Anal. Chem. 2010, 82, 7581–7587
Microfluidic Cardiac Cell Culture Model (µCCCM) Guruprasad A. Giridharan,† Mai-Dung Nguyen,† Rosendo Estrada,† Vahidreza Parichehreh,† Tariq Hamid,‡ Mohamed Ameen Ismahil,‡ Sumanth D. Prabhu,‡ and Palaniappan Sethu*,† Department of Bioengineering, Speed School of Engineering, University of Louisville, Louisville, Kentucky 40208, and Institute of Molecular Cardiology, School of Medicine, University of Louisville, Louisville, Kentucky 40202 Physiological heart development and cardiac function rely on the response of cardiac cells to mechanical stress during hemodynamic loading and unloading. These stresses, especially if sustained, can induce changes in cell structure, contractile function, and gene expression. Current cell culture techniques commonly fail to adequately replicate physical loading observed in the native heart. Therefore, there is a need for physiologically relevant in vitro models that recreate mechanical loading conditions seen in both normal and pathological conditions. To fulfill this need, we have developed a microfluidic cardiac cell culture model (µCCCM) that for the first time allows in vitro hemodynamic stimulation of cardiomyocytes by directly coupling cell structure and function with fluid induced loading. Cells are cultured in a small (1 cm diameter) cell culture chamber on a thin flexible silicone membrane. Integrating the cell culture chamber with a pump, collapsible pulsatile valve and an adjustable resistance element (hemostatic valve) in series allow replication of various loading conditions experienced in the heart. This paper details the design, modeling, fabrication and characterization of fluid flow, pressure and stretch generated at various frequencies to mimic hemodynamic conditions associated with the normal and failing heart. Proof-of-concept studies demonstrate successful culture of an embryonic cardiomyoblast line (H9c2 cells) and establishment of an in vivo like phenotype within this system. Cardiovascular disease (CVD) is the leading cause of death in the United States and claims more lives each year than the next four leading causes of death combined.1 Understanding the molecular basis of manifestations of CVD such as myocardial infarction, ischemia, hypertension, cardiomyopathy, heart failure, etc., requires multiscale, multilevel approaches. Analysis of isolated cardiac cells including myocytes, smooth muscle cells, endothelial cells and fibroblasts, free from connective tissue and contaminating cell populations enables assessment of subcellular mechanisms and signaling processes in great detail that is typically not possible using intact tissue. However, replicating the mechanical loading environment of intact cells is a challenging task fraught with * To whome correspondence should be addressed. Phone: (502) 852 0351. Email:
[email protected]. † Speed School of Engineering. ‡ School of Medicine. (1) Thom, T. Circulation 2006, 113, 85-151, doi: 10.1161/circulationaha. 105.171600. 10.1021/ac1012893 2010 American Chemical Society Published on Web 08/26/2010
intrinsic difficulties, as available in vitro techniques fail to adequately mimic the in vivo environment. The myocardium experiences passive stretching and pressure build-up during filling (preload) and actively generates mechanical force during ejection (contraction) against variable afterload. Cells in cardiac tissue therefore are under constant physical stimulation and rely on conversion of these cues into intracellular signals to control cell phenotype and muscle mass during conditions such as hypertrophy. Despite advances in isolation and culture techniques, the current state of cardiovascular in vitro models is predominantly based on glass slides or tissue culture dishes under static conditions.2 Most studies utilize isolated cells maintained in twodimensional culture forming randomly oriented cell - extra-cellular matrix (ECM) attachments. Accurate replication of the in vivo environment requires consideration of several factors. First, tissue orientation and direction of stretch are critical determinants of cardiac cell signaling, therefore randomly oriented cultures cannot replicate physiological loading. Second, fluid flow is an important determinant as shear stress strongly influences endothelial and smooth muscle phenotype and hence perfused systems are more representative of the in vivo environment. Third, cyclic build-up of pressure and stretch during preloading are critical determinants that dictate cell structure and function. Fourth, synchronous contractions that enable unloading, ensures ejection of fluid that in turn results in relaxation of built-up pressure and stretch. Some of the aforementioned issues have been previously addressed, albeit individually. Patterning of ECM proteins using techniques like microcontact printing and lithography to define areas of cell attachment and directional alignment have been accomplished.3 Several groups have reported culture of cardiac cells under conditions of perfusion using continuous and cyclic loading to mimic in vivo hemodynamic loading.4 Mechanical stretch simulating hemodynamic loads are studied using technologies like Flexcentral Flexercell Strain - FX-4000 Flexercell Tension Plus (http://www.flexcellint.com) which consists of a flexible membrane on which cardiac cells can be plated and then be subject to oscillatory deflections using a vacuum or pressure source. Electrophysiological studies have been accomplished5-7 (2) Samarel, A. M. Am J Physiol Heart Circ Physiol 2005, 289, H2291-2301, doi: 10.1152/ajpheart.00749.2005. (3) Bursac, N.; Parker, K. K.; Iravanian, S.; Tung, L. Circ. Res. 2002, 91, 4554, doi: 10.1161/01.res.0000047530.88338.eb. (4) Gupta, V.; Grande-Allen, K. J. Cardiovasc. Res. 2006, 72, 375-383, doi: 10.1016/j.cardiores.2006.08.017. (5) Klauke, N.; Smith, G.; Cooper, J. M. Anal. Chem. 2010, 82, 585-592, doi: 10.1021/ac901886j. (6) Klauke, N.; Smith, G. L.; Cooper, J. Biophys. J. 2006, 91, 2543-2551, doi: 10.1529/biophysj.106.085183.
Analytical Chemistry, Vol. 82, No. 18, September 15, 2010
7581
and such studies typically utilize systems like IonOptix Myocyte Systems (http://www.ionoptix.com) where cardiomyocytes grown on a glass slide are stimulated with an electrical impulse and analyzed for contractility via fluorescence using hardware and software developed by the manufacturer. Microfluidics based systems have also been used to create cardiomyocyte actuators and pumps as well as cell culture models.8-14 However, to the best of our knowledge the µCCCM described herein is the first system that recreates all aspects of cardiac mechanical loading (pulsatile stretch and pressure) with an integrated circulation network. The µCCCM, fabricated using standard soft-lithography techniques consists of a small (1 cm diameter) cell culture chamber on a thin silicone membrane. Continuous circulation of culture medium is maintained using a peristaltic pump. Downstream of the chamber is a collapsible pulsatile valve actuated in a pulsatile fashion using a programmable pressure generator. Closure of this collapsible pulsatile valve leads to pressure build up in the chamber which in turn leads to stretching of the thin membrane on which cells are cultured mimicking diastolic preloading in the heart. To ensure uniformity, a post is placed beneath the thin membrane such that during stretch, the portion of the membrane on which cells are cultured experiences uniform strain and the edges experience larger nonuniform strain. Further to influence outflow resistance, a passive tunable resistance hemostatic valve is placed downstream of the collapsible pulsatile valve to mimic afterload. This system is programmable and can accommodate a wide range and different combinations of operating parameters. Fluid transport and shear stress can be controlled by setting the flow rates of the pump. Pressure buildup and strain can be controlled via a combination of factors including fluid flow rate and operation of both valves. Strain is also a function of the thickness of the membrane on which cells are cultured. The µCCCM is unique in the sense that it mimics the native heart where changes in one or more variables have a cascading effect on the entire system. For example, increasing the outflow resistance by manipulating the hemostatic valve indicates a system with high afterload. This in turn results in an increase in base pressure and baseline strain within the cell culture area mimicking conditions experienced during aortic stenosis or hypertension. MATERIALS AND METHODS Cell Culture Platform Fabrication. The cell culture chamber was fabricated using a two-step process. First, a thin membrane of PDMS (Dow Corning, Midland, MI) was fabricated by mixing the prepolymer with the cross-linking agent in a ratio of 10:1 and (7) Klauke, N.; Smith, G. L.; Cooper, J. M. IEEE Trans. Biomed. Eng. 2005, 52, 531-538 , doi: 10.1109/tbme.2004.842971. (8) Camelliti, P.; Gallagher, J. O.; Kohl, P., McCulloch; A. D. Nat. Protoc. 2006, 1, 1379-1391, doi: 10.1038/nprot.2006.203. (9) Cheng, W.; Klauke, N.; Smith, G.; Cooper, J. M. Electrophoresis 2010, 31, 1405-1413, doi: 10.1002/elps.200900579. (10) Klauke, N.; Smith, G.; Cooper, J. M. Lab Chip 2007, 7, 731-739, doi: 10.1039/b706175g. (11) Klauke, N.; Smith, G.; Cooper, J. M. Anal. Chem. 2009, 81, 6390-6398, doi: 10.1021/ac9008429. (12) Klauke, N.; Smith, G. L.; Cooper, J. M. Anal. Chem. 2007, 79, 1205-1212, doi: 10.1021/ac061547k. (13) Li, X. J.; Li, P. C. H. Anal. Chem. 2005, 77, 4315-4322, doi: 10.1021/ ac048240a. (14) Tanaka, Y. Lab Chip 2007, 7, 207-212, doi: 10.1039/b612082b.
7582
Analytical Chemistry, Vol. 82, No. 18, September 15, 2010
spinning this mixture on a silicon wafer at speeds ranging from 200 to 2000 rpm on a spin-coater (Laurel Technologies, North Wales, PA) to obtain membranes of different thicknesses. Following spinning the silicon wafer was transferred to an oven (Fisher Isotemp, Florence, KY) and the PDMS was allowed to cure for 3 h at 70 °C. Separately, the PDMS prepolymer and crosslinking agent mixture were molded into a 3 mm thick layer in a Petri-dish and cured in the oven for 3 h at 70 °C. Once the crosslinking was complete a 7.5 mm ×2 cm piece of PDMS was cut out from the 3 mm thick layer and a 1 cm diameter hole was punched using a cork borer and bonded irreversibly to the thin PDMS layer on the silicon wafer following exposure to oxygen plasma in a reactive ion etcher (March Instruments, Concord, CA). This formed the cell culture chamber. Another 7.5 × 2 cm piece was also cut out and a 1 cm diameter hole was cut from the piece. The 1 cm diameter piece was taken and the diameter was reduced to 7 mm using a different cork borer to produce the post for uniform strain. The 7 mm diameter post was assembled concentrically within the punched hole and the cell culture chamber was assembled on top of this assembly. This arrangement was sandwiched between two polycarbonate plates. The top polycarbonate piece contains an inlet and outlet channel micro machined using an end-mill cutter and contains connections for inlet and outlet tubing. The two plates are clamped using four screws. Valve Fabrication. The pulsatile actuated valve was fabricated by enclosing a 3 cm long, 5 mm diameter ARGYLE Penrose Tubing which is made of latex with a wall thickness of 500 µm inside of a polypropylene T-junction which is 3 cm long, 2 cm tall, and has 1 cm diameter openings. The Penrose tube enters and exits along the top of the “T” and the outlet at the bottom is connected to a programmable pneumatic driver (LB Engineering, Germany). The inlet and outlet where the Penrose tube is inserted are sealed to ensure pressure build up within the T-junction and cause collapse of the latex tube. The volume of the Penrose tube within the T-junction was determined based on the volume of the cell culture chamber that needs to be evacuated to ensure return to baseline stretch and pressure. The valve was pneumatically actuated using a programmable pneumatic driver. The passive tunable resistance hemostatic valve (Part No. 11494, Qosina Company, Edgewood, NY) was used with a 2.79 mm diameter latex tubing. Cell Culture. H9c2 cells (ATCC, Manassas, VA), an embryonic cardiomyoblast line, were used for all experiments. These cells were initially maintained in culture using high glucose Dulbeco’s Modified Eagle’s Medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin. Prior to seeding cells within the device, the devices were sterilized and fitted with a stencil to ensure seeding was limited to a circular area (7 mm in diameter). The devices were treated with 50 mg/ mL of fibronectin for 24 h at 37 °C to promote cell adhesion. Following washing with sterile 1× phosphate buffered saline (PBS), H9c2 cells were seeded at a density of 5 × 105 cells/mL. Following seeding, the cells were allowed to attach and spread for 4 h. Following this, the medium was replaced with fresh medium and cells were cultured for 24 h. The stencil was removed and cells were either cultured under static conditions (control) or perfusion and pulsatile stretch for 24 h. Continuous circulation of the cell culture medium (flow rate: 8 mL/min)
was achieved with the use of a peristaltic pump (Ismatec, ColeParmer, Chicago, IL). Pressure Characterization. The device was assembled and the pressure build-up within the chamber was characterized for the following clinically relevant experimental conditions: (1) normal, (2) heart failure, (3) hypertension, (4) hypotension, (5) tachycardia, and (6) bradycardia. To accomplish pressure monitoring within the chamber, a small hole was punched on the side of the cell culture chamber using a 24 gauge syringe needle to access the interior of the chamber. An 18 gauge needle was then inserted into the hole to form a tight seal and attached to a passive tunable resistance hemostatic valve using luer connectors. A highfidelity micro pressure tip pressure catheter (Millar Instruments, Boston, MA) was inserted through a hemostatic valve and syringe needle such that the pressure sensing element resided in the center of the chamber. The chamber pressure data were signal conditioned and analog-to-digitally converted at a sampling rate of 400 Hz in 30 s recording epochs and stored for digital analysis with a clinically approved Good Laboratory Practicescompliant data acquisition system.15,16 The transducers were pre- and postcalibrated against known standards to ensure measurement accuracy. The chamber pressure waveforms were analyzed by using a hemodynamic evaluation and assessment research tool program17 developed in Matlab (MathWorks, Natick, Mass). Strain Characterization. Strain developed for different membrane thickness was experimentally evaluated. The membrane thickness was measured with the use of a Dektak 8 Surface Profiler (Veeco, Plainview, NY). The membrane for these experiments was modified by embedding 6 µm diameter beads (Duke Scientific, Palo Alto, CA). The setup was similar to the pressure characterization except a digital pressure gauge replaced the pressure catheter. Also, the outlet was sealed. The entire setup was placed under a microscope and an image was captured using a CCD camera. Using a syringe (Figure 2a), fluid was pumped into the chamber to generate a fixed pressure which resulted in stretching of the membrane and induce a change in the relative position of the beads. Another image was captured using the CCD camera and the relative displacement of the beads under a particular pressure was calculated by comparing the image to the control image using Metamorph Software and the strain and % strain values for different pressures and membrane thicknesses were computed. Shear Stress Simulation. Finite element software ANSYS Academic Teaching Advanced 12.0, FLOTRON module, was used to predict the wall shears stress as a function of channel height and flow rate. The system modeled using a 2D simulation and environment was verified for steady incompressible flows. The physical properties of water were applied to the fluids participating in the simulation (density (F) ) 1000 kg/m3 and dynamic viscosity (µ) ) 10 Pa*s). The inlet fluid flow rates of 8.8 mL/ min, was specified at the input, whereas the outlet was set to a fixed-pressure boundary condition and zero pressure was (15) Drew, G. A., Koenig, S. C. In Virtual Bio-Instrumentation: Biomedical, Clinical, And Healthcare Applications in Labview; Olansen, J. B., Rosow, E., Eds.; Prentice Hall: New York, 2002; Vol. 180-6. (16) Koenig, S. C.; Woolard, C.; Drew, G. D.; Unger, L.; Gillars, K. J.; Ewert, D. L.; Gray, L. A.; Pantalos, G. M. Biomed. Instrum. Technol. 2004, 38, 229–240. (17) Schroeder, M. J.; Perrault, B.; Ewert, D. L.; Koenig, S. C. Comput. Biol. Med. 2004, 34, 371–388.
Figure 1. Images of various components of the µCCCM. (a) Schematic of the setup with the cell culture chamber and oil chamber, (b) assembled platform, (c) pulsatile collapsible valve, and (d) experimental setup.
applied to the outlet. No-slip boundary conditions were applied for the channel and groove walls. The fluid domain was meshed using 2D quadrilateral element, FLUID141, to model steady state fluid. A segregated sequential solver algorithm was used; that is, the matrix system derived from the finite element discretization of the governing equation for each degree of freedom was solved separately. The mesh was refined through mesh sensitivity analyses: at each simulation, the elements showing high velocity gradients were refined, until reaching convergence of sensitive measures of the predicted quantities. Immunofluorescence and Microscopy. H9c2 were fixed with 4% paraformaldehyde in 1× PBS for 20 min, washed two times with wash buffer (1× PBS containing 0.05% Tween-20 (Fisher Scientific, Fair Lawn, NJ)) and permeabilized with 0.1% Triton X-100 (Fisher Scientific, Fair Lawn, NJ) for 5 min at room temperature. For the detection of F-actin, cells were washed two times with wash buffer, blocked with 1% BSA in 1× PBS for 30 min and incubated at room temperature for 1 h with TRITCconjugated phalloidin (1:100; Millipore, Billerica, MA). Light Diagnostics mounting fluid (Millipore, Billerica, MA) was added to the cells and cells were examined using a Nikon Eclipse TE2000-U epifluorescence microscope.
RESULTS µCCCM Components and Setup. The µCCCM consists of four components: pump, cell culture chamber, collapsible pulsatile valve and the hemostatic valve (Figure 1a) connected in series to establish a circulation loop. The collapsible pulsatile valve (Figure 1b) is responsible for ensuring cyclic changes in pressure and strain within the system. Figure 1c shows the cell culture chamber with a thin membrane at the bottom for cell culture. This is placed directly over the supporting layer containing a post to ensure uniform stretch (Figure 1c) and assembled together with fluid flow channels between polycarbonate plates to ensure leak free perfusion (Figure 1d). The hemostatic valve is downstream of the collapsible pulsatile valve and is essential to vary afterload. Analytical Chemistry, Vol. 82, No. 18, September 15, 2010
7583
Figure 2. (a) Experimental setup used for estimation of strain values at different pressures and (b) plot of % strain vs pressure for different membrane thicknesses. Table 1. Heart Rate, Systolic Fraction, Peak and End Diastolic Pressure values obtained using the µCCCM during simulation of different Physiologic Conditions
Figure 3. (a) Shown is the setup used for obtaining pressure measurements directly in the cell culture chamber. (b) A high fidelity micro pressure tip catheter was inserted inside the cell culture chamber and measurements were directly recorded.
Strain Characterization. Evaluation of strain was accomplished using the setup shown in Figure 2a. Membranes of different thicknesses (93, 139, 193, 329 µm) were evaluated for % strain (stretch) at different chamber pressures (60-140 mm of Hg). Figure 2b graphically represents pressure vs % strain for each thickness. The 139 µm thick membrane resulted in ∼20% strain for normal physiological peak pressures of 120 mm of Hg whereas the 93 µm thick membrane was capable of strains up to 60%. Pressure Characterization. The setup for pressure characterization is shown in Figure 3. Using this setup the data was continuously recorded for various conditions. The µCCCM produced a peak pressure of 123 mmHg, an end diastolic pressure of 10 mmHg, at a rate of 75 bpm, and a 40% systolic fraction. Using this model, the user can also modify different variables (operating conditions) to mimic cardiac dysfunction. Simulated heart failure had a lower peak pressure (95 mmHg) and a significantly higher end diastolic pressure (27 mmHg). Hyper- and hypotension test conditions produced normal end diastolic pressures (8-10 mmHg). Hypertension test condition had a significantly elevated peak pressure (183 mmHg) while hypotension had a significantly lower peak pressure (92 mmHg) compared to normal test condition. Simulated tachycardia had a significantly higher beat rate (200 bpm), lower end diastolic value (3 mmHg), and a higher systolic fraction (55%) compared to normal test condition. Simulated bradycardia had a lower beat rate (46 bpm), and slightly lower systolic fraction (38%) compared to normal test condition. The obtained µCCCM values and waveforms (Table 1, Figure 4) closely correlate to human left ventricular waveforms found in literature for all experimental test conditions.18-21 Shear Stress Modeling. To evaluate shear stress within the system, CFD modeling was performed. The depth of the well was varied (0.25, 2, 4, 8 mm) while the inlet flow rate was kept constant 7584
Analytical Chemistry, Vol. 82, No. 18, September 15, 2010
condition
heart rate (bpm)
systolic fraction (%)
LVP peak pressure (mmHg)
LVP end diastolic pressure (mmHg)
normal heart failure hypertension hypotension tachycardia bradycardia
75 75 75 75 200 46
40 40 40 40 55 38
123 95 183 92 112 115
10 27 8 11 3 10
∼0.024 m/s. It can be seen from the simulations (Figure 5) that for the 2, 4, and 8 mm depths, fluid flow is not significant and shear stress is ∼0 from the floor of the wells until ∼200 µm above the well (region where cells are cultured). However, for the 0.25 mm depth shear stress is significant (∼10 dyn/cm2) at 200 µm from the wall and below, in the region where cells are cultured within the device. An additional variable which has a major effect on shear stress but was not evaluated due to cascading effects on pressure and stretch is the inlet fluid flow velocity. Cell Culture within the µCCCM. H9C2 cells were cultured in static controls and within the µCCCM under pressure and stretch representing normal physiological loading. Cell cultured in static controls attain a fibroblast like morphology (Figure 6a) whereas cells cultured in the µCCCM attained a more rectangular, postage stamp -like morphology (Figure 6b). Staining with phalloidin showed random orientation of F-actin in controls in comparison to aligned F-actin in cells cultured within the µCCCM (Figure 6c and d). DISCUSSION Understanding the molecular basis of health and disease in cardiac tissue requires multiscale, multilevel approaches. Cellularlevel in vitro studies using isolated cell populations to study specific signaling pathways and molecular mediators are an (18) Koenig, S. C.; Giridharan, G. A.; Ewert, D. L.; Schroeder, M. J.; Ionan, C.; Slaughter, M. S.; Sobieski, M. A.; Pantalos, G. M.; Dowling, R. D.; Prabhu, S. D. J. Heart Lung Transplant 2008, 27, 1340–1347. (19) Travis, A. R.; Giridharan, G. A.; Pantalos, G. M.; Dowling, R. D.; Prabhu, S. D.; Slaughter, M.; Sobieski, M.; Undar, A.; Farrar, D.; Koenig, S. C. J. Thorac. Cardiovasc. Surg. 2007, 133, 517–524. (20) Guyton, A. C. HallJ. E. Textbook of Medical Physiology; Elsevier Saunders: Amsterdam, 2006. (21) Pantalos, G. M.; Koenig, S. C.; Gillars, K. J.; Giridharan, G. A.; Ewert, D. L. ASAIO J. 2004, 50, 37–46.
Figure 4. Description of terms used and simulation of various physiologic conditions using the µCCCM indicative of both normal and dysfunctional myocardium. Manipulation of different elements within the µCCCM circuit enabled recreation of various clinically relevant physiologic conditions.
important first step. This however is complicated due to the fact that cardiac tissue unlike most tissue in the body is under constant stimulation. During each heart beat, blood flows into the ventricle causing the ventricle to expand and experience a buildup in pressure and stretch. Heart muscle cells contract and eject the fluid from the chamber followed by relaxation. Cardiac endothelial cells in particular are also exposed to loading induced shear stresses whereas other cells including cardiomyocytes, smooth muscle cells and cardiac fibroblasts experience only cyclic pressure and stretch. Currently, cardiac cell culture relies on conventional cell culture techniques with limited physical stimulation and cannot replicate conditions of normal physiological loading. Further, current cell cultures are incapable of replicating different clinical conditions with varying amounts of physical loading in conditions like hypotension and hypertension.
The µCCCM was designed as an in vitro model of the left ventricle, capable of accurately replicating complex in vivo mechanical stresses associated with ventricular loading for a wide variety of clinical conditions. More importantly, the system was designed such that change in one aspect of loading has a cascading effect on other variables similar to events in vivo. This is made possible in the µCCCM by integrating four tunable components in a fluidic circuit. The pump induces fluid flow and the rate of flow can be manipulated to increase/decrease shear stress and control rate of loading. Cells are cultured on a thin membrane within the cell culture well; the thickness of the membrane plays a significant role in determining the amount of stretch for a given pressure. The collapsible pulsatile valve has several tunable variables including collapsible volume, pressure used to achieve collapse, frequency, and timing at which this pressure is applied and valve open: close ratio (diastolic: systolic). Finally, an adjustable hemostatic valve (resistance element) controls the afterload in the circuit. Again, change in one or more of these variables affect the magnitude and duration of pressure and stretch within the system. H9c2 cardiomyocytes were successfully cultured within the device. The cells not only survived physiological loads in vitro but the effects of physiological loading were clearly evident when analyzed via microscopy. Cells within the µCCCM established an in vivo postage stamp-like morphology and showed alignment of F-actin stress fibers in comparison to static controls. The µCCCM for the first time enables design of experiments where physical loads can be manipulated. Cells cultured under normal conditions can be gradually or instantaneously exposed to loads associated with cardiac dysfunction causing changes in cell structure and function. Accomplishing this at the cellular level in vitro provides the opportunity to probe in great detail the role of specific molecular mediators involved in signaling associated with manifestations of cardiovascular disease. Cells can be evaluated using microscopy directly within the µCCCM or cells or cellular contents can be extracted and evaluated for gene and protein expression. The cell culture medium can also be sampled continuously to monitor signaling through soluble factors. The µCCCM, therefore, provides an ideal platform for evaluating the effects of drugs and other adjunctive and conjunctive treatment options for recovery of cardiomyocytes following cardiac dysfunction. In certain cases the use of external support to return cardiac cells to normal physiological loads has shown to be beneficial to recovery. This scenario can also be replicated by seeding dysfunctional cells or returning cells within the µCCCM exposed to nonphysiological loads to normal loading conditions and then evaluating the structure and function of cells. Various clinical studies have supported the notion that cardiac tissue can be regenerated through transplantation of progenitor and differentiated cell populations.22-24 Although existing protocols have not achieved the goal of true regeneration, substantial physiological benefit (repair) currently can be derived from transplanting cells into the infarcted heart.25,26 Though a majority (22) Murrow, J. R.; Dhawan, S. S.; Quyyumi, A. A. 2010, 131–151. (23) Reinlib, L.; Field, L. Circulation 2000, 101, e182-187. (24) Wollert, K. C.; Drexler, H. Circ. Res. 2005, 96, 151-163, doi: 10.1161/ 01.res.0000155333.69009.63. (25) Orlic, D.; et al. Pediatric Transplantation 2003, 7, 86–88. (26) Min, J.-Y.; et al. J. Appl. Physiol. 2002, 92, 288–296.
Analytical Chemistry, Vol. 82, No. 18, September 15, 2010
7585
Figure 5. Shear Stress simulations for an inlet flow velocity of 0.024 m/s for a device with varying well depth. As can be seen from the simulations the shear stress is negligible in the region where cells are cultured for all well depths except the 0.25 mm deep well.
beating in synchrony, benefits also seem to indicate reversal in ventricular remodeling and reduction in the infarct size through physical reinforcement and paracrine signaling.27 Tracking the fate of transplanted cells and determining their true role in improving cardiac function is extremely challenging due to the heterogeneity of cells within the cardiac tissue. The µCCCM is the perfect platform to evaluate differentiation or transdifferentiation potential of various cellular populations in vitro where molecular signaling events responsible for regeneration can be discovered. It is interesting to note that various studies show that benefit can be derived from a wide variety of cardiogenic and noncardiogenic cell types including adult cardiomyocytes,28 Figure 6. Microscopy images of H9C2 cells cultured under static conditions and loading conditions (µCCCM). (a, b) phase contrast images, (c, d) staining with phalloidin to visualize intracellular F-actin.
of initial hypotheses centered on improvement in cardiac function with transplanted cells augmenting the host myocardium and 7586
Analytical Chemistry, Vol. 82, No. 18, September 15, 2010
(27) Orlic, D. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 10344-10349, doi: 10.1073/ pnas.181177898. (28) Etzion, S.; Battler, A.; Barbash, I. M.; Cagnano, C.; Zarin, P.; Granot, Y.; Kedes, L. H.; Kloner, R. A.; Leor, J. J. Mol. Cell. Cardiol. 2001, 33, 1321– 1330. (29) Taylor, D. A.; et al. Nat. Med. 1998, 4, 929–933. (30) Fujii, T.; et al. Annu. Thorac. Surg. 2003, 76, 2062–2070.
skeletal myoblasts,29 smooth muscle cells,30 fibroblasts,31 endothelial progenitors,32 mesenchymal stem cells,33 hematopoietic stem cells,34 other marrow populations,35 resident myocardial progenitors,36 and embryonic stem cells.37 The µCCCM can be used to culture various cells within the cardiac tissue including cardiomyocytes, smooth muscle cells and cardiac fibroblasts. Slight modifications to the device to increase the shear stress will make this device suitable for culturing cardiac endothelial cells. This system is also capable of coculture of two or more cardiac cell types to accomplish a more physiologically relevant cell culture model. The primary difference between cardiac cells in the heart and the µCCCM is the fact that cyclic pressures and mechanical loads are achieved by manipulating the pneumatic valve, downstream resistance, and fluid flow as opposed to cardiomyocyte contractions. Despite this limitation, the µCCCM (31) Hutcheson, K. A.; et al. Cell Transplant. 2000, 9, 359–368. (32) Kocher, A. A.; et al. Nat. Med. 2001, 7, 430–436. (33) Amado, L. C. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 11474-11479, doi: 10.1073/pnas.0504388102. (34) Orlic, D. Nature 2001, 410, 701-705, doi: http://www.nature.com/nature/ journal/v410/n6829/suppinfo/410701a0_S1.html. (35) Yoon, Y.-s.; et al. J. Clin. Invest. 2005, 115, 326–338. (36) Dawn, B.; Stein, A. B.; Urbanek, K.; Rota, M.; Whang, B.; Restaldo, R.; Torella, D.; Tang, X.-L.; Rezazadeh, A.; Kajstura, J.; Leri, A.; Hunt, G.; Varma, J.; Prabhu, S. D.; Anversa, P.; Bolli, R. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 3766-3771, doi: 10.1073/pnas.0405957102. (37) Me´nard, C.; et al. The Lancet 2005, 366, 1005–1012.
enables cardiac cell types to experience physiologic levels of pressures and loads within the native ventricle. CONCLUSION In summary, the µCCCM was developed as an in vitro model of the left ventricle. To the best of our knowledge, this is the first system that accurately replicates all aspects of physical loads and maintains a synergistic balance between pressure, stretch and frequency. Pressure, stretch and shear corresponding to normal and abnormal loading conditions have been accurately replicated. Finally, cells were cultured within this device to demonstrate proofof-concept of the ability of the µCCCM to sustain cell culture. ACKNOWLEDGMENT We thank Katie Donaldson and Thanh Huong T. Luong for help with fabrication and testing. This work was supported by the National Science Foundation under Grant No. 0814194 and via a Multidisciplinary Research Grant from the Vice President of Research at the University of Louisville. G.A.G. and M.-D.N. contributed equally to this work.
Received for review May 17, 2010. Accepted August 12, 2010. AC1012893
Analytical Chemistry, Vol. 82, No. 18, September 15, 2010
7587
