Embed Size (px)
Citation preview
Biomaterials 25 (2004) 3359–3368
ARTICLE IN PRESS
*Correspondin
E-mail addres
0142-9612/$ - see
doi:10.1016/j.bio
Effects of periodate and chondroitin 4-sulfate on proteoglycanstabilization of ostrich pericardium. Inhibition of calcification in
subcutaneous implants in rats
Beatriz Arenaza, Marian Mart!ın Maestroa, Pilar Fern!andezb, Javier Turnayc, Nieves Olmoc,Jes !us Sen!enb, Javier Gil Murd, Mar!ıa Antonia Lizarbec, Eduardo Jorge-Herreroa,*
aServicio de Cirug!ıa Experimental, Unidad de Biomateriales, Cl!ınica Puerta de Hierro, San Mart!ın de Porres 4, Madrid 28035, SpainbDepartamento de Ciencias Anal!ıticas, Facultad de Ciencias, UNED, Madrid, Spain
cDepartamento de Bioqu!ımica y Biolog!ıa Molecular, Facultad de Ciencias Qu!ımicas, Universidad Complutense, Madrid, SpaindCREB, Departamento de Ciencia de los Materiales e Ingenier!ıa Metal !urgica, Universidad Polit!ecnica de Cataluna, Barcelona, Spain
Received 11 May 2003; accepted 22 September 2003
Abstract
Chemical modification of biological materials used in the manufacture of cardiac valves tends to reduce the relatively high degree
of biodegradation and calcification of the implanted bioprostheses. The most widely used treatment to reduce biodegradability of
the valves is glutaraldehyde fixation. However, this treatment is potentially toxic and induces tissue calcification. In order to
minimize these undesirable effects, we have analyzed the effect of a pre-fixation of endogenous proteoglycans and exogenous
glycosaminoglycans, as well as the borohydride reduction influence on the different modified ostrich pericardium implants after
subcutaneous implantation in rats. The presence of calcific deposits was detected in all implanted GA-fixed samples; however,
calcification was highly reduced in both groups of periodate-prefixed materials, which showed also a very low Ca/P molar ratio.
Borohydride post-treatment of these biomaterials resulted in a significant increase in calcium phosphate precipitation, with the
appearance of calcium deposits mainly in an amorphous form even though X-ray diffraction allowed the detection of brushite- and
apatite-like crystals. Regarding tissue stability, no significant differences were found among the borohydride-untreated implants but
higher levels of matrix metalloproteinases were observed by gelatin zymography in the periodate pre-fixed materials. This increase
was partially reduced by pre-fixation of exogenous chondroitin 4-sulfate. On the other hand, borohydride post-treatment not only
increased calcification, but also reduced tissue stability and increased the presence of matrix-degrading activities.
r 2003 Elsevier Ltd. All rights reserved.
Keywords: Bioprostheses; Borohydride; Glutaraldehyde; Glycosaminoglycans
1. Introduction
Bioprosthetic heart valves derived from glutaralde-hyde-crosslinked biological tissues are used in heartvalve replacement mainly in elderly patients [1,2]. Thistype of bioprostheses presents an advantage over themechanical ones as they do not need chronic antic-oagulant therapy. Thus, they could be the first choice forheart valve replacement surgery. However, these deviceshave a shorter implanted lifetime than mechanical valvesbecause they present more easily pathologic calcificationand structural degeneration [3]. Some chemical or
g author.
s: [email protected] (E. Jorge-Herrero).
front matter r 2003 Elsevier Ltd. All rights reserved.
materials.2003.09.105
physical treatments have been developed in order tocontrol these processes and to improve the mechanicaland immunogenic properties of this biological materials.Crosslinking of the biological tissues has been the mainprocedure to obtain durable and stable bioprostheses.Up to date, glutaraldehyde (GA) crosslinking has beenthe most widely used method because it enhancesmaterial stability, decreases antigenicity and maintainssterility. However, its use promotes calcification and ispotentially cytotoxic. For this reason, alternativestabilizing reagents or procedures are being developed[4,5].
There are many factors affecting durability ofbiological prostheses: mechanical stress, calcificationand extracellular matrix biodegradation [1]. The
ARTICLE IN PRESSB. Arenaz et al. / Biomaterials 25 (2004) 3359–33683360
mechanical stress of cardiac valve cusps and bloodpressure cause progressive damage to the molecularstructure of type I collagen and induce the loss ofglycosaminoglycans (GAGs) [6]. Both processes aremajor contributing factors to material degeneration inbioprosthetic cardiac valve deterioration [7]. GAGs arenegatively charged polysaccharides of different degreesof complexity which are ubiquitous components of theextracellular matrix. Sulfated GAGs are covalentlylinked to proteins forming proteoglycans (PGs). Inpericardium, PGs are mainly composed of dermatansulfate and chondroitin sulfate [8]. In the latter, thedisaccharide unit contains glucuronate and N-acetylga-lactosamine and has usually one sulfate group perdisaccharide, which is predominantly either in the 4 or 6position on N-acetylgalactosamine. These highlycharged units not only contribute to tissue hydrationand elasticity, but also may participate in the interactionwith other extracellular matrix components. Thesenegative charges may also attract counterions; however,their role in bioprostheses calcification is controversial.For example, PGs extracted from nasal and articularcartilages retard hydroxyapatite seeded-growth in vitro[9], and addition of high concentrations of chondroitin4-sulfate inhibits the formation of hydroxyapatite incollagen gels [10]. Additionally, we have previouslydescribed that selective PGs extraction in pericardiumresults in a greater accumulation of calcium salts than inthe un-extracted tissue, with a reduced hydrothermalstability [11].
An increasing amount of interest is focused on the useof alternative chemical modifications of biologicaltissues, including PGs fixation, aimed to improve tissuebiocompatibility and calcification resistance. Anselmeet al. [12] proposed an acyl azide method to crosslink acollagen–GAGs sponge which increased the persistenceof the sponge and inhibited its calcification after 90 dayssubcutaneous implantation in rats. Crosslinking ofchondroitin sulfate to collagen gels or porous collagenmatrices has also been achieved by using carbodiimidein the absence or presence of diamines [13,14]. Amaximum of about 155mgCS/g could be immobilizedto the porous collagen matrices under these conditions[15]. Among the chemical treatments for PG fixation,periodate has been described to reduce structuraldegeneration of the bioprostheses. In addition, it hasbeen described that this treatment significantly reducescalcification of biological implants compared with GA-fixed materials after subcutaneous implantation in rats[16].
GA and periodate treatments generate aldehydegroups and, thus, some of them may remain free afterthe crosslinking reaction yielding potentially cytotoxicgroups. It has been suggested that reduction of thesegroups to alcohol with sodium borohydride maydiminish the potential cytotoxicity of the implants with
a parallel stabilization due to a concomitant reductionof the crosslinking imine bonds. Accordingly, porcinetype I collagen crosslinked by oxidized glycogen is toxicfor human fibroblasts in culture, but reduction withborohydride allows cell adhesion and proliferation onthe biomaterial [17].
Other important aspect to be considered in biologicalprostheses stability is the susceptibility of the implantsto biodegradation, where matrix metalloproteinases(MMPs) play an important role. In fact, MMPsactivities have been detected in pathological humancardiac valves and in pericardium-derived biomaterials[18–21]. MMPs are members of a family of enzymeswhich under physiological conditions contribute totissue remodeling and reparation, and play importantroles in embryo development and morphogenesis as wellas in several pathologies.
The aim of this work is the study the influence of PGsfixation in the behavior of ostrich pericardium aftersubcutaneous implantation in rats. For this purpose, wehave either fixed endogenous PGs or exogenouschondroitin sulfate by treatment with periodate fol-lowed by a final GA-stabilization. We have analyzed theinfluence of all these treatments on tissue calcificationafter implantation as well as the presence of matrixdegrading activities. An additional borohydride redu-cing treatment of the chemically modified pericardiumwas investigated.
2. Materials and methods
2.1. Materials
Ostrich pericardium was obtained directly from alocal slaughterhouse and transported to our laboratoryin sterile saline solution (0.9% NaCl, w/v). Afterwards,the tissue was cleaned to remove fat, and portions wereselected for different biochemical studies and forsubcutaneous implantation. All reagents were purchasedfrom Sigma (St. Louis, MO, USA) unless otherwisestated, and were of the highest purity available.
2.2. Chemical treatments
Ostrich pericardium was subjected to the chemicaltreatments schematically shown in Fig. 1. In the controlgroup (GA group), ostrich pericardium specimens weretreated for 24 h at room temperature with 0.626% (v/v)GA in 0.1m sodium phosphate buffer, pH 7.4.
A second group of samples were subjected to aperiodate fixation previous to the GA treatment (MPIgroup) essentially as previously described [16]. Briefly,ostrich pericardium specimens were placed into a 0.01msodium metaperiodate solution in distilled water andmaintained under constant shaking in the dark at 4�C
ARTICLE IN PRESS
Fig. 1. Chemical treatments and groups of modified ostrich pericardium
used for subcutaneous implantation in rats. GA: glutaraldehyde; CS:
chondroitin 4-sulfate; MPI: metaperiodate; BH: sodium borohydride.
B. Arenaz et al. / Biomaterials 25 (2004) 3359–3368 3361
for 24 h in a 10:1 (w/v) ratio. After periodate fixation,the samples were washed thoroughly (twice for 30min)with phosphate buffered saline (PBS: 0.15m NaCl,2.7mm KCl, 1.5mm KH2PO4, 8.1mm Na2HPO4, pH7.4) and finally incubated, as in the control group, in0.625% (v/v) GA for 24 h.
A third group was fixed with metaperiodate inidentical conditions as described for the MPI groupbut in the presence of 0.5% (w/v) chondroitin 4-sulfatefrom bovine trachea. Afterwards, samples were washedand post-fixed with GA as above.
Some of tissue specimens from the groups previouslydescribed were subjected to an additional treatment withsodium borohydride (Fig. 1) in order to reduce theimino groups formed in the crosslinking process betweencollagen and GA. For this purpose, samples were treatedfor 24 h with 0.01% (w/v) NaBH4 in deionized waterand washed three times for 30min with PBS.
2.3. Proteoglycan/glycosaminoglycan extraction and
quantification
Tissue samples were rinsed in PBS and minced. PGsand GAGs were extracted in 4m guanidine HCl insodium acetate, pH 5.8, containing protease inhibitors:1mm p-hydroxymercuribenzoate, 1mm phenylmethyl-sulfonyl fluoride, 1 mg/ml soybean trypsin inhibitor and10mm EDTA [22]. After dialysis (MWCO 12,000) at4�C for 72 h against distilled water, the PG/GAGscontent of the extracted samples was determined with aBlyscan assay kit (Biocolor Ltd., Belfast, North Ireland)and expressed as micrograms per milligram of wetweight tissue.
2.4. In vivo studies: subcutaneous implantation
Wistar rats weighing between 80 and 110 g (4–5 weeksold) were employed in subcutaneous implantation trials.
Six pericardium disks measuring 1 cm in diameter wereimplanted into the abdominal wall muscle of each of the24 animals employed in the study. Twelve animals wereused for the implantation of materials without sodiumborohydride; in each animal, two of the six implantsbelonged to the GA control group ðn ¼ 12Þ; two to theMPI group ðn ¼ 12Þ and two to the CS+MPI groupðn ¼ 12Þ: Another 12 animals were used for theimplantation of specimens with the sodium borohydridetreatment with identical distribution. The disks re-mained implanted for 30 (half of the animals) or 60days (the other half). At the end of the selected period oftime, the rats were killed by CO2 asphyxiation, and thesamples were retrieved, carefully freed of any surround-ing host tissue and rinsed with saline solution.
2.5. Calcium and phosphorous analysis
In order to determine the calcium and phosphorousaccumulation in the different implanted samples, speci-mens were rinsed in water, vacuum dried and weighed.The dry samples were hydrolyzed in 6n HCl at 100�Cfor 24 h and calcium content was quantified by atomicabsorption spectrometry (Perkin-Elmer 1100 B AtomicAbsorption Spectrophotometer) [11], and phosphorousby a molibdate complexation assay, as described else-where [23]. Calcium and phosphorous levels areexpressed as mg/g dry weight tissue.
2.6. Determination of the calcium phosphate type by
X-ray diffraction
Dry portions of the implanted materials wereanalyzed by X-ray diffraction in a Diffraktometer-1Kristalloflex (Siemens AG) and using the software EVAv3.30 Diffract AT from Socabim-Siemens for theinterpretation of the results. This software containsdiverse databases with different pattern diffractogramsfrom a great variety of substances that were used for theidentification of the peaks obtained with each sample.All the samples showed a very elastic behavior and itwas not possible to powder the tissue due to the highcollagen content. Thus, the whole samples were placedinto the equipment.
2.7. PAGE-SDS analysis
Treated tissue fragments were minced manually andsuspended in extraction buffer (50mm Tris, pH 7.4,containing 0.25% (v/v) Triton X-100 and 0.5% (w/v)SDS) at 1.25% (w/v). Samples were incubated for 48 hat 4�C, after which the tissue fragments were homo-genized for 5min using a glass mortar and pestle. Aftercentrifugation for 10min at 14,000 rpm, the super-natants were collected and analyzed by PAGE-SDSand by gelatin zymography. Equal volumes of the
ARTICLE IN PRESS
Fig. 2. Proteoglycan/glycosaminoglycan quantification in pericardium
samples after periodate fixation. Samples were subjected to a previous
fixation of PGs either in the absence (MPI) or in the presence of 0.5%
(w/v) chondroitin 4-sulfate, and were further stabilized by GA-
crosslinking. Pericardium samples without periodate fixation but with
GA-crosslinking (GA) were used as control. Extraction and determi-
nation of the overall PG content was carried out as described in
Section 2. Data are expressed as mean values7SD; statistical
significance: po0:01 (��).
B. Arenaz et al. / Biomaterials 25 (2004) 3359–33683362
supernatants from the homogenization of the differentpericardium samples were prepared for electrophoresisby addition of three-fold-concentrated loading buffer(1� loading buffer: 62.5mm Tris, pH 6.8, containing2% (w/v) SDS, 5% (v/v) 2-mercaptoethanol, 10% (v/v)glycerol and 0.02% (w/v) bromophenol blue). Sampleswere heat-denatured for 5min at 90�C and analyzed byPAGE-SDS according to Laemmli [24], using 4%stacking gels and 10% resolving gels in a Mini ProteanII unit (Bio-Rad Laboratories, Hercules, CA, USA) at25mA/gel. The gels were then routinely stained withCoomassie Brilliant Blue R250. Pre-stained markersfrom Bio-Rad were used for molecular weight determi-nation.
2.8. Gelatin zymography
Samples were prepared as described above for thePAGE-SDS analysis, but in absence of thiol reducingagents, as previously described [20,21]. Polyacrylamideresolving gels (10%) were copolymerized with 1mg/mlgelatin. Equal amounts of protein (20 mg) were loadedonto the gels without thermal denaturation and theelectrophoresis was run under standard conditions.Afterwards, the gels were washed once with 50mm Tris,pH 7.4, containing 2.5% (v/v) Triton X-100 for 30min,and twice with 50mm Tris, pH 7.4, in order to removeSDS and to allow reactivation of the gelatinaseactivities. Gels were then incubated for 48–62 hin 50mm Tris, pH 7.5, 0.15 NaCl, 10mm CaCl2, 0.1%(v/v) Triton X-100 and 0.02w/v% sodium azide.Finally, the gels were stained with Coomassie blue andwashed afterwards with 7.5% (v/v) acetic acid contain-ing 20% (v/v) methanol. The relative mobility of MMP-2 and MMP-9 was determined using conditionedmedium from HT-1080 fibrosarcoma cells (ATCC,CCL 121) [21]. The gels were subjected to densitometricanalysis to quantitate the gelatinase activity by obtain-ing volumograms on a photodocumentation systemfrom UVItec (Cambridge, UK) using the UVIBandv.97 software.
2.9. Statistical analysis
Proteoglycan quantification, calcium atomic absorp-tion spectroscopy, and phosphorous data are expressedas means7SD. In order to evaluate the statisticalsignificance of the differences among the calcium andphosphorous determinations in the different groups,first the Kolmogorov–Smirnov test was performed toensure that the normality hypothesis was not rejected ineither case. Then, intragroup comparison was carriedout using analysis of variance (ANOVA) and theNewman–Keuls multiple comparisons test; intergroupcomparison was performed by means of a two-tailed Student’s t-test for unpaired data. A significance
level of 0.05 or less was accepted as being statisticallyrelevant.
3. Results
3.1. Proteoglycan/glycosaminoglycan quantification
Periodate treatment of ostrich pericardium has beencarried out in the absence and presence of chondroitin 4-sulfate in order to fix and increase the overall PGscontent in the implants. PGs quantification data ofsamples from the control GA group and from the MPIand CS+MPI groups are shown in Fig. 2. PG contentafter guanidinium chloride extraction in periodatetreated tissue is slightly higher than that in the controlsamples, 0.03870.003 and 0.02870.004 mg/mg tissueðpo0:01Þ; respectively. However, a much more signifi-cant increase in extracted PGs (up to 0.16470.004 mg/mg tissue; po0:01) is achieved when MPI fixation isperformed in the presence of 0.5% (w/v) chondroitin 4-sulfate.
3.2. In vivo calcification studies
Accumulation of calcium and phosphate in the threegroups of implants has been analyzed after subcuta-neous implantation in rats. Fig. 3 shows the calcium andphosphorous content, as well as their molar ratio, indifferent ostrich pericardium samples (GA, MPI andCS-MPI groups) retrieved after 30 or 60 days implanta-tion. The calcium and phosphorous content (Figs. 3Aand B) is much higher ðpo0:01Þ in the GA-group at
ARTICLE IN PRESS
Fig. 3. Calcium and phosphorous content in implanted chemically
modified ostrich pericardium. Calcium (A) and phosphorous (B)
accumulated in the implants from the GA, MPI and CS+MPI groups,
expressed as mg/g of dry tissue, are determined as described in Section
2. The Ca/P molar ratio is also shown (C). Data are expressed as mean
values7SD of 12 implants per group and implantation time: 30 days
(black bars) and 60 days (empty bars). Statistical significance: po0:01(��); only significant differences between groups are shown.
Table 1
Calcium and phosphorous content in implants from borohydride-
treated pericardium groups
Group Ca (mg/g) P (mg/g) Ca/P (molar ratio)
GA�BH
30 days 60.77725.38 50.64724.41 0.9270.24
60 daysa 77.34723.67 66.6078.77 0.8870.10
MPI�BH
30 days 46.05727.60 42.98726.62 0.8570.17
60 days 44.50730.81 32.53710.56 1.0670.59
CS+MPI�BH
30 days 47.01730.30 55.20719.96 0.6670.36
60 daysa 32.26717.84 22.0076.08 1.1470.44
Note: The calcium and phosphorous content is expressed as mg/g of
dry tissue. Data represent the mean7SD of the 12 implants analyzed
for each group after 30 and 60 days implantation.aSignificant differences ðpo0:01Þ were only found between the
control GA�BH group and the CS+MPI�BH group after 60 days
implantation.
B. Arenaz et al. / Biomaterials 25 (2004) 3359–3368 3363
both implantation times compared to implants subjectedto a fixation of PGs with periodate either in the absenceor presence of exogenous chondroitin 4-sulfate (MPIand CS+MPI groups). On the other hand, nosignificant differences were found neither among thedifferent periodate fixed samples nor between samplesafter 30 or 60 days implantation. Concerning the molarCa/P ratio, both groups of periodate-fixated implantsshow a significantly lower ðpo0:01Þ ratio than the GA-
group at both studied implantation times (Fig. 3C). Inthe control group, a significant increase in this ratio isobserved from 30 to 60 days implantation (0.3370.09vs. 0.5570.11; po0:01). The highest Ca/P molar ratio isobtained in the GA-group after 60 days implantation;however, this value must correspond to amorphouscalcium phosphate as this ratio is far from those foundin mature bone (Ca/P=1.75) and in surgically explantedcalcified natural and bioprosthetic valves (1.8370.06and 1.5270.06) [25].
3.3. Effect of borohydride treatment on implant
calcification
We have analyzed the effect of this treatment on thecalcification behavior of the different groups of modifiedostrich pericardium. Table 1 shows the calcium andphosphorous content, as well as their molar ratio, ofborohydride-stabilized samples after implantation for 30and 60 days. The only significant differences among thedifferent groups and treatment times were foundbetween the control GA-group and the CS+MPI groupafter 2 months implantation ðpo0:01Þ: However, thesedifferences are not relevant taking into account that thetreatment was directed to reduce the calcium deposits inthe implanted samples. In all cases, the calcium andphosphorous levels are higher than those obtained forthe corresponding borohydride-untreated implants (Fig.3). In fact, after 1 month implantation, borohydridetreatment increases calcium content 3.2-fold in the GA-group (18.89712.27 to 60.77725.38mg Ca/g) andaround hundred-fold in the MPI (0.4470.17 to46.05727.60mg Ca/g) and CS+MPI (0.4270.19 to47.01730.30mg Ca/g) groups; similar increases wereobserved after 2 months implantation. Phosphorous
ARTICLE IN PRESSB. Arenaz et al. / Biomaterials 25 (2004) 3359–33683364
content is also increased but to a lower extent thancalcium. Regarding the Ca/P molar ratio, these valuesare close to 1, higher than those obtained in implantedpericardium samples without the borohydride treat-ment.
3.4. Characterization of calcium deposits in implanted
pericardium specimens
Calcium deposits were characterized in the borohy-dride-treated implants as they were the ones presenting ahigher calcification degree after implantation. Fig. 4shows the diffractogram obtained from ostrich pericar-dium samples after 60 days implantation. According tothe database pattern, the 2y peak with a maximumaround 32� correspond to some type of apatite,although the characteristics of the sample did not allowto identify the specific type; the small peak around 20–22� could correspond to brushite (CaHPO4 � 2H2O). Thelarge peak centered at 12� can be attributed to collagenfrom the tissue.
Ca/P molar ratios in borohydride untreated samplesare low in the control GA group (0.3370.09 and0.5570.11 after 30 and 60 days implantation, respec-tively), and very low in the periodate-prefixed implants(lower than 0.076; Fig. 3C). These calcium deposits mustcorrespond to amorphous calcium phosphates. On theother hand, borohydride treated samples not onlypresent a higher calcification degree, but also showhigher Ca/P molar ratios. In fact, these values arearound 1 (Table 1) that could correspond to brushite or
Fig. 4. X-ray diffraction spectrum of calcific deposits of implanted
ostrich pericardium. Borohydride-treated samples were analyzed as
described in Section 2 and the obtained peaks were compared to a
database diffraction pattern. The spectrum of reference hydroxyapatite
(HAP) is included.
to octacalcium phosphate [Ca/P=1.33; Ca8(PO4)6-H2 � 5H2O], that may be transient precursor phases inthe formation of apatite deposits as also described byother authors [25]. However, some apatite-like crystalstructures [Ca/P=1.66; Ca5(PO4)3OH] must exist inthese deposits, since a peak corresponding to diffractionof apatite is detected by X-ray diffraction.
3.5. Electrophoretic and gelatin zymography analyses of
pericardium implants
Fig. 5 shows the PAGE-SDS analysis of samplesextracted from pericardium specimens without or withborohydride treatment, and removed after 30 days ofsubcutaneous implantation. Results obtained after 60days implantation are very similar. The extractedproteins in the different experimental groups presentthe same band pattern, with a very intense band around68 kDa probably corresponding to albumin from thehost. Implanted pericardium from the borohydride-treated groups presents, in general, a slightly higherdegree of protein solubilization reflected in more intenseprotein bands in the gels when loading equivalentvolumes of extracted material.
We have also analyzed the presence of gelatinaseactivities in the different samples after implantation for30 and 60 days by using gelatin zymography. Fig. 6shows a representative gel with samples from theborohydride untreated groups. Since the amount of
Fig. 5. Electrophoretic analysis of proteins extracted from ostrich
pericardium implants. Surgically excised implants from the GA, MPI
and CS+MPI (30 days implantation), without and with borohydride
post-fixation, were homogenized as described in Section 2. The
supernatants were analyzed by SDS-PAGE using 10% polyacrylamide
resolving gels and the protein bands were visualized after staining with
Coomassie blue. A representative gel is shown. Molecular weight
standards are shown (MW).
ARTICLE IN PRESS
Fig. 6. Gelatin zymography of proteins extracted from homogenized
ostrich pericardium implants. Samples from the GA, MPI and
CS+MPI groups were surgically excised after 30 or 60 days
implantation, and gelatinase activities were extracted as described in
Section 2. Equal amounts of protein were loaded into each lane
without thermal denaturation. A representative overloaded gel is
shown in order to appreciate the presence of minor gelatinase
activities. The mobilities of MMP-2 and MMP-9 were established
using conditioned medium from HT-1080 fibrosarcoma cells. Mole-
cular weight standards were also loaded to estimate apparent
molecular masses of gelatinolytic activities.
Fig. 7. Densitometric analysis of MMP-2 and MMP-9 activities in
ostrich pericardium implants. Gelatin zymograms in which MMP-2
and MMP-9 appeared as clear/sharp bands were scanned and analyzed
using a photodocumentation system and volumograms from the
gelatinase activity bands were obtained. Samples were identical to
those described in Fig. 6. (A) Activity ratios between the MMP-2 and
MMP-9 doublets. The relative amount of each activity referred to the
corresponding control (GA) is shown in (B) for MMP-9 and in (C) for
MMP-2. Four different zymograms with samples from independent
implants were analyzed. Data are expressed as the mean7SD of the
corresponding values. Statistical significance: po0:05 (�), po0:01 (��);only significant differences between groups are shown.
B. Arenaz et al. / Biomaterials 25 (2004) 3359–3368 3365
protein loaded into each lane is identical, differences inthe gelatinolytic activity can be directly correlated withthe amount of metalloproteinase present in the implants.GA-fixed ostrich pericardium before implantationshows negligible gelatinase activities, as we have alsopreviously shown for other pericardium sources [20,21].However, after implantation, strong negatively stainedbands with electrophoretic mobilities equivalent toMMP-2 (B70 kDa) and MMP-9 (B90 kDa) are ob-served. These bands appear as doublets, correspondingto the pro-enzyme and to the proteolytically activatedform. Additional gelatinase activities are also detectedwith lower molecular masses, a doublet or triplet around45 kDa and an additional band close to 30 kDa; thesebands are very faint in the control GA group, but areclearly detected in the MPI-group mainly after 60 daysimplantation.
Fig. 7A shows the ratio between total MMP-2 andtotal MMP-9 activities (pro-enzyme plus active forms)obtained from the densitometric analysis of the zymo-grams. This analysis shows that the MMP-2/MMP-9ratio is lower in the periodate-treated groups comparedto the GA-group (B2 vs. 3–3.5). No significantdifferences were found in this ratio between 30 and 60days implantation within the same group. Figs. 7B andC show the relative amount of MMP-9 and MMP-2referred to the corresponding activity detected in the
control GA-group after 30 days implantation. Ingeneral, periodate treated samples show higher gelati-nase activities than the GA-group. No significantdifferences were found between gelatinase activitiesdetected after 30 and 60 days implantation in the GA-group and in the CS+MPI group, while an increase inboth activities with implantation time was observed forthe MPI-group (1.6- and 1.3-fold for MMP-9 andMMP-2, respectively). In this last group, there was alsoa significant increase in the gelatinase activities showingmolecular masses around 45 and 30 kDa (Fig. 6).
ARTICLE IN PRESSB. Arenaz et al. / Biomaterials 25 (2004) 3359–33683366
Borohydride-treated implants showed the same gen-eral gelatinolytic band pattern than the correspondinguntreated ones but with a general increase in the overallgelatinolytic activity (between 1.5 and 1.8-fold).
4. Discussion
One of the major fields in cardiac bioprosthesesresearch is the achievement of stabilizing treatments thatreduce calcification and maintain the correct structure ofthe native tissue. Several approaches, both in vivo andin vitro have been reported. The most extendedcommercial chemical modification of biological tissuesis GA crosslinking; however, this treatment is one of themajor causes for clinical failure of bioprosthetic tissuescombined with deposition of proteins and lipids andmechanical stress. Several authors have demonstratedthat GA treatment induces an structural deterioration ofcollagen molecules and promotes GAGs extraction,both processes being probably linked to an increasedcalcification [7,26].
In the present work we have analyzed the effect ofendogenous PGs fixation in ostrich pericardium byusing periodate, as described by Lovekamp andVyhavahare [16], and the influence of the addition ofexogenous chondroitin sulfate during this treatment inorder to increase the overall GAGs content in the tissue.We have achieved an increase in the total PGs content inostrich pericardium when a periodate fixation isperformed before the final GA crosslinking (around1.4-fold). Thus, this previous fixation avoids, at leastpartially, one of the undesirable effects of GA. More-over, when exogenous chondroitin 4-sulfate is addedduring periodate fixation, a high increase in the totalPGs content (more than 4-fold) can be obtained. In thisway, the PGs content of the material to be implantedcan be controlled. However, in all cases, a final GAfixation is required to confer enough mechanicalresistance to the implant by inducing covalent crosslinksbetween collagen molecules.
Other non-desirable effect of GA fixation is theintroduction of potentially toxic free aldehyde groups;this effect may be increased by periodate treatment.Periodate induces oxidative rupture of geminal diolspresent in the glucuronic acid moieties of GAGs. Freealdehydes may then react with other molecules presentin the bioprostheses, mainly with the amine groups ofthe lysine lateral chains from collagen. This crosslinkingstabilizes the interaction between PGs and othercomponents of the extracellular matrix. Some authorsare developing absorbable scaffolds composed ofcollagen-based biomaterials, crosslinked by differentmethods, and subjected to a post-treatment with sodiumborohydride [17,27]. It has been described that reductionof free aldehydes by borohydride following GA cross-
linking may be helpful to avoid this cytotoxic effect.Moreover, this reduction step stops the crosslinkingreaction and the progress of the already formed cross-links to more complex bonds, which may modify themechanical and physico-chemical properties as well asthe biodegradability of the implant. Reduction of labileimine bonds between oxidized GAGs and collagen, orbetween collagen lysine residues, has been suggested as astabilizing factor by formation of stable secondaryamines [17].
Taking into account all these considerations, we haveimplanted six different groups of chemically modifiedpericardium to analyze the effect of PGs fixation andborohydride treatment in the behavior of the implantsregarding calcification, mechanical stability and sensi-tivity to colonization by MMP secreting cells.
When calcification of the implants is analyzed, asignificant reduction of this process is observed whenendogenous pericardium PGs are pre-fixed with period-ate prior to GA stabilization (around 50-fold). However,even though periodate treatment in the presence ofchondroitin 4-sulfate increased PGs content in theimplants, no significant differences were observedbetween the MPI and the CS+MPI groups regardingcalcification at 30 or 60 days. Thus, we have confirmedthat stabilization of endogenous PGs is an importantmechanism to avoid calcification after implantation. Infact, some reports point out the protective effect ofcollagen–GAG or PGs interaction in avoiding calciumdeposits [28]. This effect is not easily explained takinginto account that GAGs present a high negative chargedensity due to the sulfate groups. However, it isinteresting that other negatively charged compounds,as heparin or sulfated poly(ethylene oxide), whencoupled to biological materials, also induce calcificationresistance [29–31]. GA treatment has been shown toremove, at least partially, PGs from pericardium [1],generating voids and cavities that may potentially trapforeign particles that may lead to nucleation centers forcalcification. Thus, pre-fixation of endogenous PGscould inhibit mineralization by avoiding the formationof these nuclei. Additionally, it can be suggested thatacidic sulfate groups might contribute to solubilizemetastable calcium deposits retarding in this way theformation of nucleation cores.
Surprisingly, when both groups of periodate pre-fixedpericardium were treated with borohydride after GAcrosslinking, the protective effect of PGs almostdisappear and calcification is strongly promoted(around 100-fold). This increase is also observed in theGA control group, but to a lower extent; however,calcification is still higher in the GA-borohydride groupcompared to the periodate-borohydride one. Proteinsfrom borohydride-treated implants are more easilyextracted after implantation than the correspondinguntreated implants, and present higher content of
ARTICLE IN PRESSB. Arenaz et al. / Biomaterials 25 (2004) 3359–3368 3367
gelatinolytic activities. Borohydride treatment has beenshown to reduce in vitro cytotoxicity of GA- andperiodate-treated collagen-based biomaterials [17,27].However, from the data herein presented, this treatment isnot advisable on GA-fixed pericardium since it must alterthe physico-chemical properties of the implants in such away that strongly promotes calcification and increases thesusceptibility of the materials to biodegradation.
Gelatinase activities, mainly MMP-2 and MMP-9,have been detected in all the implanted groups ofchemically modified pericardium. Differences betweenGA and periodate groups of implants are more evidentwhen MMP-9 is considered, whereas changes in MMP-2levels among the different groups are not so relevant.These activities must arise from host cell invasion of theimplants as all the biomaterials have been extensivelyfixed with GA, which almost totally inhibit endogenousMMP activities [20,21]. Thus, the presence of thesedegrading enzymes could be related to the biodegrad-ability of different materials. Control GA-treatedimplants are the ones presenting the lowest levels ofgelatinase activity, but show the highest calcificationdegree. On the other hand, periodate treated implantsshow a highly significant decrease in calcification withonly a slight increase in gelatinase activities. Moreover,the appearance of host gelatinase activities can bereduced by addition of chondroitin 4-sulfate duringperiodate fixation. This effect is more patent at longerimplantation times, where not only a decrease in MMP-2 and MMP-9 activities, but also the disappearance ofthe low molecular mass gelatinases, is observed.
5. Conclusion
GA-fixed ostrich pericardium presents calcificationafter subcutaneous implantation in rats. However, thisprocess can be significantly reduced by stabilization ofthe endogenous PGs or by the addition of exogenousGAGs via periodate pre-fixation. This treatment seemsto slightly enhance the susceptibility of the implants toinvasion by MMP producing host cells, but this processcan be reduced to levels similar to those detected in thecontrol GA-treated implants by exogenous GAGs pre-fixation. Borohydride treatment of GA-fixed pericar-dium, in order to diminish potential cytotoxicity ofremaining free aldehyde groups and reduction ofaldimine bonds, is not advisable under these conditionsas it dramatically enhances calcification of the bioma-terials, reduces tissue stability and increases MMPcontent after implantation.
Acknowledgements
This study was financed by grants PI020279 from FIS(Spanish Ministry of Health and Consumer Affairs),
08.9/0003.1/2001 from CAM and BCM2002-01407 fromDGI (Spanish Ministry of Education, Culture andSports).
References
[1] Schoen FJ, Levy RJ. Tissue heart valves: current challenges and
future research perspectives. J Biomed Mater Res 1999;47:
439–65.
[2] Holper K, Wottke M, Lewe T, Baumer L, Meisner H, Paek SU,
Sebening F. Bioprosthetic and mechanical valves in the elderly:
benefits and risks. Thorac Surg 1995;60:S443–6.
[3] Edmunds LH, Clark RE, Cohn LH, Grunkemeier GL, Miller
DC, Weisel RD. Guidelines for reporting morbidity and mortality
after cardiac valvular operations. The American Association for
Thoracic Surgery, Ad Hoc Liason Committee for Standardizing
Definitions of Prosthetic Heart Valve Morbidity. Ann Thorac
Surg 1996;62:932–5.
[4] Huma P, Bezuidenhout D, Torrianni M, Hendriks M, Zilla P.
Optimization of diamine bridges in glutaraldehyde treated
bioprosthetic aortic wall tissue. Biomaterials 2002;23:
2099–103.
[5] Singla A, Lee CH. Inhibition of CEM calcification by the
sequential pretreatment with ethanol and EDTA. J Biomed Mater
Res 2003;64A:706–13.
[6] Grande-Allen KJ, Mako WJ, Calabro A, Shi Y, Ratliff NB,
Vesely I. Loss of chondroitin 6-sulfate and hyaluronan from failed
porcine bioprosthetic valves. J Biomed Mater Res 2003;65A:
252–60.
[7] Vyavahare N, Ogle M, Schoen FJ, Zand R, Gloeckner DC, Sacks
M, Levy RJ. Mechanisms of bioprosthetic heart valve failure:
fatigue causes collagen denaturation and glycosaminoglycan loss.
J Biomed Mater Res 1999;46:44–50.
[8] Simionescu DT, Kefalides NA. The biosynthesis of proteoglycans
and interstitial collagens by bovine pericardial fibroblasts. Exp
Cell Res 1991;195:171–6.
[9] Chen CC, Boskey AL. The effects of proteoglycans from different
cartilage types on in vitro hydroxyapatite proliferation. Calcif
Tissue Int 1986;39:324–7.
[10] Hunter GK, Allen BL, Grynpas MD, Cheng PT. Inhibition of
hydroxyapatite formation in collagen gels by chondroitin
sulphate. Biochem J 1985;228:463–9.
[11] Jorge-Herrero E, Fern!andez P, Guti!errez M, Castillo-Olivares JL.
Study of the calcification of bovine pericardium: analysis of the
implication of lipids and proteoglycans. Biomaterials
1991;12:683–9.
[12] Anselme K, Petite H, Herbage B. Inhibition of calcification in vivo
by acyl azide cross-linking of a collagen-glycosaminoglycan
sponge. Matrix 1992;12:264–73.
[13] Hanthamrongwith M, Reid WH, Grant MH. Chondroitin-6-
sulphate incorporated into collagen gels for the growth of human
keratinocytes: the effect of cross-linking agents and diamines.
Biomaterials 1996;17:775–80.
[14] Pieper JS, Oosterhof A, Dijkstra PJ, Veerkamp JH, van
Kuppervelt TH. Preparation and characterization of porous
crosslinked collagenous matrices containing bioavailable chon-
droitin sulphate. Biomaterials 1999;20:847–58.
[15] Pieper JS, Hafmans T, Veerkamp JH, van Kuppervelt TH.
Development of tailor-made collagen-glycosaminoglycan ma-
trices: EDC/NHS crosslinking, and structural aspects. Biomater-
ials 2000;21:581–93.
[16] Lovekamp J, Vyavahare N. Periodate-mediated glycosaminogly-
can stabilization in bioprosthetic heart valves. J Biomed Mater
Res 2001;56:478–86.
ARTICLE IN PRESSB. Arenaz et al. / Biomaterials 25 (2004) 3359–33683368
[17] Rousseau CF, Gagnieu ChH. In vitro cytocompatibility of
porcine type I atellocollagen crosslinked by oxidized glycogen.
Biomaterials 2002;23:1503–10.
[18] Simionescu A, Simionescu D, Deac R. Biochemical pathways of
tissue degeneration in bioprosthetic cardiac valves. ASAIO J
1996;42:M561–7.
[19] Simionescu A, Simionescu D, Deac RFP. Matrix metalloprotei-
nases in the pathology of natural and bioprosthetic cardiac valves.
Cardiovasc Pathol 1996;5:323–32.
[20] Jorge-Herrero E, Turnay J, Calero P, Olmo N, L !opez de Silanes I,
Mart!ın Maestro M, Lizarbe MA, Castillo-Olivares JL. Calcifica-
tion and identification of metalloproteinases in bovine pericar-
dium after subcutaneous implantation in rats. J Mater Sci Mater
Med 2001;12:1013–7.
[21] Calero P, Jorge-Herrero E, Turnay J, Olmo N, L !opez de Silanes I,
Lizarbe MA, Mart!ın Maestro M, Arenaz B, Castillo-Olivares JL.
Gelatinases in soft tissue biomaterials. Analysis of different
crosslinking agents. Biomaterials 2002;23:3473–8.
[22] Gura TA, Wright KL, Veis A, Webb CL. Identification of specific
calcium-binding noncollagenous proteins associated with glutar-
aldehyde-preserved bovine pericardium in the rat subdermal
model. J Biomed Mater Res 1997;35:483–95.
[23] Chen P, Toribara T, Warner H. Microdetermination of phos-
phorous. Anal Chem 1956;28:1756.
[24] Laemmli UK. Cleavage of structural proteins during the assembly
of the head of bacteriophage T4. Nature 1970;227:680–5.
[25] Mikroulis D, Mavrilas D, Kapolos J, Koutsoukos PG, Lolas C.
Physicochemical and microscopical study of calcific deposits from
natural and bioprosthetic heart valves. Comparison and implica-
tions for mineralization mechanism. J Mater Sci Mater Med
2002;13:885–9.
[26] Nimni ME, Myers D, Ertl D, Han B. Factors which affect the
calcification of tissue-derived bioprostheses. J Biomed Mater Res
1997;35:531–7.
[27] Mao J, Zhao L, de Yao K, Shang Q, Yang G, Cao Y. Study of
novel chitosan-gelatin artificial skin in vitro. J Biomed Mater Res
2003;64A:301–8.
[28] Zimmermann B. Changes in glycosaminoglycan binding to
collagen during desmoid mineralization as revealed by different
electron-microscopic staining techniques. Acta Anat
1992;145:277–82.
[29] Lee WK, Park KD, Kim YH, Suh H, Park JCh, Lee JE, Sun K,
Baek MJ, Kim H-M, Kim SH. Improved calcification resistance
and biocompatibility of tissue patch grafted with sulfonated PEO
or heparin after glutaraldehyde fixation. J Biomed Mater Res
2001;58:27–35.
[30] Chanda J, Kuribayashi R, Abe T. Heparin in calcification
prevention of porcine pericardial bioprostheses. Biomaterials
1997;18:109–13.
[31] Chanda J, Kuribayashi R, Abe T. Heparin coupling in inhibition
of calcification of vascular bioprostheses. Biomaterials
1999;20:1753–7.
