Pre-clinical animal study design
This study enrolled 15 Yucatan minipigs ranging from 40 to 50 days of age that were assigned to three cohorts with different survival time points (Cohort 1; 365 ± 15 days, Cohort 2; 90 ± 3 days and Cohort 3; 30 ± 3 days post-implantation. American Preclinical Sciences, Minneapolis, MN IACUC protocols OIP007-IS75/OIP010-IS75). The number of animals, time points, and groups are described in Table 1. All animals enrolled in this study passed a physical exam including routine bloodwork with values that were within normal range. A surgical control animal was included for each survival time point. All surgical control animals underwent a 2 cm esophageal resection followed by a single anastomosis to reconnect the proximal and distal ends of the esophagus. All test animals underwent a 5 cm segmental esophagectomy followed by the implantation of the CEI to bridge the gap between the proximal and distal esophagus using end-to-end anastomoses.
MSC isolation and expansion
Adipose tissue (6–12 grams) was harvested from the midline pelvic region from all animals at ~28 days prior to implantation. Adipose tissue biopsies were shipped overnight at 2–8 °C in αMEM (Gibco, Grand Island, NY) + 0.1% gentamicin at 2–8 °C with temperature monitoring to the laboratory. Ad-MSCs were isolated from adipose as described32 Cells were plated at a density of 40 mg (of pre-digested tissue)/cm2 in either T75 or T150 tissue culture treated flasks in StemXVivo culture media (R&D Systems, Minneapolis, MN). Medium was refreshed every 48 h until they reached 70–80% confluence. Cells were cultured in 2-D format and then seeded at a density of 1143 cells/cm2 onto each scaffold.
MSCs were characterized via flow cytometry as previously described at the beginning of the second passage32. Briefly, cells were harvested and washed two times with wash buffer (DPBS (Ca–/Mg–) (Gibco, Grand Island, NY), 1% bovine serum albumin (Gibco, Grand Island, NY) and aliquoted into tubes containing 1 × 105 cells in 100 µL wash buffer. Primary antibodies were added to the aliquots according to Supplementary Table 1 and incubated for 30 min in the dark on ice. Tubes were washed twice with wash buffer and centrifuged at 1000 rpm for 2 min at 4 °C. Flow was performed on an Attune NxT Flow Cytometer (Thermo Fisher, Waltham, MA) and analyzed using Attune Nxt software. Percent positive events were determined against samples stained with matching isotype controls.
Cellspan esophageal implant preparation, culture, and transport
Pediatric sized scaffolds were electrospun, loaded into bioreactors and sterilized prior to cell seeding. Images of electrospun scaffolds of various sizes, scanning electron microscopy (SEM) images of the native scaffold fiber network as well as cell attachment are presented in Supplementary Fig. 4. Prior to seeding, cells were expanded out to at least the second passage. Seven days prior to surgery, cells were seeded onto a 12 mm ID × 100 mm L Cellframe™ scaffold at a density of ~4000 cells/mm2 for each implant (Supplementary Fig. 5a). In addition, for every lot of CEIs scheduled for implant, an additional sentinel CEI was prepared for Quality Control (QC) release assays. CEIs were not produced for surgical control animals. CEIs were incubated (37 °C, 5% CO2) in a custom rotating bioreactor for 6 days with media exchanges every 48 h until harvest. Upon harvest, QC assays were performed on the sentinel CEI as well as sterility assays on the implant CEI condition media. The implant CEIs were placed in shipping tubes with pre-gassed alpha MEM + HEPES (Gibco, Grand Island, NY) + 1% gentamicin (Gibco, Grand Island, NY) and cooled to 4 °C until they were transferred to pre-cooled shipping boxes. CEIs were shipped by courier to surgical sites with temperature monitors (2–8 °C).
Viability and metabolic activity of cellspan esophageal implant
Viability of cells on QC sentinel CEIs was performed as previously described32. Glucose and lactate concentrations were measured in fresh medium and medium collected on days 2, 4, and 6 from all CEIs using a Nova Stat Profile Prime analyzer (Nova Biomedical; Waltham, MA) (Supplementary Fig. 5b).
Cell Dose and cell penetration on the cellspan esophageal implant scaffold
Cell viability and cell penetration was determined on dedicated QC sentinel CEIs as previously described32, (Supplementary Fig. 5d). To determine cell numbers on scaffold, five punches were taken from along the length of QC CEIs and frozen at −80 °C. Frozen punch samples were thawed and then lysed using an SDS based buffer + proteinase K (Blood and Tissue DNA Extraction Kit, Qiagen, Germantown, MD) and subsequently homogenized using a bead mill (Thermo Fisher Scientific, Waltham, MA). DNA was extracted from lysates according to the manufacturer’s instructions (Qiagen, Germantown, MD). DNA was quantified via QuBit, using DNA BR Qubit assay kit (Thermo Fisher Scientific, Waltham, MA). Cell number was calculated by dividing DNA content by 5.1 pg DNA/diploid porcine cell and dividing by the area of a punch. Cell number on the QC was extrapolated from DNA content and calculated as cells per mm2. (Supplementary Fig. 5e).
Cellspan esophageal implant surgery and initial stent placement
All animal procedures were performed at American Preclinical Services (APS, Minneapolis, MN. Approved IACUC protocols OIP007-IS75 and OIP010-IS75). APS has the following certifications and accreditations: USDA registration number 41-R0074; AAALAC accreditation number 001236; and PHS assurance number A4586-01.
Piglets were placed under general anesthesia, prepped, and draped according to standard operating procedures. The piglet was placed in a supine position and a 5 cm transverse incision was made in a left subcostal position and the abdominal cavity was entered. The stomach was identified, and two purse-string sutures were placed. An incision was made lateral on the left to the incision site and a 14 Fr Mic tube (Kimberly Clark, Neenah, WI) was placed through the incision and tunneled to the stomach. The tube was filled with water, purse-strings secured, and the stomach was tacked to the underside of the abdominal wall. The fascia and skin were closed with absorbable sutures. Next, an oblique incision was made in the right neck and the jugular vein was isolated. The venous access port (VAP) (Kimberly Clark, Neenah, WI) was placed laterally in a subcutaneous pocket. The tubing was then tunneled to the area where the vein was accessed. The tubing was cut to size, a venotomy was made and the tubing was fed into the vein into a central position. A vicryl tie (Ethicon, Somerville, NJ) around the vein and catheter was secured to keep it in place and the port was secured subcutaneously. The Huber needle was left in place and secured, and the incision was closed with absorbable suture. Finally, a standard right posterolateral thoracotomy was performed between the fourth and fifth intercostal space. The thoracic cavity was entered, and the lung retracted medially. The esophagus was isolated, preserving the vagus nerve, and a 5 cm segment of esophagus was removed in animals receiving CEIs. A 4-0 PDS stay suture is placed proximally and distally in the esophagus above and below the area of transection. The esophagus was then transected. The seeded CEI was trimmed on each side to a 6 cm length and sewn to the proximal and distal esophageal ends. The back wall of the proximal esophagus is sewn with 4-0 PDS in a running fashion. An orogastric tube is passed into the lumen and the anastomosis is completed proximally and distally over the tube with a 4-0 PDS suture. For the surgical control animals, a 2 cm segment was removed, followed by an end-to-end anastomosis using 4-0 PDS in a running fashion. A 12 mm × 100 mm Alimaxx esophageal stent (Merit Medical, South Jordan, UT) was placed using direct vision and fluoroscopy. The ribs were re-approximated, and the muscles were closed in 2 layers with absorbable suture (Ethicon, Somerville, NJ). A chest tube was placed in the thoracic space during closure and when the muscles were re-approximated, air was evacuated from the chest by giving positive pressure breaths (Valsalva maneuver). Skin was re-approximated with absorbable suture. A summary of the procedure is found in Supplementary Fig. 2.
Animal in-life management
Animals were housed together in raised pens leading up to the initial procedure. The animals were then housed individually in raised pens with normal food and water following the adipose biopsy and other subsequent procedures. Animals were fasted for 12–24 h prior to all procedures. Animal health, including evaluation of incision site and clinical observations, body weights/condition, endoscopy/fluoroscopy, barium swallows and clinical pathology, was monitored at regular intervals. Starting day one post-implant, animals were fed a liquid nutrient diet through the implanted G-tube and allowed access to water only. Animals were slowly transitioned to a soft oral diet starting at ~21 days after implantation. Cohort 1 animals (365 ± 15 days) were transitioned to solid food starting at ~Day 35 after CEI implant. Cohort 2 (90 ± 3 days) and Cohort 3 (30 ± 3 days) maintained liquid oral diet through termination (~Day 90 and ~Day 30, respectively).
Endoscopic assessments and stent exchanges
Animals were prepped and sedated per facility standard operating procedures (SOPs). Fully covered nitinol stents manufactured by Merit Medical (South Jordan, UT) and Boston Scientific (Marlborough, MA) were exchanged as necessary up to a 23 mm diameter stent. The initial stents were deployed immediately after surgery and subsequently removed at day 21 post-implantation (Fig. 1). To accommodate for growth of the animal, the stents were exchanged every 3–4 weeks with increasing diameter sizes unless symptoms of migration or obstructions presented sooner. To visualize the esophageal stent, an endoscope with live video feed was used in conjunction with fluoroscopy. If the esophageal stent was removed, the esophagus was visualized after removal to determine if the esophagus sustained any injury during the process as well as to monitor regeneration (Fig. 2). A new stent was then inserted and deployed under fluoroscopic and endoscopic guidance. For Cohort 1, stents were permanently discontinued once the mucosal layer was fully formed and the maximum stent diameter (23 mm) was reached (between 3 and 6 months post-implant). For Cohort 2 and Cohort 3, the stents were utilized until study termination. A summary of endoscopic interventions for each cohort can be found in Table 2.
Evaluation of tissue regeneration by CT scan
Post-operative (Post-Op) CT imaging was performed on animals in all 3 cohorts (30, 90, and 365 days) at various time points, as listed in Supplementary Table 2. The esophageal stents were in place during the scanning process for the short-term assessments (up to day 90), however, the stents were removed (as per protocol) 3–6 months post implantation and therefore the CT imaging at 365 days was performed in the absence of esophageal stents. The scan recordings, DICOM image files, were analyzed by an independent medical imaging core laboratory (Medical Metrics, Inc., Houston, TX) for retrospective analysis. The CEI implant zone was determined for the first available post-op time point by identifying the anastomosis sites as surrogates for the CEI boundaries, then deriving the CEI Central Slice as the superior-inferior midpoint between the anastomosis sites. Quantitative measurements were obtained by an analyst. The analyst and radiologist were not blinded to the treatment group or time points for each subject. Quantitative measurements and qualitative assessments required identification of the CEI Implant Zone. Because the CEI was radiolucent, alternative anatomical features were used including the anastomosis sites and the carina, which were identified by the analyst (Fig. 3).
Barium swallow esophagram
Animals were brought to a procedure room, placed in a sling or on a table and fed an oral mixture of food with barium while recording fluoroscopic video at various timepoints during the study. Static images were obtained to demonstrate patency of the construct and ability of barium to traverse from the mouth to the stomach (Fig. 4).
Histopathology and immunohistochemistry
At experiment termination, the esophagus and designated representative tissues were collected and processed for histomorphology. Samples were processed by American Preclinical Services, LLC (APS). All slides were stained with hematoxylin and eosin (H&E) and with Masson’s Trichrome (MT). Slides were sent to StageBio, Inc., (Mt Jackson, VA) for imaging, histopathology review, immunohistochemical analysis, scoring and reporting. Immunohistochemical analysis was performed after deparaffinization with xylene. Antigen retrieval was performed using 0.015% citraconic anhydride at 95 °C for 15 min followed by blocking of endogenous peroxidase activity. Primary antibody was incubated at room temperature for 1 h to identify marker CK13, smooth muscle marker SM22, and axonal growth marker GAP43. Slides were then incubated with secondary antibody for 1 h at room temperature, developed using DAB and counterstained with hematoxylin. Antibodies used for immunohistochemical analysis are included in Supplementary Table 1.
Further information on research design is available in the Nature Research Reporting Summary linked to this article.