Immunogenicity of Decellularized Porcine Liver for Bioengineered Hepatic Tissue

Abstract

Liver disease affects millions of patients each year. The field of regenerative medicine promises alternative therapeutic approaches, including the potential to bioengineer replacement hepatic tissue. One approach combines cells with acellular scaffolds derived from animal tissue. The goal of this study was to scale up our rodent liver decellularization method to livers of a clinically relevant size. Porcine livers were cannulated via the hepatic artery, then perfused with PBS, followed by successive Triton X-100 and SDS solutions in saline buffer. After several days of rinsing, decellularized liver samples were histologically analyzed. In addition, biopsy specimens of decellularized scaffolds were seeded with hepatoblastoma cells for cytotoxicity testing or implanted s.c. into rodents to investigate scaffold immunogenicity. Histological staining confirmed cellular clearance from pig livers, with removal of nuclei and cytoskeletal components and widespread preservation of structural extracellular molecules. Scanning electron microscopy confirmed preservation of an intact liver capsule, a porous acellular lattice structure with intact vessels and striated basement membrane. Liver scaffolds supported cells over 21 days, and no increased immune response was seen with either allogeneic (rat-into-rat) or xenogeneic (pig-into-rat) transplants over 28 days, compared with sham–operated on controls. These studies demonstrate that successful decellularization of the porcine liver could be achieved with protocols developed for rat livers, yielding nonimmunogenic scaffolds for future hepatic bioengineering studies.

Within the United States alone, tens of thousands of patients are awaiting a liver transplant, with only a few thousand donor organs available annually.¹ This widening mismatch has led physicians and researchers to pursue alternative therapies for chronic liver disease, including in situ cell-based therapies or xenotransplantation of organs.^2–4 The field of regenerative medicine offers another approach, in which elements of both would be combined for the bioengineering of neo-organs for transplantation.^5,6

The concept of whole liver tissue engineering aims to combine patient-specific autologous hepatocytes or hepatic progenitor cells and a carrying platform, or scaffold, to allow for three-dimensional tissue growth and permit the complex cellularity of hepatic tissue. Use of decellularized organ matrices preserves the natural extracellular matrix (ECM) proteins and growth factors that guide cell attachment and proliferation in an organ-specific manner.⁷ Proper processing of the matrix scaffolds removes all cytotoxic chemicals from the decellularization process and performs complete degradation of donor nucleic acids to prevent an adverse host immune response.⁸ These bioengineered livers have the ultimate potential to surpass the current allograft gold standard.

The process begins by removing the native cellular components from a donor tissue using detergents and enzymes and leaving behind an ECM scaffold with preserved vasculature and essential biological factors. The concept has been applied to many tissues, including the heart,^9,10 lungs,^11–14 bladder,¹⁵ blood vessels,^16,17 muscle,¹⁸ intestines,^19,20 trachea,^21–23 kidney,^7,24,25 and liver.^26–30 Each detergent, enzyme, washing buffer, and sterilization technique used to decellularize a tissue can have a direct influence on the host remodeling response and functional outcome.³¹ In a previous study, decellularized matrix scaffolds were immunologically favorable up until cells were added to the scaffolding material, where proinflammatory macrophages were activated.³² To evaluate whether a decellularized tissue represents a viable scaffold option, the generated matrices need to be implanted and evaluated over time, without cells, to allow a host’s immune cells to infiltrate and respond to the material.³³ The initial response can begin as early as 2 days and last for months.³⁴ During that time, the environment from both the host itself and the degrading matrix material can influence the phenotype of the host immune cells switching between activation states that will determine the future clinical viability of the biological matrix material.^35–37 Triggering of proinflammatory macrophage activation results in the release of cytokines, growth factors, proteolytic enzymes, and reactive oxygen and nitrogen intermediates that will greatly inhibit the integration of the biomaterial with the host tissue.³⁸ Studies on the immunogenicity of decellularized whole tissue are limited, and a key criterion of transplantation viability will be evaluating the activation of host macrophages toward the classically proinflammatory phenotype (M1) or the regenerative and repair phenotype (M2).

The objectives of the current study were to generate a decellularized porcine liver by scaling up our previously established rodent perfusion protocol,²⁸ characterize the resultant scaffold, and compare the in vivo immunological response of a rodent host between allograft and xenograft decellularized liver matrices. We hypothesized that both tissues would elicit a similar host response as the result of the high levels of preservation the ECM protein structures share between species.³⁹ The generation of large-scale hepatic tissue platforms, and an understanding of the inherent immune response by a host species, will be vital in producing implantable bioengineered livers.

Materials and Methods

Apparatus and Reagents

Triton X-100, scanning electron microscopy (SEM)–grade glutaraldehyde (G5882), SDS (L5750), and all other reagents were purchased from Sigma-Aldrich (St. Louis, MO), unless otherwise stated. The PBS salts were obtained from Thermo Fisher Scientific Inc. (Pittsburgh, PA), and sodium heparin was obtained from SAGENT Pharmaceuticals (Schaumburg, IL). Masterflex L/S Digital Drive peristaltic pumps, Easy-Load II pump heads, and silicone tubing were purchased from Cole-Palmer Instrument Co (Vernon Hills, IL). Liberate Antibody Binding solution (24310) was purchased from Polysciences Inc. (Warrington, PA). Serum-free protein block (X0909), antibody diluent (S3022), dual endogenous enzyme block (S2003), and primary antibodies were purchased from Dako NA Inc. (Carpentaria, CA). Vectastain RTU ABC Reagent (PK-7100), ImmPACT DAB [diaminobenzidine] Peroxidase Substrate Kit (SK-4105), avidin/biotin blocking kit (SP-2001), and other secondary antibodies were obtained from Vector Laboratories (Burlingame, CA). Other primary antibodies were purchased from Abcam plc (Cambridge, MA), Santa Cruz Biotechnology Inc. (Santa Cruz, CA), and AbD Serotec (Raleigh, NC). An ApopTag Peroxidase In Situ Apoptosis Detection Kit (S7100) was purchased from Millipore (Billerica, MA). Fetal bovine serum, Dulbecco’s minimum essential medium F-12, and penicillin-streptomycin solution were from HyClone Laboratories (Logan, UT). Surgipath MM24 Mounting Medium (3801122) was obtained from Leica Microsystems GmbH (Wetzlar, Germany). Fischer rats F344 were obtained from Harlan Laboratories Inc. (Indianapolis, IN).

Liver Decellularization

Young adult pigs (20 to 25 kg) were i.v. heparinized with 400 U/kg, and then euthanized according to Institutional Animal Care and Use Committee guidelines. Livers (450 to 500 g) were harvested and placed on ice, and then cannulated via the hepatic artery. Organs were perfused with ice-cold PBS containing 10 U/mL sodium heparin until circulated perfusate ran clear. Livers were perfused at 50 mL/minute with increasing concentrations of Triton X-100 (1%, 2%, and then 3%) in chilled PBS, then 0.1% room temperature SDS in PBS. Livers were rinsed via the hepatic artery with chilled PBS for several days at the same flow rate, then cut into pieces (2 cm thick) and placed in PBS-containing antibiotic-antimycotic supplements on a shaker at 4°C for a further 2 to 3 days. Rat liver decellularization was performed as previously described.²⁸ To generate liver scaffolds, decellularized liver pieces were sliced into transverse sections (1 mm thick) and placed flat. Then, a biopsy was performed with a 15-mm-diameter punch. Scaffolds were placed in PBS, sterilized via 1 Mrad (10 kGy) γ irradiation, and then stored at 4°C until use.

Histological Analysis

Samples for histological staining were fixed overnight in 10% neutral-buffered formalin, then transferred to 60% ethanol before dehydration in successive alcohols, immersion in xylene, and final embedding in paraffin wax. Samples were divided into sections onto glass slides and stored at 4°C. Rehydrated sections were stained with histochemical dyes for either H&E (black/pink for nuclei/tissue, respectively) or Masson’s trichrome method (blue/black/red for collagen/nuclei/cytoplasm, respectively), according to established protocols.^40,41 All images were captured on a Leica Microsystems (Wetzlar, Germany) DM4000B upright microscope with ImagePro (MediaCybernetics, Rockville, MD) software version 6.3. For SEM, samples were fixed in 2.5% SEM-grade glutaraldehyde in PBS, then dehydrated in successive alcohols before critical point drying with carbon dioxide. Samples were sputter coated and visualized via a Hitachi S-2600N SEM with accompanying manufacturer image-capture software (Hitachi High Technologies America, Inc., Gaithersburg, MD).

Immunostaining

Antigen retrieval (when necessary) was performed on rehydrated sections with Liberate Antibody Binding solution, followed by quenching of endogenous peroxidase activity with dual endogenous enzyme block, then serum-free protein-, avidin-, and biotin-blocking steps. Primary antibodies or isotype controls were diluted in antibody diluent at the same protein concentration and incubated on sections overnight at 4°C. Negative controls lacked primary antibody. Sections were blocked again before application of secondary antibodies also diluted in antibody diluent, and incubated for 1 hour at room temperature. Slides were washed with Tris-buffered saline, incubated with ABC Reagent (Dako) for 30 minutes at room temperature, and then washed again with Tris-buffered saline. Once diaminobenzidine (DAB) chromogen was developed, sections were counterstained with hematoxylin, dehydrated, and mounted in nonaqueous mounting medium. Antibodies used are listed in Table 1, and controls are given in Supplemental Figure S1. TUNEL staining followed manufacturer’s protocols, with ABC Reagent incubation, DAB chromogen development, and hematoxylin counterstaining, as previously described.

Table 1.

Antibodies Used for Immunochemical Staining

Primary antibody	Isotype control	Secondary antibody
Anti-CD3	Mouse IgG	Biotinylated anti-mouse IgG
Dako A0452	Vector I-2000	Vector BA-2001
Anti-CD68	Mouse IgG	Biotinylated anti-mouse IgG
Abcam ab31630	Vector I-2000	Vector BA-2001
Anti-CD86	Rabbit IgG	Biotinylated anti-rabbit IgG
Abcam ab53004	Vector I-1000	Vector BA-1000
Anti-CD206	Rabbit IgG	Biotinylated anti-rabbit IgG
Abcam ab64693	Vector I-1000	Vector BA-1000
Anti-elastin	Mouse IgG	Biotinylated anti-mouse IgG
Abcam ab9519	Vector I-2000	Vector BA-2001
Anti-laminin	Rabbit IgG	Biotinylated anti-rabbit IgG
Sigma-Aldrich L9393	Vector I-1000	Vector BA-1000

Liver Scaffold Seeding

Sterilized scaffolds in PBS were equilibrated in fetal bovine serum overnight at 4°C, then soaked in hepatocyte medium (Dulbecco’s minimum essential medium F-12 with 10% fetal bovine serum and 1% penicillin-streptomycin) briefly before static seeding with hepatoblastoma (HepG2) cells. Seeded scaffolds were maintained in hepatocyte medium in a humidified environment at 37°C with 5% CO₂ up to 21 days. At harvest, scaffolds were washed in PBS and fixed for histological analysis.

In Vivo Implantation of Scaffolds

Male Fischer F344 rats (180 g) were handled and anesthetized according to Institutional Animal Care and Use Committee guidelines. Irradiated decellularized pig or rat liver scaffolds were implanted s.c. into the fat pad between the shoulder blades. Scaffolds were explanted after 2, 7, 14, and 28 days and fixed for histological analysis. Blood was collected at corresponding time points or after euthanasia for complete blood cell count analysis via a Siemens ADVIA 120 Hematology system (Siemens Medical Solutions USA Inc., Malvern, PA).

Statistical Analysis

Statistical analysis was performed using two-way analysis of variance, with Tukey’s post hoc tests for comparisons between individual groups and time. P < 0.05 was considered significant. All studies were performed in triplicate or higher. Values are expressed as means ± SD, with statistical analyses performed on GraphPad (La Jolla, CA) Prism software version 6.0a.

Results

Decellularization of the Porcine Liver

Multiple decellularization techniques exist, including diffusion for thin or laminar tissue and perfusion for larger, complex, solid organs. Herein, we used perfusion techniques applied to the rat liver with sequential stages of Triton X-100 at 1%, 2%, and 3%, followed by 0.1% SDS (Figure 1), which resulted in a marked tonal difference in the organ (Figure 1, A–C). Macroscopic and microscopic vessel branching was visible below the tissue surface (Figure 1, D and E, respectively). A histological examination demonstrated the loss of cellularity from normal to decellularized tissue, with a lack of nuclear hematoxylin staining (Figure 1, F and H) and clearance of cellular cytoplasmic keratins, leaving behind a collagenous-rich, acellular matrix, as denoted by Masson’s trichrome staining matrix (Figure 1, G and I).

Figure 1.

Porcine liver decellularization. Whole native liver (A) before perfusion with Triton X-100 (B) and SDS (C). Indicated by panels on the SDS-treated liver, the preserved major (D) and minor (E) vessels are visible at the end of the protocol. Native liver sections stained with H&E (F) and Masson’s trichrome (G) help illustrate the complete cellular removal seen in their respective decellularized counterparts (H and I). Arrows indicate preserved vessels in decellularized matrices compared with native structures. Scale bar = 100 μm (F–I).

Preservation of Liver Structures

A key rationale for the use of decellularized scaffolds is the preservation of natural components of the organ and ECM (Figure 2). Immunohistochemical staining showed the positive presence of elastin and laminin surrounding preserved vessel structures (Figure 2, A and B), and highly distinct tissue topographies were also preserved after decellularization, as demonstrated by the comparison of the capsule, parenchymal tissue, and vessel structures (Figure 2C). The preserved capsule showed a tight matrix with limited porosity (Figure 1D), enough to prevent leakage of perfused reagents (data not shown). Preserved vessel structures demonstrated a smooth and intact wall (Figure 2E). Finally, the parenchymal space of the decellularized tissue showed an open and highly porous structure, applicable for recellularization studies (Figure 2F).

Figure 2.

Preservation of discreet ultrastructural components. Positive immunostaining for vascular elements elastin (ELN; A) and laminin (LN; B). C: Scanning electron micrographs of the decellularized porcine liver showed distinct areas of the capsule, vessels, and parenchyma, shown under higher magnification in D–F, respectively. Original magnification: ×50 (C); ×200 (D–F). Scale bar = 200 μm (A and B).

Cytotoxicity of Decellularized Liver Matrix

The preparation of decellularized matrices involves cell and organelle lysis and digestion, through the application of detergents and, in some cases, enzymatic means, but the resultant matrices must demonstrate noncytotoxocity and, moreover, scaffolds must support cell growth to be considered for organ bioengineering purposes (Figure 3). HepG2 cells were statically seeded onto decellularized liver scaffolds and maintained for up to 21 days, when dense cell layers were observed. Cells at day 7 (Figure 3A) and day 21 (Figure 3B) did not exhibit apoptotic markers, as displayed in control samples with degraded DNA (Figure 3C) and confirmed by TUNEL staining. Cells were observed attached to the surfaces of matrices, with minimal penetration into the liver matrix scaffold.

Figure 3.

In vitro noncytotoxicity of porcine liver matrix scaffolds. HepG2 cells were statically seeded onto scaffolds and cultured out to 21 days to determine whether any cytotoxic compounds would be released from the decellularized matrices. TUNEL staining of the scaffolds at 7 (A) and 21 (B) days indicated minimal apoptotic response of HepG2 cells to liver scaffolds, in both close contact with scaffolds and within cell masses (arrows), compared with DNase-treated positive control samples (C). LM, liver matrix. Scale bar = 50 μm (A–C).

Immunogenicity of Liver Matrix Scaffolds

To investigate the immunogenicity of our decellularized scaffolds (rodent and porcine), we implanted naked scaffolds (Figure 4, C and D) into the s.c. dorsal adipose tissue of rats (Figure 4A). Scaffolds were recovered (Figure 4B) with surrounding tissue intact to measure cellular infiltration and host response, with no visible fibrous encapsulation of scaffolds or exudates observed at implantation sites. After 7 days (Figure 4, E and F), cells had migrated into both allogeneic (rat) and xenogeneic (porcine) scaffolding, with no signs of an inflammatory response. Over 28 days (Figure 4, G and H), host cells continued to populate each recovered scaffold, with no noticeable adverse host response surrounding the matrices. Systemic white blood cell counts were collected from animals during the experiment and compared with rodents who underwent implantation surgeries, but received no matrix implant (sham). Total white blood cell counts (Figure 4I), lymphocyte counts (Figure 4J), and monocyte counts (Figure 4K) were not significantly different between groups (P > 0.05) at 2, 7, 14, and 28 days, indicating no major systemic host response compared with normal surgical recovery. Lymphocyte count results were supported by little to no CD3⁺ T-cell activation at the implantation site, as seen over the 28 days in either group (Figure 5, A and B). Many cells that infiltrated the implanted scaffolds were positive for the pan-macrophage phenotypic marker, CD68 (Figure 5, C and D). However, as seen previously, neither the monocyte counts (as macrophage precursors) (Figure 4K) nor the cells expressing markers indicating adoption of the M1 (CD86⁺) or M2 (CD206⁺) phenotypes (Figure 5, E–H) showed an increase.

Figure 4.

Host response. In all groups, grafts were implanted s.c. in the dorsal adipose tissue of a given animal (A) and harvested intact for histological analysis (B), with no visible encapsulation or exudate at the implantation site. Arrows denote the four scaffolds implanted per animal. C and D: H&E staining of nonimplanted biopsy specimens of decellularized allogeneic (Allo; rat) and xenogeneic (Xeno; pig) livers indicated no cellular material in the scaffolds before implantation. E and F: Day 7 (d7) allogeneic and xenogeneic explants indicated cellular infiltration of the scaffolding material, with no fibrous encapsulation under microscopic examination. G and H: Day 28 (d28) explants continued to indicate cellular activity inside both allogeneic and xenogeneic scaffolding materials, with no encapsulation of the scaffold. Systemic white blood cell analysis over the length of the experiment indicated no significant difference (P > 0.05) between implanted allograft or xenograft scaffolds compared with sham–operated on animals (no scaffold implantation) in total white blood cell count (I), lymphocyte count (J), or monocyte count (K), indicating no major host response to the ECM materials. Significantly decreased counts in all groups were observed over time (P < 0.05). Scale bar = 100 μm (C–H).

Figure 5.

Cell-specific activation. Anti-CD3⁺ staining of allogeneic (Allo; A) and xenogeneic (Xeno; B) liver explant tissues on early [day 7 (d7)] and late [day 28 (d28)] time points indicated minimal or no CD3⁺ T-cell activation. CD68⁺ pan-macrophages were seen within implanted allogeneic (C) and xenogeneic (D) matrices over the duration of the experiment. Staining for macrophage polarity indicated minimal or no activation of CD86⁺ (M1) in either allogeneic (E) or xenogeneic (F) scaffolds over early and late time points. G and H: Similar results were observed for the presence of CD206⁺ (M2) macrophages, with limited positive staining in xenogeneic samples. Scale bar = 50 μm (A–H).

Discussion

In this study, we developed a successful protocol for the generation of clinically relevant sized decellularized porcine livers, as derived from our previous rodent liver decellularization method.²⁸ These studies demonstrated, after detergent perfusion and rinsing, cellular clearance and preservation of the vasculature tree and ECM proteins, as shown by our previous scaffolds^24,28 or scaffolds generated by other groups.^26,30,42 Maintenance of patent blood vessel structures provides an essential nutrient distribution network for whole organ regenerative studies. Without blood vessel structures, static transport limitations within the scaffolding material will generate necrotic pockets because cell proliferation would leave the cells most distant from nutrient sources starved.³⁸ With blood vessel structures, we avoid such a detrimental situation and instead replicate the body’s nutrient delivery and toxin removal by perfusing the vasculature at model pressure and flow conditions.

The vasculature also provides a lattice structure for endothelial cell attachment and proliferation that will be crucial for prevention of thrombosis on vascular anastomosis within the recipient. Specifically for the liver, an implantable unit will need to be directly connected to the host vasculature. A previous study demonstrated the successful decellularization and implantation²⁵ of porcine kidney scaffolds, which confirmed that an acellular matrix could withstand the strain of surgical implantation and exposure to arterial blood flow. Our liver scaffolds have also maintained the integrity of the organ’s capsule, critical for preventing increases in permeability and circulation leaks that would be detrimental in maintaining the closed-loop vasculature.

The preservation of basement membrane proteins, such as collagens, fibronectin (data not shown), elastins, and laminins, has been shown to maintain hepatic cell function for longer time periods than classic substrates.^43–45 As well as primary cells, embryonic and mesenchymal stem cells have been infused into decellularized scaffolds and shown to develop organ-specific phenotypes and functions simply through scaffold interaction.^7,42 Our present studies focused on minimizing the cytotoxicity of our scaffold material and demonstrated no contact or soluble toxicity, with maintenance of HepG2 hepatoblastoma cells over 21 days. Our follow-up research is focused on the co-culture of primary hepatic and endothelial cells on the matrix material for the development of a functional and implantable liver device.

Related to cellular engraftment, viability, and proliferation, there are ongoing investigations into the effects of γ sterilization and peracetic acid chemical sterilization on scaffold elasticity.⁴⁶ It has been reported that the elastic modulus of the underlying substrate can directly affect cell attachment, behavior, and viability.^47,48 Anecdotal observations suggested significant alteration of the physical scaffold stiffness pre-γ and post-γ sterilization, supporting previous studies.^49,50 However, specifics into the impact these changes have on tissue regeneration are limited.⁸ We believe each step in our decellularization protocol should focus only on the removal of the donor cellular material, while minimizing the impact on inherent mechanical properties of the substratum. Ongoing studies are examining alternatives to irradiation to achieve whole organ sterilization before reseeding.

Before progressing to the transplantation of regenerated liver tissue, a basic understanding of the host immune response toward the scaffolding material is essential to determine whether a decellularized biomaterial is able to affect proinflammatory macrophage activation.^38,51 Our studies assessed the response to rodent and porcine decellularized liver biopsy specimens within a rodent model (allograft versus xenograft proteins). In doing so, we reaffirmed the similarity of the scaffold components between species and reduced future concerns over matrix properties when evaluating host immune response to a recellularized scaffold. The influx of macrophages without activation, as seen with our scaffold materials, is not uncommon and may, instead, be an indication of the influence of potential factors stored in the matrix material itself.^32,35

In summary, our protocol adaptation was able to generate large-scale organ scaffolds of a clinically relevant size for further investigations of organ regeneration. Our matrices preserved essential ECM proteins for cell engraftment and function, as well as the vasculature required for nutrient distribution for whole organ reseeding. Neither our previously developed rodent scaffolds nor newly prepared porcine scaffolds elicited a host immune response, while also readily being colonized by host cells. These results indicate that a naturally derived ECM scaffold is viable for the next stages of regeneration of a bioengineered hepatic tissue.

Supplemental Data

Supplemental Figure S1

Immunocytochemistry controls. A: CD3⁺ rat thymus. B: CD68⁺ rat spleen. C: CD86⁺ rat spleen. D: CD206⁺ rat spleen. E: IgG mouse isotype staining of rat thymus. F: IgG rabbit isotype staining of rat spleen. Scale bar = 50 μm (A–F).