Prevention of calcification in bioprosthetic heart valves ...
Published on: Mar 4, 2016
Transcripts - Prevention of calcification in bioprosthetic heart valves ...
Review General Prevention of calcification in bioprosthetic heart valves: challenges and perspectives 1. Introduction d te bi 2. Current status and presentation Dan T Simionescu hi of bioprosthetic heart valves Cardiovascular Implant Research Laboratory, Department of Bioengineering, Clemson University, 501 Rhodes Research Center, Clemson, SC 29634-0905, USA o pr 3. Macroscopic alterations in explanted bioprosthetic Surgical replacement with artificial devices has revolutionised the care of tly heart valves patients with severe valvular diseases. Mechanical valves are very durable, but c tri 4. Factors that influence require long-term anticoagulation. Bioprosthetic heart valves (BHVs), devices s bioprosthetic heart valve manufactured from glutaraldehyde-fixed animal tissues, do not need long- n calcification and term anticoagulation, but their long-term durability is limited to 15 – 20 years, tio experimental models mainly because of mechanical failure and tissue calcification. Although mech- 5. Glutaraldehyde as a villain and u anisms of BHV calcification are not fully understood, major determinants are b tri treatments targeting Glut glutaraldehyde fixation, presence of devitalised cells and alteration of s di 6. Tissue components influence specific extracellular matrix components. Treatments targeted at the preven- calcification of bioprosthetic tion of calcification include those that target neutralisation of the effects of nd heart valves glutaraldehyde, removal of cells, and modifications of matrix components. 7. Expert opinion and conclusion a Several existing calcification-prevention treatments are in clinical use at g tin present, and there are excellent mid-term clinical follow-up reports available. The purpose of this review is to appraise basic knowledge acquired in the rin field of prevention of BHV calcification, and to provide directions for future P . research and development. d Lt Keywords: bioprosthetic heart valve, calcification, collagen, elastin, glutaraldehyde, mineralisation o ns Opin. Biol. Ther. (2004) 4(12):1971-1985 Expert t i li ca 1. Introduction P ub As we settle into the twenty-first century, cardiovascular diseases continue to impact morbidity and mortality, and maintain their position as the number one killer of the e y civilised world . Valvular pathology is a significant chapter of cardiovascular dis- hl eases, but very little is known about the mechanisms involved in the onset of this As pathology, and there is no medication available to limit its progression. The most f common therapeutic procedure for the treatment of valvular pathology is the surgi- h to cal replacement of defective heart valves. Diseased human heart valves are replaced by ‘engineered’ devices that include mechanical valves or valves made from biologi- ig yr cal tissues. This practice started in the early 1970s and it is estimated at present that ∼ 275,000 valve replacements are performed annually worldwide . MechanicalCop valves represent slightly more than half of these, with the remainder being tissue valves. The most physiological tissue prostheses are the pulmonary autograft valves (the result of a surgical procedure whereby the patient’s own pulmonary valve is transplanted into the aortic position) and the human allograft valves (sterilised, cryopreserved cadaveric valves obtained from humans). Allografts exhibit excellent durability after implantation, but are not readily available and represent only a small percentage of total valve replacements. Heterograft valves (xenografts), fabricated from glutaraldehyde-treated porcine aortic valves or from bovine pericardium, Ashley Publications www.ashley-pub.com represent the largest proportion of biological replacement valves. 10.1517/14712522.214.171.1241 © 2004 Ashley Publications Ltd ISSN 1471-2598 1971
Prevention of calcification in bioprosthetic heart valves: challenges and perspectives Heart valve replacement offers an excellent improvement in each treatment targets different tissue structures and not all the quality of life for thousands of patients. The main issue experimental data are available in peer-reviewed publications. that emerged during clinical investigations of replacement This paper will review basic knowledge acquired in the field of heart valves was limited durability of these devices. Reopera- prevention of BHV calcification and provide some directions tion following valve replacement surgery, for the purpose of for future research and development. retrieving and replacing the defective device, is a relatively common event and occurs within 10 – 15 years after initial 2. Current status and presentation of d te valve surgery . Clearly, a second open-heart surgery is unde- bioprosthetic heart valves bi sirable to the patient and is prone to high clinical risks. Although mechanical valves may last longer, they have a high BHVs come in different designs, shapes and colours, and o hi pr rate of thromboembolism and require an almost indefinite could be largely categorised into those that rely on a support anticoagulant therapy. Without disregarding the excellent structure (stent) for functioning and those that are used in tly results obtained with mechanical valves, the focus of this surgery without stents (stentless). Stented prostheses are con- c tri paper is on the biological tissue valves, most commonly structed from porcine aortic valves or bovine pericardium referred to as bioprosthetic heart valves (BHVs). treated with 0.2 – 0.6% neutral buffered glutaraldehyde, s n Careful analysis of explanted tissue valves has shown that which are then mounted onto supporting structures that tio the predominant aspect contributing to dysfunction of bio- mimic the valve anatomy. Glutaraldehyde, a highly reactive logical heart valve replacements is structural deterioration and u water-soluble dialdehyde, crosslinks tissue proteins by react- b tri calcification of the tissue component. Histological, ultrastruc- ing with available amine groups, and in doing so greatly tural and biochemical aspects of degenerated explanted BHVs reduces the rate of in vivo enzymatic degradation by host cells. s di are similar to those of native human diseased heart valves . Moreover, glutaraldehyde reduces the antigenicity of the tissue nd The major processes that contribute to this ‘new pathology’ and sterilises the prostheses. However, glutaraldehyde fixation of replacement heart valves are tissue calcification and is also a main cause for the lack of long-term durability of a g mechanical damage. Calcification may occur independent of BHVs (see below). tin mechanical damage , but may also be accompanied by tissue The development of BHVs has been a continuing process in rin abrasion, tearing and perforations. Numerous design and the last three decades. The first generation of BHVs were fixed manufacturing improvements have been made to reduce the with glutaraldehyde at high pressure and were not treated with incidence of mechanical damage ; however, it is too early for . P any calcification-prevention agent. These BHVs are highly d Lt a full clinical evaluation and comparison of these new designs prone to structural and calcific degeneration, and are not with earlier models. Moreover, in spite of major design devel- expected to last longer than 15 – 20 years. Existing BHVs uti- o ns opments, it is apparent that calcification remains an important lise very low (or zero)-pressure fixation and incorporate chemi- cause of dysfunction in biological valves. t i cal treatments that have been shown in animal studies to delay ca The unique pathology of BHVs presents serious challenges or prevent calcification. For stented valves, Medtronic (Minne- l i to academic scientists, industry representatives and clinicians apolis, MN, USA) employs sodium dodecyl sulfate (T6), ub alike. These challenges incorporate a convoluted interplay of α-amino oleic acid (AOA®) and toluidine blue (Intact®) as cal- P interests between basic research, commercial benefits and the cification-prevention agents targeted at valvular cusps. Other ey need to provide adequate healthcare. When compared with treatments include ethanol for the Epic® valves (St. Jude Medi- hl other biological therapies, there is a surprising paucity of basic cal, Minneapolis, MN, USA), as well as ethanol and Tween-80 s knowledge on calcification mechanisms in BHVs, as well as (XenoLogiX®) for the Carpentier-Edwards valves (Edwards o fA mechanisms of action of treatments that influence calcifica- tion. This interplay of interests is also reflected by the confus- Lifesciences Corporation, Santa Ana, CA, USA). More recently, stentless porcine valves have gained popular- g ht ing terminology assigned to the efficacy of treatments, as these ity in cardiac surgery and are considered an attractive alterna- i range from the more modest ‘mitigation’, ‘delay’, ‘retardation’ tive to stented valves, mainly because of the absence of yr or ‘reduction’ of calcification to ‘prevention’, ‘calcification- obstructive stents and strut posts. In most models available toCop resistance’, and culminate with ‘complete inhibition’ and ‘anticalcification’. This review will use the term prevention. date, the device is derived from the whole porcine aortic root (valve cusps attached to the native sinus and a portion of the Numerous approaches for the prevention of calcification in ascending aorta), fixed in glutaraldehyde at low pressure and BHVs have been ‘bench’-tested during the last 30 years, but treated with calcification-prevention reagents. The current few of them have been licensed and approved by regulatory generation of stentless BHVs, most of which incorporate agencies and have reached ‘bedside’ or clinical applications. It calcification-prevention treatments, includes the St. Jude is not the purpose of this review to rank or endorse any treat- Medical Toronto SPV® (treated with BiLinx®, ethanol for the ments aimed at prevention of calcification in BHVs. Compar- cusps and AlCl3 for the wall); Medtronic Freestyle Aortic Root ison between different treatments at this point is not realistic Bioprosthesis (AOA); Edwards Lifesciences Prima Plus because most of them have ill-defined mechanisms of action (XenoLogiX); CryoLife-O’Brien stentless porcine aortic valve and as yet unproven long-term clinical efficacy. Moreover, (without calcification-prevention treatment); Shelhigh 1972 Expert Opin. Biol. Ther. (2004) 4(12)
Simionescu implant composition. It is well known that host factors, such as age of recipient and altered calcium metabolism, are major determinants, as BHV calcification occurs at a highly accelerated rate in children and growing adults, as well as in patients with renal failure . This was explained by the pro-calcific metabolism of younger patients, as compared with adults, which includes higher levels of blood calcium, d te phosphate, bone proteins, and enhanced parathyroid bi hormone and vitamin D metabolism . In addition, a hi recent retrospective study showed that hypercholesterolaemia o pr could also be a risk factor for BHV calcification . As BHVs calcify less after being implanted in the right heart as tly Figure 1. BHV pathology. A) First-generation stented pericardial BHV, not implanted. Top view of the open valve outflow surface compared with the left heart, it is believed that mechanical c tri showing the three pericardial leaflets sutured to the support. B) stress can also influence BHV calcification. In addition, within each BHV cusp, calcification appears to be concen- s Pericardial BHV, explanted after 7 years of intracardiac n functioning, showing abrasions at the cusps base (1), commissural trated in areas of high stress . Improvements in valve designs tio tears with leaflet prolapse (2) and nodular calcification (3). that reduce valvular stress, such as flexible stents, are believed u BHV: Bioprosthetic heart valve. to reduce calcification, but definitive proof is pending future b tri clinical studies. No-React Stentless Bioprosthesis Biocor PSB/St. Jude Medical There are numerous in vitro studies showing that calcifica- s di (No-React, detoxification, surfactant/heparin). Although tion in BHVs can occur through passive mechanisms [13-20]. nd promising, follow-up durability data for these new models are However, due to inherent drawbacks of these models, more limited to only 7 – 8 years in the clinical setting, and there is relevant animal models have been developed for the study of a g conflicting evidence that the use of stentless valves results in BHV calcification, the most popular being subcutaneous tin improved clinical performance over stented bioprostheses . implantation in rats or mice, and valve replacement in sheep rin The hopes and expectations are that the low-pressure tissue or calves . Both models attain calcification of BHV tissues fixation and the addition of calcification-prevention agents in a relatively short interval (several weeks in subdermal will affect durability favourably, but many more years of . P implants and several months in intracardiac implants), and d Lt detailed clinical follow-up are required to ascertain these the morphology of calcified deposits closely resembles those expectations fully. found in calcified BHVs explanted from patients. The sub- o ns dermal model has the advantage of relative ease of use and low 3. Macroscopic alterations in explanted ti cost, but does not involve direct exposure of BHV tissue to ca bioprosthetic heart valves flowing blood and haemodynamic stress. Conversely, intracar- li diac implants in large animals more closely mimic human ub A wide variety of changes are noticeable in explanted BHVs, as implants, but are more technically demanding and very P outlined in Figure 1: abrasions at the cusps base, commissural expensive. Rat subdermal implants have been used as screen- ey tears with leaflet prolapse, and calcification. ing tests for calcification-prevention treatments of BHVs, and hl Throughout the years, attempts have been made at improv- in most studies, treatments efficient in the subdermal model s ing BHV performance at each of these levels. Implementing less also reduced calcification in circulatory implants. fA traumatic techniques for tissue mounting, the use of flexible Although animal models provide a plethora of information o stents, the covering of stents with a thin layer of pericardium about kinetics and factors involved in calcification, we still do g ht and the introduction of stentless valves [6,7] have greatly reduced not fully understand the mechanisms underlying calcification i the incidence of tears and abrasions in newer generations of of BHVs in patients. The role of host cells in BHV calcifica- yr BHVs . Despite excellent design development, structural valve tion is still a matter of debate . Prevention of cell infiltra-Cop deterioration and calcification are still considered the main cause of BHVs replacement , and may possibly represent a tion in subdermally implanted BHV cusps (by enclosing tissues in Millipore chambers before implantation) did not threat to the new generation of BHVs. Due to its great clinical affect calcification in animal models, suggesting a passive, impact, ample research efforts have been directed towards rather than cell-mediated, process . Moreover, graft rejec- understanding and reducing degenerative calcification of BHVs. tion is not thought to play a major role in bioprosthetic valve mineralisation because BHV tissues implanted in nude mice 4. Factorsthat influence bioprosthetic heart (T cell-deficient) calcify to the same degree as implants in valve calcification and experimental models normal mice . However, evidence also exists for the role of host reactions in BHV calcification. These include recent The major factors involved in BHV calcification are host- reports showing the involvement of low-level immune reac- related factors, mechanical stress, chemical treatment and tions [25-27] and the fact that inhibition of local proteases Expert Opin. Biol. Ther. (2004) 4(12) 1973
Prevention of calcification in bioprosthetic heart valves: challenges and perspectives reduces calcification . Without disregarding the importance • dye-mediated photo fixation [35-37] (PhotoFix®, Carbomed- of the above-mentioned factors, two determinants of BHV cal- ics, Austin, TX), which crosslinks proteins at the aromatic cification have received the utmost attention: glutaraldehyde amino acids level using a mild oxidation reaction catalysed fixation and tissue composition. by a light-sensitive dye • carbodiimide-based fixation  (Ultifix®, Medtronic, Min- 5. Glutaraldehyde as a villain and treatments neapolis, MN), which involves carboxyl group activation with carbodiimides and crosslinking to free amine groups targeting Glut d te bi Other non-glutaraldehyde fixation procedures include the hi Glutaraldehyde crosslinking of connective tissues, a process use of epoxy , diphenylphosphorylazide , acyl azides, used to render animal tissues inert, non-biodegradable and o cyanamide  or diisocyanates . In addition, physical pr non-antigenic, has revolutionised the surgical treatment of methods, such as ultraviolet light  and dehydration , valvular disease by providing an excellent alternative to tly were proposed for crosslinking of BHV tissues. Tissues mechanical valves and promising long-term use without the c treated with most of the above-mentioned processes exhibit tri need for lifelong anticoagulation. Paradoxically, glutaralde- good calcification-prevention properties, when compared hyde fixation also encourages calcification. The ability of glu- s with glutaraldehyde. It is assumed that these approaches are taraldehyde to induce BHV calcification has been clearly n not inducing calcification simply because they do not employ tio demonstrated by numerous experiments and has achieved a the use of glutaraldehyde. Moreover, it is not apparent ‘villain’ status . Fresh, non-crosslinked tissues are degraded bu whether cells and matrix changes, which have been described tri and resorbed rapidly after implantation, without calcification. as being responsible for calcification of BHV components, This suggests that tissues incapable of remodelling will ulti- s di are effectively being prevented using these non-glutaralde- mately calcify, and that changes induced by glutaraldehyde are hyde crosslinking procedures. For example, it is plausible to nd responsible for tissue calcification. hypothesise that all of the non-glutaraldehyde crosslinking In normal cusps, valvular interstitial cells and resident a approaches will eventually devitalise cells, but it is difficult to macrophages actively participate in maintaining tissue archi- g imagine that such cells will not undergo calcification. None tin tecture via constant remodelling and scavenging . It is of these alternative crosslinking methods are in clinical use at rin known that scavenging of cell debris, a normal physiological present. These studies suggest that the chemical nature of P tissue function, determines the fate of connective tissues. In glutaraldehyde may induce calcification, and that alternatives tissues where scavenging mechanisms are limited (such as d . to glutaraldehyde fixation warrant further investigation. Lt chemically fixed tissues), cell debris accumulate calcium ns rapidly and mediate calcification of adjacent collagen fibres. 5.2 Calcification-prevention treatments targeted The presence of macrophages in calcified atherosclerotic io at glutaraldehyde t plaques , calcified aortic aneurysms  and implanted Besides the actual glutaraldehyde crosslinks, it is apparent that ca BHVs  is good evidence for the body’s attempt to rescue i residual, loosely bound glutaraldehyde or unreacted aldehyde l and scavenge calcifying tissues; however, the amplitude of ub groups are also involved in BHV calcification. Neutralisation calcification apparently overpowers these scavenging efforts. of free aldehyde groups with compounds that possess reactive P Biomaterials derived from chemically stabilised connective ey primary amines and rinsing residual glutaraldehyde has been tissues probably calcify because of the lack of remodelling hl shown to reduce calcification in animal models. Several and scavenging. Besides preventing tissue remodelling, glu- s amino acids are included in this category, such as glutamine, fA taraldehyde fixation directly induces cell death and formation glycine, homocysteic acid [44,45] and lysine [46-49]. AOA, a of cellular debris, which in turn can serve as foci for calcifica- o potent calcification-prevention agent for cusp and pericardial ht tion (see below). Incomplete glutaraldehyde binding to tissue tissues , may be active as a glutaraldehyde-neutralising proteins also yields cytotoxic aldehyde group residuals, which i g agent, but it may also influence calcification by mechanisms yr could also induce calcification. Therefore, prevention of cal- related to the reduction of calcium diffusion through op cification was approached by using non-aldehyde crosslink- tissues . Similarly, other compounds described as calcifica- ers, neutralisation of free aldehyde groups and detoxification tion-prevention treatments have primary amine groups, suchC of glutaraldehyde residuals. as amino-biphosphonates  and toluidine blue , but it is not clear whether their effect is related to aldehyde neutralisa- 5.1 Calcification-prevention by using tion alone. Biphosphonates, for example, are also believed to non-glutaraldehyde fixation chemistry act as crystal poisons  by binding to preformed hydroxyap- In support of the glutaraldehyde hypothesis, it is remarkable atite and preventing further increase in crystal size . Ura- that tissue crosslinking and stabilisation without the use of glu- zole, a secondary amine compound [47,56], and possibly the taraldehyde (non-glutaraldehyde fixation) does not induce tis- ‘No-React’ treatment (St. Jude proprietary detoxification sue calcification in animal models . Alternative crosslinking process), may reduce calcification  by efficiently rinsing chemistries include: residual glutaraldehyde molecules from tissues. 1974 Expert Opin. Biol. Ther. (2004) 4(12)
Simionescu 6. Tissue components influence calcification of buffered glutaraldehyde. Long exposure to such low dilutions bioprosthetic heart valves of glutaraldehyde induces adequate (but not optimal) preser- vation of collagenous structures, but also produces severe cell The main tissues used for the construction of BHVs are valve alterations, which result in formation of cell debris that cusps, aortic wall and pericardium. All of these tissues are resemble matrix vesicles formed by bone cells [68,69]. The composed of specific cells embedded in an extracellular matrix degree of ultrastructural integrity of cells in BHVs correlates composed of varying proportions of collagen, elastin, non- well with their ability to nucleate calcification [70,71]. Figure 2 d te collagenous proteins and glycosaminoglycans. For example, depicts a segment of a cell membrane, putative component of bi cusp tissue is composed of ∼ 40% collagen, ∼ 4% gly- a matrix vesicle. Transmembrane proteins are embedded in a cosaminoglycans and a small amount of elastin. The adjacent aortic wall, however, is made of only ∼ 15% collagen, ∼ 50% double-layered lipid membrane, which exposes towards the o hi pr internal aspect, a high number of acidic phospholipids such as elastin and small levels of glycosaminoglycans. Pericardial tis- phosphatidyl serine. Passive transporters and ATP-mediated tly sue is almost 90% collagen. It is of interest that after glutaral- active pumps maintain normal calcium homeostasis, which c tri dehyde fixation, these tissues calcify to different extents in sustains a 10,000-fold calcium ion concentration gradient in animal models of calcification . Moreover, calcium deposi- the external compartment (EXT), as compared to the cyto- s n tion in fixed tissues can occur on different components. Devi- plasm. Calcium accumulation in matrix vesicles occurs by ill- tio talised cells are involved in calcification of all glutaraldehyde defined mechanisms, which probably involve alterations in fixed tissues used in manufacturing of BHVs ; collagen also u calcium homeostasis and massive calcium influx . Calcium b tri calcifies readily in cusp and pericardium , but elastin calci- ions may concentrate on the interior aspect of the membrane fies more severely in implants that contain portions of aortic by binding to acidic phospholipids and calcium-binding pro- s di wall . Some of these components are natural substrates of teins. Inorganic phosphate ions, produced by phosphatases, nd calcification in normal bone physiology, whereas others are combine with calcium ions immobilised onto the cell mem- only involved in pathological calcification. brane. A continuous influx of calcium ions leads to the forma- a g During bone formation and mineralisation, bone cells tion of hydroxyapatite. Calcification of cell debris in BHVs tin undergo apoptosis, release matrix vesicles that accumulate may possibly take place via these same mechanisms. rin calcium and promote collagen calcification . In vascular Some of the main calcification-prevention treatments aim pathology, it is well known that medial aortic calcinosis (cal- to extract cell lipids, thus removing potential sites of calcifica- . cification of elastic fibres in the media) is mainly associated P tion. These treatments include the use of organic solvents, d Lt with elastin, and may also involve newly formed bone such as ethanol  and chloroform/methanol , and deter- cells . Glycosaminoglycans and non-collagenous proteins gents, such as sodium dodecyl sulfate , Tween 80  and o ns seem to be involved mainly in the regulation of calcification Triton X-100 . It is therefore possible that tissues from t i processes , but these roles are still disputed. In normal which cell debris has been removed would benefit from a ca bone formation, calcium accumulation is preceded by enzy- delay in initiation of calcification. However, it is not evident l i matic degradation of proteoglycans (PGs) , suggesting that to what extent these treatments effectively reduce calcification ub they serve as a role in inhibiting and regulating the progres- of collagen or elastin, or whether they can prevent readsorp- P sion of calcification. Other regulatory proteins, such as tion of lipoproteins from blood. Moreover, there is a good ey matrix γ-carboxyglutamic acid (Gla) protein  and oste- deal of evidence for host cell infiltration in implanted hl opontin , may play a role in BHV calcification, but their BHVs [33,61,77,78]. These cells may be entrapped in the glutar- s role is still under investigation. Overall, these data suggest aldehyde-fixed matrix and eventually die, to possibly induce o fA that most matrix components present in tissues used for fab- rication of BHVs are capable of calcifying, possibly each com- late calcification of BHVs. ht ponent having different affinity for calcium and kinetics of g 6.2 Calcification of collagen and elastin – mechanisms i calcium accumulation. Due to the vital role of cells and and prevention yr matrix components, prevention of BHV calcification was Collagen fibres are present in all tissues used for BHV and pro-Cop approached by attempts to remove or extract cells, by struc- tural modification of collagen and elastin, and by stabilisation vide the architectural framework and mechanical strength required for function in a BHV. Collagen molecules (Figure 3) or addition of natural calcification inhibitors. are staggered within the collagen fibre, creating an apparently empty area, also known as a ‘hole zone’, in which PGs reside , 6.1 Calcification of glutaraldehyde fixed cells – presumably protecting collagen from calcification. This same mechanisms and prevention area has been shown to be the initial site of calcium deposition Glutaraldehyde was originally introduced as a fixative for onto bone collagen fibres [80,81]. Matrix metalloproteases transmission electron microscopy (TEM) > 35 years ago  (MMPs)  and glycosaminoglycans-degrading enzymes  can and is still in use for ultrastructural studies. Whereas high glu- degrade PGs and expose sites prone to calcification. Progression taraldehyde concentrations (2 – 3%) are used for TEM, for of collagen calcification can lead to the formation of hydroxyap- BHV preparation, tissues are treated with 0.2 – 0.6% neutral atite, which, in turn, can grow into large deposits. Expert Opin. Biol. Ther. (2004) 4(12) 1975
Prevention of calcification in bioprosthetic heart valves: challenges and perspectives EXT d te bi o hi INT pr tly 1 c tri 2 3 ATP s on 4 i ut A Ca2+ PO23- Acidic phospholipids Hydroxyapatite b s tri di a nd g tin P rin d . Lt o ns ti l i ca P ub ey B C D s hl fA Figure 2. Cell calcification. A) Schematic diagram showing hypothetical events associated with cell calcification. See details in section o ht 6.1. B) TEM picture showing a calcified cell in a 21-day subdermal implant of glutaraldehyde-fixed pericardium. Calcium deposits appear as dark needle-shaped crystals. C) and D) show higher magnifications of cell debris and matrix vesicle calcification. Original i g magnification, B) 30,000×, C) and D) 70,000×. yr ATP: Phosphatase substrate; Ca: Calcium; EXT: External compartment; INT: Internal aspect; PO: Inorganic phosphate; TEM: Transmission electron microscopy. opEthanol pretreatment of glutaraldehyde-fixed porcine aorticC valve cusps and bovine pericardium significantly reduces equally effective in preventing calcification of the BHV aortic   wall. This may reflect the relatively small contribution of colla- calcification in experimental models. This anticalcification gen to wall calcification, as compared with elastic fibres, which mechanism is hypothesised to be related to the extraction of are not affected by ethanol treatments . cholesterol and phospholipids, to permanent alterations in col- Elastic fibres are abundant structural proteins in the vascu- lagen structure , and to a putative interaction with glutaral- lar wall and also functional components of aortic cusps . dehyde . However, this effect is not limited to ethanol alone. There is ample evidence showing that degeneration and calci- Similar results were obtained with ether, methanol, chloro- fication of elastic fibres occurs in BHVs, as well as in human form/methanol and isopropanol . Despite complete removal allografts and pathological vascular calcification [31,86-90]. of lipids and alterations in collagen structure, ethanol is not Ultrastructurally, elastic fibres consist of a core composed of 1976 Expert Opin. Biol. Ther. (2004) 4(12)
Simionescu d te bi PG hi 2 1 o pr c tly s tri n u tio b Hole zone stri di 3 A a nd g tin P rin d . Lt o ns t i l ica P ub ey hl B C s o fA ht Figure 3. Collagen calcification. A) Schematic diagram showing hypothetical events associated with collagen calcification. See details in section 6.2. B) TEM picture showing a cross-section through a calcified collagen fibre in a 21-day subdermal implant of i g glutaraldehyde-fixed pericardium. Collagen fibres appear as round white structures, and calcium deposits as dark needle-shaped crystals. yr C) Shows a higher magnification of a single collagen fibre sectioned longitudinally. Original magnification, B) 30,000× and C)70,000×. op PG: Proteoglycan; TEM: Transmission electron microscopy.C crosslinked elastin molecules and a microfibrillar component (Figure 4) involves degradation of the protective coating con- located mainly around the periphery of the amorphous com- taining elastin-associated microfibrils, PGs, and exposure of ponent . The external coating, rich in acidic components the elastin core to the calcium-rich extracellular matrix milieu. such as fibrillin, is believed to protect elastin from calcifica- Consequently, elastin binds calcium ions (circles) followed by tion [92,93]. Disruption of elastin fibre integrity has been inorganic phosphate ions, leading to the formation of implicated in the initiation and progression of aortic calcifica- hydroxyapatite on the surface of the elastic fibre. tion, and ultrastructural studies have shown that in BHVs, Elastin has particular properties that set it apart from this protective coating is not fully preserved . Thus, the other matrix components. Glutaraldehyde does not chemi- hypothesised pathway by which elastin calcifies in BHVs cally react with elastin because the mature protein does not Expert Opin. Biol. Ther. (2004) 4(12) 1977
Prevention of calcification in bioprosthetic heart valves: challenges and perspectives EAMF and PG d te bi o hi pr tly EL c s tri n tio A bu s tri di a nd g tin P rin d . B Lt C o ns t i ca Figure 4. Elastin calcification. A) Schematic showing hypothetical events associated with calcification of elastic fibres. See details in l i section 6.2. B) TEM picture showing elastin-associated calcification in a 21-day subdermal implant of glutaraldehyde-fixed porcine aorta. ub Calcium deposits appear as dark needle-shaped crystals. C) Shows a higher magnification of the area outlined by dashed lines in B). P Original magnification, B) 30,000× and C) 65,000×. ey EAMF: Elastin-associated microfibrils; EL: Elastin; PG: Proteoglycans; TEM: Transmission electron microscopy. s hl possess sufficient reactive amine groups . Paradoxically, With the advent of stentless BHVs that expose large por- fA glutaraldehyde fixation is not required for elastin to tions of aortic wall to flowing blood, more effort has been to calcify . Elastin has been shown to possess a unique struc- directed towards calcification-prevention treatments targeted h ture that facilitates calcium binding . Moreover, the lack at elastin calcification. One of the most studied treatments in ig of reactivity of glutaraldehyde with elastin also implies that this category is the use of aluminium ions. This treatment is yr elastin is not protected against the activity of degradative part of the BiLinx calcification-prevention process (St. JudeCop enzymes, such as MMPs . The consequences of this lack of fixation are vital, as progressive degradation of elastin may Medical), under clinical evaluation at present. Aluminium ions bind strongly to elastin and induce conformational further aggravate calcification and sustain alterations of changes in the elastin structure, which reduces the affinity of mechanical properties of cusp  and aortic wall segments elastin towards calcium ions . Aluminium ions also partially of BHVs. Elastin degradation products could possibly elicit stabilise pure elastin against the action of degrading enzymes, immune responses and trigger unwanted reactions in host and protect it from degeneration and calcification when tested cells such as protease release and apoptosis [99-101]. In addi- in animal models . However, binding of aluminium to tion, there seems to be a direct correlation between elastin elastin does not seem to be completely irreversible , and it degradation and calcification , and local delivery of is also not yet clear what the effects of aluminium on cell- and MMP inhibitors significantly reduced calcification of collagen-mediated calcification are. In an attempt to improve subdermally implanted elastin . elastin stabilisation further, tannic acid, a polyphenol belonging 1978 Expert Opin. Biol. Ther. (2004) 4(12)
Simionescu to the galloylglucose family, was recently shown to bind to BHVs after accelerated fatigue testing, as well as after aortic elastin, resulting in an improved resistance to elastase implantation in the sheep circulatory model . Overall, and a significant reduction in calcification of glutaraldehyde- these data showed that glutaraldehyde does not stabilise GAGs. treated aorta in the rat subdermal model . Taken together, As GAGs within BHVs play an important role in maintaining these results support the hypothesis that protection of elastin proper mechanical functions, loss of GAGs may contribute from degradation has the potential to extend BHV durability significantly to degeneration of BHVs. Moreover, GAGs are in the clinical setting. hypothesised to prevent calcium deposition by chelating d te calcium and preventing hydroxyapatite nucleation . The bi 6.3 Role of non-collagenous components mechanisms responsible for GAG loss are not fully known, Non-collagenous components play important roles in calcium but it is possible that GAGs, which are not crosslinked by o hi pr homeostasis within the extracellular matrix. Bone sialopro- glutaraldehyde, may be susceptible to GAG-degrading tein, a bone protein, and bone morphogenetic protein-2, a enzymes present in cusp tissues . Covalent binding of tly member of the transforming growth factor cytokine super- GAGs, such as heparin, was shown to reduce calcification of c tri family, are known to participate in the regulation of bone BHVs in experimental models . development and maturation . Both of these proteins have Very few attempts have been made to stabilise GAGs in s n also been found in calcific aortic stenosis, indicating that BHVs. As glutaraldehyde is not effective, periodate oxidation tio valvular calcification occurs via mechanisms similar to bone was used for crosslinking of valvular GAGs. This is based on formation . Other bone-associated proteins, such as oste- u the relative specificity of periodate to oxidise geminal diols b tri opontin, osteocalcin and osteonectin, have also been demon- contained within the structure of GAGs, allowing the forma- strated in calcifying atheromatous plaques and in tion of aldehyde groups within the GAG chain . These s di experimentally induced arterial calcification; however, it is aldehyde groups would then react with amine groups found nd believed that these proteins may play a significant role in the in the lysine residues of collagen molecules, achieving GAG regulation (and not initiation) of calcification . In support fixation. Model studies using mixtures of hyaluronic acid and a g of this hypothesis, osteocalcin-deficient mice showed increased pure type I collagen showed that periodate oxidation induces tin bone formation without impairing bone resorption . GAG immobilisation onto collagen fibres. Moreover, when rin Non-collagenous components have also been shown to associ- implanted subdermally in juvenile rats, periodate-pre- ate with BHV calcification . Interestingly, osteopontin treated, glutaraldehyde-fixed cusps showed 40% less calcifi- (but not osteocalcin, bone sialoprotein or osteonectin) was . P cation as compared with glutaraldehyde alone . It is not d Lt found to be present in calcified, explanted porcine BHVs . known how various calcification-prevention treatments Moreover, glutaraldehyde-fixed cusps showed enhanced calcifi- affect the stability of GAGs. Taken together, these studies o ns cation when implanted subdermally in osteopontin-null mice, showed that GAGs are important functional components of ti strongly suggesting a regulating role for osteopontin . BHV tissues and that new methods of GAG stabilisation ca Overall, these studies point to the involvement of non-colla- deserve further investigation. li genous proteins in BHV calcification, and their application as ub a calcification-prevention strategy merits further investigation. 7. Expert opinion and conclusion P y and proteoglycans unbranched 6.4 Glycosaminoglycans Glycosaminoglycanse (GAGs) are acidic, The ‘ideal’ BHV has to fulfil a number of prerequisites that hl relate mostly to clinical durability and safety. It must be sturdy polysaccharidesspresent in all vascular connective tissues, enough to last the lifetime of a young patient (≥ 50 years), A charidesof bound to core proteins (thus forming PGs) as including valves, aorta and pericardium . These polysac-  simple to implant, widely available, and the design must not are allow regurgitant backflow. In addition, it should have a low large t h polymers that lack protein cores, such as hyaluronic incidence of thromboembolism, be fully resistant to proteoly- ig tissues, and acid. GAGs are distributed in the ground matrix of cusp and sis, be non-antigenic and non-toxic. Such a device does not yrwall and elastic are also associated with the from within the p gen surfaces of colla- exist at present.Co cusp structure hasfrom congenitalindefects, rheumatic from fibres . Loss of GAGs  Seemingly, BHV calcification after implantation is a highly been noted valves obtained unregulated process. Calcium deposition, like immunity and patients suffering fever coagulation, could be considered an acceptable protective mech- and old age, all of which are associated with valve failure . anism, with devastating effects when released from homeostatic Furthermore, GAGs are also completely lost from implanted control. In support of this hypothesis, it is noteworthy that nor- BHVs , suggesting an insufficient stabilisation by glutar- mal bones, in which calcification is highly regulated, contain a aldehyde. This is not surprising, because GAGs lack the maximum of 50% mineral content, whereas highly calcified amine functionalities necessary for crosslinking by alde- BHVs can reach > 90% mineral content [64,117]. hydes. Experimental studies have shown that GAGs are lost As described in this paper, numerous approaches for pre- during glutaraldehyde fixation of porcine cusps and vention of calcification in BHVs have been described in the walls . In addition, > 80% of tissue GAGs were lost from literature during the last 30 years, but only a few of them Expert Opin. Biol. Ther. (2004) 4(12) 1979
Prevention of calcification in bioprosthetic heart valves: challenges and perspectives Table 1. Calcification-prevention treatments: hypothetical targets and mechanisms of action. Target Treatment Mechanisms Glutaraldehyde PhotoFix, carbodiimides Non-glutaraldehyde fixation UV light, dehydration, cyanamide, acylazide, epoxy Amino-oleic acid Neutralisation and detoxification Amino-biphosphonates d te Toluidine blue bi Lysine, homocysteic acid hi Propylene glycol Urazole o Cell debris Ethanol Removal of cellular lipids pr tly Sodium dodecyl sulfate c Tween 80. tri Triton X-100 Collagen Ethanol Structural modification s n tio Elastin Aluminium chloride Structural modification Tannic acid Stabilisation b u tri Other matrix components Osteopontin Addition or preservation of calcification inhibitors Heparin, glycosaminoglycans s di Periodate Matrix enzymes MMP inhibitors Block degeneration of matrix components Hydroxyapatite growth Biphosphonates Amino-biphosphonates nd Inhibit crystal growth a Crystal poison g Block tissue transport of calcium tin Calcium transport Amino-oleic acid MMP: Matrix metalloprotease. n riWe should not overestimate the ability of glutaraldehyde to have been studied in detail from a mechanistic point of . P d Lt view. Table 1 shows a summary of hypothetical targets and equally stabilise all matrix molecules. Although glutaralde- mechanisms of action of these treatments. Some of these hyde efficiently stabilises collagen, there is ample evidence statements are backed up by adequate mechanistic data o ns that glutaraldehyde does not crosslink or stabilise elastin and (ethanol, aluminium, AOA, biphosphonates), while the ti glycosaminoglycans. The consequences of this lack of stabili- ca mechanisms of others are rather speculative. sation may be elastin degeneration and calcification, and pro- li There are numerous unresolved fundamental issues that gressive loss of glycosaminoglycans. We should also not ub deserve further exploration. First, we do not understand pre- underestimate the power of matrix-degrading enzymes P cisely what mechanisms prevent cardiovascular tissues (valves, (MMPs and glycosaminoglycan-degrading enzymes), whose ey aorta, pericardium) from calcifying under physiological con- activity in tissues is reduced by glutaraldehyde, but not com- hl ditions. Furthermore, the mechanisms by which these same pletely inhibited. Such enzymes can contribute to elastin s tissues calcify in pathology elude us. This knowledge is imper- degradation and calcification, collagen weakening and fA ative because the better we understand mechanisms of patho- removal of calcification inhibitors. With few exceptions, the o logical calcification, the easier it will be to devise calcification- role of natural inhibitors of calcification has not been fully g ht prevention treatments for BHVs. investigated and warrants further studies. i Chemical stabilisation of animal tissues is a requirement Another matter of controversy is the relevance of animal yr for the clinical use of BHVs. Despite the fact that glutaralde- models for the study of BHV calcification. Subdermal op hyde pretreatment induces calcification, there is no convinc- implants in juvenile rats and intracirculatory implants inC ing evidence yet for a better alternative. While waiting for substitute chemical crosslinkers to be made available, we larger animals may serve as screening tests for calcification- prevention treatments, but we cannot fully anticipate a clini- should maintain glutaraldehyde as the fixative of choice and cal outcome using these models alone. Furthermore, despite try to develop better treatments to prevent calcification. In numerous studies supporting the role of glutaraldehyde, cells this respect, we need to invest more effort in understanding and matrix components in the calcification of BHVs, we do how glutaraldehyde changes the properties of cells and matrix not yet have all the information to evaluate how these factors components. Because resident cells cannot renew matrix interact with host-related factors and mechanical stress. composition in BHVs, the fate and durability of BHVs As described above, most matrix components may calcify mainly depend on the mechanical and biochemical longevity independently in various physiological and pathological situa- of each component. tions, but these components exhibit different calcium-binding 1980 Expert Opin. Biol. Ther. (2004) 4(12)