1 Introduction

Organ decellularization became an intensively studied topic with a great potential in tissue engineering [1]. An ideal decellularization process enables a gentle removal of all cellular components while preserving an intact extracellular matrix (ECM) containing molecular cues like growth factors and cytokines on its surface. Such ECM or so-called scaffold has the potential to be repopulated with cells to form an artificial tissue or organ. Using the cells of the potential organ recipient should make the development of immunologically tolerated organs possible [2]. Once this strategy is successful, it could solve the problem of the growing organ shortage in transplantation medicine.

Decellularization has been successfully performed and studied in the case of many organs or tissues like lungs, heart, kidneys, liver, bone, cartilage, vessels, small intestinal submucosa, and many others [1, 3,4,5,6,7,8,9,10,11]. Some of these organs were also partially recellularized in vitro and implanted in animals [6, 7, 12]. However, achieving proper function of such organs and their lifespan after implantation is still a big challenge. The ideal decellularization method, the source of the cells, their sufficient amount, and their optimal delivery into the scaffold are some of the questions that need to be answered before getting closer to the preparation of functioning artificial organs [13, 14].

The creation of a functional vascular system is another critical step for further progress in artificial organ engineering based on native ECM. Therefore it is essential to understand the possibilities of vessel decellularization and recellularization in detail [12, 15]. Moreover, decellularized vessels themselves are a potential alternative to the grafts currently used in cardiovascular surgery [16]. As biological tissues, they should pose a lower risk of infection complications compared to synthetic prostheses [17,18,19] and their immunogenicity should be minimal compared to allografts [20, 21]. Decellularized matrix can be also used for bioink formulations for construction of vessel grafts through further processing (e.g. bioprinting or electrospinning) where other functionalization such as formation of an antithrombotic layer is possible [22, 23].

The ECM obtained by optimal decellularization preserves bio-inductive properties similar to native tissues. These are the capability of chemotaxis, cell attachment, proliferation, and activation [13, 24, 25]. These unique biomaterial properties together with the structural similarity to the target tissue enable, to a certain level, spontaneous in vivo recellularization, which has been observed also in the case of vascular structures [26,27,28]. A better understanding of spontaneous repopulation of ECM could be beneficial even in whole organ engineering. It could show which type of cells tends to recellularize particular areas of the ECM.

A number of experiments were already performed with decellularized arteries or veins implanted in the arterial system [28,29,30,31,32]. Decellularized arteries implanted in the arterial system showed successful endothelization in the majority of the cases and some of the grafts were even recellularized with smooth muscle cells (SMC) [28,29,30]. However, microcalcifications in the arterial wall or some degree of intimal hyperplasia were also noted [29, 33, 34]. In the case of decellularized veins implanted in the arterial system, only incomplete endothelization was achieved, which could be caused by different architecture and a biomechanical mismatch between the venous and arterial walls [32, 35].

Very limited experience has been gained with decellularized veins implanted in the venous system [36,37,38]. In this setting, the pressure load on the venous wall by pulsatile arterial blood flow is eliminated, which could facilitate spontaneous recellularization. The lack of such data is probably due to the lesser imperative for finding alternative grafts for venous replacement as vein reconstruction is less often needed in clinical medicine. However, for the purpose of artificial organ engineering, these findings are essential and could enable further progress. Therefore, we designed an animal model to understand the behaviour of decellularized veins in vivo under physiologic conditions. The aim was to investigate the capability of large-diameter decellularized venous grafts to be spontaneously recellularized and to reveal the potential changes in their micro-architecture after implantation. To precisely evaluate these parameters together with the immunogenicity and thrombogenicity, the decellularized vena cava grafts were compared with standard cryopreserved allografts in a porcine model.

2 Methods

2.1 Experimental animals

In total 40 pigs were used for the experiment (healthy both male and female Prestice Black-Pied pigs, 26.5 to 38.5 kg). Inferior vena cava (IVC) grafts were harvested from 20 donor pigs. Twelve of these grafts were decellularized and then orthotopically implanted in the IVC of the recipient animals from the experimental group (n = 12), whereas 8 grafts were cryopreserved and then implanted in the recipient animals from the control group (n = 8). The concept of the experiment is depicted in Fig. 1. The recipient animals were then observed for 28 days after implantation. All animals were acclimatized before the experiment, they were fed twice a day and had unlimited access to water. All the surgical procedures were performed under general anesthesia using sterile material.

Fig. 1
figure 1

Scheme of the experimental design. The preparation of the decellularized vena cava, its implantation and sampling (top part) as well as parallel procedures for control cryopreserved vena cava implants (bottom part). Created with BioRender.com

Full size image

The protocol of in vivo experiment was approved and controlled by the Ministry of Education, Youth and Sports of the Czech Republic (project code: MSMT-18870/2020–3). All the procedures involving animals were performed in compliance with the law of the Czech Republic, which is compatible with the legislation of the European Union (EU Directive 2010/63/EU for animal experiments).

2.2 Graft harvesting

The donor animals (n = 20) received premedication by subcutaneous administration of ketamine (10 mg/kg, Narkamon—Spofa, a.s., Prague, Czech Republic), azaperone (5 mg/kg, Stresnil—Jannssen Pharmaceutica NV, Beerse, Belgium), and atropine (1 mg/animal, Atropin Biotika—Hoechst Biotika, Martin, Slovak Republic). After cannulation of a peripheral vein, general anesthesia was induced by intravenous administration of propofol (1% mixture, 5–10 mg/kg/h Propofol, Fresenius Kabi, Halden, Norway), and the animals were mechanically ventilated. Nalbuphine (10 mg/animal, before the operation and repeated every 2 h during anesthesia, Nalbuphin Orpha, Chiesi CZ, Prague, Czech Republic) was used for analgesia and 0.6 mg of amoxicillin-clavulanate (Augmentin, GlaxoSmithKline Slovakia, Bratislava, Slovak Republic) was administered intravenously as prophylaxis. The IVC was dissected from the level of confluence of iliac veins to the lover margin of the liver. All lumbar veins and both renal veins were ligated and 300 international units (IU) of heparin per kg (Heparin, Zentiva, Prague, Czech Republic) were administered intravenously 5 min before IVC explantation. The explanted vein was then flushed with saline to remove all blood. In the case of grafts for decellularization, the IVC was placed in a 10% dimethyl sulfoxide (DMSO) solution and gradually frozen to − 80 °C. The allografts for animals from the control group were placed in 1:1 mixture of 6% Voluven (Fresenius Kabi, Halden, Norway) and 10% DMSO mixed in saline and then also gradually frozen to − 80 °C.

2.3 Decellularization

Decellularization of the IVC grafts was done by perfusion with 1% Triton X-100 and 1% sodium dodecyl sulfate (SDS) at room temperature. The grafts (n = 12) were thawed at 4 °C and then connected to the perfusion tube. The outflow of the vein was secured with a Luer lock adapter and the graft was placed in a 0.5 L glass bottle with inflow and outflow tubes going trough the bottle cap. The following runs of perfusion (50 ml/min) enabled decellularization of the graft: 1 run of saline for 10 min; 2 runs of Triton X-100, each for 60 min; 2 runs of SDS, each for 60 min; and 4 runs of saline, each for 15 min. The solutions were changed before each run. A photograph of the system running is in Fig. 3A including the detail of the vein connection (Fig. 3B).

DNA residue content was used to determine the decellularization efficacy. Briefly, both native and decellularized samples of IVC weighing 40–50 mg were lyophilized to obtain the dry weight used as a reference to indicate DNA residues as ng/mg dry weight. By drying, we eliminated the differences given by the water content from each sample. Then, tissues were powdered using a mortar and pestle in liquid nitrogen and homogenized in a QIAshredder column (Qiagen, Dusseldorf, Germany). Then, DNA extraction was performed with RNeasy mini kit following the manufacturer’s guidelines (Qiagen, Dusseldorf, Germany) and Qubit reader was used to obtain the final DNA amount.

2.4 Graft implantation

Recipient animals were anesthetized using the same protocol as in the case of donors. For the prophylaxis, 0.6 mg of amoxicillin-clavulanate was administered intravenously before the operation and repeated 2 h later. Firstly, port-a-cath was implanted to optimize blood sampling and postoperative care. PolyFlow polyurethane catheter was implanted in the right external jugular vein and connected to the ProPort Plastic Venous Access System (Deltec, Smiths Medical International, Ltd., UK) which was placed under the skin on the right side of the neck. Then the abdominal cavity was entered by middle laparotomy, the retroperitoneum was opened by incision of the peritoneum on the right side of the IVC, and the infrarenal part of the IVC was dissected in sufficient length. Several lumbar veins were ligated and transected and attention was paid not to injure the lymphatic vessels running close to the IVC. Five minutes after intravenous administration of heparin (100 IU/kg), the infrarenal IVC was clamped and a 1 to 1.5 cm long segment was resected and replaced with a 1.5 to 2 cm long graft. The decellularized graft was used for experimental animals (n = 12) and the cryopreserved allograft for control animals (n = 8). Both anastomoses were sewn with continuous 5–0 Prolene suture. After declamping the IVC, the retroperitoneum was left open to prevent fluid collection formation along the reconstruction. The abdominal cavity was closed with interrupted 0 nonabsorbable braided sutures, and 2–0 monofilament nonabsorbable suture was used for skin closure.

2.5 Postoperative follow-up

During the early postoperative period, the animals received infusion of crystalloid solutions (Hartman’s and 10% glucose) via the port-a-cath but the access to water was not limited. The dose of feed was gradually increased to the standard dose during the first days. All the animals received 40 mg of pantoprazole every day during the whole postoperative period. No anticoagulants or antithrombotic treatment were used after the operation. The animals were checked every morning to evaluate their overall status.

Blood samples were taken before and after the operation, and then on the 1st, 3rd, 7th, 14th, 21st, and 28th POD. Plasmatic levels of the following markers were assessed in each sample: aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase (ALP), gamma-glutamyl transferase (GGT), bilirubin, albumin, urea, and creatinine.

Computed tomography (CT) with 20 ml of intravenously administered iodinated contrast agent (Iomeron 350, Bracco, Milan, Italy) with rate of 1.8 ml/s was performed before the operation and then on the 1st, 3rd, 7th, 14th, 21st, and 28th POD.

The abdominal cavity of all the animals was explored under general anesthesia on the 28th POD and the segment of the IVC containing the implanted graft was resected.

2.6 Samples for histology

Five groups of samples were used for histological analysis: (1) native veins (resected segment of IVC of the recipient, n = 13), (2) decellularized veins before implantation (short segment taken from the graft, n = 12), (3) cryopreserved veins before implantation (short segment taken from the graft, n = 8), (4) decellularized veins after implantation (graft together with both anastomoses and parts of recipient IVC explanted 28 days after implantation, proximal part n = 7, distal part n = 7), and (5) cryopreserved veins after implantation (graft together with both anastomoses and parts of recipient IVC explanted 28 days after implantation, proximal part n = 7, distal part n = 7). The number of examined grafts after implantation was lower than the number of grafts before implantation because only the nonoccluded grafts could have been evaluated using the same methodology.

2.7 Histological processing

All venous samples were fixed in 10% neutral-buffered formalin solution. Both decellularized and cryopreserved grafts after implantation were transected in the middle between the proximal and distal anastomosis (Fig. 2A). The samples were then embedded in paraffin blocks following standard histological methods and cut into 5 μm-thick sections (Leica RM2255 microtome, Leica Biosystems GmbH, Nussloch, Germany). Transversal sections were obtained from these samples for quantitative histology. Then, the paraffin blocks with veins after implantation were dissolved in order to cut the sample with a longitudinally oriented section plane. The longitudinal sections were then histologically processed for qualitative histological analysis (Fig. 2B). All sections were deparaffinized, rehydrated and stained using the combination of four histological staining methods and four immunohistochemical reactions (Table 1).

Fig. 2
figure 2

Histological processing of graft samples explanted at the end of the experiment. Decellularized and cryopreserved grafts were transected in the middle between the proximal and distal anastomosis for quantitative histological analysis (A), longitudinal transection was then performed for qualitative histological evaluation (B)

Full size image
Table 1 Histological staining methods and quantitative parameters
Full size table

Samples of decellularized veins before implantation were not stained with four immunohistochemical reactions (anti-alpha smooth actin, anti-von Willebrand factor, anti-MAC387 and anti-Ki67) due to the absence of cells.

2.8 Microphotographs for stereological analysis

Microphotographs were taken using systematic uniform sampling of the venous wall (Supplementary Fig. 1, Table 1) [39]. The detailed location of these microphotographs is shown in Supplementary Fig. 1. The total number of microphotographs taken and quantified was 2,656.

2.9 Stereological quantification

The design of the quantitative analysis was based on previously published studies on various vessels morphology [40,41,42]. The area fractions of elastin, collagen and smooth muscle actin were quantified using the stereological point grid and Cavalieri principle [43]. The fractions were estimated as the ratio of the evaluated microscopic structures to the reference area. The settings of the used point grid were in agreement with the standards of stereological methods [44]. The two-dimensional densities of nuclear profiles, vascular profiles and MAC387-positive cells were evaluated using the unbiased counting frame [40, 45]. These densities were estimated as the number of the evaluated structures per reference area. The proliferation index was evaluated as the ratio of Ki67-positive nuclei to all cell nuclei in the reference area. At least 300 nuclei per case were calculated for the proliferation index [46]. Intima-media thickness of veins was measured using four linear probes. Stereological quantifications were performed using the Ellipse software (ViDito, Košice, Slovak Republic). For details see Supplementary Fig. 2.

2.10 Statistics

Means and sample standard deviations were used to represent the measured data. Primary analysis of the quantitative histological and biochemistry data was performed with analysis of variance (ANOVA) for repeated measurements, thus respecting the dependency of observations belonging to each animal. In order to maximize statistical power, post-hoc analysis was not performed with a complete scheme of all pairwise comparisons, but only biologically relevant comparisons were made using the t-test (either paired or two-sample) with subsequent Bonferroni correction applied to the whole family of test for each variable. In particular, histological variables were compared between all timepoint pairs within each group (3 tests for each group, 6 tests in total) and between the groups at each time point (3 tests) and the Bonferroni-corrected significance level was set to αC = 0.05 / 9 = 0.0056. The exception here was vasa vasorum, for which data was not available for the decellularized graft before implantation. The repeated measures ANOVA was thus performed using only the ‘native’ and ‘28 days after implantation’ timepoints and the post-hoc procedure comprised only 6 tests and αC = 0.05 / 6 = 0.0083. For biochemistry variables, only comparisons between groups (for each time point) were made, resulting in 8 tests and αC = 0.05 / 8 = 0.00625 for each variable. The Bonferroni correction was performed at the level of significance threshold α and if shown, the original p-values are presented. All reported p-values are two-tailed and the general level of statistical significance was set at α = 0.05. Statistical analysis was performed in Statistica (ver. 12 Cz, TIBCO Software Inc., Palo Alto, CA, USA).

3 Results

3.1 Vein decellularization

Successful decellularization process of the IVC grafts was confirmed by DNA analysis as well as by histology. DNA residues were found to be lower than 50 ng/mg dry tissue for the decellularized vein and significantly lower compared to the native vein (mean 10.63 vs 153.8, p = 0.014, unpaired t-test) (Fig. 3D). Moreover, any cells or nuclei were not found in the wall of the decellularized grafts based on evaluation of Gill’s hematoxylin stained sections (Fig. 6C), making these native scaffolds suitable for the following orthotopic implantation.

Fig. 3
figure 3

Perfusion decellularization of IVC graft using peristaltic pump (A), detail of IVC graft at the beginning of the decellularization process (B) and at the end of decellularization process (after 310 min of perfusion) (C); DNA content estimated for native veins as well as decellularized scaffolds shows significant reduction in case of decellularized tissue going below the generally accepted limit for decellularized tissue (50 ng/mg dry tissue, dashed line) (D)

Full size image

3.2 Survival of implanted animals

All the animals from the control group (n = 8) survived the whole period of 28 days after graft implantation. In the case of the experimental group, one out of 12 animals died on the 14th POD. The death occurred a few hours after the CT examination which was done in analgosedation with ketamine and azaperone. Results of the autopsy, CT, and biochemistry did not show any pathology related to the surgical procedure. The IVC graft was patent without any signs of thrombosis.

3.3 Graft thrombosis and CT results

The thrombosis of decellularized IVC graft occurred in 4 out of 12 experimental animals (33.3%). The rate of cryopreserved IVC graft thrombosis was lower, with only 1 animal out of 8 in the control group (12.5%). All observed thromboses were confirmed after explantation of the grafts at the end of the experiment. Additionally, the CT examinations enabled the detection of the time of thrombosis development. The graft thrombosis occurred during the first postoperative week in 3 animals. Exactly, it was observed on CT on the 7th POD in 2 animals and on the 1st POD in 1 animal (all animals with decellularized graft). Later incidence of graft thrombosis was observed on CT on the 14th POD in 2 animals (one with decellularized and one with cryopreserved graft). Two animals with decellularized graft thrombosis had a fluid collection compressing the IVC at the area of graft implantation. Such fluid collection was not apparent on CT scans of the other 3 animals with graft thrombosis. In total, some fluid collection near the implanted graft was present in 5 animals with decellularized grafts and in 3 animals with cryopreserved grafts. Only 2 of these collections did not spontaneously resolve and were accompanied by graft thrombosis in animals with decellularized grafts. Examples of CT results are presented in Fig. 4.

Fig. 4
figure 4

CT scans of the retroperitoneum. An asterisk indicates the IVC. Normal IVC—preoperative CT (A), the decellularized graft implanted in infrarenal vena cava is patent (orange arrow)—CT on 7th POD (B), example of rare occlusion of the decellularized graft (green arrow) compressed by large fluid collection (red triangle)—different animal, CT on 7th POD, graft thrombosis was confirmed at the end of experiment (C)

Full size image

To evaluate the graft diameter and potential stenotic complications preoperative CT was compared with the CT performed on the 28th POD. Only animals without graft thrombosis were included (experimental group with decellularized grafts n = 7, control group with cryopreserved grafts n = 7). Data from the other time points especially during the first postoperative week were influenced by perioperative changes and fluid collections affecting the IVC diameter. As most of the collections were resolved during the postoperative period, the data from the 28th POD were the most reliable. The diameter of IVC before graft implantation was comparable between the experimental and control groups (mean 11.7 mm and 11.2 mm respectively). Slight narrowing developed due to the graft implantation but the graft diameter 28 days after implantation was comparable in experimental and control animals (mean 8.1 mm and 8.6 mm respectively). The tendency to narrow was not significantly different between experimental and control animals (difference between preoperative IVC diameter and graft diameter on 28th POD—mean 3.6 mm and 2.7 mm respectively). See the data in Supplementary Fig. 3.

3.4 Macroscopic appearance of the grafts

The decellularized grafts appeared to be of slightly softer consistency compared to the cryopreserved grafts. However, the handling and sewing of anastomoses were not really different when comparing decellularized and cryopreserved grafts. Four weeks after implantation, there were adhesions in the retroperitoneum in the area of IVC reconstruction. Just a minimum of adhesions were found between the small intestine and the abdominal wall in a few animals from both groups. In the case of the animals with graft thrombosis, there were several collateral veins in the retroperitoneum which complicated the graft extraction at the end of the experiment. These collaterals were not present at the time of graft implantation. The explanted grafts were cut longitudinally and inspected from the luminal side. The inner side of the implanted grafts was of similar character as the native IVC. There were no macroscopically apparent differences between decellularized and cryopreserved grafts 28 days after implantation (Fig. 5).

Fig. 5
figure 5

Decellularized and cryopreserved grafts right after implantation (A, B), 28 days after implantation—in situ (C, D), and 28 days after implantation—view on the luminal side after the longitudinal cut (E, F). The arrows show the direction of the blood flow

Full size image

3.5 Qualitative histology analysis

The histology analysis also confirmed successful decellularization process in all 12 grafts. The wall of these grafts prior to the implantation did not contain any cells (Fig. 6). In the case of grafts after implantation, the structure of individual layers of the venous wall was relatively preserved in all the samples. Only in the case of cryopreserved grafts 28 days after implantation, the structure was altered by extensive lymphocyte infiltration. On the contrary, decellularized grafts contained only mixed inflammatory infiltrates around the suture material at the area of both anastomoses. All the decellularized as well as cryopreserved grafts were completely endothelized 28 days after implantation (Fig. 7B). Both types of grafts also contained vasa vasorum with endothelized lumen and smooth muscle actin (SMA) positive cells in tunica media. A coherent layer of SMA-positive cells was found only in the wall of the native vein. In the decellularized grafts after implantation, SMA-positive cells were distributed in small groups or separately in the venous wall. In the cryopreserved grafts, SMA-positive cells were located among lymphocyte infiltrates (Fig. 7A). Proliferating cells were located mainly in the cryopreserved grafts in the form of lymphocyte infiltrates. In the decellularized grafts, the proliferating cells were present to a lesser extent only around the suture material or in small groups in the graft wall (Fig. 7C).

Fig. 6
figure 6

Histological comparison of the structure of native (left), decellularized (middle) and spontaneously recellularized (28 days after implantation) IVC (right). Verhoeff’s green trichrome staining showing elastic fibres in black (A) and picrosirius red staining for collagen fibres (B) show signal in all, native, decellularized and recellularized samples; Gill’s hematoxylin staining detecting purple nuclei depicts nuclei in native and recellularized vein while decellularized sample is nucleus-free (C). Scale bars represent 50 μm in all images

Full size image
Fig. 7
figure 7

Microscopic appearance of the decellularized and cryopreserved grafts 28 days after implantation (longitudinal section). In the native veins, smooth muscle cells are arranged in a layer. Decellularized grafts after implantation have smooth muscle cells in groups, and cryopreserved grafts have individual smooth muscle cells scattered among lymphocyte infiltration (A). The newly formed endothelium on the grafts (B). Cryopreserved grafts have more proliferating cells than decellularized grafts (C). Scale bars 100 μm (A–C)

Full size image

3.6 Quantitative histology analysis

After qualitative evaluation, quantitative assessment was performed for ECM components elastin and collagen content as well as SMA-positive cells, vessel wall thickness, nuclear profiles, cell proliferation, presence of vasa vasorum and infiltration of macrophages. The area fraction of elastin showed a significant decrease after graft decellularization or cryopreservation and after graft implantation. The proportion of elastin was not significantly different between the groups at any time point (Fig. 8A).

Fig. 8
figure 8

Visualization of the data from quantitative histology. See comments in the text. Since the data for vasa vasorum (G) was not available for the decellularized graft before implantation, the corresponding time point was excluded from the ANOVA also for the cryopreserved samples (shown in grey). It was, however, included in the manual post-hoc procedure

Full size image

In the case of collagen, there was a significant decrease in the area fraction after the implantation of the decellularized grafts. No significant differences between the groups were observed (Fig. 8B).

The fraction of SMA-positive cells showed a tendency to decrease in cryopreserved grafts during the experiment. The decellularized grafts lost the cells during the decellularization process, but the fraction of SMA-positive cells after implantation reached a similar value as in the native vein (Fig. 8C).

The thickness of the graft wall did not change significantly during the experiment for any of the groups and was not significantly different between the groups at any time point (Fig. 8D).

In the case of decellularized grafts, the number of nuclei profiles after implantation was comparable with the number found in the native vein. This is proof of cell migration into the graft because the graft after decellularization contained no nuclei. In the case of the cryopreserved grafts, the number of nuclei profiles was significantly higher after implantation, showing a more pronounced lymphocyte infiltration in these grafts (Fig. 8E).

Proliferation index was similar in decellularized grafts after implantation compared to the native vein. Contrarily, the value was multiple times higher in cryopreserved grafts after implantation compared to the native vein (Fig. 8F).

The density of vasa vasorum profiles increased after implantation in both types of grafts compared to the number in the native vein. However, this increase was more pronounced (yet less statistically robust) in cryopreserved grafts, which had a higher density of vasa vasorum after implantation compared to the decellularized grafts (Fig. 8G). Therefore, decellularized grafts after implantation seemed to be less different from the native tissue compared to the cryopreserved grafts.

Besides the expected drop in the decellularized grafts before implantation, the fraction of MAC387-positive cells representing mainly macrophages did not show significant differences with respect to time or graft type (Fig. 8H).

To reveal any potential differences in the graft wall depending on the blood flow, sections from the proximal (caudal), as well as the distal (cranial) part of the graft, were analysed. Decellularized grafts had a greater fraction of total collagen at the proximal part compared to the distal part (mean 0.236 vs 0.169, p = 0.037, paired t-test). Cryopreserved grafts showed a higher amount of MAC 387 positive cells in the proximal part rather than in the distal part (mean 53.3 vs 25.5, p = 0.045, paired t-test). The other histological parameters did not differ between the proximal and distal area of the graft.

Complete morphologic data are summarized in Supplementary Table 1 and Supplementary Fig. 4.

3.7 Biochemistry

Only one animal from the experimental group which died on the 14th POD was excluded from the biochemistry analysis because of incomplete data. However, the available data of this animal were of similar values compared to surviving animals. There was an elevation of bilirubin on the 1st POD in animals from the control group which was not seen in animals from the experimental group. Similarly, the control animals had elevated level of ALP on the 1st POD and higher levels of ALP compared to experimental animals during the 1st postoperative week (Fig. 9). All the other biochemical parameters (AST, ALT, GGT, albumin, urea, and creatinine) did not show significant differences between the groups and stayed within normal range for the postoperative period [47].

Fig. 9
figure 9

Results of the biochemistry analysis. Only bilirubin and ALP exhibited differences between groups. Differences between individual time points were not tested in the post-hoc procedure in order to conserve statistical power for the comparison of groups

Full size image

4 Discussion

There is very limited evidence of the behaviour of decellularized veins implanted in vivo under physiologic conditions. Experiments with decellularized vessels are usually designed to evaluate potential alternatives to clinically used grafts for vascular reconstruction. Therefore, the venous grafts are mostly implanted in the arterial system [32, 48, 49]. In such a setting, the venous wall is excessively irritated due to arterial pressure which could alter the spontaneous recellularization. In the case of implantation in the venous system, the wall of the decellularized graft is usually modified prior to the implantation to prevent thrombosis (e.g. heparinization, re-endothelization) [36, 50]. Such modifications could also affect the in vivo recellularization process. To study the capacity of spontaneous repopulation of decellularized venous wall, we designed an experiment with orthotopic implantation of decellularized IVC graft without any prior anti-thrombotic modification. To date, mainly graft patency and re-endothelization were evaluated in the published studies [36, 37]. This is the first time to the author's knowledge that quantitative histological analysis of the main extracellular components forming the venous wall is performed in this setting. To evaluate the immunogenicity, we compared the decellularized grafts with cellular cryopreserved allografts.

4.1 Successful decellularization

Normal values of DNA content for porcine vein are higher than 100 ng/mg tissue, reduced to around 5ng/mg tissue in case of decellularization. It is commonly defined in literature that good decellularization leads to DNA content below 50 ng/mg tissue [51]. Our results confirm a good decellularization of the porcine IVC.

4.2 In vivo recellularization

The decellularized grafts showed the capability of spontaneous recellularization by endothelial as well as smooth muscle cells 4 weeks after implantation. The amount of SMA was not different from native IVC. The cells repopulated the appropriate locations of the venous wall including endothelization of vasa vasorum.

The in vivo repopulation of decellularized vascular structures has been documented in case of arteries implanted in the arterial system. The endothelization is usually successful, but the repopulation of the rest of the vascular wall is not usually studied in such detail [28, 34]. Decellularized ovine carotid arteries showed just limited spontaneous repopulation which was attributed to dense collagen bundles obstructing the cell migration [34].

Results corresponding to our findings concerning successful endothelization and smooth muscle cells repopulation were acquired in experiment with decellularized porcine IVC grafts that were perfused with the recipient's blood prior to implantation. The perfusion took one week, and it was performed with the intent to add the recipient’s autologous components to prevent rejection and facilitate in vivo recellularization. In the proof of concept study, there were 6 pigs included with short-term survival (3 days to 5 weeks) [37]. The following study evaluates similar grafts explanted after long-term survival of 10 recipient pigs (6 to 13 months) [38]. Compared to these studies we proved the capability of spontaneous recellularization without the need for preconditioning with the recipient's blood, which is technically demanding and increases the costs. The decellularized venous graft itself has sufficient functionality to attract specific cells in the appropriate locations including tunica media and vasa vasorum.

4.3 Histology of ECM

To reveal the potential damage to the vascular wall during decellularization we evaluated the amount of collagen and elastin as the main extracellular proteins. Moreover, changes in the amount of collagen and elastin have been documented in allogeneic venous grafts after implantation [52,53,54]. Therefore, to properly compare the performance of decellularized grafts with allografts, the quantitative analysis of collagen and elastin content was also performed 28 days after implantation. The amount of collagen and elastin was lower in decellularized as well as cryopreserved grafts after implantation when compared to native IVC but not significantly different between both graft types in any of the time points. The study of Kahle et al. showed an increase of collagen and reduction of elastin/collagen ratio accompanying venous graft disease [52]. They examined cellular allografts implanted in the arterial system. In our study, there was no increase (but rather a decrease) after implantation. Increase of collagen content in the venous wall after implantation in the arterial system was documented also by other studies [55, 56]. We hypothesize that these changes were prevented by the implantation of the venous grafts in the venous system, thus eliminating irritation by pulsatile arterial pressure. The decrease of elastin proportion, which was observed in the wall of both decellularized and cryopreserved grafts, corresponds to our previous findings. In the experiment with portal vein reconstruction using allografts originating in the portal or caval venous system, we detected a similar decrease in elastin proportion 4 weeks after graft implantation [57].

4.4 Immunogenicity

The cellular cryopreserved allografts showed major lymphocyte infiltration. As the lymphocytes and predominantly T-lymphocytes are known to be responsible for the immune response to implanted vein allografts [58, 59] this finding documents the immunogenicity of cryopreserved grafts in our study. Contrarily the decellularized grafts did not contain any lymphocyte infiltration out of the anastomotic area which proves substantially lover immunogenicity of these grafts. This finding is supported by a significantly higher proliferation index in the wall of cryopreserved grafts 28 days after implantation since the proliferating cells were present mainly in the lymphocyte infiltrates.

The presence of proliferating cells at the anastomotic area of the decellularized grafts corresponds with the finding of Osterberg et al. and is apparently caused by non-absorbable suture material [38].

The MAC387 positive cells represent mainly macrophages which are responsible for non-immune inflammation as crucial regulators of tissue healing [60]. There was no difference in the amount of MAC387-positive cells between animals with cryopreserved or decellularized grafts.

4.5 Intimal hyperplasia

Bai et al. have studied decellularized grafts of the human great saphenous vein used for venoplasty of rat IVC [36]. Their decellularized grafts were conjugated with hyaluronic acid and heparin prior to the implantation and compared with only decellularized non-coated patches of the same origin. They demonstrated that these coated decellularized patches had significantly thinner neointima compared to the uncoated patches. The intimal hyperplasia was also documented in their previous studies with pericardium patches [61, 62]. There were no signs of intimal hyperplasia in our experiments. As one of the main factors responsible for intimal hyperplasia is increased or decreased shear stress on the vascular wall [63], we conclude that intimal hyperplasia could be prevented by orthotopic implantation of decellularised venous grafts. In such situation, the physiologic perfusion is maintained and intimal hyperplasia does not occur even without any additional coating of decellularised grafts.

4.6 Thrombosis

The rate of thrombosis in our study was 12.5% in the cryopreserved grafts and 33.3% in the decellularized grafts. There is a limited number of studies describing implantation of decellularized venous grafts without any further modification in venous system. Therefore, the comparison of our results is limited. Hakkanson et al. presented 100% patency after vena cava reconstruction with decellularized caval vein grafts previously perfused by the recipient blood [37]. However, the pre-treatment with the recipient's blood was not the only difference from our model. Anticoagulation (rivaroxaban) and anti-aggregation (acetylsalicylic acid) were used in their study during the whole postoperative period [37]. Yamanaka et al. examined decellularized venous grafts of small diameter that were modified with integrin α4β1 ligand (with affinity for endothelial cells) and implanted in venous system. The authors did not mention the use of anticoagulation and they observed 62% patency after 6 months [64]. The limited graft patency in our study could be attributed to the omission of anticoagulation or anti-aggregation in the postoperative period. In our previous study focused on portal vein reconstruction with cold-stored venous allografts, we noted portal vein thrombosis in 19.2% when using the same breed of domestic pig and also no postoperative anticoagulation [57]. Decellularized veins implanted in arterial system usually show good patency. Decellularized jugular veins used as grafts implanted in the carotid artery in the canine model showed a 100% patency after 2 or 8 weeks [31, 32]. The endothelization was present only in the perianastomotic regions and the rest of the lumen was covered with a compact fibrin layer [32]. As stated in the aims, our goal was to investigate the spontaneous in vivo recellularization, not to primarily find alternative grafts for vein reconstruction. Therefore, we designed a model with minimal influence on the blood-contact interface of the grafts with intent not to alter any cell adhesion processes. Animals in our experiment received only one bolus of heparin during the operation, but no anticoagulation or anti-aggregation was used in the postoperative period.

4.7 Limitations of the study

The limitation of this study is the short-term survival period. The wall of decellularized graft was similar to the wall of the native vein 28 days after implantation. However, potential long-term changes of the graft cannot be excluded. Another limitation is the thrombosis in some of the grafts not allowing their proper quantitative histologic evaluation. At least these grafts were examined before implantation and did not differ from the non-occluded grafts. The relatively high rate of thrombosis in this study is presumably influenced by the omitting of any anticoagulation or anti-aggregation postoperatively. As described above, it was done on purpose not to influence the cell adhesion process and recellularization itself. Further studies would be needed to evaluate the effect of anticoagulation or anti-aggregation specifically.

5 Conclusion

This study proves sufficient functionality of decellularized IVC grafts enabling their in situ recellularization. 28 days after orthotopic implantation these grafts had a similar structure as the native IVC, and the immunogenicity of these grafts was substantially lower compared to the allogeneic cellular grafts. In situ recellularization has the potential to contribute to the development of artificial organs based on decellularized ECM without the need for full pre-transplant recellularization. It could also help to evaluate the quality of ECM and identify the cells attracted by the scaffold which is crucial for further tissue engineering research.