Why cellular stress suppresses adipogenesis in skeletal tissue, but is ineffective in
adipose tissue: Control of mesenchymal cell differentiation via integrin binding
sites in extracellular matrices
Vladimir Volloch⁎, Bjorn R. Olsen⁎
Department of Developmental Biology, Harvard School of Dental Medicine, Boston, MA, USA
a b s t r a c ta r t i c l ei n f o
Received 18 December 2012
Received in revised form 13 May 2013
Accepted 14 May 2013
Integrin binding sites
This Perspective addresses one of the major puzzles of adipogenesis in adipose tissue, namely its resistance to
cellular stress. It introduces a concept of “density” of integrin binding sites in extracellular matrix, proposes a
cellular signaling explanation for the observed effects of matrix elasticity and of cell shape on mesenchymal
stem cell differentiation, and discusses how specialized integrin binding sites in collagen IV-containing ma-
trices guard two pivotal physiological and evolutionary processes: stress-resistant adipogenesis in adipose
tissues and preservation of pluripotency of mesenchymal stem-like cells in their storage niches. Finally, it
proposes strategies to suppress adipogenesis in adipose tissues.
© 2013 Elsevier B.V. All rights reserved.
The “density” of integrin binding sites in the ECM and integrin-dependent mechanical stress responses . . . . . . . . . . . . . . . . . . .
Upregulation of Hsp90 activates Raf-1 and enables Raf-dependent ERK activation in mesenchymal cells . . . . . . . . . . . . . . . . . . .
In cultured mesenchymal cells, adipogenic differentiation is controlled by p38 kinase, β-catenin, and ERK . . . . . . . . . . . . . . . . . .
Adipogenic differentiation of mesenchymal cells on collagen IV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Stress-resistance of adipogenic differentiation of mesenchymal cells on DC IV is due to the effect of specialized integrin binding sites . . . . .
Specialized α1β1 integrin binding sites safeguard stress-uninterrupted adipogenesis in adipose tissues and pluripotency of stem cells in their
storage niches. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Strategies to suppress adipogenesis in adipose tissues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365
In light of a pandemic increase in the incidence of obesity, there is an
urgent need to better understand and potentially control adipogenic dif-
ferentiation. Among the central issues that need to be elucidated is the
puzzling tissue-specific resistance of adipogenesis to cellular stress. In
musculoskeletal tissues adipogenesis can be relatively easily controlled
by physical exercise, i.e. mechanical cellular stress, which is highly effec-
esis and myogenesis (Kadi, 2008; Havekes and Sauerwein, 2010; Bareja
and Billin, 2013). In contrast, adipogenesis in adipose tissue is resistant
to cellular stress and physical exercise has little effect (Askew et al.,
1975; Simonsen et al., 2004; Akpan et al., 2009) (Fig. 1). In an attempt
to shed light on these tissue-specific differences in cellular stress
responses, we consider here the behavior of human multipotent mesen-
chymal cells on two types of extracellular matrix, collagen I and collagen
IV, both major matrix components in skeletal and adipose tissues respec-
tively. As described below, such a comparison suggests that one of the
collagen IV-specific integrin binding sites plays a central role not only in
stress-resistance of adipogenesis in adipose tissue but also in preserving
the multipotency of stem and progenitor cells in their storage niches. It
also suggests strategies to suppress adipogenesis in adipose tissue.
Differentiation of multipotent mesenchymal cells into various line-
ages is determined by cytokine/growth factor stimuli and mechanical
Matrix Biology 32 (2013) 365–371
⁎ Corresponding authors.
E-mail addresses: email@example.com (V. Volloch),
firstname.lastname@example.org (B.R. Olsen).
0945-053X/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/locate/matbio
(ECM) components or generated within their cytoskeleton (Matthews et
it determines the differentiation outcome (Mauney and Volloch, 2009a,
2009b, 2010a, 2010b). The multiple components of the ECM allow for a
spectrum of cellular interactions, but many important cellular responses
to mechanical stresses are integrin-dependent. In the complex environ-
ment in vivo these responses can obviously be modulated in several
tribute to mechanisms that regulatedifferentiationof mesenchymalcells,
it is essential to isolate and analyze the effects of their engagement with
binding sites in the extracellular matrix in the absence of other control
When considering differences or similarities between cellular re-
sponses to extracellular matrices, a key component is the assortment
of integrin binding sites presented by the ECM to cellular integrin re-
ceptors. How vastly different responses can be elicited by exposure of
bone marrow-derived mesenchymal stem-like cells (hereafter re-
ferred to as mesenchymal cells) to different integrin binding sites is
well established by in vitro studies of their behavior on matrices com-
posed of native collagen I (NC I) and denatured collagen I (DC I)
(Mauney and Volloch, 2009a, 2009b), or native collagen IV (NC IV)
and denatured collagen IV (DC IV) (Mauney and Volloch, 2009b,
As components of collagen fibrils, collagen I molecules in their na-
tive structural state (NC I) present to cells up to 12 integrin binding
sites per triple-helical molecule, six for α1β1and six for α2β1 integrin
receptors (Xu et al., 2000). These integrin binding sites are triple helix
specific; upon unwinding of the helix into polypeptide constituents
(DC I), these sites are lost. Instead, seven binding sites for a different
integrin receptor, αvβ3, in each collagen I polypeptide are exposed
(Davis, 1992). These sites, known as “cryptic”, become available for
interaction with integrin receptors upon the unwinding of triple
Collagen IV molecules, in their native structural state (NC IV), con-
tain both triple helical and non-triple helical domains. Within colla-
gen IV-containing polymers, they present to cells up to six integrin
binding sites per molecule targeting three different integrin receptors:
three for α1β1, two for α2β1, and one for αvβ3 integrins (Eble et al.,
1993; Underwood et al., 1995; Colorado et al., 2000; Zarate et al.,
2004; Sudhakar et al., 2005). Denatured collagen IV (DC IV) present to
cells up to 17 integrin binding sites per polypeptide, 15 for αvβ3 and
two for α1β1 integrin receptors, which are carryovers from the
non-triple helical carboxyl domain of native collagen IV (Davis, 1992;
Marcinkiewicz et al., 2003; Sudhakar et al., 2005).
2. The “density” of integrin binding sites in the ECM and
integrin-dependent mechanical stress responses
The number of exposed integrin binding sites in the ECM per unit
cell surface area plays an important, sometimes decisive, role in elic-
itation of cellular responses. This aspect of integrin binding sites, here
referred to as “density”, manifests itself in two ways. One is the extent
of activation of a particular signal transduction pathway. For example,
the engagement of α1β1 and α2β1 binding sites strongly activates
ERK (Egan et al., 1993; Schlaepfer and Hunter, 1996; Takeuchi et al.,
1997; Mauney and Volloch, 2009a, 2009b) and activates p38 kinase
rather weakly (Klekotka et al., 2001; Mauney and Volloch, 2009a,
2009b). The more α1β1 and α2β1 sites are engaged, the higher are
the levels of activated ERK. In contrast, the engagement of αvβ3 bind-
ing sites strongly activates p38 kinase and rather weakly activates
ERK (Kaneki et al., 1999; Franklin et al., 2000; Hagemann and Blank,
2001; Dormond and Ruegg, 2003; Han et al., 2002; Mittelstadt et al.,
2005) the more αvβ3 sites are engaged, the higher are the levels of
activated p38 kinase.
The other manifestation of the effect of integrin binding site den-
sity is less obvious and is specific in collagen matrices for α2β1 bind-
ing sites. Provided a sufficient density of these sites, the exposure
of cells to collagen matrices presenting them constitutes cellular
stress and elicits activation of mechanical stress response pathways
(Mauney and Volloch, 2009b) which play an important role in cellular
physiology (Matthews et al., 2006). The inducing signal is the me-
chanicalforcegeneratedbyresistanceof matrixcontraction,a phenom-
enon which occurs both in vitro and in vivo (Grinnell and Lamke, 1984;
Tiollier et al., 1990; Schiro et al., 1991; Yoshizato et al., 1999). Cells grab
the matrix by its α2β1 sites (the “anchors”; Schiro et al., 1991; Mauney
and Volloch, 2009b) and try to contract, matrix resists and the force of
resistance elicits cellular stress response and activates stress response
genes such as heat shock protein 90 (Hsp90); the more α2β1 integrin
receptors on cell surfaces are engaged, the more force is generated
and the greater is the extent of elicited stress response in general and
of Hsp90 activation in particular. Contraction of a matrix consisting of
NC I, with six α2β1 sites per triple-helical molecule (Xu et al., 2000),
generates a substantial force and elicits strong cellular stress responses
(Mauney and Volloch, 2009b) whereas weak matrix contraction occurs
on a matrix composed of NC III (Tiollier et al., 1990) with three α2β1
binding sites per triple-helical molecule (Kim et al., 2005) and there is
neither contraction (Tiollier et al., 1990) nor elicitation of stress re-
sponses (Mauney and Volloch, 2009b) on matrices consisting of NC IV
(two α2β1 sites per molecule, Tiollier et al., 1990) or DC I (no α2β1
sites, Xu et al., 2000).
Such integrin-mediated cellular stress responses or lack thereof may
possibly explain the interesting effects of matrix elasticity and of cell
shape on differentiation of multipotent mesenchymal cells into different
lineages. In experiments reported by Engler et al. (2006), mesenchymal
stem cells were plated on native collagen I-coated gels of different rigid-
would expect the matrix (NC I) to be contracted by cells and the force
to the amount of force. When the NC I coating is very thin, its resistance
to contraction is defined by the elasticity of the underlying gel support.
Fig. 1. Cellular stress suppresses adipogenesis and stimulates osteogenesis in skeletal
but not in adipose tissue. The diagram illustrates how induction of Hsp90 stimulates
β-catenin-dependent signaling pathways resulting in osteogenesis while suppressing
adipogenesis in skeletal tissue. Induction of Hsp90 in adipose tissue does not suppress
adipogenesis. The thickness of arrows (blue) and of repression symbols (red) indicates
the strength of signals; dashed lines indicate greatly reduced signals.
V. Volloch, B.R. Olsen / Matrix Biology 32 (2013) 365–371
there is no or low stress response. With increasing rigidity of supporting
gels the resistance grows, more force is generated, and the extent of cel-
lular stress response increases; this, in turn, may determine the differen-
tiation pathway available to the cells. This potential explanation may be
tested by eliciting cellular stress response and thus inducing Hsp90 ex-
pression in cells on soft gels; this should change cell behavior to that
seen on rigid gels. Furthermore, inhibition of Hsp90 (for explanation of
the effect of Hsp90, see below) in cells on rigid gels should shift cell be-
havior to that seen on soft gels. Finally, if the gels were coated with DC
I (lacking α2β1 binding sites) instead of NC I, behavior of cells on rigid
and soft gels would be predicted to be the same because contraction-
mediated stress response will be elicited on neither.
Elicitation of cellular stress response and activation of Hsp90 could
possibly also explain the findings of McBeath et al. (2004), based on
experiments in which mesenchymal stem cells were plated on
microislands of fibronectin of different sizes layered on rigid plastic;
all analyzed islands allowed only one cell to attach. On large islands
a cell could spread, on small islands it could attach but remained
round. When cells were cultivated in media containing both osteo-
genic and adipogenic stimuli, only adipogenesis occurred on small
islands whereas only osteogenesis occurred on large islands; the au-
thors concluded that cell shape controls differentiation. However,
there is an alternative interpretation: on large islands, spread cells
contract the matrix, consequently this generates force and elicits a
cellular stress response, which is essential for osteogenesis and in-
hibits adipogenesis (Mauney and Volloch, 2009b). On small islands,
although density of binding sites on the matrix does not change, the
cell is able to use only a small fraction of its surface to engage binding
sites, therefore the amount of available integrin receptors becomes a
limiting factor, with the same effect as low density of binding sites:
loss of contractility. Consequently, stress response is not elicited, os-
teogenesis cannot occur but adipogenesis can (Mauney and Volloch,
2009a, 2009b). As in the previous example, it comes down to elicita-
tion of cellular stress response or lack thereof. This interpretation also
allows defined and verifiable predictions: (a) elicitation of stress re-
sponse by an exogenous stress will suppress adipogenesis and promote
osteogenesis in cells on small islands and enhance osteogenesis in cells
on large islands; (b) suppressionofstress response or functional inhibi-
tion of Hsp90 (see Section 3 below) or will suppress osteogenesis and
promote adipogenesis in cells on large islands; it will not affect
adipogenesis on small islands; (c) If denatured collagen I were used to
construct the islands, only adipogenesis would occur on both large
and small islands.
3. Upregulation of Hsp90 activates Raf-1 and enables
Raf-dependent ERK activation in mesenchymal cells
Activation of stress response mechanisms is essential for operation
of a signal transduction pathway initiated by stimulation of α1β1 and
α2β1 integrins and leading to ERK activation (Mauney and Volloch,
2009b) (Fig. 2). The reason is that this pathway involves activation
of Raf-1 kinase (Egan et al., 1993; Schlaepfer and Hunter, 1996;
Wary et al., 1996, 1998; Takeuchi et al., 1997). Activation of Raf-1 cru-
cially depends on Hsp90 being elevated above its normal housekeep-
ing cellular concentration (Cutforth and Rubin, 1994; van der Straten
et al., 1997; Caraglia et al., 2005); such elevation of Hsp90 occurs as
an integral part of cellular stress response as, for example, in mesen-
chymal cells exposed to elevated temperature or grown on native col-
lagen I (Mauney and Volloch, 2009b). In experiments with cells
cultured on native collagen I, inhibition of Hsp90 function causes the
same suppression of ERK activation as the use of specific inhibitors
of ERK activation (Mauney and Volloch, 2009a, 2009b).
Thatnormal housekeepinglevelsof Hsp90 in mesenchymalcellsare
not sufficientfor Raf-1 activation is evident from the analysis of ERK ac-
tivation in cells cultured on denatured collagen I. In this case, engage-
ment of αvβ3 integrins on the cell surface allows ERK activation via
several Raf-independent mechanisms (Blanco-Aparicio et al., 1999;
Kaneki et al., 1999; Saxena et al., 1999; Franklin et al., 2000; Short et
al., 2000; Hagemann and Blank, 2001; Gomez et al., 2002; Mittelstadt
et al., 2005; Noon and Lloyd, 2005; Tapinos and Rambukkana, 2005;
Wen-Sheng, 2005; Mauney and Volloch, 2009a, 2009b). However,
when mesenchymal cells cultured on DC I are subjected to thermal
stress (incubation at 39 °C), Hsp90 is elevated, and ERK activation
changes from a Raf-independent to a Raf-dependent mechanism
(Mauney and Volloch, 2009a, 2009b). Interestingly, these two path-
ways appear to be mutually exclusive; i.e. when both pathways are
available, only the Raf-dependent one is activated (Mauney and
Volloch, 2009b).ThatHsp90 andnot some otherstress responsecompo-
nent causes the change is clear from the observation that when a func-
tional inhibitor of Hsp90 is included in the experiment, ERK activation
ditions (Mauney and Volloch, 2009a, 2009b).
4. In cultured mesenchymal cells, adipogenic differentiation is
controlled by p38 kinase, β-catenin, and ERK
In mesenchymal cells adipogenic differentiation is activated by p38
kinase and inhibited by β-catenin (Engelmann et al., 1998; Jaiswal et
al., 2000; Salasznyk et al., 2004; Prestwich and McDougald, 2007;
Mauney and Volloch, 2009a, 2009b, 2010a, 2010b). Acting simulta-
neously on both p38 and β-catenin, activated ERK is a major indirect
factor controlling adipogenic differentiation. Extensive crosstalk be-
tween activated ERK and p38 kinase via protein phosphatases as well
as direct engagement (Tamura et al., 2002) results in a reciprocal bidi-
rectional seesaw-like balance between ERK and p38 phosphorylation
whereby the increase in p38 activity suppresses activation of ERK and
the other way around (Hotokezaka et al., 2002). Levels of active
(non-phosphorylated) β-catenin are primarily regulated by GSK 3β.
When GSK 3β is active, β-catenin is targeted, through phosphorylation,
for rapid ubiquitin-dependent degradation. When GSK 3β is inactivated,
β-catenin is stabilized, its levels are increased; it translocates into the
nucleus and regulates the transcription of target genes (Luo, 2009).
I and suppresses adipogenesis on denatured collagen I. The diagram illustrates how
engagement of integrin binding sites in collagen I (native-NCI; denatured-DCI) can affect
signaling pathways regulating osteogenesis and adipogenesis in mesenchymal cells. The
potential ERK-independent effect on GSK 3β-activity as discussed in the text is indicated
by the question mark. The thickness of arrows (blue) and of repression symbols (red) in-
dicates the strength of signals; dashed lines indicate greatly reduced signals.
V. Volloch, B.R. Olsen / Matrix Biology 32 (2013) 365–371
Activated ERK inactivates GSK 3β, resulting in stabilization and
upregulation of β-catenin (Ding et al., 2005) (Fig. 2).
In mesenchymal cells cultured on denatured collagen I, levels of ac-
tive p38, driven by stimulated αvβ3 integrins, are high (Mauney and
Volloch, 2009a, 2009b; Volloch, unpublished results). This keeps levels
of activated ERK and therefore of β-catenin low (Mauney and Volloch,
2009b), and in adipogenic media the extent of adipogenesis is high
(Mauney and Volloch, 2009a, 2009b). When mesenchymal cells
cultured in adipogenic media on denatured collagen I are subjected
to thermal stress (continuous incubation at 39 °C), the extent of
adipogenic differentiation is suppressed to baseline levels (Mauney
and Volloch, 2009a, 2009b). Interestingly, at these conditions levels of
activated p38 kinase remain high and inhibition of adipogenesis is
accomplished not through suppression of p38, but rather through
upregulation of β-catenin to the levels seen in mesenchymal cells cul-
tured on native collagen I (Mauney and Volloch, 2009a,2009b; Volloch,
unpublished results). The observation that β-catenin levels in mesen-
chymal cells are upregulated also in response to mechanical stress
(Sen et al., 2008) is consistent with the notion that this is a rather uni-
versal stress response component.
In mesenchymal cells cultured on native collagen I, levels of active
ERK, driven by α1β1/α2β1-initiated signaling and elevated Hsp90,
are high, levels of activated p38 are low (Mauney and Volloch,
2009a, 2009b; Volloch, unpublished results), β-catenin levels are
high (Mauney and Volloch, 2009b) and only marginal baseline
adipogenesis is seen in adipogenic media (Mauney and Volloch,
2009a, 2009b). That ERK is crucial for elevation of β-catenin even
when stress response is elicited is suggested by the observation that
inhibition of ERK in mesenchymal cells cultured on NC I results in
suppression of β-catenin levels and allows an efficient adipogenesis
even under stress conditions (Mauney and Volloch, 2009a, 2009b;
Volloch, unpublished results).
One could expect that when, in an experiment with mesenchy-
mal cells on DC I, ERK activation changes under stress from a Raf-
independent to a Raf-dependent mechanism, the levels of activated
ERK would significantly increase and those of activated p38 would
decline, but this does not happen (Mauney and Volloch, 2009a,
2009b). The basis for only a small change in the levels of active p38 ki-
naseisonlyalittlealterationinthelevelsofactiveERK despitea change
in mechanism of its activation (Mauney and Volloch, 2009a, 2009b;
Volloch, unpublished results). Raf-dependent activation of ERK, such
alent Raf-independent mechanisms (Egan et al., 1993; Schlaepfer and
Hunter, 1996; Takeuchi et al., 1997; Kaneki et al., 1999; Hagemann and
Blank, 2001; Mauney and Volloch, 2009a, 2009b). The factor which
limits Raf-dependent ERK activation on DC I matrix under stress condi-
tions is the level of available Raf-1. On NC I matrix, the engagement of
α2β1 integrins leads to activation of protein phosphatase pp2A which,
in turn, facilitates the release of Raf-1 sequestered by 14-3-3 proteins
and makes it available for interaction with and activation by Ras
(Sanders and Basson, 2004; Chetoui et al., 2006). This mechanism is
not available on DC I matrix because of its lack of α2β1 integrin binding
sites; consequently, low levels of available Raf-1 limit the extent of ERK
activation under stress conditions. Thus, in mesenchymal cells on DC I
ERK, β-catenin is elevated in one case (stress-mediated Raf-dependent
ERK activation) but not in another (Raf-independent ERK activation in
the absence of stress). This indicates that in addition to ERK, another stress
5. Adipogenic differentiation of mesenchymal cells on collagen IV
Inadipose tissues, adipogenic differentiationtakes place ina collagen
IV-containing ECM (Kuri-Harcuch et al., 1984; Nakajima et al., 1998;
Pierleoni et al., 1998; Kubo et al., 2000; Bouloumie et al., 2001). Because
of a low density of α2β1 integrin binding sites, matrix consisting of NC
IV neither mediates contraction (Tiollier et al., 1990) nor elicits
upregulation of Hsp90 (Mauney and Volloch, 2009b) in mesenchymal
cells, and their functional interactions with the matrix occur mostly
through αvβ3 integrins. When mesenchymal cells are cultured in
adipogenic media and on a matrix consisting of native collagen IV,
only low levels of p38 and marginal baseline adipogenesis can be seen
because of low density of αvβ3 integrin binding sites (Mauney and
Volloch, 2010b; Volloch, unpublished results) (Fig. 3).
However, when mesenchymal cells are cultured in adipogenic
media on a matrix composed of denatured collagen IV, the effect is
a dramatically increased adipogenic differentiation (Fig. 3) which
exceeds that observed with cells cultured on DC I by more than an
order of magnitude (Mauney and Volloch, 2010b). This is also easily
explained in terms of presentation of integrin binding sites: Increased
density of αvβ3 integrin binding sites (15 per molecule versus only
one in NC IV and seven on DC I) leads to increased levels of activated
p38 kinase (Mauney and Volloch, 2010b; Volloch, unpublished re-
sults). This raises the possibility that in adipose tissues, where
MMPs are produced and secreted by differentiating adipocytes and
were shown to be essential for adipogenesis (Bouloumie et al.,
2001), an increase in the amount of collagen IV polypeptide frag-
ments may contribute to the promotion of adipogenic differentiation.
This notion is consistent with observations that abnormalities associ-
ated with adipogenesis, such as obesity, are often associated with
overproduction of MMPs and a decline in levels of TIMPs (Chavey et
al., 2003). Obesity-associated changes in the MMP/TIMP balance, cur-
rently regarded as a consequence of the disease (Chavey et al., 2003),
would result in the generation of collagen IV-derived polypeptide frag-
ments. In lightof thedata discussed, suchtransition may constitute one
of the driving forces of the disease. Advances in in vivo visualization of
MMP activity (Artym et al., 2009; Packard et al., 2009) combined with
the recent development of techniques for in vivo visualization of colla-
genchainsbyphoto-triggered triple-helix hybridization (Lietal.,2012)
makeit possible for thefirsttime to address this notionexperimentally.
When mesenchymal cells cultured on denatured collagen IV are
subjected to thermal stress as discussed above (incubation at 39 °C),
there is, in contrast to observations with cells on denatured collagen I,
no effect on the extent of adipogenesis (Mauney and Volloch, 2010b)
despite the fact that a cellular stress response is elicited and levels of
Hsp90 are elevated (Mauney and Volloch, 2009b, 2010b; unpublished
results). There is also no elevation of β-catenin (Volloch, unpublished
results), the principal physiological inhibitor of adipogenic differentiation
(Prestwich and McDougald, 2007) (Fig. 3). It appears therefore that the
exposure to collagen IV-containing extracellular matrix confers stress re-
sistance to adipogenic differentiation in multipotent mesenchymal cells.
6. Stress-resistance of adipogenic differentiation of mesenchymal
cells on DC IV is due to the effect of specialized integrin
A clue to the underlying mechanism of this surprising resistance of
adipogenic differentiation to thermal stress when cells are cultured on
denatured collagen IV, whereas thermal stress suppresses adipogenesis
when cells are cultured on denatured collagen I, is suggested by the
comparison of integrin binding sites presented by these two matrices.
The only qualitative difference between the two is the presence of
two α1β1 integrin binding sites per molecule of DC IV but not in DC I.
These two sites are different, both structurally and functionally, from
all other α1β1 integrin binding sites which are strictly triple helix spe-
wound structural states (Xu et al., 2000). In contrast, these two sites
are located in the carboxyl non-triple helical domain of collagen IV
their target, the α1β1 integrin receptor, through a different integrin
domain than other α1β1 binding sites (Marcinkiewicz et al., 2003;
V. Volloch, B.R. Olsen / Matrix Biology 32 (2013) 365–371
Monleon et al., 2003). The engagement of these sites with α1β1
integrin receptors could therefore elicit cellular responses that are dif-
ferent from those elicited through stimulation of α1β1 integrin recep-
tors by other, “regular”, α1β1 integrin binding sites. This may explain
why β-catenin levels are not upregulated and adipogenesis is not
suppressed in thermally stressed mesenchymal cells cultured on dena-
that in thermally stressed mesenchymal cells cultured on denatured col-
lagen IV in the presence of an antibody to α1β1 integrin, levels of
β-catenin were elevated and comparable to those seen on a matrix of na-
tive collagen I (Volloch, unpublished results).
A potential mechanism for suppression of β-catenin levels in ther-
mally stressed mesenchymal cells cultured on denatured collagen IV
is suggested by experiments supporting the conclusion that α1β1
integrin binding sites in the non-triple helical carboxyl domain of col-
lagen IV suppress Raf-dependent ERK activation (Sudhakar et al.,
2005) and thus block stress-induced, ERK-mediated, inactivation of
GSK 3β (Ding et al., 2005), elevation of β-catenin and suppression
of adipogenesis (Fig. 3) . In thermally stressed mesenchymal cells cul-
tured on denatured collagen IV, levels of activated ERK are similar to
those seen in unstressed cells (Mauney and Volloch, 2009b, 2010b),
yet β-catenin is upregulated only when α1β1 integrins are blocked
under stress conditions. This observation is consistent with the con-
clusion that in addition to ERK, another stress response-associated
factor is required for GSK 3β control of β-catenin (see above) and
opens the possibility that this factor might be a component of
Raf-dependent pathways upstream of ERK.
7. Specialized α1β1 integrin binding sites safeguard
pluripotency of stem cells in their storage niches
in adipose tissues and
When mesenchymal cells are induced to differentiate on native
collagen IV, they do not undergo adipogenic or osteogenic differenti-
ation (Mauney and Volloch, 2009b, 2010b). The reason for the failure
of osteogenesis is the same as for the failure of adipogenesis: func-
tional interaction of cells with NC IV occurs mostly through low den-
sity αvβ3 integrin binding sites and this low density elicits a low
cellular response. Thus, in terms of the ability to support differentia-
tion of mesenchymal cells, native collagen IV is an inert extracellular
matrix. When thermal stress is applied to mesenchymal cells cultured
on native collagen IV in osteogenic medium, Hsp90 levels are
elevated (Mauney and Volloch, 2009b), but osteogenesis does not
occur (Mauney and Volloch, 2009b); there is also no activation of
ERK or upregulation of β-catenin (Volloch, unpublished results).
What blocks stress-mediated differentiation in cells cultured on NC
IV is the engagement of the α1β1 integrin binding sites in the carbox-
yl domain of collagen IV. Thus, suppression of Raf-dependent ERK ac-
tivation by these “special” α1β1 integrin binding sites ensures the
inertness of native collagen IV-containing matrices even under ther-
mal stress conditions.
Despite their relatively low density, α1β1 integrin binding sites in
the non-triple helical carboxyl domain of collagen IV are effective in
suppressing Raf-dependent ERK activation and elevation of β-catenin
possiblydue to their high density withinclusters of interacting collagen
IV carboxyl domains within extracellular matrices. Thus, through their
action, these α1β1 integrin binding sites in the non-triple helical carboxyl
domain of collagen IV promote two processes pivotal in physiology and
evolution: uninterrupted adipogenesis in adipose tissues even when stress
response pathways are activated, crucial for survival during periods of
limited and intermittent food supply, and preservation of pluripotency of
mesenchymal cells in their niches which are lined by a matrix containing
native collagen IV, a principal component of basal laminae (Sanes et al.,
1990; Grisham and Thorgeirsonn, 1997; Rudland et al., 1997; Mercier
et al., 2003; Blanpain et al., 2004).
8. Strategies to suppress adipogenesis in adipose tissues
lical carboxyl domain of collagen IV in the preservation of pluripotency
of mesenchymal cells in their niches remains undisputedly beneficial,
their role in preservationof adipogenesis in adipose tissues has become
detrimental in human modern societies (Neel, 1962; Prentice et al.,
2005; Stoger, 2008). Although the detailed mechanism of action
of this binding site is not yet elucidated, it is possible, based on current
knowledge, to design approaches to suppress its adipogenesis-
One strategy may be to somehow prevent their functional interac-
tion with α1β1 integrin receptors on cell surfaces, and to combine
this with elicitation of cellular stress response or even of only
Fig. 3. In mesenchymal stem cells on collagen IV matrices cellular stress is completely ineffective in promoting differentiation on native collagen IV (NC IV) or suppressing
adipogenesis on denatured collagen IV (DC IV). The diagram illustrates how engagement of integrin binding sites in collagen IV (α1(IV) and α2(IV); native-NC IV;
denatured-DC IV) can affect signaling pathways regulating osteogenesis and adipogenesis in mesenchymal cells. As explained in the text, the “α1β1” binding site in the collagen
IV carboxyl domain differs from the α1β1 binding sites in collagen I and engages the α1β1 integrin in a manner that differs (“α1β1”) from the engagement of α1β1 by binding
sites in collagen I. The potential ERK-independent effects on GSK 3β-activity as discussed in the text is indicated by the question mark. The thickness of arrows (blue) and of
repression symbols (red) indicates the strength of signals; dashed lines indicate greatly reduced signals.
V. Volloch, B.R. Olsen / Matrix Biology 32 (2013) 365–371
Hsp90 expression. This would activate the Raf-dependent ERK cas-
cade, upregulate β-catenin and consequently suppress adipogenesis.
Elevation of Hsp90 may be accomplished by physical exercise or its
equivalent or mimetic. A naturally occurring substance capable of
competing with α1β1 integrin binding sites, and preferentially with
those in the non-triple helical carboxyl domain of collagen IV, for
binding to α1β1 integrin receptors is known: obtustatin, a small, 41
amino acid-long, snake venom disintegrin and potent inhibitor of
α1β1 integrins (Marcinkiewicz et al., 2003).
Another strategy may be to use an agent or combination of agents
capable of elevating β-catenin in cells despite the suppressive effect
of α1β1 integrin binding sites in the carboxyl domain of collagen
IV; such an agent(s) should suppress adipogenesis in adipose tissues.
For example, elevation of Hsp90 in combination with an agent that
neutralizes α1β1 integrin binding sites in collagen IV would be suffi-
cient for upregulation of β-catenin. The identification of a component
of Raf-dependent ERK activation pathway required, in addition to
ERK, for GSK 3β inactivation would provide another potential target;
its activation would result in upregulation of β-catenin and suppres-
sion of adipogenesis.
Akpan, I., Goncalves, M., Dhir, R., Yin, X., Pistilli, E., Bogdanovich, S., et al., 2009. The ef-
fects of a soluble activin type IIB receptor on obesity and insulin sensitivity. Int. J.
Obes. 33, 1265–1273.
Artym, V., Yamada, K., Mueller, S., 2009. ECM degradation assays for analyzing local cell
invasion. Methods Mol. Biol. 522, 211–219.
Askew, E., Huston, R., Plopper, C., Hecker, A., 1975. Adipose tissue cellularity and lipol-
ysis. Response to exercise and cortisol treatment. J. Clin. Invest. 56, 521–529.
Bareja, A., Billin, A., 2013. Satellite cell therapy – from mice to men. Skeletal Muscle 3,
Blanco-Aparicio, C., Torres, J., Pulido, R., 1999. A novel regulatory mechanism of MAP
kinases activation and nuclear translocation mediated by PKA and the PTP-SL tyro-
sine phosphatase. J. Cell Biol. 147, 1129–1136.
Blanpain, C., Lowry, W.E., Geoghegan, A., Polak, L., Fuchs, E., 2004. Self-renewal,
multipotency, and the existence of two cell populations within an epithelial stem
cell niche. Cell 118, 635–648.
Bouloumie, A., Sengenes, C., Portolan, G., Galizky, J., Lafontan, M., 2001. Adipocyte pro-
duces matrix metalloproteinases 2 and 9. Involvement in adipose differentiation.
Diabetes 50, 2080–2086.
Caraglia, M., Marra, M., Mancinelli, F., D'Ambrosio, G., Massa, R., Giordano, A., Budillon,
A., Abbruzzese, A., Bismuto, E., 2005. Electromagnetic fields at mobile phone fre-
quency induce apoptosis and inactivation of the multi-chaperone complex in
human epidermoid cancer cells. J. Cell. Physiol. 204, 539–548.
Chavey, C., Mari, B., Monthouel, M.N., Bonnafous, S., Anglard, P., Van Obberghen, E.,
Tartare-Deckert, S., 2003. Matrix metalloproteinases are differentially expressed
in adipose tissue during obesity and modulate adipocyte differentiation. J. Biol.
Chem. 278, 11888–11896.
Chetoui, N., Gendron, S., Chamoux, E., Aoudjit, F., 2006. Collagen type I-mediated acti-
vation of ERK/MAP kinase is dependent on Ras, Raf-1 and protein phosphatase 2A
in Jurkat T cells. Mol. Immunol. 43, 1687–1693.
Colorado, P.C., Torre, A., Kamphaus, G., Maeshima, Y., Hopfer, H., Takahashi, K., Volk, R.,
Zamborsky, E.D., et al., 2000. Anti-angiogenic cues from vascular basement mem-
brane collagen. Cancer Res. 60, 2520–2526.
Cutforth, T., Rubin, G.M., 1994. Mutations in Hsp83 and cdc37 impair signaling by the
sevenless receptor tyrosine kinase in Drosophila. Cell 77, 1027–1036.
Davis, G.E., 1992. Affinity of integrins for damaged extracellular matrix: alpha v beta 3
binds to denatured collagen type I through RGD sites. Biochem. Biophys. Res.
Commun. 182, 1025–1031.
Ding, Q., Xia, W., Liu, J., Yang, J., Lee, D., Xia, J., Bartholomeusz, G., Li, Y., Pan, Y., Li, Z.,
Bargou, R., Qin, J., Lai, C., Tsai, F., Tsai, C., Hung, M., 2005. Erk associates with and
primes GSK-3beta for its inactivation resulting in upregulation of beta-catenin.
Mol. Cell 19, 159–170.
Dormond, O., Ruegg, C., 2003. Regulation of endothelial cell integrin function and
angiogenesis by COX-2, cAMP and Protein Kinase A. Thromb. Haemost. 90,
Eble, J.A., Golbik, R., Mann, K., Kuhn, K., 1993. The alpha 1 beta 1 integrin recognition
site of the basement membrane collagen molecule [alpha 1(IV)]2 alpha 2(IV).
EMBO J. 12, 4795–4802.
Egan, S., Giddings, B., Brooks, M., Buday, L., Sizeland, A., Weinberg, R., 1993. Association
of Sos Ras exchange protein with Grb2 is implicated in tyrosine kinase signal trans-
duction and transformation. Nature 363, 45–51.
Engelman, J., Lisanti, M., Scherer, P.E., 1998. Specific inhibitors of p38 mitogen-
activated protein kinase block 3 T3-L1 adipogenesis. J. Biol. Chem. 273,
Engler, A., Sen, S., Sweeney, H., Discher, D., 2006. Matrix Elasticity Directs Stem Cell
Lineage Specification. Cell 126, 677–689.
Franklin, R., Atherfold, P., McCubrey, J.A., 2000. Calcium-induced ERK activation in
human T lymphocytes occurs via p56(Lck) and CaM-kinase. Mol. Immunol. 37,
Gomez, E., Pritchard, C., Herbert, T., 2002. cAMP-dependent protein kinase and Ca2+
influx through L-type voltage-gated calcium channels mediate Raf-independent
activation of extracellular regulated kinase in response to glucagon-like peptide-
1 in pancreatic beta-cells. J. Biol. Chem. 277, 48146–48151.
Grinnell, F., Lamke, C.R., 1984. Reorganization of hydrated collagen lattices by human
skin fibroblasts. J. Cell Sci. 66, 51–63.
Grisham, J., Thorgeirsonn, S., 1997. Liver stem cells. In: Potten, C.S. (Ed.), Stem Cells.
Academic Press, London, UK, pp. 233–282.
Hagemann, C., Blank, J.L., 2001. The ups and downs of MEK kinase interactions. Cell.
Signal. 13, 863–875.
Han, Q., Leng, J., Bian, D., Mahanivong, C., Carpenter, K.A., Pan, Z.K., Han, J., Huang, S.,
2002. Rac1-MKK3-p38-MAPKAPK2 pathway promotes urokinase plasminogen
activator mRNA stability in invasive breast cancer cells. J. Biol. Chem. 277,
Havekes, B., Sauerwein, H., 2010. Adipocyte-myocyte crosstalk in skeletal muscle insu-
lin resistance. Curr. Opin. Clin. Nutr. Metab. Care 13, 641–646.
Hotokezaka, H., Sakai, E., Kanaoka, K., et al., 2002. U0126 and PD98059, specific inhib-
itors of MEK, accelerate differentiation of RAW264.7 cells into osteoclast-like cells.
J. Biol. Chem. 277, 47366–47372.
Jaiswal, R.K., Jaiswal, N., Bruder, S.P., Mbalaviele, G., Marshak, D., Pittenger, M., 2000.
Adult human mesenchymal stem cell differentiation to the osteogenic or adipogenic
lineage is regulated by mitogen-activated protein kinase. J. Biol. Chem. 275,
Kadi, F., 2008. Cellular and molecular mechanisms responsible for the action of testos-
terone on human skeletal muscle. Br. J. Pharmacol. 154, 522–528.
Kaneki, M., Kharbanda, S., Pandey, P., Yoshida, K., Takekawa, M., Liou, J.R., Stone, R.,
Kufe, D., 1999. Functional role for protein kinase Cbeta as a regulator of stress-
activated protein kinase activation and monocytic differentiation of myeloid leuke-
mia cells. Mol. Cell. Biol. 19, 461–470.
Kim, J.K., Xu, Y., Xu, X., Keene, D.R., Gurusiddappa, S., Liang, X., Wary, K.K., Hook, M.,
2005. A novel binding site in collagen type III for integrins alpha1beta1 and
alpha2beta1. J. Biol. Chem. 280, 32512–32520.
Klekotka, P.A., Santoro, S.A., Zutter, M.M., 2001. alpha 2 integrin subunit cytoplasmic
domain-dependent cellular migration requires p38 MAPK. J. Biol. Chem. 276,
Kubo, Y., Kaidzu, S., Nakajima, I., Takenouchi, K., Nakamura, F., 2000. Organization of
extracellular matrix components during differentiation of adipocytes in long-
term culture. In Vitro Cell. Dev. Biol. Anim. 36, 38–44.
Kuri-Harcuch, W., Arguello, C., Marsch-Moreno, M., 1984. Extracellular matrix produc-
tion by mouse 3 T3–F442A cells during adipose differentiation in culture. Differen-
tiation 28, 173–178.
Li, Y., Foss, C., Summerfield, D., Doyle, J., Torok, C., Dietz, H., Pomper, M., Yu, S., 2012.
Targeting collagen strands by photo-triggered triple-helix hybridization. Proc.
Natl. Acad. Sci. U. S. A. 109, 14767–14772.
Luo, J., 2009. Glycogen synthase 3 beta in tumorogenesis and cancer chemotherapy.
Cancer Lett. 273, 194–200.
Marcinkiewicz, C., Weinreb, P., Calvete, J., Kisiel, D., Mousa, S., Tuszinski, G., Lobb, R.,
2003. Obtustatin: a potent selective inhibitor of alpha1/beta1 integrin in vitro
and angiogenesis in vivo. Cancer Res. 63, 2020–2023.
Matthews, B., Overby, D., Mannix, R., Ingber, D., 2006. Cellular adaptation to mechani-
cal stress: role of integrins, Rho, cytoskeletal tension and mechanosensitive ion
channels. J. Cell Sci. 119, 508–518.
Mauney, J., Volloch, V., 2009a. Progression of human bone marrow stromal cells into
both osteogenic and adipogenic lineages is differentially regulated by structural
conformation of collagen I matrix via distinct signaling pathways. Matrix Biol. 28,
Mauney, J., Volloch, V., 2009b. Collagen I matrix contributes to determination of adult
human stem cell lineage via differential, structural conformation-specific elicita-
tion of cellular stress response. Matrix Biol. 28, 251–262.
Mauney, J., Volloch, V., 2010a. Adult human bone marrow stromal cells regulate ex-
pression of their MMPs and TIMPs in differentiation type-specific manner. Matrix
Biol. 29, 3–8.
Mauney, J., Volloch, V., 2010b. Human bone marrow-derived stromal cells show highly
efficient stress-resistant adipogenesis on denatured collagen IV matrix but not on
its native counterpart: implications for obesity. Matrix Biol. 29, 9–14.
McBeath, R., Pirone, D., Nelson, C., Bhadriraju, K., Chen, C., 2004. Cell shape, cytoskeletal
tension, and RhoA regulate stem cell lineage commitment. Dev. Cell 6, 483–495.
Mercier, F., Kitasako, J.T., Hatton, G.I., 2003. Fractones and other basal laminae in the
hypothalamus. J. Comp. Neurol. 455, 324–340.
Mittelstadt, P.R., Salvador, J.M., Fornace Jr., A.J., Ashwell, J.D., 2005. Activating p38
MAPK: new tricks for an old kinase. Cell Cycle 4, 1189–1192.
Monleon, D., Moreno-Murciano, M., Kovacs, H., Marcinkiewicz, C., Calvete, J., Celda, B.,
2003. Concerted motions of the integrin-binding loop and the C-terminal tail of the
non-RGD disintegrin obtustatin. J. Biol. Chem. 278, 45570–45576.
Nakajima, I., Yamaguchi, T., Ozutsumi, K., Aso, H., 1998. Adipose tissue extracellularmatrix:
newly organized by adipocytes during differentiation. Differentiation 63,
Neel, J., 1962. Diabetes mellitus: a ‘thrifty’ genotype rendered detrimental by ‘progress’.
Am. J. Hum. Genet. 14, 353–362.
Noon, L.A., Lloyd, A.C., 2005. Hijacking the ERK signaling pathway: Mycobacterium leprae
shuns MEK to drive the proliferation of infected Schwann cells. Sci. STKE 2005, 52.
Packard, B., Artym, V., Komoriya, A., Yamada, K., 2009. Direct visualization of protease
activity on cells migrating in three dimensions. Matrix Biol. 28, 3–10.
V. Volloch, B.R. Olsen / Matrix Biology 32 (2013) 365–371
Pierleoni, C., Verdenelli, F., Castellucci, M., Cinti, S., 1998. Fibronectins and basal lamina Download full-text
Prentice, A., Rayco-Solon, P., Moore, S., 2005. Insights from the developing world:
thrifty genotypes and thrifty phenotypes. Proc. Nutr. Soc. 64, 153–161.
Prestwich, T., MacDougald, O., 2007. Wnt/beta-catenin signaling in adipogenesis and
metabolism. (2007). Curr. Opin. Cell Biol. 19, 612–617.
Rudland, P.S., Barraclough, R., Fernig, D.G., Smith, J.A., 1997. Mammary stem cells in
normal development and cancer. In: Potten, C.S. (Ed.), Stem Cells. Academic
Press, London, UK, pp. 147–232.
Salasznyk, R.M., Klees, R.F., Hughlock, M.K., Plopper, G.E., 2004. ERK signaling pathways
regulate the osteogenic differentiation of human mesenchymal stem cells on colla-
gen I. Cell Commun. Adhes. 11, 137–153.
Sanders, M.A., Basson, M.D., 2004. Collagen IV regulates Caco-2 migration and ERK ac-
tivation via alpha1beta1- and alpha2beta1-integrin-dependent Src kinase activa-
tion. Am. J. Physiol. Gastrointest. Liver Physiol. 286, G547–G557.
Sanes, J.R., Engvall, E., Butkowski, R., Hunter, D.D., 1990. Molecular heterogeneity of
basal laminae: isoforms of laminin and collagen IV at the neuromuscular junction
and elsewhere. J. Cell Biol. 111, 1685–1699.
Saxena, M., Williams, S., Tasken, K., Mustelin, T., 1999. Crosstalk between cAMP-
dependent kinase and MAP kinase through a protein tyrosine phosphatase. Nat.
Cell Biol. 1, 305–311.
Schiro, J.A., Chan, B.M., Roswit, W.T., Kassner, P.D., Pentland, A.P., Hemler, M.E., Eisen,
A.Z., Kupper, T.S., 1991. Integrin alpha 2 beta 1 (VLA-2) mediates reorganization
and contraction of collagen matrices by human cells. Cell 67, 403–410.
Schlaepfer, D.D., Hunter, T., 1996. Evidence for in vivo phosphorylation of the Grb2
SH2-domain binding site on focal adhesion kinase by Src-family protein-tyrosine
kinases. Mol. Cell. Biol. 16, 5623–5633.
Sen, B., Xie, Z., Case, N., Ma, M., Rubin, C., Rubin, J., 2008. Mechanical strain inhibits
adipogenesis in mesenchymal stem cells by stimulating a durable b-catenin signal.
Endocrinology 149, 6065-6065.
Short, S.M., Boyer, J.L., Juliano, R.L., 2000. Integrins regulate the linkage between
upstream and downstream events in G protein-coupled receptor signaling to
mitogen-activated protein kinase. J. Biol. Chem. 275, 12970–12977.
Simonsen, L., Henriksen, O., Enevoldsen, L., Bulow, J., 2004. The effect of exercise on
regional adipose tissue and splanchnic lipid metabolism in overweight type 2 dia-
betic subjects. Diabetologia 47, 652–659.
Stoger, R., 2008. The thrifty epigenotype: an acquired and heritable predisposoition for
obesity and diabetes? Bioessays 30, 156–166.
Sudhakar, A., Nyberg, P., Keshamouni, V.G., Mannam, A.P., Li, J., Sugimoto, H., Cosgrove, D.,
Kalluri, R., 2005. Human alpha1 type IV collagen NC1 domain exhibits distinct
antiangiogenic activity mediated by alpha1beta1 integrin. J. Clin. Invest. 115,
Takeuchi, Y., Suzawa, M., Kikuchi, T., Nishida, E., Fujita, T., Matsumoto, T., 1997. Differ-
entiation and transforming growth factor-beta receptor down-regulation by
collagen-alpha2beta1 osteoblastic cells. J. Biol. Chem. 272, 29309–29316.
Tamura, S., Hanada, M., Ohnishi, M., Katsura, K., Sasaki, M. Kobayashi T., 2002. Regula-
tion of stress-activated protein kinase signaling pathways by protein phospha-
tases. Eur. J. Biochem. 269, 1060–1066.
Tapinos, N., Rambukkana, A., 2005. Insights into regulation of human Schwann cell pro-
liferation by Erk1/2 via a MEK-independent and p56Lck-dependent pathway from
leprosy bacilli. Proc. Natl. Acad. Sci. U.S.A. 102, 9188–9193.
Tiollier, J., Dumas, H., Tardy, M., Tayot, J.L., 1990. Fibroblast behavior on gels of type I, III,
and IV human placental collagens. Exp. Cell Res. 191, 95–104.
Underwood, P.A., Bennett, F.A., Kirkpatrick, A., Bean, P.A., Moss, B.A., 1995. Evidence for
the location of a binding sequence for the alpha 2 beta 1 integrin of endothelial
cells, in the beta 1 subunit of laminin. Biochem. J. 309, 765–771.
van der Straten, A., Rommel, C., Dickson, B., Hafen, E., 1997. The heat shock protein
83 (Hsp83) is required for Raf-mediated signalling in Drosophila. EMBO J. 16,
Wary, K.K., Mainiero, F., Isakoff, S.J., Marcantonio, E.E., Giancotti, F.G., 1996. The adaptor
protein Shc couples a class of integrins to the control of cell cycle progression. Cell
Wary, K.K., Mariotti, A., Zurzolo, C., Giancotti, F.G., 1998. A requirement for caveolin-1
and associated kinase Fyn in integrin signaling and anchorage-dependent cell
growth. Cell 94, 625–634.
Wen-Sheng, W., 2005. Protein kinase C alpha trigger Ras and Raf-independent MEK/
ERK activation for TPA-induced growth inhibition of human hepatoma cell
HepG2. Cancer Lett. 239, 27–35.
Xu, Y., Gurusiddappa, S., Rich, R.L., Owens, R.T., Keene, D.R., Mayne, R., Hook, A., Hook,
M., 2000. Multiple binding sites in collagen type I for the integrins alpha1beta1 and
alpha2beta1. J. Biol. Chem. 275, 38981–38989.
Yoshizato, K., Tsukahara, M., Oki, T., Hayashi, M., Obara, M., Morpho, Y., 1999. The inter-
action of cellular fibronectin with collagen during fibroblast-mediated contraction
of collagen gels. J. Investig. Dermatol. Symp. Proc. 4, 190–195.
Zarate, S., Romero, P., Espinosa, R., Arias, C.F., Lopez, S., 2004. VP7 mediates the interac-
tion of rotaviruses with integrin alphavbeta3 through a novel integrin-binding site.
J. Virol. 78, 10839–10847.
V. Volloch, B.R. Olsen / Matrix Biology 32 (2013) 365–371