Review

Guide to Collagen Characterization for Biomaterial Studies

Leah C. Abraham,1,2 Erin Zuena,1,2 Bernardo Perez-Ramirez,3 David L. Kaplan1,2
1
Departments of Chemical and Biological Engineering, Tufts University, Medford, Massachusetts 02155

2
Department of Biomedical Engineering, Bioengineering and Biotechnology Center, Tufts University,
Medford, Massachusetts 02155
3
Genzyme Corporation, BioFormulations Development, Framingham, Massachusetts 01701-9322

Received 3 January 2007; revised 19 November 2007; accepted 18 December 2007

Published online 3 April 2008 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jbm.b.31078

Abstract: The structure and remodeling of collagen in vivo is critical to the pathology and
healing of many human diseases, as well as to normal tissue development and regeneration. In
addition, collagen matrices in the form of fibers, coatings, and films are used extensively
in biomaterial and biomedical applications. The specific properties of these matrices, both in
terms of physical and chemical characteristics, have a direct impact on cellular adhesion,
spreading, and proliferation rates, and ultimately on the rate and extent of new extracellular
matrix formation in vitro or in vivo. In recent studies, it has also been shown that collagen
matrix structure has a major impact on cell and tissue outcomes related to cellular aging and
differentiation potential. Collagen structure is complex because of both diversity of source
materials, chemistry, and structural hierarchy. With such significant impact of collagen
features on biological outcomes, it becomes essential to consider an appropriate set of
analytical tools, or guide, so that collagens attained from commercial vendors are
characterized in a comparative manner as an integral part of studies focused on biological
parameters. The analysis should include as a starting point: (a) structural detail—mainly
focused on molecular mass, purity, helical content, and bulk thermal properties, (b) chemical
features—mainly focused on surface elemental analysis and hydrophobicity, and (c)
morphological features at different length scales. The application of these analytical
techniques to the characterization of collagen biomaterial matrices is critical in order to
appropriately correlate biological responses from different studies with experimental
outcomes in vitro or in vivo. As a case study, the analytical tools employed for collagen
biomaterial studies are reviewed in the context of collagen remodeling by fibroblasts. The goal
is to highlight the necessity of understanding collagen biophysical and chemical features as a
prerequisite to (a) studies with cells and tissue formation, and (b) suggest modes to establish
comparative outcomes for studies conducted in different laboratories. ' 2008 Wiley Periodicals,
Inc. J Biomed Mater Res Part B: Appl Biomater 87B: 264–285, 2008

Keywords: collagen; characterization; denatured; structure; biomaterials

INTRODUCTION as a biomaterial has led to research on use in scaffolds

Collagen Biomaterials
The prevalence of collagen in human tissues makes it a
natural choice as a polymer for biomedical materials and
tissue-engineering matrices. Collagen is the most abun-
dant protein present in mammals, composing 30% by
weight of body protein tissue.1,2 Collagen is also biode-
gradable, biocompatible, and enhances cellular penetra-
tion and wound repair.3 The potential value of collagen
Correspondence to: D. L. Kaplan (e-mail: david.kaplan@tufts.edu)
' 2008 Wiley Periodicals, Inc.
264
for ligament repair, collagen grafts for scar and burn
repair, and the engineering of osteochondral tissue.4–7
Collagen is also a target for study in diseases involving
extensive collagen remodeling, including aortic heart
valve repair and bone repair.8,9 A better understanding of
the interactions between cells and collagen should allow
for the more rational design and use of these substrates
depending on the cells, tissues, and environments
involved in vitro and in vivo. For the purpose of this
review, the focus is mainly on commercially available
collagens for use in studies of biomaterial structure and
function. While many other sources (e.g., collagen
derived from various tissues, recombinant DNA derived
 GUIDE TO COLLAGEN CHARACTERIZATION FOR BIOMATERIAL STUDIES 265

TABLE I. List of Some of the Collagen Types and Information on Chain Composition, Structure, Tissue Location and Related
Information2,10–12

Types Chain Composition Structural Details Localization Notes

I [a1(I)]2[a(I)] 300 nm, 67-nm banded fibrils Skin, tendon, bone, etc. 90% of all collagen of the
human body. Scar tissue-
the end product when
tissue heals by repair.
II [a1(II)]3 300 nm, small 67-nm fibrils Cartilage, vitreous humor Articular cartilage
III [a1(III)]3 300 nm, small 67-nm fibrils Skin, muscle, frequently Collagen of granulation tissue,
with type I and is produced quickly by
young fibroblasts before the
tougher type I collagen is
synthesized.
IV [a1(IV)2[a2(IV)] 390 nm C-term globular domain, All basal lamina Basal lamina
nonfibrillar
V [a1(V)][a2(V)][a3(V)] 390 nm N-term globular domain, Most interstitial tissue, Most interstitial tissue, assoc.
small fibers assoc. with type I with type I
VI [a1(VI)][a2(VI)][a3(VI)] 150 nm, N1C term. globular Most interstitial tissue, Most interstitial tissue, assoc.
domains, microfibrils, assoc. with type I with type I
100-nm banded fibrils
VII [a1(VII)]3 450 nm, dimer Epithelia Epithelia
VIII [a1(VIII)]3 130 nm, N1C term. Globular Some endothelial cells Some endothelial cells
domains
IX [a1(IX)][a2(IX)][a3(IX)] 200 nm, N-term. Globular Cartilage, assoc. with type II Cartilage, assoc. with type II
domain, bound proteoglycan
X [a1(X)]3 150 nm, C-term. Globular Hypertrophic and mineralizing Hypertrophic and mineralizing
domain cartilage cartilage
XI [a1(XI)][a2(XI)][a3(XI)] 300 nm, small fibers Cartilage Cartilage
XII a1(XII) 75-nm triple helical tail, central Interacts with types I and III Interacts with types I and III
globule, three 60-nm globule
arms
Mainly types I–IV have been utilized to varying degrees in tissue-engineering related biomaterials studies, with some efforts on the other types shown.

collagens, collagen-like peptides) and variations of colla-
gen (e.g., crosslinked, with telopeptides, and related var-
iations) are available to researchers who isolate their own
materials, many of the core analytical tools would remain
similar to those described here in the context of commer-
cially available prepared sources.
There are over 20 known types of collagen (Table I).11
The fibril-forming (fibrillar) collagens include collagen type
I [a1(I)]2a2 (I) comprising fibril bone, skin, tendons, liga-
ments, cornea, and internal organs, accounting for 90% of
body collagen; collagen type II [a1(II)]3 comprising fibril
cartilage, intervertebral disc, notochord, and vitreous humor
of the eye; collagen type III [a1(III)]3 comprising fibril
skin, blood vessels, and internal organs; collagen type V
[a1(V)]2a2(V) and a1(V) a2(V) a3(V) fibril (with type I)
comprising tissue similar to those for type I collagen; colla-
gen type XI a1(XI) a2(IX) a3(XI) fibril (with type II) com-
prising tissue similar to collagen type II.11 For all collagen
types, each collagen chain of 1000 amino acids is com-
posed of three left-handed a helix chains that twist together
to form the right-handed helix of the collagen molecule.2,13
The collagen molecule is about 300 kDa, composed of
10% each of proline and hydroxyproline, and has glycine
present at every third amino acid position.13
A variety of commercial collagen sources are used for
tissue-engineering applications as well as for cell studies
that require collagen matrices. Table II lists some of the
more common commercial collagen sources used in these
types of studies. These materials, often from poorly speci-
fied preparations, make comparisons between studies of
various collagen materials difficult. An additional issue is
that collagen source materials are often prepared differently
in each laboratory, complicating further attempts at com-
parisons of biological outcomes. One of the aims of this
review is to summarize the characteristics of collagen
source materials to draw common themes in terms of how
the source material and the physical features of the source
relate to biological outcomes. To accomplish this goal, we
have focused on the characterization of commercially avail-
able collagen preparations in order to highlight some of the
complications and strategies that can be employed toward a
working ‘‘guide’’ for assessment of these materials.
Vendors, such as Sigma Aldrich, continue to refer to
collagens both by the widely used ‘‘modern’’ type classifi-
cation system defined by the chain types (type I 5
(a1[I])2a2[I])1, type II 5 (a1[II])3)10 and by types as defined
in the early 1970s by the researchers who first separated
collagen chains (Miller type II cartilage).28 The earlier col-
lagen classifications tended to be defined more by the tissue
type from which the collagen had been extracted than the
collagen chain content.

Journal of Biomedical Materials Research Part B: Applied Biomaterials

 266

TABLE II. Collagens and Associated Data Available From Vendors, Unless Otherwise Indicated

Collagen Material Source Preparation Characteristics Tissue Engineering Application

Vitrogen Collagen Corp, Palo A Bovine dermal 99.9% pure collagen by SDS Collagen scaffold for tendon repair14
lto, CA collagen dissolved in polyacrylamide gel electrophoresis
0.012N HCl in conjunction with bacterial
collagenase sensitivity and silver
staining
BD matrigelTM matrix BD Biosciences Extracted from EHS mouse Solubilized basement membrane Cell attachment and differentiation in
sarcoma, a tumor rich in preparation. Major component is 3T3-F442A preadipocytes15,16
ECM proteins laminin (56%), followed by
collagen IV (31%), heparan sulfate
proteoglycans, and entactin (8%)
Roche type I rat tail 1 Roche Purified from rat tail tendon by a Collagen from rat tail, mainly of type Collagen remodeling by
179 179 modification of the method of I collagen fibroblasts19,20
Bornstein17; Michalopoulos and
Pitot18
Roche Collagen S from calf Roche Collagen is purified from calf skin by Collagen: [98%, collagen type 1: Comparison to collagen extracted
skin 1 098 292 extraction with 0.5M acetic acid, [95%, collagen type III: \5% from tissues21
pH 2.5
Sigma bovine type I calf skin Sigma Aldrich Prepared from calf skin. Further 0.1% (1 mg21 mL) solution of calf Recommended for use as a cell
sterile solution details are not provided by the skin collagen in 0.1M acetic acid culture substratum at 6–10 lg cm22.
vendor22 Not suitable for 3D gel formation
Sigma bovine type I calf skin Sigma Aldrich Prepared by a modification of the Solubility noted as 5 mg21 mL in Synthesis and characterization of a
BioChemika soluble procedure of Gallop and Seifter23 water by vendor, hazy, colorless model extracellular matrix that
ABRAHAM ET AL.

and viscous induces partial regeneration of adult

mammalian skin
Sigma bovine nasal septum Sigma Aldrich Prepared by a modification of the Type II collagen (Bornstein and Traub Activity of stachyrase A against
pepsin extraction procedure of classification22) collagen25
Niyibizi et al.24
Sigma bovine type I Achilles Sigma Aldrich Prepared by the method of Einbinder J Type I collagen (Bornstein and Traub Suitable for use as a substrate for
tendon and Schubert26 classification22) collagenase
Sigma chicken sternal Sigma Aldrich Prepared by a modification of the This collagen has been tested in Recommended for use as a cell
method of Trentham D.E., et al.27 culture with mammalian cells to culture substratum at 6–10 lg cm22
verify it is low in endotoxin content.
Miller classification type II28
Sigma human type I Sigma-Aldrich Prepared from human skin by 95% (SDS-PAGE) type I, acid- Preparation of soluble collagen23
modification of Gallop and Seifter23 soluble
Characterization of collagen matrices pertinent to collagen trafficking and remodeling. As mentioned earlier, the analytical assessments described are confined primarily to commercially available sources of collagen. This focus
allows those interested in collagen isolation from tissue sources using various extraction and modification protocols to employ methods from the literature or from their own labs, but to then to use the analytical ‘‘guide’’ provided herein
to assess their collagen preparations.

Journal of Biomedical Materials Research Part B: Applied Biomaterials

 GUIDE TO COLLAGEN CHARACTERIZATION FOR BIOMATERIAL STUDIES
The rationale behind this review is the remarkable impact
collagen matrix structure has on cell and tissue outcomes
from recent studies of cellular aging as well as on retention of
stem cell differentiation potential.19,20,29,30 Collagen bioma-
terials have been used in many tissue-engineering applica-
tions. For example, osteoblast-like cells (Saos-2) adhere
more effectively and proliferate on xenogenic bone biomate-
rial containing collagen fibers compared to deproteinated
bone.31 Layering of collagen sheets and scaffolds seeded
with cardiomyocytes has enhanced cell survival with macro-
scopic pulsation similar to that of native heart tissue.32 Colla-
gen-based scaffolds are also prominent in the field of dermal
repair.33–35 In each of these reports, the characteristic of the
collagen biomaterial used in the experiments was important
to the success of the biological system under study. Further-
more, the widespread use of collagen biomaterials in many
biomedical applications with different rates and extents of
degradation, and where different release profiles of therapeu-
tics are sought,3,36,37 suggest that there exist important rela-
tionships between collagen structure and function in the
biomedical context. With such a significant impact of colla-
gen features on biological outcomes, it becomes essential to
consider an appropriate set of characterization tools so that
collagens are characterized in a comparative manner as an in-
tegral part of any study focusing on biological outcomes. Of
particular interest to the biopharmaceutical industry is to
have standard analytical tools and procedures that could be
used in characterization, scale-up, and comparability analysis
of delivery systems based on collagen.
The interaction of cells with a biomaterial, particularly at
the surface, determines adhesion and spreading and conse-
quently plays an important role in determining the pathways of
cellular differentiation, growth, and survival.38 In addition, a
biomaterial matrix for tissue engineering must have porosity
and mechanical stability suitable for the target cells and tissue
functions.34 For example, the goal of engineering blood vessels
using tissue engineering approaches has led to the utilization
of synthetic biopolymers to approximate the three-layered
structures present in native arteries.39 The tailoring of biomate-
rial properties to mimic those of the target tissue is desired to
help increase the chances of success of tissue-engineering
implants.36 In order to successfully mimic these chemical and
physical features of native tissues with collagen, both the tis-
sues and biomaterials must be extensively characterized. Char-
acterization tools for collagen must consider structural,
morphological, and chemical features because of the impor-
tance of physical and chemical factors on cell responses. In
addition, these features directly influence rates and extent of
collagen remodeling in vitro and in vivo, thus playing a major
role on functional outcomes. Bulk and surface properties need
to be considered because of its impact on stability, mechanical
performance, and cell interactions.
Objective
Despite the large number of studies with designed biomate-
rial surfaces, there remains the need to engineer biomateri-
Journal of Biomedical Materials Research Part B: Applied Biomaterials
267
als that can provide both surface and bulk requirements for
tissue-engineering matrices and also promote desired cell
responses in vivo for tissue repair. Although collagen is fre-
quently used as a biomaterial, the understanding of colla-
gen biomaterial characteristics as a function of cellular
responses is far from complete, particularly when consid-
ered in the light of the prominent role this family of fibrous
protein plays in vivo. The objective of this review is to es-
tablish a more consistent basis of collagen characterization,
so that biological responses can be more accurately related
to differences in structure, morphology, and chemistry.
These types of relationships are critical in order to optimize
and control cell and tissue outcomes on collagen-based bio-
materials, as well as to better predict and control rates and
extent of integration of in vitro prepared and/or grown tis-
sues in vivo. This insight should lead to the better design
and control of matrix structural and morphological features,
resulting in more predictable and relevant cell and tissue
outcomes in vitro and in vivo. We present information on
collagen biomaterial use and characterization along with
our results related to characterization of collagen matrices
pertinent to collagen trafficking and remodeling. As men-
tioned earlier, the analytical assessments described are con-
fined primarily to commercially available sources of
collagen. This focus allows those interested in collagen
isolation from tissue sources using various extraction and
modification protocols to employ methods from the literature
or from their own labs, but then to use the analytical ‘‘guide’’
provided herein to assess their collagen preparations.
ANALYTICAL METHODS
Overview
The characterization of collagen is divided into three major
areas in this review: (a) structural detail—mainly focused
on molecular mass, purity, helical content, and bulk ther-
mal properties, (b) chemical features—mainly focused on
surface elemental analysis and hydrophobicity, and (c) mor-
phological features at different length scales. In total, this
suite of analytical assessments of collagens can provide a
consistent basis for comparison of materials and thus bio-
logical outcomes with these materials. While this list of
characterization methods suggested for collagen biomateri-
als is not exhaustive, it provides a starting point for consis-
tency in analysis. Table III includes a technique summary
listing for each method, the collagen format needed, solu-
tion concentrations, information gathered, and major limita-
tions of the method.
Structural Information
Mass Spectroscopy. Collagens proteins have been
reported with molecular masses from 28340 to 300 kDa.41
Mass spectroscopy can be employed to provide molecular
mass data on collagen as well as the identification of
telopeptides and other potential contaminants in collagen
268 ABRAHAM ET AL.

preparations that can directly impact cell functions.42 Ma-

gels, sponges

Films: 15–780

Resolution of
Films, fibers,
trix-assisted laser desorption ionization time-of-flight

Topography
lg cm22
ESEM

features
(MALDI-TOF) mass spectroscopy is commonly used. Pro-
tein fragments are dissolved in an organic acid, then dried
onto matrices—most often metals or ceramics.10 After exci-
tation with a laser, the protein fragments are accelerated in
gels, sponges

Artifacts due to
an electric field.4 The detector identifies the proteins and
Films: 15–780

dehydration
Films, fibers,

Topography
fragments by their mass and charge.11 Protein samples are
lg cm22
SEM

typically applied at 10 lM. The sample is spotted onto to

a metal MALDI plate after dissolution in a solution of
water and organic with a crystallized molecules such as
3,5-dimethoxy-4-hydroxycinnamic acid (sinapinic acid), a-
Films or fibers

Films: 15–780

topography

equipment
cyano-4-hydroxycinnamic acid (alpha-cyano or alpha-ma-

Specialized
lg cm22

Molecular

trix), or 2,5-dihydroxybenzoic acid (DHB).42–49

AFM

Collagen chains (94 kDa40 to 98 kDa42 exceed the

working range of many mass spectrometers. Therefore,
digestion of the 300 kDa collagen to 94- to 98-kDa
Hydrophobicity
Films: 15–780

fragments is required. The lack of data reported for

Contact Angle
Films or fibers

smoothness

digested collagen and the complication due to its high mo-

lg cm22

Hydration,
surface

lecular mass provide significant challenges in mass spec-

troscopy assessments of collagens; thus, gel electrophoresis
is more commonly utilized (see later). Complications also
arise from low-molecular-mass collagen telopeptides (6–
composition
Films: 15–780

14 kDa42) that usually require a separate MALDI-TOF ma-

Films, fibers,

equipment
Specialized
lg cm22

trix for appropriate analysis. Mass spectroscopy is also use-

XPS

Atomic

ful to track bone remodeling and the formation of new

gels

bone collagen, where the presence of telopeptides is com-

mon.45 The purity of the collagen source material can also
be assessed using mass spectrocscopy.46
Dehydrated film

heat capacity
temperature,
in DSC pan

Denaturation

Large mass

Sodium Dodecyl Sulfate Polyacrylamide Gel Electro-

DSC

needed
5 mg dry
weight

phoresis. Sodium dodecyl sulfate polyacrylamide gel elec-

trophoresis (SDS-PAGE) is most commonly used to assess
collagen source material purity and breakdown. SDS-PAGE
allows visualization of protein fragments by loading of pro-
tertiary structure

software limited

tein samples in the wells of a thin gel, then using electric

film on plate
0.125 lg lL21

Secondary and

Deconvolution
weight film
100 lg dry

voltage to drive the protein fragments through the gel. The

solution or
cuvette or
CD
Solution in

smallest protein fragments are least impeded by the gel ma-

trix and travel furthest through the gel. Coomassie blue or
silver stain are commonly used to visualize the protein
TABLE III. Characterization Techniques for Collagen

bands. Small sample amounts from nanogram to microgram

are needed, and molecular mass banding patterns are
Denaturation
SDS-PAGE

10–20 lg,

obtained. Subsequent Western blots can be used to assess the

Solution

water

specificity of collagen type using monoclonal antibodies.47 A

Size

summary of collagen materials used in tissue engineering,

including the characterization of molecular mass, is shown in
(urea treated)

Table IV. SDS-PAGE gels commonly used for collagens are

Spectroscopy

10 lM, water

specialized
Single strand

equipment

4–20% polyacrylamide. Collagen samples can be loaded

Mass range,
Mass

solution

directly in dilute acid solutions [0.1% glacial acetic acid

(GAA)].
Mass

The materials described earlier were purified by vendors

and provided as purified single type collagens. For samples
Material format

that are purified in the laboratory directly from tissues,

Concentration,

limitations
Information

modified methods such as interrupted electrophoresis can

Description

solvent
Collagen

provide an additional tool to separate different collagen

Major

types from a single tissue. Interrupted electrophoresis

begins with a collagen sample in the well. After the bands

Journal of Biomedical Materials Research Part B: Applied Biomaterials

 GUIDE TO COLLAGEN CHARACTERIZATION FOR BIOMATERIAL STUDIES 269

TABLE IV. Measures of Molecular Mass of Collagens

Collagen Material Technique/Findings Reference

Proteins synthesized on oocytes grown Correlation of mass spectroscopy with polyacrylamide 48
with radio-labeled proline gels for molecular weight analysis
Collagen extracted from rat tail SDS-PAGE for molecular size as a function of UV exposure 49
Sigma human type I collagen SDS-PAGE for molecular size as a function of MMP and 50
TIMP cleavages
Rat tail tendon type I collagen SDS-PAGE for molecular size as a function of MMP cleavages 51
Collagen extracted from human hip bones MALDI-TOF MS. to show the presence of C-telopeptides 42
Collagen extracted from fetal calf skin tissue Infrared (IR)-MALDI TOF MS detection of collagen triple helix 40
Type II collagen from fetal bovine cartilage SDS-PAGE and MS to characterize gelatinase B degradation 46
of collagen type II
N-terminal propeptide of human procollagen Size determination of PINP by MS and SDS-PAGE 52
(PINP) from amniotic fluid

migrate into the gel the current is interrupted, b-mercapto-
ethanol is then added to the wells and incubated, causing
the collagen helices to unwind at rates related to the disul-
fide content of the specific collagens. When this method
was applied to collagens isolated from human skin samples,
the migration of a1[III] chains was delayed when compared
to a1[I] chains, allowing resolution.53
Circular Dichroism. Circular dichroism (CD) utilizes the
differential absorption of circular polarized light in an
asymmetrical environment to assess structure.54 The amide
bonds of a protein in highly ordered regions such as a heli-
ces and b sheets have specific optical activity due to orien-
tation.54 CD has commonly been employed as a technique
to characterize the helical content of collagen.24,54,56 The
helical nature of collagen is responsible for the important
structural properties of tissues and in scaffolds used in tis-
sue-engineering. For example, CD has been used to confirm
collagen incorporation and structure in polymer–collagen
electrospun matrices designed as scaffolds for soft tissue-en-
gineering applications.57 CD has been used to assess the sus-
ceptibility of collagen to ultraviolet light based on loss of
helicity,58 to confirm the presence of collagen helical content
for collagen-like peptides56,59 and to study the enzymatic hy-
drolysis of collagen due to MMP-related reactions.51 CD has
also been used to characterize variation in collagen structure
in specific skeletal diseases such as osteogenesis imper-
fecta.60,61 Thermal denaturation melts the collagen, thus dis-
rupting the triple helix, and is usually an irreversible process
because of the complex self-assembly involved in proper
collagen chain associations and registry.62
The helical state of collagen in biomaterial and disease
applications can have significant effect on the cellular
remodeling responses. In fibroblastic matrix remodeling
in vitro, the rates of remodeling are greater, and cell health
is improved on denatured (wound-like) collagen versus
non-denatured (native-like) collagen.19 These observations
suggest that the presence of denatured collagen in a tissue-
engineering matrix might promote active remodeling neces-
sary for integration of implants. In several collagen disease
states including osteoporosis, osteogenesis, and bone metas-
tases, the remodeling of collagen plays an important role in
the pathology of the disease.46,63,64 In osteoarthritis, the
reduction in mechanical properties of subchondral bone has
been associated with an increase in denatured collagen.65
Since the helical content of collagen is critical to cell
responses both in vitro and in vivo, establishing the helical
content of collagen biomaterials is necessary. A summary
of collagen biomaterials used in tissue engineering charac-
terized by CD is shown in Table V.
The quality of CD data depends on sample concentration
and temperature. Sample concentrations must be controlled
and should be low enough (\0.125 mg mL21 for collagen)
to avoid saturation of the detector. Sufficient temperature
control is also required to avoid denaturation under experi-
mental conditions. Ellipticity data noting the angle of
polarization of light, reported in millidegrees, can be con-
verted to mean residue ellipticity [degree cm2 dmol21];
however, the molecular mass of the sample is required for
further conversion to molar ellipticity [dL mol21 dm21].
Since most commercially available collagens do not specify
molecular mass, this must be determined experimentally in
order to report mean residue ellipticity data (see section
above on molecular mass determination). Despite the com-
plications associated with collecting and comparing data on
collagen structure by CD, this remains a powerful tool to
assess helicity and thus degree of naturation/denaturation of
a sample used in biological studies. Thus, ellipticity, deter-
mined by CD, provides a measure of this structural feature,
and thus an assessment of the native/denatured state of the
collagen preparation.
Several methods for calculating helix content are avail-
able. For example, fhelix 5 [h]obs 222/(240,000[1 2 2.5/chain
length]).66 Data conversion from millidegrees to mean resi-
due ellipticity using [h]222 5 h/(molar concentration 3 15
residues), in deg cm2 dmol21, allowing the calculation of
(observed) (max) (max)
helix content as [h]222 /[h]222 , where [h]222 is given
by 240,000(1 2 2.5/n), and n is the number of amino
acids in the peptide (Chakrabartty et al. 1991).71 Other
calculations rely on deconvolution programs to evaluate
helical content.73,74 In each case, the accuracy of the quan-
titative assessment of helicity depends strongly on the ac-

Journal of Biomedical Materials Research Part B: Applied Biomaterials

270 ABRAHAM ET AL.

TABLE V. Circular Dichroism Assessments of Collagens and Associated References

Collagen Material Technique/Findings Reference

Collagen-like helices from streptococcal proteins Unfolding of the helix was observable at 220 nm 66
after heat denaturation
Calf skin collagen and gelatin Lower intensity, redshifted CD for heat-denatured 67
collagen at 220 nm
Chick type I procollagen Tm of 428 C 58
Fetal calf skin collagen Characterization of collagen helix reduction with 68
denaturation
Collagen-like peptides Confirmation of the presence of triple helix 56
Collagen type I from bovine skin or rat tendon Characterization of collagen helix reduction with 51
denaturation
Bovine calf skin type I collagen and peptides Confirmation of the presence of triple helix 55
Collagen-like peptides Changes in helicity with side group modifications 59
Calf skin collagen Characterization of collagen and collagen peptides, 69
reduction of helix with heat treatment
Calf skin collagen also prepared with crosslinks Characterization of collagen helix reduction with 70
denaturation and SDS-PAGE
Sigma bovine type I, electrospun with polymers CD spectroscopy of released collagen confirmed collagen 57
incorporation and preservation of collagen structure

curacy of the solution concentration, molecular mass, and
amino acid content. These data can often be problematic
for collagens, as they are often not well defined. A major
limitation to the interpretation of CD data for collagens and
other fibrous proteins is that the current algorithms used for
conversion of signals to structural information are based on
globular proteins.
Differential Scanning Calorimetry. Differential scanning
calorimetry (DSC) provides direct determination of en-
thalpy (DH) by measuring the temperature dependence of
partial heat capacity.75 The difference in electrical energy
required to raise the temperature of the sample versus that
to raise the temperature of the reference solvent (buffer) is
normalized by the heating rate to calculate the difference
in heat capacity.75 With known masses and temperature
changes, the sample heat capacity and melting temperature
can be calculated.75
DSC is frequently used to characterize the bulk thermal
characteristics of a biomaterial,76 including crosslinked col-
lagen.77 Thermal properties of collagen-based scaffolds
provide information on transitions in the structural state,
reflecting initial primary (chemistry) sequence, structural
state, and degree of crosslinking, and also purity of sam-
ples. There is a wide variation in the protocols used to col-
lect data by DSC. Most often the researcher determines the
apparent Tm, as the thermal unfolding of large proteins like
collagen is usually irreversible. Incomplete drying of sam-
ples can lead to errors in determining the melting tempera-
ture (Tm).78 The rate of sample heating can impact thermal
transitions. For example, collagen-like peptides are heated
at rates no greater than 0.18C min21 for accurate determi-
nation of Tm. Thermal equilibration of collagen may take
as much as 40 min, requiring a heating rate of 0.0048C
min21.62 Despite these issues, rates of heating used in
many DSC studies of collagen are often as fast at 108C
min21.79 A summary of collagens analyzed by DSC is
shown in Table VI.
Chemical Information
X-ray Photoelectron Spectroscopy. X-ray photoelec-
tron spectroscopy (XPS) is commonly employed to charac-
terize the atoms present on the uppermost 10 nm of a

TABLE VI. Differential Scanning Calorimetry Assessments of Collagens

Collagen Material Technique/Findings Reference

Isenglass collagen from fish, bovine hide DSC for Tm’s 80
collagen
Collagen extracted from bovine tendon Higher denaturation temperatures for crosslinked 33
collagen scaffolds
Collagen helix-like peptides DSC for enthalpy change and Tm for various peptides 56
Rat skin collagen Evaluation of crosslinking, age-related changes 76
Rat tail tendon, albino rats Evaluation of collagen film and solution thermal transitions 81
as a function of UV irradiation
Adult bovine femur bone Tm changes as a function of c-irradiation 82
Collagen extracted from bovine muscle Tm changes as a function of sample moisture content 78
Human amnion Thermal characterization as a function of crosslinking 77

Journal of Biomedical Materials Research Part B: Applied Biomaterials

 GUIDE TO COLLAGEN CHARACTERIZATION FOR BIOMATERIAL STUDIES
TABLE VII. X-ray Photoelectron Spectroscopy Analysis of Collagens
Collagen Material
Sigma type I calf skin bovine collagen
Collagen films on poly(e-caprolactone)
Type I bovine collagen (KNC SemedS collagen powder)
from Kensey Nash
material’s surface.83 After bombarding the material with X-
rays, measuring the emitted photoelectron numbers and
energies provides the basis for determination of the
amounts and chemical identities, respectively, of the atoms
present on the material’s surface.83,84 Because of the crude
preparation steps necessary to isolate collagen from animal
tissue, it is important to characterize the chemical composi-
tion to assess contaminants, material homogeneity such as
in film formation in tissue culture wells, and surface modi-
fications with adsorbed or chemically coupled growth, se-
rum, or adhesion factors. The elemental composition of
collagen biomaterials can be determined by XPS,85,86 and
this technique can also be used to confirm the biosynthesis
of collagen by cells.87 XPS has been used to determine the
presence of absorbed serum proteins on tissue-engineering
surfaces,88 and for the presence of a collagen coating to
improve the biocompatibility of titanium implants for bone
growth.89 A summary of collagens characterized by XPS is
shown in Table VII.
Contact Angle. The hydrophobic character of biomate-
rial surfaces influences cell adhesion and spreading, and
these surfaces are often characterized using water contact
angle measurements. Furthermore, contact angle is com-
monly employed as an indicator of surface chemical modi-
fication reactions to track successful chemical coupling
reactions, reflective of a change in surface hydrophobicity/
hydrophilicity. Using an contact angle goniometer, the
angle of contact of a small drop (sessile drop method) of
fluid placed on the surface of interest can be measured.92
Contact angle data for prepared collagen surfaces helps
predict and explain cell attachment data. Fibroblasts adhere
and proliferate preferentially on hydrophilic surfaces with
contact angles below 578.93,94 Surfaces with a contact angle
of 708 and a collagen-grafted polyethylene water contact
angle of 438 supported optimal fibroblast proliferation.95
Contact angles of ultrapure water on 1.0 mg mL21 Cellgen
IPG type I collagen cast films varied from 418 to 718.96
Collagen coatings applied to poly(e-caprolactone) films for
use as implants displayed contact angles that confirmed
changes in hydrophobicity upon grafting collagen to the
TABLE VIII. Contact Angle Assessment of Collagens
Collagen Material
Collagen-grafted polyethylene
Poly(e-caprolactone) grafted with Sigma calf skin
collagen type I
Collagen extracted from rat tail
Journal of Biomedical Materials Research Part B: Applied Biomaterials
271
Technique/Findings Reference
Film coating of polystyrene surfaces via C/N/O bond energies 90
Presence of collagen films by [N]/[O] ratios 91
Collagen-coated titanium surface [C]69.2 [O]17.1 [N]12.6
Collagen source material [C]69.1 [O]17.5 [N]11.7 89
surface.91 The contact angle has also been related to the
synthesis of new collagen, with an increase in contact angle
from 428 to 1168.95 A summary of collagen material in tis-
sue-engineering applications characterized by contact angle
is shown in Table VIII.
Morphological Information
Surface Morphology. Cells respond to surface morphol-
ogy or roughness. In order to characterize the surface to-
pology of a collagen-based biomaterial, a variety of surface
imaging tools can be used including atomic force micros-
copy (AFM), scanning electron microscopy (SEM), envi-
ronmental scanning electron microscopy (ESEM), and light
microscopy. Each technique offers advantages and disad-
vantages associated with sample preparation and resolution,
thus often combinations of several microscopy techniques
are considered. The collection of SEM data requires appro-
priate sample preparation such as sputter coating with gold,
which can dampen surface resolution. The soft nature of
collagen can result in cracking of the coating during imag-
ing. Therefore, ESEM is often a more suitable choice, since
the problems with SEM are avoided, and resolution is usu-
ally sufficient although not as good as SEM.
Atomic Force Microscopy. Nanometer-scale resolution
is achieved with AFM, providing input on scales related to
surface receptors and protein interactions. Contact imaging
via tapping mode using the microscopic probe on the sur-
face of a sample is accompanied by measurements of force
deflection of the cantilever on which the probe is mounted,
to generate the readout of surface morphology. AFM has
been used to characterize the surface roughness of collagen
coated with polystyrene and oxidized polystyrene,80 and to
characterize surface roughness with addition of collagen
films on poly(e-caprolactone).91 The presence of collagen
castings from 0.5 mg mL21 bovine collagen intended to
mask poly-e-caprolactone hydrophobicity increased mean
surface roughness (Ra) from 46 to 60 nm.91 AFM has also
been used to pattern surfaces with collagens and collagen
peptides in the dip pen lithography mode with line resolu-
Technique/Findings Reference
Water contact angle 438 6 38 95
Hydrophobicity characterization, collagen-grafted polymer 458 91
Changes in contact angle with crosslinking 97
272 ABRAHAM ET AL.

TABLE IX. Atomic Force Microscopy Analysis of Collagens

Collagen Material Technique/Findings Reference

Sigma type I calf skin collagen 67-nm banding and 150 nm diameter collagen fibrils via AFM 98
Collagen films on poly(e-caprolactone) Increased roughness (Ra 5 60 nm) with addition of collagen films 91
Collagen from bovine vertebrae 67-nm banding and 50–200 nm diameter collagen fibrils via AFM 99
Type I bovine skin collagen AFM to confirm build up of film coatings 100
Dentin collagen fibers Size distribution and repeat distances 101
Bovine dermal collagen Alignment of collagen fibers with AFM tip 102
Sigma type I calf skin bovine collagen Film coating of polystyrene surfaces via AFM roughness 90

tion to 30- to 50-nm line widths.74 A brief summary of col-
lagens characterized by AFM is shown in Table IX.
Scanning Electron Microscopy. Depending on the spe-
cific model, SEM magnification of 3,0003 to 30,0003 can
be achieved and can be used to image the collagen sub-
strate and cells grown on these surfaces. A sputter-coated
thin layer of gold is required to facilitate imaging of the
surface. For example, the shape of human lung fibroblasts,
IMR-90 cells, growing on collagen has been related to cell
age using SEM, with older cells exhibiting a larger more
irregular morphology.103 Formulas have been developed
relating ratios of maximum to minimum cell length to char-
acterize cell morphology in the study of stromal cell
spreading.104 In collagen scaffolds engineered for artificial
dermis applications, SEM has been used to quantitate the
pore sizes of the scaffolds as well as the extent of collagen
crosslinking.33,105 The orientation of osteoblasts along pat-
terned collagen surfaces was examined using SEM to iden-
tify patterns conducive to bone formation.97 The adhesion
and spreading of bone marrow stem cells on silk biomate-
rial fibers for ligament repair has been imaged directly by
SEM.106 A summary of collagen biomaterials characterized
by SEM is shown in Table X.
Environmental Scanning Electron Microscopy. ESEM
uses a high vacuum, high humidity chamber to image sam-
ples without sputter-coating. Both native and denatured col-
lagen-cell samples have been imaged. For example, to
characterize self-assembled fibrils of collagen composites
for bone-tissue engineering, ESEM was used.108 In tissue-
engineering arterial constructs with collagen coatings,
ESEM was used to image cells.111 A summary of collagen
material in tissue-engineering applications characterized by
ESEM is shown in Table XI.
Light Microscopy. More routine observations of colla-
gen-cell constructs are frequently made with light micros-
copy. For example, light microscopy was used to image the
growth of chondrocytes on a collagen type I/III matrix
designed to improve regenerative capacity of hyaline artic-
ular cartilage.113 Light microscopy techniques are common
and provide at least a gross morphological assessment of
material features and cell interactions as a starting point for
the assessment of biological responses, such as cell adhe-
sion, spreading, and replication (Table XII).
CASE STUDY—CELLULAR REMODELING
COLLAGEN BIOMATERIALS
To illustrate the importance of characterization of collagen
biomaterials, the relationship between collagen biomaterials
and matrix remodeling by human cells will be described.
These studies illustrate the importance of understanding
collagen structure related to cellular aging,19 the retention

TABLE X. Scanning Electron Microscopy Analysis of Collagens

Collagen Material Technique/Findings Reference

Sigma type I collagen–chitosan matrices SEM morphology characterization of collagen crosslinking 107
Collagen extracted from bovine tendon Pore size 50–150 lm, porosity rate 94% 33
Type I collagen extracted from equine tendon SEM to characterize the morphology of the spray-dried 108
collagen composite for bone-tissue engineering
Collagen extracted from rat tail Cell alignment and orientation in comparison to collagen 97
crosslinking
Collagen type I from bovine Achilles tendons Collagen crosslink morphology dependency on freeze drying 39
temperature
Porcine temporomandibular joint disc SEM characterization of collagen fibers 109
2.6% collagen gel (Matrix Pharmaceutical) as SEM to characterize the contact of how alveolar bone with the 110
an adenovirus delivery vehicle dental implant surface
Type I collagen from bovine tendon Assessment of scaffold crosslinking in the presence of various 111
amino acids

Journal of Biomedical Materials Research Part B: Applied Biomaterials

 GUIDE TO COLLAGEN CHARACTERIZATION FOR BIOMATERIAL STUDIES 273

TABLE XI. Environmental Scanning Electron Microscopy Analysis of Collagens

Collagen Material Technique/Findings Ref.

Sigma rat tail type I ESEM to determine extent of cell coverage on collagen endothelial 111
arterial graft constructs
Cell generated ESEM evaluation of collagen fibrotic bundles in the formation of new 112
tissue
Type I collagen extracted from equine ESEM to characterize the self-assembled fibrils of collagen composite 108
tendon for bone-tissue engineering

of differentiation potential of human stem cells toward
bone116 and adipose tissue, impact on phagocytosis,30 and
impact on rates of matrix remodeling to generate new
extracellular matrices.30 These biological outcomes and the
impact of collagen structure/morphology and chemistry on
these outcomes highlights the importance of appropriate an-
alytical characterization of biomaterial substrates for the
study of biological relevance.
Collagen Preparation
Details regarding reagents and related background can be
found in the earlier referenced papers. Rat tail collagen
type I was purchased from Roche Chemicals (Indianapolis,
IN) and collagen films were prepared as we have previously
reported.19 Briefly, collagen was dissolved at 2–88C in ster-
ile filtered 0.1% GAA at 5–10 mg mL21 over at least
3 days for complete dissolution. Once dissolved, the solu-
tion of collagen is diluted to working concentrations just
before use to generate the nondenatured (native) biomate-
rial surfaces. Denaturation is accomplished by a 60-min
treatment in a 508C water bath and confirmed by CD.20
Surface morphology, because of changes in collagen con-
centration, suggests that positive fibroblast growth14 occurs
on surfaces formed from collagen at 78 lg cm22, and addi-
tional concentrations were also applied to tissue culture
plastic (TCP) wells for study. The materials are dried at
room temperature in a vacuum drying oven, typically for
12–48 h, until no liquid remains. Also used in the collagen
evaluation gel (Figure 2) were type I bovine collagen from
Sigma (St. Louis, MO) and human placental collagen from
Calbiochem (San Diego, CA). After the initial assessments
of these various collagen sources, primarily by SDS-PAGE
for this case study, we selected just one collagen commer-
cial source for the remaining characterization assessments,
with a few exceptions, along the lines of the guide. The
exception to this plan was light microscopy.
Cells
IMR-90 human lung primary fibroblasts were purchased
from the American Type Culture Collection (ATCC, Mana-
ssas, VA) and cultured at 378C and 5% CO2 in 20% fetal
bovine serum, 77% eagle minimum essential medium
(MEM), 1% penicillin–streptomycin liquid, 1% L-gluta-
mine-(200 mM), and 1% MEM nonessential amino acids
solution (10 mM). Cells were split at confluence 1:10.
Cells with fewer than 30 cumulative population doubling
levels (PDL) were designated ‘‘young,’’117 and cells with
more than 48 cumulative PDLs were designated as
‘‘aged.’’118 Metabolism of several proteins is twofold
higher in very young cells (PDL 5 22) versus old cells (PDL
5 48).118 Cells were harvested at 70–80% confluence for
inoculation of the collagen surfaces. The IMR-90 cells were
selected for their high rates of collagen synthesis and distinct
morphological changes that occur with aging.119–122
Molecular Mass
The collagen samples were analyzed on an Applied Biosys-
tems Voyager-DE Pro MALDI mass spectrometer in linear
mode (Tufts University Core Protein Chemistry Facility,
Boston, MA). Sample preparation included 10-min heating
at 438C in 8M urea followed by buffer exchange to water
by dialysis. Matrices were DHB and sinipinic acid depend-
ing on molecular mass of the collagen sample, such as the

TABLE XII. Light Microscopy Analysis of Collagens

Collagen Material Technique/Findings Ref.

2.6% collagen gel (Matrix Pharmaceutical) as Light microscopy with histological staining to note the 110
an adenovirus delivery vehicle formation of hew bone
Extracted collagen type II from porcine costa Micrographs of chondrocyte attachment to various ratio 114
polymer:collagen scaffolds
Type I/III collagen bilayer Light microscopy to show layering and porosity of 115
bilayer membrane
Rat tail type I collagen from BD Biosciences Examination of composite layers of dermal tissue 35
engineering construct

Journal of Biomedical Materials Research Part B: Applied Biomaterials

274 ABRAHAM ET AL.

Figure 1. Mass spectroscopy of collagen. Collagen MALDI 4000–100,000 Da shows the major col-
lagen peaks with no evidence of telopeptides. The a chains 94 kDa are visible, and the scan con-
firms the presence of the expected size. Small amounts of typical cleavage products from MMP-1
activity ([1/4] and [3/4] size) are present. The absence of telopeptide products or other size a
chains indicates a relatively pure collagen sample.

presence of telopeptides or other contaminants. Figure 1 tive of degradation in the solution samples. Upon partial
shows MALDI data for the collagen, with expected masses digestion with collagenase, all three sources of collagens
at 94,366.41, 47,365.12, and 31,585.99. Calculated theoreti- showed a reduction in the typical collagen bands (sizes of
cal values for a1(I) 93,915 Da and a2(I) 94,910.7 Da are
expected to vary with species, extent of glycosylation, and
extent of pepsin digestion.40 Collagen 1 alpha chain subu-
nits dissociate under heat treatment (necessary to dissociate
the three alpha chains for MALDI) giving masses roughly
around 31,100 and 45,500, identified by Dreiseward et al. as
a31 and a2140. Molecular masses of C-telopeptides of the a1
chain of type 1 collagen are reported as follows: type I colla-
gen teloPeptide (ICTP) (a1Ca1Ca1H) 10,279 Da, divalent
a1Ca1H 5967 Da, divalent a1Ca2H 6,037 Da, monovalent
a1C smaller 3,730 Da and larger 4,326 Da, and histidinohy-
droxylysinonorleucine HHL crosslinked a1Ca2Ha1H (skin)
7,024 Da.42 The N-terminal propeptide of procollagen type I
(PINP) masses have been reported from 14,313 to 14,360.8
Da.52 No evidence of telopeptides was observed by MALDI
in the collagen samples prepared earlier.

Sodium Dodecyl Sulfate Polyacrylamide Gel

Electrophoresis
To characterize collagen molecular mass for comparison
with MALDI and to attain more quantitative information
(via densitometry), the samples were run by SDS-PAGE
(Figure 2). Type 1 collagens from three vendors (Roche
type I rat tail collagen, Sigma type I bovine collagen, and
Calbiochem type I human collagen) were compared and
were run as untreated solutions and also after treatment
with collagenase. The sources of collagen from Sigma and
Roche had fewer small peptide bands on the gels, indica-
Figure 2. Sodium dodecyl sulfate polyacrylamine gel for source col-
lagen. The gel shows lanes 1–10 (left to right): 1, mark 12 standard;
2, Sigma bovine collagen; 3, roche rat tail collagen; 4, Calbiochem
human placental collagen; 5, mark 12 standard; 6, collagenase
digested Sigma bovine collagen; 7, collagenase digested Sigma bo-
vine collagen; 8, collagenase digested roche rat tail collagen; 9, col-
lagenase digested Calbiochem human placental collagen; and 10,
Calbiochem collagenase (vs. type I collagen). The schematic bar to
the right denotes the expected positions and chain combinations
for collagen. The Sigma and Roche collagen samples are relatively
pure and nondegraded. The reduction of all of the main bands of all
three collagens with collagenase supports the presence of collagen
in the sample.

Journal of Biomedical Materials Research Part B: Applied Biomaterials

 GUIDE TO COLLAGEN CHARACTERIZATION FOR BIOMATERIAL STUDIES
Figure 3. Circular dichroism on collagen solutions. Shown are the
elipticity data for collagen on solutions of 0.125 mg mL21 collagen.
Samples were (—) Denatured at 0.5 mg mL21, (- -) Denatured at
1.0 mg mL21, and (h) Nondenatured. Typical alpha helical protein
structure is observed for the nondenatured collagen. The expected
reductions in CD amplitude at 195 and 221 nm are seen for the
denatured collagen samples. This helical reduction is observed for
samples irrespective of the concentration of the sample during
denaturation.
roughly 100 kDa)52 and an accompanying appearance of
lower molecular mass bands on the gel. The differences in
purity of the three commercial sources of collagen suggest
that additional purification may be appropriate for some of
these source materials depending on the nature of the bio-
logical studies to be conducted.
Circular Dichroism
CD data were collected using a Jasco J 710 Spectropo-
larimeter (Easton, MD) at 258C from 190 nm to 260 nm
with 0.05-nm step resolution, 10 nm min21 collection, an
accumulation rate of 4, response of 16 s, band width at
1.0 nm, and sensitivity of 50 millidegrees. Data were con-
sidered valid for the range of the instrument when the
‘‘HT’’ (Jasco labeling of voltage) photomultiplier voltage
was below 650 V. CD for collagen solutions was collected
at a concentration of 0.125 mg mL21. For collagen films,
drops of 100 lL of 0.5 lg lL21 collagen were dried on
quartz Suprasil (QS) 0.01-mm flat cuvette plates from
Hellma (Plainview, NY) for analysis.
CD was used to examine helical content as a function
of heat denaturation (temperature melts), as a reflection of
structure and chemistry. For both native and denatured
(1 h, 508C heat treated) collagen, the helical content of
the collagen solutions and the films was confirmed by CD
(Figure 3). The CD curve for the nondenatured collagen
shows the expected profile for a helical collagen molecule
including maxima and minima at 221 and 197 nm, respec-
tively. Both of the denatured collagen samples show a sig-
nificant decrease in the positive peak at 221 nm as well as
a significant reduction in the negative peak at 197 nm.
These data confirm the helical content of the nondena-
Journal of Biomedical Materials Research Part B: Applied Biomaterials
275
tured collagen and also indicate a significant reduction of
helical content for denatured collagen that had been heat-
treated. There was no significant difference in helicity as
a function of the concentration at which the collagen was
heat-treated. These same data are presented in Table XIII
after deconvolution using the CDNN program.73 Using a
back propagation network model with a single hidden
layer between input and output, the CDNN program cal-
culated five different secondary structure fractions (helix,
parallel and antiparallel beta-sheet, beta-turn, and random
coil).73
Although Figure 3 shows a reduction in ellipticity at
h222 as is expected for the denatured collagen samples,
Table XII demonstrated the complication of using deconvo-
lution programs based on globular proteins to evaluate CD
data of collagen samples. The CDNN program calculates
less helical content in the nondenatured samples compared
to the denatured samples, an incorrect conclusion based on
visual inspection of Figure 3. Quantification of alpha helix
and helix reduction requires knowledge of the molar con-
centration and the number of residues in the sample.71,72
Errors in estimating these values may also lead to some
errors in applying the CDNN algorithms. The more likely
source of error is in the reference proteins associated with
the CDNN program. The calculations of secondary struc-
ture depend on comparisons to model proteins (globular)
embedded in the CDNN program.73 For accurate ‘‘auto-
mated’’ calculation of molecule shape, a program with
fibrillar reference molecules is needed.
CD was applied directly to collagen films (Figure 4)
composed of 100 lL of 0.5 lg lL21 (50 lg) collagen-dried
over a surface area of 0.8 cm diameter. For the 0.25-
cm2 film area examined, film density was 199 lg cm22.
The spectra for these films had the profile expected for col-
lagens. Likely because of film opacity there was a shift of
the spectra with the maxima shifted to 225 nm and the
minima shifted to 206 nm. However, there remained clear
evidence for significant reduction in helical content for the
films prepared from denatured collagen when compared to
the films prepared from nondenatured collagens. Using the
CDNN program on the film data yielded similar results to
those for the liquid collagen samples.
TABLE XIII. Deconvolution of Circular Dichroism Data Using
CDNN74 Program
% Structurea
Secondary
Structure Denatured at Denatured at
Type Nondenatured 0.5 mg mL21 1.0 mg mL21
% helix 5.9 12.4 10.6
% antiparallel 61.8 42.5 47.8
% parallel 4.6 7.4 6.9
% beta turn 20.8 16.3 16.2
% random coil 6.9 21.5 18.6
a
Each CD sample was run at 0.125 mg mL21. The concentrations listed above
are the concentrations at which the samples were during the 508C denaturing heat
treatment.
276 ABRAHAM ET AL.
Figure 4. Circular dichroism on collagen solutions. Shown are the
elipticity data for collagen films, where 100 lL of 0.5 mg mL21 of
collagen was dried directly on quartz Suprasil (QS) 0.01-mm flat
cuvette plates: (- -) Denatured (—) Gelatin (h) Nondenatured. Typical
alpha helical protein structure is observed for the nondenatured
collagen. The expected reductions in CD amplitude near 195 and
221 nm are seen for the denatured collagen sample. The slight
offset in wavelength is expected due to the width of the cuvette on
which the films were dried. These data confirm that the dried colla-
gen films used in many tissue-engineering studies have similar heli-
cal characteristics to their counterpart solutions.
Differential Scanning Calorimetry
Several attempts were made to characterize collagen ther-
mal transitions using a TA Instruments temperature modu-
lated DSC, TA2920 MDSC (New Castle, DE). For cooling,
a TA Instruments liquid nitrogen cooling accessory
(LNCA) was used. Dry nitrogen gas was purged into the
TMDSC cell at a flow rate of 20 mL min21. The standard
DSC was carried out with a heating rate of 58C min21
from 220 to 2008C and a cooling rate of 208C min21.
Roughly 5 mg of a 0.5 mg mL21 sample of dried nondena-
tured collagen film was added to an aluminium DSC pan.
A similar weight empty reference pan was used as a con-
trol (Figure 5). Although a thermal transition is suggested
by the data, integration for calculation of Tm is not possi-
ble. This limitation is due to the level of noise in the data,
a function both of the difficulty of adding sufficient weight
of collagen into a sample pan and also the sensitivity of
the instrument. As described earlier, with a slower heating
rate (not available on all DSC systems) or a more advanced
DSC system, improved thermal transitions could be deter-
mined, as are reported in the literature.
X-ray Photoelectron Spectroscopy
Sample surfaces were characterized using a Surface Sci-
ence Instruments (Mountain View, CA) Model SSX-100
XPS. Each sample was subjected to triplicate elemental
scans: 1000 nm, resolution 4, window 100 eV, at varied
positions in the well. Scans were conducted with a charge
neutralizer flood gun at 5 eV and with a nickel wire mesh
over the sample surface to prevent charging. After survey
scans to identify elements present, environmental scans for
C, N, and O were conducted. XPS was used to verify the
C/N/O ratios expected for uncontaminated TCP and colla-
gen films.29 To assess contaminants in the collagen prepa-
ration, the presence of the expected collagen C/N/O ratios,
and film coverage over all areas of sampling, XPS analysis
was conducted on denatured and nondenatured collagen
films [Figure 6(a,b)]. The elemental scans showed only car-
bon (C), nitrogen (N), and oxygen (O). The ratios of C, N,
and O were those expected for TCP and collagen. Detec-
tion is possible to about 0.1 atomic %, with accuracy of
element concentrations at less than 10 lM.123 All sample
concentrations of collagen, denatured and nondenatured,
showed similar C/N/O ratios to confirm that the collagen
matrices were free from significant contamination with sili-
cone, which can sometimes be a problem in collagen prep-
arations, and that the coatings on the plates were
continuous.
Contact Angle
Static contact angles were measured using a Rame-Hart
NRL CA contact angle Gonimeter (Mountain Lakes, NJ).
Measurements were made minimally in triplicate by the
sessile drop method. Hydrophobicities of denatured and
nondenatured collagen films of concentrations 15.6 lg
cm22 to 780 lg cm22, were assessed via water contact
angle measurements (Figure 7). From collagen concentra-
tions 32.5–468 lg cm22, there was an increase in contact
angle with increasing collagen concentration for both the
denatured and the nondenatured samples. The increase in
hydrophobicity with increasing concentration of collagen
may be related to the accompanying increase in surface
roughness. Both the surface roughness and the hydropho-
bicity increases with concentration may be related to the
better cell growth and survival as observed for cells grown
on 78 and 156 lg cm22 collagen matrices.19 An increase
Figure 5. Differential scanning calorimetry shown is the heat flow
data for DSC of roughly 5 mg of a 0.5 mg mL21 collagen sample.
Data were collected in a TA Instruments temperature modulated
DSC, TA2920 MDSC (New Castle, DE). The graph demonstrates the
challenges in collecting data for collagen samples due to equipment
limitations or sample preparation/amounts.
Journal of Biomedical Materials Research Part B: Applied Biomaterials
 GUIDE TO COLLAGEN CHARACTERIZATION FOR BIOMATERIAL STUDIES 277

Figure 6. (a) X-ray photoelectron spectroscopy on denatured collagen films. Surface content of the
following atoms is shown (j) Carbon, ( ) Nitrogen, (h) Oxygen. TC is tissue culture plastic. GAA is
tissue culture plate with 0.1% GAA (the collagen solution buffer) as control. The presence of colla-
gen-associated atoms in samples at each film thickness confirms that for each area of the film
tested, the plate is fully covered with collagen. The absence of significant content of noncollagen
atoms indicates a lack of contamination of the samples. (b) X-ray photoelectron spectroscopy on
nondenatured collagen films. Surface content of the following atoms is shown (j) Carbon, ( ) Nitro-
gen, ( ) Oxygen.

in hydrophobicity found with increasing collagen concen-
tration may relate to the phenomena of ‘‘pillars’’ in surface
topology leading to increased surface roughness.124 These
data may help explain the poorer cell growth observed on
higher collagen concentrations.19
Atomic Force Microscopy
AFM imaging was performed in tapping mode on a Dimen-
sion 3100 Nanoscope III with tapping mode etched silicon
probes (TESP), SMP, and DNP20 contact mode probes
(Digital Instruments, Santa Barbara, CA). TESP probes
have a cantilever length of 225 lm and a spring constant
of 1–5 N m21. Rotated TESP AFM tips have the same
spring constants and cantilever lengths as the TESP probes,
Figure 7. Contact angle on collagen films. Shown is the water con-
tact angle for the following surfaces: (h) tissue culture plastic; ( )
glacial acetic acid; ( ) denatured collagen; (j) nondenatured colla-
gen. The increase in hydrophobicity with collagen concentration
may be related with increasing surface roughness at higher collagen
concentrations and illustrates the challenges in assessing surface
energy of films where surface morphology (smoothness) is an issue.
except that the tips rotated at a 158 angle to allow for bet-
ter visualization of high-aspect ratio features. Imaging was
achieved with 5- to 100-lm scan widths at rate of 0.5–
1.0 Hz. Data were collected on undisturbed collagen sam-
ples in phosphate buffered saline (PBS) in a 35-mm dish.
Large differences in surface roughness as a function of col-
lagen concentration were observed. The phase data surface
images typical of a 780 lg cm22 denatured collagen sam-
ple are shown in Figure 8. From the angled image showing
a portion of the surface in an edge on view, it can be
observed that the collagen surface has extensive roughness.
From the perspective of the cell, this 100 lm by 100 lm
collagen surface offers a bed of blunt spikes. The increas-
ing roughness observed with increasing collagen concentra-
tion corresponded to decreasing cell viability on the highest
collagen concentrations.19
To determine whether the increase in hydrophobicity
with increasing concentration correlated to surface rough-
ness, AFM was conducted with samples of varying concen-
trations. The AFM height images for 0.1% GAA on TCP
(GAA/TCP) and on denatured collagen concentrations of
15.6 to 780 lg cm22 are shown in Figure 9(a–g) . Each
scan is for a 100-lm square area. The height scales are
200 nm. The GAA on TCP plate surface indicates a regular
pattern of small spikes. All 35-mm TCP plates wet, dry,
and with and without GAA, showed similar patterns. Even
at the lowest collagen concentrations the collagen films
filled the valleys of the TCP topology and create a
smoother surface. Also visible in the collagen film images
are a series of more frequent surface topology spikes with
increasing concentration. The 15.6 and 31.5 lg cm22 sam-
ples show increasing smoothing of the TCP topology and a
small number of 50-nm to 200-nm spikes. For the samples
prepared from 78 and 156 lg cm22, there was no evidence
of the TCP topology, and the 50-nm to 200-nm spikes
associated with the collagen film are more frequent. The

Journal of Biomedical Materials Research Part B: Applied Biomaterials

278 ABRAHAM ET AL.

Figure 8. AFM image of collagen surface roughness. Digital Instruments Nanoscope, (left) 100-lm
scan size, 0.5003-Hz scan rate, 256 samples, phase data image, 908 data scale; (right) 30-lm scan
size, 1.0-Hz scan rate, 256 samples, phase data image, 908 data scale. The surface roughness and
scale of surface topology of a high concentration collagen film is observed.

468 and 780 lg cm22 collagen samples have mountain-like
spikes up to and exceeding 200 nm in their topology cover-
ing most of the film surface. This increasing surface rough-
ness observed visually can also be quantitated using root
mean square (RMS) roughness calculations in the Digital
Instruments (DI) software (Figure 10). RMS roughness
increased significantly from 29 for GAA on TCP to 71 for
the denatured collagen films at 780 lg cm22. The increas-
ing surface roughness may help explain cell growth and
survival on the samples prepared from the solutions of col-
lagen containing 78 and 156 lg cm22.19 To confirm that
surface topology was due to the collagen film, any colla-
gen-related structure should be altered by heat treatment at
658C and to a greater extent at 858C. In Figure 11(a–c) a
156 lg cm22 denatured collagen sample is shown before
and after 65 and 858C heating intended to reduce surface

Figure 9. (a–g) Denatured collagen (0, 16, 31, 78, 156, 468, 780 lg cm22) surface roughness. Digi-
tal Instruments Nanoscope, 100-lm scan size, 0.5003-Hz scan rate, 512 samples, height data
image, 200-nm data scale. The increasing surface roughness of collagen films with increasing colla-
gen concentration is observed. [Color figure can be viewed in the online issue, which is available at
www.interscience.wiley.com.]

Journal of Biomedical Materials Research Part B: Applied Biomaterials

 GUIDE TO COLLAGEN CHARACTERIZATION FOR BIOMATERIAL STUDIES 279

fraction of the films survived processing to permit the col-

Figure 10. Collagen surface roughness quantitation of AFM data.
Route mean square (RMS) roughness as quantified by the Digital
Imaging AFM software was calculated for 100-lm areas, then aver-
aged over three samples at each collagen concentration. The
increase in RMS roughness with increasing collagen concentration
is shown. Statistical analysis was preformed for 3–10 samples at
each concentration.
roughness due to the collagen structure by ‘‘melting out’’
the topology. The heating steps resulted in statistically sig-
nificant reduced roughness. These observations support the
hypothesis that the roughness observed in nonheated films
was due to the collagen structure.
Scanning Electron Microscopy
Cells that had been grown on collagen surfaces were fixed
with 2.5% glutaraldehyde in 0.1M sodium cacodylate, then
washed with 0.1M sodium cacodylate.107 Samples were
dehydrated by soaking in a graduated series of alcohol
washes. Samples were sputter-coated using a Polaron
SC502 Sputter Coater (Fisson Instruments, UK). SEM
images were collected using a JEOL LSM-840-A SEM
(Peabody, MA). Some of the denatured collagen film sam-
ples collapsed under the force of the sputter coating and
the vacuum of the SEM imaging; thus, these films lacked
sufficient structural integrity for this analysis. A sufficient
lection of images. The roughness of nondenatured collagen
samples for GAA on TCP and for the collagen samples
prepared from 78, 156, 312, and 780 lg cm22 are shown
both with and without cells via SEM imaging (Figure 12).
Images are shown at 3,0003 and 1,0003 to depict both
single cells and groups of cells. Both PDL 33 and PDL 48
cells grown on nondenatured collagen are shown. For both
the image rows without cells, increased surface topology
roughness can be observed with increasing collagen con-
centration. For both the PDL 33 and the PDL 48 cells, bet-
ter adhesion to the collagen films, more contact points
(suggesting matrix phagocytosis), and fewer irregular cells
shapes were observed at the middle collagen concentrations.
The cells grown on 156 lg cm22 nondenatured collagen
were observed with the fewest of the age-related morphol-
ogy indicators, demonstrated fewer senescence related fea-
tures and also supported by assessments of cell morphology,
biochemistry, and transcript measures related to these age-
related outcomes.23,108 These observations correspond to
quantitation of senescence-associated b-galactosidase assays
for cell function.19 Since cell death and low proliferation
rates are a common problem in cell expansion and use on
biomaterial matrices, identifying collagen matrix properties
that lead to improved cell heath may allow for the design of
more successful tissue engineering constructs.
The SEM images of the collagen surfaces support, along
with the AFM images, increased surface roughness of the
collagen surfaces with increasing concentration used in film
preparation. The SEM images also allow assessment of cell
morphology that is not available via AFM. IMR-90 fibro-
blasts showed significant changes in cell morphology with
age. Increases in cell size, cell surface roughness, decreased
proliferation rates, and increased senescence-associated b-
galactosidase expression are all associated with aging in
IMR-90 cells.125 From the SEM images, it can be con-
cluded that the cells grown on the highest concentrations of
nondenatured collagen were more phenotypically aged than
cells grown on the 156 lg cm22 nondenatured collagens.

Figure 11. (a–c) Denatured collagen (156 lg cm22) surface roughness—(a) nonheated (b) reduced
by heating at 658C, and (c) reduced by heating at 858C. Digital Instruments Nanoscope, 100-lm
scan size, 0.5003-Hz scan rate, 512 samples, height data image, 200-nm data scale. The reduction
in surface roughness with heating is shown to demonstrate that the surface roughness observed is
related to collagen structure. [Color figure can be viewed in the online issue, which is available at
www.interscience.wiley.com.]

Journal of Biomedical Materials Research Part B: Applied Biomaterials

280 ABRAHAM ET AL.

Figure 12. Scanning electron microscopy on nondenatured collagen films. For all six rows of
images, the collagen concentration increases (left to right) 0.1% GAA and 0, 78, 156, 312, and
780 lg mL21 collagen. The rows (top to bottom) show 10-lm size images of collagen with no cells,
10-lm size passage 19 ‘‘old’’ cells, 10 lm size passage 13 ‘‘young’’ cells, 30 lm size images of col-
lagen with no cells, 30 lm size Passage 19 ‘‘old’’ cells, and 30 lm size passage 13 ‘‘young’’ cells.
Increasing surface roughness is observed as a function of increasing collagen concentration. Cells
appear ‘‘younger’’ (smaller, smoother edges, fewer processes) on the lower collagen concentrations
versus the higher collagen concentrations. Cells similarly appear ‘‘younger’’ on the denatured colla-
gen versus the nondenatured collagen.

Journal of Biomedical Materials Research Part B: Applied Biomaterials

 GUIDE TO COLLAGEN CHARACTERIZATION FOR BIOMATERIAL STUDIES 281

Figure 13. Environmental scanning electron microscopy on collagen-cell surfaces. (A) P21 cells,
0.1% glacial acetic acid, (B) P21 cells, denatured collagen, 156 lg cm22, (C) P21 cells, denatured
collagen, 468 lg cm22, (D) P21 cells, denatured collagen, 780 lg cm22, (E) P21 cells, nondenatured
collagen, 156 lg cm22, (F) P21 cells, denatured collagen, 468 lg cm22, (G) P21 cells, nondenatured
collagen, 780 lg cm22. Similar to SEM, increasing surface roughness is observed as a function of
increasing collagen concentration. Cells appear ‘‘younger’’ (smaller, smoother edges, fewer proc-
esses) on the lower collagen concentrations versus the higher collagen concentrations. Cells simi-
larly appear ‘‘younger’’ on the denatured collagen versus the nondenatured collagen. In ESEM, the
lack of a gold coating allows for better visualization of the cellular reaction to the collagen surfaces.
Individual cells can be observed to be spreading on ‘‘favorable’’ collagen surfaces or contracting
from ‘‘nonfavorable’’ collagen surfaces.

Environmental SEM
ESEM images were collected on a FEI model Quanta 200
ESEM (Hillsboro, OR) with a tungsten filament and a pelt-
ier stage. Working distances were 6–10 mm, tilt was 258,
temperature was 48C, and pressure was 2.75 Torr. ESEM
was used to observe cellular interactions with collagen mat-
rices with as little sample processing (fewer artifacts) as
possible. ESEM images for GAA on TCP and collagen
samples from 156, 468, and 780 lg cm22 with cells are
shown in Figure 13. For all collagen concentrations, the
cells grown on denatured collagen samples are more regu-
larly shaped and appear more attached to the underlying
collagen matrices, and may appear more closely in contact
with the substrate due to phagocytosing of the collagen ma-
trix. The cells grown on the denatured collagen are simi-
larly healthy when compared to those on the GAA on
tissue culture plastic wells. The cell in the 780 lg cm22
image is particularly notable for its nonattached appear-
ance, as if it is trying to remove itself from the surface.
The images show some of the collagen surface characteris-
tics as well as the cellular interactions with the collagen
substrates. This provides some assessment of the morphol-
ogy of the IMR-90 cells in relation to cell age, which may
be helpful in selecting collagens for tissue-engineering mat-
rices with the goal of promoting better cell health and sur-
vival. Compared with SEM, ESEM offers benefits in
avoiding sputter coating with metal that can generate arti-
facts, and generates fewer vacuum-induced artifacts. How-
ever, there is generally a loss in resolution with ESEM,
which for cell characterization studies is usually not an
issue in terms of gross morphological assessments.
Optical Microscopy
Cell images were captured using a Zeiss Axiovert S100
microscope (Thornwood, NY) equipped with a Sony
Exwave HAD 3CCD color video camera (Shinagawa, To-
kyo, Japan). Images were processed with Scion Image for
windows v4.0.2 software (Fredrick, MD). Image overlays
for fluorescence images used Adobe Photoshop 5.0 or Corel
Photo-Paint 10 software. In order to determine the extent
of cell growth and the general morphology of the cells,
light microscopy images were collected. Several light mi-
croscopy images of IMR-90 cells on Sigma and Roche type
I collagen at various concentrations are shown in Figure 14.
Although the light microscope images are not as detailed
as those from AFM and SEM, they provide immediate,
low-cost images, that can be collected in most laboratories.

Journal of Biomedical Materials Research Part B: Applied Biomaterials

282 ABRAHAM ET AL.

Figure 14. Light microscopy images shown are IMR-90 cells on (A) 10 mg mL21 Sigma collagen,
(B) 5 mg mL21 Sigma collagen, (C) 1 mg mL21 Sigma collagen, (D) 0.1 mg mL21 Sigma collagen,
(E) tissue culture plastic, (E) 1 mg mL21 Roche Collagen, and (F) 1 mg mL21 Roche Collagen. Light
microscopy is a convenient tool to show cell numbers and general morphology, while little informa-
tion is gained about the collagen surface. [Color figure can be viewed in the online issue, which is
available at www.interscience.wiley.com.]

Additionally, light microscopy images can be taken without
destructive preparation of the samples, allowing for further
cell culture postimaging.
Case Study—Conclusions
Several analytical tools were used to determine details of col-
lagen structure, morphology, and chemistry in this ‘‘case
study.’’ These methods were used to assess increases in colla-
gen roughness and the observance of cell morphologies with
relationship to helicity and topography. The cells grown on
denatured collagens appeared healthier than those grown on
nondenatured collagens and higher collagen substrate con-
centrations. We have also related these results to the cell
aging and collagen trafficking results reported elsewhere.19,30
We have retained a focus in the analytical guide and in
the case study on commercial sources of collagens to pro-
vide a starting point for assessments that can be considered.
Many researchers prefer to isolate their own collagen, such
as from tendons or rat tails. The isolation procedures for
these extractions are well-described in the literature. Once
carried out, similar analytical tools as described in this pa-
per, as a guide to assessments of the isolated collagens, can
be considered for these tissue-derived sources of materials.
In a similar fashion, variations in the presence of telopepti-
des, contaminating ECM components, or crosslinking pro-
cedures, are some issues that may be encountered, which
can at least in part be assessed with the tools outlined in
this guide.

SUMMARY
The characterization of biomaterial matrices is essential to
the design of intelligent tissue engineered matrices as well
as to provide comparative data among studies from differ-
ent laboratories. Cell responses to collagen matrices depend
on many features of these biomaterials, including secondary
and tertiary structure, chemical composition, hydrophobic-
ity, and surface roughness, among others. Techniques that
provide more thorough characterization of the matrices are
crucial to identifying optimal collagen-based biomaterials
for matrix fabrication and to relate the data to cell biology.
We have described the benefits and limitations of several
of the methods of characterization that can be applied to
collagen biomaterials and provide an initial basis for intra-
and interstudy comparisons. These types of biomaterial
characterizations are beneficial to guide the design of bio-
material tissue-engineering matrices and for the evaluation
of cellular responses to these matrices.
REFERENCES
1. Burgeson RE, Nimni ME. Collagen types. Molecular structure
and tissue distribution. Clin Orthop 1992;282:250–272.
2. Patino MG, Neiders ME, Andreana S, Noble B, Cohen RE.
Collagen: An overview. Implant Dent 2002;11:280–285.
3. Friess W. Collagen—Biomaterial for drug delivery. Eur J Pharm
Biopharm 1998;45:113–136.
4. Burke JF, Yannas IV, Quinby WC Jr, Bondoc CC, Jung WK.
Successful use of a physiologically acceptable artificial skin
in the treatment of extensive burn injury. Ann Surg 1981;194:
413–428.
5. Eaglstein WH, Falanga V. Tissue engineering and the devel-
opment of Apligraf, a human skin equivalent. Cutis 1998;62(1
Suppl):1–8.
6. Meaney Murray M, Rice K, Wright RJ, Spector M. The effect
of selected growth factors on human anterior cruciate liga-
ment cell interactions with a three-dimensional collagen-GAG
scaffold. J Orthop Res 2003;21:238–244.
7. Angele P, Kujat R, Nerlich M, Yoo J, Goldberg V, Johnstone
B. Engineering of osteochondral tissue with bone marrow
mesenchymal progenitor cells in a derivatized hyaluronan-gel-
atin composite sponge. Tissue Eng 1999;5:545–554.

Journal of Biomedical Materials Research Part B: Applied Biomaterials

 GUIDE TO COLLAGEN CHARACTERIZATION FOR BIOMATERIAL STUDIES
8. Brekken RA, Sage EH. SPARC, a matricellular protein: At
the crossroads of cell-matrix. Matrix Biol 2000;19:569–580.
9. Blair HC, Zaidi M, Schlesinger PH. Mechanisms balancing
skeletal matrix synthesis and degradation. Biochem J 2002;
364(Pt 2):329–341.
10. Hay ED. Cell Biology of Extracellular Matrix. New York:
Plenum; 1991. xvii, 468 pp.
11. Alberts B, Johnson A, Lewis J, Raff M, Roberts K, Walter P.
Molecular Biology of the Cell. New York: Garland Science;
2002.
12. Ricard-Blum S, Dublet B, van der Rest M. Unconventional
Collagens. Grenoble Cedex: Oxford University Press; 2000.
155 pp.
13. Badii F, Howell NK. Elucidation of the effect of formalde-
hyde and lipids on frozen stored cod collagen by FT-Raman
spectroscopy and differential scanning calorimetry. J Agric
Food Chem 2003;51:1440–1446.
14. Awad HA, Butler DL, Harris MT, Ibrahim RE, Wu Y, Young
RG, Kadiyala S, Boivin GP. In vitro characterization of mes-
enchymal stem cell-seeded collagen scaffolds for tendon
repair: Effects of initial seeding density on contraction
kinetics. J Biomed Mater Res 2000;51:233–240.
15. Zimmermann WH, Melnychenko I, Eschenhagen T. Engi-
neered heart tissue for regeneration of diseased hearts. Bioma-
terials 2004;25:1639–1647.
16. Viravaidya K, Shuler ML. The effect of various substrates on
cell attachment and differentiation of 3T3-F442A preadipo-
cytes. Biotechnol Bioeng 2002;78:454–458.
17. Bornstein MB. Reconstituted rattail collagen used as substrate
for tissue cultures on coverslips in Maximow slides and roller
tubes. Lab Invest 1958;7:134–137.
18. Michalopoulos G, Pitot HC. Primary culture of parenchymal
liver cells on collagen membranes. Morphological and bio-
chemical observations. Exp Cell Res 1975;94:70–78.
19. Abraham LC, Vorrasi J, Kaplan DL. Impact of collagen struc-
ture on matrix trafficking by human fibroblasts. J Biomed
Mater Res A 2004;70:39–48.
20. Volloch V, Kaplan D. Matrix-mediated cellular rejuvenation.
Matrix Biol 2002;21:533–543.
21. Kleinman HK, McGarvey ML, Liotta LA, Robey PG, Trygg-
vason K, Martin GR. Isolation and characterization of type IV
procollagen, laminin, and heparan sulfate proteoglycan from
the EHS sarcoma. Biochemistry 1982;21:6188–6193.
22. Bornstein P, Traub W. The Chemistry and Biology of Colla-
gen. New York, NY: Academic Press; 1979.
23. Gallop PM, Seifter S. Preparation and properties of soluble
collagens. Meth Enzymol 1963;6:635–641.
24. Niyibizl C, Fietzek PP, van der Rest M. Human placenta type
V collagens. Evidence for the existence of an alpha 1(V)
alpha 2(V) alpha 3(V) collagen molecule. J Biol Chem 1984;
259:14170–14174.
25. Kordula T, Banbula A, Macomson J, Travis J. Isolation and
properties of stachyrase A, a chymotrypsin-like serine proteinase
from Stachybotrys chartarum. Infect Immun 2002;70:419–421.
26. Einbinder J, Schubert M. Binding of mucopolysaccharides
and dyes by collagen. J Biol Chem 1951;188:335–341.
27. Trentham DE, Townes AS, Kang AH. Autoimmunity to type
II collagen an experimental model of arthritis. J Exp Med
1977;146:857–868.
28. Miller EJ. Isolation and characterization of a collagen from
chick cartilage containing three identical a chains. Biochemis-
try 1971;10:1652–1659.
29. Abraham LC, Dice JF, Lee K, Kaplan DL. Phagocytosis and
remodeling of collagen matrices. Exp Cell Res 2007;313:
1045–1055.
30. Abraham LC, Dice JF, Finn PF, Mesires NT, Lee K, Kaplan
DL. Extracellular matrix remodeling—Methods to quantify
cell–matrix interactions. Biomaterials 2007;28:151–61.
Journal of Biomedical Materials Research Part B: Applied Biomaterials
283
31. Basle MF, Grizon F, Pascaretti C, Lesourd M, Chappard D.
Shape and orientation of osteoblast-like cells (Saos-2) are
influenced by collagen fibers in xenogenic bone biomaterial.
J Biomed Mater Res 1998;40:350–357.
32. Shimizu T, Yamato M, Kikuchi A, Okano T. Cell sheet engi-
neering for myocardial tissue reconstruction. Biomaterials
2003;24:2309–2316.
33. Zhang L, Ma D, Wang F, Zhang Q. The modification of scaf-
fold material in building artificial dermis. Artif Cells Blood
Substit Immobil Biotechnol 2002;30:319–332.
34. Ng KW, Khor HL, Hutmacher DW. In vitro characterization
of natural and synthetic dermal matrices cultured with human
dermal fibroblasts. Biomaterials 2004;25:2807–2818.
35. Torkian BA, Yeh AT, Engel R, Sun CH, Tromberg BJ, Wong
BJ. Modeling aberrant wound healing using tissue-engineered
skin constructs and multiphoton microscopy. Arch Facial Plast
Surg 2004;6:180–187.
36. Friess W. Collagen in drug delivery and tissue engineering.
Adv Drug Deliv Rev 2003;55:1529–1530.
37. Knowles GC, McKeown M, Sodek J, McCulloch CA. Mecha-
nism of collagen phagocytosis by human gingival fibroblasts:
Importance of collagen structure in cell recognition and inter-
nalization. J Cell Sci 1991;98(Pt 4):551–558.
38. Diener A, Nebe B, Luthen F, Becker P, Beck U, Neumann
HG, Rychly J. Control of focal adhesion dynamics by mate-
rial surface characteristics. Biomaterials 2005;26:383–392.
39. Buijtenhuijs P, Buttafoco L, Poot AA, Daamen WF, van Kup-
pevelt TH, Dijkstra PJ, de Vos RA, Sterk LM, Geelkerken
BR, Feijen J, Vermes L. Tissue engineering of blood vessels:
Characterization of smooth-muscle cells for culturing on col-
lagen-and-elastin-based scaffolds. Biotechnol Appl Biochem
2004;39(Pt 2):141–149.
40. Dreisewerd K, Rohlfing A, Spottke B, Urbanke C, Henkel W.
Characterization of whole fibril-forming collagen proteins of
types I, III, and V from fetal calf skin by infrared matrix-
assisted laser desorption ionization mass spectrometry. Anal
Chem 2004;76:3482–3491.
41. Schonherr E, Witsch-Prehm P, Harrach B, Robenek H, Rau-
terberg J, Kresse H. Interaction of biglycan with type I colla-
gen. J Biol Chem 1995;270:2776–2783.
42. Eriksen HA, Sharp CA, Robins SP, Sassi ML, Risteli L, Ris-
teli J. Differently cross-linked and uncross-linked carboxy-ter-
minal telopeptides of type I collagen in human mineralised
bone. Bone 2004;34:720–727.
43. Vorm O, Reopstorff P, Mann M. Improved Resolution and
very high senisitivity in MALDI TOF of matrix surfaces
made by fast evaporation. Anal Chem 1994;66:3281–3287.
44. Gusev A, Wilkinson W, Proctor A, Hercules D. Improvement
of signal reproducibility and matrix/comatrix effects in
MALDI analysis. Anal Chem 1995;67:1034–1041.
45. Demers LM, Costa L, Lipton A. Biochemical markers and
skeletal metastases. Cancer 2000;88(12 Suppl):2919–2926.
46. Van den Steen PE, Proost P, Grillet B, Brand DD, Kang AH,
Van Damme J, Opdenakker G. Cleavage of denatured natural
collagen type II by neutrophil gelatinase B reveals enzyme
specificity, post-translational modifications in the substrate,
and the formation of remnant epitopes in rheumatoid arthritis.
Faseb J 2002;16:379–389.
47. Wei L, Sun XJ, Wang Z, Chen Q. CD95-induced osteoar-
thritic chondrocyte apoptosis and necrosis: Dependency on
p38 mitogen-activated protein kinase. Arthr Res Ther 2006;8:
R37.
48. Bonner WM, Laskey RA. A film detection method for trit-
ium-labelled proteins and nucleic acids in polyacrylamide
gels. Eur J Biochem 1974;46:83–88.
49. Miles CA, Ghelashvili M. Polymer-in-a-box mechanism for
the thermal stabilization of collagen molecules in fibers. Bio-
phys J 1999;76:3243–3252.
284 ABRAHAM ET AL.
50. Ellerbroek SM, Wu YI, Stack MS. Type I collagen stabiliza-
tion of matrix metalloproteinase-2. Arch Biochem Biophys
2001;390:51–56.
51. Tam EM, Moore TR, Butler GS, Overall CM. Characteriza-
tion of the distinct collagen binding, helicase and cleavage
mechanisms of matrix metalloproteinase 2 and 14 (gelatinase
A and MT1-MMP): The differential roles of the MMP hemo-
pexin c domains and the MMP-2 fibronectin type II modules
in collagen triple helicase activities. J Biol Chem 2004;
279:43336–43344.
52. Brandt J, Krogh TN, Jensen CH, Frederiksen JK, Teisner B.
Thermal instability of the trimeric structure of the N-terminal
propeptide of human procollagen type I in relation to assay
technology. Clin Chem 1999;45:47–53.
53. Sykes B, Puddle B, Francis M, Smith R. The estimation of
two collagens from human dermis by interrupted gel electro-
phoresis. Biochem Biophys Res Commun 1976;72:1472–
1480.
54. Greenfield NJ. Methods to estimate the conformation of pro-
teins and polypeptides from circular dichroism data. Anal
Biochem 1996;235:1–10.
55. Consonni R, Zetta L, Longhi R, Toma L, Zanaboni G, Tenni
R. Conformational analysis and stability of collagen peptides
by CD and by 1H- and 13C-NMR spectroscopies. Biopolymers
2000;53:99–111.
56. Mizuno K, Hayashi T, Peyton DH, Bachinger HP. Hydroxyla-
tion-induced stabilization of the collagen triple helix. Acetyl-
(glycyl-4(R)-hydroxyprolyl-4(R)-hydroxyprolyl)(10)-NH(2)
forms a highly stable triple helix. J Biol Chem 2004;279:
38072–38078.
57. Stankus JJ, Guan J, Wagner WR. Fabrication of biodegradable
elastomeric scaffolds with sub-micron morphologies. J Biomed
Mater Res A 2004;70:603–614.
58. Hayashi T, Curran-Patel S, Prockop DJ. Thermal stability of
the triple helix of type I procollagen and collagen. Precautions
for minimizing ultraviolet damage to proteins during circular
dichroism studies. Biochemistry 1979;18:4182–4187.
59. Babu IR, Ganesh KN. Enhanced triple helix stability of colla-
gen peptides with 4R-aminoprolyl (Amp) residues: Relative
roles of electrostatic and hydrogen bonding effects. J Am
Chem Soc 2001;123:2079–2080.
60. Peltonen L, Palotie A, Hayashi T, Prockop DJ. Thermal sta-
bility of type I and type III procollagens from normal human
fibroblasts and from a patient with osteogenesis imperfecta.
Proc Natl Acad Sci USA 1980;77:162–166.
61. Vomund AN, Braddock SR, Krause GF, Phillips CL. Potential
modifier role of the R618Q variant of proalpha2(I)collagen in
type I collagen fibrillogenesis: In vitro assembly analysis. Mol
Genet Metab 2004;82:144–153.
62. Leikina E, Mertts MV, Kuznetsova N, Leikin S. Type I colla-
gen is thermally unstable at body temperature. Proc Natl
Acad Sci USA 2002;99:1314–1318.
63. Wang X, Li X, Bank RA, Agrawal CM. Effects of collagen
unwinding and cleavage on the mechanical integrity of the
collagen network in bone. Calcif Tissue Int 2002;71:186–
192.
64. Ellerbroek SM, Wu YI, Overall CM, Stack MS. Functional
interplay between type I collagen and cell surface matrix met-
alloproteinase activity. J Biol Chem 2001;276:24833–24842.
65. Day JS, Van Der Linden JC, Bank RA, Ding M, Hvid I,
Sumner DR, Weinans H. Adaptation of subchondral bone in
osteoarthritis. Biorheology 2004;41:359–368.
66. Xu Y, Keene DR, Bujnicki JM, Hook M, Lukomski S. Strep-
tococcal Scl1 and Scl2 proteins form collagen-like triple heli-
ces. J Biol Chem 2002;277:27312–27318.
67. Jenness DD, Sprecher C, Johnson WCJ. Circular dichroism of
collagen, gelatin, and poly(proline) II in the vacuum ultravio-
let. Biopolymers 1976;15:513–521.
68. Brown EM, Dudley RB, Elsetinow AR. A conformational
study of collagen as affected by tanning procedures. J Am
Leather Chem Assoc 1997;92:225–233.
69. Rossi A, Zuccarello LV, Zanaboni G, Monzani E, Dyne KM,
Cetta G, Tenni R. Type I collagen CNBr peptides: Species
and behavior in solution. Biochemistry 1996;35:6048–6057.
70. Yamauchi K, Takeuchi N, Kurimoto A, Tanabe T. Films of
collagen crosslinked by S S bonds: Preparation and charac-
terization. Biomaterials 2001;22:855–863.
71. Chakrabartty A, Kortemme T, Baldwin RL. Helix propensities
of the amino acids measured in alanine-based peptides with-
out helix-stabilizing side-chain interactions. Protein Sci 1994;
3:843–852.
72. Cochran DA, Doig AJ. Effect of the N2 residue on the stabil-
ity of the a -helix for all 20 amino acids. Protein Sci 2001;
10:1305–1311.
73. Bohm G, Muhr R, Jaenicke R. Quantitative analysis of protein
far UV circular dichroism spectra by neural networks. Protein
Eng 1992;5:191–195.
74. Wilson DL, Martin R, Hong S, Cronin-Golomb M, Mirkin
CA, Kaplan DL. Surface organization and nanopatterning of
collagen by dip-pen nanolithography. Proc Natl Acad Sci
USA 2001;98:13660–13664.
75. Jelesarov I, Bosshard HR. Isothermal titration calorimetry and
differential scanning calorimetry as complementary tools to
investigate the energetics of biomolecular recognition. J Mol
Recogn 1999;12:3–18.
76. Flandin F, Buffevant C, Herbage D. A differential scanning
calorimetry analysis of the age-related changes in the thermal
stability of rat skin collagen. Biochim Biophys Acta 1984;
791:205–211.
77. Kumar TR, Shanmugasundaram N, Babu M. Biocompatible
collagen scaffolds from a human amniotic membrane: Physi-
cochemical and in vitro culture characteristics. J Biomater Sci
Polym Ed 2003;14:689–706.
78. Rochdi A, Foucat L, Renou JP. Effect of thermal denaturation
on water–collagen interactions: NMR relaxation and differential
scanning calorimetry analysis. Biopolymers 1999;50:690–696.
79. Knott L, Bailey AJ. Collagen biochemistry of avian bone:
Comparison of bone type and skeletal site. Br Poult Sci 1999;
40:371–379.
80. Hickman D, Sims TJ, Miles CA, Bailey AJ, de Mari M,
Koopmans M. Isinglass/collagen: Denaturation and functional-
ity. J Biotechnol 2000;79:245–257.
81. Sionkowska A, Kaminska A. Thermal helix–coil transition in
UV irradiated collagen from rat tail tendon. Int J Biol Macro-
mol 1999;24:337–340.
82. Kubisz L, Mielcarek S, Jaroszyk F. Changes in thermal and
electrical properties of bone as a result of 1 MGy-dose c-irra-
diation. Int J Biol Macromol 2003;33:89–93.
83. Hanley L, Kornienko O, Ada ET, Fuoco E, Trevor JL. Sur-
face mass spectrometry of molecular species. J Mass Spec-
trom 1999;34:705–723.
84. Merrett K, Cornelius RM, McClung WG, Unsworth LD, Shear-
down H. Surface analysis methods for characterizing polymeric
biomaterials. J Biomater Sci Polym Ed 2002;13:593–621.
85. Dewez JL, Doren A, Schneider YJ, Rouxhet PG. Competitive
adsorption of proteins: Key of the relationship between sub-
stratum surface properties and adhesion of epithelial cells.
Biomaterials 1999;20:547–559.
86. Dewez JL, Schneider YJ, Rouxhet PG. Coupled influence of
substratum hydrophilicity and surfactant on epithelial cell ad-
hesion. J Biomed Mater Res 1996;30:373–383.
87. Zambonin G, Losito I, Triffitt JT, Zambonin CG. Detection of
collagen synthesis by human osteoblasts on a tricalcium phos-
phate hydroxyapatite: An X-ray photoelectron spectroscopy
investigation. J Biomed Mater Res 2000;49:120–126.
Journal of Biomedical Materials Research Part B: Applied Biomaterials
 GUIDE TO COLLAGEN CHARACTERIZATION FOR BIOMATERIAL STUDIES
88. Healy KE, Ducheyne P. Hydration and preferential molecular
adsorption on titanium in vitro. Biomaterials 1992;13:553–
561.
89. Morra M, Cassinelli C, Cascardo G, Cahalan P, Cahalan L,
Fini M, Giardino R. Surface engineering of titanium by colla-
gen immobilization. Surface characterization and in vitro and
in vivo studies. Biomaterials 2003;24:4639–4654.
90. Pamula E, De Cupere V, Dufrene YF, Rouxhet PG. Nanoscale
organization of adsorbed collagen: Influence of substrate
hydrophobicity and adsorption time. J Colloid Interface Sci
2004;271:80–91.
91. Cheng Z, Teoh SH. Surface modification of ultra thin poly
(epsilon-caprolactone) films using acrylic acid and collagen.
Biomaterials 2004;25:1991–2001.
92. Kwok DY, Gietzelt T, Grundke K, Jacobasch H-J, Neumann
AW. Contact angle measurements and contact angle interpre-
tation, Part 1: Contact angle measurements by axisymmetric
drop shape analysis and a goniometer sessile drop technique.
Langmuir 1997;13:2880–2894.
93. Altankov G, Grinnell F, Groth T. Studies on the biocompati-
bility of materials: Fibroblast reorganization of substratum-
bound fibronectin on surfaces varying in wettability. J Biomed
Mater Res 1996;30:385–391.
94. Brocchini S, James K, Tangpasuthadol V, Kohn J. Structure–
property correlations in a combinatorial library of degradable
biomaterials. J Biomed Mater Res 1998;42:66–75.
95. Tamada Y, Ikada Y. Fibroblast growth on polymer surfaces and
biosynthesis of collagen. J Biomed Mater Res 1994;28:783–789.
96. Higuchi A, Tamiya S, Tsubomura T, Katoh A, Cho CS,
Akaike T, Hara M. Growth of L929 cells on polymeric films
prepared by Langmuir-Blodgett and casting methods. J Bio-
mater Sci Polym Ed 2000;11:149–168.
97. Ber S, Torun Kose G, Hasirci V. Bone tissue engineering on
patterned collagen films: An in vitro study. Biomaterials 2005;
26:1977–1986.
98. Paige MF, Rainey JK, Goh MC. A study of fibrous long spac-
ing collagen ultrastructure and assembly by atomic force mi-
croscopy. Micron 2001;32:341–353.
99. Hassenkam T, Fantner GE, Cutroni JA, Weaver JC, Morse
DE, Hansma PK. High-resolution AFM imaging of intact and
fractured trabecular bone. Bone 2004;35:4–10.
100. Zhang J, Senger B, Vautier D, Picart C, Schaaf P, Voegel JC,
Lavalle P. Natural polyelectrolyte films based on layer-by
layer deposition of collagen and hyaluronic acid. Biomaterials
2005;26:3353–3361.
101. Habelitz S, Balooch M, Marshall SJ, Balooch G, Marshall GW Jr.
In situ atomic force microscopy of partially demineralized human
dentin collagen fibrils. J Struct Biol 2002;138:227–236.
102. Jiang F, Khairy K, Poole K, Howard J, Muller DJ. Creating
nanoscopic collagen matrices using atomic force microscopy.
Microsc Res Tech 2004;64:435–440.
103. Cristofalo VJ, Pignolo RJ. Replicative senescence of human
fibroblast-like cells in culture. Physiol Rev 1993;73:617–638.
104. Qiu Q, Sayer M, Kawaja M, Shen X, Davies JE. Attachment,
morphology, and protein expression of rat marrow stromal
cells cultured on charged substrate surfaces. J Biomed Mater
Res 1998;42:117–127.
105. Ma L, Gao C, Mao Z, Zhou J, Shen J. Enhanced biological
stability of collagen porous scaffolds by using amino acids as
novel cross-linking bridges. Biomaterials 2004;25:2997–3004.
106. Chen J, Altman GH, Karageorgiou V, Horan R, Collette A,
Volloch V, Colabro T, Kaplan DL. Human bone marrow stro-
mal cell and ligament fibroblast responses on RGD-modified
silk fibers. J Biomed Mater Res A 2003;67:559–570.
107. Tan W, Krishnaraj R, Desai TA. Evaluation of nanostructured
composite collagen–chitosan matrices for tissue engineering.
Tissue Eng 2001;7:203–210.
Journal of Biomedical Materials Research Part B: Applied Biomaterials
285
108. Tampieri A, Celotti G, Landi E, Sandri M, Roveri N, Falini
G. Biologically inspired synthesis of bone-like composite:
Self-assembled collagen fibers/hydroxyapatite nanocrystals.
J Biomed Mater Res A 2003;67:618–625.
109. Detamore MS, Orfanos JG, Almarza AJ, French MM, Wong
ME, Athanasiou KA. Quantitative analysis and comparative re-
gional investigation of the extracellular matrix of the porcine
temporomandibular joint disc. Matrix Biol 2005;24:45–57.
110. Dunn CA, Jin Q, Taba M Jr, Franceschi RT, Bruce Rutherford
R, Giannobile WV. BMP gene delivery for alveolar bone engi-
neering at dental implant defects. Mol Ther 2005;11:294–299.
111. Baguneid M, Murray D, Salacinski HJ, Fuller B, Hamilton G,
Walker M, Seifalian AM. Shear stress preconditioning and tis-
sue engineering based paradigms for generating arterial sub-
stitutes. Biotechnol Appl Biochem 2004;39:151–157.
112. Ferrando JM, Vidal J, Armengol M, Huguet P, Gil J, Manero
JM, Planell JA, Segarra A, Schwartz S, Arbos MA. Early
imaging of integration response to polypropylene mesh in ab-
dominal wall by environmental scanning electron microscopy:
Comparison of two placement techniques and correlation with
tensiometric studies. World J Surg 2001;25:840–847.
113. Ehlers EM, Fuss M, Rohwedel J, Russlies M, Kuhnel W,
Behrens P. Development of a biocomposite to fill out articular
cartilage lesions. Light, scanning and transmission electron
microscopy of sheep chondrocytes cultured on a collagen I/III
sponge. Anat Anz 1999;181:513–518.
114. Hsu SH, Tsai CL, Tang CM. Evaluation of cellular affinity
and compatibility to biodegradable polyesters and type-II col-
lagen-modified scaffolds using immortalized rat chondrocytes.
Artif Organs 2002;26:647–658.
115. Cherubino P, Grassi FA, Bulgheroni P, Ronga M. Autologous
chondrocyte implantation using a bilayer collagen membrane: A
preliminary report. J Orthop Surg (Hong Kong) 2003;11:10–15.
116. Mauney JR, Kaplan DL, Volloch V. Matrix-mediated reten-
tion of osteogenic differentiation potential by human adult
bone marrow stromal cells during ex vivo expansion. Bioma-
terials 2004;25:3233–3243.
117. Garfinkel S, Wessendorf JH, Hu X, Maciag T. The human dip-
loid fibroblast senescence pathway is independent of interleu-
kin-1 a mRNA levels and tyrosine phosphorylation of FGFR-1
substrates. Biochim Biophys Acta 1996;1314:109–119.
118. Chen KY, Chang ZF. Age dependency of the metabolic conver-
sion of polyamines into amino acids in IMR-90 human embry-
onic lung diploid fibroblasts. J Cell Physiol 1986;128:27–32.
119. Low RB, Hildebran JN, Absher PM, Stirewalt WS, Arnold J.
Comparison of the use of isotopic proline vs. leucine to mea-
sure protein synthesis in cultured fibroblasts. Connect Tissue
Res 1986;14:179–185.
120. Hildebran JN, Airhart J, Stirewalt WS, Low RB. Prolyl-
tRNA-based rates of protein and collagen synthesis in human
lung fibroblasts. Biochem J 1981;198:249–258.
121. Kulju KS, Lehman JM. Increased p53 protein associated with
aging in human diploid fibroblasts. Exp Cell Res 1995;217:
336–345.
122. Mu XC, Higgins PJ. Differential growth state-dependent regu-
lation of plasminogen activator inhibitor type-1 expression in
senescent IMR-90 human diploid fibroblasts. J Cell Physiol
1995;165:647–657.
123. Briggs D. Surface Analysis of Polymers by XPS and Static
SIMS. Cambridge, UK: Cambridge University Press; 1998.
124. Patankar NA. Mimicking the lotus effect: Influence of double
roughness structures and slender pillars. Langmuir 2004;20:
8209–8213.
125. Cristofalo VJ. Cellular biomarkers of aging. Exp Gerontol
1988;23:297–307.