INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX REFERENCES

IT HAS BEEN SHOWN THAT HEART valves are living tissue containing several varieties of actively metabolizing cells (16, 17), yet little information exists regarding the magnitude of their metabolic activity or factors that affect it. In addition, the amount of oxygen needed by the tissue to sustain its function is unknown and little is understood regarding the oxygen supply routes of this tissue. With the increasing interest in the creation of tissue-engineered bioprosthetic heart valves (23, 24), such data is important in the design of a successful valve substitute.

Our previous work (30) has suggested that oxygen is supplied to aortic valve cusp tissue via two complementary pathways, in a manner similar to that seen in large arteries and veins. In those vessels, oxygen diffuses into the tissue from both the vessel lumen and also through a circulatory system within the adventitia. In contrast, for aortic valve tissue, oxygen can diffuse through both the fibrosal and ventricularis surfaces because both surfaces are exposed to arterial blood. However, when cusp thickness limits oxygen delivery to the cells by simple diffusion, our results imply that an intrinsic circulation is necessary within the central layer (the spongiosa) to compensate for unmet oxygen demands. That circulation may be the equivalent of the vasa vasorum of the aortic root.

To better appreciate the roles of these oxygenation pathways, we must first gain an understanding of the metabolic demands of the tissue. A good approximation of the overall metabolic rate and oxygen demands of a tissue can be provided by the measurement of its rate of oxygen consumption (O₂) (25). The other major factor that influences oxygen delivery is the permeability of the tissue to oxygen. The amount of oxygen that can be transported through a tissue is the product of diffusivity (DO₂) and solubility () divided by the thickness of the tissue (L), i.e., DO₂ · /L. Therefore, by determining how thick the cusp is, by how readily oxygen can diffuse in from the surfaces of the cusp, and by having some knowledge of tissue oxygen solubility, we can quantify the ability of an oxygenation route to sustain the metabolic needs of this tissue. In addition, if both O₂ and DO₂ of the tissue are known, we can also indirectly assess the role played by the intrinsic circulation of the cusps.

DO₂ and O₂ can be measured experimentally using an oxygen sensor. To measure diffusivity, the tissue is exposed to a step change in oxygen level at one surface and the transient change in oxygen at the other closed surface is measured with the oxygen sensor. DO₂ is proportional to the time constant of this oxygen transient. To measure O₂, the exposed surface of the tissue is covered with an oxygen-impermeable barrier (e.g., glass coverslip) and the fall of the oxygen level with time is measured with the sensor. Consumption is proportional to the rate of oxygen decrease. It is important to remember that such measurements of DO₂ and O₂ are temperature sensitive. Because "correction factors" for temperature can lead to poor representative values (2), such experiments should be carried out at a physiological temperature if they are to reflect the in vivo situation.

On the basis of the work by Ellsworth and Pittman (10), we designed a chamber to carry out oxygen measurements of both aortic valve cusp DO₂ and O₂ parameters at the physiological temperature of 37°C with the use of a Clark polarographic oxygen sensor. The current output of a Clark sensor is proportional to PO₂ in the medium at the surface of the sensor. We performed our experiments using porcine aortic valve cusps because these valves are similar to human valves (19) and are used as bioprosthetic replacement devices (9).

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX REFERENCES

Cusp preparation. Seven porcine hearts were freshly obtained from a local abbatoir and returned (on ice slush) to the laboratory. One cusp from each aortic valve was randomly dissected and rewarmed up to 37°C in a physiological balanced salt solution immediately before being mounted in the sample chamber for DO₂ and O₂ measurements.

The sample chamber. A sample oxygen chamber (Fig. 1) was designed for DO₂ and O₂ measurements based on the design of Ellsworth and Pittman (10). The design consisted of an acrylic chamber to house the cusp with a gas inlet and outlet. A Clark PO₂ sensor (model 733, Diamond General Development; Ann Arbor, MI) entered through a tight-fitting O-ring at the base of the chamber. The surface of the Clark sensor was positioned to be flush with the bottom surface of the chamber so that the tissue could be placed directly on top of the sensor. The Clark sensor recorded changes in tissue PO₂ throughout the experiment. The chamber was built with a tightly fitting lid so that the chamber could be easily sealed. The temperature of the chamber was measured using a thermocouple. The thermocouple controlled the gas inlet heater, which maintained the chamber and the tissue inside it at 37 ± 0.2°C. All gases were humidified before entering the chamber by being bubbled through distilled water and heated to 33°C using a water bath. This ensured that the tissue moisture level remained constant throughout the experiments and the increase in gas temperature from 33 to 37°C prevented the buildup of excess condensation within the experimental setup.

View larger version (33K): [in this window] [in a new window]   Fig. 1.   Schematic of the experimental setup. Gas from the tanks is humidified with the use of a water bath set to 33°C, flow is controlled through the flowmeter, and the gas is heated just before entry into the chamber. It enters through the lid, which is sealed during experimental measurements. The temperature of the chamber is controlled at 37 ± 0.2°C through the thermocouple in the chamber, which controls the gas heater. The Clark polarographic oxygen sensor enters into the base of the chamber through an O-ring and the tissue lies directly on top of it. The voltage of the Clark electrode is controlled by the picoammeter, which also relays the current readings of the electrode to a virtual chart recorder through an analog (A)-to-digital (D) interface.

DO₂ measurement. The DO₂ measurement was determined using the method described by Mahler (14). The PO₂ on the lower surface of the cusp (ventricularis side) was continuously recorded while the cusp tissue was allowed to equilibrate with a heated humidified gas containing 30% oxygen. Once a constant current reading was recorded by the sensor, the PO₂ in the chamber (upper surface of the cusp) was changed in a stepwise fashion to 21% oxygen (Fig. 2). Recording of the polarographic sensor current continued until a new steady state was reached. The PO₂ measured by the sensor at the bottom surface of the chamber remained >6 mmHg, i.e., above the critical PO₂ necessary to maintain uniform O₂ consumption (5).

View larger version (21K): [in this window] [in a new window]   Fig. 2.   A schematic showing the experimental approach for oxygen diffusivity (DO2) and oxygen consumption (O2) measurements. In both instances, the three-layered cusp structure is laid over a Clark polarographic oxygen sensor. For the DO2 measurement, after the PO2 within the tissue has equilibrated, tissue is subjected to a step change in oxygen at the open surface (fibrosa). In the case of the O2 measurement, after PO2 within the tissue has equilibrated, the coverslip is applied to create a 0 PO2 boundary condition at the fibrosa surface. Experiments take place in a sealed chamber.

Calculation of DO₂. As described by Mahler (14), the time course for the new steady state after a step change in PO₂ is predominantly monoexponential in nature and the rate constant (k) of this function is related to the diffusivity of the tissue through the equation k = ²DO₂/4L², where k is the rate constant (in s¹), DO₂ is the diffusion coefficient (in cm²/s), and L is the thickness of the tissue (in cm). Therefore, a semilogarithmic plot was generated for each run by plotting the log of the PO₂ change versus time and k was determined as the slope of the line of best fit through the data points (Fig. 3). Thickness of the cusp area over the sensor was measured after the consumption experiments. The method for measuring tissue thickness is described in Tissue thickness measurement.

View larger version (16K): [in this window] [in a new window]   Fig. 3.   The drop in PO2 as recorded by the Clark oxygen sensor after gas on the fibrosa surface of the cusp goes through a step change from 30% down to 21% PO2. Note that the PO2 has been normalized. This is the recording for sample 6. Inset shows the semilog plot from which the monoexponential rate constant (k) was calculated for determination of tissue DO2.

O₂ measurement. To measure O₂, we followed the technique of Takahashi et al. (25). The cusp was first allowed to equilibrate in the gas with 21% oxygen. Once a steady state was reached, as recorded by the PO₂ electrode on the lower surface of the tissue, the chamber was quickly opened and a coverslip was dropped onto the exposed surface of the cusp (Fig. 2). This prevented any oxygen exchange at the top surface of the tissue with the gas in the chamber. The chamber lid was then resealed and 100% N₂ was blown through the chamber to keep the tissue warmed to 37°C during the measurement. The current output of the sensor was recorded while PO₂ within the tissue fell toward 0, as the tissue consumed all the available oxygen.

Calculation of O₂. In a closed system (i.e., no external source of oxygen) O₂ can be calculated by using the formula O₂ = P/t, where O₂ is the consumption rate of the tissue (in ml O₂ · ml tissue¹ · s¹),  is the oxygen solubility (in ml O₂ · ml tissue¹ · mmHg¹), and P/t is the change in PO₂ over time (in mmHg/s). This method is based on the one used for measuring O₂ in cell suspensions (13). Oxygen solubility of the aortic valve tissue was estimated based on the assumption that solubility of a tissue is equivalent to the solubility of the water fraction of a tissue. To get a closer approximation, the solubility of oxygen in plasma was used. This solution, which contains proteins, has a solubility of 2.816 × 10⁵ ml O₂ · ml tissue¹ · mmHg¹ at 37°C, which is ~90% the value of water at 37°C (7), and is much more similar to tissue than water. Previous work in our laboratory (22) has shown that porcine aortic valve tissue is 90% water. Therefore, the oxygen solubility value used for the determination of the O₂ within porcine aortic valve tissue was 90% that of oxygen solubility in plasma at 37°C (2.534 × 10⁵ ml O₂ · ml tissue¹ · mmHg¹). To determine the rate of change in PO₂ over time, the recorded current was changed to PO₂ using the calibration curves generated at the beginning of the experiments. By using the rate of fall in PO₂ versus time and the solubility of the cuspal tissue, O₂ was calculated.

Tissue thickness measurement. To determine cusp thickness without compressing the tissue during measurement, which is one of the main obstacles of mechanical measurement techniques (12, 28), we employed a radiographic thickness imaging technique developed in our laboratory (29). After DO₂ and O₂ measurements, the cusps were placed in a physiologically balanced salt solution until thickness imaging could be performed. All radiographic imaging was carried out within a few hours of polarographic measurements. Briefly, the imaging system contained a 200-µm beryllium window-fixed tungsten anode microfocus source (Kevex X-Ray; Scotts Valley, CA) and a digital X-ray detector consisting of a cesium iodide input phosphor coated onto a fiber-optic taper bound to a charged-coupled device detector (Hamamatsu; Bridgewater, NJ). It was developed to allow for imaging of specimens up to 24 × 32 mm in the x and y directions and optimized so that discrete 50 × 50-µm (pixels) areas within a sample could be analyzed. Mean thickness values were obtained for the area of the cusp that was draped directly over the polarographic oxygen sensor. Thickness values were used in the determination of porcine aortic valve cusp DO₂ values.

Finite difference modeling of tissue PO₂ profiles. To gain an understanding of the oxygen supply within the native aortic valve cusp, a finite difference mathematical model was developed based on Fick's Law for diffusion. This would allow for in vivo tissue oxygen profiles to be created, based on our experimental in vitro DO₂, O₂, and thickness data. MATLAB, a commercially available numeric computation and visualization software package (The MathWorks; Natick, MA), was used to produce and run the simulations. The parameters of the model included a time step of 0.025 s with a node spacing of 2 × 10² cm. To evaluate the sensitivity of the approach, a model was designed to mimic the in vitro experiment. Results within 2% of the experimental values for DO₂ and O₂ were obtained, giving validity to our modeling approach. The assumption was made that PO₂ in the surrounding environment of the cusp would be 100 mmHg in vivo. The first model was created assuming that the cusp was a homogenous one-layered structure. Matched experimental values for DO₂, O₂, and tissue thickness for each sample were used and a solubility equal to 90% that of plasma at 37°C were incorporated to generate tissue PO₂ profiles as shown in Fig. 4A.

View larger version (14K): [in this window] [in a new window]   Fig. 4.   A: in vivo PO2 profile determined from the finite difference model of a homogenous cusp using experimentally determined mean thickness, DO2, and O2 values. B: in vivo PO2 profile determined from the finite difference model of a three-layered cusp using approximated values for the fibrosa, ventricularis, and spongiosa (see MATERIALS AND METHODS).

5

5

1

1

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX REFERENCES

Rate constants of the diffusivity measurements were in the range of 4.2-9.6 × 10³ s¹ for the seven samples. Mean thickness of the tissue over the oxygen sensor, where the measurements were made, ranged from 4.55 to 8.63 × 10² cm (Table 1). With the use of matching rate constants and mean thickness values for each sample, the mean DO₂ for porcine aortic valve tissue was found to be 1.06 × 10⁵ ± 1.535 × 10⁶ cm²/s. A sample recording of the change in current with time reflecting the step change in oxygen used to determine DO₂ measurements is shown in Fig. 3. The graph inlay shows the semilogarithmic plot of this data from which k can be acquired.

After a coverslip was applied to the exposed surface of the cusp, oxygen is consumed by the tissue until a level of 0 PO₂ is reached throughout the tissue. This rate of fall in oxygen was recorded by the Clark oxygen sensor in contact with the lower surface of the tissue and was determined to be within a range of 0.68-2.13 mmHg/s. By using a solubility value of 2.534 × 10⁵ ml O₂ · ml tissue¹ · mmHg¹, the mean O₂ for the tissue was found to be 3.05 × 10⁵ ± 1.37 × 10⁵ ml O₂ · ml tissue¹ · s¹.

The oxygen profiles for the one-layer model were determined for each sample using their individual experimental values for thickness, DO₂, and O₂. The predicted minimum PO₂ values at the center of the spongiosa in vivo for a surface PO₂ of 100 mmHg for each cusp can be found in Table 2. The mean minimum PO₂ at the cusp center for the one-layer model under these conditions was 42 mmHg. In contrast, when the cusps were each modeled as a three-layered structure, more closely representing the in vivo situation, and solubility and diffusivity values reflected tissue layer composition, their individual oxygen profiles returned lower minimum PO₂ values at the maximum distance from the cusp surfaces. These values are also reported in Table 2. The mean minimum PO₂ value of 27 mmHg predicted using the three-layered model created a 26% larger PO₂ gradient than that determined using the one-layer model and a paired Student's t-test found this difference to be statistically significant (P  0.001). Examples of the mean oxygen profiles for both the homogenous and heterogeneous tissue models can be seen in Fig. 4.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION APPENDIX REFERENCES

Present recommendations for the creation of a tissue engineered bioprosthetic valve involve a biodegradable scaffold seeded with the cells of the recipients (31). The long-term goal is to create a self-sustaining valve implant containing living cells enabling ongoing tissue repair without immune rejection. To achieve this goal, several issues are being investigated, including the most suitable harvest source for these cells, the cell density necessary to produce a stable structure, and the appropriate scaffolding material. The final device must embody the flexibility and mechanical strength needed to withstand the substantial physical forces applied during the cardiac cycle (24). Because the ultimate goal is also to create a durable implant that will survive the lifetime of the patient, careful consideration must be given to all parameters of native aortic valve function. For example, not only is it important to seed the implant with the appropriate cells and density, it is also crucial that an adequate oxygen supply be available to these cells to allow for optimal function.

To date, the three-layered structure of aortic valve cusps is well defined, as are the physical properties and orientation of its major structural components, collagen and elastin (3, 6, 22). The cellular content of the cusps is less well understood, but several recent studies (16-18) have provided clarification. For these interstitial cells, an oxygen delivery system is certainly necessary, and aortic valve metabolic requirements have not yet been quantified. A blood supply within the valve has been identified previously and discussed intermittently over the past several decades. Its distribution and relation to tissue requirements has been unclear, but we (30) have recently shown that the aortic valve capillary network is contained within the thickest regions of the cusps, where oxygen diffusion from the cusp surfaces alone, is unlikely to provide an adequate environment to sustain cellular activity. The addition of such a vascular bed to a tissue-engineered valve would substantially increase the level of complexity and it would therefore be desirable to avoid that requirement if possible. A full understanding of the metabolic activity of valve cusps is therefore necessary to address this issue.

Using the Clark oxygen sensor technique, we have been able to determine both oxygen diffusion and consumption values of porcine aortic valve tissue. The mean value of 1.06 × 10⁵ cm²/s we found for DO₂ at 37°C was well within the range of in vivo values reported in the literature for other vascular structures such as the dog femoral artery (8) and the aorta of various species (4). Our reported O₂ values ranged from 1.73-5.39 × 10⁵ ml O₂ · ml tissue¹ · s¹ and were anywhere from 59% to 87% lower than the mean in vivo value reported for the dog femoral artery (8). Because of the similar nature of these vascular structures (18) and the reported rates of turnover within the valves (20, 26), it is possible that our in vitro measurement has provided an underestimation of aortic valve cusp O₂ values.

To better understand the implications of our experimental results, we developed the finite difference model to examine oxygen profiles within the tissue. In our experiments, we treated the aortic valve cusps as homogeneous structures by determining an overall diffusion and consumption value. When these experimental results were incorporated into a one-layer cusp model and oxygen profiles within the cusp tissue were determined, it would seem that oxygen supplied by diffusion from the surfaces of the cusp was adequate to sustain the metabolic function of the tissue. However, examining the cusp as a homogenous tissue is a gross oversimplification and much information exists with regard to the composition of the aortic valve as a three-layered structure (3, 11). Therefore, we developed a second model comparing the oxygen profile when the three layers of the cusp were taken into account. This latter model required that diffusivity and solubility values of the three layers be approximated based on layer composition. Because of a recent report by Rocca et al. (18) on uniform cellular distribution within the tissue, tissue O₂ was taken as uniform throughout the structure. The resulting oxygen profiles confirmed that there was a statistically significant difference in oxygen permeability if the valve were modeled as a three-layered structure. Furthermore, even with a conservative O₂ value, a PO₂ of 0 was reached in the center of a cusp. These results predict that oxygen diffusion from the cusp surfaces alone may be inadequate to support cell viability and that the microcirculation previously identified within the spongiosa is necessary to act as an additional oxygen source. If myoglobin (or another O₂ carrier) were present within the tissue, the impact of facilitated diffusion would have to be added to the model. To date, we have found no reports in the literature identifying myoglobin within aortic valve cusp tissue. It is also possible that there is a nonlinear distribution of DO₂, solubility, and O₂ across the layers of the cusp and that a three-layered model is still an oversimplification.

Because our studies were necessarily carried out in vitro rather than in vivo, our estimates of O₂ may be lower than that present in a functioning valve. In addition to the transport time of the whole heart on ice, the time delay between tissue excision to tissue assessment of ~90 min may be responsible for stunning the interstitial cells, resulting in an overall decrease in O₂. Furthermore, there have been reports demonstrating the presence of contractile smooth muscle bundles (1) and innervation within the valve (15). This suggests an active component to valve motion in vivo that would not be present in the in vitro studies, which could also contribute to an underestimation of O₂. For these reasons, our results likely reflect the minimum oxygen requirements of the valve tissue and the oxygen needs in vivo may be significantly higher.

In addition, cusp thickness is extremely variable and tissue thickness has a significant impact on both the calculation of DO₂ and the tissue oxygen profiles. In our experiments, we only made thickness measurements of tissue lying directly over the oxygen sensor and thereby only utilized tissue thickness in a small area of the cusps. If the thickness of the entire cusp is examined, as shown for one sample in Fig. 5, it becomes apparent that there are much thicker regions than those we have analyzed. The impact of these larger diffusion distances would be lower PO₂ values within the center of the cusps.

View larger version (104K): [in this window] [in a new window]   Fig. 5.   A thickness map of an aortic valve cusp (sample 6) using the radiographic technique discussed in MATERIALS AND METHODS. The thickness area where the polarographic oxygen sensor was approximately located is outlined in black. The mean thickness of this region was used for the calculation of tissue DO2.

If an optimal bioengineered replacement valve were to be developed with a potential durability matching patient lifespan, that tissue must have the ability to sustain adequate oxygenation. These experiments give initial insight into the oxygen properties of porcine aortic valve tissue and, utilizing the modeled oxygen profiles, reinforce the concept that a microcirculation is necessary within the spongiosa as an additional oxygenation source. Of course, the microcirculation may also be used to carry away potentially harmful metabolic products, with lower diffusivity than oxygen, from the deeper tissue. These values of oxygen diffusion and consumption will also allow for comparison to be made with tissue-engineered prototypes before implantation to examine similarities between the designed implants and native tissue.