Esophageal grafts were harvested from human deceased donors following a brain-death or Maastricht category III circulatory arrest. Non-opposition, informed consent was obtained from the donors’ families for research purposes. The graft retrieval authorization was granted by the “Agence de la Biomédecine”, grant number PFS18-018 in Saint-Louis Hospital, Université Paris Cité, France. Under sterile conditions, 15 cm of thoracic esophagi were removed either by thoracotomy or by transhiatal approach in case of only intra-abdominal organ retrieval. The esophageal segment was harvested between two clamps so that there was a minimum digestive contamination (Fig. 1a,b). Esophageal retrieval was the last performed, in order to avoid contamination of the other grafts.
Decellularization of human esophagi
The protocol was adapted from the one previously developed by our team in a porcine model15. It consisted of five major steps: decontamination, decellularization with a detergent treatment to remove cells, rinsing and removal of the residual detergent, incubation with DNase followed by a rinsing step before preservation. All these steps were carried out under sterile condition in a laminar flow hood.
Immediately after retrieval in the operation room, esophagi were cleaned with Povidone-Iodine (Betadine® dermique 10%, Meda Pharma, Paris, France) and rinsed with an antibiotic and antimycotic (ATB/ATM) solution composed of physiologic serum 500mL, Gentamicin 160 mg (Panpharma, Boulogne Billancourt, France), Clindamycin 300 mg (Fresenius Kabi, Bad Homburg vor der Höhe, Germany), Vancomycin 250 mg (Canonsburg, PA, USA) and 50 mg of Amphotericin B (Fungizone®, Bristol-Myers Squibb, New-York, USA). Twenty esophagi were then transported to our laboratory in the same ATB/ATM solution at 4 °C. At reception, the solution was renewed and the esophagi were decontaminated at 25 °C, under an agitation of 225 rpm for 24 h. The objective was to remove most of the microbial contamination before the decellularization.
After the decontamination step, the esophagi were rinsed 3 times in sterile water for 1h under agitation to remove the residual ATB/ATM. The esophagi were then placed in a bioreactor (Synthecon, Texas, USA) a rotating perfusion system that facilitates decellularization by mechanical action. The procedure was performed in sterile conditions under a laminar flow hood. The bioreactor was composed of a support for the esophagus (Fig. 1c) placed in a rotating glass cylinder, allowing a continuous unidirectional flow irrigating the lumen of the graft. This flow inside the lumen was generated with a 500 mL bottle connected in series to a peristaltic pump and the cylinder. The cylinder and the bottle were filled with a solution of 2% Sodium Dodecyl Sulfate (SDS, Euromedex, France) and 5 mM Ethylene Diamine Tetraacetic Acid (EDTA, Euromedex, France). The SDS/EDTA solution was renewed after 24 h and the esophagus was treated for a total of 72 h under constant rotation (27 rpm) and flow inside the lumen at 27 mL/min (Fig. 1d).
Treated esophagi were then rinsed inside the cylinder with sterile water for four cycles of one hour and one last cycle of two hours with the same parameters of flow and rotation.
To remove the residual SDS, esophagi were then incubated with an ion-exchange resin Amberlite® XAD16N (Sigma, France) under agitation at 25 °C, 225 rpm, in an orbital agitator (Intelli‐mixer, ELMI, Riga, Latvia).
To remove DNA, DHE were incubated for 3h at 37 °C and 4 rpm on the rotating agitator with 10 U/mL of DNase (Pulmozyme®, Roche, Boulogne‐Billancourt, France) in 22.5 mL of Phosphate Buffer Saline (PBS) with calcium and magnesium (Eurobio, Courtaboeuf, France). The DNase was then rinsed in PBS/EDTA and DHE were then stored at 4 °C in PBS, before further characterization.
Optimization of the SDS-removal method
In order to optimize SDS removal step for a clinical grade application, ion-exchange resin was replaced by an activated charcoal cartridge, commonly used for toxic elimination in intensive care units. The SDS released from the DHE is adsorbed by the cellulose-coated activated charcoal.
After SDS treatment and initial rinsing steps, the esophagi were placed in a 500 mL bottle filled with sterile water and connected in series to the Adsorba 300C (Baxter, USA) cartridge forming a closed circuit with a peristaltic pump (27 mL/min). The bottle was placed on a warming magnetic agitator Cimarec™ (Thermo Fisher Scientific Inc., Waltham, USA) at 30 °C and 100 rpm and the solution was continuously filtered for 72 h (Fig. 1e). The final protocol is resumed in the Fig. 1f.
Characterization of the DHE
Sterility was evaluated on samples of esophagi before and after ATB/ATM treatment and at the end of the whole process. Samples were incubated in Schaedler broth (Biomérieux SA, Craponne, France) for 10 days and then seeded into Chocolate agar + PolyViteX (Biomérieux SA) for aerobic culture, Columbia agar + 5% sheep blood (Biomérieux SA) for anaerobic culture, and Sabouraud chloramphenicol gentamicin agar (Bio‐Rad Inc, Hercules, USA) for fungal culture. The bacterial media were incubated at 37 °C for 8 days and the fungal medium at 30 °C for 11 days.
Histology, elastin quantification and scanning electron microscopy
Analysis of the general structure and residual cells was carried out both on native and decellularized esophagi. Samples were fixed in 4% buffered Formalin (PFA, Alfa Aesar™, Thermofisher, Germany) then embedded in paraffin. Five micrometers-thick sections were stained with hematoxylin–eosin–saffron (HES) for the general structure and residual cells, picrosirius red for the collagen and orcein for the elastic fibers. Images were obtained with a LEICA DM6 microscope.
Elastin in native esophagus and DHE (n = 2), was quantified using Fastin™—Elastin assay kit (Biocolor life science assays, UK), according to the manufacturer’s instructions. The mucosa/submucosa and the muscle/adventice were mechanically separated using forceps. Humidity was withdrawn by pressing samples in a sterile compress. They were then cut into small pieces weighing between 30 and 80 mg. Elastin was solubilized from the tissue in 750 µl of 0.25 M oxalic acid, at 100 °C for one hour. Two consecutive extractions per sample were performed. The extracted elastin was precipitated and centrifuged at 13000 g. It was then stained with Dye Reagent for one hour, centrifuged, and the supernatant was eliminated. Elastin bound dye was then solubilized in Dye Dissociation reagent. In parallel to the samples, known concentrations of elastin standard, ranging from 0 to 750 µg/ml were used to establish the standard curve.
To calculate the total elastin concentration in the tissue, the sum of the two extractions was divided by the weight of the samples. The final concentration was represented as µg of elastin/mg of wet tissue. Results were then compared with an unpaired t-test, using Prism software (GraphPad).
To perform the scanning electron microscopy (SEM) analysis, circumferential segments of native and decellularized esophagi were fixed in glutaraldehyde 2% (Sigma) diluted in cacodylate buffer 0.1 M (Alfa Aesar™, Thermofischer, Germany), pH 7.3 at 4 °C for 24 h. Samples were placed without any drying/metallization onto a cool stage (Deben) and cooled down to − 25 °C. SEM micrographs were acquired with a Scanning Electron Microscope Gemini SEM 300 (Zeiss) operating in variable pressure mode (60 Pa) at 10 keV.
DNA quantification and electrophoresis on agarose gel
To evaluate the efficiency of the decellularization, DNA extraction and quantification were performed. Samples of DHE before and after DNase treatment were lyophilized, weighed and digested overnight by proteinase K at 56 °C. The DNA was extracted with DNeasy Blood and Tissue kit (Qiagen, Hilden, Germany) according to the manufacturer’s protocol. Extracted DNA was quantified by Nanodrop (Thermofisher) and the results were expressed as ng/mg of dry tissue. The DNA fragments’ length was evaluated after electrophoresis for 1h at 150 V in a 2% agarose gel with SYBR® safe (Thermofisher) and analyzed with iBright1500 (Invitrogen). Image acquisition was done on auto-exposure mode, for an exposure time of 23 ms.
Glycosaminoglycans (GAG) purification and quantification
Total sulfated glycosaminoglycans (GAGs) were extracted from dried/lyophilized native and DHE samples and purified as previously described19. Total sulfated GAGs were quantified using 1,9 dimethylmethylene blue procedure (DMMB, Sigma Aldrich, Saint Louis, USA). Absorbance was read at 656 nm on a Spark® multimodal plate reader (TECAN, Switzerland) giving the optical density (OD) for each well.
Immunostaining of the GAGs was performed as previously described20. Slides of native and decellularized esophagus were incubated for 10 min with 50‐mM NH4Cl and saturated with 3% bovine serum albumin/PBS for 45 min. GAGs were then stained overnight at 4°°C with phage display antibodies linked to VSV (vesicular stomatitis virus) sequences: EV3C3V (1:5) binds to heparan sulfate (HS) and LKN1 (1:5) binds to dermatan sulfate (DS), an epimerized form of chondroitin sulfate (CS) also named CS‐B. All dilutions were performed in PBS 1X / bovine serum albumin 3%. Bound antibodies were detected with rabbit anti‐VSV polyclonal antibody followed by incubation with a goat anti‐rabbit antibody coupled to FITC. Controls of immunostaining were obtained with anti‐VSV and goat anti‐rabbit FITC staining without primary antibodies. The slides were also stained with 4′,6‐diamidino‐2‐phenylindole (DAPI) to observe cell nuclei.
Fluorescent intensity of the images was acquired with the LSM 800 microscope (Zeiss, Oberkochen, Germany) and analyzed with the software ZEN Blue edition V 2.1.
Heparin/GAG ELISA competition assay towards heparin binding proteins (HBP)
GAG abilities to interact with HBP were evaluated by an ELISA based competition assay as described21. Briefly, 96- wells ELISA plates (Costar) are coated with heparin-BSA conjugate and optimized doses of Recombinant Human HBP (FGF-2; VEGF) were added to each well, together with increasing concentration of extracted GAG. This allows to quantify the abilities of soluble GAG tested to interact with HBP and compete with immobilized-heparin. After washing steps, HBP remaining bond to coated heparin-BSA on plate are quantified by specific antibody sandwiches according to classical ELISA detection21. The maximal binding (100%) was determined in presence of the HBP and in absence of extracted GAG. Inhibitory Competitive dose (IC50) was defined as GAG concentration (µg/ml) that inhibit 50% of the HPB binding to immobilized heparin. IC50 value obtained for Heparin positive control samples was used to define a relative affinity of 100% and to report % of affinities of other GAG. For instance, a tenfold higher IC50 value of GAG sample “x” as compared to GAG sample “y” for an HBP corresponds to a tenfold lower affinity of GAG “x” as compared to GAG “y”, since tenfold higher concentration of “x” than “y” are necessary to compete for HBP binding to heparin.
In-vitro biocompatibility assays
Cytotoxicity analyses on Balb/3T3 cells
The cytotoxicity of the DHE was evaluated according to the European Standard ISO 10993-5-200922. Balb/3T3 clone A31 cells (ATCC, U.S.A) were cultured in Dulbecco’s modified eagle medium (DMEM) High Glucose 4,5 g/L (ATCC, U.S.A.) supplemented with 10% bovine calf serum (BCS, ATCC) and 1% ATB/ATM. Cells were seeded at 15 000 cells/cm2 in a 48 wells plate and incubated at 37°C for 24 h. DHE samples of 5 mm diameter were prepared with a biopsy puncher (Kai medical, Solingen, Germany) and incubated with 200 µL/well of extraction medium (same composition as the Balb/3T3 culture medium but supplemented with 2% BCS). At 24 h, the supernatant was used to replace the culture medium of the Balb/3T3 cells and culture medium of the control cells was renewed. Cells were cultured for 72 h and then detached by a trypsin treatment. Viability was evaluated by Annexin V-FITC and 7AAD staining (Beckman Coulter France, Villepinte, France) according to the manufacturer’s instructions. Results were analyzed by flow cytometry using the Attune™−NxT cytometer (ThermoFiscer scientific).
Lymphocyte proliferation assay
To assess the in vitro immunogenicity of the DHE, human peripheral blood mononuclear cells (hPBMCs), were isolated on Ficoll gradient and stained with carboxyfluorescein succinimidyl ester fluorescent 5 μM (CFSE, Life Technologies, Carlsbad, USA) according to the manufacturer’s instructions. Samples of 5 mm diameter of DHE were placed in a 96 wells plate. PBMCs were seeded at 104 cells/well in a culture medium composed of Roswell Park Memorial Institute glutamax (RPMI, Gibco® Waltham, MA, USA) supplemented with 10% FBS and 1% ATB/ATM. PBMCs were seeded alone, in contact with fragments of DHE or with 5 μg/mL of phytohemagglutinin‐L (PHA-L, Roche) as a positive control. After 5 days, cell proliferation was analyzed by measuring the fluorescence intensity decrease by flow cytometry with the Attune™−NxT (Thermofisher, USA). The percentage of division (PD) was calculated by using the formula:
PD = divided events/(divided events + undivided events).
F of the DHE with human mesenchymal stromal cells (hMSCs)
In order to perform a biocompatibility test of the DHE and to show its ability to support cell attachment, lack of cytotoxicity, migration and proliferation, we seeded bone marrow hMSCs at passage 2 on samples of DHE. The hMSCs were isolated and characterized as previously described15,23.
Segments of 2 cm of three DHE were prepared and cultured in a 50 mL tube Cellstar® CELLreactor™ (Greiner bio one) filled with a suspension of 5.106 cells in 30 mL of MSC culture medium (Fig. 5c). Samples were placed on a rotating agitator (Intelli‐mixer, ELMI, Latvia) at 1 rpm and 37 °C. Medium culture was renewed after 24 h to remove unattached cells and then every 72 h. Samples were cultured for 14 days, then fixed with PFA 4% and analyzed by HES coloration.
Qualitative evaluation of metabolic activity by colorimetric test after DHE seeding
Qualitative evaluation of metabolic activity after cell seeding on DHE were analyzed using CellTiter 96® AQueous Non-Radioactive Cell Proliferation Assay (MTS, Promega, USA). About one cm2 of DHE (3 before and 3 after cryopreservation) after 14 days of culture with hMSC were incubated at 37 °C, 5% CO2 for 1 h into 600 µL culture medium with 1% of MTS solution. Control group consist to DHE without cell. The purple color observed in the DHEs correspond to the presence of living cells.
hMSC quantification after DHE seeding
hMSC quantification was performed on six samples of DHE (3 before and 3 after cryopreservation) after 14 days of culture with hMSC. About one cm2 of DHE seeded with hMSC was incubated during 5 min with propidium iodide at 30 µg/ml (Sigma Aldrich) labelling nucleus of dead cells then incubated during 5 min with Hoechst 33,345 at 10 µg/ml (Sigma Aldrich) labelling nucleus of all cells. After several PBS washings, four images were acquired at four locations using a fluorescence microscope (Spinning disc, Nikon). All images are a Z-stack projection of 10 to 20 sections spaced from 3 µm on thickness of 30 to 60 µm. Images were processed using the software Image J to count viable cells and determine hMSC concentration per mm3 of DHE.
Residual SDS quantification
The concentration of residual SDS was measured with a spectrophotometric method, based on the use of a carbocyanine dye (Stains-all) the color of which changes from intense fuchsia to yellow upon addition of SDS.
Buffers and reagent solutions
Stains-all was dissolved in N, N-dimethylformamide to give a stock solution of 2.0 mg/ml. Working solutions were diluted 1:20 with MilliQ water into dark amber Falcon™ tube and stored in the dark at 2–8°C.
Standards and sample preparation
SDS standard concentrations were prepared by dilution of the 1% stock solution with PBS. Six SDS standard solutions were prepared, ranging from 0.01 to 0.00025% SDS and stored at room temperature. The linearity of the curve representing absorbance versus concentration was validated prior to the dosage.
Samples of DHE were lyophilized and stored at 4°°C. Pieces up to 30 mg were weighed and placed into 2 mL tubes with 1 mL of PBS. Samples were grinded with a Tissue Lyser (Qiagen) at 50 rpm for 15 min and placed in an incubator for 48 h at 37°C. Supernatants were collected and analyzed for the presence of SDS.
The concentration of the SDS were also quantified in the PBS in which DHE were preserved after decellularization (500 mL/5cm long DHE).
Spectrophotometric analysis of detergent concentration
All microtiter assays were performed in a Dulbecco’s PBS buffer.
PBS buffer (210 µL/well) was dispensed using a multichannel pipet. 15 µl of each standard, or sample was dispensed into the well of a 96 wells “ultra-low UV” (Corning). A multichannel pipet was then used to dispense 75 µl/well of Stains-all working solution. The microtiter plate was immediately placed in the incubator/reader and shaken for 20 s, allowed to incubate unshaken for 20 s, and then read for absorbance using a 450 nm filter with a microplate reader “Varioskan Lux” (Thermo Scientific). Results were then standardized as μg of SDS/mg of dry DHE.
Biomechanical properties of DHE using uniaxial traction assays as previously described15 were performed. Two frozen/thawed and 2 unfrozen DHE were used: DHE were cut into approximately 1 cm wide and 3 cm long strips, in the longitudinal and in the transverse directions.
Prior to the experiment, the dimensions of each sample were measured using a caliper. The tensile tests were performed at a strain rate around 0.6% s−1, continuously, until the rupture of the sample. During each test, the displacement of the grips and the force were recorded every second. Images were also acquired to verify that no slipping occurred during the assay.
The force was divided by the initial section of the sample to obtain the nominal stress. Stretch was determined through the machine displacement, divided by the initial sample length. Four parameters were extracted from the nominal stress versus stretch curve24: the tangent modulus of the linear region, the heel region length, the failure stretch and the ultimate tensile stress.
Development and evaluation of a cryopreservation protocol
In order to create a ready to use DHE bank, a cryopreservation method was validated. This procedure was carried out under sterile conditions. DHE were immersed in 100 mL RPMI (Invitrogen) with 10% dimethylsulfoxid (DMSO, WAK) and 0.8% human albumin. DHE and cryoprotectant solution were packed in capton-teflon bags (Hemofreeze, MedHem Science, NL). Samples were slowly cooled, at − 1 to − 2°C/min between 4 and − 10°C, − 2°C/min between − 10 and − 30°C, − 5°C/min between − 30 and − 50°C then − 10°C/min down to − 160°C, in Freezal (Air Liquide, France). DHE were cryopreserved for at least 3 weeks and thawed for further analyses. For thawing, they were kept at room temperature for 8 min then immersed in a water bath of 40°C until complete defrosting. The bags were then decontaminated, thawed DHE taken out and immersed in a series of decreasing concentrations of DMSO (8%, 4% and 2% respectively). They were then preserved in physiological serum at 4°C before further evaluation.
To evaluate the safety and incidence of the cryopreservation method on the quality of the esophagus, we analyzed and compared samples before cryopreservation (DHE stored in PBS at 4°C) and after thawing (cryo-DHE). For that, we assessed sterility, cytotoxicity, lymphocyte proliferation assay, histology with HES staining, SEM and biomechanical properties as previously described.
Statistical analysis was performed with SPSS Statistics IBM Corp. (Released 2019. IBM SPSS Statistics for Macintosh, Version 26.0. Armonk, NY: IBM Corp). Data were expressed as proportions (%) and mean ± standard deviation. Univariate analysis was done with a Mann–Whitney U-test for the quantitative data (DNA, GAG and SDS quantification, as well as cytotoxicity analysis). Results of the elastin quantification were compared with an unpaired t-test, using Prism software (GraphPad).
Statistical analysis was performed using Jamovi for the biomechanical tests. Prior analysis, some values were removed: 2 heel-region for which the sample was already stretched initially, and 1 sample with abnormal linear part. T-tests were performed to compare the effect of the direction of the loading or the effect of the freezing on heel region length and tangent modulus.
Ethics approval and consent to participate
Esophageal grafts were harvested from human deceased donors following a brain-death or Maastricht category III circulatory arrest. Non-opposition consent was obtained for research purposes. The graft retrieval authorization was granted by the “Agence de la Biomédecine”, Grant Number PFS18-018 in Saint-Louis Hospital, Université Paris Cité, France.