The Scaffold Derived from GGTA1 and CMAH Gene Knockout Pigs Generated by Gene Editing Provoked Little Inflammatory Responses in a Humanized Pig Model

The porcine small intestinal submucosa ECM (SIS-ECM) has been used as a supportive scaffold for healing in a variety of tissues. However, the outcomes of its application are far from satisfactory. The possibility of generating a porcine small intestinal submucosa extracellular matrix (SIS-ECM) that is fully biocompatible was investigated. Samples of SIS-ECM were prepared from either domestic wild-type (WT) or double-gene knockout (dKO) pigs without antigenic responses to α-Gal and N-glycolylneuraminic acid (Neu5Gc). The scaffolds, which were sutured into cuts made in the longissimus muscle of the dKO pigs, were expected to exhibit xeno-reactions through natural antibodies as in a human response. A process was established to manufacture ameliorating acellular porcine SIS-ECM implants with consistent quality characteristics, which were assured by analyzing the level of residual DNA, the glycosaminoglycan content, and histochemical stains. Once implanted, the acellular SIS-ECM from the WT pigs caused a significant increase in serum IL-6 levels in the dKO recipient pigs, indicating a host defense through immune reactions. The levels remained unchanged when preparations from the dKO pigs were used. The pathological score of the multinuclear giant cells of the dKO (WT) group (2.3±0.5) was significantly greater than that of the dKO (dKO) group (0.7±0.3). The IL-6 levels and pathological evidences suggested that dKO pigs without α-Gal or Neu5Gc antigenic causing lower inflammatory response can serve as biocompatible SIS-ECM donors or as animal models for testing anthropomorphic immune responses to biomedical devices. Porcine small intestine submucosa (SIS) is not an acellular collagenous matrix and contains porcine DNA:


Background
The extracellular matrix (ECM) is a noncellular structure required for the process of forming a new networks and tissues and has now become a successful and widely used medical material in various forms. The porcine small intestinal submucosa ECM (SIS-ECM) has been used as a supportive scaffold for healing in a variety of tissues, including arterial and venous tissues, tendons and wound closures [1][2][3][4]. The porcine SIS-ECM consists of collagen, proteoglycan glycosaminoglycan, glycoprotein, and growth factors, and it has been approved as a commercial biomaterial for a variety of clinical applications [5,6]. However, the outcomes of its application are far from satisfactory; prominent noninfectious swelling and severe pain at the site of implantation often accompany its use because of the contamination of porcine DNA [7] and the Gal epitope, which both evoke immune responses [8]. In general, matrix decellularization can be achieved by processing efforts; the side effects from using the SIS-ECM in clinical applications is approached based on the consequences of a deleterious innate antigen response.
The α-Gal epitope is the chief antigen responsible for hyperacute rejection in xenotransplantation but there is a further non-Gal antigenicity concern. In vertebrates, glycans usually have terminal sialic acids that function as markers in normal cells and are recognized by a variety of receptors (e.g., siglecs) [9] to mediate intercellular and intracellular communication. However, the sialic acid family includes several derivatives of the nine-carbon sugar neuraminic acid. In pigs, apes, and Old World nonhuman primates, N-glycolylneuraminic acid (Neu5Gc) is the non-Gal antigen primarily expressed but in humans it is N-acetyneuraminic acid (Neu5Ac) [10]. These innate anti-Neu5Gc antibodies inevitably provoke chronic inflammation that chiefly contribute to the side effects from the use of porcine SIS-ECM [11].
Depleting α-Gal and converting Neu5Gc to Neu5Ac of human form to diminish the 4 antigenicity can be achieved only by knocking out the responsible genes in the pig genome. The α-Gal free or GGTA1 knockout (KO) pigs [12] and Neu5Ac (Neu5Gc free) or CMAH KO pigs [13] have already been generated through our continuous efforts with adopted CRISP/Cas9 gene editing techniques. Thus, the humanized porcine (HP) SIS-ECM that can be obtained from the dKO pigs can also serve as an animal model for antigenicity testing. In accordance with these possibilities, this study was designed to investigate and compare the healing process of artificially induced skin wounds dressed with either HP or WT SIS-ECM in dKO pigs that simulate the immune response of human. The allotropic dressing preparations were expected to reveal the innate antigens suspected of causing side effects and enhance the understanding of the advantages of using multigene KOs in more biocompatible scaffolds or to generate tissues.

Animals and Animal care
The GGTA1 KO and CMAH KO pigs were generated by Chung et al. [12] and Tu et al. [13], respectively, by CRISPR/Cas9 gene editing. In this study, the animals were generated by cross breeding GGTA1 KO boars with CMAH KO sows. All animals were reared in a station that was free from specific pathogens (atrophic rhinitis, Mycoplasma hyopneumoniae, pseudorabies, Actinobacillus pleuropneumoniae, swine dysentery, scabies, classical swine fever, foot and mouth disease and porcine reproductive and respiratory syndrome). The pigs receiving the SIS were housed in individual cages of 2 m 2 /20 kg pig body weight, under artificially lighting (450-600 lux for 9 h per day) and sunlight through windows. The animals were fed a restricted (4% body weight) commercial diet formulated to meet the requirements recommended by the National Research Council (2012) and had free access to water. The donors were scarified by sedation, electrical stunning and exsanguinated.

5
The porcine jejunums of the WT and KO animals were obtained from the experimental slaughterhouse. The intestines were ligated at each end of the jejunum and the ileum by cotton threads and incised. The jejunums and ileums were procured, placed into an icefilled container and transferred to a nearby laboratory for further processing. Permission to undertake the SIS implantation experiments was granted by the institution (IACUC Approval Nos. 105129 and 107060, Agriculture Technology Research Institute) where the animal study was conducted.Blood samples were collected from the anterior vena cava of all pigs under restraining at days 1, 2, 5, 7 and 14 of the implantation of the SIS-ECM (day 0) and placed in tubes without additives to prepare serum for measuring serum IL-6 concentration. The pathologic study tissues were harvested from the sacrificed recipients at week 1, 2 and 4 after implantation.

Preparation of the SIS-EMC
The porcine SIS-EMC was prepared by the procedure including mechanical disassociation, detergent treatment, and sterilization by peracetic acid. The jejunum samples were carefully removed from the excess mesentery tissue and inverted. Then, the samples were thoroughly and repeatedly rinsed by reverse osmosis water to remove debris. The jejunum samples were cut into lengths of approximately 20 cm and frozen at -30℃ before use. The SIS was isolated using a well-established method of preparation [14,15].
The decellularization protocol was performed by immersing the SIS (3×20 cm) in 100 mL 0.2% SDS saline solution for 4 h under room temperature with moderate shaking. After changing to a new SDS solution, a secondary decellularization step was performed under the same conditions for 16 h. The decellularized SIS was removed and then thoroughly rinsed with deionized water, and perfused for around 15 min with deionized water until the water reached pH 7.0. ± 0.5.
The decellularized SIS membrane (80 cm 2 ) was cut into 2×2 cm sections and placed into 6 sterile tubes with 50 mL of 0.1% peracetic acid (v/v) for 2 h with gentle shaking. After sterilization, the peracetic acid was cleaned from the decellularized SIS membrane by five times flushing with 50 mL of saline. The sterilized, decellularized SIS membrane successfully passed the bioburden test.

DNA Quantification
About 50 mg SIS samples immersed in 400 μL of lysis buffer (20 mM Tris-HCl, pH=8.0; 40 mM EDTA-Na; 80 mM NaCl; and 0.2% SDS) were minced into small pieces with sterile scissors. The tissues were incubated at 56℃ and digested with 20 mg/mL proteinase K until the tissue pieces could no longer be seen in the solution. A total of 900 μL of equilibrated phenol was added, and the solution was mixed gently by inversion for 10 min.
The tubes were centrifuged at 13,000×g for 10 min at room temperature. The upper aqueous layer was pipetted to another clean tube. Next, the extraction procedure was conducted with phenol/chloroform (1/1, v/v) and then with chloroform. The upper aqueous DNA solution was mixed with 600 μL of isopropanol and was centrifuged at 13,000×g for 10 min to precipitate the DNA. The DNA pellet was washed with 70% ethanol and then 100% ethanol and dried at 55℃ in an oven. The DNA was quantified using the Qubit ® dsDNA HS assay kit (Invitrogen, USA) and measured with the Qubit ® fluorometer (Invitrogen, USA).

Nested PCR for detecting porcine DAP 12 DNA
Nested PCR was used to detect the porcine DAP 12 DNA in the SIS-ECM samples [7]. For the first amplification step, a pair of outer primers (forward: 5'-AATGGACTCCTGCCT CTC-3' and reverse: 5'-CCTGGACAAGGCTGAAAC-3') was used in the PCR assay and was expected to produce a 426 bp amplicon. The resulting product from the initial PCR was used as a template for a second set of primers (forward: 5'-AGAAAAGCAAAGGGGTGGGG-3' and 7 reverse: 5'-AAGGGTGGACAGCAAGGTCC-3'). The nested pair of primers was expected to amplify a 223 bp amplicon by PCR. Fifty microliters of the PCR amplification mixture was prepared in a PCR tube containing 5 µL of the target DNA sample and 1 µL of each primer (200 pmol). The DNA was denatured at 94℃ for 2 min and then amplified by 30 cycles with annealing temperature at 54℃. After the amplification, 10 µL from each PCR product solution was loaded onto a 2.0% agarose gel, which was then stained with ClearVision DNA stain (Protect Technology Enterprise Co., Ltd. Taiwan), and then, the products were observed under UV light. Every PCR run included positive (DNA from untreated porcine small intestine) and negative (water) controls.

Histology of the SIS-ECM
All samples were fixed in 10% formalin solution for 24 h prior to staining with hematoxylin and eosin (H&E). The fixed samples were embedded in paraffin and sectioned for staining, and representative images were captured with an upright microscope (Nikon Eclipse E600).

Experiment Design and Intramuscular Implantation of the SIS-ECM
Growing pigs of 15~20 kg body weight, were used for this study. Atropine (0.1 mL /kg) and 0.5 mg Strenil (Janssen Pharm., Belgium) were injected intramuscularly. Animals were then anesthetized with an intramuscular injection of 2.0-8.8 mg/kg of Zoletil (Virbac S. A, FRA).
Under sterile conditions, the SIS-ECM (2 × 2 cm) prepared from the WT, CMAH KO ( CMAH KO) and the dual (GGTA1/CMAH) KO (dKO) pigs were individually implanted into the longissimus of WT and dKO pigs. A full-thickness skin incision of 3×3 cm in an "∟" shape was created, and the deep fascia was opened. Then, the SIS-ECM was implanted into the longissimus and fixed with size 5-0 nylon sutures. The wound was closed with size 5-0 nylon sutures.
For the pathology analysis, six implants were implanted into each pig (WT and dKO) with three SIS-ECM implants from the dKO pigs on the right side and three SIS-ECM implants from the WT pigs on the left side.
After the indicated number of days, all animals were euthanized and exsanguinated. The tissues surrounding the implanted SIS-EMC were cut from the muscle and placed in 10% neutral buffered formalin.
To evaluate the effect of the serum IL-6 levels, three SIS-ECM implants from the same pigs were implanted into WT, CMAH KO or dKO pigs into the same side of the back of each pig by the same procedure described above. Blood samples were collected from pigs on the indicated day for serum IL-6 measurement.

Pathology analysis
The tissues surrounding each test device were also preserved in 10% formalin for histopathological examination. The tissues were trimmed as necessary and embedded in paraffin. One section approximately 4 μm thick was prepared from each site. The slides were stained with hematoxylin and eosin (H&E) and assessed by light microscopy. The histologic evaluation index included multinuclear giant cells (MGC), inflammatory cell infiltration, granuloma formation, and neovascularization. The histopathologic grading of the local lesion response was based on the (4-point) scoring system presented by Shackelford [16]: grade 0, within normal limits; grade 1, minimal; grade 2, slight; grade 3, moderate; and grade 4, severe.

Determination of IL-6 concentration
All blood samples were remained at 4°C and allowed to clot. After centrifugation at 1500×g, the serum samples were collected and stored at -20°C until analyzed. The concentration of IL-6 was determined in duplicate using a Swine IL-6 Pig ELISA kit (REF:ESIL6) from Thermo Fisher Scientific (Maryland, USA). All assays were performed according to the manufacturer's protocols. The detection limit of the kit was 45 pg/mL.

Preparation of the SIS-EMC
The efficacy of the decellularization was evaluated by histological observation and the DNA content. Histological staining with H&E revealed the presences of visible and distinct nuclei in the small intestinal submucosa before ( Figure 1A) and after ( Figure 1B Table 1). The DNA content levels were not significantly different than those obtained by the peracetic acid sterilization procedure. The purified DNA was tested by nested PCR for the DAP12 gene, and no DNA fragment of more than 200 bp could be observed for either the PCR or the nest-PCR amplicons (data not shown).
The SDS treatment did not significantly affect the glycosaminoglycan content of the SIS-EMC samples; and more than 90% glycosaminoglycan remained. However, the samples treated with peracetic acid showed a significant reduction of 66.2~68.4% in glycosaminoglycan content ( Table 2). The amount of glycosaminoglycan and the ratio of its reduction during the SIS-EMC preparation process were similar to the values in the samples from 4-to 5-month-old pigs ( Table 2). A small intestine sample from 13-month-old pigs was used to prepare the SIS-ECM, and the glycosaminoglycan content was 16.22±1.01, 15.05±0.45, and 10.56±0.47 μg/mg in the dKO (un), dKO (SDS) and dKO (All) samples, respectively, near to those from younger pigs.

Analysis of Neu5Gc and Neu5Ac
Tissues from the cornea, ear, tail and small intestine were analyzed. In WT pigs, both Neu5Ac and Neu5Gc were detected in the tissues, but only Neu5Ac could be found in the tissues from the CMAH KO pigs. The content of Neu5Gc was less than that of Neu5Ac in the cornea, ear, and tail of WT pigs, but the Neu5Gc content was more than that of Neu5Ac in the small intestine of WT pigs (Figure 2A). According to the results of the HPLC analysis, no Neu5Gc was not found in the tissues from the CMAH KO pigs. In this study, the process of decellularizing the WT SIS-EMC reduced the Neu5Gc content but obviously the Neu5Gc remained in the SIS-EMC ( Figure 2B). Although Neu5Gc is a minor component in most tissues, including corneal ( Figure 3C), none of the various decellularized processes could entirely remove it. Some processes, including treatment by N2 (showed in Figure 3D), SDS (showed in Figure 3E), formic acid, phospholipase and tonicity (data not show), were adapted to decellularize porcine cornea, but they all failed to remove Neu5Gc in the corneal products.

Pathology analysis
The findings from the various parameter evaluations of each lesion at different time courses (1-, 2-and 4-week) are summarized in Table 3 The indications of proliferation (granulation) were based on granuloma and neovascular formations ( Table 3). The fibroblasts and endothelial cells were proliferating and maturing in the tissues with granulonma formation from weeks 2 to 4. The prominent vascularization of the granulomas indicated sufficient maintenance of a healing response.
However, there were no significant differences in granuloma formation to indicate tissue maturation or blood vessel neovascularization among the groups. On the other hand, no 13 evidence for chronic active inflammation or surrounding muscle necrosis was noted during the entire study period.

Changes in Serum IL-6 levels after SIS-ECM implantation
Except on day 7, the serum IL-6 concentrations were remained unchanged in all samples

Discussion
In previous reports, the α-Gal epitope is persistently present within the ECM derived from the tissue of porcine origin despite a vigorous decellularization process [18][19][20]. In this study, Neu5Gc was also detected in the WT pig SIS-ECM that was produced by the decellularization process. For comparison, several commercial ECM products were analyzed, and Neu5Gc was detected in all the products (data not shown) proving our dKO SIS-ECM is novel with biocompatibility.
During evolution, the human CMAH gene was mutated and does not express intrinsic Neu5Gc; however, human serum contains anti-Neu5Gc antibodies, which represents the majority of the non-Gal antibodies that limit xenotransplantation [21,22]. Similar to Neu5Gc, α-Gal cannot be synthesized by humans due to a loss-of-function mutation in the GGTA1 gene, but all humans carry anti-α-Gal antibodies. It is worth noting that α-Gal has a single carbohydrate xenogeneic antigen (Galα1-3−Galβ1-4−GlcNAc-R) [23], but there are 14 multiple Neu5Gc-glycans on glycoproteins and glycolipids. These findings agreed with the report that the complexity of Neu5Gc should be seriously considered in xenotransplantation [11]. Our results showed that acellular SIS-ECM allografts implanted into dKO when compared to WT or pigs caused lower histopathologic scores of the lesions and unchanged serum IL-6 concentration, indicated no serious immune response. A previous study demonstrated that porcine-derived scaffold containing the α-Gal epitope induced a serum antibody response, but this α-Gal epitope had no adverse effects on the host remodeling response [24]. A recent article reported that, in a nonhuman primate model, decellularized WT and α-Gal KO porcine lung scaffold implants differed little in terms of immunological responses, although the latter showed delayed immune cell infiltration and downregulated chronic T-cell-mediated reactions to the scaffold [25].
During wound healing, platelets release factors throughout the formation of the hemostatic wound plug and initiate the inflammatory response [26]. This inflammatory response is, in part, mediated by proinflammatory cytokines, such as IL-6, which are produced by fibroblasts and macrophages in response to the factors released in the hemostatic wound plug [27]. These proinflammatory cytokines recruit inflammatory cells, including polymorphonuclear cells, monocytes, and macrophages, into the wound. The inflammatory cells are then activated and produce more inflammatory cytokines, causing additional inflammation and cytokine production. These inflammatory cells also produce growth factors that stimulate both fibroblast proliferation and modify the ECM, leading to scar formation [28], which exerted an important effect in the remodeling response [29]. A study on the implantation of non-cross-linked acellular matrices (scaffold) showed that IL-6 positive cells were increased at day 14 and successively decreased thereafter [30]. This IL-6 expression was observed in the location surrounding the wound site in the woundhealing experiments [30]. During transplantation, the serum level of IL-6 increases in association with inflammatory reactions, including allograft rejection [31][32][33].
Furthermore, IL-6 has been investigated extensively in clinical transplant patients since a relatively large amount of IL-6 is released during the early phase of immune activation [34,35]. Blocking both TNF-α and IL-6 was, therefore, beneficial in a pig-to-baboon kidney xenotransplant and that a blockade of two or even more cytokines together could further reduce xenograft rejection [36].
In this study, there was no obvious increase in IL-6 levels in the serum of all the experimental animals 2 days after SIS-ECM implantation. The lower levels of serum IL-6 on day 2 may result from the acellular matrix implanted into the muscles of pigs and caused no infection in the surgical wounds. The serum IL-6 level was increased on day 7 and then decreased to a level comparable to that before implantation ( Figure 5). These data agreed with the histological analysis of the tissues near the implanted SIS-ECM. In our results, the SIS-ECM from the WT and KO pigs were implanted into the muscles of the pigs, and the tissues were examined after 1-, 2-, and 4-week ( Table 3). The results from the histopathologic analysis revealed that the tissues surrounding the SIS-ECM exhibited a prominent host tissue response, with areas of dense accumulation of inflammatory cells along with multinucleated giant cells at 1-week, and the cell infiltration gradually decreased by 2-and 4-week. The novel finding was that different IL-6 levels were detected in the sera from different experimental groups. The Neu5Gc content in the SIS-ECM from the WT pigs was analyzed by HPLC, and a considerable amount of Neu5Gc was detected.
The presence of the α-Gal epitope in biologic scaffolds has been investigated for porcine SIS-ECM [37]. In this study, Neu5Gc and α-Gal were xeno-antigens from the WT or the KO pigs, as their SIS-EMC were implanted into the dKO pigs, the WT SIS-EMC was implanted into the CMAH KO pigs, and the CMAH KO SIS-EMC was implanted into dKO pigs. Each group of recipient animals increased their serum IL-6 levels. In contrast, in the group of dKO pigs which the dKO SIS-EMC were implanted, the serum IL-6 level changed little. A previous report [30] on the biocompatibility of porcine skin scaffolds, as assessed 3, 14, 21 and 90 days after their implantation in rats, indicated that the implants were surrounded by polymorphonuclear cells on day 3 and then were replaced by a notable number of IL-6 positive cells by day 14. The time course of cell infiltration, IL-6 expression, and tissue remodeling were all delayed in comparison with the responses shown by the data of our study. The discrepancy may result from the difference of animal used (rats versus pigs) and the implant site selected (subcutaneous tissue versus muscle).
Macrophages are one of the broadest ranging cell types capable of polarizing entirely from being contributors to tissue inflammation (M1 macrophage) toward being contributors to wound healing (M2 macrophage) [38]. Classical M1 macrophages are induced in response to lipopolysaccharide-expressing pro-inflammatory cytokines such as TNFα, IL-6, and IL-1β, of which all contribute to tissue inflammation and osteoclastogenesis. The M2 macrophages are induced by IL-4 and IL-13 and typically produce TGF-β and arginase, both of which are implicated in the tissue repair process. Histopathologically, the role of multinucleated giant cells around certain classes of biomaterials has been directly linked to xeno-rejection, but a recent article suggested that it is necessary to better characterize them scientifically and appropriately as M1-MGC and M2-MGC accordingly [39].
Obtained results clearly suggested that SIS-ECM derived from dKO pigs showed little antigenic response to α-Gal or N-glycolylneuraminic acid, as determined by the histopathologic scores of the local response lesions and serum IL-6 concentration in the implanted dKO pigs. Those animals can not only serve as biocompatible SIS-ECM donors but also be used as animal models for testing anthropomorphic immune responses to biomedical devices.

Conclusions 17
The SIS-ECM derived from dKO pigs showed little antigenic response to α-Gal or Nglycolylneuraminic acid and therefore has observed lower inflammatory response in the implanted pigs. We have shown that the acellularized SIS-ECM derived from KO pigs performed superior and present a promising future of clinical application. A novel animal model for antigenic response to α-Gal or N-glycolylneuraminic acid has also established.
into the dKO pigs. The histopathological scores were graded 0-4 as follows: Grade 0, within normal limits; Grade 1, minimal; Grade 2, slight; Grade 3, moderate; and Grade 4, severe. * Statistically significant difference between the control and treated groups at p<0.05. Figure 1 Histological