
AF cells were allowed to spread on fibronectin
substrates (overnight in low-serum media) and
immunostained with a5 polyclonal (ab1928, Chemicon) and b1 antibodies, staining was intense and
localized to focal contacts as shown in Figure 3.
Together these results strongly suggest that a5b1
is a functional receptor involved in attachment and
spreading on fibronectin for IVD cells.
Cell-Mediated Collagen Gel Contraction
AF cells seeded in Type I fibrillar collagen gels
were found to actively contract the free-floating
JOURNAL OF ORTHOPAEDIC RESEARCH JUNE 2007
gels, with reductions of over 30% in gel area for
control gels after 18 h (no antibody: 33% contraction; nonblocking IgG antibody: 31% contraction),
as shown in Figure 4. Preincubating cells with
function-blocking antibodies for a1 and b1 integrin
subunits inhibited gel contraction almost completely (a1: 7% contraction; b1: 4%), indicating a1b1
is critical for forming mechanically functional
integrin interactions with fibrillar Type I collagen
in AF cells.
IVD Cell Attachment to Laminin
NP cells were found to attach readily and in high
numbers on laminin-1 coated substrates. In contrast, AF cell attachment to laminin was very
minimal (<15% of NP control values, compared to
10% attachment on BSA-coated surfaces). NP cell
attachment to laminin was partially inhibited by
blocking the a6 (64% inhibition) and b1 (46%)
integrin subunits (Fig. 2D), with a significant
difference detected between these two integrin
DOI 10.1002/jor
INTEGRIN-MEDIATED CELL–MATRIX INTERACTIONS IN THE IVD
835
Figure 3. Immunofluorescent staining of AF cells attached to fibronectin-coated
substrates and stained for (A) a5 and (B) b1 integrin subunits. Arrows indicate brightly
staining focal adhesion sites. Scale bar ¼ 20 mm.
function-blocking conditions. Because of very low
attachment numbers for control AF cells, differences in AF cell attachment due to integrin blocking were not evaluated.
While performing laminin attachment experiments, it was observed that the size distribution of
NP cells attaching to laminin substrates appeared
to vary depending on blocking condition. Since the
NP region of the immature IVD has been reported
to contain a phenotypically mixed cell population
that may include both large, vacuolated ‘‘notochordal-like’’ cells as well as smaller chondrocyte-like
cells,26 we quantified the size distribution of
attached cells for different blocking conditions.
Cell diameters for NP cells attached to laminin
were found to range in diameter from 10–50 mm,
with peak cell numbers between 10–20 mm and a
second peak occurring between 25–30 mm (Fig. 5).
Blocking the a6 subunit prevented attachment of
larger cells to laminin, with the majority of
attached NP cells being smaller than 20 mm.
Blocking the b1 subunit did not completely eliminate larger NP cells from attaching to laminin.
Identification and Arrangement of
Laminin in IVD Tissues
The findings of a strong role for laminin in promoting NP but not AF cell attachment motivated
Figure 4. AF cell-mediated collagen gel contraction (mean SEM) and inhibition by
integrin function-blocking antibodies. Cells suspended in collagen gels and allowed 18 h
to contract. n ¼ 3 separate cell isolations, 3–4 replicates per condition. *ANOVA,
significant difference from control and IgG, p < 0.0001.
DOI 10.1002/jor
JOURNAL OF ORTHOPAEDIC RESEARCH JUNE 2007
836
GILCHRIST ET AL.
Figure 5. Distribution of cell size for NP cells attached to laminin-coated substrates
for unblocked control cells and cells preincubated with function-blocking antibodies to
either integrin a6 or b1. Asterisks denote cell diameter groupings where blocking
condition deviates significantly (p < 0.001) from control.
additional studies to localize the distribution of
laminin in the extracellular matrix of the tissues
studied. An antibody recognizing the laminin g1chain (present in 10 of the 16 known laminin
isoforms27) was used to identify laminin in porcine
IVD tissues. Tissue from the NP region stained
highly positive for laminin, with tissue often
separating into nests of notochordal-like cells.
This staining was found to co-localize with the a6
(Fig. 6A) integrin subunit for NP tissue samples.
The innermost region of the porcine AF showed
staining for laminin that was distinctly pericellular, with a6 also present in these cells (Fig. 6B).
Faint streaks of laminin staining were found in the
mid- and outer-AF, as shown in Figure 6C, with
almost all cells negative for the a6 subunit.
Additionally, staining for the b4 integrin subunit
revealed high expression that co-localized with
laminin in the NP (data not shown), but b4 was not
highly expressed in either inner or outer AF
regions.
DISCUSSION
The results of this study provide new information
revealing a set of functional integrin cell-surface
JOURNAL OF ORTHOPAEDIC RESEARCH JUNE 2007
receptors that mediate interactions with extracellular matrix proteins in the IVD. A major
finding of this study was that of NP, but not AF,
cellular interactions with laminin, with evidence
that these interactions may even differ for cell
subpopulations within the NP. Cells from the NP
region were found to attach readily to laminin
substrates, mediated in part by a6 and b1 integrin
subunits. NP cells expressed much higher levels of
the a6 subunit both in situ, as well as in isolated
cells analyzed via flow cytometry. In contrast, AF
cells did not readily attach to laminin and only
rounded cells located in the innermost regions of
the AF showed positive staining for laminin and
the a6 subunit. Cells of the NP that remained
attached to laminin after blocking the function
of a6 were found to be primarily smaller cells,
suggesting that a population of large NP cells
mediates interactions with laminin solely via the
a6 subunit. Blocking the b1 subunit did not
prevent attachment of all larger cells, however,
indicating that these cells may have other means
of attachment to laminin (e.g., a6b4). These
findings may be relevant to observations of
notochordal cell remnants in the immature NP
tissue, that are documented as being larger and
DOI 10.1002/jor
INTEGRIN-MEDIATED CELL–MATRIX INTERACTIONS IN THE IVD
837
Figure 6. Immunohistochemical staining detecting the presence of laminin (green)
and a6 integrin (red) in porcine IVD tissue, with counterstaining for cell nuclei (blue). NP
tissue (A) often separated into nests of cells staining highly positive for laminin. Regions
of bright laminin staining typically coincided with bright a6 staining (arrows). Distinct
pericellular staining for laminin was also found in inner AF region (B), with cells also
staining positively for a6. The outer AF region (C) stained faintly for laminin, with very
little a6 staining. Scale bars ¼ 50 mm.
present in cell clusters.26,28 The results presented
here suggest that NP cell–laminin interactions
mediated by the a6 integrin may be a unique
feature of this immature NP region that contains a
high proportion of notochordal-like cells.
Integrins involved in IVD cell attachment to
collagen substrates appeared to be similar for cells
from both AF and NP regions of the IVD, with a
potential role for additional as-yet unidentified
receptors. For cells from both regions, attachment
to Type II collagen was found to be mediated
DOI 10.1002/jor
primarily by the a1b1 receptor. These findings
differ somewhat from a recent study21 where rat
NP cell attachment to Type II collagen was
inhibited by approximately 40% with a blocking
antibody to the a2 subunit. In the present study, we
detected no significant attachment inhibition on
Type II collagen with a blocking antibody to a2b1,
corresponding with our previous immunostaining
of porcine IVD tissues22 where significant levels of
a2 were not detected in situ. In contrast to Type II
collagen, IVD cell attachment to Type I collagen
JOURNAL OF ORTHOPAEDIC RESEARCH JUNE 2007
838
GILCHRIST ET AL.
was only partially (50%) inhibited when blocking
the a1 integrin subunit, although blocking b1
entirely inhibited attachment. In addition, blocking a2b1 only slightly (19%) reduced adhesion of NP
cells and had no detectable affect on AF cells. These
findings point towards roles for other collagenbinding integrins in mediating the IVD cell attachment to Type I collagen. Other known collagenbinding integrins are the less studied a10b129 and
a11b130 integrins, which have been localized to
cartilaginous tissues including findings for the a10
subunit in the inner AF of the mouse19 and a11
subunit in the outer AF of human and mouse
embryos.20 One possible explanation for only
partial inhibition on Type I collagen is that both
AF and NP tissues contain subpopulations (e.g.,
chondrocyte-like inner AF cells) that utilize a10 or
a11 subunits. Tools to assay a10 and a11 protein
expression and functional interactions with collagens remain limited and were not available for
this study.
In contrast to cell attachment experiments,
findings from collagen gel contraction experiments
provided evidence of a dominant functional role for
the a1 and b1 integrin subunits in mediating
mechanical interactions with fibrillar Type I
collagen. Treatment with blocking antibodies to
either subunit inhibited AF cell-mediated gel
contraction almost completely, suggesting that
even though other b1-containing integrins may be
involved in attachment to monomeric Type I
collagen, a1b1 is necessary to promote fibrillar
collagen gel contraction. The a1b1 integrin has
previously been shown to have a higher affinity for
monomeric and beaded-filament collagens (e.g.,
Type VI collagen) than for Type I fibrils,31,32 but
also to participate in Type I collagen fiber contraction for some mesenchymal cell types.33,34 Thus,
our own findings, and those of prior studies,
suggest that the role of the a1b1 integrin in fibrillar
collagen interactions may be cell-type or context
specific. These fibrillar interactions may be important in IVD cell-directed remodeling or repair of the
collagen matrix.
Cell assays on fibronectin substrates suggest
that the a5b1 integrin mediates IVD cell attachment for both NP and AF cells in porcine tissues.
These results are similar to those for articular
chondrocytes, where blocking a5b1 also prevents
cell attachment to fibronectin.35 In chondrocytes,
the a5b1 integrin also binds proteolytic fragments
of fibronectin, which have been shown to initiate a
distinct signaling cascade with downstream effects
that may lead to matrix degradation or remodeling.36,37 Recent studies present evidence that a
JOURNAL OF ORTHOPAEDIC RESEARCH JUNE 2007
similar pathway may exist in disc,38,39 and thus the
a5b1 integrin may be a key receptor for cellmediated responses to matrix degradation in the
IVD.
The experiments performed to assess functional
roles of integrins in this study required that IVD
cells be isolated from their surrounding tissue and
cultured for a short time in vitro. Removing cells
from their native environment introduces the
potential for artifact, including alterations to
integrin receptor expression as a result of enzymatic isolation or culture environment (e.g., serum
growth factors, altered oxygen tension, rigid twodimensional culture surface). We attempted to
limit time in culture, with cells cultured for only
2–6 days after isolation. Additionally, flow cytometry experiments were undertaken to understand how integrin expression varied upon
isolation and over these times in culture, and to
compare with expression patterns noted in disc
tissues via immunohistochemistry previously.22
We found that integrin expression upon isolation
(Day 0) was low but increased markedly after a
short time in culture, which may represent a need
for cells to ‘‘recover’’ from the cell isolation
procedure, or could also reflect changes induced
by the cell culture environment. However, flow
cytometry experiments indicated that notable
tissue-level expression patterns were retained for
cells in vitro (e.g., differential expression of a6
between AF and NP cells, high expression levels of
a1, a5, b1 subunits in both cell types), providing
confirmation that phenotypic expression patterns
identified with tissue immunostaining persisted
upon cell isolation and culture.
The present study identifies functional cell–
matrix interactions via integrins for cells of the
IVD, with the intriguing finding that interactions
with laminin may be uniquely important for cells of
the immature NP. This study provides a catalog of
integrin receptors important for interactions with
collagens, fibronectin, and laminin for IVD cells
from different tissue regions; however, the biological significance of these receptors in the IVD
remains to be investigated. Additionally, further
work is necessary to characterize these interactions in human IVD tissues, where ‘‘notochordallike’’ cells of the NP region disappear within the
first decade of life. Identifying the downstream
cellular responses regulated by these cell–matrix
interactions, as well as changes in these interactions that occur during aging and disc degeneration
of the human IVD, may yield important insights
into the mechanisms of IVD pathologies and
suggest novel treatment avenues.
DOI 10.1002/jor
INTEGRIN-MEDIATED CELL–MATRIX INTERACTIONS IN THE IVD
ACKNOWLEDGMENTS
This research was supported in part by grants from
the NIH (R01AR47442, EB002263). The scientific contributions of Liufang Jing, Peter Truskey, and Steve
Johnson are greatly appreciated.
REFERENCES
1. Urban JP. 2002. The role of the physicochemical environment in determining disc cell behaviour. Biochem Soc
Trans 30:858–864.
2. Roughley PJ. 2004. Biology of intervertebral disc aging
and degeneration: involvement of the extracellular matrix.
Spine 29:2691–2699.
3. Adams JC, Watt FM. 1993. Regulation of development and
differentiation by the extracellular matrix. Development
117:1183–1198.
4. DeLise AM, Fischer L, Tuan RS. 2000. Cellular interactions and signaling in cartilage development. Osteoarthritis Cartilage 8:309–334.
5. Labat-Robert J. 2004. Cell-matrix interactions in aging:
role of receptors and matricryptins. Ageing Res Rev 3:233–
247.
6. Oegema TR Jr, Johnson SL, Aguiar DJ, et al. 2000.
Fibronectin and its fragments increase with degeneration
in the human intervertebral disc. Spine 25:2742–2747.
7. Eyre DR, Muir H. 1976. Types I and II collagens in
intervertebral disc. Interchanging radial distributions in
annulus fibrosus. Biochem J 157:267–270.
8. Roberts S, Menage J, Duance V, et al. 1991. 1991 Volvo
Award in basic sciences. Collagen types around the cells of
the intervertebral disc and cartilage end plate: an
immunolocalization study. Spine 16:1030–1038.
9. Yu J. 2002. Elastic tissues of the intervertebral disc.
Biochem Soc Trans 30:848–852.
10. Hayes AJ, Benjamin M, Ralphs JR. 2001. Extracellular
matrix in development of the intervertebral disc. Matrix
Biol 20:107–121.
11. Hynes RO. 1992. Integrins: versatility, modulation, and
signaling in cell adhesion. Cell 69:11–25.
12. Danen EH, Sonnenberg A. 2003. Integrins in regulation
of tissue development and function. J Pathol 200:471–
480.
13. Loeser RF. 2000. Chondrocyte integrin expression and
function. Biorheology 37:109–116.
14. Giancotti FG, Ruoslahti E. 1999. Integrin signaling.
Science 285:1028–1032.
15. White DJ, Puranen S, Johnson MS, et al. 2004. The
collagen receptor subfamily of the integrins. Int J Biochem
Cell Biol 36:1405–1410.
16. Belkin AM, Stepp MA. 2000. Integrins as receptors for
laminins. Microsc Res Tech 51:280–301.
17. Hayes AJ, Benjamin M, Ralphs JR. 1999. Role of actin
stress fibres in the development of the intervertebral disc:
cytoskeletal control of extracellular matrix assembly. Dev
Dyn 215:179–189.
18. Zhu Y, McAlinden A, Sandell LJ. 2001. Type IIA procollagen in development of the human intervertebral disc:
regulated expression of the NH(2)-propeptide by enzymic
processing reveals a unique developmental pathway. Dev
Dyn 220:350–362.
DOI 10.1002/jor
839
19. Camper L, Holmvall K, Wangnerud C, et al. 2001.
Distribution of the collagen-binding integrin alpha10beta1
during mouse development. Cell Tissue Res 306:107–
116.
20. Tiger CF, Fougerousse F, Grundstrom G, et al. 2001.
Alpha11beta1 integrin is a receptor for interstitial collagens involved in cell migration and collagen reorganization on mesenchymal nonmuscle cells. Dev Biol 237:
116–129.
21. Risbud MV, Guttapalli A, Albert TJ, et al. 2005. Hypoxia
activates MAPK activity in rat nucleus pulposus cells:
regulation of integrin expression and cell survival. Spine
30:2503–2509.
22. Nettles DL, Richardson WJ, Setton LA. 2004. Integrin
expression in cells of the intervertebral disc. J Anat 204:
515–520.
23. Baer AE, Wang JY, Kraus VB, et al. 2001. Collagen gene
expression and mechanical properties of intervertebral
disc cell-alginate cultures. J Orthop Res 19:2–10.
24. Gullberg D, Tingstrom A, Thuresson AC, et al. 1990. Beta 1
integrin-mediated collagen gel contraction is stimulated by
PDGF. Exp Cell Res 186:264–272.
25. Theodore PR, Simon AR, Warrens AN, et al. 2002. Porcine
mononuclear cells adhere to human fibronectin independently of very late antigen-5: implications for donorspecific tolerance induction in xenotransplantation. Xenotransplantation 9:277–289.
26. Trout JJ, Buckwalter JA, Moore KC, et al. 1982.
Ultrastructure of the human intervertebral disc. I.
Changes in notochordal cells with age. Tissue Cell 14:
359–369.
27. Aumailley M, Bruckner-Tuderman L, Carter WG, et al.
2005. A simplified laminin nomenclature. Matrix Biol 24:
326–332.
28. Meachim G, Cornah MS. 1970. Fine structure of juvenile
human nucleus pulposus. J Anat 107:337–350.
29. Camper L, Hellman U, Lundgren-Akerlund E. 1998.
Isolation, cloning, and sequence analysis of the integrin
subunit alpha10, a beta1-associated collagen binding
integrin expressed on chondrocytes. J Biol Chem 273:
20383–20389.
30. Velling T, Kusche-Gullberg M, Sejersen T, et al. 1999.
cDNA cloning and chromosomal localization of human
alpha(11) integrin. A collagen-binding, I domain-containing, beta(1)-associated integrin alpha-chain present in
muscle tissues. J Biol Chem 274:25735–25742.
31. Jokinen J, Dadu E, Nykvist P, et al. 2004. Integrinmediated cell adhesion to type I collagen fibrils. J Biol
Chem 279:31956–31963.
32. Tulla M, Pentikainen OT, Viitasalo T, et al. 2001. Selective
binding of collagen subtypes by integrin alpha 1I, alpha 2I,
and alpha 10I domains. J Biol Chem 276:48206–48212.
33. Racine-Samson L, Rockey DC, Bissell DM. 1997. The role
of alpha1beta1 integrin in wound contraction. A quantitative analysis of liver myofibroblasts in vivo and in primary
culture. J Biol Chem 272:30911–30917.
34. Gotwals PJ, Chi-Rosso G, Lindner V, et al. 1996. The
alpha1beta1 integrin is expressed during neointima
formation in rat arteries and mediates collagen matrix
reorganization. J Clin Invest 97:2469–2477.
35. Loeser RF, Carlson CS, McGee MP. 1995. Expression of
beta 1 integrins by cultured articular chondrocytes and in
osteoarthritic cartilage. Exp Cell Res 217:248–257.
36. Forsyth CB, Pulai J, Loeser RF. 2002. Fibronectin
fragments and blocking antibodies to alpha2beta1 and
JOURNAL OF ORTHOPAEDIC RESEARCH JUNE 2007
840
GILCHRIST ET AL.
alpha5beta1 integrins stimulate mitogen-activated protein
kinase signaling and increase collagenase 3 (matrix
metalloproteinase 13) production by human articular
chondrocytes. Arthritis Rheum 46:2368–2376.
37. Homandberg GA. 1999. Potential regulation of cartilage
metabolism in osteoarthritis by fibronectin fragments.
Front Biosci 4:D713–D730.
JOURNAL OF ORTHOPAEDIC RESEARCH JUNE 2007
38. Aota Y, An HS, Homandberg G, et al. 2005. Differential
effects of fibronectin fragment on proteoglycan metabolism
by intervertebral disc cells: a comparison with articular
chondrocytes. Spine 30:722–728.
39. Anderson DG, Li X, Balian G. 2005. A fibronectin fragment
alters the metabolism by rabbit intervertebral disc cells
in vitro. Spine 30:1242–1246.
DOI 10.1002/jor