Individualized tissue‐engineered veins as vascular grafts: A proof of concept study in pig

Personalized tissue engineered vascular grafts are a promising advanced therapy medicinal product alternative to autologous or synthetic vascular grafts utilized in blood vessel bypass or replacement surgery. We hypothesized that an individualized tissue engineered vein (P‐TEV) would make the body recognize the transplanted blood vessel as autologous, decrease the risk of rejection and thereby avoid lifelong treatment with immune suppressant medication as is standard with allogenic organ transplantation. To individualize blood vessels, we decellularized vena cava from six deceased donor pigs and tested them for cellular removal and histological integrity. A solution with peripheral blood from the recipient pigs was used for individualized reconditioning in a perfusion bioreactor for seven days prior to transplantation. To evaluate safety and functionality of the individualized vascular graft in vivo, we transplanted reconditioned porcine vena cava into six pigs and analyzed histology and patency of the graft at different time points, with three pigs at the final endpoint 4–5 weeks after surgery. Our results showed that the P‐TEV was fully patent in all animals, did not induce any occlusion or stenosis formation and we did not find any signs of rejection. The P‐TEV showed rapid recellularization in vivo with the luminal surface covered with endothelial cells. In summary, the results indicate that P‐TEV is functional and have potential for use as clinical transplant grafts.

. In contrast, the management of venous diseases has evolved little during the same period (Onida & Davies, 2016).
Many diseases of the venous system can be treated pharmacologically, but for some conditions reconstructive surgery are required to restore venous anatomy and function. Vital venous structure may be lost due to thrombosis, trauma or surgical tumour resections.
Sometimes autologous veins are available as grafts in vascular reconstructions, but for many patients a biological individualized vascular graft would be of great importance to enable restoration of crucial venous function.
For venous reconstructions, different grafts substitutes have been evaluated including Polytetrafluorethylene grafts, cryopreserved or decellularized allogenic veins and vein conduits constructed of autologous peritoneum or bovine pericardium.
A promising technology to improve current vascular grafting is decellularization of native blood vessels from human or animal donors followed by reconditioning with autologous material to generate individualized tissue-engineered vascular grafts. Decellularization employs chemical, enzymatical and/or physical methods to remove all donor cells from the extracellular matrix (ECM; Gilbert et al., 2006).
The ECM possesses an ideal three-dimensional structure for cell attachment and proliferation due to abundance of functional proteins, surface structure and natural viscoelastic behavior to withstand blood pressure. Decellularized blood vessels have been evaluated as arterial vascular grafts (Lawson et al., 2016;Schaner et al., 2004) and most clinical trials utilizing decellularized vascular grafts have focused on providing vascular access for patients requiring hemodialysis. Even though some biological grafts, for venous reconstructions have shown promising results, most grafts have had long term patency problems (Chemla & Morsy, 2009;Das et al., 2011;Gui et al., 2009;Quint et al., 2011;Wystrychowski et al., 2014).
There is still a need for further optimizing biological grafts for clinical use and one way to optimize the vessel wall regeneration could be reconditioning of decellularized veins with the recipient's autologous blood. Perfusion of functional organs with whole blood or blood-derived solutions prior to conventional transplantation has been shown beneficial for preserving the organ physiology and functionality (Hosgood & Nicholson, 2011;Nicholson & Hosgood, 2013;Steen et al., 2016). The ambition is to prepare the graft with autologous components preventing rejection and facilitating rapid in vivo cellularization of the graft after implantation.
VERIGRAFT is developing clinical-grade grafts under the trade name personalized tissue-engineered vein (P-TEV). We have used the P-TEV protocols of decellularization and reconditioning to produce porcine individualized tissue-engineered grafts for preclinical testing (referred to as P-TEV in this text).
To study safety and functionality of P-TEV, these grafts were evaluated in an in vivo porcine model of vena cava transplantation for up to 35 days. The results showed efficient graft recellularization and full patency throughout the whole study. -819 to a perfusion bioreactor system ( Figure 1a) and decellularized using protocols previously described (Simsa et al., 2018(Simsa et al., , 2019. Vessels were perfused at 100 ml/min, 37°C and 115 rpm agitation for 2 h each with 1% Triton X-100 (Merck Millipore), 1% Tri (n-butyl)phosphate (Merck Millipore) and 40 U/mL deoxyribonuclease (VWR), all containing 0.5% antibiotic-antimycotic (AA, Thermo Fisher Scientific). Between each reagent change, vessels were washed 3 � 5 min in H 2 O. This process was repeated for seven days. For overnight incubation, the veins were perfused with a 5 mM ethylenediaminetetraacetic acid (EDTA; Merck Millipore) solution and the perfusion and agitation were lowered to 70 ml/min and 50 rpm respectively. The vessels were washed with 5 mM EDTA for 48 h and PBS for 24 h, both containing 0.5% AA. After decellularization, the vessels were removed from the bioreactor, biopsies were taken for DNA quantification, histology, and immunohistochemistry before peracetic acid sterilization and final washes in PBS under sterile conditions. The sterilized decellularized veins were either frozen in PBS at −80°C or directly used for subsequent whole blood perfusion. One vein was decellularize for each animal to be transplanted.

| Preparation of P-TEV
From the leg of each pig to be transplanted with P-TEV, 50 ml peripheral venous whole blood was collected in heparin vacutainers (BD) one week pre-surgery under sedation with Tiletamine/zolazepam (Zoletil 50 mg/ml + 50 mg/ml, 0.06 ml/kg) and Dexmedetomidine (Dexdomitor 2.5 mg). Blood samples were stained with Türk's solution (Merck) and leukocytes were counted in a Bürker chamber to ensure a normal physiological status. 25 ml blood was mixed immediately with 25 ml STEEN Solution (XVIVO Perfusion), 0.5% Antibiotic-antimycotic Mix (Thermo Fisher Scientific), 80 ng/ml vascular endothelial growth factor (Cellgenics) and 10 ng/ml fibroblast growth factor (R&D Systems). 5 μg/ml acetylsalicylic acid (Sigma-Aldrich) was added to terminally inhibit all contained thrombocytes. The complete blood solution was added to the decellularized blood vessel inside a reconditioning bioreactor, which allowed circulation of the blood solution through and around the vessel in a vertical position at 2 ml/min (Figure 1b-e). This process was performed in a laminar flow hood at room temperature for seven days, during which the glucose level was measured and kept between 3 and 8 mM. After perfusion, the graft was harvested, rinsed, biopsied for DNA quantification and histology and kept in PBS + AA until use (maximum two hours). During blood procurement and production, a system was used to allow traceability of the autologous pig blood and the resulting graft to allow administration in an individualized, strictly autologous manner.

| Animal model for transplantation of P-TEV
The in vivo experiments were performed after prior approval from the local ethics committee for animal studies at the administrative court of appeals in Gothenburg, Sweden. Eight female pigs (bodyweight 54-69 kg) of a mixed breed of Yorkshire, Hampshire and Swedish Pigham were used in this study; two for sham surgery (the pig's own vena cava was cut and re-sutured) and six for P-TEV transplantation. The pigs were cared for in accordance with regulations for the protection of laboratory animals and the pigs were housed together before and after surgery. The animals were housed together with two-three pigs in a room of 28 m 2 . The rooms had bedding consisting of wood shavings and straw and the pigs were fed twice daily and had free access to water. The general condition of the pigs is observed daily during the acclimatization period as well as after operation. Acetylsalicylic acid (Trombyl, Pfizer) 160 mg was given orally once daily during the whole study starting six days before surgery and rivaroxaban (Xarelto, Bayer) 2 mg/kg was given orally twice daily during the whole study starting the day before surgery. All surgical procedures were carried out under

| Immunohistochemistry
Immunohistochemical staining was performed on paraffin sections by rehydrating samples following standard procedures and performing

| Scanning electron microscopy
Tissue samples were rinsed in PBS (Medicago) and submerged in 2.5% glutaraldehyde fixative (Sigma-Aldrich) at room temperature.

| Rotational thromboelastometry
Rotational thromboelastometry (ROTEM) is a method which detects clot formation in blood samples. ROTEM measurement was performed to analyze the influence of the intrinsic (INTEM, activated by trauma inside the vascular system) and extrinsic (EXTEM, activated by trauma outside the vascular system) coagulation pathways. Also, the influence of fibrinogen (FIBTEM) and heparin (HEPTEM) were analyzed.
Effects on blood coagulation in the presence of native (n = 4), decellularized (n = 8) and reconditioned (n = 8) vein tissue were analyzed. Vena cava segments were excised from cadaveric pigs of approximately 120 kg, stored in ice-cold PBS and frozen at −80°C.
Decellularization and reconditioning was performed as described above. Venous blood for reconditioning was received from four different human donors. Reconditioned blood vessels were harvested the same day as the ROTEM analysis was performed. 8 mm diameter tissue pieces were cut out from the vessels with tissue punchers and stored in PBS −/− until use. Blood for ROTEM analysis was collected from the same donors used for reconditioning, in sodium citrate tubes (BD). A single piece of vein tissue was added to 1.2 ml blood (from the same donor as used for reconditioning) and incubated at 37°C for 1h, after which ROTEM measurement was performed. All The analysis was performed with ROTEM Delta software.

| Statistical analysis
SPSS was used for statistical analysis. Statistical method used was Wilcoxon Signed-Rank Test for paired test of vein wall thickness before and after decellularization. In other cases, Kruskal-Wallis with Mann-Whitney post hoc was used. p < 0.05 was considered statistically significant. The data are reported as average ± Standard Error of the Mean (SEM).

| Decellularization and reconditioning of vascular grafts
One vein was decellularized and reconditioned for each pig to be transplanted and the process was successful for all six animals. Histological analysis of the native veins showed normal morphology and distribution of cells, whereas the decellularized vein segments lacked HÅKANSSON ET AL. presence of nuclei (Figure 2a-g). The morphology of the ECM appeared to remain intact after decellularization and there was no significant difference in vein wall thickness before and after decellularization (p = 0.7, data not shown) as measured on hematoxylin and eosin stained sections (Figure 2a-d). En face DAPI staining revealed no cell nuclei present on the luminal side after decellularization ( Figure 2g). Lymphocyte count in the autologous peripheral whole blood, used for reconditioning, were between 12 and 22 � 10 6 lymphocytes/ml, which is considered in the normal physiological range (Luke, 1953). After reconditioning, blood cells were identified on the surface (Figure 2h)  After the initial surgery with conventional technique, holding the intestines aside using wetted gauze while locating vena cava, one P-TEV transplanted pig had to be euthanized 15 days post-surgery due to intestinal complications. At dissection, the intestines had massive adhesions, probably due to the handling of the intestines during surgery, and the symptoms of the pig indicated ileus.

| Surgical transplantation of the vascular graft and patency after up to five weeks in vivo
Changing to retroperitoneal surgery technique was successful, although during surgery of one sham operated pig, the peritoneum broke and the intestines had to be handled by wetted gauze, as previously. The same pig had to be euthanized seven days postsurgery due to indicated ileus which was probably caused by the handling of the intestines. The remaining six pigs operated with retroperitoneal surgery technique (one sham and five P-TEV transplanted pigs), had no signs of ileus and no intestinal adhesions were found at dissection after euthanization.
Due to complications not related to the vena cava transplantation, two more P-TEV transplanted pigs had to be euthanized at early time points. One got limp on the left front leg due to a fractured kneecap and had to be euthanized 17 days post-surgery.
The other pig was euthanized three days post-surgery due to suspected hernia. In summary, the different time points for euthanization were the following: Sham (7 and

| Recellularization of P-TEV in vivo
Already at three days post-surgery, cells had started to repopulate the P-TEV graft. At 17 days post-surgery the P-TEV graft was well recellularized and after 5 weeks the number of cells in the P-TEV appeared to be equal to the native tissue. Importantly, no intimal hyperplasia was observed in the P-TEV, which in this study was predicted as a major potential complication (Figure 4). DNA content of the P-TEV was significantly increased during the time in vivo from 10.2 ± 8.0 ng/mg tissue after reconditioning to 55.1 ± 6.5 ng/mg tissue post-in vivo (p < 0.001).
It is not clear whether the P-TEV were recellularized from the connecting native vena cava via the anastomosis, from the adjacent surrounding soft tissue, from the blood in the vessel lumen or from a combination of all sites together. However, already at three days post-surgery both the proximal, center and distal parts of the P-TEV were occupied with cells ( Figure 5 a-

| DISCUSSION
Tissue-engineered vascular grafts obtained by decellularization steps are a potential alternative to current vascular grafts. In this study, we used decellularization and reconditioning to produce P-TEV and evaluated the safety and efficacy of the grafts in a porcine in vivo vena cava transplantation model for up to 35 days. All P-TEV were fully patent and did not show any signs of occlusion. Recellularization of the P-TEV was observed already from day three post-surgery, and five weeks post-surgery the P-TEV were equivalently cellularized compared with native vena cava, with the lumen covered with flattened CD31-positive cells, a marker for endothelial cells.
There are known differences between vascular regeneration mechanisms in animal and humans (Byrom et al., 2010;Sánchez et al., 2018) and it is not possible to conclude that the recellularization in humans will be as rapid as in pigs. However, animal models are our best options to study whole tissue regeneration, including immune response, and our results with full recellularization after 5 weeks indicated that P-TEV is a promising alternative for vascular grafting and motivates to proceed with long term studies towards clinical trial. Decellularized vascular grafts for arterial reconstruction have been described in multiple previous studies, while studies assessing grafts for venous reconstructions are absent (Katzman et al., 2005;Kovalic et al., 2002;Simsa et al., 2018Simsa et al., , 2019. Regarding arterial grafts, various studies suggests that recellularization of decellularized grafts with endothelial cells (ECs) prior to transplantation improves graft functionality and prevents thrombosis (Lin et al., 2018). For recellularization of vascular grafts, different cell sources have been applied in the past, including somatic ECs (Dahan et al., 2017;Quint et al., 2011), endothelial progenitor cells (Kaushal et al., 2001;Melchiorri et al., 2016;Quint et al., 2011), induced pluripotent stem cell derived ECs (Nakayama et al., 2015) or -825 embryonic stem cell derived ECs (Shi et al., 2013). Disadvantages of these methods are the required growth and expansion of cells in vitro, which could induce genetic variations and mutations as well as increase variability. Other disadvantages include regulatory hurdles, high cost, limited availability of cells and time-consuming processes.
In the present study, we utilized a decellularization method previously described (Simsa et al., 2018). Biomechanical data from this study showed unchanged maximum tensile strength and burst pressure of the decellularized scaffolds compared with native blood vessels, whereas an increase in stiffness of the vein was observed.
Quantifications of ECM proteins in the same study showed unchanged levels of insoluble collagen and a decrease in soluble collagen and glycosaminoglycans (GAGs). The basement membrane protein laminin was also reduced compared with native tissue. Our hypothesis was that shielding the ECM surface from thrombosis and neointimal formation can be achieved by preconditioning steps, that is by perfusion of the vascular graft with peripheral whole blood. The rationale behind this is that decellularized blood vessels present different collagen types, which have often been denatured to some degree.
Collagen is a known activator of the coagulation cascade, by initiating platelet binding and activation through the intrinsic coagulation pathway (Böer et al., 2015;Boeer et al., 2014;Wong & Griffiths, 2014), and contributes to the pathogenesis of venous thrombosis. By perfusing decellularized blood vessels with peripheral whole blood, the luminal surface of the grafts gets covered with different factors potentially important for avoiding clotting protein adhesion, quiescence of platelet activation and facilitating recellularization and acceptance of the vascular grafts. Data from the ROTEM analysis showed increased clotting time, increased clot formation time and decreased alpha angle which indicates better hemocompatibility and thereby a slower activation of the intrinsic coagulation system for the reconditioned tissue compared with native and decellularized tissue samples. This effect could be due to remnants of heparin in the P-TEV since addition of heparinase (HEPTEM analysis) removed these differences between the samples. Importantly, the reconditioning did not affect the coagulation properties in any negative way, but rather increased the clotting time which can be part of prevention of thrombosis. In addition, the growth factors in the reconditioning solution have the ability to bind to the P-TEV and can potentially support recellularization and differentiation of stem cells and promote tissue regeneration (Ikada, 2006;Ullah et al., 2019). In a separate bioinformatics study, autologous blood proteins bound to the graft after reconditioning are being analyzed. Preliminary proteomics data indicate that more than 150 proteins bind to the decellularized scaffold during perfusion with autologous peripheral whole blood (data not shown). Also, autologous blood cells attaches and the content of dsDNA in the graft increase during reconditioning (Figure 2). The DNA content of the P-TEV post-in vivo (55.1 ± 6.5 ng DNA/mg wet tissue) was lower compared with native vena cava tissue (Figure 2), however this was probably due to the formation of scar tissue in the healing process from the surgery which was not possible to dissect from the original P-TEV. DNA quantification of native tissue close to the anastomosis, with attached scar tissue, showed similar amounts (53.3 ± 3.5 ng DNA/mg wet tissue).
A limitation of the study is that a decellularized scaffold was not included as a control for the in vivo study. However, the aim of the study was not to compare functionality and mechanistical differences between decellularized scaffolds and P-TEV but only to show safety and functionality of the P-TEV since the literature already have shown problems with decellularized scaffolds, as described in the introduction.
Before application of novel medical products on humans, animal trials are required to assess safety and efficacy. While smaller animals lack physiological similarity to humans, common animal models for vascular grafts include dogs, rabbits, sheep and pigs (Mehran et al., 1991). In the present study, pig was chosen for its similar cardiovascular system compared with humans. Even though pig is known to have a higher incidence of intimal hyperplasia formation in graft studies compared with other model animals (Liu et al., 2018), no intimal hyperplasia was observed in any of the P-TEV in our study.
The thin vein wall facilitates ECM graft penetration and recellularization. The observation of gradual recellularization of the P-TEV  (Ketchedjian et al., 2005;Martin et al., 2005;Quint et al., 2011). So far, no successful studies in a relevant large animal have been published for grafts in the venous system. In our study, no vessel wall thickening, and no intimal hyperplasia was observed. A gradual recellularization in vivo during physiological condition may mimic the normal restoration of vascular cellular injuries.
The venous grafts in this study did not contain valves due to that these are absent in large veins of the pigs. For future use in humans to treat venous diseases such as chronic venous insufficiency, CVI, we believe that in vivo recellularization can restore normal morphology and function of vein valves. While other studies on the arterial side show patency issues when grafts were implanted without a large initial cell population on the intima (Dahan et al., 2017;Quint et al., 2011), our methodology is functional even on the venous side, with retained patency, graft acceptance in the body and efficient in vivo recellularization. Taken together, our results support that the tested approach is a promising alternative for the next generation of vascular grafts.
In the present study we showed that decellularized blood vessels, reconditioned with peripheral whole blood, can be transplanted into a pig animal model. By analyzing transplanted P-TEV after 3 days (n = 1), 2 weeks (n = 2) and 4-5 weeks (n = 3) we can conclude that recellularization appeared rapidly, already after 3 days, and the grafts were cellularized to a level comparable with the native vein tissue after 4-5 weeks with a protective endothelial layer on the luminal graft surface. All P-TEV were fully patent and without signs of coagulation, thrombosis, intimal hyperplasia or rejection. These results motivate to proceed with P-TEV in long term studies to further validate the product for clinical studies.

ACKNOWLEDGMENTS
We want to acknowledge the staff at the Department of Experimental Biomedicine at Gothenburg University. This study was partly network. The company VERIGRAFT AB holds a patent on peripheral whole blood perfusion of decellularized tissues and did also finance the project.

CONFLICT OF INTEREST
Lachmi Jenndahl, Tobias Gustafsson-Hedberg, Robin Simsa and Raimund Strehl were employees of the company VERIGRAFT which contributed financially, including salaries and study costs, and by providing laboratory space.