Prevention of Bioprosthetic Heart Valve Calcification by Ethanol Preincubation
Efficacy and Mechanisms
Background Calcification of the cusps of bioprosthetic heart valves fabricated from either glutaraldehyde cross-linked porcine aortic valves or bovine pericardium frequently causes the clinical failure of these devices. Our investigations studied ethanol pretreatment of glutaraldehyde cross-linked porcine aortic valves as a new approach to prevent cuspal calcification. The hypothesis governing this approach holds that ethanol pretreatment inhibits calcification resulting from protein structural alterations and lipid extraction.
Methods and Results Results demonstrated complete inhibition of calcification of glutaraldehyde-pretreated porcine bioprosthetic aortic valve cusps by 80.0% ethanol in rat subdermal implants (60-day ethanol-pretreated calcium level, 1.87±0.29 μg/mg tissue compared with control calcium level, 236.00±6.10 μg/mg tissue) and in sheep mitral valve replacements (ethanol-pretreated calcium level, 5.22±2.94 μg/mg tissue; control calcium level, 32.50±11.50 μg/mg tissue). The mechanism of ethanol inhibition may be explained by several observations: ethanol pretreatment resulted in an irreversible alteration in the amide I band noted in the infrared spectra for both purified type I collagen and glutaraldehyde cross-linked porcine aortic leaflets. Ethanol pretreatment also resulted in nearly complete extraction of leaflet cholesterol and phospholipid.
Conclusions Ethanol pretreatment of glutaraldehyde cross-linked porcine aortic valve bioprostheses represents a highly efficacious and mechanistically based approach and may prevent calcific bioprosthetic heart valve failure.
Bioprosthetic heart valves fabricated from glutaraldehyde-pretreated porcine aortic valves or bovine pericardium were initially preferred because of reduced thrombogenicity over mechanical heart valve prostheses. However, it has become evident that calcification is one of the most frequent causes of BPHV failure,1 2 3 and this has led to reduced clinical usage. Various anticalcification strategies—including pretreatment of the valves with either metallic salts, detergents, or bisphosphonates, coimplants of polymeric controlled release drug delivery systems, and covalent attachment of anticalcifying agents—have been investigated with animal models.4 5 6 7 8 All these approaches have been partially effective in preventing calcification in rat subdermal implants. However, only SDS and AOA used as pretreatments have been found to be partially effective in preventing calcification in both rat subdermal and sheep circulatory implant studies. Nevertheless, the mechanism of inhibitory approaches is incompletely understood. At present, there is no proven effective approach for preventing BPHV calcification in the clinical setting.
In the present paper, we report an efficacious approach in preventing calcification of BPHVs using ethanol pretreatments. Our working hypothesis is that ethanol pretreatment inhibits bioprosthesis calcification through an interaction of membrane-lipid removal and ethanol-induced collagen structural changes. Because cell membrane–oriented calcification and collagen calcification are the prominent pathophysiological features of bioprosthetic tissue mineralization, inhibition of both processes may be a prerequisite for high therapeutic efficacy.
The objective of the present study was to investigate the mechanism of ethanol pretreatment for preventing BPHV calcification. Our goals were to (1) assess the comparative effects of ethanol pretreatment on the lipid content of bioprosthetic leaflets, (2) assess the effect of ethanol pretreatment on leaflet collagen conformation by the use of IR, (3) determine the effects of ethanol on tissue morphology and stability, and (4) investigate the dose response and comparative efficacy of ethanol for preventing bioprosthetic leaflet calcification in rat subdermal implants and sheep mitral valve replacements.
Absolute 200-proof ethanol was obtained from McCormick Distilling Co Inc. HEPES was obtained from Sigma Chemical Co. Glutaraldehyde as an 8.0% EM-grade solution was obtained from Polysciences. Type I collagen (99.9% pure) was obtained from Collagen Corp (as a sterile solution of pepsin-solubilized bovine dermal collagen dissolved in 0.012 N HCl, 3 mg/mL). Ketamine hydrochloride (Aveco) and Rompun (Haver) were used for rat anesthesia.
Bioprosthetic Tissue Preparation
Glutaraldehyde Fixation Protocols
Fresh porcine aortic heart valves were shipped on ice from St Jude Medical, Inc, to the laboratory within 24 hours of slaughter. Cusps and aortic wall samples were separated and cross-linked in a 0.6% glutaraldehyde solution (diluted from 8.0% aqueous solution) buffered with 50 mmol/L HEPES, pH 7.4, then transferred after 24 hours to 0.2% glutaraldehyde solution in the same buffer for storage at room temperature for at least 1 week to achieve complete fixation.
Ethanol Exposure Conditions
Ethanol Pretreatment Before Cross-linking for Rat Subdermal Studies
Fresh porcine aortic valve leaflet tissues were placed in aqueous ethanol solutions (10 mL per valve leaflet) with concentrations of 20.0%, 40.0%, 60.0%, 80.0% (vol/vol), and 100.0% (nonbuffered) prepared in HEPES buffer, pH 7.4, as above. The solutions were kept on a shaker bath at room temperature (25°C) for 24 hours; then tissues were rinsed free of ethanol in an excess volume of HEPES buffer and subjected to the cross-linking as described above.
Ethanol Pretreatment After Glutaraldehyde Cross-linking for In Vitro and Rat Subdermal Studies
Glutaraldehyde cross-linked leaflet tissues were placed in ethanol solutions (with concentrations and volumes as described above) prepared in HEPES buffer, pH 7.4. The pretreatment was continued for 24 hours; then the tissues were rinsed free of ethanol with saline before use in in vitro studies and rat subdermal implantation.
Similarly, glutaraldehyde-pretreated leaflet samples were treated with HEPES buffer (10 mL per valve leaflet) as a control and chloroform-methanol (2:1) for 24 hours on a shaker bath at 25°C. After pretreatment, all tissue samples were equilibrated in HEPES buffer, pH 7.4, for in vitro studies or in sterile saline for in vivo studies for at least 24 hours to remove traces of solvents before further use.
Ethanol Pretreatment Protocol for Stent-Mounted Bioprostheses for Sheep Implants
Clinical-grade, 25-mm stent-mounted bioprostheses of a proprietary design (Bioimplant) were provided by St Jude Medical, Inc. Glutaraldehyde-pretreated bioprostheses for sheep implants were exposed to ethanol (80.0% ethanol in HEPES buffer, pH 7.4, at room temperature for 72 hours) as necessitated by manufacturing sterilization protocols. Control bioprostheses for sheep mitral valve replacement were prepared identically but without ethanol exposure. All bioprostheses were washed in sterile saline before implantation.
DSC (model DSC7, Perkin Elmer) was used to assess the thermal denaturation temperature of the tissue proteins. Wet tissue samples (6 to 8 mg) were sealed hermetically in DSC pans and scanned from 30°C to 100°C at 10°C/min. The DSC instrument was standardized regularly with metallic indium. Peak maxima were reported as the thermal denaturation temperature by use of the average of at least three separate runs.
Hydrated, glutaraldehyde–cross-linked porcine aortic valve leaflets and pure type I collagen films (either control or ethanol pretreated) were used for FTIR studies. Pure collagen films were obtained as follows. Type I collagen solution was adjusted to pH 7.4 according to the procedure provided by the manufacturer and then placed in a Petri dish (10 mL solution of 3 mg/mL concentration in a 100-mm-diameter Petri dish). The solution was gelled at 37°C for 1 hour and left uncovered in a hood until dry. The film was washed with distilled water to remove the salts and then cross-linked with glutaraldehyde with the procedure discussed above for the leaflets.
A Perkin Elmer (model 1740) FTIR spectrometer was equipped with a horizontal attenuated total reflectance ZnSe cell (ATR, SpectraTech Inc). A water spectrum was obtained as a reference and then subtracted from the sample by the instrument-controlled computer. The collagen films and leaflets were equilibrated in double-distilled water and blotted on a filter paper. The hydrated samples were then placed on an ATR cell and pressed with a sponge to obtain good surface contact between the cell and the sample. Fifty scans were obtained with a resolution of 4 cm−1. The spectra were then deconvoluted in the region of 1800 to 1000 cm−1 with the Perkin Elmer software program to resolve the different amide bands. Three to five replicates were performed for each sample.
Implant and Explant Procedures
Male weanling rats (50 to 60 g, CD, Sprague-Dawley, Charles River Laboratories, Burlington, Mass) were anesthetized by an intramuscular injection of ketamine and Rompun (4:3), 0.001 cm/g body weight. Two subdermal pouches were created (one ventral and one dorsal) on each rat as described previously,9 and two tissue samples (washed with sterile saline to remove any traces of solvents and glutaraldehyde) were implanted in these pouches. The rats (five per group) were killed by CO2 asphyxiation after specified time periods (21 or 60 days). The leaflet samples were explanted and rinsed with saline solution. A major portion of each sample was used for quantitative calcium and phosphorous determination, whereas representative samples were fixed for morphological studies (see below). Femurs from representative rats were harvested and fixed in 10.0% neutral buffered formalin for morphological bone growth studies.
Clinical-grade porcine bioprosthetic valves (St Jude Medical, Inc, as above; size, 25 mm) were implanted in the mitral position of Dorset or Colombian cross-bred female sheep (3 to 4 months old, 28 to 38 kg). The detail techniques of this procedure are described in previous publications.10 Preoperatively, the sheep were vaccinated against Clostridium C and D and received ticarcillin disodium and gentamicin sulfate. Valves were implanted into the mitral position under halothane/isoflurane inhalation anesthesia via a standard left thoracotomy through the fourth intercostal space with the aid of cardiopulmonary bypass. Subcutaneous heparin sulfate (1000 U) was given twice a day for 2 days; piroxicam (10 mg orally) was given each day beginning on postoperative day 1, but no anticoagulation was administered long term. Sheep used for this study were fed diets containing the recommended daily calcium intake for young developing sheep. Approximately 2 weeks after the operation, the sheep were transferred to a farm, where they were monitored. Typical implant duration was 150 days. All sheep received humane care in compliance with the procedures formulated by the NIH (NIH publication No. 86-23, revised 1985).
Explanted bioprosthetic tissue samples (leaflets and aortic wall) from either rats or sheep were washed with an excess of saline followed by distilled water, lyophilized, and weighed. The following procedure was used for the determination of calcium. Each sample (15 to 40 mg) was hydrolyzed in 1 mL of 6 N HCl in a test tube at 70°C for 4 to 6 hours in a closed container. The solution was then completely evaporated at 70°C with a continuous flow of nitrogen in the tube. The residue left in the test tube was then dissolved in 1 mL of 0.01 N HCl. The calcium content of each leaflet was based on the aliquot concentration of each hydrolysate by use of atomic absorption spectroscopy according to the established procedures.9
Amino Acid Analysis
Leaflet samples were treated under the conditions described above and then reequilibrated in HEPES buffer, pH 7.4. The samples were lyophilized and weighed. Typically, a 3- to 5-mg tissue sample (dry weight) was hydrolyzed in 6 N HCl at 70°C in a closed container to obtain a hydrolysate. The hydrolysate was evaporated to dryness at 70°C in a stream of nitrogen. The residue left in the tube was dissolved in 0.01 N HCl and used for amino acid analysis on an automated amino acid analyzer (model 420H, Applied Biosystems).
Glutaraldehyde-fixed leaflet samples (control and ethanol pretreated) were extracted with chloroform/methanol 2:1 (vol/vol).11 Total cholesterol was determined enzymatically with fluorometric detection.12 Total phospholipid phosphorous was determined by the modified Fiske-SubbaRow method13 with a Fiske-SubbaRow reagent obtained from Fisher Scientific Co. Total cholesterol and phospholipids were expressed as nanomoles per milligram of dry tissue weight.
Multiple representative ethanol-pretreated (20.0%, 40.0%, 60.0%, and 80.0%) and untreated control aortic valve tissue samples explanted from the rat subdermal study were fixed in 25.0% glutaraldehyde and 2.0% paraformaldehyde in cacodylate buffer, pH 7.4 (Karnovsky's fixative), dehydrated through graded ethanols, and embedded in glycolmethacrylate (GAM, Polysciences Inc). Femurs retrieved from rats at the time of death were fixed in 10.0% neutral buffered formalin and similarly embedded. For light microscopic examination, sections cut 3 μm thick were stained with hematoxylin and eosin for overall morphological features and the von Kossa stain for calcium phosphates.
For ultrastructural morphology, representative specimens of 80.0% ethanol–pretreated tissue, both unimplanted and explanted from the rat subdermal site after 21 and 60 days, were fixed as above; postfixed in 2.0% osmium tetroxide, dehydrated ethanol, and propylene oxide; and embedded in Poly/Bed 812 (Polysciences). Sections were cut 60 nm thick, stained with lead citrate and uranyl acetate, and examined with a JEOL-100CX transmission electron microscope (JEOL) at an accelerating voltage of 80 kV.
All data are expressed as mean±SEM. For lipid and protein analyses, at least five different samples were used for each measurement. For in vivo calcification studies in either the rat subdermal or sheep mitral valve replacement model, at least 10 different samples were analyzed for each group. Statistical significance was established with ANOVA. Probability values of P≤.01 were considered significant.
The present study has demonstrated that ethanol pretreatment of BPHVs is highly efficacious in preventing calcification in both rat subdermal implants and sheep mitral valve replacement models without adversely affecting tissue stability and morphology.
In Vitro Material Assessment
All tissue samples were equilibrated in saline after the ethanol pretreatments before DSC was performed. The control leaflets cross-linked with glutaraldehyde and equilibrated in saline at pH 7.4 showed a thermal denaturation temperature of 88.3±0.56°C. For various pretreatments ranging from 40.0% to 100.0% ethanol, the temperatures were found to range from 87.6°C to 87.8±0.43°C, following ≤5 minutes of rehydration. Thus, ethanol pretreatment did not significantly change the temperatures.
Lipid Extraction Results
A threshold was observed for ethanol pretreatments in removing lipids from the tissue (Table 1⇓). At a concentration of 40.0%, the ethanol pretreatment did not remove cholesterol and phospholipids from the leaflet samples. However, at a concentration of ≥60.0%, ethanol pretreatments were highly efficacious in extracting cholesterol and phospholipids from the tissue. The 60.0% ethanol pretreatment extracted up to 98.0% of the cholesterol and 71.0% of the phospholipids, whereas the 80.0% ethanol pretreatment removed 99.0% of the cholesterol and 94.0% of the phospholipids. The other solvent included in the study (chloroform-methanol [2:1], a known lipid extraction medium11 ) was also found to be very effective in extracting almost all cholesterol and most of the phospholipids from the leaflet tissue. Thus, the 80.0% ethanol pretreatment was found to be as effective as the positive controls in removing cholesterol and phospholipids from the tissue. The relationship of the various delipidation protocols to calcification inhibition is described below.
Amino Acid Analyses
Total amino acid analysis of the leaflet samples after ethanol pretreatments (Table 2⇓) demonstrated neither significant bulk extraction of total protein nor alteration in amino acid composition, as assessed by percent distribution of major amino acids per 100 amino acid residues, as a result of ethanol pretreatments.
An overlay of the deconvoluted IR spectra of glutaraldehyde–cross-linked unimplanted leaflets (control) and glutaraldehyde–cross-linked, 30.0% and 80.0% ethanol–pretreated, unimplanted leaflets is shown in Fig 1A⇓. For the unimplanted glutaraldehyde-pretreated leaflets not exposed to ethanol (control), the band at 1656 cm−1 had a higher intensity than the band at 1631 cm−1. After 80.0% ethanol pretreatment the bands (1658 and 1632 cm−1) in the amide I region were shifted slightly, along with changes in the relative band intensities; it was observed that the 1632-cm−1 band had a relatively greater intensity than the 1658-cm−1 band. No changes were observed in the amide II region or any other part of the spectrum. Spectra for leaflets pretreated with 30.0% ethanol for 24 hours (Fig 1A⇓) show no remarkable changes compared with control untreated leaflets. When type I collagen films were subjected to IR in the hydrated state, similar changes in the amide I region were observed with the 80.0% ethanol treatment (Fig 1B⇓).
In Vivo Calcification Assessment in Rat Subdermal Studies
Ethanol Concentration Effects on Calcification
In 21-day rat subdermal implants, a nearly complete inhibition of calcification of leaflets was noted for ethanol pretreatment (24-hour incubation) at concentrations of ≥50.0% (Fig 2⇓). No concentration-dependent dose response was found. Instead, little to no calcification occurred above a threshold level of ethanol (50.0%). We also performed 21-day rat subdermal calcification studies for the leaflets pretreated with 80.0% ethanol for 72 hours. There was no improvement in the anticalcifying efficiency with the longer pretreatment time periods, although this pretreatment also prevented calcification (calcium level, 3.96±1.85 μg/mg tissue). Thus, 80.0% ethanol provides optimal inhibition over the range of 24 to 72 hours of pretreatment. A 60-day rat subdermal group (24-hour ethanol pretreatment) was also studied. For this implant duration, controls calcified severely (calcium level, 236±6.1 μg/mg tissue). The 80.0% ethanol (in HEPES, pH 7.4, 24 hours) pretreatment was most effective, with complete inhibition of calcification with the calcium levels comparable to unimplanted bioprosthetic tissue (calcium level, 1.87±0.29 μg/mg tissue), whereas the 60.0% ethanol pretreatment was partially effective (calcium level, 28.5±12.0 μg/mg tissue). Thus, the 80.0% ethanol pretreatment was found to be the best condition for preventing leaflet calcification in both the 21- and 60-day rat subdermal models.
Delipidation Conditions (Effects on Calcification)
Because the ethanol pretreatment was demonstrated to be effective in preventing calcification of leaflets, we sought to study inhibition of calcification by pretreatment of leaflets with chloroform-methanol (2:1), which was also as efficient as ethanol in extracting phospholipids and cholesterol (Table 1⇑). We performed 21-day rat subdermal studies for leaflets pretreated with chloroform-methanol (2:1) solutions for 24 hours. Although chloroform-methanol (2:1) was a very efficient delipidation solution, only partial inhibition calcification was seen (calcium level, 65.3±14.1 μg/mg tissue).
Ethanol Exposure Before Glutaraldehyde Cross-linking
As Fig 3⇓ shows, there was significant inhibition of calcification after the ethanol exposure (≥60.0%), regardless of the fact that pretreatments were carried out before glutaraldehyde cross-linking. However, the inhibition of calcification was not as complete as the ethanol pretreatment after routine glutaraldehyde cross-linking. Thus, these results suggest that the mechanism of action of ethanol must be due in part to interactions with glutaraldehyde.
In Vivo Calcification Assessment in Sheep Studies
A controlled comparison of orthotopic porcine aortic BPHV replacements in the mitral position was carried out that compared the 80.0% ethanol–pretreated group with a series of control implants prepared identically but without ethanol exposure. All implants functioned for 150 days. Ethanol pretreatment completely prevented valve leaflet calcification compared with controls not exposed to ethanol (Table 3⇓). Ethanol-pretreated leaflets had calcium levels comparable to unimplanted leaflets (calcium level, 2.8±0.7 μg/mg tissue). These results indicate that ethanol pretreatment was also highly effective in preventing calcification in a circulatory model of sheep mitral valve replacement.
Figs 4⇓ and 5 summarize the morphological results. In all ethanol groups, neither inflammation nor other untoward tissue reaction was noted (Fig 4A⇓); in the 60.0% and 80.0% ethanol–pretreated groups, calcification was almost completely inhibited, with only focal mineral noted (Fig 4B⇓). In agreement with the chemical data from subcutaneous explants, calcification of 20.0% and 40.0% ethanol–pretreated cusps was severe and identical to that of control untreated cusps (Fig 4C⇓). Mitral valves pretreated with 80% ethanol that were explanted from sheep at 5 months had no to minimal focal cuspal mineralization and only focal aortic wall calcification (Fig 4D⇓); in contrast, untreated valves had marked nodular cuspal and aortic wall mineralization (Fig 4E⇓). By electron microscopy, unimplanted 80.0% ethanol–pretreated tissue had a morphology of cells and extracellular matrix, including collagen, typical of conventional, high-pressure glutaraldehyde-fixed porcine aortic valves; as previously described for porcine bioprosthetic valve tissue treated with antimineralization agents that extract lipid, the glutaraldehyde-devitalized connective tissue cells intrinsic to the cusps appeared somewhat indistinct because of vacuolization of the cytoplasm or membrane dissolution14 (Fig 5A⇓). Electron microscopy also demonstrated that 80.0% ethanol–pretreated tissue explanted from the rat after 60 days subdermal was devoid of mineralization in both cells and collagen (Fig 5B⇓). Moreover, no abnormalities were noted in the femoral growth plate morphology of rats with 60-day 80.0% ethanol–pretreated implants (data not shown).
The key findings in the present study are as follows. First, ethanol pretreatment of BPHVs was found to completely inhibit leaflet calcification both in rat subdermal implants (60 days) and in sheep circulatory implants as orthotopic mitral valve replacements (150 days). Second, the mechanism of ethanol inhibition of calcification is not completely understood. However, the findings of the present study may serve to elucidate. A number of chemical changes were caused by the ethanol pretreatment, including an efficient extraction of cholesterol and phospholipid, and significant changes in collagen structure. Specifically, the infrared amide I region demonstrated a prominent and irreversible change, which may be indicative of structural alterations relating to the resistance to calcification.
Although ethanol has not previously been described as an agent that might be useful for preventing cardiovascular calcification, the present results are not completely unexpected in view of prior work related to ethanol and physiological mineralization. Cell biology studies with ethanol have demonstrated prominent abnormalities in the cellular metabolism of calcium in bone line cells and fibroblasts.15 16 In other studies, it has been reported that the presence of ethanol tends to fluidize membranes or disorder acyl chains of phospholipids, which affect many cellular activities.17 Ethanol is also well known for its disturbance of normal cellular membrane handling of ions.17 Furthermore, ethanol significantly affects calcium phosphate nucleation and phase transformations because of its interactions with water.18
In addition, the changes in the amide I region of IR noted in the present study have been observed by others using solvents other than ethanol. For example, Brodsky found that collagen dissolved in the solvent hexafluoro-isopropanol undergoes a threefold drop in molecular weight and loses its triple helical conformation.19 Furthermore, it is very likely that the combined effect of the ethanol-induced collagen change and the other chemical changes induced by ethanol (lipid extraction, altered water status) result in an unfavorable milieu for calcium phosphate formation and phase transformations.
Ethanol and BPHV Calcification Mechanisms
Lipid Extraction Considerations
To understand the mechanism of action of ethanol in preventing BPHV calcification, leaflet samples were analyzed for total lipid and cholesterol content before and after pretreatment (Table 1⇑). Ethanol (at a concentration of ≥50.0%) was a very efficient extractor of both cholesterol and phospholipids, with nearly complete extraction of both. Membrane-bound phospholipids are considered to be donors of phosphorous in the initial stages of mineralization of BPHVs because of hydrolysis by alkaline phosphatase.20 Complete removal of phospholipids, which are initial sites of calcification, may partially explain the mechanism of action of ethanol. However, the results with chloroform-methanol (2:1) demonstrated that this delipidation regimen resulted in the complete extraction of both total cholesterol and phospholipid (Table 1⇑). Nevertheless, this was associated with a less complete inhibition of calcification (calcium level, 65.3±14.1 μg/mg tissue) than ethanol in a 21-day rat subdermal model. Thus, these data indicate that although lipid extraction may play a part in the mechanism of action of ethanol, lipid extraction alone cannot completely explain the anticalcification efficacy of ethanol, and ethanol may be altering the other factors that influence mineralization.
Ethanol and Material Effects Considerations
The DSC results obtained in this study suggest that ethanol pretreatment did not alter the thermal shrinkage temperature of the tissue, indicating no changes in material stability (ie, cross-link status). This is important because some pretreatments, such as polyoxyethylene ether (Triton X-100) and N-lauryl sarcosine, have been found to cause marked tissue weakening that resulted in cuspal perforations after 20 weeks of implantation in the mitral position in sheep.21 The thermal shrinkage data will not guarantee an extended material stability, and long-term accelerated-wear testing studies need to be performed.
Because the major part of the bioprosthetic leaflets is made up of type I collagen, spectroscopic methods used in the past for collagen can be applied to the leaflets. The triple helical conformation of collagen has been studied through various spectroscopy procedures including FTIR spectroscopy, circular dichroism, and Raman spectroscopy.19 22 23 24 25 26 27 Furthermore, variations on these techniques have been particularly useful for investigating protein structural changes. Previous IR spectroscopy investigations with collagen demonstrated characteristic structural changes, especially in the carbonyl–amide I stretching region induced by various solvent exposures, including alcohols and other dehydrating conditions.19
In our work, aortic valve leaflet tissue samples were used for IR analysis. The leaflets were 0.3 to 0.4 mm thick and not transparent for transmission IR analysis. This necessitated the use of a horizontal ATR accessory to obtain the IR spectra. The spectra obtained were deconvoluted in the 1800- to 1000-cm−1 region to examine the differences in the collagen conformation resulting from ethanol pretreatment. Previous reports have shown that major changes in band position and absorbance occur in these regions because of denaturation or fibrillogenesis processes.23 Bands observed in the amide I region (1690 to 1625 cm−1) are attributed to the amide carbonyl stretching in collagen.19 27 The amide I band of collagen can be resolved into several components, indicating nonequivalent amide carbonyl bonds in the triple helix.23 26 In particular, the 1660-cm−1 band was assigned to prolyl carbonyls directed inside the triple helix (capable of intramolecular hydrogen bonding within the triple helix), whereas a band at 1630 cm−1 was assigned to amide carbonyls directed outward that are capable of intermolecular hydrogen bonds with water molecules. The significant decrease in the ratio 1660/1630 cm−1 is interpreted to indicate complete denaturation of the triple helix. In this study, major changes were observed in the amide I region of the spectra after the 80.0% ethanol pretreatment. When the IR spectra were deconvoluted, it was found that the amide I region consisted of two bands for glutaraldehyde cross-linked leaflets (1658 and 1632 cm−1). After 80.0% ethanol pretreatment, the ratio 1658/1632 cm−1 was decreased (increase in intensity of the 1632-cm−1 band), suggesting some sort of coil expansion and/or denaturation of the triple helix (Fig 1A⇑). Tissue samples were cross-linked with glutaraldehyde before ethanol treatment so that the collagen triple helix was stabilized. After ethanol treatment, tissue samples were equilibrated in an aqueous medium for several days before the IR spectra were obtained. Thus, the spectra indicate that ethanol treatment (80.0%, 24 hours) resulted in conformational changes in the collagen molecule despite the fact that it was cross-linked and thus stabilized. The resulting change was irreversible. On the other hand, 30.0% ethanol treatment for 24 hours, which did not prevent calcification, also did not show any spectral changes compared with control leaflets (Fig 1A⇑).
Leaflet tissues have many other components besides type I collagen (the major component). To delineate the fact that changes seen in IR spectra by ethanol treatment in leaflet tissue are not due to changes in other components, pure type I collagen films were prepared and exposed to the similar ethanol pretreatment. IR spectra of hydrated collagen films showed amide I band changes similar to those seen in leaflet tissues caused by 80.0% ethanol treatment (Fig 1B⇑). These collagen results raise questions about the prior changes that could be present in the BPHV leaflets exposed to ethanol before glutaraldehyde fixation. These experiments were beyond the scope of the present study but would be an important subject for future work.
Aortic Wall Calcification and Ethanol Effects
Aortic wall calcification has recently become of interest because of the increasing use of stentless BPHVs consisting of the aortic valve apparatus and ascending aortic root. These devices are particularly useful for aortic valve replacements in the presence of a very narrow annulus and for other similarly difficult surgical situations. In this stentless bioprosthetic configuration, the aortic wall segment is relatively great compared with a stented valve, and the aortic wall calcific deposits are not contained within a Dacron-wrapped stent. Thus, there has been increasing concern about the importance of aortic wall calcification as a failure mode to be anticipated in the future as a clinical outcome for patients implanted with stentless valves.28 Although aortic wall calcification was beyond the scope of the present investigations, our electron microscopy observations described above demonstrated that ethanol pretreated valves typically had only focal aortic wall calcification; in contrast, bioprosthetic valves that were not exposed to ethanol had marked nodular cuspal and aortic wall mineralization. Thus, these results suggest that ethanol may have some efficacy for inhibiting aortic wall calcification. This would be in contrast to the observations that most agents investigated thus far that appear to be effective in preventing valve leaflet calcification, such as AOA,28 failed to prevent aortic wall calcification. Future investigations of ethanol pretreatment in this regard are thus warranted.
Comparisons With Other Methods and Mechanisms
All inhibition strategies under investigation focus on one of the host, implant, or mechanical determinants of calcification. For example, detergent pretreatment (SDS, Tween 80) of bioprosthetic tissue may inhibit their calcification by removing nucleation sites (phospholipids) from the substrate.5 Detergents also have been considered to affect the membrane surface charge and collagen fibrils. However, it has been observed that certain surfactants can result in a severe decrease in the durability of the BPHVs.29 Also, replacement orthotopic and conduit-mounted heart valves treated with various surfactants have shown inconsistent efficacy of the anticalcification treatment.21 Recently, use of 2-AOA as a detergent to prevent BPHV calcification has been investigated.6 28 30 It has been hypothesized that 2-AOA binds to free aldehyde groups of tissue-bound glutaraldehyde as a result of a Schiff base reaction. The results showed that AOA significantly reduced leaflet calcification in left ventricular apicoaortic shunts in a sheep model. Reduced calcium diffusion through AOA-treated leaflets compared with controls was thought to be the part of the mechanism of action of AOA. However, the issues of the in vivo stability and the possibility of degradation of AOA remain unresolved.
An alternative strategy under study is the use of anticalcifying agents, such as bisphosphonates, to block the growth of developing hydroxyapatite crystals. However, such compounds, when used systemically, have irreversible adverse effects on bone and calcium metabolism.31 Thus, a site-specific controlled delivery approach was sought for these drugs. The slow-release drug delivery systems for bisphosphonates, when placed adjacent to the bioprosthetic leaflet implant, profoundly inhibited glutaraldehyde-fixed porcine aortic valve leaflet calcification in a rat subdermal model, with no adverse effects of the drug.32 More recently, a synergistic effect of the two drugs, ethanehydroxybisphosphonate and FeCl3, released from polymeric matrices inhibited aortic wall calcification in a rat aortic allograft model.7 However, such controlled release systems have been found to be less effective in preventing leaflet calcification of porcine-based BPHV.
The use of covalently linking anticalcification agents, such as APD, to leaflet tissue has been shown to be effective in a rat subdermal model.8 However, circulatory studies of APD-pretreated BPHVs have failed to show inhibition of calcification.21 The most likely cause for this failure may be the instability of the glutaraldehyde linkage to the bioprosthetic tissue and unstable APD binding to the residual aldehyde functions.
Aluminum is associated with osteomalacia in renal dialysis patients. This observation led to strategies using metallic salts containing Al3+ and Fe3+ to prevent calcification. Cuspal incubation in solutions of trivalent ions such as Al3+ and Fe3+ have been shown to markedly inhibit calcification of subcutaneously implanted bioprosthetic tissue in the rat at very low concentrations (0.001 mol/L).33 Morphological studies of bioprosthetic tissue pretreated with aluminum and ferric chloride reveal localization of both Al3+ and Fe3+ to devitalized cells. Thus, these agents may inhibit membrane-linked calcification events.
The use of protamine sulfate has also been investigated.34 Covalent bonding of protamine to bioprosthetic tissue is hypothesized to inhibit implant calcification because it imparts a net positive surface charge, thereby repelling calcium ions. Polyacrylamide treatment was thought to prevent phospholipid penetration into valve tissue by forming a hydrogel within the connective tissue. However, this protective effect was lost if the treated bioprostheses were mechanically cycled before implantation of their leaflet.21
Thus far, there has been no confirmation that efficacy in preventing BPHV calcification in animal model studies translates to clinical success. A number of concerns in this area have not been addressed. These include factors such as the markedly different lipoprotein profile in human subjects compared with sheep and rats, marked differences in calcium metabolism between experimental animals and humans, coexisting cardiovascular or valvular disease, and marked differences in coagulation status between sheep and humans. Nevertheless, the microscopic and ultrastructural pathologies of BPHV calcification are comparable to clinical material and animal model explants.5 9 10 Therefore, it is reasonable to proceed with clinical studies of ethanol-pretreated leaflets on the basis of the results reported in these experiments.
Synergistic strategies have been reported using other anticalcification approaches, such as bisphosphonates combined with metal ions as described above.30 The potential synergy of ethanol with other anticalcification approaches is unknown. Because the ethanol mechanism is specifically based on protein-lipid interactions, it is highly likely that synergy could be expected if the ethanol approach were combined with a specific calcium phosphate crystal antagonism strategy, such as that involved with bisphosphonates. The potential therapeutic benefits of dietary ethanol ingestion by BPHV recipients, combined with ethanol pretreatment, are also unclear. These considerations will be the subject of further investigations.
Because ethanol pretreatment effectively inhibited calcification of glutaraldehyde-pretreated porcine aortic leaflets in both subdermal implants and mitral valve replacements, this approach warrants examination for its clinical effectiveness. On the basis of the results of our FTIR studies and lipid extraction assays, the mechanism for ethanol inhibition of calcification may be due to an interaction of structural protein conformation changes, lipid extraction, and altered water status. Finally, ethanol pretreatment resulted in no significant alterations in BPHV leaflet material characteristics and morphology as assessed by shrinkage temperature and light and electron microscopy. This suggests that ethanol pretreatment would hypothetically not affect BPHV mechanics, although this remains to be investigated.
Selected Abbreviations and Acronyms
|APD||=||amino propyl hydroxybisphosphonate|
|BPHV||=||bioprosthetic heart valve|
|DSC||=||differential scanning calorimetry|
|FTIR||=||Fourier-transform infrared spectroscopy|
This work was supported in part by grants from the NIH (HL-38118) and St Jude Medical, Inc. We thank Ardith Bates for her assistance in preparing the manuscript. We also thank William Mirsch of St Jude, Inc, for his expert advice in planning the circulatory studies.
- Received June 13, 1996.
- Revision received October 14, 1996.
- Accepted October 21, 1996.
- Copyright © 1997 by American Heart Association
Folch J, Lees M, Sloane-Stanley GH. A simple method for the isolation and purification of total lipids from animal tissues. J Biol Chem. 1957;226:497-509.
Gamble W, Vaughan M, Kruth HS, Avigan J. Procedure for determination of free and total cholesterol in micro- or nanogram amounts suitable for studies with cultured cells. J Lipid Res. 1978;19:1068-1070.
Bartlett GR. Phosphorus assay in column chromatography. J Biol Chem. 1959;234:466-468.
Ramp WK, Demaree DN. Inhibition of net calcium efflux from bone by ethanol in vitro. Am J Physiol. 1984;246:C30-C36.
Baraga J, Feld MS, Rava RP. In situ optical histochemistry of human artery using near infrared Fourier transform Raman spectroscopy. Proc Natl Acad Sci U S A. 1992;89:3473-3477.
Jakobsen RJ, Brown LL, Hutson TB, Fink DJ, Veis A. Intermolecular interactions in collagen self-assembly as revealed by Fourier transform infrared spectroscopy. Science. 1983;220:1288-1290.
Chen W, Schoen FJ, Levy RJ. Mechanism of efficacy of 2-amino oleic acid for inhibition of calcification of glutaraldehyde-pretreated porcine bioprosthetic heart valves. Circulation. 1994;90:323-329.
Jones M, Eidbo EE, Hilbert SL, Ferrans VJ, Clark RE. The effects of anticalcification treatments on bioprosthetic heart valves implanted in sheep. Trans Am Soc Artif Intern Organs. 1988;34:1027-1030.
Hirsch D, Drader J, Pathak YV, Yee R, Schoen FJ, Levy RJ. Synergistic inhibition of the calcification of glutaraldehyde pretreated bovine pericardium in a rat subdermal model by FeCl3 and ethanehydroxydiphosphonate: preincubation and polymeric controlled release studies. Biomaterials. 1993;14:705-711.
Levy RJ, Wolfrum J, Schoen FJ, Hawley MA, Lund SA, Langer R. Inhibition of calcification of bioprosthetic heart valves by local controlled-release diphosphonate. Science. 1985;228:190-192.
Gott JP, Chi P, Dorsey LMA, Jay JL, Jett GK, Schoen FJ, Girardot J-M, Guyton RA. Calcification of porcine valves: a successful new method of antimineralization. Ann Thorac Surg. 1992;53:207-216.