Hemocompatibility improvement of perfusion-decellularized clinical-scale liver scaffold through heparin immobilization

Whole-liver perfusion-decellularization is an attractive scaffold–preparation technique for producing clinical transplantable liver tissue. However, the scaffold’s poor hemocompatibility poses a major obstacle. This study was intended to improve the hemocompatibility of perfusion-decellularized porcine liver scaffold via immobilization of heparin. Heparin was immobilized on decellularized liver scaffolds (DLSs) by electrostatic binding using a layer-by-layer self-assembly technique (/h-LBL scaffold), covalent binding via multi-point attachment (/h-MPA scaffold), or end-point attachment (/h-EPA scaffold). The effect of heparinization on anticoagulant ability and cytocompatibility were investigated. The result of heparin content and release tests revealed EPA technique performed higher efficiency of heparin immobilization than other two methods. Then, systematic in vitro investigation of prothrombin time (PT), thrombin time (TT), activated partial thromboplastin time (APTT), platelet adhesion and human platelet factor 4 (PF4, indicates platelet activation) confirmed the heparinized scaffolds, especially the /h-EPA counterparts, exhibited ultralow blood component activations and excellent hemocompatibility. Furthermore, heparin treatments prevented thrombosis successfully in DLSs with blood perfusion after implanted in vivo. Meanwhile, after heparin processes, both primary hepatocyte and endothelial cell viability were also well-maintained, which indicated that heparin treatments with improved biocompatibility might extend to various hemoperfusable whole-organ scaffolds’ preparation.

a portal implantable functional tissue-engineered liver using a decellularized liver scaffold (DLS) and hepatocytes in rats 8 . Recently, this process has been scaled up, generating biocompatible scaffolds at a clinically relevant scale 9 . Although heterotopic transplantation of these acellular scaffolds is possible, the scaffolds remain thrombogenic, even with anti-coagulation following transplantation 10 . The thrombogenicity of the scaffold is one of the main obstacles to research on engineering clinical transplantable liver tissue 11 .
Surface-immobilized heparin on collagenous tissue has been reported to prevent thrombus formation 12 . We have developed a layer-by-layer (LBL) self-assembly heparin-coating technique to electrostatically immobilize heparin on rat DLSs 8 . Wang et al. 13 reported that heparin can be covalently immobilized on collagenous tissue by activating its carboxyl groups using a water-soluble carbodiimide via multi-point attachment (MPA). In addition, Larm et al. 14 developed a procedure in which nitrite-depolymerized heparin may be covalently immobilized on polyurethane tubing via end-point attachment (EPA). Heparin covalently immobilized on the surface of biomaterials through MPA or EPA has been shown to retain its antithrombogenic activity 13,14 .
The present study was intended to provide a method for producing blood compatibility with DLSs. We immobilized heparin using LBL electrostatic immobilization (Fig. 1A-left) and covalent immobilization via MPA (Fig. 1A-middle) or EPA (Fig. 1A-right). We evaluated the performance of the heparinized scaffolds by measuring immobilized heparin contents and release rates, coagulation time, platelet adhesion and activation, cell culture in vitro and implantation in vivo (Fig. 1B). These hemocompatibility improvement studies are the first step toward generating an engineered transplantable functional liver decellularized whole organ scaffolds.
Hereafter, /h-LBL, /h-MPA, and /h-EPA refer to DLSs with heparin immobilized using the LBL, MPA and EPA techniques, respectively. Non-heparinized DLS was used as control.

Results
Perfusion decellularization of porcine livers. Twenty-two whole livers were perfusion decellularized successfully out of 24 isolated livers from mini-pigs. The perfusion whole-organ decellularization process yielded a fully decellularized porcine liver with a translucent appearance (see Supplementary  Fig. S1A online). Histological evaluation revealed no remaining nuclei or cellular material throughout the entire thickness of the graft (see Supplementary Fig. S1B online). 4,6-diamidino-2-phenylindole (DAPI) staining (see Supplementary Fig. S1C online) and scanning electron microscope (SEM) analysis (see Supplementary Fig. S1D online) showed that the liver cells were completely removed, whereas the collagen and fibers were maintained. The remnant DNA extract was subjected to agarose gel electrophoresis to reveal the length of DNA fragments in the decellularized liver scaffold (DLS), which can barely be seen by visual inspection (see Supplementary Fig. S2A online). The DNA content in the normal tissue was 9.1 ± 0.2 μ g/mg, whereas the DNA content in DLS was 0.03 ± 0.01 μ g/mg (see Supplementary  Fig. S2B online).

Properties of DLSs.
Histological and immunohistochemical staining revealed that, compared to normal liver tissue, sulfated glycosaminoglycans (GAGs), collagen fibers, elastic fibers, and reticular fibers were well retained (see Supplementary Fig. S3A online) and ECM components, including collagen I and IV, laminin, and fibronectin, were preserved in DLSs (see Supplementary Fig. S3B online). The collagen content was almost 8-fold higher DLSs than in normal liver tissue (see Supplementary Fig. S3C online). Moreover, nearly 70% of the GAGs were retained in DLSs compared to normal liver tissue (see Supplementary Fig. S3D online).

Elimination of xenogeneous antigens in DLSs.
The porcine DNA sequences of porcine endogenous retrovirus (PERV), PERV polymerase (PERV Poly), swine leukocyte antigen DR alpha (SLA-DRA), swine leukocyte antigen 2 (SLA-2), sus scrofa cytochrome b (SsCytb) and, porcine beta actin (β -actin) were detected in native organs. Otherwise, these DNA sequences were absent in the ECM after decellularization ( Fig. 2A). The immunohistochemistry for α -Gal showed although native materials were found to be Gal positive, the most remarkable reduction of the Gal epitope can be observed in the DLSs (Fig. 2B).

Homogeneous heparin immobilization on the DLSs.
To compare the efficiency of three heparinized methods, the different heparin solutions with the same heparin concentrations (1, 0.5 or 0.25 g/L) were perfusion through portal vein (PV) at the same flow rate (100 mL/min) for 4 h in total. Analyzed samples were taken from 12 different sites (S1 to S12) as shown in Fig. 3A, which involved the porta hepatis to margin of the right, median and left lobes of the whole liver. Based on the heparin content results of S1 to S12 from each liver with different heparin treatments, the coefficients of variations (CV%) for three methods were 4.09% (LBL), 1.09% (MPA) and 1.63% (EPA) respectively at 1 g/L heparin preparation (Fig. 3B). Table 1 shows three types of heparinized scaffolds gained different contents of heparin immobilization. The /h-EPA scaffold gained the highest heparin content (p < 0.001) at 1 g/L heparin preparation. Heparin-toluidine blue (TB, C.I. Basic Blue 17) interaction was employed as analysis of functional groups present on the polyanion chain. The colorimetric assay of heparin by means of metachromatic dyes TB is easy and quick. The heparinized DLSs showed even and positive staining with TB throughout the entire matrix, whereas control DLSs did not stain with TB (Fig. 3C). Compared with the /h-LBL and /h-MPA scaffolds, the inside of /h-EPA scaffolds were more deeply stained. SEM imaging showed that the surface morphology of /h-EPA group DLSs was similar to the non-heparinized DLSs (Fig. 3D).
However, to compare the surface properties of the tissues immobilized with heparin using different methods impartially, the amounts of heparin immobilized on all test samples were intentionally adjusted to be approximately the same via shorten the heparin treatment time of /h-LBL and /h-EPA scaffolds. Hereafter, the heparin contents of /h-LBL, /h-MPA, and /h-EPA scaffolds were 10.36 ± 0.06, 10.25 ± 0.08 and 10.51 ± 0.05 μ g/mg dry weight, respectively, for the following experiment tests. and sustained release (Fig. 4B) tests showed that heparin was quickly released from the /h-LBL, /h-MPA, and /h-EPA scaffolds on day1, followed by slow release for 7 days. After 7-day accumulativer or sustained release test, above 33% or 42% of heparin retained in /h-EPA scaffolds, compared with original heparin content, respectively.

Heparinized treatments reduced platelet adhesion and activation. Platelet adhesion studies
were conducted with the purpose of determining the blood compatibility of each test sample, because platelet adhesion is one of the most important procedures during blood coagulation on prosthetic surfaces 16 . Field counts of adherent platelets were made using representative micrographs at similar magnifications to meter the differences observed using SEM. Fig. 6A shows examples of SEM micrographs of platelets adhered to test samples. The number of platelets adhered to all test scaffolds were counted and are compared in Fig. 6B. The results showed that the heparinized DLSs retained significantly fewer platelets than the control scaffolds (p < 0.05). In addition, fewer platelets adhered to the /h-EPA scaffolds (117 ± 13 platelets/15000 μ m 2 ) than to the /h-LBL (177 ± 14 platelets/15000 μ m 2 ) and /h-MPA scaffolds (163 ± 11 platelets/15000 μ m 2 ) (p < 0.05). However, no significant differences were observed between the /h-LBL and /h-MPA scaffolds (p > 0.05).
Platelet factor 4 (PF4) was used to detect platelet activation by modified biomedical materials 17 . Figure 6C shows the PF4 concentrations in plasma after the scaffolds contacted with blood; it could be observed that for the heparinized DLSs the PF4 concentration showed a prominent decrease compared with normal plasma or control DLS. Combined with the numbers of the adhered platelets, these results indicated that platelet activation on the heparin treated DLSs was greatly suppressed.
In vivo implantation of DLSs. The mechanical properties of the vasculature deprived of the endothelial layer supported the surgical reconnection of the vessel stumps of the scaffold to the recipient's renal vein (RV) and splenic vein (SV). After connection with RV, blood flowed well within the whole scaffold (Fig. 7A). DLSs were filled with blood within 5 min, which swelled gently and acquired a color similar to normal livers (Fig. 7B). Satisfactory blood flow throughout the heparinized DLSs was observed for 60 minutes (Fig. 7C), whereas no blood flow throughout the control DLSs was observed after 20 min perfusion. No leaks were detected and no adverse events were recorded during surgery. At the end of perfusion, the graft was harvested for histologic analysis. H&E staining revealed vascular structures of Heparinized DLSs maintained good cellular compatibility. Rat primary hepatocytes and human umbilical vein endothelial cells (HUVECs) were cultured in vitro to measure the cytocompatibility of the heparinized DLSs. H&E staining revealed that primary hepatocytes aggregated together to form multi-layer cell structure after 3 days seeded on both control and heparinized DLSs (Fig. 8A). Immunohistochemical analysis of cell proliferation showed that approximately 80% of the cultured cells stained positive with proliferating cell nuclear antigen (PCNA) antibody (Fig. 8B). Morphologies of the HUVECs cultured for 3 days on the control and heparinized DLSs are shown in Fig. 8C. It was observed that the HUVECs exhibited a typical flattened morphology on the scaffolds. Immunohistochemical analysis of cell proliferation showed that approximately all of the cells stained positive with PCNA (Fig. 8D). The results indicated heparinized DLSs could support cell adhesion and growth as the same as control scaffolds.

Discussion
Development of liver grafts using traditional tissue engineering approaches has been hampered by a lack of oxygen and nutrient diffusion, which limits the thickness of the cell scaffolds to sheet-like structures 18 . Perfusion decellularization allows liver tissue regeneration of an intact liver structure at a clinically relevant scale by meeting metabolic demands via intact vasculature. In the foreseeable future, the optimal approach for producing whole-liver grafts is to repopulate the decellularized matrix scaffold with parenchymal and nonparenchymal cells and conditioning the graft in vitro through perfusion with nutrients 19 . The repopulated graft can be heterotopically transplanted as an auxiliary support to a dysfunctional organ 11 .
From our own perspective, human livers will be the preferred source of organs for clinical trials of decellularization techniques; however, if this process is used for commercial-scale production, pig livers, which are easy to obtain and their quality is easy to ensure, could be preferable in order to reduce costs and for ease of standardization and quality assurance. In this study, SDS-based perfusion decellularization was applied to whole porcine livers to generate xenoantigen-reduced and biocompatible organ scaffolds that preserve the ECM composition and architecture, at a clinically relevant scale.
Although orthotopic or heterotopic transplantation of these acellular scaffolds is possible, the scaffolds are thrombogenic even with anti-coagulation 10,20 . The prothrombotic properties of collagen are a major drawback in applications for DLSs. Endothelial coverage of the lumen of the vasculature is essential to prevent thrombosis and to provide proper vascular function. Re-endothelialization of the vascular network of the DLSs would be a major step on development of DLSs 21 .However, in practice, endothelial cells must be delivered in a manner that appropriately localizes them to the vascular conduit surfaces and the hepatic sinusoid cavity, and achieving near-perfect endothelial cell coverage of the vasculature in the scaffold is challenging. For this reason, most attempts at transplantation of whole-liver scaffolds have been limited to only a few hours 11 . In a previous study, our group developed a promising solution to deposit heparin layers on the DLSs, rendering the whole-liver graft blood-compatible 8 . In the present study, the characteristics of the porcine DLSs immobilized with heparin via LBL, MPA, and EPA were estimated. The non-heparinized counterpart was used as control. We build on our previous work to show that perfusion-decellularized DLSs can be heparinized effectively, reducing the thrombogenicity of the scaffolds.
In order to scale-up the method of heparin immobilization adapted from previous thin tissues 13,22 and rodent studies 8 to porcine whole-liver scaffolds, the solution perfusion was employed instead of immersion. According to the literatures, the final concentrations of heparin solution were 2 g/L 8 , 4.2 g/ L 13 Table 1. Heparin contents of different heparinized DLSs (μ g/mg dry weight, n = 3, ± SD)   that EPA is more efficiency of heparin immobilization on DLSs than LBL or MPA method (Table 1). In addition, 2 g/L heparin solution for LBL 8 and 3.3 g/L heparin solution for EPA 22 were tested on DLSs, the heparin concentrations immobilization on DLSs were reached 12.82 ± 1.13 μ g/mg dry weight and 20.67 ± 0.25 μ g/mg dry weight (n = 3), respectively. Obviously, LBL modification process is the most complicated and tedious. Furthermore, the heparin contents can be conveniently adjusted through heparin solution perfusion time.
Heparin is an anionic linear polysaccharide consisting of two repeating disaccharide units: D-glucosamine-L-iduronic acid and D-glucosamine-L-glucuronic acid 23 . Heparinized biological tissues have been used to manufacture blood-contacting prostheses 12 . The heparinization method applied for producing the /h-LBL scaffold was developed by our group. Positively charged polymer PDADMAC was used as a "bridge" to capture the negatively charged heparin and to adsorb to the collagen surfaces via an LBL self-assembly technique. The heparin can be covalently immobilized to collagenous scaffold by activating its carboxyl groups using carbodiimide. The immobilized heparin on DLSs can be MPA because of the many free carboxyl groups in heparin for /h-MPA. The immobilization of nitrite-depolymerized heparin can be covalent EPA on the DLSs as /h-EPA. Nitrous acid can make the heparin partially depolymerize and yield a terminal free-aldehyde functional group on the anhydro mannose unit 14 . The partially depolymerized heparin molecules were linked to the free amino groups on the scaffold via EPA using this terminal free-aldehyde functional group 14 . Heparin may have been sustained in /h-MPA and /h-EPA more than in /h-LBL because the heparin covalently immobilized on the scaffold interacted more strongly than the heparin electrostatically immobilized on the scaffold. The above results reveal that heparin can release slowly from the scaffolds, and that the scaffolds therefore possess excellent anticoagulation capacity (Fig. 4).
PF4 has been reported to play a vital role in platelet aggregation 24 , then, PF4 was used to detect platelet activation by biomedical materials 17 . Platelet adhesion on the scaffolds is of primary importance for blood compatibility. Before platelet adhesion, plasma proteins are quickly adsorbed on the scaffold surface, and the adsorbed proteins strongly affect subsequently blood-material surface interactions. Fibrinogen appears to be an important component of this adsorption process. This plasma protein is known to be capable of mediating platelet activation, aggregation and adhesion via direct interactions with platelet receptors such as GP IIb/IIIa 25 . The ability of surface-immobilized heparin on collagenous tissue to prevent thrombus formation through disturbed fibrinogen adsorption has been reported in the  12,22 . Furthermore, in the /h-EPA tissue, heparin molecules were covalently immobilized on the tissue surface like bristles (Fig. 1A-right). The fibrinogen molecules that were adsorbed on swinging bristles, rather than on the tissue surface, may then be easily washed away during the deionized water rinsing procedure prior to desorption. Field counts of scanning electron images showed greater adhesion of platelets on the surfaces of the nonheparinized tissues than on their heparinized counterparts (Fig. 6A).
The most direct test of the anti-thrombotic benefit of these techniques is perfusion of the whole organ to whole blood and quantification of thrombus formed in the whole organ. With or without heparin treated DLSs were transplanted into the porcine receipt with successful reperfusion of the scaffolds, which acquired a liver-like macroscopic appearance. Notably, no extravasation occurred either during the operative period, which suggests that under normal circumstances vessel watertight sealing depends mainly on the integrity of the scaffold's bare vasculature rather than the endothelium. Despite the pre-administration of anticoagulation prophylaxis in DLSs, the whole vascular network of control DLSs was occluded by massive clots within 20 min after perfusion. Heparinized DLSs remained blood flow throughout the whole intraoperative observation time of at least 1 hour. However, the blood flow rate decreased significantly from the inflow to the outflow of heparinized DLSs, due to the complete absence of hepatocytes and endothelium, massive blood entered and trapped in the original liver parenchyma and hepatic sinus areas. These results indicate that DLSs pretreated with heparin can decrease thrombus formation and increase the graft patency. However, engineering an intact, functional whole-liver graft will require recellularizing liver parenchyma with hepatocytes and covering the vascular bed with endothelial cells to restore the unique, functional characteristics of the hepatic microvascular circulation and to prevent coagulation 20,26 . Heparin treated DLSs did not compromise their cytocompatibility, which could be reendothelialized with HUVECs and recellularized with rat primary hepatocytes as non-treated DLSs 9,19,26 .
Prevention of blood coagulation is difficult to achieve, and failure to do so hinders long-term in vivo testing of recellularized liver grafts. This concern may be overcome by immobilizing heparin on the DLSs, as shown in this study. The three types of heparinized DLSs showed different efficiency of heparin immobilization. The heparinized scaffolds showed improved anticoagulation and cytocompatibility compared to the control scaffold both in vitro and in vivo test. The /h-EPA scaffold showed a dramatic improvement in hemocompatibility compared to the other scaffolds. Whole-organ tissue engineering is complex and involves many technically challenging steps that must be performed successfully in series. Although heparinization of the grafts is a major improvement, the efficacy and safety of this technique still require further longtime in vivo testing.

Conclusion
This report provides an efficient method for immobilization of heparin on DLSs, which could provide a solution to the problem of anticoagulation ability in liver engineering. The data presented here show some advantages in the use of the EPA heparinization method to improve blood and cell compatibility of DLSs. These hemocompatibility improvement studies are a first step toward generating a transplantable functional liver graft using DLSs.

Materials and methods
Animals. 24 livers were isolated from male Bama miniature pigs weighing 10-15 kg for perfusion decellularization. 12 Bama miniature pigs weighing 23-25 kg were employed as DLSs transplantation receipts. A 220 g male Sprague-Dawley rat was used for hepatocyte isolation. All animals were obtained from the Animal Experiment Center of Sichuan University (Chengdu, China). All animal experiments were performed in accordance with the Animal Welfare Act and approved by the institutional animal care and use committee at Sichuan University.
Perfusion decellularization of porcine liver. The surgeries were performed under ketamine (6 mg/ kg body weight, administered IP, Kelun, Chengdu, China) and xylazine (10 mg/kg IP, Kelun) anesthesia. Under deep anesthesia, a laparotomy was performed and the liver was exposed. After systemic heparinization through the inferior vena cava, the hepatogastric ligament was carefully dissected. The proximal PV was catheterized. The hepatic artery and common bile duct were ligated and transected. All perihepatic ligaments were severed. Simultaneously, the liver was slowly perfused with 2 L deionized water containing 0.1% EDTA (Kelun) through a cannula in the PV, and the SHIVC was transected, allowing outflow of the perfusate. Following blanching, the liver was stored at − 80 °C overnight. The liver was perfused with 1% Triton X-100 (Amresco, Solon, OH, USA) for 3 h and then by 1% SDS (Promega, San Luis Obispo, CA, USA) in deionized water at a rate of 200 mL/min for 6 h after thawing. This was followed by 3 h of perfusion with 1% Triton X-100 to remove residual SDS. Subsequently, the liver was washed with 20 L of distilled water to remove residual detergent, followed by infusion of 40 L of phosphate-buffered saline (PBS) at 200 mL/min.

Polymerase chain reaction (PCR) analysis. Genomic DNA content in native and decellularized
livers was normalized to the initial weight of the sample. The sequences of primers were listed in Table 2. PCR products were loaded on 2% agarose gel and visualized by the GoldView nucleic acid staining solution.

Gene
Abbr. Perfusion heparin treatment. LBL technique. The /h-LBL scaffolds were prepared using the LBL self-assembly technique for heparin immobilization, as previously described 8 . The polyelectrolyte polydiallyldimethylammonium chloride (PDADMAC, Mw = 100-200 kDa, positively charged, Sigma, USA), which is oppositely charged to heparin (Kelong, Chengdu, China), was used. Typical adsorption conditions were used to form multiple layers on the surfaces of DLS. Following a 30-min perfusion of PDADMAC or heparin (1 g/L of PDADMAC or 2, 1, 0.5 or 0.25 g/L heparin in PBS) for each step at 100 mL/min through PV. After each perfusion, the DLS were washed via PBS perfusion for 10 min to eliminate residual polyelectrolytes. The eight-bilayer (PDADMAC/heparin) 8  Each heparin immobilized DLSs were sampled from 12 sites showed in Fig. 3A and cut into 300-μ m thick sections with a diameter of 1 cm and used in the following in vitro experiments.
Heparin contents test. The quantity of heparin immobilized on the three DLSs was determined as Smith et al. 28 described.
Heparin release rate test. To determine heparin release, the three different types of heparinized DLSs were each fixed at the bottom a well in a 24-well plate and immersed in PBS solution (2.5 mL, pH 7.4) for 7 days. To determine the accumulative release, solution (2.5 mL) was periodically collected from each well and replaced with fresh solution, and the heparin content released each day was recorded. To determine the sustained release, the solution was collected without replacement until the final time point. The heparin concentration in PBS solution was determined as described 28 .

In vitro coagulation time tests.
To evaluate the antithrombogenicity of the heparinized DLSs, PT, APTT and TT were measured by a semiautomatic blood coagulation analyzer CA-50 (Sysmex Corporation, Kobe, Japan) as described 29,30 . Platelet adhesion. The experimental procedure employed in the platelet adhesion study for DLSs was reported by Ko et al. 31 . Scanning electron microscopy (SEM, JSM-7500F, JEOL, Japan) was performed to count the total number of adhered platelets on DLSs 22 . Platelet activation. Platelet activation (PF4) was evaluated via commercial enzyme-linked immunosorbent assay (ELISA) kit (E-(Hu)(11)-04605, Yanyu, Shanghai, China). Whole blood was incubated with the heparinized or control DLSs for 2 h and centrifuged at 1000 × g (4 °C) for 15 min to obtain plasma. Then detections were conducted according to the manufacturer's instruction.
In vivo implantation of heparinized DLSs. To examine the anti-thrombotic ability of the heparinized DLSs under whole blood perfusion, we performed an auxiliary transplantation of the DLS into a recipient porcine model. Considering the blood volume of recipient, the right lobe and left lobe were removed, the median lobe was preserved with PV and SHIVC. Before implanted, 100 mL saline with 25 IU/mL heparin was injected into DLS through PV. The DLSs was implanted in the infrahepatic space by using the recipient left renal vein (RV) and splenic vein (SV) as an inflow and outflow, respectively. After positioning of a clamp on the recipient's RV and SV, scaffold's PV and SHIVC were anastomosed end to end using silicone cuff, respectively. After reperfusion, scaffolds were monitored for 1 hour before retrieved for pathology studies.
Cell culture. Hepatocytes were isolated from a 220 g male Sprague-Dawley rat by a two-step perfusion method as our previously described 32 . All harvests yielded hepatocytes with viability exceeding 95% by trypan-blue dye exclusion. HUVECs were cultured in DMEM medium supplemented with 10% fetal bovine serum (FBS, Hyclone, America). To evaluate the cytocompatibility of the scaffolds, primary hepatocytes or HUVECs (2.5 × 10 4 cells/cm 2 ) were cultured inside the control, /h-LBL, /h-MPA, and /h-EPA scaffolds (n = 3) in 24-well culture plates, respectively. The viable cells cultured on each test specimen were examined by H&E staining at 3 days after cell seeding. The proliferation of the cells was evaluated using the PCNA (1:2000, rabbit polyclonal IgG, ab18197; abcam, Shanghai, China) immunohistochemistry staining.
Statistical analysis. The coagulation time tests of the heparin-treated and non-heparin-treated scaffolds were compared using a χ 2 test. For comparison of heparin contents and cell number on the scaffolds, a one-way analysis of variance (ANOVA), followed by a Student-Newman-Keuls test for variability, was performed. The data are presented as mean ± SD, and p < 0.05 was considered statistically significant.