Summary
Repair of long-gap esophageal atresia is associated with a high degree of complications. Tissue engineering on a scaffold of a bioresorbable material could be a solution. The aim of the present study was to investigate the in vivo tissue engineering of smooth muscle cells and epithelium on a poly-ε-caprolactone mesh in rabbit esophagus.
Twenty female rabbits had a window of 0.6 × 1 cm cut in the abdominal part of the esophagus. The defect was covered with a poly-ε-caprolactone mesh. The rabbits were killed on postoperative day 28–30, and mesh with surrounding esophagus was removed for histological examination.
Fifteen rabbits survived the trial period. Six had no complications and had the mesh in situ. They all had ingrowth of epithelial and smooth muscle cells and an almost completely degraded mesh. Nine rabbits developed pseudo-diverticula.
It proved possible to engineer both epithelial and smooth muscle cells on the poly-ε-caprolactone mesh in spite of a fast mesh degradation. The latter may be the explanation to the development of pseudo-diverticula; this is a problem that needs attention in future experimental trials.
Introduction
The surgical treatment of esophageal atresia (EA) aims for a primary anastomosis within the first days of life.1 In approximately 10% of cases, the distance between the esophageal segments is too long for primary anastomosis. There are several techniques to treat long-gap atresia, e.g. delayed primary anastomosis, growth by traction (Foker's method), gastric, jejunal or colonic interposition.1 However, all these methods are associated with a relatively high risk of early and late complications. Alternative treatment strategies are warranted, and tissue engineering seems to be an attractive option.
The scaffold for tissue engineering in the esophagus must meet several demands. First, it must be able to avoid leakage. Second, it must allow tissue ingrowth, gradually replacing the scaffold with native esophageal tissue including epithelium, muscle, nerves and vascular cells. Poly-ε-caprolactone (PCL) is an interesting material because it prevents major leakage and has a slow degradation time of approximately 24 months when only degraded by hydrolysis.2 However, studies have shown that PCL has a shortened degradation time in different biotic environments. Body temperature, microorganisms, acid and digestive enzymes enhances degradation.3,4 Other factors that influence the degradation of implanted material are the mechanical forces added to the material and the foreign body response.5
In vitro studies have shown that PCL allows growth of rat esophageal epithelial cells,6 and in vivo studies have shown that PCL has a very good biocompatibility and allows bone regeneration in rabbits.7 However, there are no in vivo studies regarding esophageal tissue engineering on PCL meshes to our knowledge.
The aim of the present study was to investigate the in vivo tissue engineering on a PCL mesh in rabbit esophagus. The hypothesis was that it would be possible within 28 days to engineer both smooth muscle tissue and esophageal epithelium.
Materials and Methods
Animals
Twenty two-to-three month old female New Zealand White rabbits, with a weight varying from 2.1–2.8 kg, were included in the study. The rabbits were kept individually at standard laboratory conditions, and had free access to food, hay and water at all times. All food was autoclaved. Postoperatively, the rabbits were fed with both normal and soaked food to ensure appropriate food intake. The rabbits were operated on day 1 and planned killed on day 28–30, after which the mesh and surrounding esophagus were removed and examined histologically.
Anesthesia
Pre-anesthesia comprised of 0.15 mL/kg mixture of fentanylcitrat 0.315 mg/mL and fluanisone 10.0 mg/mL and 1 mg/kg midazolam administered subcutaneously. After sedation, the rabbits were intubated with an uncuffed tube size 2.5 to 3.5.
The rabbits were anesthetized with 2–4% sevoflurane via a ventilator at a respiratory frequency of 18 per minute, and with a tidal volume of 30–35 mL mix of oxygen and atmospheric air. Perioperative analgesia consisted of 5 mL/hour of fentanyl administered intravenously. During surgery, the rabbits were placed on a heating pad set to 32°C to prevent hypothermia.
Postoperatively, a single dose of 0.4 mL/kg mixture of sulfadoxine 200 mg/mL and trimethoprim 40 mg/mL was given subcutaneously as prophylaxis against infection. To prevent postoperative dehydration, 5 mL of isotonic saline and 5 mL isotonic glucose was given subcutaneously. Postoperative analgesia consisted of 0.125 mg/kg buprenorfin given subcutaneously.
Meshes
The sterile porous meshes were made from 80 PCL. PCL is a bioresorbable polyester approved by the US Food and Drug Administration (FDA) and the (EU/FDA). The meshes were made by a noncontrolled precipitation reaction leaving a non-organized surface. The meshes were without shape but flexible in all directions making it possible to retain the tubular shape of the esophagus once implanted.
The thickness of the meshes was measured by scanning electron microscopy to approximately 270 μm (Fig. 1).
Scanning electron microscopy (SEM) of luminal face of the poly-ε-caprolactone (PCL) mesh.
Surgical procedures
An upper midline laparotomy gave access to the distal part of the esophagus. A window of approximately half the circumference (0.6 × 1 cm) was cut in the esophageal wall (Fig. 2). The defect was covered by the PCL mesh (Fig. 3), which was sutured to the edges of the esophageal defect with single resorbable monofilament sutures. The mesh and the suture line were covered by a surgical patch (TachoSil®, Takeda Pharmaceuticals International GmbH, Zurich, Switzerland) to prevent initial leakage from the suture line. TachoSil is a surgical patch used for hemostasis and tissue sealing. It is coated with the human coagulation factors thrombin and fibrinogen and when applied creates an air and liquid-tight seal. Muscle and fascial layers in the abdominal wall were closed with single resorbable sutures, and the skin was closed intracutaneously with a continuous resorbable monofilament suture.
The window cut in the esophageal wall.
The operation window covered by the poly-ε-caprolactone (PCL) mesh.
On postoperative day 28–30, the rabbits were pre-anesthetized with 0.3 mL/kg mixture of fentanylcitrat 0.315 mg/mL and fluanisone 10.0 mg/mL and 2 mg/kg midazolam and killed with a lethal dose of pentobarbital, and tissue samples including mesh and surrounding esophagus were removed for analysis.
Histopathological analysis
All tissue samples were fixed in a 4% formaldehyde solution, embedded in paraffin and sliced in 3 μm thick slices, which were stained with hematoxylin and eosin (HE), alpha smooth muscle actin (α-SMA) and desmin. The slices were all made from the most central part of the original defect in the esophagus. HE was used as an overview staining to identify epithelium coverage, mesh degradation and to identify immune response represented by macrophages and foreign body giant cells (FBGCs). α-SMA and desmin were used to identify growth of smooth muscle cells and smooth or skeletal muscle cells, respectively. All tissue samples were examined under a light microscope at ×10 to ×400. Quantification of tissue growth was estimated by visual evaluation by two independent pathologists. Identification of enteric neurons was inconsistent and insecure, and therefore not included in the analysis.
Results
Study flow and the main findings from the macroscopic and the microscopic examination are illustrated in Figure 4.
Flow chart of the macroscopic and microscopic findings.
Macroscopic findings
Fifteen rabbits completed the study. In all animals, a minor weight loss during the first week postoperatively was observed, but all gained sufficient weight during the rest of the trial period. Six rabbits had the mesh in situ and no signs of anastomotic leakage or mesh defects. The esophagus showed signs of slight stenosis in two of these rabbits. Nine of the 15 rabbits had developed a pseudo-diverticulum at the mesh site. The pseudo-diverticula were all more or less dome shaped with sizes varying from 1.5 cm*1 cm*1 cm to 3.5 cm*3.5 cm*4 cm. Macroscopically, the pseudo-diverticula differed from the original esophageal tissue being thicker and more rigid and looking like granulation tissue.
Inside the pseudo-diverticula food residuals and caseous necrosis was found. The meshes were more or less intact either still partly attached to the edge of the operation window or inside the pseudo-diverticula.
Five of the 20 rabbits died during the trial period – four died within the first week and the last within 11 days postoperatively. Of these, two had lost almost 20% of their preoperative weight; one of these had the mesh in situ and no pathological explanation of the failure to thrive, and the other had an anastomotic leakage. A third rabbit suffered from respiratory difficulties due to pneumonia and had developed a pseudo-diverticulum at the site of the mesh. A fourth had an abscess in close relation to the liver but had the mesh in situ and no signs of leakage. In the fifth rabbit, the mesh was found intact but dislocated from the esophagus intra-abdominally.
Microscopic findings
In the six rabbits without any complications, we found that the mesh was almost completely degraded, leaving only residuals of the mesh identified as areas of dense dark purple stained clutched fiber-like material (Fig. 5). A continuous epithelial layer covering the operation window with inflammatory reaction and deposition of granulation tissue were found in all these cases. Staining with α-SMA showed a layer of smooth muscle cells just below the epithelium (Fig. 6). Staining with desmin showed, as expected, a layer of smooth muscle just below the epithelium as found with α-SMA staining, followed by a layer of striated muscle (Fig. 7).
Black arrow: Mesh residuals. Yellow arrow: Foreign body giant cell (FBGC) containing mesh material. Hematoxylin and Eosin (HE) × 100.
Red arrow: Newly formed blood vessels. Green arrow: Newly formed disorganized smooth muscle tissue covering the operation window. Alfa smooth muscle actin (α-SMA) × 10.
Same sample as Figure 6. Between purple arrows: Newly formed striated muscle tissue. Green arrow: Thin layer of newly formed smooth muscle tissue covering the operation window. Desmin × 10.
In the three animals, a discontinuous layer of disorganized and spread smooth muscle cells equivalent to the operation window was found, whereas it was continuous in other three animals. All samples contained macrophages, and in four samples FBGCs were found (Fig. 5). In two of these, the FBGCs contained mesh residuals. Generally, the newly formed epithelial and muscular tissue was disorganized compared to native esophagus.
Practically no invasion of muscle cells, epithelial cells or macrophages were found in the mesh from the nine rabbits that had developed pseudo-diverticula.
In three of the five rabbits that were killed before end of the trial period (postoperative day 4–11), the mesh was found in situ. Histological examination showed varying degrees of cell invasion into the mesh. Macrophages but no FBGCs were found in the tissue surrounding the mesh.
Discussion
After 28–30 days, we found that the mesh was almost completely degraded and replaced by a layered continuum of epithelium and smooth muscle cells the latter disorganized in varying degrees.
Different materials have been implanted in the esophagus in a variety of animal models. Freud et al.8 found that nonabsorbable materials such as lyophilized dura mater (Lyodura), polytetraflouroethylene (PTFE) and polyethylene terephthalate (Dacron) only allows ingrowth of mucosal cells and scar tissue and no ingrowth of muscle cells in dogs 12 months after implantation. The use of these materials was associated with a high risk of complications including stricture.8 In two of the rabbits in our study, we found signs of slight stenosis with no clinical symptoms. Lynen Jansen et al.9 used a rabbit model and found that implantation of a mesh made of the nonabsorbable material polyvinylidene fluoride in the abdominal part of esophagus allowed complete mucosal and initial regeneration of the muscle layer within 3 months. Other studies have shown it is possible to grow esophageal muscle cells and epithelium on uncoated and coated resorbable Vicryl® (Ethicon Endo-Surgery INC., Blue Ash, Cincinnati, OH, USA) meshes within 3 to 6 months, respectively.9,10 Shinhar et al.10 used an absorbable collagen coated Vicryl mesh in the cervical part of esophagus in 24 dogs. They found that the mesh was resorbed within 4 weeks and at that time was substituted with scar tissue and epithelium. They found signs of muscle regeneration in only few instances after 3 months. After 6 months, they described complete regeneration of the esophageal wall, and they reported a low mortality rate. Lynen Jansen et al.9 used an uncoated Vicryl mesh in the abdominal part of esophagus in rabbits and found that the material allowed ingrowth of both epithelium and muscle cells after 3 months. A high rate of mortality and complications compared to our study was reported. Biological absorbable scaffolds have also been used as scaffolds for tissue engineering in the esophagus. Lopes et al.11 implanted a porcine small intestinal submucosa patch and tube in esophagus in a rat model. Both epithelial and muscle tissue could be seen in the patch group after 5 months. None of the rats in the group with tube implant survived the trial period because of complications. As described earlier, we found that the PCL mesh used in our study allowed massive ingrowth of both epithelium and smooth muscle cells within a month. We also found that the PCL mesh was almost completely degraded at this time.
Nine rabbits developed a pseudo-diverticulum. A possible explanation to this may be that the degradation of the mesh has been too fast for proper ingrowth of new tissue leaving room for minor leakage. Lynen Jansen et al.9 also found that fast degradation could be an explanation of the high complication rate in the Vicryl group. Since PCL in a sterile environment is approximately 24 months to degrade by hydrolysis,2 the finding of almost completely degraded mesh was rather surprising. The environment in the esophagus is far from sterile containing saliva with enzymes and bacteria passing through with every swallow. Identification of the bacteria present could have shown if any of these had known degradative abilities like, e.g. pseudomonas lipase.4 The motility of the esophageal wall results in stretch and tension of the mesh, which increase the degradation rate5 and due to the location of the defect close to the gastroesophageal junction acid reflux could act as a catalyst in the hydrolysis of the ester bonds in PCL. The finding of FBGCs showed that the inflammatory and the foreign body response also played a role in the fast degradation of the mesh. Other studies have shown that PCL implanted in sterile environments (e.g. bone) in vivo – where a foreign body response would be present – does not degrade as fast as found in our study.7 All these parameters point in the direction of a form of synergy of the different ways of degradation leaving the PCL mesh degraded much faster in the esophagus than seen in, e.g. bone. Another parameter that could influence the degradation is the use of TachoSil® (Takeda Pharmaceuticals International GmbH, Zurich, Switzerland). In rats, the use of TachoSil on intestinal anastomoses resulted in an increased rate of complications such as increased inflammation, ileus and perforation.12
The creation of a circumferential defect in the thoracic part of the esophagus would have been more representative to the clinical situation. The primary endpoint was to investigate the characteristics of tissue growth in the PCL mesh with a new production method, and a thoracotomy was considered to be too stressful for the animals with a suspected high mortality rate.
The only problem with the PCL mesh used in our study seems to be the fast degradation and therefore risk of leakage and development of pseudo-diverticula. Collagen coating might be a solution10 which should be tested in future studies in addition to studies where the scaffold is cultured with smooth muscle, epithelial or stem cells in vitro. In addition to the different coatings and cultures, the PCL material should be tested in a tubular graft to better imitate the esophageal shape, and therefore the clinical situation.
Acknowledgments
Thanks to the Faculty of Health Science, University of Southern Denmark; MD Katrine Urth, Ph.D. Jeeva Sellathurai, Profesor Henrik Daa Schroeder, Dept. of Pathology, Odense University Hospital; Veterinarian Henrik Saxtorph and Head of Biomedical Laboratory Peter Bollen, The Biomedical Laboratory, University of Southern Denmark. Odense University Hospital Research Foundation financed the study.
References
Author notes
P. Diemer and S. Markoew contributed equally both in carrying out the project and in the writing process.
