Extracellular Matrix Fibronectin Stimulates the Self-Assembly
of Microtissues on Native Collagen Gels
Carlos A. Sevilla, M.S.,1,2Diane Dalecki, Ph.D.,1,2and Denise C. Hocking, Ph.D.1–3
Fibronectin is an adhesive glycoprotein that is polymerized into extracellular matrices via a tightly regulated,
cell-dependent process. Here, we demonstrate that fibronectin matrix polymerization induces the self-assembly
of multicellular structures in vitro, termed tissue bodies. Fibronectin-null mouse embryonic fibroblasts adherent
to compliant gels of polymerized type I collagen failed to spread or proliferate. In contrast, addition of fibro-
nectin to collagen-adherent fibronectin-null mouse embryonic fibroblasts resulted in a dose-dependent increase
in cell number, and induced the formation of three-dimensional (3D) multicellular structures that remained
adherent and well-spread on the native collagen substrate. An extensive fibrillar fibronectin matrix formed
throughout the microtissue. Blocking fibronectin matrix polymerization inhibited both cell proliferation and
microtissue formation, demonstrating the importance of fibronectin fibrillogenesis in triggering cellular self-
organization. Cell proliferation, tissue body formation, and tissue body shape were dependent on both fibro-
nectin and collagen concentrations, suggesting that the relative proportion of collagen and fibronectin fibrils
polymerized into the extracellular matrix influences the extent of cell proliferation and the final shape of mi-
crotissues. These data demonstrate a novel role for cell-mediated fibronectin fibrillogenesis in the formation and
vertical assembly of microtissues, and provide a novel approach for engineering complex tissue architecture.
in tissue development have emerged.1Cell function is con-
trolled through complex and dynamic interactions between
cells and their surrounding extracellular matrix (ECM). This
3D network is composed of collagens, glycoproteins, and
proteoglycans, and provides an adhesive substrate for the
organization of cells into tissues.2Fibronectin is a principal
component of the ECM and plays a key role in early em-
bryonic development.3In the ECM, fibronectin is organized
as an extensive network of elongated, branching fibrils. So-
luble fibronectins are assembled (polymerized) into insoluble
ECM fibrils via a tightly regulated, cell-dependent process
that can be rapidly up- or downregulated.4Most adherent
cells assemble a fibronectin matrix via a similar cellular
mechanism involving the actin cytoskeleton and integrin
receptors.4In turn, ligation of ECM fibronectin fibrils by cell
surface receptors stimulates actin-dependent processes, in-
cluding cell spreading,5growth,6and migration,7,8promotes
cell cohesion,9,10and regulates morphogenetic movements in
Xenopus embryos.11Additionally, fibronectin matrix poly-
merization promotes the co-deposition of collagen fibrils into
any of the recent advances in regenerative medicine
have been made as details of the mechanisms involved
the ECM,12stimulates collagen fibril contraction,13and in-
creases the tensile strength of cell-embedded collagen gels.14
These studies suggest that fibronectin matrix polymerization
affects embryonic development and tissue morphogenesis
through bidirectional control of ECM and cytoskeletal or-
ganization and function.
Cellular self-assembly describes the aggregation of indi-
vidual cells into spheroids or mixed populations of cells into
multi-layered tissue-like structures.15Self-assembly is es-
sential during development and morphogenesis, and is
emerging as an attractive approach to engineering complex
3D tissues. During self-assembly, cell–cell and cell–ECM
adhesions generate mechanical forces and chemical signals
that control cell motility, cell proliferation, and tissue struc-
ture.9,16–18The balance of adhesive and anti-adhesive forces
is thought to be a critical determinant in tissue morphogen-
esis, and may play a role in cellular self-assembly.19,20Using
this premise, we developed a model of self-assembled
microtissue by combining compliant, low-adhesive type I
collagen substrates21with the cell-mediated formation of
highly adhesive fibronectin matrix fibrils.22Addition of
fibronectin to collagen-adherent fibronectin-null mouse em-
bryonic fibroblast (FN-null MEFs) triggered cell proliferation
and induced the formation of multicellular, dome-shaped
1Department of Biomedical Engineering, University of Rochester, Rochester, New York.
2The Rochester Center for Biomedical Ultrasound, University of Rochester, Rochester, New York.
3Department of Pharmacology and Physiology, University of Rochester, Rochester, New York.
TISSUE ENGINEERING: Part A
Volume 16, Number 12, 2010
ª Mary Ann Liebert, Inc.
microtissues, referred to as ‘‘tissue bodies.’’ Altering fibro-
nectin and collagen concentrations influenced the extent of
cell proliferation as well as the ultimate shape of the tissue
bodies. These studies demonstrate the importance of ECM
composition and organization in tissue self-assembly and
provide a useful model for studying cell motility, growth,
and morphogenesis in three dimensions.
Materials and Methods
Human plasma fibronectin was isolated from Cohn’s
fraction I and II.23Fibronectin fragments (60, 120, and 160/
180kDa) were generated proteolytically, as described.24La-
minin was from BD Biosciences. Recombinant vitronectin
was produced in bacteria and purified on heparin-Sepharose
(GE Healthcare).25Type I collagen was extracted from rat tail
tendons using acetic acid and precipitated with NaCl.26Re-
combinant His-tagged functional upstream domain (FUD,
also referred to as pUR-4) and the control peptide, Del29
(provided by Dr. Deane Mosher, University of Wisconsin,
Madison, WI), were expressed in bacteria and purified on
nickel-Sepharose (GE Healthcare).27Antibodies and their
sources are as follows: anti-fibronectin clone 9D228(a gift
from Dr. Deane Mosher, University of Wisconsin, Madison,
WI); anti-fibronectin clone L8, (a gift from Dr. Michael
Chernousov, Weis Center for Research, Geisinger Clinic, PA);
anti-bromodeoxyuridine (BrdU) mAb, Alexa594-conjugated
goat anti-mouse IgG, and Alexa Fluor488-conjugated goat
anti-rabbit IgG (Invitrogen); and nonimmune mouse IgG and
polyclonal anti-fibronectin (Sigma).
FN-null MEFs6(provided by Dr. Jane Sottile, University of
Rochester, Rochester, NY) were cultured on collagen I-coated
dishes under fibronectin- and serum-free conditions using a
1:1 mixture of Cellgro?(Mediatech) and Aim V (Invitrogen).
Native type I collagen gels were prepared by mixing
(DMEM; Life Technologies), type I collagen, and 1?DMEM
on ice such that the final mixture contained 0.25–2.5mg/mL
collagen and 1?DMEM. The pH of the collagen solution was
adjusted to 7.2–7.4 with 0.1N NaOH. Aliquots (0.21mL/
cm2) of the collagen mixture were added to wells of 48-well
tissue culture plates or 35mm tissue culture dishes and al-
lowed to polymerize overnight at 378C and 8% CO2. Col-
lagen gels were then equilibrated for 1h at 378C in 8% CO2
with a 1:1 mixture of Aim V/Cellgro. Polymerized collagen
gels had an average thickness of 2mm. To form thin films of
monomeric collagen, tissue culture wells were coated over-
night at 48C with 1mg/mL collagen in 0.02N acetic acid, and
then washed three times with ice-cold phosphate-buffered
Dulbecco’s modifiedEagle’s medium
Cell growth assays
Monodispersed FN-null MEFs cells were seeded on po-
lymerized collagen gels (5.26?104cells/cm2) in 48-well tissue
culture plates. Control collagen gels received an equal vol-
ume of media without cells. Cells were allowed to adhere to
the collagen substrates for 4h at 378C and 8% CO2. Various
concentrations of fibronectin (6.25–200nM) or an equal vol-
ume of the vehicle control, PBS, were then added to wells. In
some experiments, vitronectin (50nM), laminin (50nM), or
fibronectin fragments (400nM) were added to wells. In other
experiments, 9D2 mAb, L8 mAb, or control IgG (208nM), and
FUD or Del29 peptides (25–125nM) were added at the time of
fibronectin addition. Cells were incubated for up to 6 days at
378C and 8% CO2, and cell number was determined using 3-
(MTT; USB).29Briefly, MTT was added to a final concentra-
tion of 5.3mM per well and cells were incubated for 4h. The
medium was removed and collagen gels were digested with
collagenase (from Clostridium histolyticum, type-I; Sigma) at a
final concentration of 206 units/mL. Upon digestion of the
collagen gels, acidified isopropanol (0.04N HCl) was added to
dissolve the formazan crystals. Absorbance values were ob-
tained using a spectrophotometer, and the absorbance at the
reference wavelength (700nm) was subtracted from the ab-
sorbance at the test wavelength (570nm). Relative cell number
was determined by subtracting the average absorbance ob-
tained from ‘‘no-cell’’ collagen gels from absorbances obtained
with cell-seeded collagen gels.
Bright-field images were obtained using an Olympus
BX60 microscope and photographed using a Spot digital
camera (Diagnostic Instruments). To obtain single-channel
fluorescent images, FN-null MEFs were seeded on poly-
merized collagen gels (5.26?104cells/cm2) in 35mm tissue
culture dishes. Four hours after seeding, fibronectin (25nM)
was added to wells and cells were incubated for 6 days. Cells
were fixed and permeabilized with 48C acetone/methanol
(1:1) for 8min at ?208C. Fibronectin was observed using an
anti-FN polyclonal antibody followed by an Alexa488-labeled
goat anti-rabbit secondary antibody. After staining, cells
were examined with an Olympus BX61W confocal micro-
scope equipped with an Olympus PlanF1 immersion objec-
tive (40?, 0.8 numerical aperture).30Fluorescence images
were acquired by illuminating the sample with a 20-mW
argon laser and imaging with a Nipkow disk confocal head
and intensified charge-coupled device camera (XR Mega 10;
Stanford Photonics). Images were recorded to a DVD re-
corder and processed offline using ImageJ (NIH).
Multi-channel imaging within microtissues was per-
formed using multiphoton microscopy. To observe actively
proliferating cells, BrdU (100mM; BD Biosciences) was added
to wells on day 1, 18h before fixation. Cells were then fixed
with 4% paraformaldehyde and permeabilized with 0.05%
Triton X-100. Fibronectin was observed using an anti-FN
polyclonal antibody followed by Alexa488-labeled goat anti-
rabbit secondary antibodies. Cells were co-stained with an
anti-BrdU monoclonal antibody and Alexa594-labeled goat
anti-mouse secondary antibodies. Cell nuclei were labeled
with 40,60-diamidino-2-phenylindole (DAPI) (60nM). Images
were collected using an Olympus Fluoview 1000AOM-MPM
microscope equipped with a 25?, 1.05NA water immersion
lens (Olympus). BrdU, labeled with Alexa594, and DAPI were
simultaneously excited at 780nm using a femtosecond Mai
Tai HP Deep See Ti:Sa laser (Spectra-Physics). The emitted
3806 SEVILLA ET AL.
collagen gels (1mg/mL) and allowed to adhere for 4h. (A) Cells were treated with increasing concentrations of fibronectin
(FN; 0–200nM). Cell number was determined on day 6. Data are presented as fold increase in cell number over 0nM
FN?standard error of the mean (SEM) of three experiments performed in triplicate. *Significantly different from 0nM FN,
ANOVA, p<0.05. (B) Cells were treated with 100nM fibronectin or an equal volume of the vehicle control, PBS. Cell number
was determined at various times. Data are presented as fold increase in cell number over 0h?SEM of three experiments
performed in triplicate. *Significantly different from PBS, Student’s t-test, p<0.05. (C) Representative phase-contrast images
of cells treated with either 100nM PBS or an equal volume of PBS for various times. Images were acquired at 0, 24, 72, and
144h. Scale bar, 50mm. ANOVA, analysis of variance; FN-null MEFs, fibronectin-null mouse embryonic fibroblast, PBS,
Effect of fibronectin on cell proliferation and microtissue formation. FN-null MEFs were seeded on polymerized
3808 SEVILLA ET AL.
fluorescence was separated using a dichroic mirror (505nm
long-pass), a 609nm bandpass filter (#FF01-609/54–25;
Semrock) for BrdU, a 460nm bandpass filter (#FF01-460–80;
Semrock) for DAPI, and detected using two bi-alkaline
photomultiplier tubes. Fibronectin labeled with Alexa488was
excited at 900nm and filtered with a 519nm bandpass filter
(#BA495-546; Olympus) using the same apparatus.
Tissue body heights were measured using a confocal mi-
croscope and CCD camera (Dage-MTI CD72). The fine ad-
justment knob on the microscope was calibrated to 1mm/
gradation and the calibration was verified using microspheres
(8mm mean diameter?0.11mm standard deviation; Bangs
Laboratories). Tissue body heights were measured by re-
cording the number of gradations needed to move from the
plane of focus at the cell–collagen gel interface to the peak of
the tissue body. For each experiment, 35mm tissue culture
dishes were divided into three equal regions, and heights from
10 isolated tissue bodies within each region were measured.
Growth assays and tissue body height measurements are
expressed as mean?standard error of the mean (SEM) and
represent one of at least three separate experiments per-
formed in triplicate. Statistical comparisons were performed
using either one-way analysis of variance followed by Tu-
key’s post test or Student’s t-test for unpaired samples, as
appropriate. Results were considered statistically significant
Fibronectin increases cell proliferation and induces
tissue body formation
Cell proliferation is an integrated response to physical and
chemical signals arising from cell adhesion to ECM proteins.
Both the macromolecular composition and 3D structure of
complex ECMs influence cell proliferative responses. Pre-
vious studies have shown that fibronectin enhances the
growth rate of cells seeded onto monomeric collagen.6To
determine the effect of fibronectin on the proliferation of cells
adherent to polymeric collagen, FN-null MEFs were seeded
on 2-mm-thick gels of native collagen and cultured in the
absence or presence of fibronectin. FN-null MEFs do not
express fibronectin and have been adapted to grow under
serum-free conditions.6Thus, the use of FN-null MEFs in the
current study allows us to characterize cell behavior in the
complete absence of fibronectin and to distinguish the effects
of soluble fibronectin from ECM fibronectin. Addition of fi-
bronectin to FN-null MEFs adherent to native collagen gels
resulted in an increase in cell number with increasing con-
centrations of fibronectin (Fig. 1A). Beginning with a fibro-
nectin concentration of 25nM, there was a statistically
significant increase in cell number compared to control
samples (Fig. 1A).
To quantify fibronectin-induced cell growth as a function
of time, collagen-adherent cells were treated with either
100nM fibronectin or the vehicle control, PBS. Cell number
was determined on each of 7 days. FN-null MEFs adherent to
polymerized collagen and cultured in the absence of fibro-
nectin did not proliferate over the 144h period (Fig. 1B;
þPBS). In contrast, differences in cell number between
vehicle-control and fibronectin-treated cells were significant
beginning 72h after fibronectin addition (Fig. 1B; þFN). At
144h after fibronectin addition, there was an *2-fold in-
crease in cell number compared to PBS controls (Fig. 1B).
Large networks of well-spread cells formed on the surface of
collagen gels within 24h of fibronectin addition (Fig. 1C). By
72h, fibronectin had induced the formation of 3D, multicel-
lular structures (Fig. 1C). Taken together, these data indicate
that fibronectin stimulates cell spreading, proliferation, and
microtissue formation of cells adherent to polymerized type I
Microtissue morphology is dependent
on fibronectin concentration
Effects of fibronectin on microtissue morphology were
assessed initially by phase-contrast microscopy. Addition of
fibronectin to FN-null MEFs adherent to collagen gels re-
sulted either in the formation of compact 3D structures,
termed tissue bodies (Fig. 2A, 12.5–50nM fibronectin), or
flattened sheet-like structures with broad surface areas (Fig.
2A, 100 and 200nM fibronectin), depending on fibronectin
treatment concentration. Addition of 6.25nM fibronectin to
collagen-adherent FN-null MEFs had no effect on cell mor-
phology compared to the vehicle control, PBS (compare Fig
2A with Fig. 1C; þPBS). To quantify the effect of fibronectin
concentration on tissue body height, cells were fixed on day
6 and heights were measured using confocal microscopy.
Tissue body height exhibited a biphasic response to fibro-
nectin with peak height occurring at 25nM fibronectin (Fig.
2B). As the concentration of fibronectin was increased from
25 to 200nM, a dose-dependent decrease in tissue body
height occurred (Fig. 2B). The average height of the flattened
sheet-like structures formed in response to 200nM fibro-
nectin was *3 times that observed in the absence of fibro-
nectin (0nM FN; Fig. 2B), indicating that these sheet-like
structures are composed of multiple layers of cells. Cell
height in the presence of 6.25nM fibronectin was not sig-
nificantly different from that observed in the absence of fi-
bronectin (Fig. 2B). The contrasting effects of fibronectin
concentration on cell proliferation (dose dependence, Fig.
1A) and tissue body height (biphasic, Fig. 2B) indicate that
the extent of vertical height is not determined solely by the
extent of cell proliferation.
collagen gels for 4h. Cells were then treated with various concentrations of fibronectin (FN; 0–200nM). (A) Representative
phase-contrast images of cells on day 6. Scale bar, 50mm. (B) Heights of tissue bodies were determined on day 6. Data are
presented as average height?SEM and represent one of two experiments performed in triplicate. *Significantly different
from 0nM FN, ANOVA, p<0.05. (C) Collagen-adherent cells were treated with 25nM fibronectin and tissue body heights
were determined at various times. Data are presented as average height?SEM and represent one of two experiments
performed in triplicate. *Significantly different from 0h, ANOVA, p<0.05.
Effect of fibronectin concentration on height of tissue bodies. FN-null MEFs were allowed to adhere to polymerized
To measure the vertical expansion of fibronectin-induced
tissue bodies with respect to time, FN-null MEFs were see-
ded on polymerized collagen gels and treated with 25nM
fibronectin to induce tissue body formation. Tissue body
height was measured over the course of 6 days. A monotonic
increase in tissue body height was observed as time in-
creased (Fig. 2C). A significant increase in tissue body height
was observed within 48h of fibronectin addition (Fig. 2C).
Tissue bodies assembled in response to a single treatment of
25nM fibronectin reached a peak height of *50mm within 5
days (Fig. 2C).
Intact fibronectin is required for cell proliferation
and tissue body formation
To determine whether stimulation of cell proliferation and
tissue body formation is a unique response to fibronectin, the
effects of other ECM proteins on proliferation and tissue
body formation were assessed. Incubation of collagen-
adherent FN-null MEFs with equal molar concentrations of
either vitronectin or laminin had no effect on cell prolifera-
tion (Fig. 3A). In addition, neither vitronectin nor laminin
triggered the formation of tissue bodies (Fig. 3B). These data
suggest that cell proliferation and tissue body formation on
polymerized collagen I is not a general property of ECM
Ligation of integrin receptors by fibronectin fragments can
stimulate actin cytoskeleton organization and activate intra-
cellular signaling events that control cell proliferation.31To
determine whether fibronectin fragments stimulate the pro-
liferation of cells adherent to polymeric collagen, FN-null
MEFs were seeded on polymerized collagen gels in the ab-
sence or presence of intact fibronectin or equal molar con-
(integrin-binding), or 160/180-kDa (collagen- and integrin-
binding) fibronectin fragments. Incubation of cells with fi-
bronectin fragments had no effect on cell proliferation (Fig.
4A), and did not promote cell spreading or tissue body for-
mation (Fig. 4B). These data indicate that the interaction of
cells with integrin- and/or collagen-binding fragments of
fibronectin alone is not sufficient to stimulate cell prolifera-
tion or tissue body formation.
Cells in tissue bodies assemble an extensive
3D fibronectin matrix
The ability of cells to spread and assemble a fibrillar fi-
bronectin matrix is influenced by the rigidity of the under-
polymerized collagen gels (1mg/mL) were treated with 50nM sp. fibronectin (FN), vitronectin (VN), or sp. laminin (LN), or
an equal volume of PBS. (A) Cell number was determined on day 6. Data are presented as mean fold increase in cell number
over PBS control?SEM of three experiments performed in triplicate. *Significantly different from þPBS, ANOVA, p<0.05.
(B) Representative phase-contrast images of cells after 6 days. Scale bar, 50mm.
Cell proliferation and tissue body formation in response to extracellular matrix proteins. FN-null MEFs adherent to
3810SEVILLA ET AL.
merized collagen gels (1mg/mL) were treated with either 200nM fibronectin or equal molar concentrations of various FN
fragments. (A) Cell number was determined on day 6. Data are presented as mean increase in cell number over PBS
control?SEM of three experiments performed in triplicate. *Significantly different from þPBS, ANOVA, p<0.05. (B) Rep-
resentative phase-contrast images of cells after 6 days in culture. Scale bar, 50mm.
Effect of fibronectin fragments on cell proliferation and tissue body formation. FN-null cells adherent to poly-
polymerized collagen gels (1mg/mL) were treated with 25nM fibronectin. After a 6 day incubation, gels were fixed and
permeabilized. The fibronectin matrix was observed along the z-axis at 10mm step sizes using a confocal microscope. Arrows
denote elongated fibronectin fibrils. Arrowheads indicate short, stitch-like fibrils. Images represent one of three experiments
performed. Scale bar, 50mm.
Cells in tissue bodies polymerize an extensive three-dimensional fibronectin matrix. FN-null MEFs adherent to
MATRIX-INDUCED MICROTISSUES 3811
lying substrate.32,33To determine whether FN-null MEFs
adherent to polymerized collagen gels can assemble soluble
fibronectin into matrix fibrils, the formation of fibronectin
fibrils within tissue bodies was assessed using confocal im-
munofluorescence microscopy. As shown in Figure 5, fibro-
nectin staining was observed throughout the tissue body.
Elongated fibronectin matrix fibrils were localized to the
periphery of the tissue bodies and were observed primarily
at the interface of the tissue body and the collagen gel (Fig. 5;
20 and 30mm; arrows). Fibronectin staining was also visible
within the central region of the tissue body where it formed
shorter stitch-like fibrils that appeared to co-localize with
regions of cell–cell contact (Fig. 5; 30mm; arrowheads).
A comparison of the time courses for fibronectin-induced
cell growth (Fig. 1C) and tissue body height (Fig. 2C) reveals
that tissue bodies form before the onset of significant changes
in cell number, suggesting that cells assembled into tissue
bodies may be actively proliferating. BrdU incorporation and
multiphoton immunofluorescence microscopy were used to
determine the location of actively proliferating cells relative
to tissue bodies and fibronectin fibrils. Three-dimensional
reconstruction of fibronectin-induced microtissues demon-
strates that cells that had assembled into tissue bodies were
actively proliferating (Fig. 6A, B). Fibronectin staining was
clearly visible throughout the tissue body with fibronectin
fibrils extending out onto the collagen substrate (Fig. 6A) and
co-localizing with proliferating cells (Fig. 6B, arrows).
Cell proliferation and tissue body formation depend
on fibronectin matrix formation
To determine whether assembly of ECM fibronectin fibrils
is necessary for cell proliferation and tissue body formation
in response to fibronectin, studies were conducted using two
different approaches to inhibit fibronectin matrix assembly.
The 9D2 and L8 mAbs inhibit fibronectin matrix polymeri-
zation without affecting the initial association of fibronectin
with the cell surface,28or fibronectin–integrin interactions.7
The 49-residue peptide (FUD) from Adhesin F1 of Strepto-
coccus pyogenes inhibits fibronectin matrix assembly by
blocking the binding of the amino-terminus of fibronectin to
cell surfaces.27To determine whether inhibitors of fibronec-
tin matrix assembly block fibronectin-mediated cell prolif-
eration and tissue body formation, FN-null MEFs adherent to
polymerized collagen gels were treated with fibronectin in
the presence or absence of 9D2 mAbs, L8 mAbs, or FUD
peptides. Addition of 9D2 or L8 mAbs inhibited fibronectin-
dependent cell proliferation (Fig. 7A) and tissue body
formation (not shown). In contrast, nonimmune IgG had no ef-
fect on the fibronectin-mediated responses (Fig. 7A, and not
shown). Similarly, addition of FUD peptides to fibronectin-
treated, collagen-adherent cells inhibited cell proliferation
(Fig. 7B) and tissue body formation (Fig. 7C) in a dose-
dependent manner. The control peptide, Del29, did not alter
the cellular responses to fibronectin (Fig. 7B, C), demon-
strating the specificity of the FUD peptide. Addition of FUD
or Del29 peptides to collagen-adherent cells in the absence of
fibronectin had no effect on basal cell number or cell mor-
phology as compared to PBS-treated controls (Fig. 7B, C).
Addition of 125nM FUD peptides to fibronectin-treated
cells completely inhibited tissue body formation (Fig. 7C),
whereas addition of 25nM FUD altered the final morphology
adherent to polymerized collagen gels were treated with
25nM fibronectin. After a 24h incubation, BrdU was added
and cells were incubated for an additional 18h. Fibronectin
(green) was observed using an anti-FN pAb followed
by Alexa488-conjugated anti-rabbit IgG. Proliferating (BrdU-
positive) cells (red) were observed using an anti-BrdU
mAb followed by Alexa594-conjugated anti-mouse IgG. Cell
nuclei were stained with DAPI (blue). (A) Fibronectin, BrdU,
and DAPI were observed along the z-axis at 1mm step sizes
using two-photon microscopy. Collected images were then
projected onto the z-plane using ImageJ. Scale bar, 50mm. (B)
Merged image showing fibronectin, DAPI, and BrdU stain-
ing along one slice of the tissue body shown in A. Arrows
point to fibronectin-BrdU colocalization. Scale bar, 50mm.
Cell proliferation in tissue bodies. FN-null MEFs
3812 SEVILLA ET AL.
polymerized type I collagen. After 4h, cells were treated with either PBS or fibronectin (FN; 25nM). In (A), 9D2, L8, or
nonimmune mouse IgG (208nM) was added to fibronectin-treated wells. Data presented as fold increase in cell number over
PBS?SEM of three experiments performed in triplicate. *Significantly different from PBS, ANOVA, p<0.001. (B) FUD or
Del29 (25 and 125nM) peptides were added to fibronectin- and PBS-treated wells. Data are presented as fold increase in cell
number over PBS?SEM and represent one of three experiments performed in triplicate. *Significantly different from PBS,
ANOVA, p<0.05. (C) Representative phase-contrast images of cells treated with either PBS or FN in the absence or presence
of FUD or Del29 peptides. Images were collected on day 6 of culture and represent one of three experiments performed. Scale
bar, 50mm. FUD, functional upstream domain.
Cell proliferation and tissue body formation depend on FN matrix polymerization. FN-null MEFs were seeded on
of tissue bodies and led to the formation of smaller, more
compact structures (Fig. 7C). These data indicate that fibro-
nectin matrix polymerization is essential for promoting pro-
liferation and self-assembly of cells adherent to polymerized
Fibronectin-induced cell proliferation and tissue body
formation are influenced by collagen organization
To determine whether the macromolecular organization of
the collagen substrate is important for tissue body formation
in response to fibronectin, FN-null MEFs were seeded onto
either polymerized or monomeric collagen in the presence of
fibronectin. Addition of fibronectin to cells adherent to po-
lymerized collagen induced the formation of tissue bodies
(Fig. 8). In contrast, addition of fibronectin to cells adherent
to monomeric collagen resulted in monolayer growth of cells
(Fig. 8). As reported previously,6fibronectin enhanced the
growth rate of FN-null MEFs adherent to monomeric colla-
gen (data not shown). Together, these data indicate that fi-
substrate collagen structure, whereas formation of 3D tis-
sue bodies in response to fibronectin requires polymeric
Increasing the concentration of collagen in polymerized
collagen gels increases the linear modulus, failure stress, and
fibril density, without affecting fibril diameter.34To deter-
mine how substrate properties of polymerized collagen affect
is independent of
fibronectin-induced cell proliferation and tissue body for-
mation, FN-null MEFs were seeded on polymerized collagen
gels of increasing collagen concentration (0.25–2.5mg/mL),
in the absence or presence of 25nM fibronectin. In the ab-
sence of fibronectin, cell proliferation was unaffected by in-
creasing collagen concentration (Fig. 9A). In contrast,
increasing the collagen concentration of the polymerized gel
decreased fibronectin-induced cell proliferation in a dose-
dependent manner (Fig. 9A). Similarly, the morphology (Fig.
9B) and height (Fig. 9C) of fibronectin-induced tissue bodies
were dependent on collagen concentration. Tissue bodies
formed on 0.25mg/mL collagen gels were characterized by
broad extensions surrounding a central 3D structure (Fig.
9B). In contrast, tissue bodies formed on 2.5mg/mL collagen
gels formed compact structures with few cell extensions onto
the collagen substrate (Fig. 9C). Compared to tissue bodies
formed on 1mg/mL collagen, tissue body height was less on
gels composed of 0.25 and 0.5mg/mL collagen (Fig. 9C),
consistent with the more flattened morphology observed in
Figure 9B. Although the final shape of tissue bodies differed
as collagen concentrations increased from 1.0 to 2.5mg/mL,
the vertical heights were not different (Fig. 9C). A compari-
son of Figure 9A and C at 2.5mg/mL collagen indicates that
fibronectin induced the formation of 3D structures in the
absence of cell proliferation.
Thus far, our data indicate that increasing concentrations
of substrate collagen inhibit the growth-promoting effects of
fibronectin fibrils. To determine whether increasing the
concentration of exogenous fibronectin could reverse the
inhibitory effects of high concentrations of substrate colla-
gen, FN-null MEFs were seeded on 2.5mg/mL collagen gels
and treated with up to 800nM fibronectin. Addition of in-
creasing concentrations of fibronectin to cells on 2.5mg/mL
collagen gels resulted in a dose-dependent increase in cell
proliferation over basal levels (Fig. 10A) and similarly, re-
versed the morphological effects of high collagen concen-
trations, leading to tissue bodies with progressively flatter
morphologies (Fig. 10B).
In this study, we have used FN-null MEFs adherent to
native collagen gels to investigate the functional role of
fibronectin fibrillogenesis on cell proliferation, cellular self-
assembly, and microtissue morphology. Our data demon-
strate that fibronectin specifically stimulates proliferation
and self-assembly of cells adherent to native collagen fibrils
into tissue-like structures by a mechanism that requires fi-
bronectin matrix assembly. These responses were specific to
intact fibronectin, as neither fibronectin fragments nor other
ECM proteins triggered microtissue formation. Fibronectin-
induced tissue body formation occurred in the absence of cell
microtissue formation. FN-null MEFs cells were seeded on
either thick gels of polymerized collagen or thin coats of
monomeric collagen, and incubated with fibronectin (25nM)
for 4 days. Representative phase-contrast images are shown.
Scale bar, 50mm.
Fibrillar collagen is required for fibronectin-induced
meric collagen gels of increasing collagen concentrations (0.25–2.5mg/mL), were treated with fibronectin (FN; 25nM) or an
equal volume of PBS. (A) Cell number was determined on day 6. Absorbance values were normalized to values obtained for
þPBS on 0.25mg/mL collagen. Data are presented as average increase in cell number?SEM of three experiments performed
in triplicate. *Significantly different from respective þPBS control, p<0.05. (B) Representative phase-contrast images of cells
after 6 day in culture. Scale bar, 50mm. (C) Heights of tissue bodies were determined using confocal microscopy. Data are
presented as average height?SEM and represent one of three experiments performed in triplicate. *Significantly different
from 1mg/mL collagen, ANOVA, p<0.05.
Cell proliferation and tissue body formation depend on collagen concentration. FN-null MEFs, adherent to poly-
3814 SEVILLA ET AL.
proliferation, indicating that proliferation and cellular self-
assembly are independent responses to fibronectin matrix
Unlike tissue spheroids, in which balls of compacted cells
form floating aggregates,15cells at the collagen gel interface
of fibronectin-induced tissue bodies were adherent and well
spread on the collagen substrate. Altering either fibronectin
or collagen concentration influenced the extent of cell pro-
liferation as well as the ultimate shape of the microtissue.
Increasing fibronectin concentrations induced a monotonic
MEFs, adherent to 2.5mg/mL polymerized collagen gels, were treated with increasing concentrations of fibronectin (25–800nM),
or an equal volume of PBS (0). (A) Cell number was determined on day 6. Absorbance values were normalized to values obtained
for 0nM fibronectin. Data are presented as an average increase in cell number?SEM of three experiments performed in triplicate.
Increasing fibronectin concentrations reverse effects of collagen on cell proliferation and microtissue shape. FN-null
3816SEVILLA ET AL.
increase in cell proliferation, but had a biphasic effect on
microtissue height, where higher fibronectin concentrations
produced flatter, sheet-like structures with broad bases. In
contrast, increasing collagen concentrations resulted in a
dose-dependent inhibition of fibronectin-induced cell prolif-
eration, but increased tissue body height, producing mono-
lith-like structures with small bases. Taken together, these
data indicate that the relative proportion of collagen and
fibronectin fibrils polymerized into the ECM influences the
extent of cell proliferation and the final shape of micro-
Cellular self-assembly in response to matrix fibronectin
occurred only on polymeric collagen gels and not on mo-
nomeric collagen substrates, implying that the macromolec-
ular organization of collagen fibrils plays a permissive role in
fibronectin-induced microtissue formation. Indeed, results
from our study provide support for the concept that cellular
self-assembly may be controlled through local differences in
adhesive and anti-adhesive forces that occur between cells
and their substratum. A similar phenomenon was reported
recently for bovine aortic endothelial cells wherein decreased
cell–substrate adhesivity corresponded to increased cell–cell
adhesion and endothelial network formation.18Others have
demonstrated that decreasing substrate rigidity permits the
concurrent development of cell–matrix and cell–cell adhe-
sions, while increasing substrate rigidity favors cell–matrix
adhesions.19,35As such, the biphasic effect of fibronectin on
tissue body height observed in the current study may be a
consequence of enhanced rigidity or adhesivity in response
to increased fibronectin fibril formation that, in turn, in-
creased cell–matrix adhesions and decreased cell–cell adhe-
sions, leading to flatter structures.
In contrast to the biphasic effect of fibronectin concentra-
tion on tissue body morphology (Fig. 2), increasing the col-
lagen concentration of the polymerized substrate reduced
fibronectin-mediated cell spreading and increased tissue
body height, leading to more compact structures (Fig. 9).
Increasing the collagen concentration of polymerized type I
collagen gels increases both the stiffness of the gel and the
density of the collagen fibrils, but does not affect collagen
fibril diameter.34Others have shown that increasing sub-
strate stiffness while maintaining ligand density promotes
cell spreading and reduces cellular self-assembly.19On the
other hand, peak cell spreading on compliant collagen gels
occurs at intermediate collagen densities36and is reduced at
high collagen fibril densities.37Hence, in the current study,
the reduction in tissue body area that occurred in response to
increasing collagen concentrations was likely due to effects of
increasing collagen fibril density, which in our model, pre-
dominated over the expected cellular response to increasing
substrate stiffness. Importantly, these data indicate that cel-
lular self-organization is an integrated response to chemical
and mechanical signals arising from both fibronectin and
collagen fibrils in the ECM. As such, it may be possible to
direct the formation and shape of engineered tissues by
spatially and/or temporally manipulating the composition
and organization of the supporting ECM.
An extensive fibronectin matrix was polymerized by cells
throughout the microtissues and was not limited to areas in
direct contact with the collagen substrate or to well-spread
cells. At least two distinct morphologies of fibronectin ma-
trices were associated with tissue bodies. Elongated, fibrillar
fibronectin appeared localized to outer regions of the tissue
body and co-localized with proliferating cells. Fibrillar fi-
bronectin staining was also observed in tracks leading away
from tissue bodies, suggesting active matrix polymerization
by cells during migration to the central structure. In addition,
a pericellular form of the fibronectin matrix localized to
central regions of the tissue bodies and was associated with
nonproliferating cells. We hypothesize that regional varia-
tions in the organizational patterns of fibronectin fibrils, due
in part to local variations in tension or rigidity, may give rise
to distinct fibronectin matrices and, in turn, distinct cell
Smooth muscle cells and fibroblasts seeded on native
collagen fibrils fail to spread and display a reduced prolif-
erative capacity associated with increased expression of
growth inhibitory signaling molecules.38,39Similarly, in the
present study, FN-null MEFs seeded on polymeric collagen
gels in the absence of fibronectin failed to spread and did not
proliferate. In contrast, initiation of cell-mediated fibronectin
matrix polymerization overcame the inhibitory effect of po-
lymerized collagen and promoted cell proliferation. The
mechanisms by which fibronectin matrix polymerization
initiates cell proliferation on fibrillar collagen are not known.
Growth-promoting intracellular signaling events initiated by
cell adhesion to fibronectin fibrils40may simply override
growth-inhibitory signals induced by fibrillar collagen. Al-
ternatively, binding of fibronectin to collagen41may physi-
cally insulate growth-inhibitory epitopes of fibrillar collagen
from cells or, conversely, form a pro-migratory complex that
permits cell migration over the collagen substrate8to initiate
Several cell types, including endothelial cells, fibroblasts,
and myocytes, sense and respond to substrate rigidity. For
example, fibroblasts adherent to flexible collagen-coated
polyacrylamide gels show reduced cell spreading and higher
rates of motility than cells on more rigid collagen-coated
substrates, which spread and form stable focal contacts.33
Further, substrate rigidity and intracellular cytoskeletal ten-
sion generation strongly influence fibronectin fibril forma-
tion.32,42In turn, the formation of a fibronectin matrix
enhances actin cytoskeletal tension13,43and increases the
mechanical tensile properties of cell-embedded collagen
gels.14In the current study, the relative proportion of colla-
gen and fibronectin fibrils polymerized into the ECM influ-
enced the formation and shape of microtissues, with
increasing fibronectin levels leading to progressively flatter
structures. These studies indicate that the local balance of
ECM-derived forces, influenced in part by the extent of fi-
bronectin fibril formation, contributes to the microenviron-
ment of the cell to locally influence cellular behaviors
essential for tissue morphogenesis.
This work was supported by Grants EB008368 and
EB008996 and from the National Institutes of Health. C.A.S.
was supported by NIH predoctoral fellowship F31AR057675.
We gratefully acknowledge the support of Dr. Karl Ka-
sischke and the URMC Multiphoton Core Facility.
No competing financial interests exist.
1. Lauffenburger, D.A., and Griffith, L.G. Who’s got pull
around here? Cell organization in development and tissue
engineering. Proc Natl Acad Sci USA 98, 4282, 2001.
2. Hynes, R.O., and Yamada, K.M. Fibronectins: multifunc-
tional modular glycoproteins. J Cell Biol 95, 369, 1982.
3. George, E.L., Georges-Labousse, E.N., Patel-King, R.S.,
Rayburn, H., and Hynes, R.O. Defects in mesoderm, neural
tube and vascular development in mouse embryos lacking
fibronectin. Development 119, 1079, 1993.
4. Magnusson, M., and Mosher, D.F. Fibronectin: structure,
assembly, and cardiovascular implications. Arterioscler
Thromb Vasc Biol 18, 1363, 1998.
5. Gui, L., Wojciechowski, K., Gildner, C.D., Nedelkovska, H.,
and Hocking, D.C. Identification of the heparin-binding
determinants within fibronectin repeat III1: role in cell
spreading and growth. J Biol Chem 281, 34816, 2006.
6. Sottile, J., Hocking, D.C., and Swiatek, P.J. Fibronectin ma-
trix assembly enhances adhesion-dependent cell growth. J
Cell Sci 111, 2933, 1998.
7. Hocking, D.C., and Chang, C.H. Fibronectin polymerization
regulates small airway epithelial cell migration. Am J Phy-
siol Lung Cell Mol Physiol 285, L169, 2003.
8. Sottile, J., Shi, F., Rublyevska, I., Chiang, H.Y., Lust, J., and
Chandler, J. Fibronectin-dependent collagen I deposition
modulates the cell response to fibronectin. Am J Physiol Cell
Physiol 293, C1934, 2007.
9. Robinson, E.E., Foty, R.A., and Corbett, S.A. Fibronectin
matrix assembly regulates alpha5beta1-mediated cell cohe-
sion. Mol Biol Cell 15, 973, 2004.
10. Salmenpera, P., Kankuri, E., Bizik, J., Siren, V., Virtanen, I.,
Takahashi, S., et al. Formation and activation of fibroblast
spheroids depend on fibronectin-integrin interaction. Exp
Cell Res 314, 3444, 2008.
11. Rozario, T., Dzamba, B., Weber, G.F., Davidson, L.A., and
DeSimone, D.W. The physical state of fibronectin matrix
differentially regulates morphogenetic movements in vivo.
Dev Biol 327, 386, 2009.
12. Sottile, J., and Hocking, D.C. Fibronectin polymerization
regulates the composition and stability of extracellular ma-
trix fibrils and cell-matrix adhesions. Mol Biol Cell 13, 3546,
13. Hocking, D.C., Sottile, J., and Langenbach, K.J. Stimulation
of integrin-mediated cell contractility by fibronectin poly-
merization. J Biol Chem 275, 10673, 2000.
14. Gildner, C.D., Lerner, A.L., and Hocking, D.C. Fibronectin
matrix polymerization increases the tensile strength of a
model tissue. Am J Physiol Heart Circ Physiol 287, H46, 2004.
15. Mironov, V., Visconti, R.P., Kasyanov, V., Forgacs, G.,
Drake, C.J., and Markwald, R.R. Organ printing: tissue
spheroids as building blocks. Biomaterials 30, 2164, 2009.
16. Duguay, D., Foty, R.A., and Steinberg, M.S. Cadherin-
mediated cell adhesion and tissue segregation: qualitative
and quantitative determinants. Dev Biol 253, 309, 2003.
17. Dean, D.M., and Morgan, J.R. Cytoskeletal-mediated tension
modulates the directed self-assembly of microtissues. Tissue
Eng Part A 14, 1989, 2008.
18. Califano, J.P., and Reinhart-King, C.A. A balance of chem-
istry and mechanics regulates endothelial network forma-
tion. Cell Mol Bioeng 1, 122, 2008.
19. Guo, W.H., Frey, M.T., Burnham, N.A., and Wang, Y.L.
Substrate rigidity regulates the formation and maintenance
of tissues. Biophys J 90, 2213, 2006.
20. Ryan, P.L., Foty, R.A., Kohn, J., and Steinberg, M.S. Tissue
spreading on implantable substrates is a competitive out-
come of cell-cell vs. cell-substratum adhesivity. Proc Natl
Acad Sci USA 98, 4323, 2001.
21. McDaniel, D.P., Shaw, G.A., Elliott, J.T., Bhadriraju, K.,
Meuse, C., Chung, K.H., et al. The stiffness of collagen fibrils
influences vascular smooth muscle cell phenotype. Biophys J
92, 1759, 2007.
22. Morla, A., Zhang, Z., and Ruoslahti, E. Superfibronectin is a
functionally distinct form of fibronectin. Nature 367, 193, 1994.
isolationof human plasmafibronectin. Thromb Res 27,1, 1982.
24. Hocking, D.C., Smith, R.K., and McKeown-Longo, P.J. A
novel role for the integrin-binding III-10 module in fibro-
nectin matrix assembly. J Cell Biol 133, 431, 1996.
25. Wojciechowski, K., Chang, C.H., and Hocking, D.C. Ex-
pression, production, and characterization of full-length vi-
tronectin in Escherichia coli. Protein Expr Purif 36, 131, 2004.
26. Windsor, L.J., Havemose-Poulson, A., Yamada, S., Lyons,
J.G., Birkedal-Hansen, B., Stetler-Stevenson, W.G., et al.
Matrix Metalloproteinases. New York: Wiley & Sons, Inc.,
2002, pp. 10.8.6–10.8.7.
27. Tomasini-Johansson, B.R., Kaufman, N.R., Ensenberger,
M.G., Ozeri, V., Hanski, E., and Mosher, D.F. A 49-residue
peptide from adhesin F1 of Streptococcus pyogenes inhibits
fibronectin matrix assembly. J Biol Chem 276, 23430, 2001.
28. Chernousov, M.A., Fogerty, F.J., Koteliansky, V.E., and
Mosher, D.F. Role of the I-9 and III-1 modules of fibronectin
in formation of an extracellular matrix. J Biol Chem 266,
29. Mosmann, T. Rapid colorimetric assay for cellular growth
and survival: application to proliferation and cytotoxicity
assays. J Immunol Methods 65, 55, 1983.
30. Sumagin, R., and Sarelius, I.H. A role for ICAM-1 in main-
tenance of leukocyte-endothelial cell rolling interactions in
inflamed arterioles. Am J Physiol Heart Circ Physiol 293,
31. Howe, A., Aplin, A.E., Alahari, S.K., and Juliano, R.L. In-
tegrin signaling and cell growth control. Curr Opin Cell Biol
10, 220, 1998.
32. Halliday, N.L., and Tomasek, J.J. Mechanical properties of
the extracellular matrix influence fibronectin fibril assembly
in vitro. Exp Cell Res 217, 109, 1995.
33. Pelham, R.J., Jr., and Wang, Y.-L. Cell locomotion and focal
adhesions are regulated by substrate flexibility. Proc Natl
Acad Sci USA 94, 13661, 1997.
34. Roeder, B.A., Kokini, K., Sturgis, J.E., Robinson, J.P., and
Voytik-Harbin, S.L. Tensile mechanical properties of three-
dimensional type I collagen extracellular matrices with
varied microstructure. J Biomech Eng 124, 214, 2002.
35. Tsai, J., and Kam, L. Rigidity-dependent cross talk between
integrin and cadherin signaling. Biophys J 96, L39, 2009.
36. Engler, A., Bacakova, L., Newman, C., Hategan, A., Griffin,
M., and Discher, D. Substrate compliance versus ligand
density in cell on gel responses. Biophys J 86, 617, 2004.
37. Elliott, J.T., Tona, A., Woodward, J.T., Jones, P.L., and Plant,
A.L. Thin films of collagen affect smooth muscle morphol-
ogy. Langmuir 19, 1506, 2003.
38. Koyama, H., Raines, E.W., Bornfeldt, K.E., Roberts, J.M., and
Ross, R. Fibrillar collagen inhibits arterial smooth muscle
proliferation through regulation of Cdk2 inhibitors. Cell 87,
39. Schor, S.L. Cell proliferation and migration on collagen
substrata in vitro. J Cell Sci 41, 159, 1980.
3818SEVILLA ET AL.
40. Danen, E.H., and Yamada, K.M. Fibronectin, integrins, and
growth control. J Cell Physiol 189, 1, 2001.
41. Ingham, K.C., Brew, S.A., and Isaacs, B.S. Interaction of fi-
bronectin and its gelatin-binding domains with fluorescent-
labeled chains of type I collagen. J Biol Chem 263, 4624, 1988.
42. Zhong, C., Chrzanowska-Wodnicka, M., Brown, J., Shaub,
A., Belkin, A.M., and Burridge, K. Rho-mediated contractil-
ity exposes a cryptic site in fibronectin and induces fibro-
nectin matrix assembly. J Cell Biol 141, 539, 1998.
43. Corbett, S.A., and Schwarzbauer, J.E. Requirements for a5b1
integrin-mediated retraction of fibronectin-fibrin matrices. J
Biol Chem 274, 20943, 1999.
Address correspondence to:
Denise C. Hocking, Ph.D.
Department of Pharmacology and Physiology
University of Rochester
601 Elmwood Ave.
Rochester, NY 14642
Received: May 26, 2010
Accepted: July 19, 2010
Online Publication Date: September 3, 2010