In-vivo performance of high-density collagen gel tubes for urethral regeneration in a rabbit model

In-vivo performance of high-density collagen gel tubes for urethral regeneration in a rabbit model

Biomaterials 33 (2012) 7447e7455 Contents lists available at SciVerse ScienceDirect Biomaterials journal homepage:

2MB Sizes 1 Downloads 127 Views

Biomaterials 33 (2012) 7447e7455

Contents lists available at SciVerse ScienceDirect

Biomaterials journal homepage:

In-vivo performance of high-density collagen gel tubes for urethral regeneration in a rabbit model Lionel A. Micol a, Luis F. Arenas da Silva b, Paul J. Geutjes b, Egbert Oosterwijk b, Jeffrey A. Hubbell a, Wout F.J. Feitz b, Peter Frey a, c, * a

EPFL, Laboratory for Regenerative Medicine and Pharmacobiology, Institute for Bioengineering, School of Life Sciences and School of Engineering, Ecole Polytechnique Fédérale de Lausanne (EPFL), CH-1015 Lausanne, Switzerland RUNMC, Department of Urology, Radboud University Nijmegen Medical Centre, 6500 HB Nijmegen, The Netherlands c CHUV, Department of Pediatric Urology, Rue du Bugnon 46, 1011 Lausanne, Switzerland b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 23 April 2012 Accepted 28 June 2012 Available online 15 July 2012

Congenital malformations or injuries of the urethra can be treated using existing autologous tissue, but these procedures are sometimes associated with severe complications. Therefore, tissue engineering may be advantageous for generating urethral grafts. We evaluated engineered high-density collagen gel tubes as urethral grafts in 16 male New Zealand white rabbits. The constructs were either acellular or seeded with autologous smooth muscle cells, isolated from an open bladder biopsy. After the formation of a urethral defect by excision, the tissue-engineered grafts were interposed between the remaining urethral ends. No catheter was placed postoperatively. The animals were evaluated at 1 or 3 months by contrast urethrography and histological examination. Comparing the graft caliber to the control urethra at 3 months, a larger caliber was found in the cell-seeded grafts (96.6% of the normal caliber) than in the acellular grafts (42.3%). Histology of acellular and cell-seeded grafts did not show any sign of inflammation, and spontaneous regrowth of urothelium could be demonstrated in all grafts. Urethral fistulae, sometimes associated with stenosis, were observed, which might be prevented by urethral catheter application. High-density collagen gel tubes may be clinically useful as an effective treatment of congenital and acquired urethral pathologies. Ó 2012 Elsevier Ltd. All rights reserved.

Keywords: Collagen Smooth muscle cell Urinary tract In-vivo test Hydrogel

1. Introduction Several congenital disorders of the urethra such as hypospadias or epispadias, as well as acquired pathologies such as urethral strictures, can severely impair its function. Hypospadias is a frequent genital birth defect, requiring surgical reconstruction and affecting about 1 in 300 male newborns in Europe [1]. Urethral stricture affects 1 in 1000 men after their sixties and is usually treated by incision and dilation, although a recent study suggests that a reconstructive procedure by urethroplasty should be considered more often [2]. Such repair is usually undertaken by tubularizing the urethral plate, the foreskin or the buccal mucosa. However, if the urethral plate is underdeveloped or the latter two

* Corresponding author. EPFL, Laboratory for Regenerative Medicine and Pharmacobiology, Institute for Bioengineering, School of Life Sciences and School of Engineering, Station 15, Ecole Polytechnique Fédérale de Lausanne (EPFL), CH-1015 Lausanne, Switzerland. Tel.: þ41 21 693 96 62; fax: þ41 21 693 96 85. E-mail addresses:, [email protected] (P. Frey). 0142-9612/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved.

tissues are used, grafting can be associated with severe long-term complications; urethroplasties have been shown to lead to shortor long-term complications in up to 54% of the cases [3]. Therefore, tissue-engineered urethral grafts may offer surgical alternatives with potentially improved outcomes [4]. Several examples of pre-clinical studies in this field have shown that cell-seeding of tubular matrices is necessary for healthy regeneration of the urethra without subsequent stricture. In 2007, Fu et al. [5] described an experimental model of urethral defect repair, using collagen matrices of bladder submucosal origin. They compared matrices seeded with foreskin epidermal cells with acellular matrices and their results clearly showed that only cellseeded matrices maintained a wide urethral caliber without stricture formation. In 2010, Nuininga et al. [6] reported, also in a rabbit model, the regeneration of a urethral defect with acellular, growth factor-loaded type I collagen tubular matrices. These matrices showed a better tissue regeneration, despite a poorer functional outcome due to a relative narrowing of the urethral caliber, compared to matrices without inclusion of growth factors. In 2007, Fossum et al. [7] demonstrated hypospadias repair in human


L.A. Micol et al. / Biomaterials 33 (2012) 7447e7455

patients with autologous in vitro cultured urothelium. Also, RayaRivera et al. [8] reported a recent study of posterior urethra repair with autologous-cell-seeded PLGA grafts. The functional outcome of their five-year follow-up showed a good recovery of the voiding capacity as assessed by urodynamic measurements and demonstrated that tissue engineering is a viable alternative to classical reconstructive surgery. The production of cell-seeded constructs is, however, usually time-consuming, as a one- or two-week incubation period is necessary for the cells to populate the full matrix. We previously have developed high-density collagen gel tubes (hdCGTs) that are suitable for urinary tract tissue engineering and can be homogenously cell-seeded and ready for implantation within hours [9]. Hence, the hdCGTs with the seeding process have the potential to drastically shorten the production time of cell-seeded tissueengineered urinary tract grafts. To demonstrate their efficiency as urinary tract grafts, we have investigated the ability of hdCGTs to regenerate over 3 months an iatrogenic urethral defect in an in-vivo rabbit model. Their in-vivo performance was assessed at 1 and 3 months after implantation by urethrography, regarding the lumen caliber, and by histology for analysis of tissue regeneration and possible inflammatory reaction. 2. Material and methods 2.1. Animals In-vivo experiments were performed on 16 New Zealand white male rabbits (2.5e3.5 kg; Charles River Laboratories, Sulzfeld, D) distributed into four groups of four rabbits each. The animal experiment (with submission number RU-DEC 2010142) was approved by the Animal Ethics Committee of the Radboud University Nijmegen Medical Centre (Nijmegen, NL). The rabbits were housed individually in a temperature-controlled cage with 50e55% humidity for a 12 h lightedark cycle with free access to standard laboratory chow and tap water. After a minimum of 2wk quarantine period, the rabbits underwent bladder biopsy. 2.2. Preoperative preparation Antibiotic prophylaxis was carried out with 8 mg/kg Enrofloxacin (Bayer B.V., Mijnrecht, NL) administered by subcutaneous injection 15 min before induction of anesthesia. Anesthesia was induced by intramuscular injections of 0.1 mg/kg Medetomidine (Janssen Cilag GmbH, Neuss, D) and 10 mg/kg Ketamine (Eurovet animal health care B.V., Bladen, NL) and maintained with isoflurane 2% and 35% O2 after intubation and fixation of an endotracheal canula. This was followed by abdominal or genital shaving and surgical desinfection, for either bladder biopsy or graft implantation, respectively. Pain medication was given in the operation room combining intravenous injection of 20 mg/kg Buprenorphine (RB pharmaceuticals Ltd., Buckskin, UK) and subcutaneous injection of 4 mg/kg Carprofen (Pfizer animal health B.V., Capelle a.d. Ijssel, NL). 2.3. Bladder biopsy All 16 rabbits underwent bladder biopsy as follows. A 2-cm median skin incision was made starting 2 cm above the pubic bone. The linea alba was then opened on a 2cm length, through which the bladder was exposed and a w1  0.8  0.3-ccm fullthickness fusiform biopsy of the anterior bladder wall was excised using scissors (Fig. 1A). Each specimen was kept at 4  C in 2 ml of transport medium which was composed of Hanks balanced salts solution (HBSS) with Ca2þ and Mg2þ supplemented with 100 mM 4-(2-hydroxyethyl)-1 Piperazine ethane sulfonate (HEPES), 1% (w/v) PenicillineStreptomycin, and 0.1% (w/v) Aprotinin. All sutures were purchased from Ethicon (Johnson&Johnson Medical B.V., Amersfoort, NL). The bladder was then sutured in two layers (mucosa and detrusor muscle) with MonocrylÒ 6-0 by running sutures. The abdominal muscle layer was then adapted in the median plane with 2stitches of VicrylÒ 4-0 and the aponeurosis of the abdominal wall closed with VicrylÒ 4-0 continuous sutures. The subcutis was closed with two single VicrylÒ 4-0 stitches and the skin closure with intradermic continuous stitches of MonocrylÒ 5-0.

2.5. Cell isolation and culture Bladder biopsy specimens were transferred and kept at 4  C overnight in 2 ml stripper medium, which was composed of filtered HBSS without Ca2þ and Mg2þ, 10 mM HEPES, 0.1% Aprotinin, 1% Penicillin/Streptomycin, 2.4 U/ml dispase II (Roche Diagnostics, Almere, NL). This overnight step allowed for enzymatic digestion of the lamina basalis and the next day fat, vessels, and peritoneum were removed by micro-dissection followed by gentle removal of the mucosa with tweezers. The remaining tissue was minced in small parts and further incubated with 1.5 U Liberase blendzyme (Roche Diagnostics, Almere, NL) in 1 ml stripper medium at 37  C in 5% CO2 atmosphere for 90 min. The tissue extract was further filtered with a 70mm cell strainer, collected in 50 ml tube, and centrifuged at 500  g for 8 min. After the supernatant had been carefully discarded, the pellet was resuspended in 3 ml smooth muscle cell medium (SMCM; Sciencell, Carlsbad, USA) supplemented with 1% (w/v) penicillin-streptomycin, 2% (v/v) fetal bovine serum, and 1% (v/v) SMC growth supplement (Sciencell, Carlsbad, USA). Cells were then transferred into one T-25 culture flask per specimen, cultured separately for each rabbit with 3 media changes per wk, passaged once after 10 d, split into four T-75 culture flasks, and eventually trypsinized and counted with a hematocytometer before integration into the graft one day prior to implantation. 2.6. Formation of the collagen gel tube The urethral grafts used were high-density collagen gel tubes (hdCGT) formed as previously described [9]. Briefly, 10 MEM was added to 5 ml of 2.06 mg/ml sterile acid-soluble rat-tail type I collagen at a 1:9 ratio and the mix was neutralized with 1 M NaOH. Once neutralized, 300 ml of cell culture medium with (for cellular hdCGTs) or without (for acellular hdCGTs) 3.13$107 SMCs/ml was added within 2 s. The resulting solution was mixed and directly poured into the mold. After 30 min, allowing polymerization at room temperature, the resulting tubular collagen hydrogel was removed from its mold. To form compressed hdCGT with a collagen concentration of w30 mg/ml [9], the previously obtained tubular collagen hydrogel was kept on its mandrel and deposited on a nylon mesh lying, itself, on five layers of double-layer tissue paper (Weita, Arlesheim, CH). The nylon mesh and the tissue papers were folded around the tubular collagen hydrogel and then suspended to allow water extraction from the tubular collagen hydrogel for 30 min. All solid materials used above were autoclaved prior to use. After formation, hdCGTs were kept at 37  C in 5% CO2 atmosphere overnight, prior to implantation. The next day, a single 2 cm-long segment of each hdCGT was cut out and used as urethral graft for implantation (Fig. 1D and E). The applied grafts were 2 cm long, and had a 3 mm internal diameter. Shorter constructs could not be produced, as they would be unstable during one of the intermediate processing steps. In cellular grafts 3$106 bladder SMCs were incorporated. When the amount of cells obtained from bladder biopsies after 1 month of culture was insufficient to produce 5-cm long full-length cell-seeded hdCGTs, full-length hdCGTs were produced in order to comprise one cell-seeded end and one unseeded end. The latter was discarded, without any consequences for the quality of the remaining end, when, prior to graft implantation, a 2-cm segment was cut off and used as urethral graft. 2.7. Graft implantation Four wk after the bladder biopsy and 1 d after graft preparation, acellular or cellular urethral grafts were implanted in the rabbits with the following procedure adapted from previous work [6]. After preoperative preparation, the penis was exposed cranially using a ProleneÒ 6-0 holding suture, attached to the glans. An 8 F hydrophilic catheter (Lofric, Astra Tech, Mölndal, SE) was inserted through the urethra into the bladder. Under 2.5 optical magnification, a 1-cm incision was performed on the ventral surface of the penis, starting 1.5 cm proximally to the external meatus. Dissection of the corpus spongiosus followed, and a 1 cm urethral segment (and corpus spongiosus) also located 1.5 cm proximal to the external meatus was excised to create a critical-sized defect (Fig. 1B and C). The two urethral ends were marked with ProleneÒ 6-0 sutures. The catheter was withdrawn, however, left in the distal urethra, and a 2 cm-long acellular or cell-seeded urethral graft was cut inserted onto the catheter (Fig. 1D and E). After the catheter had been placed back into the bladder, the graft was placed at the location of the excised urethra and anastomosed to the ends of the native urethra using a straightforward sealing procedure. In brief, 0.2 ml of pre-warmed (37  C) fibrin glue (Tisseel, Baxter, Volketswil, CH) was deposited around each anastomosis (Fig. 1F). Finally, the skin was sutured with MonocrylÒ 6-0 continuous sutures (Fig. 1G) and the catheter was removed. All sutures were purchased from Ethicon (Johnson&Johnson Medical B.V., Amersfoot, NL). The postoperative care scheme was identical to that described for bladder biopsy surgery.

2.4. Postoperative antibiotic therapy and analgesia 2.8. Retrograde contrast urethrography Antibiotic therapy was continued for a total of 3 d with 5 subcutaneous injections every 12 h of 4 mg/kg Enrofloxacin and pain medication was continued with one single subcutaneous injection of 20 mg/kg Buprenorphine (RB pharmaceuticals Ltd., Buckskin, UK) 12 h after surgery and one daily subcutaneous injection of 4 mg/ kg Carprofen for 5 more days.

Animals were submitted to radiography and sacrificed 1 or 3 months after the urethral surgery, depending on the groups. Anaesthetized rabbits were positioned on their back, their penis was stretched caudally using a ProleneÒ 6-0 suture attached to the glans and a control radiography of the pelvic area was performed.

L.A. Micol et al. / Biomaterials 33 (2012) 7447e7455


Fig. 1. Macroscopic views illustrating bladder biopsy and graft implantation in the rabbit: A. Bladder biopsy. B. Defect formation by urethra dissection. C. Implantation site after urethral segment removal. D. Full collagen gel tube (hdCGT). E. A 2 cm-long hdCGT segment was used as a graft and was slid onto the LoFricÒ catheter. F. Fibrin glue was used to anastomose the graft to the two ends of the native urethra. G. View of penis after its closure, before removal of the catheter.

Then, a catheter was inserted into the urethral meatus and contrast medium, consisting of Iobitridol (Xenetix 300, Guerbet Nederland B.V., Gorinchem, NL) diluted with normal saline at a 1:1 ratio, was injected into the urethra to perform a retrograde urethrogram. All images were collected with a Philips BV-25 C-arm image intensifier (Philips, Eindhoven, NL). Thereafter, the animals were sacrificed using a lethal pentobarbital injection.

2.10. Statistics

2.9. Histology

3. Results

After sacrifice, the rabbit penis was cut at its base and fixed in 4% (v/v) formalin in phosphate buffer saline (PBS) at pH 7.4 for 24 h. Formalin-fixed samples were embedded in paraffin and serial sections of 8 mm were generated and stained with hematoxylin and eosin or Masson’s trichrome [10]. Samples were analyzed by bright-field microscopy with a Leica M205C stereomicroscope set at 0.78 magnification (Olympus, Volketswil, CH) for overviews and an Olympus AX70 microscope used with its 2 and 10 sample lenses (Olympus, Volkeltswil, CH) for microscopic views.

Statistical analysis of the data was performed, using Prism 5 (GraphPad Software, La Jolla, USA) for Windows. Significant differences between data sets were determined by one-way and two-way ANOVA as well as one sample t-test. p < 0.05 was considered significant: *p < 0.05.

All rabbits survived bladder biopsy and urethral graft implantation. At the time of harvesting bladder biopsies for smooth muscle cell isolation and culture no complications were observed, except macroscopic hematuria in one rabbit that resolved spontaneously after 2 d. The mean specimen weight was 231.5 mg (S.D.: 56.9 mg). Each bladder biopsy allowed acquisition of 8.7$106 (range


L.A. Micol et al. / Biomaterials 33 (2012) 7447e7455

3.2e17.4$106) bladder SMCs after 1 month in culture, as confirmed by immunohistochemistry for alpha-smooth muscle actin. SMCs were homogenously distributed in the obtained hdCGTs as assessed by control histology at the time of implantation (data not shown) which was in accordance with our previous study [9]. All rabbits survived urethral graft implantation and could void within 48 h after surgery, except one that voided 72 h after urethral graft implantation without subsequent complications.

62.69  18.1% and 42.4  34.3% of the normal urethral caliber, respectively. Regarding cell-seeded grafts a similar caliber was found at 1 month (45.7  43.62%), while at the 3 months evaluation of the lumen caliber was very close to the one of the normal urethra (96.6  12.5%). The values for acellular grafts were statistically different (p < 0.05) from the theoretical mean of 100%, which was not the case for cell-seeded grafts. 3.2. Microscopic appearance of urethral tissue component

3.1. Analysis of urethrograms Retrograde contrast urethrography was performed to analyze the lumen of the grafts and to measure their respective caliber. The obtained urethrograms allowed for the detection of complications ranging from fistulae to stenosis (as illustrated in Fig. 2). Of all 16 investigated rabbits, 5 had a normal urethrogram. In the remaining 11 rabbits partial narrowing of the urethra was observed 7 times, associated 3 times with a fistula. Isolated fistulae were observed twice and two rabbits showed complete stenosis, both associated with fistulae (in bar graph presented in Fig. 3A). The rabbits implanted with a cellular graft and evaluated at 3 months showed the fewest complications, as only a single fistula was detected in this group. In all other groups a minimum of three out of four rabbits had an abnormal urethrogram. The narrowest lumen calibers of the grafts were measured on the urethrograms (Fig. 3B). Comparing the average graft caliber to a control urethra, acellular grafts evaluated at 1 and 3 months were

Urethral graft remodeling of the collected rabbit penises was analyzed by bright-field light microscopy. In particular urothelial lining, lumen shape, smooth muscle, and connective tissue regeneration, as well as inflammatory response were evaluated. Hematoxylin-eosin (HE) staining was performed to study the presence of epithelial and inflammatory cells, while Masson’s trichrome (MT) staining allowed highlighting of extracellular matrix distribution and smooth muscle cell (SMC) ingrowth as well as detection of urothelial umbrella cells. In the applied MT staining extracellular matrix appears light to dark blue, endothelial cells and nuclei purple, smooth muscle and epithelial cells dark pink, and red blood cells light red, while umbrella cells and keratin are stained in dark red. 3.3. Microscopic analysis of controls Given the sparse quality images of rabbit urethra available in literature, we first analyzed normal histology of the rabbit urethra.

Fig. 2. Urethrograms: A. Control urethrogram from a rabbit that did not undergo critical defect formation. B. Negative image of the same urethrogram as in A with two white lines delimiting the side of the penis and white arrows pointing at the 1 cm-spaced ladder (1), the base of the penis (2), the apex of the penis (3), and the catheter (4). C. Urethrogram at 3 months after cellular graft implantation with forceps demarcating the graft implantation site. D. Urethrogram at 3 months after cellular graft implantation, showing a mild narrowing of the graft caliber. E. Urethrogram at 3 months after cellular graft implantation, revealing the presence of a fistula (5). F. Urethrogram at 1 month after cellular graft implantation, revealing a complete stenosis (6) of the graft. In figures CeF black arrowheads delimitate the graft.

L.A. Micol et al. / Biomaterials 33 (2012) 7447e7455

Fig. 3. Rabbit urethrogram findings at evaluation. A. Complications revealed by urethrogram: no complication (white), urethral narrowing (<66.7% control; gray), complete stenosis (black), fistula (hatching). B. Graft patency at evaluation. Narrowest lumen calibers of grafts were measured on urethrograms at one or three months and are plotted as a percentage of the urethra caliber of control rabbits not submitted to any surgery. Error bars represent standard deviations (n ¼ 4). Statistical significance (one sample t-test towards the theoretical mean of 100%) is shown as follow: *p < 0.05.

In control rabbits, the native urethral lumen was star-shaped and bordered by a typical urothelial lining under which a thin lamina propria was present and surrounded by longitudinal SMC bundles. Around this layer of longitudinal SMC bundles several lacunar vascular structures were observed in the encompassing connective tissue layer. Externally, sparse circular SMC bundles were also apparent. These features are depicted in Figs. 4 and 5A and B. 3.4. Microscopic analysis of graft lumina In longitudinal-sections of the samples collected at 1 and 3 months, a straight and open lumen was conserved in most cases all along the graft (Fig. 4CeF and I and J). In cross-sections, a centered and star-shaped lumen was distinguishable (Fig. 5CeF and I and J). Other findings, such as a discontinuous lumen in longitudinalsections (Fig. 4G and H) or a lateralized and rather circular lumen in


cross-sections (Fig. 5G and H) were, however, observed in some cases. At times, a communication between the lumen and the ventral penile skin (Fig. 4E and F), or a slanted lumen, oriented perpendicularly to the penile axis (Fig. 4G and H), was also identified. Despite the diverse microscopic aspect of graft lumina, a straight, centered, and star-shaped lumen was usually present at the graft site. The luminal side of all grafts was inspected for the presence of epithelial cells. As illustrated in Fig. 4CeJ, an epithelium could always be observed, which was newly lining the full-length of the luminal side of the implanted grafts. This epithelium presented the typical histological pattern of the urothelium, including 3e5 cell layers and the presence of its characteristic superficial umbrella cells (black arrowheads in the inserts of Fig. 4C, E and I) stained in dark red by MT staining (Fig. 5D, F, H and J and their respective inserts). These findings were observed both in acellular and cellseeded grafts, whether evaluated at 1 or 3 months after implantation. In 5 rabbits, which were distributed in all 4 experimental groups, a single 200e500 mm focus of metaplasia in a pluristratified epithelium, which was sometimes partially keratinized, could be detected (Fig. 5G and H, Supplementary Fig. 1A). These metaplastic foci were always observed in a ventral position in the mid-urethral graft. Additionally, some sediment-like inclusions surrounded by urothelium were present in the lamina propria (Supplementary Fig. 1B). Although rare foci of metaplasia were observed in a few rabbits, all grafts presented urothelial lining within 1 month of implantation. Occasionally, in samples harvested after 1 month, the epithelial lining was locally absent or composed of only a single cell layer. Since such findings were not observed in the samples harvested at 3 months, it could be explained by ongoing regeneration. We also found, in rare cases, foci of metaplasia towards a pluristratified keratinized epithelium (Fig. 5G and H, Supplementary Fig. 1A). This finding, both in 1 and 3 months samples, that did not present a fistula, was always located at mucosal areas that were not surrounded by SMCs. Another frequent microscopic observation was the presence of sediment-like inclusions (Supplementary Fig. 1B) compatible urine sediments, originating from the very dense and sediment-containing rabbit urine, collected in urothelial crypts. The sediment-like inclusions could also have been urethral gland secretions, however urethral glands were only observed at a more proximal location in controls (data not shown). In two cases, submucosal fresh hemorrhage could be observed in the distal urethra following catheterization for retrograde urethrogram (Supplementary Fig. 1D). 3.5. Microscopic analysis of smooth muscle and connective tissue SMC distribution was studied on MT staining of cross-sections. At 1 month after implantation, SMCs, demonstrated in dark red, were sparse in both acellular and cell-seeded grafts (Fig. 5D and H respectively); whereas, at 3 months after implantation, a larger amount of SMCs, organized in longitudinal bundles were visible in both acellular and cell-seeded grafts (Fig. 5F and J respectively). These SMC bundles were localized on the dorso-lateral side of the urethra but never on the ventral side of the urethra. SMC distribution was therefore not clearly different between acellular and cell-seeded grafts; SMCs were, however, denser at 3 months compared to 1 month after implantation. Connective tissue regeneration was investigated in stained crossand longitudinal-sections with both HE and MT. At 1 month, large hypostained areas could be distinguished in the connective tissue surrounding the lumen of both acellular (Fig. 5C and D) and cellseeded (Fig. 5G and H) grafts. They appeared in slightly lighter pink in HE staining of cross-sections (Fig. 5C and G) and in clearly lighter


L.A. Micol et al. / Biomaterials 33 (2012) 7447e7455

Fig. 4. Overviews of longitudinal-sections of rabbit penises: A&B. Non-operated penises (controls) in hematoxylin-eosin (HE) and Masson’s trichrome (MT) stainings for A and B respectively. CeF. Acellular urethral grafts (black arrowheads delimitate graft) at 1 (C&D) and 3 months (E&F) in HE (C&E) and MT (D&F) stainings. Note fistula formation (white arrowhead). GeJ. SMC-seeded urethral grafts (black arrowheads delimitate graft) at 1 (G&H) and 3 months (I&J) in HE (G&I) and MT (H&J) stainings. Note fistula formation (white arrowhead). Asterisks show lacunar vascular structures. Scale bar is 5 mm.

blue-green, if not almost white in MT-stained cross-sections (Fig. 5D and H). At 3 months, such hypostained areas were no longer distinguishable in HE stained cross-sections (Fig. 5E and I), while some smaller hypostained areas were still detectable in MT-stained crosssections (Fig. 5F and J). In all specimens, independently from the

evaluation time, lacunar vascular structures were rarely present at the graft site compared to the controls and to the remaining native urethral segments of the treated animals (Figs. 4 and 5). Small blood vessels, however, could be recognized at graft sites. Sporadic stigmata were found in the connective tissue of urethral grafts, such as

L.A. Micol et al. / Biomaterials 33 (2012) 7447e7455


Fig. 5. Overviews of cross-sections of rabbit penises: A&B. Non-operated penises (controls) in HE (A) and MT (B) stainings. CeF. Acellular urethral grafts at 1 (C&D) and 3 months (E&F) in HE (C&E) and MT (D&F) stainings. GeJ. Cell-seeded urethral grafts at 1 (G&H, gray arrows show epidermal metaplasia) and 3 months (I&J) in HE (G&I) and MT (H&J) stainings. Note in G presence of focus of pluristratified keratinized epithelium. Asterisks show lacunar vascular structures. Black scale bars are 5 mm. White scale bars are 1 mm. Inserts for HE staining show urothelium and apical umbrella cells (black arrowheads) and inserts for MT staining show submucosal SMCs (white arrowheads). Inserts scale bars are 50 mm.

sebaceous glands associated with possible hair follicles (Supplementary Fig. 1A) or epidermal cysts (Supplementary Fig. 1C). These stigmata were always present on the ventral side of the urethra. In summary, urethral grafts displayed hypostained connective tissue

areas at 1 month, which seemed to be resolving at 3 months, and fewer vascular structures compared to untreated controls. Inflammatory response, evaluated by HE staining, against the urethral grafts were not observed, neither in acellular nor in cell-


L.A. Micol et al. / Biomaterials 33 (2012) 7447e7455

seeded ones. Scarce foci of inflammatory cells could, however, be detected at graft sites in five of the eight rabbits implanted with cellular grafts (Supplementary Fig. 1E), which was not the case in acellular grafts. On one histological slide of a cell-seeded urethral graft at 3 months, a bundle-shaped inflammatory focus, measuring 1000  300 mm, was observed lateral to the graft (Supplementary Fig. 1F). Moreover, giant cells were surrounding the remaining, often encapsulated, suture material (data not shown). To conclude, despite the rare presence of localized inflammatory reaction in cellseeded grafts, no inflammatory reaction was found against the hdCGT used as the urethral graft. 4. Discussion In this in-vivo study, we have shown the good performance of high-density collagen gel tubes, hdCGTs [9] applied as urethral grafts in terms of regeneration, biocompatibility, and functionality in a rabbit model. Our results are of importance, since urethral defects greater than 1 cm in length have been shown by Dorin et al. to be critical, as they are too large to allow spontaneous regeneration by cellular ingrowth from the anastomotic edges and to prevent stenotic fibrosis [11]. Furthermore, our cell-seeded hdCGTs were implanted already one day after their formation, which outperforms other SMC-seeded implants requiring incubation for 7 d minimum. In the case of Raya-Rivera et al., a total of 4e7 wk was necessary for complete graft preparation [8]. The main limiting factor of our study remains the amount of time (4 wk) necessary to expand the SMCs prior to formation of the graft. The isolation of SMCs from micro-dissection and further digestion of the stroma of a bladder biopsy, similarly used by others authors [12,13], is the most widely used method to obtain bladder SMC populations but still requires several weeks of subsequent cell expansion. Further improvements on cell isolation and expansion are necessary to allow for a quicker availability of the cell-seeded graft to treat human patients. We tested both acellular and cell-seeded hdCGTs as urethral grafts. The procedure of water extraction of the collagen gel tube resulted in improved mechanical stability of the construct. To prevent any host versus graft reaction against the cell-seeded hdCGTs, we exclusively used autologous bladder SMCs, previously isolated from a bladder biopsy and expanded in vitro. Although only a 1 cm defect had been generated, we implanted a 2-cm long urethral graft. The application of a longer graft was chosen to limit tensile stress, generated by the retracting native urethra, on the graft itself and especially on its anastomoses. As the collagen graft could not be sutured, fibrin glue was used to perform the anastomoses. In addition the fibrin glue also acted as a sealant preventing urine leakage and subsequent inflammation of the urethral graft. Biocompatibility of any graft or implant inserted into a living body is crucial, and our hdCGTs showed excellent biocompatibility. They are convenient substrates for cellular attachment and allow for re-epithelialization of the luminal side by urothelial cells. Furthermore, they do not trigger any generalized inflammatory reaction. The sparse foci of inflammatory cells were observed adjacent to suture materials or, in one case, to a possible rabbit hair remnant visible at 3 months (Supplementary Fig. 1E and F). Compared with the results of Nuininga et al. showing clear inflammatory reaction towards their cross-linked collagen matrices [6], our hdCGTs did not trigger any inflammatory response. Our results showed that, compared to control urethrae, the grafts maintained similar lumina and showed the typical starshape. They all displayed spontaneous urothelial lining. The mucosal folds bordering the star-shaped lumen were, however, less marked in grafts than in controls, which could be linked to the lower amount of SMCs in the surrounding tissue, since the presence

of SMCs could also play a role for a proper urothelial differentiation, induced by cellular cross-talk. Epithelial lining was present continuously along the graft including typical apical umbrella cells of the stratified urothelium, despite the exclusion of urothelial cells from our graft preparations. We can hence conclude that spontaneous urothelial regeneration occurred along the graft lumen. Such urothelial regeneration could have arisen by two different mechanisms: either by colonization from the edges of native urothelium adjacent to the graft, or by seeding, during voiding, of urine-derived progenitor cells which have been demonstrated to give rise to urothelial cells [7]. Other groups have obtained various results regarding urothelial lining of their urothelial grafts. For example, Nuininga et al. obtained an epithelial lining of 3e5 cell layers compatible with urothelium on their cross-linked collagen matrices implanted in rabbits [6], while Raya-Rivera et al. seem to have obtained variable results with epithelia that are more often appearing as a pluristratified epithelium with their polyglycolic acid mesh scaffolds implanted in human patients [8]. The presence of SMCs had a significant positive effect on the urethra caliber at 3 months (Fig. 3B), which tended to be the opposite effect of that at 1 month. The contraction of the seeded SMCs could explain the smaller caliber of cell-seeded grafts at 1 month with subsequent resolution over time. The variances associated with urethral caliber measurements were relatively high for acellular grafts evaluated at 3 months and cellular grafts evaluated at 1 month. This was due to the non-exclusion of samples with complete stenosis from these two groups. On the other hand, no obvious SMC density difference between acellular and cell-seeded grafts could be seen in histology. Also, in both acellular and cellular grafts vascular lacunar structures were rarely present. This can be explained by the complete removal of the venous plexus of the urethra at the site of the graft during urethral defect formation. Since, SMC were present only dorso-laterally, this venous plexus may be required to promote SMCs survival by preventing ischemia. The dorso-lateral localization of SMCs in regenerated urethrae observed in this study also seems to be consistent with the results obtained by Nuininga et al. with cross-linked collagen matrices [6]. Furthermore, SMCs were essentially present on the dorsal side of the urethra, close to the highly vascularized corpus cavernosum. Therefore, SMCs present in cell-seeded grafts probably did not survive long after implantation. Since vascular supply seems to be of importance for the functionality and survival of SMCs, it could be advantageous to design constructs that also promote endothelial cell growth. Histological appearance of a discontinuous (Fig. 4G and H) or of a lateralized lumen (Fig. 5G and H) was a clear sign of stenosis, already shown by the respective urethrograms of two rabbits. These two stenoses, one associated with an acellular graft and the other one with a cell-seeded graft, were distal and provoked proximal fistula formation, preventing urinary retention. We have also noticed fistula formation for both cellular and acellular grafts without associated stenosis (Fig. 3A). In cellular grafts, only one urethro-cutaneous fistula was found at the evaluation 3 months postoperatively. Fistulae, previously conspicuous by retrograde urethrography, were also evident at microscopic examination. Communication between the urethral lumen and the ventral penile skin (Fig. 4E and F) or a slanted lumen was compatible with urethro-cutaneous fistulae. Several factors could be responsible for the formation of fistulae such as intraluminal overpressure, due to stenosis or anastomotic leakage. Extrinsic pressure of edematous surrounding tissues after wound closure may provoke fistula formation. Mechanical resistance of our hdCGTs has been demonstrated for axial and radial extension; however, resistance to shear stress has never been investigated. If proven needed, hybrid collagen gel tube combined with a fast degradable polymer could

L.A. Micol et al. / Biomaterials 33 (2012) 7447e7455

be produced. This will need to be further investigated, especially regarding the biocompatibility. The fibrin glue, used for anastomosis, can be expected to be resistant to urine as it has been reported in hypospadias repair in human patients [14,15] and cannot be responsible for fistula formation. However, the possibility of intraluminal presence of fibrin glue or further action of Factor XIII, a natural crosslinker present in fibrin glue, after graft implantation cannot be totally excluded since the catheter was directly removed after wound closure. This situation could promote stenosis at the anastomotic site. Nevertheless, all the factors listed could possibly be controlled in future studies, if prolonged urethral catheterization is applied. Use of urethral catheters or tutors would be in accordance with common surgical practice in hypospadias surgery. If the urethral catheter could prevent luminal narrowing and stenosis of both acellular and cell-seeded hdCGTs, acellular hdCGTs would then become of clear superiority as urethral grafts since they do not require cell isolation and expansion steps, but had similar histological outcome at 3 months compared with cellseeded grafts. 5. Conclusion Functional urethral reconstruction was performed in 16 rabbits, using either acellular or autologous-cell-seeded high-density collagen gel tubes (hdCGTs). Our results showed excellent graft biocompatibility as well as complete urothelialization of the straight and star-shaped lumen. Complications such as fistula and related-stenosis were evident in approximately half of the animals; however, these could be potentially overcome by simply maintaining a catheter in the urethra after surgery. We have showed that our hdCGTs could regenerate a rabbit urethral segment and that our technology can potentially become an alternative treatment for urethral pathologies such as hypospadias. Acknowledgments The authors gratefully acknowledge Dr. Jessica Sordet-Dessimoz (Histology Core Facility, Station 15, EPFL, CH-1015 Lausanne), Jose Artacho (BiOP platform, Station 15, EPFL, CH-1015 Lausanne), Thijs Eijkenboom and Dorien Tiemessen (Nijmegen Center for Molecular Life Sciences, UMC St Radboud, NL- 6500 HB Nijmegen), as well as Alex Hanssen, Maikel School, and Wilma Janssen-Kessels (Centraal Dierenlaboratorium, UMC St Radboud, NL- 6500 HB Nijmegen) for their technical contribution to this study and, eventually, Dr. Fabio


Aloisio (Institut für Tierpathologie, Vetsuisse-Fakultät, Pf 8466, CH3001 Bern) for his expertise and pieces of advice and Dr. Kristen Lorentz (LMRP, Station 15, EPFL, CH-1015 Lausanne) for reviewing the manuscript. This research was funded by EU-FP6 project EuroSTEC (Soft Tissue Engineering for Congenital birth defects in children; reference: LSHB-CT-2006-037409) as well as the Swiss National Science Foundation (reference: 323530-123716). Appendix A. Supplementary data Supplementary data related to this article can be found online at References [1] Dolk H, Loane M, Garne E. The prevalence of congenital anomalies in Europe. In: Posada de la Paz M, Groft SC, editors. Rare diseases epidemiology. Netherlands: Springer; 2010. p. 349e64. [2] Anger JT, Buckley JC, Santucci RA, Elliott SP, Saigal CS. Trends in stricture management among male medicare beneficiaries: underuse of urethroplasty? Urology 2011;77:481e5. [3] Nuininga JE, De Gier RPE, Verschuren R, Feitz WFJ. Long-term outcome of different types of 1-stage hypospadias repair. J Urol 2005;174:1544e8. [4] Atala A. Engineering organs. Curr Opin Biotechnol 2009;20:575e92. [5] Fu Q, Deng CL, Liu W, Cao YL. Urethral replacement using epidermal cellseeded tubular acellular bladder collagen matrix. BJU Int 2007;99:1162e5. [6] Nuininga JE, Koens MJW, Tiemessen DM, Oosterwijk E, Daamen WF, Geutjes PJ, et al. Urethral reconstruction of critical defects in rabbits using molecularly defined tubular type i collagen biomatrices: key issues in growth factor addition. Tissue Eng Part A 2010;16:3319e28. [7] Fossum M, Svensson J, Kratz G, Nordenskjold A. Autologous in vitro cultured urothelium in hypospadias repair. J Pediatr Urol 2007;3:10e8. [8] Raya-Rivera A, Esquiliano DR, Yoo JJ, Lopez-Bayghen E, Soker S, Atala A. Tissue-engineered autologous urethras for patients who need reconstruction: an observational study. Lancet 2011;377:1175e82. [9] Micol LA, Ananta M, Engelhardt EM, Mudera VC, Brown RA, Hubbell JA, et al. High-density collagen gel tubes as a matrix for primary human bladder smooth muscle cells. Biomaterials 2011;32:1543e8. [10] Bancroft JD, Gamble M. Theory and practice of histological techniques. 6th ed. Philadelphia: Elsevier Health Sciences; 2007. [11] Dorin RP, Pohl HG, De Filippo RE, Yoo JJ, Atala A. Tubularized urethral replacement with unseeded matrices: what is the maximum distance for normal tissue regeneration? World J Urol 2008;26:323e6. [12] Baker SC, Southgate J. Towards control of smooth muscle cell differentiation in synthetic 3D scaffolds. Biomaterials 2008;29:3357e66. [13] Kimuli M, Eardley I, Southgate J. In vitro assessment of decellularized porcine dermis as a matrix for urinary tract reconstruction. BJU Int 2004;94:859e66. [14] Gopal SC, Gangopadhyay AN, Mohan TV, Upadhyaya VD, Pandey A, Upadhyaya A, et al. Use of fibrin glue in preventing urethrocutaneous fistula after hypospadias repair. J Pediatr Surg 2008;43:1869e72. [15] Barbagli G, De Stefani S, Sighinolfi MC, Anino F, Micali S, Bianchi G. Bulbar urethroplasty with dorsal only buccal mucosal graft and fibrin glue. Eur Urol 2006;50:8.