Accepted Manuscript
Title: Histological Evaluation of Explanted Tissue Engineered Bovine
Pericardium (CardioCel®)
Author: Sudesh Prabhu, Jane E Armes, Douglas Bell, Robert Justo, Prem
Venugopal, Tom Karl, Nelson Alphonso
PII: S1043-0679(17)30163-6
DOI: http://dx.doi.org/doi: 10.1053/j.semtcvs.2017.05.017
Reference: YSTCS 986
To appear in: Seminars in Thoracic and Cardiovascular Surgery
Please cite this article as: Sudesh Prabhu, Jane E Armes, Douglas Bell, Robert Justo, Prem
Venugopal, Tom Karl, Nelson Alphonso, Histological Evaluation of Explanted Tissue
Engineered Bovine Pericardium (CardioCel®), Seminars in Thoracic and Cardiovascular Surgery
(2017), http://dx.doi.org/doi: 10.1053/j.semtcvs.2017.05.017.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service
to our customers we are providing this early version of the manuscript. The manuscript will
undergo copyediting, typesetting, and review of the resulting proof before it is published in its
final form. Please note that during the production process errors may be discovered which could
affect the content, and all legal disclaimers that apply to the journal pertain.
1
Histological evaluation of explanted tissue engineered bovine pericardium (CardioCel®) 1
Sudesh Prabhu MCh1,2,3
, Jane E Armes FRCPA2,3,4
, Douglas Bell MBBS2, Robert Justo 2
FRACP1,2,3
, Prem Venugopal FRCS1, Tom Karl FRACS
2,5,Nelson Alphonso FRACS
1,2,3 3
1 Queensland Paediatric Cardiac Services, Lady Cilento Children’s Hospital, 4
Brisbane, Australia 5
2 School of Medicine, University of Queensland, Brisbane, Australia 6
3 Mater Research Institute, University of Queensland, Brisbane, Australia 7
4 Department of Pathology, Mater Health Services, Brisbane, Australia 8
5 Cardiac Surgery, Johns Hopkins All Children’s Hospital, St. Petersburg, Florida, 9
USA 10
11
Corresponding Author: 12
Nelson Alphonso 13
PO Box 3474, Level 7F, Clinical Directorate 14
Lady Cilento Children’s Hospital, 501 Stanley Street, Brisbane, QLD, 4101, Australia 15
Tel: +61 7 30683486 Fax: +61 7 30684229 16
E-mail: [email protected] 17
18
19
Page 1 of 22
2
Category: Congenital heart disease 20
Word count: 3488 21
Declaration of conflict of interest: The authors declare no potential conflicts of interest with 22
respect to the research, authorship and/or publication of this article 23
Funding: The authors received no financial support for the research, authorship and/or 24
publication of this article 25
26
Page 2 of 22
3
ABSTRACT 27
Objectives 28
CardioCel is bovine pericardium which is subjected to a novel anti-calcification tissue 29
engineering process. We present the histopathological findings of human explants of 30
CardioCel that were used in operations for congenital heart disease in children. 31
Methods 32
Six explants were identified from 140 patients undergoing CardioCel implants from 33
October 2012 to March 2015. CardioCel explants were evaluated histologically using 34
hematoxylin and eosin, Masson’s trichrome and immunohistochemical staining. 35
Results 36
A variable inflammatory response was seen in the surrounding native tissue, but not 37
within the CardioCel graft in any of the explants. Neo-intimal layer of varying thickness 38
developed on the visceralsurface of 5 CardioCel explants with endothelializationof the 39
longest duration explant. A granulation tissue layer developed on the parietalsurface of the 40
graft (consistently thicker than the neo-intima). Maintained collagen fibre architecture 41
(laminated) and variable fibroblastic invasion (which increased with the age of the implant) 42
were identified in all six cases. Scattered capillary vessels were noted in the majority of the 43
explants with new collagen fibres in one, suggesting early remodeling. Calcium was seen in 44
one explant at the interface of the graft and inflammatory response on its parietal surface. 45
Conclusions 46
Evidence of graft remodeling was noted in the majority of the explants without 47
inflammatory cells or calcification within the explanted graft material. A noticeable feature 48
was the differential thickness of the host reaction to the parietal compared with the visceral 49
Page 3 of 22
4
surface of the graft. We will continue to evaluate CardioCel as a cardiovascular substitute for 50
extracardiac and intracardiac reconstruction. 51
Keywords: remodeling; bovine pericardium; histopathology; CardioCel; congenital heart 52
diseases 53
Abstract word count:250 54
55
Page 4 of 22
5
Glossary of abbreviations 56
ECM – extra cellular matrix 57
CHD – Congenital heart disease 58
IHC – immunohistochemistry 59
TAPVD – total anomalous pulmonary venous drainage 60
CT scan – computerized tomography 61
VSD – ventricular septal defect 62
IAA – interrupted aortic arch 63
TOF – tetralogy of Fallot 64
65
66
Page 5 of 22
6
PERSPECTIVE STATEMENT 67
The inflammatory response to CardioCel is limited only to adjacent native tissue. 68
Fibroblast ingrowth is a uniform characteristic. Neo-intimal formation occurs commonly. 69
Early remodeling was demonstrated in the majority. Overall, the histological findings are 70
encouraging and we will continue to evaluate CardioCel as a cardiovascular substitute for 71
extracardiac and intracardiac reconstruction. 72
73
CENTRAL MESSAGE 74
CardioCel grafts show evidence of remodeling, neo-intima formation (visceral surface) and 75
inflammatory response (parietal surface). 76
77
Central Figure 78
Remodelling of a CardioCel transannular patch graft after 502 days of implantation, seen as a 79
fibroblastic infiltrate (†), surrounded by newly interposed, pale, eosinophilic collagen (*) 80
between the brightly eosinophilic graft collagen layers. Neovascularisation (∇) is also 81
demonstrated (H&E, original magnification x40). 82
83
Page 6 of 22
7
INTRODUCTION 84
Glutaraldehyde preserved bovine pericardium was first used in 1977 to construct a 85
prosthetic heart valve 1and has found widespread usage as a cardiovascular substitute for 86
intracardiac and extracardiac repairs. Fresh (unfixed) bovine pericardial implants undergo 87
inflammation and partial digestion without mineralization. Cross-linking collagen with 88
glutaraldehyde stabilizes fibers, adding strength and durability, while reducing antigenicity. 89
However, the limitations of most glutaraldehyde- preserved bovine pericardium are well 90
known, and include degeneration and calcification which causes rigidity and shrinkage. 91
Calcification of glutaraldehyde cross-linked bovine pericardium is predominantly promoted 92
by the presence of cell remnants in the extracellular matrix (ECM), residual membrane-93
associated acidic phospholipid and non-viable interstitial cells rendered inactive by the 94
glutaraldehyde treatment. Calcification has also been observed in the ECM proteins (collagen 95
and elastin) and it has been suggested that collagen serves as a nucleation site for calcium 96
phosphate, independent of devitalized cells2. The amount of incorporated glutaraldehyde 97
determines the degree of cross-linkage, with a high uptake associated with greater 98
calcification3. Antigenicity, immunological reactions and the inflammatory process may also 99
contribute to calcification as evidenced by lower incidence of calcification in glutaraldehyde 100
treated autologous tissue than heterologous tissues 4. 101
CardioCel is bovine pericardium which is subjected to an anti-calcification tissue 102
engineering process (ADAPT TEP) developed by AdmedusRegen Pty Ltd (Perth, Western 103
Australia). ADAPT TEP includes steps to reduce cytotoxicity by removing lipids, all cells 104
and cell remnants, nucleic acids (DNA, RNA) and α-Gal (galactosyl) epitopes. Crosslinking 105
to maintain strength and elasticity is achieved with a low concentration of monomeric 106
glutaraldehyde. Detoxification is further promoted by non-glutaraldehyde sterilization and 107
storage in a glutaraldehyde free solution5, 6
. 108
Page 7 of 22
8
Remodeling of an implantable biological scaffold with native tissue ingrowth, 109
endothelialization and neovascularization in humans remains an elusive goal and such a 110
scaffold would undoubtedly represent the ideal biological material for repair of congenital 111
heart defects (CHD). Animal studies have demonstrated migration of autologous cells into a 112
CardioCel scaffold with endothelialization and neovascularization 7, 8
. Remodeling has not 113
been demonstrated in human implants. We present the histopathological and 114
immunohistochemical (IHC) findings of 6 human explants of CardioCel patches that were 115
used initially in operations for CHD in children. 116
METHODS 117
Subjects 118
Data for patients undergoing CardioCel implants from October 2012 to March 2015 was 119
retrieved from our departmental database. Patients who had explantation of CardioCel (partial 120
or complete) were identified. Inclusion criteria were (1) a congenital cardiac anomaly in 121
patients aged 1 day to 18 years, (2) CardioCel used as a cardiovascular tissue substitute, (3) 122
subsequent surgery (any indication) with explantation of CardioCel patch (partial or 123
complete). 124
Clinical data were collected for all patients from electronic health records and paper 125
charts. Patients were consented and ethics committee approval was obtained 126
(HREC/15/QRCH/93). 127
Histologic examination 128
CardioCel explanted at re-operation was fixed in formalin. Hematoxylin and eosin (H&E) 129
stain was used for assessment of grade and type of inflammation, type of inflammatory cells, 130
including eosinophil response, degree of inflammatory cell infiltration into CardioCel 131
collagen layers, neovascularization and neo-intima formation. Collagen characteristics were 132
Page 8 of 22
9
studied using H&E and Masson’s trichrome stains. All specimens underwent 133
immunohistochemical staining. Endothelial cells were identified using antibodies to CD34 134
and Factor-VIII. IHC was also used to discern the subtypes of inflammatory cells, which 135
included T lymphocytes (CD3), helper T lymphocytes (CD4), suppressor T lymphocytes 136
(CD8), B lymphocytes (CD20), plasma cells (CD79a), macrophages and histiocytes (CD68). 137
Unimplanted CardioCel was used as control and after fixation in formalin was studied 138
using H&E and Masson’s trichrome stains. 139
For the purpose of description of the histological response to CardioCel, we divided the 140
surgical specimen into 3 zones:(1) native tissue response on the parietal (rough/outer) surface 141
of the graft, (2) reactive change within the CardioCel graft itself and (3) native tissue 142
response on the visceral (smooth/inner) surface of the graft. 143
RESULTS 144
During the study period, 140 patients underwent various cardiovascular procedures in 145
which CardioCel was used as a cardiovascular substitute. Fifteen patients (10.7%) underwent 146
re-operations (excluding mediastinal exploration for bleeding). The CardioCel patch was 147
explanted (1 partial and 5 complete) in 6 patients (4%). The indications for re-operation were 148
related to the CardioCel patch in 2 patients and not related to CardioCel in the remaining 4 149
patients. 150
Subjects 151
Case 1 (10 days in situ): An 8 year old child presenting with severe mitral insufficiency 152
underwent mitral valve repair. The posterior mitral leaflet (PML) was enlarged with a 153
CardioCel patch. On the 8th postoperative day (POD), the patient was re-operated for severe 154
mitral insufficiency and low cardiac output syndrome secondary to an atrial tachyarrhythmia 155
Page 9 of 22
10
resistant to amiodarone and cardioversion. Mitral valve replacement with a 25mm bileaflet 156
mechanical prosthesis with chordal preservation, de Vega tricuspid valve annuloplasty and 157
radiofrequency bi-atrial maze were performed. The CardioCel patch used for augmentation of 158
the PML was explanted. 159
Case 2 (67 days in situ): A 3-day old neonate with obstructed mixed total anomalous 160
pulmonary venous drainage (TAPVD) underwent repair on day 3 of life with a Warden 161
procedure, atrial septectomy and construction of an intra-atrial baffle using CardioCel. At 60 162
days follow up, echocardiography, cardiac catheterization and computerized tomography (CT 163
scan) demonstrated severe obstruction within the intra-atrial baffle and suprasystemic right 164
ventricular (RV) systolic pressures. At re-operation, the patch of CardioCel used to create the 165
intra-atrial baffle was noted to be thickened and shrunken. The patch was completely 166
explanted and replaced with glutaraldehyde treated homograft pericardium. 167
Case 3 (134 days in situ): A 3 day old neonate underwent a Yasui operation for type-B 168
interrupted aortic arch (IAA) with ventricular septal defect (VSD) and subaortic obstruction. 169
The IAA was repaired by direct anastomoses of the descending aorta to the ascending aorta 170
posteriorly with anterior augmentation using a CardioCel patch. The child subsequently 171
underwent 2 attempts at balloon angioplasty for severe stenosis of the arch anastomosis 172
before undergoing re-operation at 5 months. There was a thick neo-intima related to the 173
CardioCel patch. The CardioCel patch was explanted and replaced with pulmonary artery 174
homograft. The neo-intima got separated and was not sent for histological examination. 175
Case 4 (272 days in situ): A 4 month old child with tetralogy of Fallot (TOF) underwent 176
takedown of a Blalock-Taussig shunt and repair of TOF with augmentation of the anterior 177
pulmonary valve leaflet using glutaraldehyde treated autologous pericardium. CardioCel was 178
used to close the VSD and 2 separate additional patches were used to augment the left 179
Page 10 of 22
11
pulmonary artery (LPA) and the transannular right ventricular outflow tract (RVOT). The 180
patient was re-operated at 12 months for severe pulmonary regurgitation and severe tricuspid 181
regurgitation. CT scan demonstrated focal narrowing of the origin of the LPA. Dense 182
pericardial adhesions were noted between native pericardium and the transannular CardioCel 183
patch (Figure 1A). The neo-pulmonary valve was stiff and immobile and completely covered 184
with neo-intima on both surfaces. The neo-intima extended under the transannular CardioCel 185
patch into the left pulmonary artery. The neo-intima was loosely adherent to the overlying 186
CardioCel patch (Figure 1B). There was a shelf of neo-intima causing localized narrowing at 187
the origin of the LPA where the two CardioCel patches used for the transannular repair and 188
the LPA repair were sutured together. Both CardioCel patches and autologous pericardial 189
neo-pulmonary valve were completely explanted. 190
Case 5 (428 days in situ): A child with common arterial trunk underwent repair on day 40 of 191
life (detachment of branch pulmonary arteries [PA] from common arterial trunk, closure of 192
VSD, closure of inter atrial communication and interposition of 12mm valved conduit from 193
RV to PA). The child was re-operated at 12 months for severe distal RV-PA conduit 194
obstruction. The PA bifurcation was reconstructed using a CardioCel patch and the RV-PA 195
conduit was replaced with a 15mm pulmonary homograft using an anterior hood of 196
CardioCel for the proximal anastomosis. Thirteen months later the patient presented with 197
severe narrowing of the proximal segment of the pulmonary homograft without stenosis of 198
the branch PAs. Balloon angioplasty caused a contained rupture necessitating reoperation. At 199
reoperation there was no narrowing of the anastomosis between the homograft and the 200
reconstructed PA bifurcation. The luminal surface of the CardioCel patch at the bifurcation 201
was covered with neo-intima. The bifurcation was enlarged by resecting a portion of the 202
CardioCel patch. 203
Page 11 of 22
12
Case 6 (augmented anterior pulmonary valve leaflet in situ for 292 days, transannular patch 204
[TAP] -peripheral section 502 days, central section 292 days): A 3 month old child 205
underwent primary repair of TOF using CardioCel for the TAP and closure of the VSD. The 206
child was reoperated 7 months later for a residual VSD and pulmonary insufficiency. 207
CardioCel was used to augment the anterior pulmonary leaflet and the TAP centrally. The 208
child underwent a second reoperation 9 months later for severe pulmonary and tricuspid 209
insufficiency and origin stenosis of both branch PAs. The entire valve complex and TAP 210
were explanted. 211
Histological examination 212
CardioCel controls were uniformly 400 µm in thickness. Two distinct surfaces were 213
identifiable; a smoother visceral surface and a rougher parietal surface. Native collagen fibres 214
were arranged in layers and ‘ghost vessels’ (remnants of bovine blood vessels) were seen 215
(Figure 2). 216
In our seriesthe parietal surface of each of the explanted CardioCel grafts was covered by 217
a layer of granulation tissue showing neovascularization and containing fibroblasts and a mild 218
to moderate inflammatory response (Figure 3). The thickness of this layer varied from 0.1mm 219
(Case 1) through to approximately 1mm (e.g., Cases 2, 5 and 6) and did not appear to be 220
temporally related. IHC showed that the infiltrating lymphocytes were predominantly T 221
lymphocytes (CD3 positive) with a lesser B lymphocyte infiltrate (CD20 positive; Figure 3 - 222
B1, B2), apart from Case 4 (which included a foreign body multinucleate giant cell response) 223
where T and B lymphocytes were approximately equal in number. A semi-quantitative 224
assessment revealed that there were either equal numbers of T helper and suppressor cells 225
(CD4 and CD8 positive, respectively) or an excess of T helper cells compared with T 226
suppressor cells. None of the cases had an excess of T suppressor cells. Plasma cells were 227
Page 12 of 22
13
sparse in each case. Eosinophils were present in all cases, but the quantity varied (from 228
<1/high power field in Case 5 to >100/high power field in Case 6) and the degree of 229
eosinophil infiltrate did not appear to be temporally related. 230
The macrophage inflammatory component varied from mild to marked within the 231
thickened surface layer. In Case 2 and Case 4, CD68-positive macrophages were arranged 232
between the granulation tissue and the graft and in Case 4 there was a classical foreign body 233
giant cell response (Figure 4A, 4B& 4C). This case was the only case in which dystrophic 234
calcification was identified (Figure 4D). This calcification was closely aligned with the 235
macrophages, between the granulation tissue layer and the CardioCel graft. Calcium was not 236
seen within the graft material in any of the explants. There was no infiltration of any type of 237
inflammatory cell into the collagen laminae of any of the CardioCel explants. 238
In all but one case (Case 1) there was an identifiable neo-intima on the visceral surface of 239
the CardioCel patch (this was not sent for histology in Case 3). This layer was composed of 240
neo-vascularised fibromyxoid substance with interposed fibroblasts (Figure 5A). There was 241
minimal to no inflammation within this layer. The thickness of this layer was always less than 242
that seen on the opposite, parietal surface of the graft and measured between 0.02mm (Case 243
4) and 0.4mm (Case 6). Using IHC, we were able to identify endothelial cells in the explant 244
with the longest insitu implant duration (Case 6; Figure 5). CD34 and Factor VIII IHC 245
positive endothelial cells were identified on the surface of the neo-intima covering the 246
luminal surface (blood interface) of the TAP. Endothelial cells were also identified on both 247
neo-intimal surfaces of the CardioCel patch which was used to augment the anterior leaflet of 248
the pulmonary valve (explanted on day 292). In Case 1 (explanted on day 10), no neo-intima 249
was identified. However, a single, discontinuous layer of cells were identified on the visceral 250
surface. 251
Page 13 of 22
14
In general, there was no patch breakdown and the architecture of laminated native 252
collagen fiberswas maintained within all 6 human explants. Neovascularization adjacent to 253
the CardioCel patch was seen in all explants. Variable fibroblastic invasion was identified in 254
all six cases (Figure 6), including Case 1, where the graft had been in place for only 10 days 255
before explantation. This was associated with scattered capillary vessels (Figure 5). The 256
degree and extent of fibroblast infiltration into the graft appeared, to increase with the age of 257
the implant. Indeed, the longest implant (Case 6) had a fibroblastic infiltrate throughout all 258
layers of the CardioCel graft (Figure 6). A function of fibroblasts is to produce collagen 259
fibres and these were seen lying between the CardioCel laminated collagen bundles in 260
association with the fibroblastic infiltrate in the longest-standing graft (Case 6) (Figure 5). 261
This finding is consistent with early remodeling in the CardioCel patch. 262
DISCUSSION 263
CardioCel is in an early phase of evaluation as a cardiovascular substitute in humans 264
(implantation began in 2008). The Queensland Paediatric Cardiac Service (QPCS) began 265
using CardioCel in October, 2012. 266
In a study from Melbourne, Brizard et al 7 implanted CardioCel in a subcutaneous pocket 267
created in the dorsal area of 6-week old male Albino Wistar rats. The rats were sacrificed at 8 268
(n=5) and 16 (n=5) weeks. The CardioCel patches demonstrated improved biostability (tested 269
using shrinkage temperature) and durability (tested using resistance to enzymatic 270
degradation), as well as reduced calcification compared with bovine pericardium fixed with 271
0.6% glutaraldehyde, cryopreserved human pericardium rapidly fixed with 0.6% 272
glutaraldehyde and ADAPT treated cryopreserved human pericardium. Histology at 8 weeks 273
(n=5) demonstrated host fibroblasts in the collagen matrix. By 16 weeks (n=5) there was an 274
increased number of host fibroblasts and the presence of functioning neo-capillaries on the 275
Page 14 of 22
15
edges of the explant without visible calcification. In a second study by the same investigators 276
CardioCel was used for reconstruction of mitral valve (PML was replaced by CardioCel) and 277
pulmonary valve (one of theleaflets was replaced by CardioCel) in a juvenile sheep model 8. 278
After 7 months of implantation there was a continuous endothelial lining visible on the blood 279
interface. A few neo-capillaries, myofibroblasts, monocytes, and cells with smooth muscle 280
cell phenotype were visible within the CardioCel. There was no calcium deposition or 281
thickening of the implant. More recently CardioCel was used for complete trileaflet 282
replacement of the aortic valve in a surgical sheep model 9. At 6 months 8 of 9 leaflets were 283
intact and pliable with diminished mobility and macroscopic calcification of only one cusp. 284
Histology identified microscopic calcification of2 additional cusps. The original CardioCel 285
structure was well preserved. Large portions of the leaflets were covered with a neo-intima 286
on both surfaces with endothelialization of many parts of the neo-intima. There was 287
infiltration of new fibroblasts into the CardioCel at the base of the cusps. There were 288
macrophages within the neo-intima but no inflammatory cells were demonstrated within the 289
CardioCel. 290
A South African single center, prospective, non-randomized clinical study with 30 291
pediatric patients with various cardiac defects, demonstrated that there were no implant 292
related events leading to death or morbidity in the 30-day post-operative period 10
. 293
Echocardiography assessment at 6 and 12 months showed intact anatomic and 294
hemodynamically stable repairs, and no visible calcification. The authors concluded that 295
CardioCel can be safely and efficaciously used as a cardiovascular substitute for surgical 296
repair of both simple and complex congenital cardiac defects. In our series of 140 patients, 6 297
(4.2%) had CardioCel explanted at subsequent reoperation. CardioCel appeared to perform 298
satisfactorily from an anatomic and hemodynamic standpoint in the majority of patients. In 2 299
(cases 2, 3) CardioCel was mainly responsible for the indication for reoperation and in this 300
Page 15 of 22
16
sense these 2 patients can be considered to represent failures of the prosthesis. Both patients 301
were neonates. We have used CardioCel in 15 additional neonates (20 implants) [arch repair 302
(n=7), PA reconstruction (n=5), VSD patch (n=5), intra-ventricular baffle in DORV (n=1), 303
ASD patch in arterial switch operation (n=1) and aorto-pulmonary window (n=1)]. The 304
overall failure probability is comparable to other prosthetic materials 11
. 305
In our study there was no inflammatory infiltrate into any of the grafts even after 502 days 306
of implantation. The inflammatory response appeared to be confined to the granulation tissue 307
which developed on the graft’s parietal surface;the thickness of whichwas variable and 308
exceeded the thickness of the CardioCel patch in all cases except Case 1. The thickness of 309
this layer did not appear to be temporally related though measurement of thickness can be 310
affected by tangential embedding during sectioning. In 5 of our cases there was an 311
identifiable neo-intima on the visceral surface of the graft. The neo-intima was uniformly 312
thinner than the granulation tissue layer on the opposite surface and there was minimal to no 313
inflammation within this layer. This demonstrates that there is a graft dependent different 314
host response on its parietal compared with its visceral surface regardless of the site of 315
implantation, intra-cardiac or extra-cardiac position or exposure to systemic or pulmonary 316
pressures. When used as an intra-cardiac baffle in neonates, it is therefore possible that this 317
layer could contribute to luminal compromise as seen in one of our cases (Cases 2). We 318
recommend that CardioCel is implanted with the visceral surface designated as the luminal 319
surface wherever possible. 320
In the study on juvenile sheep cited above, no histological evidence of calcification was 321
noted after 210 days 8. In the study from South Africa, post-operative echocardiography at 6 322
and 12 months did not show calcification 10
;however echocardiography will not identify the 323
presence of microscopic calcium deposit. We observed focal calcification in 1 of our patients 324
Page 16 of 22
17
(case 4, 272 days Figure 4C). The calcification was aligned with macrophages and situated 325
between the granulation tissue layer covering the parietal surface and the CardioCel graft 326
used as TAP. As this reaction was present only on the external surface it is possible that it 327
represents a more generalized post-operative pericarditis than a foreign body reaction to 328
CardioCel itself (12, 13). The calcification did not have any functional consequence as seen 329
in the aortic valve leaflet in the sheep model 9. Calcification was not demonstrated in any of 330
the other 5 implants, 2 of which were in situ for longer periods of time (428 days and 502 331
days). 332
CardioCel is intended for use as a biological scaffold. Its durability and strength as 333
compared to autologous pericardium has been previously demonstrated in animal studies 7. 334
Previous human and animal studies 8, 9, 10
have demonstrated its capacity to provide structural 335
integrity when used for cardiac repairs. Evidence of early remodeling has been demonstrated 336
in animal studies 7, 8, 9
but has never been previously demonstrated in humans. Development 337
of a neo-intima is important for the remodeling process. A recognizable layer of neo-intima 338
was seen on the surface of all our explants, excepting Case 1, which had only been in situ for 339
10 days. In one case the neo-intima was peeled off at the time of explantation and could not 340
be demonstrated on histology.In the juvenile sheep model cited above, there was a continuous 341
endothelial lining visible on the external blood interface after 7 months 8. In the surgical 342
sheep model many parts of the neo-intima covering the CardioCel valve leaflets were 343
endothelialized by 6 months9. We demonstrated endothelial cell formation on one implant 344
with the longest duration of implantation (16 months). Endothelial cells are very fragile and 345
can be easily disrupted during histopathological sampling. It is possible that the development 346
of an endothelial cell lining in humans is temporally related to the number days in situ and to 347
the site of implantation with quicker endothelialization in valve reconstruction. Both animal 348
models cited above demonstrated the development of neo-capillaries on the edges with 349
Page 17 of 22
18
migration of the animal’s own fibroblasts into the CardioCel scaffold 7, 9
. All our explants 350
demonstrated fibroblastic infiltration associated with neovascularization that appeared to be 351
temporally related. In the longest implant (Case 6) there was a fibroblastic infiltrate 352
throughout all layers of the CardioCel graft (Figure 6) and new collagen fibres were seen 353
lying between the laminated collagen bundles of CardioCel. This finding is consistent with 354
early remodeling in the CardioCel patch. 355
Complete remodeling into a three-layered vessel wall, valve or inter-atrial septum was not 356
seen in our series. It is possible that this is a time-related phenomenon that can be only be 357
proved or disproved from further explants. 358
CONCLUSION 359
We demonstrate that the CardioCel graft incites a thick granulation tissue response on its 360
parietal surface and a thinner, neo-intima on its visceral surface, a finding which may be of 361
importance when considering the site of implant of a CardioCel graft. The inflammatory 362
response to CardioCel is limited to the granulation tissue response and there was no 363
inflammatory cell infiltration into the graft collagen lamina. Calcification did not occur in the 364
CardioCel patch. Temporally related fibroblast ingrowth into the graft was shown to be a 365
uniform characteristic. This feature, together with the development of a neo-intimal layer 366
indicates thatearly remodelingwas occurring in our explanted CardioCel grafts. 367
ACKNOWLEDGEMENT 368
We thank Janelle Johnson, for her help with data management. 369
370
Page 18 of 22
19
REFERENCES 371
1. Ionescu MI, Tandon AP, Mary DA, Abid A. Heart valve replacement with the 372
Ionescu-Shiley pericardial xenograft. J Thorac Cardiovasc Surg 1977; 7: 31-42 373
2. Strange G, Brizard CP, Karl TR, Neethling L. An evaluation of Admedus’ tissue 374
engineering process treated (ADAPT) bovine pericardium patch (CardioCel) for the 375
repair of cardiac and vascular defects. Expert Rev Med Devices 2015; 12: 135-141 376
3. Golomb G, Schoen FJ, Smith MS, Linden J, Dixon M, Levy RJ. The role of 377
glutaraldehyde-induced cross-links in calcification of bovine pericardium used in 378
cardiac valve bioprostheses. Am J Pathol 1987; 127: 122-130 379
4. Vincentelli A, Latremouille C, Zegdi R et al. Does glutaraldehyde induce calcification 380
of bioprosthetic tissues? Ann Thorac Surg 1998; 66: S255-258 381
5. Neethling WM, Hodge AJ, Clode P, Glancy R. A multi-step approach in anti-382
calcification of glutaraldehyde-preserved bovine pericardium. J Cardiovasc Surg 383
(Torino) 2006; 47: 711-718 384
6. Neethling WM, Yadav S, Hodge AJ, Glancy R. Enhanced biostability and 385
biocompatibility of decellularized bovine pericardium, crosslinked with an ultra-low 386
concentration monomeric aldehyde and treated with ADAPT. J Heart Valve Dis 2008; 387
17: 456-463 388
7. Neethling WML, Brizard CP, Firth L, Glancy R. Biostability, durability and 389
calcification of cryopreserved human pericardium after rapid glutaraldehyde 390
stabilization versus multistep ADAPT treatment in a subcutaneous rat model. Eur J 391
Cardiothorac Surg 2014; 45: e110-e117 392
8. Brizard CP, Brink J, Horton SB, Neethling WML et al. New engineering treatment of 393
bovine pericardium confers outstanding resistance to calcification in mitral and 394
Page 19 of 22
20
pulmonary implantation in a juvenile sheep model. J Thorac Cardiovasc Surg 2014; 395
148: 3194-3201 396
9. Meuris B, Ozaki S, Neethling W, Vleeschauwer SD, Verbeken E, Rhodes D, 397
Verbrugghe P, Strange G. Tri-leaflet aortic valve reconstruction with a decellularized 398
pericardial patch in a Sheep Model. J Thorac Cardiovasc Surg 2016; DOI: 399
10.1016/j.jtcvs.2016.05.024 400
10. Neethling WML, Strange G, Firth L, Smit FE. Evaluation of a tissue engineered 401
bovine pericardial patch in paediatric patients with congenital cardiac anomalies: 402
initial experience with the ADAPT treated CardioCel patch. Interactive 403
CardioVascular and Thoracic Surgery 2013; 17: 698-703 404
11. Schreiber C, Eicken A, Vogt M et al. Repair of interrupted aortic arch: results after 405
more than 20 years. Ann Thorac Surg 2000; 70: 1896-1900 406
12. Oh KY, Shimizu M, Edwards WD, Tazelaar HD, Danielson GK. Surgical pathology 407
of the parietal pericardium: A study of 344 cases (1993-1999). Cardiovasc Pathol 408
2001; 10: 157-168 409
13. Talreja DR, Edwards WD, Danielson GK et al. Constrictive pericarditis in 26 patients 410
with histologically normal pericardial thickness. Circulation 2003; 108: 1852-1857 411
412
413
Page 20 of 22
21
FIGURE LEGENDS 414
Figure 1: Intraoperative photographs (case 4): A - Fairly dense pericardial adhesions between 415
native pericardium and the transannular CardioCel patch. B - Loosely adherent neo-intima on 416
the luminal surface of the transannular CardioCel patch. 417
Figure 2: Control: Non-implanted CardioCel graft showing laminated collagen fibres. One 418
surface is rough (upper) i.e. parietal, whilst the opposite is smooth, i.e. visceral (A: H&E, 419
original magnification x20). Non-implanted CardioCel graft in which residual bovine 420
vascular outlines can be identified (B: Masson’s Trichrome, original magnification x 20). A 421
smooth (lower, visceral) and rough (upper, parietal) surface are again noted (* Remnants of 422
bovine blood vessels - ghost vessels). 423
Figure 3: A. Photomicrographs of all 6 patients (H&E, original magnification x 10) 424
demonstrating response to the parietal (rough) surface in the form of granulation tissue of 425
variable thickness (arrow) which contained an inflammatory infiltrate, predominantly T 426
lymphocytes, with a lesser population of B lymphocytes (B1- T lymphocytes demonstrated 427
by CD3 IHC and B2 - B lymphocytes demonstrated by CD20 IHC, original magnifications x 428
20). Neo-intima (*) formation on the visceral (smooth) surface can be seen in cases 2, 4, 5 429
and 6. No discernible neo-intima was seen in case 1 and in case 3 the neo-intima was 430
separated intraoperatively and not sent for histology. 431
Figure 4: Case 4, transannular patch A and B: Macrophages arranged between the granulation 432
tissue and the graft, with a classical foreign body giant cell response (* ; H&E, original 433
magnification x 20). C: Histiocytes, including multi-nucleated foreign body giant cells, 434
expressing the macrophage marker CD68 (anti-CD68 IHC, original magnification x 40). D: 435
Dystrophic calcification (arrows) was closely aligned with the macrophages, between the 436
granulation tissue layer and CardioCel. 437
Figure 5: Remodeling (case 6, 502 days) 438
A: CardioCel transannular patch graft with neo-intima (lower) on the visceral side of the graft 439
and a thicker granulation tissue layer, including an inflammatory cell infiltrate, on parietal 440
surface of the graft (H&E, original magnification x10). B: Remodeling of the graft seen as 441
fibroblastic infiltrate (†), surrounded by new pale, eosinophilic collagen (*) between the 442
brightly eosinophilic graft collagen layers. Neovascularisation (∇) is also demonstrated 443
Page 21 of 22
22
(H&E, original magnification x40). C: Endothelium (↑) extending into the graft (Factor VIII 444
IHC, original magnification x 10). D: Neo-intima which has formed on the visceral surface of 445
the graft showing a surface layer of endothelial cells (↑) (CD34 IHC, original magnification x 446
20). E: Native pulmonary valve leaflet composed of fibromyxoid tissue with interspersed 447
fibroblasts and overlying endothelial cells. There is no inflammatory cell infiltrate nor 448
neovascularisation (H&E, original magnification x 10). 449
Figure 6: Fibroblastic infiltration Photomicrographs of all 6 patients (H&E, original 450
magnification x 20) demonstrating unchanged thickness of CardioCel patches and temporally 451
related fibroblast infiltration (arrow). Infiltration of fibroblasts seen as early as 10 days (case 452
1); and the longest implant (506 days, case 6) demonstrating fibroblastic infiltration 453
throughout the layers of the CardioCel patch 454
455
456
Page 22 of 22