high resolution imaging of collagen organisation … · high resolution imaging of collagen...

Journal of

ARTICLE IN PRESS

www.elsevier.com/locate/yjsbi

Journal of Structural Biology xxx (2007) xxx–xxx

StructuralBiology

High resolution imaging of collagen organisation and synthesis usinga versatile collagen specific probe

Ralf A. Boerboom a, Katy Nash Krahn a, Remco T.A. Megens b,Marc A.M.J. van Zandvoort b, Maarten Merkx c, Carlijn V.C. Bouten a,*

a Department of Biomedical Engineering, Soft Tissue Biomechanics and Engineering, Eindhoven University of Technology,

PO Box 513, 5600 MB Eindhoven, The Netherlandsb Department of Biophysics, Cardiovascular Research Institute Maastricht, University of Maastricht, Maastricht, The Netherlands

c Department of Biomedical Engineering, Laboratory of Macromolecular and Organic Chemistry, Eindhoven University of Technology,

Eindhoven, The Netherlands

Received 2 January 2007; received in revised form 20 April 2007; accepted 21 April 2007

Abstract

Collagen is the protein primarily responsible for the load-bearing properties of tissues and collagen architecture is one of the maindeterminants of the mechanical properties of tissues. Visualisation of changes in collagen three-dimensional structure is essential in orderto improve our understanding of collagen fibril formation and remodelling, e.g. in tissue engineering experiments. A recently developedcollagen probe, based on a natural collagen binding protein (CNA35) conjugated to a fluorescent dye, showed to be much more specificto collagen than existing fluorescent techniques currently used for collagen visualisation in live tissues. In this paper, imaging with thisfluorescent CNA35 probe was compared to imaging with second harmonic generation (SHG) and the imaging of two- and three-dimen-sional collagen organisation was further developed. A range of samples (cell culture, blood vessels and engineered tissues) was imaged toillustrate the potential of this collagen probe. This images of collagen organisation showed improved detail compared to images gener-ated with SHG, which is currently the most effective method for viewing three-dimensional collagen organisation in tissues. In conclu-sion, the fluorescent CNA35 probe allows easy access to high resolution imaging of collagen, ranging from very young fibrils to moremature collagen fibres. Furthermore, this probe enabled real-time visualisation of collagen synthesis in cell culture, which provides newopportunities to study collagen synthesis and remodelling.� 2007 Elsevier Inc. All rights reserved.

Keywords: Collagen visualisation; Real-time; Two-photon microscopy; CNA35; Second harmonic generation (SHG)

1. Introduction

Living tissues are composed of cells embedded inextracellular matrix (ECM), the latter mainly consistingof proteoglycans, collagen and elastin. Collagen is the mainload-bearing component within the tissue, while the elastinprovides elasticity to the tissue and the proteoglycans givethe tissue its swelling capacity. In load-bearing tissues col-lagen is abundantly present and the mechanical propertiesdepend on the collagen fibre architecture, e.g. collagen fibre

1047-8477/$ - see front matter � 2007 Elsevier Inc. All rights reserved.

doi:10.1016/j.jsb.2007.04.008

* Corresponding author. Fax: +31 40 244 7355.E-mail address: [email protected] (C.V.C. Bouten).

Please cite this article in press as: Boerboom, R.A. et al., High resoluStruct. Biol. (2007), doi:10.1016/j.jsb.2007.04.008

orientation, collagen fibre content and collagen fibre cross-linking (Billiar and Sacks, 2000a,b; Dahl et al., 2005).Studying the active change in collagen architecture is thefocus of diverse fields of research, including developmentalbiology, biomechanics and tissue-engineering. Progress inthese fields requires further elucidation of collagen fibrilformation and remodelling processes by imaging the localthree-dimensional (3D) collagen organisation.

Collagen fibres and bundles can be visualised in livingtissues without the use of specific probes, through tech-niques such as polarised light, phase contrast microscopy(de Campos Vidal, 2003) and differential interference con-trast microscopy (Petroll et al., 2004). Techniques for 3D

tion imaging of collagen organisation and synthesis using ..., J.

mailto:[email protected]

2 R.A. Boerboom et al. / Journal of Structural Biology xxx (2007) xxx–xxx

ARTICLE IN PRESS

visualisation of collagen without the use of specific probesinclude autofluorescence (Voytik-Harbin et al., 2001; Zipfelet al., 2003), confocal reflection microscopy (Hartmannet al., 2006; Voytik-Harbin et al., 2001; Brightman et al.,2000; Wolf and Friedl, 2005) and second harmonic gener-ation (SHG) using femtosecond pulsed infrared excitation(Cox et al., 2003; Campagnola et al., 2002; Zoumi et al.,2002; Konig and Riemann, 2003). Several tissue constitu-ents possess intrinsic autofluorescence properties, enablingthem to be visualised using confocal or multi-photonmicroscopy without the use of added probes. However,autofluorescence is not specific enough and autofluores-cence intensity is relatively low, especially when it is usedin combination with fluorescent probes (Zipfel et al.,2003; Richards-Kortum and Sevick-Muraca, 1996). Simi-larly, with confocal reflection microscopy it is difficult todiscriminate between the constituents in cultured tissueconstructs due to the absence of spectral information.SHG is only expressed by ordered non-centrosymmetricmaterials (e.g. collagen) and is used to detect the orienta-tion and distribution of mature and immature collagenfibril segments (Williams et al., 2005; Konig et al., 2005).The strong forward scattered SHG (Williams et al., 2005;Cox et al., 2003; Campagnola et al., 2002) enables detailedvisualisation of collagen organisation within various tis-sues. However, depending on the tissue properties forwardscattered SHG is not always feasible. Furthermore, SHGrequires high laser power, which increases the risk of collat-eral damage to cells and tissue.

In view of the shortcomings of these existing techniques,we recently developed a collagen specific fluorescent probe(Krahn et al., 2006). This fluorescent CNA35 probe con-sists of a part of a bacterial collagen binding proteindomain (CNA35), which is covalently bound to a commer-cially available fluorescent dye. Using solid phase bindingassays and immunohistological staining we previouslyshowed that this probe was more specific for collagen thendichlorotriazinyl aminofluorescein (DTAF) (Krahn et al.,2006). The fluorescent CNA35 probe showed affinity forboth fibrillar and non-fibrillar collagen. In this study wecontinue the development of this fluorescent CNA35probe. By comparing it directly to SHG and applying itto a wide variety of samples the use of this fluorescentprobe at different structural levels is demonstrated. Weshow that this probe reveals more detail in the collagenorganisation compared to SHG, including visualisation ofvery small collagen fibrils. Furthermore, the probe allowsfor real-time monitoring of collagen synthesis, thusenabling us to study the active change in collagenorganisation.

2. Methods

2.1. Fluorescent CNA35 collagen probe

A vector coding for the collagen binding domain A ofStaphylococcus aureus was transformed into Escherichia

Please cite this article in press as: Boerboom, R.A. et al., High resolStruct. Biol. (2007), doi:10.1016/j.jsb.2007.04.008

coli and expression of this collagen binding domain wasinduced, as described in Krahn et al. (2006). This domain(CNA35) was purified and subsequently labelled witheither oregon green 488 (CNA35-OG488; Invitrogen, TheNetherlands) or Alexa Fluor 488 (CNA35-AF488; Invitro-gen, The Netherlands). The fluorescent CNA35 probe wasapplied to four different sample types.

2.2. Imaging a mouse carotid artery

A Swiss mouse was euthanised by a mixture of O2/CO2

gas. Carotid arteries were isolated and mounted on twoglass micropipettes in a perfusion chamber (Hilgers et al.,2003; Megens et al., 2007b) filled with 10 ml phosphate buf-fered saline (PBS; Sigma, USA) containing the fluorescentprobe(s). A transmural pressure of 80 mm Hg was appliedin order to mimic physiological pressure. Experiments wereapproved by the local ethics committee on the use of labo-ratory animals. Procedures followed were in accordancewith the institutional guidelines.

SYTO44 (Invitrogen, the Netherlands), eosin (Invitro-gen, the Netherlands), and CNA35-OG488 were used asspecific fluorescent markers for DNA/RNA, elastin, andcollagen, respectively (van Zandvoort et al., 2004). Allprobes were applied extraluminally, were excitable withtwo-photon microscopy, and exhibited emission spectrawith maxima at 480, 560, and 520 nm, respectively. Forcomparison with SHG, the labelling solution containedCNA35-OG488 [1.0 lM] in PBS. For imaging of the caro-tid artery, a mixture of SYTO44 [1.5 lM], CNA35-OG488[1.0 lM] and eosin [0.25 lM] in PBS was used.

For imaging a Nikon E600FN upright microscope(Nikon Corporation, Japan), coupled to a standard Biorad2100 MP multiphoton system (Biorad, Great-Britain) wasused (van Zandvoort et al., 2004). A 140-fs pulsed Ti: sap-phire laser (Spectra Physics Tsunami, USA) was tuned andmode-locked at either 800 nm (fluorescence) or 840 nm forSHG. A 60· water dipping objective with a 2.0 mm work-ing distance was used for imaging in upright geometry(numerical aperture (NA) 1.0, Nikon). For the SHG exper-iment, two photomultiplier tubes (PMT) were used todetect the emitted (fluorescent and SHG) signals. Thechannel of PMT1 was tuned at 400–500 nm in order todetect SHG and the channel for PMT2 was tuned at500–560 nm for detecting the CNA35-OG488. For imagingof the carotid artery using the combination of fluorescentmarkers, three PMTs were used. The channels of the threePMTs were tuned as follows: 470–480 nm, SYTO44(PMT1); 500–520 nm, CNA35-OG488 (PMT2); 590–610 nm, eosin (PMT3). Separate images were obtainedfrom each PMT (coded blue, green, and red, respectively)and combined into a single image.

2.3. Imaging an engineered cardiovascular construct

Human vena saphena (HVS) myofibroblasts wereobtained from patients and expanded using regular cell

ution imaging of collagen organisation and synthesis using ..., J.

R.A. Boerboom et al. / Journal of Structural Biology xxx (2007) xxx–xxx 3

ARTICLE IN PRESS

culture methods (Schnell et al., 2001). Cell culture mediumconsisted of advanced Dulbecco’s Modified Eagle Medium(DMEM; Gibco, USA) supplemented with 10% fetalbovine serum (FBS; Biochrom, Germany), 1% L-glutamax(Gibco, USA) and 0.1% gentamycin (Biochrom, Ger-many). A rectangular shaped (0.5 · 2.5 · 1.0 mm3) non-woven polyglycolic acid (PGA) scaffold (density72.76 mg/cm3; Cellon, Luxemburg) was coated with 1%(w/v) poly-4-hydroxybutyrate (P4HB; Symetis Inc., Swit-zerland) in tetrahydrofuran (THF; Merck, Germany)(Hoerstrup et al., 2000). After vacuum drying for 48 hthe scaffold was placed in ultraviolet light for 1 h. Thescaffold was then placed in 70% ethanol for 4–5 h toobtain sterility. The ethanol was allowed to evaporateovernight and the scaffold was washed three times inphosphate buffered saline (PBS, Sigma, USA). The scaf-fold was placed in tissue culture medium overnight tofacilitate cell attachment. The tissue culture medium con-sisted of DMEM supplemented with 10% FBS, 1% L-glutamax, 0.3% gentamycin and L-ascorbic acid 2-phos-phate (0.25 mg/l; Sigma, USA). Myofibroblast cells atpassage 7 were seeded on these scaffolds with fibrin gel(5 mg/ml; Sigma, USA) as a cell carrier (Mol et al.,2005). Approximately 3–5 million cells were used andthe construct was subsequently cultured in tissue culturemedium in a humidified incubator (37 �C and 5% CO2).The medium was changed every 3–4 days.

Cell Tracker Blue CMAC (CTB; Invitrogen, the Nether-lands) and CNA35-OG488 were used as specific fluorescentmarkers for cell cytoplasm and collagen, respectively. CTBwas excitable with two-photon microscopy and exhibitedan emission spectrum with a maximum at 466 nm. ThePGA/P4HB scaffold exhibited autofluorescence below500 nm and above 560 nm. Labelling of the tissue was per-formed with tissue culture medium, which was supple-mented with CTB [15.0 lM] and CNA35-OG488[3.0 lM]. The CTB solution was applied for 5 h, followedby CNA35-OG488 [3.0 lM] for 16 h. The samples werethen placed in tissue culture medium.

An inverted Zeiss Axiovert 200 microscope (Carl Zeiss,Germany) coupled to an LSM 510 Meta (Carl Zeiss, Ger-many) laser scanning microscope was used to image the tis-sue engineered construct. A chameleon ultra 140-fs pulsedTi:Sapphire laser (Coherent, USA), was tuned to 760 nm toimage the applied fluorescent probes. A 63· water-immer-sion objective (1.2 NA; Carl Zeiss, Germany) was used andthe channels for the three PMTs were defined as follows:435–485 nm, CTB and scaffold (PMT1); 500–530 nm,CNA35-OG488 (PMT2); 608–640 nm, scaffold (PMT3).

2.4. Real-time imaging of collagen synthesis

For the real-time experiment HVS cells were grownto passage 6 and plated on round glass coverslips (Men-zel-Glaser, Germany) at a concentration of 10,000 cells/cm2. The cells were cultured in regular cell culturemedium.


Cell Tracker Orange CMRA (CTO; Invitrogen, theNetherlands), and CNA35-AF488 (Invitrogen, The Neth-erlands) were used as specific fluorescent markers for cellcytoplasm and collagen, respectively. AF488 was usedbecause of its superior photostability compared toOG488. The emission peaks of the CTO and CNA35-AF488 are situated at 572 and 520 nm, respectively. 24 hafter seeding cells were incubated in cell culture mediumwith CTO [4 lM] (37 �C and 5% CO2) for 45 min. The cov-erslip was mounted in a bioreactor (Gawlitta et al., 2007)which enabled control of CO2 (5%) and temperature(37 �C). The bioreactor was filled with tissue culture med-ium containing CNA35-AF488 [0.475 lM] and to minimiseevaporation the medium was covered with a coating ofmineral oil (Sigma, USA).

The CTO and CNA35-AF488 were imaged using thesame LSM510 laser scanning microscope as mentioned inSection 2.3 using a 63· oil-immersion objective (NA 1.4;Carl Zeiss, Germany). CTO and CNA35-OG488 wereexcited using a Helium–Neon laser (488 nm) and an ArgonLaser (543 nm), respectively. Two PMTs were used and thechannels were tuned as follows: 505–530 nm, CNA35-OG488 (PMT1); 585 nm and longer, CTO (PMT2). Afterseeding, the cells were rested for 24 h. After these 24 hthe incubator setup was placed on the microscope andthe cell culture was subsequently studied for 52 h withthe labelling solution present. The microscope settings(gain, laserpower) were based on a 48 h old culture, corre-sponding to 24 h after the start of imaging. This was doneto prevent saturation of the signal with increase in collagencontent while simultaneously allowing the detection of suf-ficient detail early on in the culture. Images were recordedevery 60 min. Furthermore, using an identical experimentalsetup another 24 h old cell culture was used to image a sin-gle cell and its surrounding collagen.

3. Results

The fluorescent CNA35 probe was applied to severaltypes of cultures in order to evaluate its imaging propertiesfor collagen.

Fig. 1 shows an image of the collagen organisation inthe tunica adventitia (outer layer of the artery) of a struc-turally intact (non-viable) mouse carotid artery, obtainedusing both SHG (A) and the collagen probe (B). The sig-nal originating from the CNA35-OG488 bound to the col-lagen fibres in the mouse carotid artery aligns perfectlywith the SHG signal, which originates from the collagenfibres in the same focal plane and confirms the specificityof the probe for the collagen fibres. The image obtainedusing the CNA35-OG488 shows much more fine struc-tures compared to the image obtained with SHG. Atgreater depth, the intensity of the emitted CNA35 signalis higher, as is apparent from the fact that a much lowerlaser power (5% versus 40%) is needed to generate signalsof similar strength using two-photon LSM and SHG,respectively.


Fig. 1. Mouse carotid artery, (A) SHG signal of collagen (green), (B)fluorescence signal of collagen probe (green, CNA35-OG488) in a mousecarotid artery recorded at the same focal position obtained with two-photon laser scanning microscopy (15 lm deep). The image size is512 · 512 pixels and the picture measures 103 · 103 lm.


ARTICLE IN PRESS

In Fig. 2A a schematic overview of a structurally intactcarotid artery is shown, which also illustrates the origin ofthe consecutive two-photon images (x/y-scans, B–E). Inthese images collagen fibrils (yellow), cell nuclei (blue) andelastin (red) are shown. (B) and (C) show the tunica adven-titia, which consists primarily of collagen fibrils, fibroblast-like cell nuclei and a few elastin fibres. The tunica media (D–E) consists of collagen, elastin and smooth muscle cell(SMC) nuclei. SMCs are primarily oriented perpendicularto the longitudinal axis of the artery. Small collagen fibrilsare clearly visible in between the smooth muscle cell nuclei.Note that collagen and elastin are distinguishable.

In order to show that the probe enables visualisation ofvery small and young collagen fibrils, we imaged an engi-


neered cardiovascular construct of approximately 1 mmthickness after three weeks of culture. Fig. 3 and its Supple-mentary movie show detailed two-photon images of theconstituents of the engineered tissue constructs: HVS cells(blue), collagen fibrils (green) and PGA/P4HB scaffold(purple). The collagen fibrils are in an early stage of devel-opment and are still very thin compared to the collagenfibrils in the tunica adventitia (Fig. 1). Collagen fibrils typ-ically have a diameter in between 10–500 nm (Fratzl, 2003;Ushiki, 2002). Bearing in mind that the visualised fibrils arevery young, these fibrils most likely have a diameter that issmaller than the optical resolution of 0.4 lm. Notably, thecollagen fibrils show a wavy pattern, as is observed innative collagenous tissues, and the fibrils are closely associ-ated with the surface of the cells (indicated with arrows inFig. 3). This close association to the cell surface is mostclearly seen in the Supplemented movie.

In order to demonstrate the opportunities of this fluo-rescent CNA35 probe for high resolution imaging, theprobe was applied to a cell culture. Zooming in to a singlecell (Fig. 4), we can see very detailed collagenous micro-structures (green) originating from the cell (red). In redthe cell cytosol is visible, which clearly outlines the darkercell nucleus and the cell organelles. Very small collagenfibres and fibre networks are distributed over the entireimaging plane surrounding the cell. The collagen probe isalso attached to very small and round structures withinthe cell cytosol, which are not observed in the cell nucleus.These structures are most likely the result of pinocytosis, inwhich endocytic vesicles non-specifically take up smalldroplets of extracellular fluid and any material dissolvedin it (Lodish et al., 1995). Another possibility could bethe existence of fibropositors of collagen, as was suggestedby Canty and Kadler (2005).

To study its applicability in real-time imaging of colla-gen fibril formation, CNA35-AF488 was used in a cell cul-ture experiment. Fig. 5A–F show six confocal images fromthe same spatial and focal position obtained over a periodof 52 h. In the images and Supplementary movie interestingareas are indicated with white circles. These images andespecially the Supplemented movie show that the cells(red) move around and actively secrete collagen (green)into their surroundings. The first image shows a very lim-ited amount of collagen, which could be either newlyformed or still attached to the cell and left over from theprevious culture. Over time the collagen content increasesand fibre-like structures start to appear, which is veryclearly shown within the white circles. Some cells movearound with the collagen fibres closely attached to theirsurface and form larger collagen aggregates, other cellsrelease the collagen into the surroundings. Most impor-tantly the images show the ability to monitor changes incollagen organisation over time. However, several newlysynthesised collagen fibres remain detached from the cellsurface and float in the culture surroundings. In order todemonstrate if CNA35 affected cell-collagen interaction,collagen type I coated culture plates were seeded with myo-


Fig. 2. (A) Schematic overview of carotid artery, illustrating the scanning direction (B–E) through the tunica intima, media and adventitia. (B–E)Subsequent two-photon images through the mouse carotid artery (z-stack) imaged at a depth of 17, 27, 36 and 52 lm, respectively. Cell nuclei (Syto44,DNA/RNA) are shown in blue, collagen (CNA35-OG488) in yellow and elastin (eosin) in red. Image size is 512 · 512 pixels and the picture measures205 · 205 lm.

Fig. 3. Representative image of tissue engineered construct (30 lm deep).Human vena saphena cells are shown in blue (Cell Tracker Blue), collagenin green (CNA35-OG488) and scaffold is shown in purple (autofluores-cence). Image size is 512 · 512 pixels and the picture measures202 · 202lm. Indicated with arrows is the close association of collagenwith the cell surface. These arrows are also indicated in the Supplementarymovie, which shows images throughout the tissue. This movie clearlyshows the close association of the cells and the collagen.

Fig. 4. High magnification image of myofibroblast cell in monolayerculture 1 day after seeding. The cell cytosol is shown in red (Cell TrackerOrange) and collagen is shown in green (CNA35-OG488). Image size is952 · 952 pixels and the picture measures 95 · 95 lm.


ARTICLE IN PRESS

fibroblasts in the presence and absence of CNA35-OG488(Supplementary data). In a long-term experiment no effecton cell attachment was observed and in a short-term exper-iment an increase in attachment was observed. In addition,in the presence of high concentrations of CNA35-OG488


(3.0 lM), tissue engineered cardiovascular constructsshowed similar compaction to unstained tissue engineeredconstructs. These data suggest that there is no negativeeffect of CNA35 on cell-collagen interaction.

4. Discussion

In this paper we further explored the performance of arecently developed collagen specific probe (Krahn et al.,


Fig. 5. (A–F) Collagen production of cells in cell culture, studied over time (0–52 h). The white circles indicate areas where active formation of fibre-likestructures is taking place, which is most clearly demonstrated in the Supplemented movie of real-time collagen synthesis. Collagen secretion is apparentover the entire area. The cell cytosol is shown in red (Cell Tracker Orange) and collagen is shown in green (CNA35-AF488). The time (in minutes) from thestart of the experiment is shown in the upper left corner. Image size is 512 · 512 pixels and the picture measures 205 · 205lm.


ARTICLE IN PRESS

2006). The fluorescent CNA35 probe was applied to sam-ples with different levels of maturity and demonstratedimproved properties with regard to visualisation of differ-ent structural levels of collagen organisation. The probeshowed an improvement in detail and enabled real-timemonitoring of collagen synthesis.

A mouse carotid artery was imaged using both SHG andfluorescently labelled CNA35. Fig. 1A and B show that thesignal from the probe colocalises with the SHG signal,which confirms the collagen specificity of the probe (Krahnet al., 2006). Furthermore, it shows that the CNA35 probereveals much more detail (fine structures) than the back-scattered SHG signal. This can possibly be explained bydifference in signal strength obtain with SHG and imagingwith fluorescent CNA35. The small fibrils probably do notgenerate enough SHG to be visualised, whereas with fluo-rescent enhancement this is possible. Image formation withSHG and fluorescent probes using two-photon laser scan-ning microscopy is different (Gauderon et al., 2001; Mor-eaux et al., 2001), e.g. resulting in differences inpolarisation information. Furthermore, SHG can only begenerated in non-centrosymmetric assemblies of chiralmolecules. SHG is strongest in forward scatter geometry(Williams et al., 2005; Cox et al., 2003; Konig et al.,2005), however, in certain dense and thick tissues (e.g. largearteries) it is difficult to obtain images with the forwardscattered SHG of collagen fibres (Boulesteix et al., 2006;Megens et al., 2007b). This limits the use of forward scat-tered SHG and requires additional techniques to image col-lagen in backward geometry, which is the geometry used inthis study. Apparently, the SHG in backward geometry is


much weaker than the fluorescence signal of the collagenprobe. The stronger signal of the collagen probe enablesus to use lower laser power for imaging, which reducesthe risk of tissue destruction and allows the use of a widerange of laser powers to image deeper into the tissue. In tis-sue engineered constructs, images have been recorded110 lm deep and imaging was not limited by the diffusionof the probe. In addition, Megens et al. (2007a) applied theCNA35-OG488 probe to muscular and elastic arteries andshowed limited labelling in viable elastic arteries by thepresence of intact endothelium and elastic laminae. Thisdemonstrated that the availability of the probe in specifictissue regions depends on the tissue composition, the con-centration of the applied probe, the amount of time theprobe is applied to the tissue and the way the probe isapplied to the tissue (e.g. intra- and extraluminally). Micro-tubules and skeletal muscle exhibit SHG as well, albeit at amuch lower signal compared to collagen (Zipfel et al.,2003). The probe has been tested for specificity by usingsolid phase binding assays (Krahn et al., 2006) and immu-nohistological staining, which suggested that the probebinds specifically to collagen with a different affinity forthe different collagen types (I–VI). A possible drawbackto the use of fluorophores is the risk of photo-bleaching,whereas intrinsic tissue properties like SHG do not exhibitphoto-bleaching.

In addition, an artery was stained to visualise the cells,the elastin network and the collagen network. A detailedimage of the collagen organisation within the tissue isobtained, showing images of collagen fibres ranging froma few microns thick (�4.0 lm) to sub-micron resolution



ARTICLE IN PRESS

(�0.4 lm). The collagen probe revealed small collagenfibres in between the smooth muscle cells in the tunicamedia of the carotid artery, which are much more difficultto observe with the use of SHG (Zoumi et al., 2004; Bou-lesteix et al., 2006; Schenke-Layland et al., 2005, 2006).

The probe was applied on three weeks old tissue engi-neered cardiovascular constructs in order to show its use-fulness in visualizing the developing collagen fibrilswithin live young tissues. After three weeks of culture theconstruct predominantly consists of cells, glycosaminogly-cans, collagen and PGA scaffold (Mol et al., 2005). Thecells, the collagen and the PGA scaffold were imaged.Fig. 3 shows that the collagen fibrils of the tissue engi-neered construct are relatively thin compared to fibrils inmature native artery (Figs. 1 and 2) and that these fibrilsexhibit a very diffuse organisation compared to nativefibrils. Mature native fibrils are bundled in fibres ratherthan in individual fibrils. The benefit of the fluorescentprobe used here is that it binds to individual collagen fibrils(Zong et al., 2005) and allows collagen fibrils with sub-res-olution diameter to be imaged. Collagen fibres typicallyhave a diameter in between 10–500 nm (Ushiki, 2002). Sev-eral successful attempts have been performed to studyECM remodelling in time by using confocal reflectionmicroscopy in 3D fibrin lattices and fibroblast populatedcollagen lattices (Hartmann et al., 2006; Brightman et al.,2000; Voytik-Harbin et al., 2001; Wolf and Friedl, 2005).However, with confocal reflection microscopy alone it isnot possible to distinguish between different constituents,whereas with multiple fluorescent labels in combinationwith laser scanning microscopy this is possible. Collagenassembly by cells has previously been quantified usingFITC labelled rat-tail collagen (Johnson and Galis, 2003)to monitor the remodelling and structure of exogenous col-lagen. The study presented here is the first to look at fluo-rescently enhanced endogenous collagen.

Fig. 4 show that the probe enables detailed visualisationof collagen structures at high magnification and providesnew opportunities to visualise collagen at the cellular level.The fluorescent collagen probe also allows real-time visual-isation of collagen synthesis. This was demonstrated byseeding myofibroblast cells on top of a glass coverslipand studying these cells over time in the presence of theprobe. The probe was coupled to a more stable fluorescentdye to prevent possible bleaching of the probe. Myofibro-blasts are involved in wound healing and pathological pro-cesses and predominantly synthesise type I and type IIIcollagen (Hinz et al., 2001). The real-time collagen synthe-sis shows similar features to real-time elastin fibrillogenesis,which has been studied with confocal laser scanningmicroscopy (Kozel et al., 2006; Czirok et al., 2006). Onecan see collagen fibrils that are closely associated to the cellsurface and that are subsequently transferred to otherextracellular collagen fibres. However, more detailed stud-ies have to be performed to elucidate the exact process ofcollagen fibre formation. Potential drawbacks to the colla-gen probe are its effects on cell-collagen interaction and


fibril formation. However, the collagen probe did not showany negative effect on cell-collagen interaction. A theoreti-cal drawback of this probe is that it binds directly to thecollagen. By binding to the collagen the probe potentiallyprevents proper fibril formation. Krahn et al. (2006) showthat the dissociation constant (Kd) of CNA35 is approxi-mately 10�7–10�6, which ensures that the binding is suffi-ciently strong but not irreversible. When mediumcontaining CNA35-OG488 is removed from the cell cultureand replaced with plain culture medium, a new equilibriumwill form between the probe bound to the collagen and theplain culture medium. Due to the absence of probe in theculture medium this equilibrium will shift towards the cul-ture medium, which effectively decreases the concentrationof probe bound to collagen.

In summary the present study has shown that theCNA35-based collagen probe enables high resolutionimaging of collagen organisation and synthesis in greatdetail. The probe can be used to study very young and thincollagen fibrils in cell culture and more mature collagen innative tissue. Furthermore, this study is the first to reporton real-time monitoring of collagen synthesis. The use offluorescent CNA35 for imaging of collagen results in moredetail compared to SHG while using lower laser powers.The probe is very flexible in its design. Different probes(e.g. fluorescent, MRI) can be coupled to the collagen bind-ing domain, which provides new possibilities to study col-lagen organisation and remodelling with high resolution.

Appendix A. Supplementary data

Supplementary data associated with this article can befound, in the online version, at doi:10.1016/j.jsb.2007.04.008.

References

Billiar, K.L., Sacks, M.S., 2000a. Biaxial mechanical properties ofthe native and glutaraldehyde-treated aortic valve cusp: Part II—A structural constitutive model. J. Biomech. Eng. 122 (4), 327–335.

Billiar, K.L., Sacks, M.S., 2000b. Biaxial mechanical properties of thenatural and glutaraldehyde treated aortic valve cusp: Part I—Exper-imental results. J. Biomech. Eng. 122 (1), 23–30.

Boulesteix, T., Pena, A.M., Pages, N., Godeau, G., Sauviat, M.P.,Beaurepaire, E., Schanne-Klein, M.C., 2006. Micrometer scale ex vivomultiphoton imaging of unstained arterial wall structure. Cytometry A69 (1), 20–26.

Brightman, A.O., Rajwa, B.P., Sturgis, J.E., McCallister, M.E., Robinson,J.P., Voytik-Harbin, S.L., 2000. Time-lapse confocal reflection micros-copy of collagen fibrillogenesis and extracellular matrix assemblyin vitro. Biopolymers 54 (3), 222–234.

Campagnola, P.J., Millard, A.C., Terasaki, M., Hoppe, P.E., Malone,C.J., Mohler, W.A., 2002. Three-dimensional high-resolution second-harmonic generation imaging of endogenous structural proteins inbiological tissues. Biophys. J. 82 (1 Pt 1), 493–508.

Canty, E.G., Kadler, K.E., 2005. Procollagen trafficking, processing andfibrillogenesis. J. Cell Sci. 118 (7), 1341–1353.

Cox, G., Kable, E., Jones, A., Fraser, I., Manconi, R., Gorrell, M.D.,2003. 3-dimensional imaging of collagen using second harmonicgeneration. J. Struct. Biol. 141 (1), 53–62.


http://dx.doi.org/10.1016/j.jsb.2007.04.008

http://dx.doi.org/10.1016/j.jsb.2007.04.008


ARTICLE IN PRESS

Czirok, A., Zach, J., Kozel, B.A., Mecham, R.P., Davis, E.C., Rongish,B.J., 2006. Elastic fiber macro-assembly is a hierarchical, cell motion-mediated process. J. Cell Physiol. 207 (1), 97–106.

Dahl, S.L.M., Rucker, R.B., Niklason, L.E., 2005. Effects of copper andcross-linking on the extracellular matrix of tissue-engineered arteries.Cell Transplant. 14 (10), 861–868.

de Campos Vidal, B., 2003. Image analysis of tendon helical superstruc-ture using interference and polarized light microscopy. Micron 34 (8),423–432.

Fratzl, P., 2003. Cellulose and collagen: from fibres to tissues. Curr. Opin.Colloid Interface Sci. 8, 32–39.

Gauderon, R., Lukins, P.B., Sheppard, C.J., 2001. Optimization ofsecond-harmonic generation microscopy. Micron 32 (7), 691–700.

Gawlitta, D., Li, W., Oomens, C.W.J., Baaijens, F.P.T., Bader, D.L.,Bouten, C.V.C, 2007. The relative contributions of compression andhypoxia to development of muscle tissue damage: an in-vitro study.Ann. of Biomed. Engng. 35 (2), 273–284.

Hartmann, A., Boukamp, P., Friedl, P., 2006. Confocal reflection imagingof 3D fibrin polymers. Blood Cells Mol. Dis. 36 (2), 191–193.

Hilgers, R.H.P., Bergaya, S., Schiffers, P.M.H., Meneton, P., Boulanger,C.M., Henrion, D., Levy, B.I., Mey, J.G.R.D., 2003. Uterine arterystructural and functional changes during pregnancy in tissue kalli-krein-deficient mice. Arte-rioscler. Thromb. Vase. Biol. 23 (10), 1826–1832.

Hinz, B., Mastrangelo, D., Iselin, C.E., Gabbiani, G., 2001. Mechanicaltension controls granulation tissue contractile activity and myofibro-blast differentiation. Am. J. Path. 159 (3), 1009–1020.

Hoerstrup, S.P., Sodian, R., Daebritz, S., Wang, J., Bacha, E.A., Martin,D.P., Moran, A.M., Guleserian, K.J., Sperling, J.S., Kaushal, S.,Vacanti, J.P., Schoen, F.J., Mayer, J.E.J., 2000. Functional livingtrileaflet heart valves grown in vitro. Circulation 102 (19 Suppl. 3), 44–49.

Johnson, C., Galis, Z.S., 2003. Quantitative assessment of collagenassembly by live cells. J. Biomed. Mater. Res. A 67 (3), 775–784.

Konig, K., Riemann, I., 2003. High-resolution multiphoton tomographyof human skin with subcellular spatial resolution and picosecond timeresolution. J. Biomed. Opt. 8 (3), 432–439.

Konig, K., Schenke-Layland, K., Riemann, I., Stock, U.A., 2005.Multiphoton autofluorescence imaging of intratissue elastic fibers.Biomaterials 26 (5), 495–500.

Kozel, B.A., Rongish, B.J., Czirok, A., Zach, J., Little, C.D., Davis, E.C.,Knut-sen, R.H., Wagenseil, J.E., Levy, M.A., Mecham, R.P., 2006.Elastic fiber formation: a dynamic view of extracellular matrixassembly using timer reporters. J. Cell. Physiol. 207 (1), 87–96.

Krahn, K.N., Bouten, C.V.C., van Tuijl, S., van Zandvoort, M.A.M.J.,Merkx, M., 2006. Fluorescently labeled collagen binding proteinsallow specific visualization of collagen in tissues and live cell culture.Anal. Biochem. 350 (2), 177–185.

Lodish, H., Baltimore, D., Berk, A., Zipursky, L.S., Matsudaira, P.,Darnell, J., 1995. Molecular Cell Biology, Third Ed. ScientificAmerican Books, Oxford, England.

Megens, R.T.A., Oude Egbrink, M.G.A., Cleutjens, J.P.M., Kuijpers,M.J.E., Schiffers, P.H.M., Merkx, M., Slaaf, D.W., van Zandvoort,M.A.M.J., 2007a. Imaging collagen in intact viable healthy andatherosclerotic arteries using fluorescently-labeled cna35 and two-photon laser scanning microscopy. Mol. Imaging, in press.

Megens, R.T.A., Reitsma, S., Schiffers, P.H.M., Hilgers, R.H.P., Mey,J.G.R.D., Slaaf, D.W., Oude Egbrink, M.G.A., van Zandvoort,


M.A.M.J., 2007b. Two-photon microscopy of vital murine elasticand muscular arteries, combined structural and functional imagingwith subcellular resolution. J. Vase. Res. 44, 87–98.

Mol, A., Driessen, N.J.B., Rutten, M.C.M., Hoerstrup, S.P., Bouten,C.V.C., Baaijens, F.P.T., 2005. Tissue engineering of human heartvalve leaflets: a novel bioreactor for a strain-based conditioningapproach. Ann. Biomed. Eng. 33 (12), 1778–1788.

Moreaux, L., Sandre, O., Charpak, S., Blanchard-Desce, M., Mertz, J.,2001. Coherent scattering in multi-harmonic light microscopy. Bio-phys. J. 80 (3), 1568–1574.

Petroll, W.M., Cavanagh, H.D., Jester, J.V., 2004. Dynamic three-dimensional visualization of collagen matrix remodeling and cytoskel-etal organization in living corneal fibroblasts. Scanning 26 (1), 1–10.

Richards-Kortum, R., Sevick-Muraca, E., 1996. Quantitative opticalspectroscopy for tissue diagnosis. Annu. Rev. Phys. Chem. 47 (NIL),555–606.

Schenke-Layland, K., Madershahian, N., Riemann, I., Starcher, B.,Halbhuber, K.J., Konig, K., Stock, U.A., 2006. Impact of cryopres-ervation on extracellular matrix structures of heart valve leaflets. Ann.Thorac. Surg. 81 (3), 918–926.

Schenke-Layland, K., Riemann, I., Stock, U.A., Konig, K., 2005. Imagingof cardiovascular structures using near-infrared femtosecond multi-photon laser scanning microscopy. J. Biomed. Opt. 10 (2), 024017.

Schnell, A.M., Hoerstrup, S.P., Zund, G., Kolb, S., Sodian, R., Visjager,J.R., Grunenfelder, J., Suter, A., Turina, M., 2001. Optimal cell sourcefor cardiovascular tissue engineering: venous vs. aortic humanmyofibroblasts. Thorac. Car-diovasc. Surg. 49 (4), 221–225.

Ushiki, T., 2002. Collagen fibers, reticular fibers and elastic fibers. Acomprehensive understanding from a morphological viewpoint. Arch.Histol. Cytol. 65 (2), 109–126.

van Zandvoort, M.A.M.J., Engels, W., Douma, K., Beckers, L., Egbrink,M.O., Daemen, M., Slaaf, D.W., 2004. Two-photon microscopy forimaging of the (atherosclerotic) vascular wall: a proof of conceptstudy. J. Vase. Res. 41 (1), 54–63.

Voytik-Harbin, S.L., Rajwa, B., Robinson, J.P., 2001. Three-dimensionalimaging of extracellular matrix and extracellular matrix-cell interac-tions. Methods Cell Biol. 63 (NIL), 583–597.

Williams, R.M., Zipfel, W.R., Webb, W.W., 2005. Interpreting second-harmonic generation images of collagen I fibrils. Biophys J. 88 (2),1377–1386.

Wolf, K., Friedl, P., 2005. Functional imaging of pericellular proteolysisin cancer cell invasion. Biochimie 87 (3-4), 315–320.

Zipfel, W.R., Williams, R.M., Christie, R., Nikitin, A.Y., Hyman, B.T.,Webb, W.W., 2003. Live tissue intrinsic emission microscopy usingmultiphoton-excited native fluorescence and second harmonic gener-ation. Proc. Natl. Acad. Sci. USA 100 (12), 7075–7080.

Zong, Y., Xu, Y., Liang, X., Keene, D.R., Hook, A., Gurusiddappa, S.,Hook, M., Narayana, S.V.L., 2005. A ’Collagen Hug’ model forStaphylococcus aureus CNA binding to collagen. EMBO J. 24 (24),4224–4236.

Zoumi, A., Lu, X., Kassab, G.S., Tromberg, B.J., 2004. Imaging coronaryartery microstructure using second-harmonic and two-photon fluores-cence microscopy. Biophys J. 87 (4), 2778–2786.

Zoumi, A., Yeh, A., Tromberg, B.J., 2002. Imaging cells and extracellularmatrix in vivo by using second-harmonic generation and two-photonexcited fluorescence. Proc. Natl. Acad. Sci. USA 99 (17), 11014–11019.


high resolution imaging of collagen organisation … · high resolution imaging of collagen...

Documents