the time to compute a full displacement map was reduced to less than10 seconds for a 256x256 pixel image. The most likely traction force that can cause this displacement wasfound by generating a matrix of the Boussinesq equations in two dimensions, which describe thedisplacements at every point in a purely elastic half space that are produced by a force shear to the surfaceacting over a defined area (Butler et al., 2002; Dembo and Wang, 1999; Marganski et al., 2003). By limitingtraction nodes to locations inside of the resting cell outline, an over-specified system was generated in whichmore displacement vectors were known than traction locations. By using a method of Tikhonov regularization,which balances the least squares error of a displacement fit to traction magnitudes and directions with a forcebalance, torque balance and penalty for high magnitudes, a traction map is generated that minimizes theeffects of noise and non-uniformity of the displacement map due to non-uniformity in the random distribution ofmarkers in the gel.C3.1.f. Collagen Gel Preparation. Collagen and neutralizing solution (100 mM Hepes, pH 7.3 in 2* PBS) arekept at 4 C until needed. Adult cardiac myocytes are isolated and allowed to aggregate by gravity at thebottom of a 15ml tube. Cells are diluted to 1 million 100,000 per ml. Collagen gels have a final concentration of2 mg/ml. Media, neutralizing solution and then collagen are added to the tube containing the cells working onice. The final solution quickly becomes viscous and takes on a yellow/pinkish hue. We pipette 8 ml perstretching or traction force plate and place the plate in the incubator, at 37 C. The collagen polymerizes aroundthe cells in 30-60 minutes, and 5 ml media is then carefully added on top of the gel. Media is changed every 3-214 Knowlton, Kirk U.4 days. For immunoblotting of adult cardiac myocytes cultured in 3D collagen gels, cells are detached with 0.5mM EDTA in calcium and magnesium-free PBS and media is removed. Next 1 ml 2* lysis buffer is added tothe gel, the cells are lysed and the lysate centrifuged at 4 C for 12 minutes. 30 ul 3* Laemmli sample buffer isadded to 60 ul supernatant, heated for 3 minutes at 100 C, and loaded into an SDS-PAGE gel for proteinanalysis.NEW METHOD DEVELOPMENT: We will extend these traction force microscopy methods (demonstrated insection C2.1.c) from neonatal rat ventricular myocytes to apply to them neonatal and adult mouse ventricularmyocytes. Having previously successfully plated and stretched mouse neonatal ventricular myocytes (Knoll etal., 2002), we expect these traction force methods to apply to mouse neonatal myocytes without significantmodification. However, for adult myocytes, we will also need to add a layer of collagen gel containing the adultmyocytes on top of the polyacrylamide substrate to achieve adequate cell-matrix adhesion. Given that we haveseen isolated adult myocytes can be adhered at their ends with poly-lysine and generate physiological systolictension without detaching (Bluhm et al., 1995), and given that we have also seen that neonatal myocytes cangenerate physiological active stresses when adhered polyacrylamide gels (section C2.1.C), we are confidentthat conditions can be created that will allow adult murine ventricular myocytes to be sufficiently adhered to asufficiently gel to generate measurable substrate deformations and thus to calculate traction forces. Cells willembedded in a collagen gel containing poly-lysine to increase cell attachment strength. The collagen gel will beformed by mixing 1.55 mg/ml collagen type I (BD Biosciences) in 100 mM HEPES buffer at pH 7.3 and placingthis solution at 37 C for 30 minutes to gel. Another gel solution will be made containing the same concentrationof collagen (1.55 mg/ml) in the BDM-containing heart buffer solution plus 2.5% 500 nm diameter fluorescein-coated polystyrene beads (Polysciences), 0.05 mg/l poly-L-lysine-succinimide (sigma) and a myocytesuspension of 10,000 cells/cm2 based on the gel surface area. This solution will be pipetted on top of theoriginal solution and allowed to sit at room temperature for 30 minutes before placing at 37 C for 30 minutes.This low-temperature gelation allows the cells and beads to settle to the lower surface along one plane tofacilitate microscopy and analysis. When the gel has polymerized, it will be washed 3 times with Tyrodessolution containing 2 mM calcium and allowed to sit in that solution until the cells begin to contract.C3.1.g. Myocvte Micropatterninci. We have developed two methods for micropatterning deformablemembranes for engineering patterned myocytes. The protocols combine photolithographic, microfluidic andmicro-grooving techniques, to micropattern cell adhesion sites on elastic membranes, with cardiac myocyteand fibroblast isolation and culture. The resulting in vitro models of cardiac tissue mimic important structuraland functional aspects of native ventricular myocardium, and show improved preservation of cell-shape,alignment and connectivity, while providing control over cell-matrix adhesion and mechanical environment (e.g.application of stretch). Both of these approaches are described in detail in a new article in press in NatureProtocols (Camelliti et al., 2006). The original microfluidic patterning technique (also known as microcontactprinting) uses flat silicone rubber membranes, onto which stripes are deposited that consist of extra-cellularmatrix proteins to guide cell attachment (Gopalan et al., 2003). This technique is shown in Fig. 12A andinvolves a reusable microstructured silicone rubber stamp, which is cast from photolithographically etchedelectronics-grade silicon metal wafers. The stamp is manually sealed onto the target membrane, formingmicrofluidic channels for spatially-restricted application of matrix protein solution. After drying of the solution,the stamp is carefully removed, leaving a patterned growth substrate on the membrane. It is possible todouble-print structures, for example to create regularly intersecting patterns of parallel lines, therebytransforming pseudo-ID structures (lines of cultured cells) into a pseudo-2D representation of tissue (grids).Isolated cardiac cells have been successful plated onto such tracks, made of collagen. This procedureprovides good spatial resolution and definition of spatial anisotropy, but results in sparse cultures.Microfabrication of deep microstructures is achieved by molding a replica of polydimethylsiloxane (PDMS) frometched silicon wafers. The mold is then used to create a pattern of collagen, laminin, fibronectin, or a mixtureby microfluidics. After 5 minutes, the microfluidic mold is removed and excess matrix protein washed away,leaving the pattern on the substrate. Myocytes are then cultured onto the membrane or dish as describedabove. On the patterned substrata, cells adhere, spread, grow and take on an elongated tissue-like phenotype.Non-specific absorption of proteins from serum is minimal since the cells are serum starved during thestretching protocols (to prevent non-stretch-induced cellular growth). Cellular adhesion away from the215 Knowlton, Kirk U.substrate patterns can also be reduced by modifying the hydrophobic surfaces with polyethylene oxidepolymers (Li et al., 1996) or covalently coupled polyethyleneglycol (PEG). Spin negative photoresistle to UVthrough mask Transparency mask^Develop exposed photoresist Mlcropatterned SI wafer[Cast POMS mold PDMS main with micro-channels PDMS mold on membrane Fill channelswith collagen h arc In IncubatorRemove PDMS moldCollagen tracks on membraneCells on parallelcollagen tracks POMS mold on membrane'Fill channels with collagen 1 h3T*C in Incubator1 Remove PDMS mold 40'C overnightCriss-cross collagen Plate cellatracks on membrane Cells on criss-crossA collagen .tracksFigure 12. A. Micropatterning procedures. B. Equibiaxial and non-equibiaxial cell stretcher assemblyThe second method for micropatterning involves fabricating grooves into the silicone membrane itself (Fig. 13).With this method, cells are more confluent on the membrane, and thus more protein is available for analysisfrom each stretcher. Photolithographic patterning of silicon wafers is carried out as described previously(Bhatia et al.,1998). Briefly, negative photoresist is coated on a silicon wafer, exposed to ultraviolet light (365nm) through a transparency mask, and developed using high resolution printing of patterns with required widthand separation between the parallel lines (10 u,m parallel lines and 10 UJTI spacing). Polydimethylsilioxane(PDMS) is prepared from a mixture of 2 liquid components (Sylgard 186 Kit, Dow Corning), poured ontodeveloped wafer as a thin layer and after degassing, and cured. The resulting membranes are coated with there uired ECMfor aligning Itie myocytes. Figure 13. Novel myocyte patterning assay for Membrane micropatteminn improved quantitative measurements of protein changes due to chronic cell stretching. A PDMS PDMS Mold mold is fabricated and used to make thin silicone membranes with the same alignment pattern. A collagen (or other substrate) layer is Silicone membrane applied to the surface, and myocytes applied to the membrane. The cells align with the membrane and take an elongated, rod-like Collagen coating phenotype, and stretch patterns (longitudinal and/or transverse) are then applied to these cell preparations.Myocytes alignwith patternC3.1.h. Stretchers and Membranes. Silastic deformable membranes are prepared with a collagen type I andother substrate coatings similar to that used by Sadoshima et al. (Sadoshima et al.,1992). The cells (2 * 105cells/cm2) are plated on patterned or unpatterned silicon membranes coated with collagen type-1 and equippedon equibiaxial (uniform) or non-equibiaxial (elliptical) stretcher devices as shown in Fig. 12B (Lee et al., 1996).The cells are serum-depleted for 24 hr followed by passive stretch for an additional 24 hr. Total RNA isextracted and subjected to RNA protection assay (RPA) using Direct RPA system (Ambion Inc.). Mouse BMP216 Knowlton, Kirk U.coding sequences were amplified by RT-PCR and used as the template for a BMP riboprobe synthesis withmouse GAPDH as a control. Micropatterning allows aligned myocyte cultures to be subjected to non-equibiaxial stretch on the elliptical stretcher (Gopalan et al., 2003), which applies anisotropic strain to cells in aconstant ratio (e.g. 2:1). If cells are patterned with their axis aligned with one axis of the elliptical stretcher, theratio of principal strains is defined by the ratio of minor to major diameters of the ellipse. With these devices,the magnitude of strain is precisely controlled at any value and calibrated for each stretcher.C3.1.I. Cell Preparations for Histology. In order to evaluate the sarcomeric definition and organization inmyocytes plated on gels of different rigidities, cells are fixed and stained for markers such as a-actinin at 48hours and 7 days after initial plating. Cells are fixed by washing with cold (4 C) PBS followed immediately byimmersion in cold 0.75% glutaraldehyde for 20 minutes. Cells membranes are permeabilized with 0.1% TritonX100 (Sigma) for 5 minutes. Antibodies are diluted in 1% bovine serum albumin (BSA, Gemini) and added tothe fixed cells for 60 minutes at room temperature. Secondary antibodies of tetra-rodamine isothiocyanate(TRITC)-conjugated goat anti-rabbit (Jackson Immunoresearch, West Grove, PA) are diluted in 1% BSA andadded for 30 minutes. Cells are imaged with a confocal microscope at NCMIR or by phase-contrast andepifluorescence using our Nikon Eclipse TE300 inverted microscope and Photometries Cascade 51 2X CCDcamera and acquired using Metamorph software (Universal Imaging Corporation). In addition, cell area andaspect ratio (long axis/short axis) are computed by segmenting cells with ImageJ software (NIH).C3.1.I. Ratiometric Imagine] of Intracellular Calcium Transients. Isolated cells are labeled with Fura-2(Grynkiewicz et al., 1985). Fura-2 calcium indicator (Molecular Probes) is diluted to a concentration of 5 uM ina high-calcium Tyrodes buffer (130 mM NaCI, 5.4 mM KCI, 2 mM CaCI, 1 mM MgCI2, 0.3 mM Na2HPO4, 10mMHEPES, 5.5 mM Glucose at pH 7.4) and added at room temperature for 30 minutes. Cells are then washed 3times with fresh Tyrodes buffer and placed in Tyrodes buffer for the duration of the experiment. Cells areimaged using a Nikon Eclipse TE3000 inverted microscope and stimulated to beat using the same procedureas in the traction force measurements. A phase-contrast image of the resting myocytes is taken in order toconstrain the analysis of the image. Using a Lambda DG-4 high-speed filter changer (Sutler Instruments), cellsare illuminated with alternating 380 nm and 340 nm light and the fluorescence at 510 nm is imaged using aPhotometries Cascade 51 2X CCD camera and acquired using Metamorph software (Universal ImagingCorporation). The filters are switched and a new image is acquired every 15 ms for 4.5 seconds (300 images).In order to calibrate the Fura-2 response, an equal amount of high-calcium Tyrodes buffer containing 20 uMcalcium ionophore-4-bromo-A23187 (Molecular Probes) is added to saturate the Fura bound to calcium andimages are acquired at 340 nm and 380 nm excitation/ 510 nm emission. The media is removed and mediacontaining 10% ethylenediaminetetraacetic acid (EDTA, Sigma) is added in order to chelate all calcium andresult in a baseline Fura fluorescent signal. Again, images are acquired at a 340 nm and 380 nm excitation/510 nm emission.Images are analyzed using custom macros programmed in Metamorph or ImageJ (NIH) to split the imagesequence of alternating images at 340 and 380 nm excitation into 2 stacks, compute the ratio of those stacksand compute the average value of that ratio for each image inside the cell outline for manually outlined cellareas. The concentration of free calcium in the cell is then computed using the following formula: -R] max /where kd is the dissociation constant (0.14 uM), Q is the ratio of the total average fluorescence inside the celloutline at 380 nm excitation, R is the ratio of fluorescence at 340 nm excitation to the fluorescence at 380 nmexcitation in each image, Rmin is the same ratio upon addition of EDTA and f?ma* is the same ratio upon additionof the calcium ionophore.C3.1.k. FRET Imaging of Intracellular Signaling Events. Neonatal cardiac myocytes are transfected with arecombinant FRET reporter, such as the A-kinase activity reporter AKAR2 (Zhang et al., 2005; Zhang et al.,2001) 1 day after isolation using FuGeneS transfection reagent (Roche Diagnostics) and imaged 2 days aftertransaction (-5% transfection efficiency). For adult cells we will used viral-mediated gene transfer. AKAR2 is arecombinant protein composed of the yellow (YFP) and cyan (CFP) mutants of GFP, a protein kinase Asubstrate, and a phosphothreonine-binding domain. PKA-mediated phosphorylation of AKAR2 increases FRET 217 Knowlton, Kirk U.from CFP to YFP, allowing real-time fluorescence imaging of endogenous PKA activity by the yellow/cyanemission ratio. Fixed myocytes expressing reporter are labeled for a-actinin to confirm normal cell morphology.Cells are washed with Hanks buffered saline solution with 20 mM HEPES buffer (pH 7.2) 20 minutes prior toimaging and kept at 35 degrees C with a stage heater (Carel) during imaging. Transfected myocytes areimaged on a Nikon Eclipse TE300 inverted microscope equipped with a 60X Plan Apochromat objective,Photometries Cascade 512F CCD camera, and MetaFluor 6.2 software (Universal Imaging Corporation).Imaging is performed using a 430/25 nm excitation filter (for CFP) and simultaneously recording CFP (470/30nm) and YFP emissions (535/30 nm) with a DualView emission splitter (Optical Insights; Chroma filters).Images are acquired with 1 s exposure every 5 s. To obtain emission ratio time-courses, YFP or CFP emissionintensities for each image are averaged over a region of interest, background subtracted, and then yellow/cyanemission ratio is calculated, normalized by the ratio before reagent application. For emission ratio images, asimilar approach is used on a pixel-by-pixel basis with a 5x5 pixel median filter using MetaMorph software(Universal Imaging Corporation).C3.2. Isolated Murine Trabeculae and Papillary MusclesC3.2.a. Muscle Isolation and Mounting. Male or female mice are anesthetized with Isoflurane. After cervicaldislocation, the chest is opened and the heart is arrested by intracardiac injection of low calcium, highpotassium cardioplegic solution. Then the heart is rapidly removed, cannulated, and perfused with modifiedHepes buffered solution containing (mM): 137.2 NaCI, 15.0 KCI, 1.2 MgCI2, 2.8 mM Na acetate, 10 taurine, 1.0CaCb, 10mM glucose, and 10mM Hepes in equilibrium with 100% O2. The right ventricle is opened and aright ventricular papillary muscle is dissected free by first cutting a small section of the valve, to which themuscle is attached by way of the cordea, and then cutting a small section of the septal wall, to which themuscle it attached at the base. Papillary muscles are chosen based on their geometry. Typically papillarymuscles chosen for these studies are long, thin, and unbranched. Studies are not done on left ventricular (LV)papillary muscles because these muscles are thick and diffusion of oxygen and nutrients to the core of thesemuscles may be limited. Additionally, ANP and BMP induction may be effected by a decreased oxygen supplyin LV papillary muscles and hence these specimens may not have a normal stretch induced response (Chen,2005; Sabatine et al., 2004).The muscles are mounted in the cardiac tissue culture chamber, containing the same Hepes buffered solutionused for the dissection. Muscles are mounted between a basket attached to a force transducer and actuatorcontrolled micromanipulator, and a stationary titanium hook like extension. Oxygen is flow over the solutionduring the mounting procedure, to prevent hypoxia. All portions of the system, for experiments longer than 6hours in time, are steam sterilized prior to each experiment. Additionally, all solutions are filter sterilized. TheHepes buffered solution is then exchanged for a modified M199 cell culture media containing (mM): 2.0 L-carnitine, 5.0 creatine, 5.0 taurine, 2.0 L-glutamine, 0.2% albumin, 100 IU/ ml penicillin, 0.1 mg/mlstreptomyocin, 10mM Hepes, and 50 ug/ml of insulin in equilibrium with 5% CO2and 95% Oa.The temperatureof the muscle chamber is maintained at 34 C. The pH of the solution is maintained between 7.4 and 7.5. Themodified Hepes buffered solution is exchanged for modified M199 media in steps, such that calciumconcentration is slowly increased from 1.0 mM to 1.75 mM. Once the solution is replaced, the muscle isstimulated via the hook and a platinum electrode positioned within close proximity of the muscle's base.Muscles are left at slack length and stimulated at 0.2 Hz for one hour (Janssen et al., 1998). After the musclehas equilibrated, it is either stretched between 85-95% of Lmax, where Lmax is defined as the length of themuscle at which the muscle produces the greatest active force, or left at slack length for the duration of theexperiment (2 hr, 5 hr, 12 hr). Muscle lengths are acquired with an LVDT in parallel with the micromanipulator.The surface of each muscle can be marked with titanium dioxide markers, such that local muscle deformationscan be observed during the experimental procedure. Muscle dimensions are acquired by video capture,calibrated with a phantom of known dimensions placed at the focal plane of the specimen.C3.2.b. Mechanical Testing. Right ventricular papillary muscles are mounted between a stationary hook and anIsometric Harvard Apparatus force transducer (724490), which is capable of measuring force ranges between0-0.5g and has an accuracy of 1% (<1 mg). The Harvard Apparatus force transducer is attached to aNewport 460P series, high precision, linear modular ball bearing stage, whose position in the x/y/z direction iscontrolled by a micromanipulator, which is connected to a Newport motorized actuator (CM-12CC), having aresolution less than 100 nm, an incremental motion less than 500 nm, and a speed range between 50-500 218 Knowlton, Kirk U.Urn/sec. In order to measure muscle lengths and distances by which the muscle is stretched with a precisionon the order of 1 |xm, a LVDT (Omega LD310-10) having an accuracy of 400 nm and a linear range of 20 mm,is mounted in parallel with the actuator.Using this experimental setup, uniaxial muscle forces are recorded while pacing the muscle at 1 Hz and duringcontinuous stretching of the muscle up to Lmax. Muscle lengths are simultaneously acquired and localdeformations of the muscle are imaged by a CCD camera (COHU Inc.). Digital video recordings of muscledeformation are synchronized with acquired force data. Lagrangian uniaxial strain measurements (E=1/2(A2-1),where Ais the stretch ratio) are calculated with respect to the slack length of each muscle. Lagrangian stressesare calculated by dividing acquired force data by the initial cross sectional area of each muscle.C3.2.C. Real-Time PCR Procedures. At the end of each muscle experiment, the muscles are immediatelysubmerged in RNA later solution, which functions to stabilize mRNA in intact tissue samples and containsRNase inactivating reagents (Quiagen). Right ventricular papillary muscles are completely isolated from theirattachments, the valve and septal wall. Then the muscles are homogenized and RNA is extracted from thesetissues, using Quiagen's RNeasy Micro RNA purification kit. Given that total RNA extraction from these smallspecimens is low (-10 ng/|jJ) and that DNase treatment has been shown to degrade and or destroy nucleicacid extracts, DNase treatment is not applied to RV papillary muscle specimens (Jemiolo and Trappe, 2004).Reverse transcription (RT) is accomplished using the Invitrogen Super Script III cDNA synthesis kit, which isoptimized for low inputs of RNA. Finally, quantitative PCR is performed using an Applied Biosystems ABI 7700real time thermal cycler, AB Taqman Universal Master Mix with UNG, with AB pre-made primers and taqmanprobes for ANP, BMP, and GAPDH.ANP and BNP gene expression is normally induced by increased mechanical loads in isolated cardiacmyocytes as well as intact cardiac tissue (Ruwhof et al., 2000). GAPDH housekeeping gene expression doesnot change in these experiments. For each amount of RNA tested, duplicate Ct values are obtained andaveraged. Then the AACt method is applied to the averaged Ct values obtained, in order to quantify the foldchange of ANP or BNP relative to the unstretched control samples. Within each gene expression assay,appropriate controls are also included. No RT controls are performed for each amount of RNA, testing for anygenomic DNA contamination. Also negative controls, where no cDNA is added to the reaction, are performedin order to test the purity of the enzyme master mix as well as the primer and probe mixture. Data arepresented as mean 1SEM. Because mRNA concentrations are log-normally distributed and requirenonparametric statistics, a two group t-test is performed on the ACt values or the raw output data obtainedfrom the thermal cycler, between stretched and unstretched experimental groups. Significance is set at the P <0.05 level, with appropriate corrections for multiple comparisons.C3.2.d. Antibody Staining and Confocal Microscopy. Papillary muscles that are not used for gene expressionanalysis are stretched to some known % stretch and fixed in 2% PFA (para-formaldehyde) diluted in 50% PBS(phosphate-buffered saline) and 50% modified Hepes buffered cardioplegic solution. After 10-20 minutes, thesolution is replaced with 4% PFA in 100% PBS. Muscles are allowed to fix for an additional 10-20 minutes andthen transferred into Tissue-Tek O.C.T. compound. They are frozen on dry ice and stored at -80 C. Using acryostat, tissue samples are then sectioned (10-15 u,m sections). For ANP protein analysis, tissue cross-sections are prepared, while for titin epitope, staining longitudinal sections are prepared. After a series ofblocking and washing steps, tissue sections are then stained with appropriate antibodies (e.g. IgG antimouseANP antibodies are used to look at ANP protein expression in stretched and unstretched specimens, IgGmonoclonal anti-mouse a-actinin (Sigma) antibodies stain for sarcomeric z-disks, monoclonal antimouse IgM9D10 antibodies stain for the PEVK region on titin (Hybridoma Bank), and polyclonal anti-goat IgG telethoninantibodies stain for Tcap (Santa Cruz Biotechnology), a protein adjoined to the N-terminal end of titin).Appropriate secondary antibodies conjugated to either Alexa 568 or Alexa 488 were also used. For titin epitopestaining procedures, the tissues were double stained with either a-actinin and Tcap or a-actinin and 9D10. ABiorad confocal microscope is used to image the sections with filter sets appropriate for the excitation of Alexa-568 and Alexa-488 simultaneously. Distances between titin epitopes and the z-disk were determined usingFFT analysis as we reported earlier for skeletal muscle fibers (Shah et al., 2004).219Knowlton, Kirk U.C3.3. Isolated Perfused Whole Mouse Heart PreparationsC3.3.a. Isolated Mouse Heart Preparations. Mice are administered an intraperitoneal injection of heparin(100USP Units), anesthetized with isoflurane and sacrificed by cervical dislocation. The heart is rapidly excised,washed in cold cardioplegic solution, and the aorta cannulated on a 20-gauge stainless steel cannula. Theremaining tissue surrounding the aorta is tied closed using 5-0 suture to seal all vasculature to the heart. Theheart is mounted on a retrograde Langendorff perfusion apparatus (Radnoti). The coronaries are perfused withan oxygenated modified Krebs-Henseleit solution: 24.9 mM NaHCOs, 1.2 mM KHaPO^ 11.1 mM dextrose, 1.2mM MgSO4, 4.7 mM KCI,118 mM NaCI, and 2.55mM CaCI2 (Grupp et al., 1993). The perfusate is bubbledwith 95% O2-5% CO2 gas, maintained at 35-37 C, and a flow rate of 1-2 ml/min is achieved by applying aconstant pressure of 70 mmHg. The coronaries are cleared and the heart is allowed to beat freely andequilibrate for 15 minutes. Surface EGG electrodes are set in place and recordings are taken throughout theexperiment. 95% 02 / 5% COj The heart is paced with a digital stimulator (DS8000, World Precision Instruments). It is paced epicardially from either Retrograde Perfuston the atria or ventricle with a bipolar platinum electrode at a Reservoir constant current two times threshold. For isovolumic Langendorff preparations, a fluid filled balloon is inserted intoAntarograde the left ventricle through the mitral valve. The balloon, PerfusionReservoir consisting of plastic wrap at the end of a 20-gauge blunt needle, is connected to a syringe infusion pump and inline with a fluid-filled pressure transducer. The balloon is inflated Compliance Chamber and deflated to an EDP of 30 mmHg to precondition the Moan Aortic P Measurement tissue. For the working heart preparation, the heart contracts Outflow Resistance against a resistive-compliant network consisting of a chamber f. Cardiac Output with an air bubble and a flexible tube with an adjustable clamp (Fig. 14).The heart is kept immersed in warm (37 C) Collection Reservoir KHS throughout the experiment except when marker deformation is being videotaped. The pressure in the aortic Right Atrtal outflow line is monitored using a transducer (Millar Pacing Loads Instruments, Houston, TX).Video Camera Figure 14. Schematic of isolated ejecting heart apparatus in the VCR Core A lab. (From Karlon ef a/. (Karlon et al.,2000))C3.3.b. Non-Homogeneous Epicardial Strain Analysis. Non-homogeneous epicardial strain distributions aremapped in the isolated mouse heart using high-resolution optical imaging of 50-100 small titanium dioxidemarkers arrayed on the epicardium with a short-bristled brush. In addition to LV strains, we have also usedthese methods in our lab to map regional distributions of septal strains in isolated working and non-workingmouse hearts (Karlon et al.,2000) and to reconstruct right ventricular strains during systole (Lorenzen-Schmidtet al.,2005). A bipolar stimulating electrode will be used to pace the heart from the right atrium and a high-fidelity Millar pressure-volume catheter will be used to measure ventricular volume and pressure. A high-resolution CCD camera will be used to image the epicardium during isovolumic or ejecting contractions at fullrange of end-diastolic volumes. Markers will be tracked to sub-pixel accuracy throughout the cardiac cycleusing the object tracking tools of the MetaMorph (Universal Imaging) software package. The markers aremapped to a bicubic Hermite finite element mesh and a least-squares fit of marker displacements is used tocompute two-dimensional non-homogeneous strain fields using software developed in Dr. McCulloch'slaboratory (Karlon et al.,2000; Mazhari et al., 1998). Briefly, the epicardial surface is modeled with 6-8 bi-cubicHermite prolate spheroidal finite elements, and two-dimensional reference marker coordinates measured fromthe video images are projected on to the surface. The deformed coordinates are then projected on to the samematerial points in the models, which are then refitted by least squares resulting in a parametrically deformedmesh from which regional strains can be interpolated at each frame. This approach has also been used inconjunction with optical mapping of electrical activity in the isolated perfused rabbit heart (Sung et al.,2003). Inaddition to computing isovolumic strains (which are substantial), it is possible to obtain strains representative ofejection phase wall motions in the non-working heart, by using a reference state at an end-diastolic frame atmid-high ventricular volume and a deformed state at an end-systolic frames at lower ventricular volume. These220 Knowlton, Kirk U.techniques can all be applied to the right ventricle as well as the left, which will be important for Project 2(Chen).C3.3.C. Optical Mapping of Action Potential Propagation. A 0.8 ml bolus of the voltage-sensitive dye di-4-ANEPPS (25 uM) is injected into the perfusion line. The ventricular epicardium is illuminated by two 470 nmwavelength, 3.6-W 40-LED cluster bulbs (Ledtronics, Inc.). Fluorescence emission is passed through a >610nm high-pass filter, focused with a fast 50 mm lens (1:0.95, Navitar) and recorded by a 12-bit charge-coupled-device (CCD) camera (CA-D1-0128T, Dalsa). Images of a 9x9 mm2 field of view are acquired at 950 framesper second and a spatial resolution of 64x64 pixels.The time series of fluorescent images provides temporal intensity signals at each pixel. Each signal isnormalized with respect to baseline intensity (AF/F) and inverted to better represent a cardiac action potential.Spatial phase-shift and temporal median filtering (Sung, 2001) are implemented to reduce noise in the signal.Activation time is determined as the time of maximum slope (dF/dt) during the action potential upstroke. Thisprovides an activation map over the ventricular epicardium showing the rate and path of action potentialpropagation. Spatial gradients of activation time are used to calculate local apparent conduction velocity vectorat each pixel (Bayly et al., 1998). Additionally, the curvature of the activation wavefront is calculated as thespatial gradient of the normalized conduction velocity vector field. Time of repolarization is calculated byidentifying the peak signal and determining when the action potential has recovered 20%, 50%, and 80% backto baseline. Action potential duration (APD) is then defined as the time difference between repolarization andactivation. APD dispersion is determined as the standard deviation of APD over the surface of the heart.Attenuation of motion during optical mapping studies is necessary for the analysis of repolarization. ModifiedKrebs-Henseleit solution with 1 mM CaCb and 15 mM of the electromechanical uncoupler 2,3-butanedionemonoxime (BDM) is perfused during data acquisition to eliminate motion artifact. However, it has beendemonstrated that BDM prolongs action potential duration and slows conduction velocity in the isolated murineheart (Baker et al., 2004). We have recently developed a combination of ratiometry and a motion trackingalgorithm to eliminate motion artifact without the use of chemical uncouplers. Since the emission spectrum ofthe potentiometric dye is wavelength shifted with altered membrane potential, upright and inverted opticalaction potentials can be recorded at green (560 nm) and red (620 nm) wavelengths respectively. A wavelengthsplitting optical setup allows us to record fluorescence emission at low and high wavelengths simultaneouslyusing a single CCD camera, a technique originally applied to ratiometric microscopy (Kinosita et al., 1991).Image registration and ratio calculation of the two wavelengths amplifies the magnitude of the recorded opticalaction potential and corrects for differences in pixel intensity values due to heterogeneous dye loading oruneven excitation lighting. To correct for myocardial motion during contraction, a Lucas-Kanade optical flowalgorithm is used to compute a vector field of frame-to-frame pixel displacements that is used to perform a non-rigid sub-pixel resolution transformation of deformed frames back to the reference configuration. Combined,these techniques effectively correct for tissue displacement and local motion artifact in the optical recordings.C3.3.d. Programmed Stimulation. Restitution kinetics and vulnerability to arrhythmia are assessed in theisolated mouse heart by programmed S1-S2 stimulation protocol. The isolated heart is paced from the leftventricular epicardium via a bipolar platinum electrode with a digital stimulator (World Precision Instruments,DS8000). The ventricle is paced at a basic cycle length of 200ms (S1-S1) for >20 beats followed by apremature stimulus (S2) (Baker et al., 2000; Lerner et al., 2000). The S1-S2 cycle length is decreased until theeffective refractory period is reached, i.e. when the S2 stimulus no longer induces an action potential. Duringthis protocol, optical mapping is used to analyze action potential morphology (amplitude and duration) andconduction velocity of the S2 induced action potential (Knollmann et al., 2006). The relationship between thesemeasurements and the S1-S2 diastolic interval can be used to assess the restitution kinetics of the mouseheart. Premature action potentials during short S1-S2 cycle lengths can induce arrhythmias in susceptiblemedia. The arrhythmia vulnerability of transgenic hearts can be assessed by the incidence of ventriculartachycardias during this protocol (Lerner et al., 2000). Additionally, burst pacing or a train of several high-frequency S2 stimuli can induce arrhythmias in the isolated heart (Fig. 15).C3.3.e. Electrical Space and Time Constants. The effective space and time constants directly aggregatemyocardial electrical resistance and capacitance and are measured in the isolated mouse heart using similarmethods to those of Poelzing et al. (Poelzing et al., 2005). A 125-micron diameter Teflon-coated platinum221 Knowlton, Kirk U.unipolar electrode was used to pace the mid-lateral LV free wall. After steady-state pacing from the electrode(100 paced beats), a 1-mA cathodal stimulus was delivered for 250 ms, 35 ms after the final drive train pacingstimulus, during the refractory period of the previous action potential, resulting in a cathode-break stimulusupon release of the 250 ms pulse. Signals are normalized by baseline action potential amplitude so that thetissue stimulus response could be compared across space. The common mode signal (mean signal distal fromthe stimulus
Showing the most recent 10 out of 144 publications
