4
Cell–Nanostructure Interactions

Joachim P. Spatz

4.1

Introduction

Cell–cell and cell–extracellular matrix (ECM) adhesion is a complex, highly regulated pro-
cess which plays a crucial role in most fundamental cellular functions including motility,
proliferation, differentiation, and apoptosis [1, 2]. Focal adhesions are the primary cellular
domains responsible for surface adhesion. These are complex multimolecular assemblies
consisting of transmembrane proteins, the integrin receptors, and cytoplasmic proteins,
such as vinculin, paxillin, and focal adhesion kinase (for an overview, see Figure 4.1)
[3–9]. Cell binding to the ECM results in local accumulation of integrins, cytoplasmic pro-
teins, which form focal adhesion clusters (FAC), and reorganization of the actin cytoske-
leton, which generates forces to the underlying substrate by the work of myosin molecular
motors [10].

Despite enormous progress and challenging studies in the field of cell adhesion during

the past 50 years [3], numerous questions concerning the signaling of focal adhesion
which are based on single protein assembly remain unsolved. We have only partial knowl-
edge of the existence of hierarchical and cooperative arrangements and synergetic interac-
tions between focal adhesion proteins. Likewise, we have a very limited understanding of
the function of cell adhesion, and of the significance of focal adhesion cluster size, shape,
characteristic length scales between proteins in a FAC, and protein assembly dynamics.
Future knowledge in this area, when combined with tools that control these processes
in focal adhesion, would allow for the tuning of cell adhesion and its associated signaling,
with molecular precision.

Integrin clustering into linear objects is clearly observed using either differential inter-

ference contrast (DIC) microscopy or fluorescent optical microscopy following immuno-
histochemical staining of a specific protein involved in FAC formation [11]. However, it
is not yet known how the adhesion and signaling of cells is coordinated by integrin clus-
tering, integrin–integrin separation distances and integrin pattern geometries in cell
membranes, nor how many integrins are necessarily involved in the formation of stable
adhesion [12]. Control of these structural arrangements in the cell membrane is offered by
adhesive nanotemplates or substrate topographies, and points clearly to the value of

53

Nanobiotechnology. Edited by Christof Niemeyer, Chad Mirkin
Copyright

c 2004 WILEY-VCH Verlag GmbH & Co. K aA, Weinheim

ISBN 3-527-30658-7

G

nanostructures in an understanding of molecular dimensions and processes in focal ad-
hesion formation.

Integrins are heterodimers formed by the noncovalent association of a and b subunits;

the b subunit recognizes the RGD (Arginine-Glycine-Aspatate) motif, a sequence which is
present in many ECM proteins. The a and b tails form together a V-shaped flexible struc-

54

4 Cell–Nanostructure Interactions

Figure 4.1

Schematic depicting the complexity of

the main molecular domains of cell–matrix adhe-
sions by Geiger [3]. The primary adhesion receptors
are heterodimeric (a and b) integrins, represented
by orange cylinders. Additional membrane-asso-
ciated molecules enriched in these adhesions (red)
include syndecan-4 (Syn4), layilin (Lay), the phos-
phatase leukocyte common antigen-related receptor
(LAR), SHP-2 substrate-1 (SHPS-1) and the uro-
kinase plasminogen activator receptor (uPAR).
Proteins that interact with both integrin and actin,
and which function as structural scaffolds of focal
adhesions, include a-actinin (a-Act), talin (Tal),
tensin (Ten) and filamin (Fil), shown as golden
rods. Integrin-associated molecules in blue include:
focal adhesion kinase (FAK), paxillin (Pax), integrin-
linked kinase (ILK), down-regulated in rhabdomyo-
sarcoma LIM-protein (DRAL), 14-3-3b and caveolin
(Cav). Actin-associated proteins (green) include

vasodilator-stimulated phosphoprotein (VASP),
fimbrin (Fim), ezrin–radixin–moesin proteins
(ERM), Abl kinase, nexillin (Nex), parvin/actopaxin
(Parv) and vinculin (Vin). Other proteins, many of
which might serve as adaptor proteins, are colored
purple and include zyxin (Zyx), cysteine-rich protein
(CRP), palladin (Pall), PINCH, paxillin kinase linker
(PKL), PAK-interacting exchange factor (PIX),
vinexin (Vnx), ponsin (Pon), Grb-7, ASAP1,
syntenin (Synt), and syndesmos (Synd). Among
these are several enzymes, such as SH2-containing
phosphatase-2 (SHP-2), SH2-containing inositol
5-phosphatase-2 (SHIP-2), p21-activated kinase
(PAK), phosphatidylinositol 3-kinase (PI3K), Src-
family kinases (Src FK), carboxy-terminal src kinase
(Csk), the protease calpain II (Calp II) protein
kinase C (PKC). Enzymes are indicated by lighter
shades. For details, see Ref. [3].

ture of an estimated lateral diameter of 80–120 Å [13, 14], with the head ranging between
57–73 Å for aIIbb

3

-integrin, and

Z90 q 60 q 45 Å for a

v

b

3

-integrin [15, 16]. Ligand bind-

ing affinity is influenced by the conformational changes in the receptor caused by the ex-
tracellular environment, and also by the interactions with the cytoplasmic proteins [17].

A prominent example where spatial arrangement of RGD ligands in a defined and rigid

geometry fulfills important functions is given by the adenovirus. Its capsid contains the
penton base protein, a pentamer that promotes virus entry in cells via a

v

-integrins. In

fact, the penton base protein presents RGD sequences on the tips of a regular stiff pen-
tagon having a side length of

Z60 Å [18] (Figure 4.2). Since only the pentameric form

mediates integrin-specific adhesion of nonactivated lymphoid cells, while the monomeric
does not, the polyvalent binding of several integrins to the RGD sequences is mandatory
and assumed to be controlled by the rigidity of this protein [19]. This rigidity of the virus
template also allows the adenovirus to escape neutralization by IgG antibodies of the
immunosystem directed against RGD integrin receptor sites due to steric hindrance.
Thus, the 60 Å separation between RGD sequences demonstrates a minimal distance
necessary for integrin heads to bind simultaneously but which is too small for the IgG
to neutralize the virus.

The adenovirus is a major example where nature demonstrates regulation of cell inter-

actions with the extracellular site by single ligand pattern of specific geometry, and this
length of scale opens challenging opportunities for understanding and tailoring cell func-

55

4.1 Introduction

Figure 4.2

Left: A cryoelectron microscopy image

reconstruction of the adenovirus binding complex.
The penton base capsomers at the icosahedral
vertices are shown in yellow, the reconstructed
portion of the flexible fibers in green, the remaining
capsid density in blue and the Fab density in ma-
genta. The complex is viewed along an icosahedral
3-fold axis. The scale bar is 100 Å. Right: The ade-
novirus penton base protrusions. (A) Top view of
the penton base (yellow) and fiber (green), along

with weak protrusion density (red). (B) Side view of
the external portion of the penton base with fiber.
(C) Top view of the penton base, showing the dis-
tances between weak protrusions. (D) An enlarge-
ment of a single penton base protrusion with a loop
modeled from the crystallographic RGD peptide.
The arginine side chain is shown in blue, and the
aspartic acid side chain in red. The loops are
shaded with a transparency gradient to denote
motion. The scale bars are 25 Å [18].

tions with ultrahigh sensitivity. The formation of a regular and stiff ligand template is a
basic requirement for mimicking such molecularly defined adhesive “keys” that are set by
the spatial distribution of ligands for single integrin occupation. As described in the next
section, methods developed from nanotechnological systems already approach fidelity and
functionality to provide these requirements for control of cell adhesion on a molecular
level.

4.2

Methods

Making use of advanced opportunities from material sciences for the identification, loca-
tion, and systematic manipulation of molecular components at interfaces identifies great
potential in the development of new materials for biophysical and biochemical investiga-
tions, and particularly in the field of cell adhesion. In principle, cellular adhesion studies
on the nanoscale may be divided into the reaction of cells to variations in substrate topo-
graphy, or to the presence of a chemical contrast along a substrate. While topography in-
duces surface roughness and such greater adhesive areas, the substrate’s chemical con-
trast points to an opportunity of controlling transmembrane and intracellular molecules,
as well as protein distributions.

A group of researchers at Glasgow University demonstrated the response of cells not

only to micrometric but also to nanometric scale topography [20–22]. The formation of
nanotopography was explored using methods based on polymer demixing; that is, demix-
ing of polystyrene and polybromostyrene, where nanoscale islands of reproducible height
were fabricated. The islands were shown to affect cell spreading compared with planar
surfaces, where morphological, cytoskeletal, and molecular changes in fibroblast reaction
to 13 nm-high islands were observed. It should be noted that this topography length scale
is on the order of roughness which is induced by adhesive proteins such as fibronectin or
lamin as well as collagen fibers that are dominant within the ECM. The cellular responses
were characterized using methods such as scanning electron microscopy, fluorescence
microscopy, and gene microarray. The results showed that cells respond to the islands
by broad gene up-regulation, notably in areas of cell signaling, proliferation, cytoskeleton,
and the production of ECM proteins. Results obtained with microscopy confirmed the mi-
croarray findings and highlighted several corresponding points between cytoskeletal, mor-
phological, and genetic observations. Moreover, they also showed a synergy from the for-
mation of focal contacts to the up-regulation of genes required for fibroblast differentia-
tion. These findings indicate that increased cell attachment and spreading is required
for up-regulation proliferation and matrix synthesis.

One prerequisite when preparing model systems to demonstrate the role of spatial

ligand distributions and ligand concentration in cell adhesion in vitro is the availability
of a nonadhesive surface; this permits specific cellular responses to be attributed entirely
to the interaction with specific adhesion-mediating ligands. Polyethylene glycol (PEG) or
polyethylene oxide (PEO) -based substrates are widely used as biologically inert interfaces,
and recent developments have included the grafting of high molecular weight PEG [23]
and star-shaped PEG macromolecules to substrates [24, 25], or the use of oligo(ethylene
oxide) functionalized self-assembled monolayers (SAMs) [26, 27].

56

4 Cell–Nanostructure Interactions

The surface concentration and spatial distribution of cell-adhesive ligands in such a bio-

logically inert PEG or PEO background may be controlled statically by mixing bioactive
macrosystems with unsubstituted molecules [28], or dynamically by the electrochemical
control of ligand release [29]. The modification of inert polymers with a cell-adhesive
motif that often contains the amino acid sequence RGD (which is also found in fibronec-
tin) and its influence on cell adhesion and signaling has been recently reviewed in detail
by Kessler et al. [30].

Cell attachment to interfaces depends on many factors, including the affinity and spe-

cificity of surface-bound ligands to integrins, the mechanical strength of ligand support
and linkage, spacer length, overall ligand concentration, and ligand clustering [30]. As
an example, the number of attached cells is clearly correlated to RGD surface density,
as indicated by a sigmoidal increase with RGD concentration [31]. This suggests that
there is a minimum ligand density for cell response. As a general rule, a higher RGD sur-
face density is related to intense cell spreading, cell survival, and focal contact formation.
Ever since the early days of RGD-mediated cell adhesion, discussions have been ongoing
on the subject of how many RGD molecules are required to induce not only cell attach-
ment but also cell spreading and focal contact formation. Hubbell produced a benchmark
result by showing amounts as low as 1 fmol RGD ligands cm

–2

to be sufficient for cell

spreading, and as low as 10 fmol cm

–2

to be sufficient for the formation of focal contacts

and stress fibers [32]. In these studies, RGD molecules were covalently bound to glyco-
phase glass coverslips via NH

2

-terminal primary amines. Smart macromolecular designs

of PEG molecules such as PEG-stars allow the control of an average number of RGD li-
gands per star (one, five, or nine YGRGD peptides per star) as demonstrated by Griffith
and Lauffenburger in a series of publications and illustrated in Figure 4.3 [25, 33–35]. Cell
adhesion and movement was observed for 1000 ligands mm

–2

, or more [25]. The macromo-

lecular approach has the advantage that the large and flexible chains may account for dif-
ferent cell-binding activities that are probably caused by local enrichment of ligands at the
cell membrane and anchoring compliance. However, it may not account for control of pre-
cise ligand clustering, as the ligand template may be not well-ordered, as well as being too
flexible. This permits ligands to cluster by pure chemical affinity, or cells to arrange li-
gands at their convenience [34]. This situation does not apply to the adenovirus, where
the control of cell function is achieved by the arrangement of single ligands in patterns
on a rigid template.

57

4.2 Methods

Figure 4.3

Schematic illustration of star polymer as a tether to present ligand (shaded oval) in a manner

in which the total average concentration (left to right) and the spatial distribution, from homogeneous to
highly clustered (top versus bottom), can be independently varied (from Ref. [25]).

A clear understanding of how the adhesion and signaling of cells depend on the com-

position, size, and distribution of FACs has long been limited to patterning studies of
ligands on submicrometer patches. By using micro contact printing [36], surfaces pat-
terned with adhesive and nonadhesive domains have been prepared at scales down to
the micrometer level. These surfaces were then successfully used to control geometrically
both cell shape and viability [36]. The results of these studies indicated strongly that cell
shape and integrin distribution are able to control the survival/apoptosis of cells, and can
switch between these two basic programs of the cells.

Even smaller adhesion pattern were prepared using dip-pen nanolithography, where

cell-adhesive patches (retronectin) of 200 nm diameter and 700 nm separation still
showed attachment of cells (Figure 4.4) (see also Chapter 19) [37].

The control of defined spacing between adhesive ligands on interfaces at protein length

scales between 10 and 200 nm over large surface areas remains a challenge. However, this
is the length scale on which protein clustering in focal adhesions occurs when cells ad-
here to interfaces. The exact spatial control of receptor clustering in the cell membrane
on this length scale is demanding for challenging concepts from nanotechnology which
offer a rigid nanoadhesive pattern with flexible geometries over extended surface areas
at rather low production costs.

A substrate-patterning strategy based on the self-assembly of polystyrene-block-poly

(2-vinylpyridine) (PS-b-P2VP) diblock copolymer micelles covers the indicated length
scale; that is, diblock copolymer micelle lithography [38–42]. PS-b-P2VP diblock copoly-
mers form reverse micelles in toluene. The core of a micelle consists of associated
P2VP blocks which complex HAuCl

4

if this is added to the micellar solution. Dipping

and retracting a substrate from such a solution results in uniform and extended monomi-
cellar films supported by the substrate. Each micelle contains approximately the same
quantity of Au. Treating these films with a gas plasma results in the deposition of highly

58

4 Cell–Nanostructure Interactions

Figure 4.4

Cell adhesion to a

pattern of retronectin fabri-
cated by dip-pen nanolitho-
graphy (DPN) [37].
(A) Diagram describing the
cell adhesion experiment on
the DPN-generated pattern.
The total patterned area is
6400 mm

2

. The alignment

marks were generated by
scratching a circle into the
backside of the Au-coated
glass substrate. (B) Topogra-
phy image (contact mode) of
the retronectin protein array.
(C) Large-scale optical micro-
scope image showing the
localization of cells in the
nanopatterned area.
(D) Higher-resolution optical
image of the nanopatterned
area, showing intact cells.

regular Au-nanodots, thereby forming a rather perfect hexagonal pattern on solid-state
interfaces such as glass or Si-wafers. A preparation scheme is presented in Figure 4.5A.
The scanning electron microscopy (SEM) images in Figure 4.5B–E show Au-dots as
white spots which are arranged in pattern on Si-wafers by self-assembly of polystyrene-
block-poly(2-vinylpyridine(HAuCl

4

)

0.5

)

diblock

copolymer

micelles,

that

is

PS-b-

P[2VP(HAuCl

4

)

0.5

]. The nanoscopic patterns consist of Au-nanodots (3, 5, 6, or 8 nm in

diameter) with spacings between dots of 28, 58, 73, and 85 nm respectively adjusted by
the molecular weights of PS-b-P2VP and the amount of HAuCl

4

added to the micellar

solution. A side view of the Au-nanodots on a Si-wafer is shown in the high-resolution
electron transmission microscopy image in Figure 4.5F.

59

4.2 Methods

Figure 4.5

Micellar block copolymer lithography

and biofunctionalization. (A) Scheme of diblock
copolymer micelle lithography. (B–E) Extended
Au-nanodot pattern are displayed using scanning
electron microscopy [38]. Uniform Au-nanodots
(bright spots) of: (B) 3 nm by PS(190)-

b-

P[2VP(HAuCl

4

)

0.5

](190); (C) 5 nm by PS(500)-

b-

P[2VP(HAuCl

4

)

0.5

](270); (D) 6 nm by PS(990)-

b-

P[2VP(HAuCl

4

)

0.5

](385); and (E) 8 nm by PS(1350)-

b-P[2VP(HAuCl

4

)

0.5

](400) deposited onto Si-wafers

are shown. The number in brackets refers to the
number of monomer units in each block which
control the separation between Au-dots. These

varied between (B) 28, (C) 58, (D) 73 and (E) 85 nm.
The Au-dots form extended, nearly perfect hexa-
gonally-close packed pattern as indicated by the
Fourier transformed images (inset) which show
second-order intensity spots. (F) Biofunctionaliza-
tion of the Au-nanodots pattern [43]. The Au-dots
are presented as side view micrographs using a
high-resolution transmission electron microscope.
Molecules and Proteins are drawn schematically.
Since the Au-dot is sufficiently small, it is most
likely that only one integrin transmembrane recep-
tor occupies one dot.

These nanostructures serve as chemical templates for the spatial arrangement of RGD-

based ligands, as shown schematically in Figure 4.5F [43]. Biofunctionalization of the in-
terface comprises binding a polyethylene oxide layer to the silicon oxide substrate between
the Au-nanodots to avoid any nonspecific adsorption of proteins or parts of a cell mem-
brane. Subsequently, the Au-dots are functionalized by RGD ligands that bind selectively
to Au from a solution containing these ligands. In this study, cyclic RGD molecules have
been used, that is c(RGDfK)-thiol, as has been synthesized by the group of Kessler (TU
Munich). c(RGDfK)-thiols contain the cell-adhesive RGD sequence which is recognized
by a

v

b

3

-integrin with high affinity [44, 45]. In Figure 4.5F, Au-dots and integrins are

drawn approximately to scale, indicating that the size of a Au-nanodot provides dimen-
sions of a chemical anchor point to which, potentially, only one integrin can bind. This
is a very valuable tool, as the pattern dimensions and geometries control the assembly
of single integrins to form the basis of a focal adhesion cluster. Thus, uniform patterning
of extended substrate areas by diblock copolymer micelle lithography provides access to an
important length-scale for cell-adhesion studies that is hardly accessible with other tech-
niques.

In Figure 4.6, MC3T3-osteoblasts were seeded on glass and examined after one day

using optical phase-contrast microscopy. Only three-fourths of the glass substrate area
was patterned with Au-nanodots, with different spacings between the dots. The Au-dots
were functionalized by c(RGDfK)-thiols, and the free glass was passivated by PEG. A
line of cells marks the borderline of the nano-pattern area (white arrows). The right

60

4 Cell–Nanostructure Interactions

Figure 4.6

Optical phase-contrast mi-

croscopy images of MC3T3-osteoblasts
on nanopatterns of different spacing
[43]: (A)

Z58 nm; (B) Z73 nm. Cells

mark the borderline of nanostructures.
Extending the separation between indi-
vidual dots from 58 nm to 73 nm
causes failure in cell spreading. (C)
Cells attach also to Au-nanodots that
were not covered by c(RGDfK)-thiols,
but cell adhesion and spreading is truly
limited, as in the case of 73-nm spac-
ing between dots.

side was entirely passivated against cell adhesion; thus, cell adhesion and attachment is
only observed on the left side of the images. When plated on Au-nanodot patterns with
various spacing, functionalized by c(RGDfK)-thiols, MC3T3-osteoblasts show different ad-
hesion behavior. It is clear that cells spread very well on the 58-nm (Figure 4.6A) patterns,
appearing as they do on uniformly RGD- or fibronectin-coated surfaces (not shown). On
the other hand, hardly any cell spreading is observed on substrates with 73-nm spaced
nanodots (Figure 4.6B). Quiescent and migrating cells are also visible. Quiescent cells
present a rounded shape which causes strong scattering of light, while migrating cells
are usually characterized by extended filopodia (see arrows). These observations have
been repeated with additional cell types, i. e., REF52-fibroblasts, 3T3-fibroblasts and
B16-melanocytes, indicating a universal characteristic cell-adhesion behavior. Figure
4.6C shows MC3T3-osteoblasts on Au-nanodots separated by 58 nm and not conjugated
to c(RGDfK)-thiols. Cell spreading on these surfaces is rather poor, and few cells remain
attached after gentle rinsing.

The molecular formation of focal contacts and the assembly of actin stress fibers in

MC3T3-osteoblasts adhering to these nanopattern substrates were investigated by cultur-
ing cells for one day, fixing and staining them against vinculin and actin. Figure 4.7 pre-
sents confocal micrographs where c(RGDfK) covered Au-dots had spacings of (A) 58 and
(B) 73 nm. It is obvious that the pattern with adhesive c(RGDfK) Au-nanodot separations
of 58 nm establishes well-constituted, quite long vinculin clusters (shown as green) and
well-defined actin stress fibers (shown as red). The adhesion area of these cells is a factor
of

Z4 greater. Fairly blurred images of vinculin and actin distribution were obtained when

Au-dots were not covered by c(RGDfK)-thiols (not shown), or the separation between the
dots was 73 nm (B). It is also of note that only these cells which remained on substrates
after fixing and staining could establish either strong or stable adhesion (see Figure 4.6A),
or cells which could at least form nonstable adhesions (shown by red arrows in Figure
4.6B). All other cells (yellow arrows in Figure 4.6B) were washed off by the fixing and
staining process.

The increase in dot separation distances causes a decrease in global dot density. There-

fore, the observed limitation of cell adhesion at increased dot separation could be reasoned
either on the global density of c(RGDfK)-thiol covered Au-dots or on the local dot-to-dot
distance. In order to address this issue, “micro”-nanostructured interfaces were created
as described in Ref. [38]. This technique allows for deposition of a defined number of
Au-nanodots in a confined area of the substrate. The surfaces were designed such that
the global dot density was 90 dots mm

–2

, and thus significantly smaller than in all cases

of extended Au-dot pattern (280 dots mm

–2

, with 58 nm and 190 dots mm

–2

with the 73-nm

separated dots). The local dot density was organized in 2

q 2 mm

2

patches of 58-nm

spaced dots was 280 dots mm

–2

(Figure 4.8A). Figure 4.8B shows a bright-field optical

micrograph taken 3 hours after plating of MC3T3-osteoblasts on the substrate. Clearly, the
cells are confined only to the structured area, and the process of cell spreading was ad-
vanced (as shown in the inset). After 24 hours (C), well-spread cells are present in this
area, whereas cells located outside the frame (indicated by arrows) are poorly spread. Fig-
ure 4.8D shows a confocal fluorescent micrograph of a cell from (C) after immunohisto-
chemical staining for vinculin (green) and actin (red), thereby demonstrating the confine-
ment of focal adhesion to the square pattern and the origin of actin stress fibers from

61

4.2 Methods

there. In this case, the average focal adhesion length was 2.6

e 0.9 mm; this value is be-

tween the side length of one square pattern and its diagonal. The distribution in focal
adhesion length is remarkably narrow and displays the confinement by a square. Cells
do not adhere to all squares, but in some areas a separation distance between focal adhe-
sions of 1.5 mm is recognized as shown in the inset of Figure 4.8D. If cultured on pattern
uniformly structured with dots separated by 58 nm, these cells form focal adhesion
lengths with a mean value of 5.6

e 2.7 mm (Figure 4.7A).

These adhesion experiments indicate that local dot–dot separation, rather than global

dot density, was critical for inducing cell adhesion and focal adhesion assembly. Thus,
for example, the dot density located under cells attached to the “micro”-nanostructured

62

4 Cell–Nanostructure Interactions

Figure 4.7

A pair of confocal fluorescent optical

micrographs of MC3T3-osteoblasts stained for vin-
culin (green) and actin (red) [43]. Cells interacting
with Au-nanodot patterns with Au-dot spacing of (a)
58 nm and (b) 73 nm. (C) Scheme of biofunctio-
nalized nanopattern to control integrin clustering
in cell membranes: Au-dots are conjugated with
c(RGDfK)-thiols and areas between cell adhesive
Au-dots are passivated by PEG against cell adhe-

sion. Therefore, cell adhesion is mediated entirely
via c(RGDfK)-covered Au-nanodots. A separation of
Au/RGD dots by

j73 nm causes limited cell at-

tachment and spreading and actin stress fiber for-
mation because of restricted integrin clustering.
This is indicated by failure of focal adhesion acti-
vation (FA–), whereas distances between dots of
J 58 nm caused focal adhesion activation (FA+).

squares, consisting of 58-nm separated dots is considerable lower than that of dots located
underneath cells attached to a substrate, uniformly patterned by dots, separated by
j 73 nm. Nevertheless, cells did form focal adhesions on the former surface, but failed
to do so on the latter. This is summarized schematically in Figure 4.7C.

4.3

Outlook

Cell attachment to interfaces depends on many factors, such as the affinity and specificity
of surface-bound ligands to integrins, the mechanical strength of ligand support and link-
age, spacer length, overall ligand concentration, and ligand clustering [30]. This survey of
challenging investigations concerned with cell adhesion on nanostructured interfaces con-

63

4.3 Outlook

Figure 4.8

MC3T3-osteoblast adhesion on

“micro”-nanostructures occupied by c(RGDfK)-
thiols [43]. (A) SEM image of “Micro”-nanostruc-
tures: SEM-micrograph of 5-nm Au-dots separated
by 58 nm in a hexagonally-close packed pattern lo-
calized in 2

q 2 mm squares which are separated by

1.5 mm [38]. The bright-field optical micrograph of
adhesive MC3T3-osteoblasts on pattern shown in

(A) is covering the area in the marked box after (B)
3 h and (C) 24 h of cell culture. (D) Fluorescent
optical micrograph of MC3T3-osteoblast showing
the location of FA by staining for vinculin (green).
FA appear as small strip-like, bright green objects
on the pattern in (A). The actin filaments are seen
in red.

cludes that highly intriguing cellular processes are stimulated and controlled by substrate
nanotopography and spatial ligand patterning for single integrin receptor occupation.
Thus, nanoadhesive patterns offer the unique opportunity to define length scales in multi-
molecular complexes within focal adhesions, with unprecedented resolution as small as a
single protein. Variations in nanoadhesive site organization, including alterations in
ligand template pliability and presentation of small dot clusters, for example pairs or tri-
plets, may shed light on the minimal molecular number of an effective integrin cluster
necessary to obtain cell attachment, spreading or migration, and also of the possible pat-
tern-specific features (molecularly defined adhesion “keys”) that trigger cell adhesion-
based signaling [43].

References

64

4 Cell–Nanostructure Interactions

[1]

H. M. Blau, D. J. Baltimore, J. Cell Biol.
1991, 112, 781-783.

[2]

E. Ruoslahti, B. Obrink, Exp. Cell. Res.
1996, 227, 1-11.

[3]

B. Geiger, A. Bershadsky, R. Pankov, K. M.
Yamada, Nature Rev. Mol. Cell Biol. 2001,
2, 793.

[4]

E. Zamir, B. Geiger, J. Cell Sci. 2001, 114,
3583; E. Zamir, B. Geiger, J. Cell Sci. 2001,
114, 3577.

[5]

D. R. Critchley, Curr. Opin. Cell Biol. 2000,
12, 133.

[6]

F. G. Giancotti, E. Ruoslahti, Science 1999,
285, 1028.

[7]

R. O. Hynes, Cell 1987, 4, 549.

[8]

S. Miyamoto, S. K. Akiyama, K. M. Yamada,
Science 1995, 267, 883.

[9]

S. Levenberg, B. Z. Katz, K. M. Yamada,
B. Geiger, J. Cell Sci. 1998, 111, 347.

[10]

N. Q. Balaban, U. S. Schwarz, D. Rivelin, P.
Goichberg, G. Tzur, L. Sabanay, D. Mahalu,
S. Safran, A. Bershadsky, L. Addadi,
B. Geiger, Nature Cell Biol. 2001, 3, 466.

[11]

B. Alberts, D. Bray, J. Lewis, M. Raff,
K. Roberts, J. D. Watson, Molecular Biology
of the Cell, 3rd edn., Garland Publishing,
New York, USA, 1994.

[12]

L. A. Lasky, Nature 1997, 390, 15-17.

[13]

J. W. Weisel, C. Nagaswami, G. Vilaire,
J. S. Bennett, J. Biol. Chem. 1992, 267,
16637-16643.

[14]

M. V. Nermut, N. M. Green, P. Eason,
S. S. Yamada, K. M. Yamada, EMBO J.
1988, 7, 4093-4099.

[15]

J. P. Xiong, T. Stehle, B. Diefenbach,
R. Zhang, R. Dunker, D. L. Scott, A. Joa-
chimiak, S. L. Goodman, M. A. Arnaout,
Science 2001, 294, 339.

[16]

E. M. Erb, K. Tangemann, B. Bohrmann,
B. Müller, J. Engel, Biochemistry 1997, 36,
7395-7402.

[17]

J. P. Xiong, T. Stehle, R. Zhang, A. Joa-
chimiak, M. Frech, S. L. Goodman, M. A.
Arnaout, Science 2002, 296, 151-155.

[18]

P. L. Stewart, C. Y. Chiu, S. Huang, T. Muir,
Y. Zhao, B. Chait, P. Mathias, G. R. Ne-
merow, EMBO J. 1997, 16, 1189-1198.

[19]

D. G. Stupack, E. Li, S. A. Silletti,
J. A. Kehler, R. L. Geahlen, K. Hahn,
G. R. Nemerow, D. A. Cheresh, J. Cell Biol.
1999, 144, 777-788.

[20]

M. J. Dalby, S. J. Yarwood, M. O. Riehle,
H. J. H. Johnstone, S. Affrossman, A. S. G.
Curtis, Exp. Cell Res. 2002, 276, 1–9.

[21]

M. J. Dalby, M. O. Riehle, H. Johnstone,
S. Affrossman, A. S. G. Curtis, Biomaterials
2002, 23, 2945–2954.

[22]

M. J. Dalby, S. Childs, M. O. Riehle, H. J. H.
Johnstone, S. Affrossman, A. S. G. Curtis,
Biomaterials 2003, 24, 927–935.

[23]

D. L. Elbert, J. A. Hubbell, Biomacromole-
cules 2001, 2, 430-441.

[24]

G. B. Brown, Spatial Control of Ligand
Presentation on Biomaterials Surfaces, in:
Materials Science and Engineering. MIT,
Cambridge, 1999, pp. 165.

65

References

[25]

G. Maheshwari, G. Brown, D. A. Lauffen-
burger, A. Wells, L. G. Griffith, J. Cell Sci.
2000, 113, 1677.

[26]

M. Mrksich, G. M. Whitesides, ACS Symp.
Ser. 1997, 680, 361.

[27]

R. L. C. Wang, H. J. Kreuzer, M. Grunze,
J. Phys. Chem. B 1997, 101, 9767-9773.

[28]

C. Roberts, C. S. Chen, M. Mrksich,
V. Martichonok, D. E. Ingber, G. M.
Whitesides, J. Am. Chem. Soc. 1998, 120,
6548-6555.

[29]

W. S. Yeo, C. D. Hodneland, M. Mrksich,
ChemBioChem 2001, 2, 590-592.

[30]

U. Hersel, C. Dahmen, H. Kessler,
Biomaterials 2003, 24, 4385.

[31]

M. Kantlehner, P. Schaffner, D. M. J. Fin-
singer, A. Jonczyk, B. Diefenbach, B. Nies,
G. Holzemann, S. L. Goodman, H. Kessler,
ChemBioChem. 2000, 1, 107-114.

[32]

S. P. Massia, J. A. Hubbell, J. Cell Biol.
1991, 114, 1089-1100.

[33]

L. Y. Koo, D. J. Irvine, A. M. Mayes, D. A.
Lauffenburger, L. G. Griffith, J. Cell Sci.
2002, 115, 1423-1433.

[34]

D. J. Irvine, A. M. Mayes, L. G. Griffith,
Biomacromolecules 2001, 2, 85-94.

[35]

D. J. Irvine, A. V. G. Ruzette, A. M. Mayes,
L. G. Griffith, Biomacromolecules 2001, 2,
545.

[36]

C. S. Chen, M. Mrksich, S. Huang, G. M.
Whitesides, D. E. Ingber, Science 1997, 276,
1425.

[37]

K. B. Lee, S. J. Park, C. A. Mirkin, J. C.
Smith, M. Mrksich, Science 2002, 295,
1702.

[38]

R. Glass, M. Arnold, M. Möller, J. P. Spatz,
Adv. Funct. Mater. 2003, 13, 569; R. Glass,
M. Möller, J. P. Spatz, Nanotechnology 2003,
14, 1153.

[39]

J. P. Spatz, S. Mößmer, M. Möller, T. Her-
zog, A. Plettl, P. Ziemann, Langmuir 2000,
16, 407.

[40]

J. P. Spatz, S. Mößmer, M. Möller,
Chem. Eur. J. 1996, 2, 1552-1555.

[41]

J. P. Spatz, A. Roescher, M. Möller,
Adv. Mater. 1996, 8, 337.

[42]

J. P. Spatz, S. Sheiko, M. Möller,
Macromolecules 1996, 29, 3220-3226.

[43]

M. Arnold, E. Ada Cavalcanti-Adam,
R. Glass, J. Blümmel, W. Eck, M. Kant-
lehner, H. Kessler, J. P. Spatz, Chem Phys
Chem, in press.

[44]

M. Pfaff, K. Tangemann, H. Kessler, J. Biol.
Chem. 1994, 269, 20233.

[45]

R. Haubner, D. Finsinger, H. Kessler,
Angew. Chem. 1997, 109, 1440.