Validation Study of a Mitral Valve Organ Culture System

It has been observed that alterations in normal ECM structure and mechanotransduction play a role in heart valve diseases. Current paradigms in valve biology, cell culture and animal models, are inadequate for study of the agents that cause these alterations. A mitral valve organ culture system has been designed that stimulates mitral valves in a simulated in vivo environment for periods of 4-6 weeks. We hypothesize that a mitral valve exposed to its native mechanical stresses over this period will maintain its ECM structure and cell viability, making this a valid system for studying disease agents. The first part of this validation study involved analyzing freshly excised mitral valves and testing them for collagen, GAG, and DNA content using the hydroxyproline, uronic acid, and Hoechst DNA assays, respectively, in their anterior, posterior, and chordal regions. The discovered amounts were consistent with data from previous studies and with knowledge of regions of tensile and compressive loads in the valve. This data will be reinforced with histology and mechanical testing and later compared with an analysis of biochemical and mechanical data for valves involved in medium-term static and dynamic runs in this bioreactor. Once validated, this system is a powerful tool for the study of mitral valve disorders. Background The mitral valve is the valve that controls blood flow between the left ventricle and atrium. The valve consists of two leaflets, the anterior leaflet and the smaller posterior leaflet, which are attached to each other by commissural sections and to the wall of the heart by a ring known as the annulus. A series of cord-like tendons, called chordae tendineae, are attached close to the free edges of both leaflets and to projections of the ventricular wall, known as papillary muscles. During systole, the leaflets coapt together to stop blood flow from the ventricle back into the atrium and the chordae tendineae are pulled taut, preventing the leaflets from prolapsing into the atrium. It has been observed that alterations in ECM microstructure and cell interactions of the mitral valve are responsible for a deviation from this normal function. Valvular heart disease was responsible for about 20,000 deaths and contributed to another 40,000 in 2006. Mitral valve disorders are present in about 2% of the population and were responsible for 43,000 hospitalizations last year. In 2003, an estimated 95,000 inpatient valve procedures were performed. On average, a valve procedure costs about $118,000 and has an in-hospital death rate of 5.6%. [1] These disorders may also play a role in other cardiovascular diseases. Mitral valve dysfunction, ranging from mild to very severe, occurs in 55 to 75% of patients with congestive heart failure. Even mild mitral regurgitation drastically reduces the survival rate of CHF patients. It has been shown that mitral valve dysfunction in CHF results from changes in the ECM composition of the valve tissue that lead to altered valve mechanics [2]. A greater understanding of the pathogenesis of these disorders is necessary to develop better techniques of combating heart failure and other possibly valve-related diseases. An ex vivo system has been designed and built specifically to assess the effects of heart failure agents on valve morphology, but can still serve as a versatile tool in the overall study of valve pathogenesis. The system involves culturing whole mitral valves for periods of days to weeks under native mechanical stresses experienced in vivo. It serves a role that other paradigms of valve biology have difficulty in doing. Cell culture has easily controlled conditions and is inexpensive but is limited to the study of simple cell interactions. Animal models allow for holistic study but lack the control and flexibility an ex vivo system offers. Results of animal models can also be affected by the many other factors present in a live organism. This system incorporates all the interactions that take place within the tissue and the mechanical stresses of its environment without the many drawbacks of other research tools. The aspects of the system were designed to encourage the valve to retain its native behavior. A bladder pump, consisting of a silicon membrane that separates an air-filled chamber from a medium-filled one, is used to produce pulsatile flow. A computer-controlled regulator increases air pressure, increasing pressure in the medium and pushing fluid through a mechanical aortic valve (Carbomedics, Austin, Texas) to mimic systole. When air pressure is less than fluid pressure, the system mimics diastole and medium flows back into the ventricular chamber through the mitral valve. A physiological flow rate of 5 L/min and a pressure peak of 120 mmHg are achieved. Pressure is measured using a pressure transducer (Cole Parmer, Vernon Hills, Illinois) in the medium-side of the ventricular chamber and flow rate is measured using an ultrasonic flow meter (Transonic Systems, Ithaca, New York) on the outside of the flow loop. The medium used in the system, similar to that used in the tissue engineering of heart valves, is optimized to properly nourish the valve tissue. Even the positioning of the valve in the chamber was developed to closely mimic in vivo geometry. Changes in valve geometry due to cardiac conditions have been shown to alter valve function.[3] A sterile venting filter ensures that medium contains appropriate levels of oxygen and carbon dioxide for cell survival. Figure 2,3,4: (top left) ventricular chamber separated by silicon membrane, (top right) view of entire bioreactor, The mechanical features of this system properly simulate the mitral valve’s physiological conditions but the biochemical aspects of the system have yet to be validated. We hypothesize that physiological mechanical stresses will influence ECM and cell interactions to help maintain the native structure and composition of the mitral valves. This in turn causes the valve to behave normally and induces stresses upon itself that further maintains its structure. If in vitro cultured valve structure and behavior is similar enough to that of native valves, this system is appropriate for the study of disease agents and their effect on the mitral valve. The first part of this validation study involves dissecting freshly excised valves from porcine hearts and analyzing them for biochemical composition and cellularity. This data will be reinforced with histology and later compared to a similar analysis of valves cultured during dynamic and static runs in this system. Materials/Methods Sample Preparation: 12 hearts from porcine subjects of unknown age were acquired from a local abattoir, stored in a cool environment, and dissected within several hours. It has been shown that valves Fig. 5: (Left) Bioreactor setup and running inside incubator. (Previous page) A diagram representing various stages of air bladder pumping. Fluid flows into chamber through mitral valve during diastole and fluid flows out through a mechanical aortic valve during systole. Red arrows represent fluid flow and blue arrows represent air flow. retain their bio-mechanical properties for at least 5 days when stored at 4C. For fresh valves, hearts are dissected in a clean, but not sterile environment. For the organ culture system, hearts are dissected under a sterile hood to prevent bacterial contamination of the system. Once excised, chordae were removed, strut chordae were taken for testing, and the valves were divided into anterior and posterior leaflets. Tissue samples were stored in phosphate buffer solution at -20C. Wet weights of the samples were found. Samples were lyophilized for 16 hours and weighed again for dry weight. Dried samples were placed in centrifuge tubes and rehydrated with 1 mL of 100mM ammonium acetate. Samples were minced using fine scissors. A 100 μL aliquot of proteinase-K solution (10 mg/mL) was added to each sample. These were then left to digest at 60C for 16 hours. If not fully digested, another aliquot of proteinase-K solution was added as needed. Sample tubes were heated with a temperature of at least 70C for 30 minutes to denature the proteinase-K. Samples were then frozen and lyophilized to concentrate samples to less than 1 mL. Once concentrated, ammonium acetate was added to bring sample solution volume to 1 mL. A 0.5 mL aliquot was taken for the Hoechst DNA assay and the rest was centrifuged at 5000 RPM for 3 min for the Uronic Acid and Hydroxyproline assays. Fluorometric (Hoechst) DNA Assay: Materials: NaCl/Pi Buffer (233.76 g 4M NaCl, 7.1 g 50 mM Na2HPO4, 0.58 g 2 mM EDTA, 0.2 g 0.02% NaN3 in 1 L water) Pi Buffer (7.1 g 50 mM Na2HPO4, 0.58 g 2 mM EDTA, 0.2 g 0.02% NaN3 in 1 L water) Dye Stock Solution (Bisbenzimide (Hoechst 33258) 0.2 mg in 1 ml nanopure water) DNA Stock Solution (Calf thymus DNA, 100 μg/ml in nanopure water) Procedure: A consistent level of DNA is found in all cells so average DNA amount can be used as a measure of cellularity. First, samples were vortexed and a working dilution of 1:50 for each sample was prepared by combining 10 μL o f sample and 490 μL of Pi buffer. Final dilutions were selected so that ng/mL, based on average ng DNA per mg tissue weight, would fall within range of standards. Two different dilutions were used for each sample to check for repeatability. Sample aliquot and Pi buffer were added to 1.5 mL centrifuge tube. Samples were sonicated in water bath sonicator for 3 minutes (Pulse: 30 seconds on, 10 seconds off). DNA stock solution was diluted in Pi buffer to make 1 mL of 10 μg/mL solution. Standards were prepared from this working dilution. Bisbenzimide solution was diluted 1:100 with NaCl buffer. Diluted dye was added based on level of dilution. Samples with dye remain stable for at least 16 hours if protected from light. After being vortexed, samples and standards were stored in darkness for 30 minutes. Samples and standards were quickly aliquoted in triplicate into a black 96 well plate and read in a fluorocolorimeter with excitation set to 356 nm and emission set to 458 nm. Average DNA mass was found based on sample concentrations and dilutions. Cell count was estimated by calculating that mass as percentage of dry weight. Determination of Hydroxyproline Content: Materials: 6N HCl Chloramine-T Reagent Aldehyde/Perchloric acid solution Procedure: Hydroxyproline is an important amino acid present in collagen and can be used to measure collagen content. Centrifuged samples were hydrolyzed with 6N HCl for 4 hrs at 115C and then allowed to cool. Caps were removed and remaining HCl was boiled off at 115 C. Hydrolyzate was dissolved in dH2O to give a concentration less than 300 μg/mL OH-proline. 20 μL sample aliquots were added to duplicate glass tubes and 20 μl aliquots of each standard was added to one glass tube. All samples and standards had 250 μL of Chloramine-T reagent added, were vortexed, and allowed to sit at room temperature for 20 minutes. 250 μl of Aldehyde/Perchloric acid solution was added to all samples and standards. All tubes were vortexed in placed in 60C water bath for 15 minutes. Samples were removed from water bath, vortexed, and aliquoted, in triplicate, into a 96-well plate and read in a fluorocolorimeter at an absorbance of 558 nm. Blumenkranz Uronic Acid Assay: Materials: Sulfuric Acid Tetraborate (2.384 g Borax per 500 ml concentrated sulfuric acid stored at 4C) Hydroxyphenyl reagent ( 0.15% m-phenylphenol in 0.5% NaOH stored at -20C) Procedure: Hexuronic acid is a component of all proteoglycans and can therefore be used as measure of GAG content. Sample dilutions were chosen so that results would be within range of standards. Assuming that 1 mg wet weight of tissue contains roughly 1 μg of uronic acid, amount in samples was estimated and divided by sample solution volume. All samples were diluted to a concentration less than 100 μg/mL. Samples and standards were prepared in 5 mL glass tubes. Sulfuric acid tetraborate was added to all tubes based on required dilution. Samples and standards were added to tubes, vortexed, and heated at 100 C for 5 minutes. After being cooled to room temperature, half of standard volumes were transferred to duplicate tubes to serve as blanks. 10 μL of hydroxyphenyl reagent were added to standards and samples but not the blanks. All tubes were vortexed and transferred in triplicate to a 96-well plate. Plate was read in fluorocolorimeter at 532 nm with background subtraction of 750 nm. Blank absorbance values were subtracted from sample absorbance values. Net absorbance was used to find concentration which was multiplied by the dilution factor to give uronic acid concentration. Finally, total starting volume and tissue mass was used to determine uronic acid concentration in μg per mg tissue wet weight. Results GAG Content: 12 mitral valves were tested for uronic acid content. A portion of the samples is represented in Table 1. As expected, GAG content was found to be highest in the posterior Fig. 3, 4: (left) Spectramax M2 used for all assays, (right) water bath sonicator used for Hoechst assay leaflet and least in the chordae. Chordal GAG content was the lowest of the three sections tested, with anterior GAG content being only slightly higher. Table 1: GAG content based on uronic acid mass percentage of sample wet weight. Collagen Content: 8 valve samples were measured for collagen content. A set of these samples are represented in Table 2. Collagen content was found to be highest in the chordal samples but often lowest in the anterior leaflet samples. Variability inconsistent with these overall observations was found between samples but this variability stayed consistent between trials of the same samples. This means that valves from different specimens show unique characteristics, probably due to unique in vivo environments, or an error in procedure was made that altered the biochemical content of the samples. 0.00% 0.10% 0.20% 0.30% 0.40% 0.50% 0.60% 0.70% 0.80% 0.90% 1.00% 3 4 5 6 Average U ro ni c A ci d % o f W et W ei gh t Anterior Posterior Chordae