near-infrared fluorescence imaging of mouse myocardial microvascular endothelium using cy5.5-lectin...

FULL ARTICLE

Near-infrared fluorescence imaging of mousemyocardial microvascular endothelium usingCy5.5-lectin conjugate

Cecilia Nguyen1, Saro Bascaramurty2, Bozena Kuzio2, Lori Gregorash2,Valery Kupriyanov3, and Olga Jilkina*; 4

1 University of Manitoba, Department of Oral Biology, Winnipeg, MB, R3T 2N2, Canada2 Institute for Biodiagnostics, National Research Council of Canada, 435 Ellice Avenue, Winnipeg, MB, R3B 1Y6, Canada3 Retired. Former affiliation: Institute for Biodiagnostics, National Research Council of Canada, 435 Ellice Avenue, Winnipeg, MB,

R3B 1Y6, Canada4 Institute for Biodiagnostics, National Research Council of Canada, and University of Manitoba, Department of Oral Biology,

435 Ellice Avenue, Winnipeg, MB, R3B 1Y6, Canada

Received 3 November 2011, revised 12 January 2012, accepted 26 January 2012Published online 24 February 2012

Key words: Cy5.5, Lycopersicon esculentum lectin, mouse, fluorescence microscopy, cardiac imaging techniques,endothelial cells

Æ Supporting information for this article is available free of charge under http://dx.doi.org/10.1002/jbio.201100119

# 2012 by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Journal of

BIOPHOTONICS

Early View publication onwww.wileyonlinelibrary.com(issue and page numbers not yet assigned;citable using Digital Object Identifier – DOI)

Cy5.5-lectin, a non-toxic conjugate, combines the bene-fits of near-infrared (NIR) imaging, such as significantreduction of background fluorescence and increasedtissue depth penetration, with its affinity for vascularendothelial cells. When compared to endothelial stainingmethods using FITC-lectin and ICAM2 antibodies,Cy5.5-lectin was confirmed to specifically bind endothe-lial cells and produce a fluorescence signal both in real-time and post-infusion. Ex-vivo experiments with iso-lated hearts demonstrated that binding was limited toperfused areas of the myocardium. With mouse in-vivotail-vein injections, other organs such as the liver, spleen,and kidney were also stained and yielded similar qualityimages of the heart.

A 50 mm section through a mouse heart perfused withCy5.5-lectin and excited at 680 nm.

* Corresponding author: e-mail: [email protected], Phone: 1-(204)-984-6558, Fax: 1-(204)-984-7036

J. Biophotonics 1–14 (2012) / DOI 10.1002/jbio.201100119

1. Introduction

In cardiovascular research, near-infrared (NIR)fluorescence imaging has emerged as a new opticalimaging technique [1]. There are many advantages toimaging within the NIR spectral range (700–900 nm).Firstly, at these wavelengths there is very little inter-ference from endogenous fluorescent chromophores,such as NAD(P)H and flavins as well as very littleinterference from quenchers such as cytochromes, allof which mostly absorb and fluoresce in the spectralregion between 359–650 nm [2]. Secondly, low tissueabsorption (by hemoglobin and myoglobin) withinthe NIR range means that light penetration can beseveral millimetres [3].

These characteristics are desirable in the imagingof endothelial cells and the endothelial networkwithin intact organs. Endothelial cells form a unicel-lular layer in all blood vessels; therefore they play akey role in vascular health due to their function as abarrier between blood and interstitial tissue fluid, aswell as the controller of blood pressure through therelease of vaso-dilators and constrictors [4, 5]. En-dothelial dysfunction often leads to vascular compli-cations in patients with diabetes mellitus, hyperten-sion, atherosclerosis and more recently, obesity [6–11]. It has also been suggested that the interactionsinvolving endothelial cells are dependent on theirmicroenvironments, which in an in-vivo study, wouldremain intact [12]. Currently, however, there is noreadily available optical probe capable of imagingendothelial cells in the NIR range in live organs,without subjecting them to the optical clearance pro-cedures. A probe for endothelial cells in the NIRrange can also complement current non-NIR probessuch as fluorecein isothiocyanate (FITC)-lectin con-jugate, allowing, in principle, for sequential imagingof the same organ following pharmacological inter-vention as Cy5.5 and FITC fluoresce in differentspectral regions.

This paper reports the development and testingof the fluorescent NIR dye conjugate, Cy5.5-lectin,as a probe for endothelial cells. Cy5.5 N-hydroxy-succinimide (NHS) ester is a cyanine derivative thatis excited at 675 nm and fluoresces at 694 nm (GEHealthcare-Amersham). The dye binds to the freeamino groups of lysine residues as well as the N-term-inal amino group of proteins such as lectin throughacylation [13]. Cy5.5-NHS esters and conjugates havebeen used successfully in previous in-vivo studiesand no adverse or toxic effects have been reported[14, 15]. Previous studies have demonstrated thatLycopersicon esculentum (LEA, tomato) lectin bindspecifically to polylactosamine oligosaccharides, ofwhich endothelial cell surface carbohydrates aremostly comprised, thus making it an effective probefor endothelial cells [16–18]. This specificity is cru-cial in the design of a NIR probe for endothelial

cells. From studies analyzing the amino acid compo-sition of LEA lectin, it is known that this specific lec-tin has approximately 11 lysine residues and thus isable to react with the NHS-ester group to form aconjugate [19]. LEA lectin has been shown to be re-latively non-toxic, a characteristic that is crucial inthe development of probes for in-vivo imaging [19,20]. This also opens the possibility for use of the con-jugate as a deposition flow tracer in different speciesas well as in potential clinical applications [21]. Cur-rently, mostly intracellular deposition flow tracersare employed for this purpose [22–24]. Compara-tively, Cy5.5-lectin is a vascular micro-flow imagingprobe that would reflect the micro-flow distributionmore accurately and because of its ability to gener-ate images through thicker tissues, Cy5.5-lectin couldbe used in physiological studies of microvascular en-dothelium function. For instance, endothelial stainingcan be used to analyze cardiac micro-vessel branch-ing and filling. The latter may change under condi-tions of higher workload and hypoxia, or in a diseasethat manifests itself as diffuse-type damage (e.g. dia-betes, hypertension, inflammation associated micro-emboli etc.) as opposed to localized-type damagecaused by occlusion of large coronary vessels. Micro-emboli are frequently found in patients who diefrom sudden death and currently, this kind of myo-cardial damage is only confirmed in biopsies or bypost-mortem histology [25].

Our specific aims in this paper were to first,synthesize the Cy5.5-lectin conjugate. The secondaim was to test the binding of the conjugate to en-dothelial cells by injecting the dye into the cardiactissue using both in-vivo and ex-vivo methods. Thethird aim was to test the ability of the conjugate toaid in real-time imaging of normal and abnormalheart perfusion.

2. Materials and methods

All procedures in this study conform with the “Guideto the Care and Use of Experimental Animals” pub-lished by the Canadian Council on Animal Care(Ottawa, Canada, 1993).

2.1 Reagents

The following materials were acquired: Monofunc-tional and bisfunctional Cy5.5 N-hydroxysuccinimide(NHS)-ester reactive dyes (GE Healthcare Ltd.,Buckinghamshire, UK); Lycopersicon esculentum(LEA) lectin, fluorescein isothiocyanate (FITC)-lec-tin conjugate and L-lysine monohydrochloride (Sig-ma, St. Louis, MO); Bovine serum albumin (BSA)

C. Nguyen et al.: Mouse endothelial cell labelling using Cy5.5-lectin conjugate2

Journal of

BIOPHOTONICS

# 2012 by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.biophotonics-journal.org

(Lifeblood Medical, Inc, Freehold, NJ); 10% neutralbuffered formalin (EMD Chemicals Inc, Gibbstown,NJ); Fluoro-Gel aqueous mounting medium (Elec-tron Microscopy Sciences, Hatfield, PA); PharmingenAnti-Rat Ig HRP Detection Kit and rat anti-mouseintercellular adhesion molecule 2 (ICAM2) antibo-dies (BD Sciences, San Diego, CA). All other chemi-cals used were of analytical grade.

2.2 Synthesis

2.2.1 Synthesis of Cy5.5-lectin conjugate

Cy5.5-lectin conjugates were synthesized by reacting3.4� 10�4 M of FluoroLink Cy5.5 monofunctional orbisfunctional NHS–ester reactive dye with 6 mg ofLEA lectin, which was dissolved in 6 ml of 0.1 Msodium carbonate-bicarbonate buffer (pH 9.3). Theinitial dye-to-protein (D/P) ratio was approximately20–25 moles of dye per mole of protein. The reac-tion mixture was allowed to react over a period of4–6 hrs at 25 �C in the dark with mixing. Afterwards,4 rounds of dialysis were used to remove any re-maining free dye. Spectra/Por 7 regenerated cellu-lose dialysis membrane (MWCO ¼ 25 000; SpectrumLaboratories, Inc., Rancho Dominguez, CA) and thefollowing dialysis buffers were used: the first tworounds used 0.15 M NaCl and the last two roundsused 0.01 M phosphate buffered saline (PBS) with0.01% NaN3. Dialysis was performed on ice withstirring for 4 hrs after which the dialysate was trans-ferred into fresh buffer and refrigerated (4 �C) over-night. Cy5.5-lectin conjugate was stored in the darkat 4 �C until needed. The quality of the conjugatewas determined by estimating the final D/P ratio.

Applying the Beer–Lambert law, the D/P ratiowas calculated using the following equation:

D=P ¼ A675A280 protein

econjlconjðA280 � 0:18 A675Þ Cproteinð1Þ

The absorbance of the conjugate was measured at280 nm and 675 nm using a microplate reader (BioTekInstruments, Inc. Winooski, USA) and defined inEq. (1) as A280 and A675 respectively. The pathlength,lconj, was calculated by measuring the absorbance ofthe conjugate at 900 nm and 977 nm. The extinctioncoefficient of Cy5.5-NHS ester, defined as econj, is250 000 M�1 cm�1 (GE Healthcare, manufacturermanual). Since the extinction coefficient of the lectinwas unknown, the absorbance and concentration ofa solution of un-reacted protein was used and de-fined here as A280 protein and Cprotein. The factor of0.18 accounts for the absorption of the dye at280 nm, which is approximately 18% of the maxi-mum dye absorbance at 675 nm. Comparisons were

also made between the binding efficiency of mono-functional Cy5.5 and bisfunctional Cy5.5 to LEAlectin, as both were readily available. Smaller scaledreactions were also performed when needed as de-scribed in the text.

2.2.2 Synthesis of Cy5.5-BSAand Cy5.5-lysine conjugate

Synthesis was performed under the same conditionsas described for Cy5.5-lectin. Modifications weremade for the Cy5.5-lysine reaction in order to ensurecomplete blockage of the dye reactive group(s). Theamount of lysine added was 100-fold the mole ofdye. Purification steps for the Cy5.5-lysine reactioninvolving dialysis were not performed due to thesmall size of lysine (MW ¼ 146.19 Da) and the lim-itations of the dialysis membrane. Cy5.5-lysine wasthen used to test the ability of dialysis as a methodto remove free dye and to determine if any non-spe-cific binding by the dye occurred once infused.

2.2.3 Measuring Cy5.5 fluorescencein solutions

In the dark, a 650 nm laser was placed against thequartz cuvette containing a solution of Cy5.5 orCy5.5-lectin, causing excitation of the chromophores.Emitted fluorescence was detected at 90� using a fi-ber optic cable connected to a spectrometer (ControlDevelopment, Inc., South Bend, IN). Single-readfluorescence spectra between 480–1100 nm with aDl ¼ 1 nm were acquired. The program CDI Spec32(Control Development, Inc., South Bend, IN) wasused for data collection and processing.

2.3 Ex-vivo dye loading intoLangendorff-perfused mouse hearts

2.3.1 Mouse heart isolation and perfusion

Male and female C57BL6 mice (body weight ¼ 33.0� 5.8 g, heart weight ¼ 0.19 � 0.04 g, n ¼ 31), bredat the Institute for Biodiagnostics animal facility,were anesthetized with 0.25 ml a pentobarbital so-lution (65 mg/ml) [26]. The hearts were quickly re-moved and perfused in a Langendorff mode withKrebs-Henseleit buffer (KHB) as described else-where [27]. The KHB solution consisted of (in mM):118 NaCl, 4.7 KCl, 1.2 MgSO4, 2.5 CaCl2, 0.5 EDTA,11 Glucose, 25 NaHCO3 and 1.5 Na-pyruvate, which

J. Biophotonics (2012) 3

FULLFULLARTICLEARTICLE

# 2012 by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheimwww.biophotonics-journal.org

was aerated with a mixture of 95% O2 and 5% CO2

at 36 �C. The coronary flow rate provided by a pe-ristaltic pump was 3 ml/min. The conjugates weremixed with KHB upon entering the hearts using acalibrated syringe pump set to an infusion rate of3 ml/hr. Infusion time was either 10 or 20 minutesdepending on the concentration of the injected probe.The concentration of the Cy5.5-lectin, Cy5.5-BSAand FITC-lectin conjugates within the syringe were100 mg/ml for the 10 min injection and the 50 mg/mlfor the 20 min injection. The mechanical function ofthe heart was measured by introducing a water-filledballoon into the left-ventricle (LV) of the heart. Theballoon was connected to a pressure transducer inter-faced to a Digi-Med Model-210 heart performanceanalyzer (Micro-Med, Louisville, KY) to monitorheart rate, LV systolic pressure, LV end-diastolicpressure, and perfusion pressure. Effluent producedduring the perfusion period with the probe was col-lected in order to estimate the amount of the conju-gate that was retained by the heart.

2.3.2 Optical point spectroscopy of perfusedmouse hearts

During Langendorff perfusion, the absorbance spec-tra of the heart were monitored using the fiber opticilluminator Oriel model 77501 (Stratford, CT) inconjunction with a spectrometer (model PDA-512,Control Developments Inc., South Bend, IN) [28].Spectra within the 400–1100 nm range were acquiredwith a 0.1 s integration time, a sample average of 120and a delay of 48 s, which resulted in approximatelyone spectrum per minute. Pseudo-optical density(POD) was calculated according to the formula [28]:

POD ¼ �log ðIs=I0Þ ð2ÞThe variables Is and I0 represent the intensity of lightdiffusely reflected back by the sample and the lightreflected by the Spectralon standard respectively.The presence of oxy-myoglobin peaks at 545 and581 nm was an indicator for adequate oxygenationof the myocardial tissue [29].

2.3.3 Left anterior descending (LAD) arteryligation

Methods as described by Jilkina et al. were modifiedfor mice [22]. Briefly, mouse hearts were excised andperfused in a Langendorf mode (described in 2.3.1).To avoid motion artifacts during imaging sessions,high potassium KHB (24 mM KCl) was used toarrest the hearts in place of regular KHB. The LADartery was ligated with a 4-O silk surgical suture

(Ethicon, Inc. Somerville, NJ). Optical point spectro-scopy was used to determine quality of the ligationand degree of perfusion restriction. Areas unaffectedby the ligation continued to have a normal oxygena-tion profile compared to the poorly oxygenated li-gated area downstream of the LAD.

2.4 Tail-vein injection (in-vivo dye loading)

Mice were anesthetized with isofluorane (3–4% in-duction, 1.5–2% maintenance) in 100% oxygen at0.6 l/min [30]. Injection protocol was based on thatdescribed by McDonald et al. [31]. The lateral tail-vein was located and injected with approximately200 ml of the selected conjugate at concentrations be-tween 0.5 mg/ml to 2.0 mg/ml. The injected solutionwas allowed to re-circulate up to 5 min before har-vesting the heart. The mouse was then given an intra-peritoneal injection of 0.25 ml pentobarbital solution(65 mg/ml), and removed from gas anesthesia pend-ing a midline thoracotomy and harvest of the heartand other tissues.

2.5 Histological staining

Harvested tissue samples were frozen into blocks byusing labeled standard TissueTek cryomolds (SakuraFinetek USA, Inc., Torrance, CA) filled with Shan-don Cryomatrix (Thermo Fisher Scientific, Cheshire,UK), and floated in liquid N2. Blocks were wrappedin foil and stored at �80 �C until sectioning.

2.5.1 Fluorescent slides

Frozen sections at 10 to 100 mm thicknesses were col-lected with minimum lighting on Fisher Plus Goldslides using a Leica CM3050S cryostat (Nussloch,Germany). The sections were fixed in 10% neutralbuffered formalin for 10 minutes and then rinsedwith filtered water. Sections were coverslipped usingFluoro-Gel aqueous mounting medium and weresealed at the edges with nail polish to prevent dry-ing. Slides were stored in the dark at 4 �C.

2.5.2 Hematoxylin and Eosin (H and E)staining

H and E staining methods were based on standard pro-tocols provided by Dako and are available online at<www.dako.com/08066_12may10_webchapter13.pdf>(Dako Canada Inc., Burlington, Canada).


Journal of

BIOPHOTONICS


2.5.3 Immunohistochemistry withintercellular adhesion molecule 2(ICAM2) staining

Immunohistochemistry methods were based on thestandard procedure provided by BD Biosciences(San Diego, CA). Briefly, cryo-sections were col-lected for immunohistochemical staining at a thick-ness of 6 mm on charged slides. Antibodies were usedat 1/25 and 1/50 dilutions. Rat anti-mouse ICAM2antibodies and biotinylated goat anti-rat immunoglo-bulin were, respectively, the primary and secondaryantibodies used.

2.6 Microscopy

Both bright field and fluorescent microscopy wereperformed using a Zeiss Observer.Z1 non-confocalsystem (Carl Zeiss Canada Ltd, Toronto). The objec-tive lenses used were: EC “Plan-neofluar” 10�/0.3DICI (1.11 mm optical resolution) and “N-Achro-plan” 40�/0.65 (0.51 mm optical resolution). Two ex-citation modes were used unless indicated otherwisein the text. For sections labeled with FITC-lectin, theColibri LED-Module 470 nm with filter set 62 HE B/G/HR R was used (manufacturer preset). For sec-tions labeled with Cy5.5-lectin, the Colibri LED-Module neutral white with filter set 32 AF 680 (exci-tation BP 665/45, emission 725/50, beam splitter FT695) was used. Images were captured using Zeiss Ax-ioCam MR Rev3 (1388� 1040 standard mono resolu-tion, digital gain factor of 2 with an index of 1) andZeissAxioCamICc3 (2080� 1540 8-bit RGB colourresolution, digital gain factor of 2 with an index of 1).

2.7 Intensity correlation analysis

Images were processed and analyzed using AxioVi-sion v.4.7.2.0 and ImageJ v.1.44. Using ImageJ withthe Intensity Correlation Analysis ImageJ plug-indescribed and developed by Li et al. [32], pixel inten-sities for fluorescent images were measured over aselected region of interest (ROI). The resulting plotprofile was then used to determine signal to back-ground noise.

2.8 Real-time Imaging

The method and equipment used was previously de-scribed by Gussakovsky et al. [33] Briefly, the heartwas illuminated using six 635 nm 5 mW laser diodesand images were captured using an infrared-sensitive

CCD camera (RS Roper Scientific NTE/CCD-512-EBTF.GR-1, Tucson, AZ), with a Nikon Micro AF60lens at f/8 and a Schott RG 695 glass filter. The cam-era was placed approximately 5 cm away from theheart. Images captured every 30 s with an exposuretime of 2500 ms for 70 minutes (20 min baseline,30 min infusion, 20 min washout) were then pro-cessed using Matlab (The MathWorks, Natick, MA).

2.9 Statistics

Unpaired student T-test was used for data compari-son calculations. Differences were considered statis-tically significant when P � 0.05. Data are presentedas means � standard deviation.

3. Results

3.1 Synthesis of Cy5.5-lectin conjugates

To demonstrate the effectiveness of the reaction be-tween Cy5.5-NHS ester and LEA lectin as well asthe effectiveness of dialysis as a means to removeexcess dye, Cy5.5-lectin conjugation reactions wereconducted in the absence or presence of 100-fold ex-cess of lysine. Figure 1A shows the absorbance spec-tra of post-dialysis Cy5.5-lectin conjugate (i) as wellas the post-dialysis product of Cy5.5-NHS ester con-jugation reaction blocked with lysine (ii). The absorb-ance spectra acquired from the lysine-blocked reac-tion showed no peak at 675 nm, indicating removalof the small Cy5.5-lysine conjugate through the dia-lysis membrane. The protein peak at 280 nm due tolectin is present in both post-dialysis reactions. Thefluorescence spectra (Figure 1B) showed similar re-sults with the Cy5.5-lectin conjugate reaction show-ing a strong peak at 694 nm compared to the almostcomplete lack of peak in the lysine-blocked reaction.Conjugation of Cy5.5-NHS ester to lectin resulted ina Cy5.5-labelled product that had a D/P ratio of 2.90� 1.54 (n ¼ 6).

3.2 Binding of Cy5.5-lectin to endothelialcells in mouse hearts perfusedin Langendorff mode

3.2.1 Visible-NIR point spectroscopyof Cy5.5-lectin perfused mouse hearts

To demonstrate specific binding of Cy5.5-lectin toendothelial cells, isolated mouse hearts were per-fused in a Langendorff mode with KHB containing




Cy5.5-lectin (3.3� 10�8 mol/g wet tissue). Cardiacfunctional data revealed that infusion of Cy5.5-lectindid not adversely affect the heart function. FollowingCy5.5-lectin infusion, no change in the pressure rateproduct was observed: 6116.2 � 3033.7 beats/min

� mmHg versus 5013.0 � 2603.1 at baseline (n ¼ 7).The oxygenation of the heart as seen in the oxygena-tion profile (Figure 2) was adequate as demonstratedby the presence of two oxy-myoglobin peaks at 545and 581 nm before, during and post-infusion. POD

Figure 1 Cy5.5-lectin conjugation reaction in the presence and absence of free lysine. (A) i) Absorbance spectra of Cy5.5-lectin conjugate measured as described in “Materials and Methods” Section 2.2.1 (A) ii) Absorbance spectra of Cy5.5-NHS ester reacted with lectin in the presence of free lysine (in 100-fold excess). (B) Fluorescence spectra, acquired asdescribed in “Materials and Methods” Section 2.2.3, of the products after dialysis with the black spectrum representing thereaction product between activated dye and LEA lectin (corresponds to the absorbance spectra Ai) and the grey spectrumrepresenting the reaction product blocked by excess lysine (corresponds to the absorbance spectra Aii). The small peak ataround 650 nm is due to scatter of the laser light.

Figure 2 Pseudo-optical density(POD) spectra of a representa-tive normal mouse heart perfusedwith Krebs-Henseleit buffer andCy5.5-lectin as described in “Ma-terials and Methods” Section 2.3.2Spectra were taken prior toCy5.5-lectin infusion as a baseline(thick line) and after the 20 min.(dashed line) and 30 min. (thinline) mark during the infusionperiod. For clarity of the presen-tation, spectra were offset to zeroat 725 nm. Oxy-myoglobin, Cy5.5and water peaks are indicated witharrows.


Journal of

BIOPHOTONICS


spectra acquired during Cy5.5-lectin conjugate infu-sion showed the steadily increasing POD peak at680 nm due to accumulation of Cy5.5-lectin in thehearts (Figure 2B).

3.2.2 Histology of Cy5.5-lectin ex-vivoperfused mouse hearts

Colocalization analysis of images captured fromhearts infused with both Cy5.5-lectin and FITC-lec-tin showed that the staining patterns from the twoconjugates superimposed (Figure 3A). The samestructures appeared to be stained by both conjugatespixel for pixel (Figure 3B). Due to the type of ob-jective lens used, automatic sequential imaging atthe two wavelengths was limited because chromaticaberration occurred. Manual correction post-imagingwas performed so that the images more closely over-lapped so that quantitative analysis could be per-formed (Figure 3B). Specificity of Cy5.5-lectin bind-ing to endothelial cells was demonstrated by replacingCy5.5-lectin with Cy5.5-BSA. BSA did not bind anystructures in the hearts and subsequently, images ofthe mouse heart post-infusion with Cy5.5-BSAshowed a lack of fluorescence signal (Figure 3C). Asa non-fluorescent control, endothelial cells were alsoidentified via immunohistochemistry staining usingICAM2 antibodies. The brown stain indicated thelocation of the endothelial cells to be along the wallsof larger vessels (Figure 3D). Smaller capillaries werealso visible as small brown spots. Complete networksof capillaries that normally measure 5–10 mm in dia-meter were invisible due to the thickness of thesections (6 mm). Overall, staining pattern from car-diac tissue labelled with Cy5.5-lectin, FITC-lectin orICAM2 antibodies were similar thus demonstratingspecific binding of LEA lectin to the same structures,namely vascular endothelial cells. General structuralimages were also acquired in order to establish abaseline (Figure 3E). Background noise due mainlyto autofluorescence at the 470 nm was noticeablymore prominent than that at 680 nm. To demon-strate the reduction of background fluorescence inthe NIR spectral range, hearts infused with Cy5.5-lectin and FITC-lectin and prepared for fluorescencemicroscopy were excited at 470 nm, 590 nm and680 nm (Figure 4). At 470 nm and 590 nm, back-ground fluorescence for 10 mm sections was promi-nent; however at 680 nm, background fluorescencein comparison to the signal from Cy5.5-lectin wasminimal thus allowing for experimentation withthicker tissue sections. Increasing section thicknessprovided depth to the images at 680 nm (Figure 5).In thin 10 mm sections, capillaries appeared as spotsand speckles in the image while at 30 mm these spotsappeared as a visible network. Increasing the section

thickness to 70 mm resulted in images where the ca-pillary network was still clearly visible. At 100 mmhowever, the images became too busy to identify in-dividual capillaries. Comparatively, in the thickersections imaged at 470 nm, stronger background noisesignificantly decreased the visibility of individual ca-pillaries (not shown). Exposure times for the imagescaptured at 680 nm also decreased from 7773 ms to2525 ms as section thicknesses increased.

3.3 Real-time fluorescence imaging

Imaging of Cy5.5-lectin accumulation in intact heartswas successfully performed during infusion of Cy5.5-lectin conjugates into mouse hearts perfused in aLangendorff mode (Figures 6, 7). To eliminate move-ment artifacts, the hearts were arrested with high-po-tassium KHB cardioplegic solution. High potassiumcardioplegic arrest preserved cardiac energetics andstructure [34]. Control hearts showed fairly uniformedstaining with a linear increase in fluorescence duringthe infusion period of 30 minutes. Average peak in-tensity of fluorescence was 150 a.u. (n ¼ 8). A ki-netic curve was generated by comparing the pixel in-tensity from the same ROIs within approximately140 images (1 image per 30 s) that were capturedduring the infusion (Figure 7). A linear fit of theslopes during the infusion period resulted in R2 orcorrelation values close to 1 (0.992 � 0.004, n ¼ 4).Images captured during the washout period indi-cated that no significant loss of conjugate occurredduring the 20 minutes because there was little or nochange in fluorescence intensity (slope of linear fitwas near 0) (Figure 7). This demonstrates that thebinding between endothelial cells and Cy5.5-lectin isstable. To compare staining of endothelium in theperfused versus non-perfused areas of mouse hearts,surface images of normal and LAD-ligated heartswere acquired. LAD ligation was confirmed by opti-cal point spectroscopy. Post-ligation, the area nor-mally perfused by the LAD showed a POD spectrumlacking the two oxy-myoglobin peaks at 545 and581 nm (figure not shown). Instead, a deoxy-myoglo-bin peak at 553 nm was seen (figure not shown). Incaptured images, these areas showed very little in-crease in fluorescence intensity from the baselinereadings. Average peak intensity of fluorescence was35 a.u. (n ¼ 6).

3.4 Histology and imaging of in-vivo staining

Images collected from mouse hearts stained using in-vivo methods showed staining patterns similar tothose found when using the ex-vivo injection of the





Journal of

BIOPHOTONICS


conjugate (Figure 8). Exposure times for images ex-cited at 670 nm were approximately 14 s for 10 mmheart slices, which was higher than the average of 7 sfor 10 mm sections from hearts stained using ex-vivomethods. Weaker signal is attributed to decreasedCy5.5-lectin per gram of tissue (1.8� 10�10 moles ofCy5.5-lectin/g of tissue) as compared to the 3.3�10�8 mol/g achieved with the Langendorff mode.

4. Discussion

4.1 Specificity of Cy5.5-lectin binding

Binding of Cy5.5-lectin is directed by the LEA lec-tin component of the conjugate (MW �71 kDa) asshown by the lack of binding in the experiments withCy5.5-BSA (Figure 3C) and Cy5.5-lysine (see sup-plemental material 1). BSA is a serum protein (MW�68 kDa) that does not specifically bind to endothe-lial cells and has been shown to remain briefly in cir-culation before being cleared by metabolic organs

3

Figure 3 (online color at: www.biophotonics-journal.org) Colocalization of Cy5.5-lectin labelling with FITC-lectin andICAM2 labelling for endothelial cells. (A) Images show a 10 mm cross-section through a mouse heart stained with bothFITC-lectin (lexcitation at 494 nm, lemission at 518 nm) and Cy5.5-lectin (lexcitation at 675 nm, lemission at 694 nm) conjugates.Images are shown in at 100� magnification and then focused on a large vessel at 400� magnification. All images wereacquired in black and white and then pseudo-colored. FITC was excited at 470 nm and Cy5.5 was excited at 680 nm (See“Materials and Methods” Section 2.6 for complete imaging parameters). Each image was captured in separate channelsand then layered to show regions of overlapping signals, seen here as orange. The white dashed line indicated the ROIselected for the colocalization analysis. (B) The graph shows a similar profile for both the FITC-lectin signal as well as theCy5.5-lectin signal with intensities peaking at the similar pixels over a total distance of 1800 pixels. (C) Control imageobtained from a 10 mm cross-section through a mouse heart perfused and infused with Cy5.5-BSA under the conditionsdescribed in “A”. Exposure time was 7773 ms. Note the lack of fluorescence signal. (D) i) An example of stained vesselsfrom a mouse heart section labelled with ICAM2 antibodies at 100� total magnification and ii) at 400� total magnifica-tion. A negative control section from the same heart was only stained with hematoxylin (viewed at 400� total magnifica-tion). (E) Section of mouse heart tissue stained with H and E at 400� total magnification. The dark blue dots are nucleiand the magenta-red areas indicate cytoplasm and extracellular matrix.

Figure 4 Reduction of autofluo-rescence within the NIR spectralrange. At 590 nm excitation,images of cross-sections throughmouse hearts were captured usingthe Colibri LED-Module 590 nmexcitation laser with filter set62 HE B/G/HR R preset forHcRed by Zeiss. At 470 nm and680 nm, images were captured asdescribed in section 2.6. The cam-era exposure time for the auto-fluorescence only image at 470 nmwas 1086.6 ms. Comparatively, theFITC-lectin image was exposed for200 ms and the Cy5.5-lectin imageat both 590 nm and 680 nm was ex-posed for 7773 ms. All images areat 100� magnification. Note thedifference in background signaleven though the pattern of stainingremains the same.




[35]. With Cy5.5-lysine, the reactive groups of thedye were blocked and again, no non-specific bindingwas seen. With Cy5.5-lectin, binding is limited to

vascular endothelial cells. Qualitatively and quantita-tively, staining patterns seen with Cy5.5-lectin matchthe staining pattern of a proven probe for endothe-lial cells such as FITC-lectin (Figure 3A and B).Other cells such as red blood cells were not labelledwith Cy5.5-lectin since no signal was seen in bloodsamples collected from tail-vein experiments post-in-jection (not shown). Also, binding of Cy5.5-lectinconjugate is limited to areas of the heart that areperfused. This was seen in the lack of signal fromthe LAD ligated areas of the heart.

4.2 Synthesis

The D/P ratio achieved for the reaction, 2.90 � 1.54(n ¼ 6), was within the 2–6 moles of dye per mole ofLEA lectin range that commercial products such asFITC-lectin tend to have. By choosing to use themonofunctional Cy5.5-NHS ester instead of the bis-functional form of the dye, oligomer formation wasprevented which guaranteed a clean end-productpost-dialysis and decreased chances of self-quench-ing due to dye-dye interactions [36]. Although dialy-sis is a simple and basic purification method, it wasproven to be sufficient for this purpose through ex-perimentations with lysine-blocked reactions.

Figure 5 Cross-sections through a mouse heart infused withCy5.5-lectin conjugate at 100� magnification and excited at680 nm. Section thicknesses were set at 10 (a), 35 (b), 70 (c)and 100 mm (d). Exposure times also decreased accordinglyfrom 7773 ms, to 5922 ms, to 2712 ms, to 2525 ms. Note thelarger vessels that can be seen in the thicker sections.

Figure 6 Images captured from real-time imaging of infusion of Cy5.5-lectin into arrested mouse hearts under normalLangendorff perfusion (A) and LAD-ligation conditions (B) as described in “Materials and Method” Section 2.3. Undernormal conditions, the fluorescence image shows a fairly uniform signal. In the ligated heart, a dark region below theligation point area shows little fluorescence. Non-cardiac objects are also seen such as the cannula and water-filled balloonportion of the function probe (small black arrows). Post-infusion 40 mm cross-sections images were also captured using amicroscope (C) and (D), normal and LAD-ligated respectively). Images are a mosaic of images captured sequentially at100� magnification and then stitched together. The dashed line on the normal heart under white light indicates the ap-proximate location from which the short-axis tissue section was taken and imaged. Boxes shown in the post-infusion imagefor the LAD ligated heart indicates the regions of interest (ROIs) used to calculate the kinetic curves in Figure 7. ROIsare numbered 1–4 starting at the top left and moving counter-clockwise. RV: Right Ventricle, LV: Left Ventricle.


Journal of

BIOPHOTONICS


4.3 Advantages of using Cy5.5-lectinconjugates

The minimal amount of background interferencewithin the NIR-range allowed for imaging of thickersections of upward to 100 mm and even imaging ofthe whole heart as shown (Figure 6), which is impos-sible with non-NIR dye-conjugates. Images capturedfrom real-time ex-vivo cardiac imaging sessions were,however, limited by the camera resolution but canstill clearly indicate differences in perfusion. In thecontrol heart, normal microvascular perfusion het-erogeneity that is known to exist in healthy heartswas observed [37]. The fluorescence captured in thepoorly perfused area of the LAD-ligated heart is lar-gely due to light scatter and tissue transparency inthe NIR range. Indeed, NIR light can penetrate car-diac tissue for up to 4 mm, which is almost the fullthickness of the mouse heart [3]. From Figure 8, thebinding was very stable (over 30 min ex-vivo); there-fore, redistribution or diffusion into the non-perfusedareas was unlikely thus allowing Cy5.5-lectin conju-gate to be used a deposition microperfusion tracer.First-pass flow tracers (e.g., indocyanine green, ICG)transiently pass through the organ (e.g., heart) andthe tracer’s distribution and/or clearance kinetics isdetermined during this time and used to evaluate theperfusion [38]. For this type of tracer, clearance ki-netics is very important. In contrast, a deposition flow

Figure 7 A representative kineticsgraph showing Cy5.5-lectin bind-ing in a LAD-ligated mouse heartbased on data shown in Figure 6.ROI 1 and ROI 2 are from thewell-perfused region of the heartwhereas ROI 3 and ROI 4 arefrom the region of poor perfusiondue to the ligation. The fluores-cence intensity is a mean of pixelintensities collected from the ROIselected in the set of imagescaptured real-time. A negligibleamount of Cy5.5-lectin is lost inthe washout since fluorescence in-tensity only slightly drops (slope� 0) after the end of the infusionperiod.

Figure 8 A 30 mm cross-section through mouse cardiac tis-sue infused with Cy5.5-lectin using the in-vivo tail-vein in-jection as described in “Materials and Method” Section2.4. The exposure time was 13 seconds. Note the similarstaining pattern between this image and the images fromex-vivo staining (Figure 5). The 30 mm cross-sectionthrough liver and kidney tissue was labeled and preparedin a similar fashion to the heart. The image of the spleenwas acquired from a 50 mm section. The exposure time forthe liver, kidney, and spleen images were 24 seconds. Allimages were acquired at 100� total magnification. BV:blood vessel, MZS: marginal zone sinuses, WP: white pulp,CV: central vein, S: sinusoids, RC: renal corpuscles.




tracer (e.g., microspheres) is retained by the tissue; itcan be later detected in the tissue and quantitated[21]. For fluorescent deposition tracers, fluorescenceimaging can be used to evaluate the perfusion [33].

Although there was a difference in signal inten-sity between in-vivo and ex-vivo infusion methods,mostly due to a difference in dye content per gramof tissue, the overall quality of the signal produced issufficient for the detection methods used. Once thetissue is processed into slides, the fluorescent signalremains stable for several days with proper storage.However, decay of the signal could be acceleratedby the imaging process, which sometimes requiredrepeated excitation of the dye thus causing photo-bleaching. Variations in signal intensity could be dueto various effects such as quenching. Based on absor-bance spectra there is some minimal self-quenchingthat occurs for Cy5.5-lectin (determined by the in-creased shoulder peak at approximately 630 nm post-conjugation in Figure 1). Although the fluorescenceintensity is not optimal due to the self-quenching, itis still strong enough for the ex-vivo and in-vivo ima-ging used. It is possible that in-vivo protease activitycould cause the cleavage of the LEA lectin moleculeand recover the fluorescence lost due to self-quench-ing however, binding of the probe would then becompromised.

4.4 Future applications based on in-vivoimaging

Use of the conjugate is not limited to the heart aswe observed staining of the microvascular endothe-lium in the liver, kidney, and spleen during in-vivoexperiments (Figure 8). The non-toxic nature of theCy5.5-lectin conjugate allows it to have potentialuses in in-vivo imaging that could also expand be-yond the mouse to other animal models.

5. Supplemental online material

Supplemental material 1: Control image obtainedfrom a 10 mm cross-section through a mouse heartperfused and infused with Cy5.5-lysine. Exposuretime for the image was 7773 ms which is comparableto other control and experimental images. Note thelack of fluorescence signal, which indicates a lack ofnon-specific binding due to the Cy5.5 component ofthe conjugate. It can be concluded then that anybinding by Cy5.5-lectin is therefore due to the lectincomponent of the conjugate.

Supplemental material 2: Real-time fluorescenceimaging of healthy and LAD-ligated mouse hearts

perfused with Cy5.5-lectin (1.8� 10�5 M) for 30 mincan be watched as videos; each second of the videocorresponds to 2 minutes of experimental time.

Cecilia Nguyen holds a B.Sc.degree with honours fromthe University of Winnipeg.She worked as a summerstudent at the Institute forBiodiagnostics within theNational Research Councilof Canada between 2009and 2010. Currently, she is aM.Sc. student in the Depart-

ment of Oral Biology at the University of Manitoba.

Saro Bascaramurty is aTechnical Officer at the In-stitute for Biodiagnostics,National Research CouncilCanada since 1994. Sheholds a Diploma in MedicalLab Technology with a widerange of extensive experi-ence in the clinical & re-search fields. Currently, she

is responsible for validating data from various non-in-vasive techniques with the traditional histological meth-ods.

Bozena Kuzio received herM.Sc. in biology and is aTechnical Officer at the Na-tional Research Council ofCanada – Institute for Bio-diagnostics since 1996. Herexpertise includes non-inva-sive and surgical studies ofanimal hearts as well as bio-chemical assays.

Lori Gregorash received abachelor of agriculture fromthe University of Manitobaand is an Animal HealthTech. She has worked at theInstitute for Biodiagnosticssince 1994 with the SurgicalServices team.


Journal of

BIOPHOTONICS


Acknowledgements The research was supported, in part,by operating grants from Manitoba Health Research Coun-cil and the Heart and Stroke Foundation of Manitoba toDr. O. Jilkina. Ms. C. Nguyen is a recipient of the ManitobaGraduate Scholarship.

References

[1] E. R. McVeigh, Circ. Res. 98, 879–886 (2006).[2] J. L. Kovar, M. A. Simpson, A. Schutz-Geschwender,

and D. M. Olive, Anal. Biochem. 367, 1–12 (2007).[3] S. P. Nighswander-Rempel, V. V. Kupriyanov, and

R. A. Shaw, J. Biomed. Opt. 10, 024023 1–7 (2005).[4] N. S. Umapathy, Z. H. Fan, E. A. Zemskov, I. B. Alieva,

S. M. Black, and A. D. Verin, Vascul. Pharmacol. 52,199–206 (2010).

[5] R. F. Furchgott and P. M. Vanhoutte, FASEB J. 3,2007–2018 (1989).

[6] Z. Bagi, A. Koller, and G. Kaley, Am. J. Physiol.Heart Circ. Physiol. 285, H1404–H1410 (2003).

[7] E. B. Okon, A. W. Y. Chung, P. Rauniyar, E. Padilla,T. Tejerina, B. M. McManus, H. Luo, and C. van Bree-men, Diabetes 54, 2415–2423 (2005).

[8] M. R. Hoenig, C. Bianchi, A. Rosenzweig, and F. W.Sellke, Curr. Vasc. Pharmacol. 6, 292–300 (2008).

[9] R. Ross, N. Engl. J. Med. 340, 115–126 (1999).[10] S. Sitia, L. Tomasoni, F. Atzeni, G. Ambrosio, C. Cor-

diano, A. Catapano, S. Tramontana, F. Perticone,P. Naccarato, P. Camici, E. Picano, L. Cortigiani,M. Bevilacqua, L. Milazzo, D. Cusi, C. Barlassina,P. Sarzi-Puttini, and M. Turiel, Autoimmun. Rev. 9,830–834 (2010).

[11] A. Romero-Corral, F. H. Sert-Kuniyoshi, J. Sierra-Johnson, M. Orban, A. Gami, D. Davison, P. Singh,S. Pusalavidyasagar, C. Huyber, S. Votruba, F. Lopez-Jimenez, M. D. Jensen, and V. K. Somers, J. Am. Coll.Cardiol. 56, 662–666 (2010).

[12] E. Dejana, M. Simionescu, and H. Wolburg, Cell Tis-sue Res. 335, 1–3 (2009).

[13] Y. G. Ko and P. X. Ma, J. Colloid Interface Sci. 330,77–83 (2009).

[14] A. A. Bogdanov, C. P. Lin, and H. W. Kang, Pharm.Res. 24, 1186–1192 (2007).

[15] D. Citrin, A. K. Lee, T. Scott, M. Sproull, C. Menard,P. J. Tofilon, and K. Camphausen, Mol. Cancer Ther.3, 481–488 (2004).

[16] L. Laitinen, Histochem. J. 19, 225–234 (1987).[17] P. L. Debbage, S. Seidl, A. Kreczy, P. Hutzler, M. Pa-

velka, and P. Lukas, Histochem. Cell Biol. 114, 259–275 (2000).

[18] S. Mazzetti, S. Frigerio, M. Gelati, A. Salmaggi, andL. Vitellaro-Zuccarello, Eur. J. Histochem. 48, 423–428(2004).

[19] M. S. Nachbar, J. D. Oppenheim, and J. O. Thomas, J.Biol. Chem. 255, 2056–2061 (1980).

[20] C. Banchonglikitkul, J. D. Smart, R. V. Gibbs, S. J.Donovan, and D. J. Cook, J. Drug Target 10, 601–606(2002).

[21] F. W. Prinzen and J. B. Bassingthwaighte, Cardiovasc.Res. 45, 13–21 (2000).

[22] G. Munch, S. McKay, E. Gussakovsky, B. Kuzio, V. V.Kupriyanov, and O. Jilkina, J. Biomed. Op. 16, 1–13(online) (2011).

[23] R. C. Marshall, P. Powers-Risius, B. W. Reutter, J. P.O’Neil, M. La Belle, R. H. Huesman, and H. F. Van-Brocklin, J. Nucl. Med. 45, 1950–1959 (2004).

[24] G. Arsos, A. Kyparos, E. Moralidis, D. Kyparos,S. Georga, S. Sotiriadou, C. Matziari, and C. Karakat-sanis, Physiol. Res. 58, 21–28 (2009).

[25] G. Heusch, R. Schulz, M. Haude, and R. Erbel, J. Mol.Cell Cardiol. 37, 23–31 (2004).

[26] O. Jilkina, B. Xiang, B. Kuzio, J. Rendell, and V. V.Kupriyanov, Magn. Reson. Med. 53, 1172–1176 (2005).

[27] O. Jilkina, B. Kuzio, J. Rendell, B. Xiang, and V. V.Kupriyanov, J. Mol. Cell Cardiol. 41, 893–901 (2006).

[28] O. Jilkina, M. Glogowski, B. Kuzio, P. Zhilkin, E. Gus-sakovsky, and V. V. Kupriyanov, Int. J. Cardiol. 149,315–322 (2011).

[29] S. J. Millar, B. W. Moss, and M. H. Stevenson, MeatSci. 42, 277–288 (1996).

[30] J. Donovan and P. Brown, in: In Current Protocols inImmunology, edited by J. E. Coligan, A. M. Kruisbeek,D. H. Margulies, E. M. Shevach, and W. Strober (JohnWiley and Sons, Hoboken, NJ, 1998), pp. 1.4.1–1.4.5.

Valery Kurpiyanov is a grad-uate of the Moscow Lomo-nosov State University. Hehas a Ph.D. degree in bio-chemistry and a Doctor ofScience degree in biochem-istry/cardiology. He was aSenior Research Officer (re-tired now) at the Institutefor Biodiagnostics of theNational Research Councilof Canada. He studied ap-plications involving MR and

near-infrared imaging to further understand and diag-nose cardiac diseases.

Olga Jilkina received herPh.D. in biochemistry fromUniversity of Manitoba, Ca-nada. She is a ResearchOfficer at the Institute forBiodiagnostics at NationalResearch Council of Canadaand Adjunct Professor inthe Department of OralBiology at the University of

Manitoba. She is involved in studies of cardiac metabo-lism using NMR spectroscopy and optical spectroscopyand imaging.




[31] G. Thurston, P. Baluk, A. Hirata, and D. M. McDonald,Am. J. Physiol. 271, H2547–H2562 (1996).

[32] Q. Li, A. Lau, T. J. Morris, L. Guo, C. B. Fordyce, andE. F. Stanley, J. Neurosci. 24, 4070–4081 (2004).

[33] E. Gussakovsky, B. Kuzio, Y. Yang, and V. V. Kupriya-nov, J. Biophotonics 4, 277–287 (2011).

[34] O. Jilkina, B. Kuzio, G. J. Grover, C. D. Folmes, H. J.Kong, and V. V. Kurpiyanov. Biochim. Biophys. Acta1618, 39–50 (2003).

[35] Y. Hama, Y. Koyama, P. L. Choyke, and H. Kobayashi,J. Biomed. Opt. 12, 034016–17 (2007).

[36] M. Ogawa, C. A. S. Regino, P. L. Choyke, and H. Ko-bayashi, Mol. Cancer Ther. 8, 232–239 (2009).

[37] T. Matsumoto and F. Kajiya, Basic Res. Cardiol. 96,547–552 (2001).

[38] V. V. Kupriyanov, S. Nighswander-Rempel, andB. Xiang, J. Mol. Cell Cardiol. 37, 947–957 (2004).


Journal of

BIOPHOTONICS


near-infrared fluorescence imaging of mouse myocardial microvascular endothelium using cy5.5-lectin...

Documents