Biocompatibility Evaluation Criteria for Novel Xenograft Materials: Distribution and Quantification of Remnant Nucleic Acid and Alpha-Gal Epitope

Objective: Commercially decellularized xenogeneic scaffolds are currently employed as for the healing of diseased tissues. Strangely enough their use is permitted even in absence of any assessment of the elimination of xenogeneic cell material, as the alpha-Gal epitopes. In addition, the decellularization procedures are not monitored to prove the elimination of the calcific potential associated to the nucleic acids remnants. The currently treatment with glutaraldehyde is unable to grant a complete immuno-tolerance of implanted xenogeneic tissues, reducing but not eliminating the immunogenicity particularly for the alpha-Gal epitope (the major hindrance for the success of xenotransplantation). Recently, our group has extensively reported studies focused on the evaluation of biocompatibility properties of xenogeneic tissues. In this report we are performing this investigation and nucleic acid detection to novel xenogeneic tissues that have shown very promising preclinical/clinical results in different areas of application.


Introduction
Due to the ageing of human populations, the number of elderly patients suffering from chronic diseases or end-stage organ failure is constantly increasing. Physicians treat organ or tissue loss by transplanting organs from a compatible individual (homotransplantation) performing surgical reconstruction or using mechanical or bioprosthetic devices. Unfortunately, the request for organ and tissue (heart, kidney, lung, pancreas, cardiac valves, cartilage, etc…) far exceeds the number of available donors (Table 1). Additionally tissue substitutes have demonstrated a limited durability (heart valve substitutes, ligament and tendons, cartilage) and artificial prosthetic devices may badly affect patient's quality of life (VADS, pacemaker, hemo-dialysis, etc.).
The lack of homologous substitutes has led biomedical researchers to explore the use of animal donors as an attractive and unlimited source of biological devices (xenotransplantation). Scaffolds derived from mammalian tissues have been used to repair or replace a variety of damaged or diseased tissues including cardiac [1,2], esophageal [3,4], dermal [5], musculotendinous [6,7] and ocular [8], among others. Results of preclinical and clinical studies have varied from success [9][10][11] to complete failure [12][13][14][15][16][17]. The host response to these materials can depend on several aspects such as the original species (e.g. porcine, equine, bovine, fish), the tissue from which the extracellular matrix (ECM) has been isolated (e.g., dermis, small intestine, pericardium, cornea), the mechanical loading, the onset of biologic reaction and the effects of trophic factors to which the scaffold is exposed following implantation and host blood perfusion. Moreover, tissue treatment and decontamination/sterilization methods applied during manufacturing process can also markedly affect remodelling and functional outcome.
antigens and remnants of nucleic acids. Such treatments are commonly based on the use, under controlled conditions, of physical agents, detergents, enzymes and/or chemicals compounds with minimal or limited effects on architecture, ultrastructure, mechanical integrity and biological activity of the residual ECM scaffolds. However some of these treatments have been unable to completely remove remnants of cellular material such as nucleic acids and membrane fragments [18][19][20][21][22][23]. Such residues, entrapped among the ECM network, are able to activate innate and acquired immune reaction, and lead to acute failure of the implanted graft with subsequent fibrosis, scarring and encapsulation, up to the establishment of a chronic inflammatory reaction [24]. DNA remnants have been advocated as the main cause of "inflammatory reaction" and calcification phenomena (due to the presence of charged phosphoric acid ester groups) following the implantation of xenograft-derived scaffolds in orthopaedic or cardiac applications [12,25]. Notwithstanding the effort to completely eliminate each single whole-structured cell [26], antigenic determinants may persist in membrane residues and still be capable of reacting (e.g. the xenoantigen alpha-Gal considered the major hindrance for the success of xenotransplantation) [15,27]. This sugar moiety is expressed in most mammalian tissues, excluding humans and higher primates [28]. In humans, continuous antigenic stimulation by the gastrointestinal flora results in production of anti alpha-Gal antibodies accounting for 1% of the circulating immunoglobulins, both IgMs and IgAs [29,30]. Once a xenogenic tissue is recognized, the complement cascade is activated triggering endothelial cell dysfunction, platelet aggregation and vascular thrombosis [31].
The most common approach to preservation and manufacturing of bioprosthetic substitutes (pericardial tissue patches, biological heart valves prostheses, ligaments and tendons) is glutaraldehyde (GLU) fixation. GLU treatment has been considered, for more than 30 years, as the standard chemical process able to ensure tissue biocompatibility, increasing mechanical strength and allowing sterilization and safe storage of heart valve bioprostheses. Unfortunately, GLU fixation does not provide the complete masking of all xenoantigens, in particular the alpha-Gal epitope [32,33], eliciting an anti-αGal specific IgM immunoresponse once the bioprosthetic substitute is implanted in humans [34].
Over the last four years, our group has developed a solid experience on assessing the biocompatibility of xenogeneic tissues, based on the quantification and distribution of such xenogeneic epitope [26,31,35]. Accordingly we have investigated the content and distribution of alpha-Gal as well as of nucleic acid in a variety of novel xenogeneic tissues used or intended for use in clinical practice, that have shown very promising outcomes in different areas of application. The investigated tissues were derived from different animal species and were subjected to different manufacturing procedure. Producer and characteristics of the tested materials are reported in Table 2.

Materials and Methods
All procedures and manipulation of tissues were always carried out in sterile conditions. From each type of tissue 10 samples were obtained for the alpha-Gal quantification, 3 samples for the immunofluorescence assay and 4 samples for the total DNA purification and quantification.

Tuna cornea samples collection
Six tuna cornea (TC) were directly removed from six freshly caught Atlantic Bluefin and placed in sterile cold phosphate buffered saline (PBS, Sigma, St. Louis, USA) in the presence of a full complement antibiotic solution (Penicillin 100 U/ml, Streptomycin 100 mg/ml and Amphotericin B 250 μg/ml) during transport to the laboratory. All blood residues were removed by sterile PBS washing and samples preserved in 0.3% GLU solution for 12 hours. Tuna cornea fixation was performed according to the technique described by Parravicini et al. [36]. To prevent folds, the specimens were inserted and locked inside a frame, and immersed in 0.2 M acetate buffer (pH, 4.6) with 0.3% of GLU for 15 days. The corneal material was then treated for 24 hours with the same acetate buffer in 4% formaldehyde solution, for sterilization and removal of residual fixative traces.

CorMatrix samples collection
Samples of CorMatrix ® ECM™ for cardiac tissue repair have been provided by SIC System (Italian distributor). Three decellularized SIS-ECM sheets (7 cm × 10 cm), each from a different product lot, were extracted from their packaging and placed in phosphate buffered saline (PBS), pH 7.4 at room temperature for 25 minutes to allow proper rehydration as required by the "Instruction for Use" provided by the manufacturer.

Equine pericardium samples collection
Two equine pericardial (EP) patches model XAG-400 (Edwards Lifesciences, Irvine, USA) treated with GLU were extracted from their packaging and subjected to three consecutive washes with sterile phosphate buffered saline (PBS), pH 7.4 at room temperature for 4 minutes to remove the excess of the preserved solution as required by the "Instruction for Use" provided by the manufacturer.

Alpha-Gal ELISA test: tissue preparation
Each sample from TC, SIS and EP, was gently blotted on special filter paper (Whatman International Ltd, Maidstone, England) and its weight recorded (weight range 30-50 mg). As previously reported by our group [35], the GLU fixation procedure performed on the TC and EP samples is responsible for aspecific interaction between the epitopes and the antibody used for their detection. In order to avoid such confounding results, previously GLU-treated, TC and EP tissues were subjected to a "GLU inactivation procedures" actually under patent (European Application no. EP13154536.0), before performing the ELISA test. Conversely, SIS does not require such processing because supplied by the manufacturer as decellularized tissue not treated with GLU.
Subsequently, specimens were further minced and incubated with the primary anti-alpha-Gal antibody M86 (Axxora, Nottingham, UK) [1:50] for 2 hours at 37°C with gentle stirring and finally centrifuged at 14,340 g for 30 minutes at 4°C.

Alpha-Gal ELISA test
The adopted ELISA test is briefly outlined [31]. A Polisorp 96 well plate (Nunc, Rochester, NY), was coated with 50 μl of alpha-Gal/ BSA (Bovine Serum Albumine -Dextra Laboratories, UK), 5 μg/ml, for 2 hours at 37°C. After washing three times with PBS, the blocking procedure was performed using 250 μl per well of 1% HSA (Human Serum Albumine -Sigma, St. Louis, USA) in PBS for 2 hours at room temperature in darkness; then wells were washed three times as above.
A set of wells was incubated with 100 μl of supernatant derived from the different samples (TC, SIS and EP) overnight at 4°C in darkness. After washing, the secondary HRP-conjugate antibody [1:500] (Dako Cytomation, Denmark) was loaded. Finally, 100 μl of horseradish peroxidase substrate buffer (phosphate-cutrate buffer with sodium perborate plus o-phenylenediamine; Sigma, St. Louis, USA) was added to each well for 30 minutes at room temperature in darkness. The plate absorbance was measured by Titertek Multiskan Plus plate reader at 450 nm.

Total DNA determination
Samples (weight range 15-20 mg) were processed for total DNA extraction and purification by the DNaesy Blood & Tissue Kit (Qiagen, Texas, USA). Briefly, samples are first lysed using proteinase K (Qiagen, Texas, USA). The lysate is loaded onto minispin column and buffer was previously adjusted to provide ideal DNA-binding conditions. During centrifugation, DNA is selectively bound to the DNeasy membrane as contaminants pass through. Remaining contaminants and enzyme inhibitors are removed in two wash steps and DNA is then eluted in buffer. The subsequently quantification was performed by a NanoDrop 1000 Spectrophotometer (Thermo Scientific, Delaware, USA).

Statistical analysis
Data are expressed as mean ± SD. Significance of the differences between samples of the same or from different sheets were determined by one way ANOVA. Differences were considered statistically significant when p<0.05.

Alpha-Gal ELISA test
As reported in Table 3, the ELISA test did not reveal the presence of alpha-Gal xenogeneic epitopes within the TC and SIS samples tissue. Conversely the GLU-fixed EP samples showed an average amount of 9.7 ± 4.9 × 10 10 epitopes each 10 mg of wet weight (n=10, p>0.05).

Immunofluorescence analysis
The presence of the alpha-Gal epitopes in EP tissue was confirmed also by immunfluorescence analysis (Figure 3). At lower magnification ( Figures 3A and 3C) the epitope distribution seems to be limited to the endothelial layer, while it resulted to be present in high amount also in the inner stroma as evidenced in Figure 3B. TC and SIS tissues did not exhibit alpha-Gal residues, confirming the data previously obtained by the ELISA test. As shown in Figures 1 and 2, the lack of any red fluorescent signals is a clear indication of the absence of alpha-Gal in this tissue. A diffuse distribution of nucleic acids material was a common observation for all the analyzed tissues, as confirmed by DAPI  staining. This finding was unexpected in the SIS-ECM sample ( Figure  2) because provided as decellularized tissue by the manufacturer.

Total DNA determination
Despite SIS-ECM was previously subjected to decellularizing procedures, the average amount of total DNA content, was similar to GLU-fixed TC specimens: 6.64 ± 1.34 μg and 6.20 ± 0.74 μg/10 mg of wet tissue respectively (Figure 4). A greater amount, 15.30 ± 1.61 μg/10 mg, was exhibited by EP. Differences between samples of the same tissue were not statistically significant (p>0.05).

Tuna cornea
TC has several intrinsic characteristics that make it suitable for bioapplication such as its strength, size, ready availability and histologic composition. Previous results of physical tests indicated that TC stroma features biomechanical characteristics similar to those of bovine pericardium already in use for bioprosthetic heart valve manufacturing [36]. These functional properties, combined with low antigenicity [37], make TC a potential biomaterial for the production of heart valve substitute as well as tissue patches. However the presence of the alpha-Gal gene in ocular surface tissue could be decisive for the rejection of the implanted xenograft [38]. Up to now, a qualitative evaluation of such xenoantigen in corneal tissue has been limited to three fish species [39][40][41]: our results implement this knowledge. In fact the low antigenic profile of TC is determined by the complete absence of the alpha-Gal epitope as confirmed by our performed immunofluorescence analysis ( Figure 1) and ELISA test (Table 3). While the lack of such antigen ensures a better biocompatibility, the widespread presence of the nucleic acids material in the GLU-treated biomaterial, will expose the same tissue to a high probability of calcification occurrences similarly to the current biological heart valve substitutes [42] (even if these latter maintain a considerable amount of alpha-Gal epitope able to react with circulating human antibodies [35]). The low amount of DNA content revealed in the tissue (Figure 2) is justified by its low cell density which characterizing this type of matrix. These results are likely to explain, at least in part, the excellent in-vivo and in-vitro results reported by the use of this tissue as a material for heart valve substitute manufacturing [36]. Otherwise, the implant performed in a sheep model by Parravicini R et al. [36] does not appears considerably representative for the evaluation of the calcific potential of this biomaterial due to the insufficient follow up (6 months).

CorMatrix™ ECM
Features like a great level of graft resorption and remodelling, an high content of trophic factors as well as an excellent biomechanical behaviour (very similar to that of glutaraldehyde treated bovine pericardial tissue) [43], have identified the acellular porcine SIS-ECM as an optimal source of material to manufacture cardiovascular biological devices. SIS-ECM has been shown to promote cells attachment and proliferation (epidermal cells, fibroblast [44], smooth muscle cells [45], cardiomyocytes and myofibroblasts [46]), recruitment of mononuclear cells with an M2 phenotype [24] and graft vascularization [47]. However, to meet the requirements of optimal biocompatibility, a proper decellularization treatment could lead to a non-thrombogenic alpha-Gal epitope-free acellular scaffolds, able to potentially avoid inflammatory and calcification phenomena. A previous report by McPherson and colleagues clearly confirmed the recognition of the αGal epitope in porcine SIS, and its prompt binding with human IgG [48]. Both the ELISA test (Table 3) and the immunofluorescence analysis ( Figure 1) performed in this study highlight as the decellularization treatment devised by CorMatrix ® ECM™ is able to achieve the elimination of the resident xenogeneic alpha-Gal antigens. While the alpha-Gal-free feature can be considered as a indication of good decellularization treatment (thus confirming also the removal of the residues of cell membranes), the large amount of nucleic acids   debris found within the CorMatrix ® ECM™ should be considered alarming, as liable for potential initiation of the calcification process [15,25]. The porcine acellular CorMatrix ® ECM™ has been shown to maintain an average DNA amount (6.96 ± 1.6 μg/10 mg) equal to half of that originally present in a porcine native tissue (11.48 ± 2.4 μg/10 mg) [49]. These DNA residues, spreading throughout the thickness of the SIS-ECM tissue (Figure 1), are likely prone to calcify as observed in acellular SIS-ECM graft covering the internal side of porcine pulmonary valved stents [45]. Similar results were reported also by Zheng MH et al.
[12] and Walton JR et al. [14] for whome porcine SIS membranes (Restore™SIS, US-FDA clearance) failed to repair human rotator cuff: histological examination revealed that the xenograft material was not a cell-free collagenous matrix but contained a multiple cell layers [12] and DNA [14]. Finally, the incomplete biomaterial inertness seems to be another cause of failure for porcine SIS (SurgiSIS, Cook Inc.) in alloplastic esophageal replacement [13]. An investigation addressed to evaluate if such relevant residual DNA remnants could comprise endogenous retrovirus sequences, as previously found in acellular porcine heart valve scaffolds [50], is still in progress. Further studies are performing in order to assess how these nucleic residues can be correlated to the failure of the SIS-derived material in high-strain environments [51].

Equine pericardium
EP xenograft patch is a relatively new product employed in various medical fields such as musculoskeletal, neurological and cardiovascular surgery [52][53][54][55][56][57][58]. It has also been described for use in repair of rotator cuff defects [59,60], as well as for the treatment of diabetic foot ulcers [61]. Finally patch of EP tissue were applied either for the coverage [17] or for the self-made manufacture of heart valves substitutes [62]. Like many of the xenogeneic tissues intended for clinical use, even the EP patches are supplied after GLU fixation.
While GLU treatment is the most commonly used crosslinking agent able to reduce tissue degradation and to preserve anatomic integrity, strength and flexibility, it represent also a potential calcific factor especially when the tissue is subjected to a high mechanical stress. In fact, the attempt to reconstruct arterial vessels or tendons with an EP patch, showed no positive results in condition of increasing blood pressure or mechanical loading. As evidenced by Yamamoto H and colleagues, dilatation of the EP tube graft did develop as a result of long-term exposure to arterial pressure [63]. Similarly, degeneration of an EP valve conduit was occurred much earlier than that of the standard bovine or porcine bioprosthetic heart valve substitutes (at 3 to 5 years after the intervention), with a constant increase of the pressure gradient across the conduit during the follow up [64].
The treatment with GLU, in addition, was not efficient to ensure a complete immunotolerance of the implanted xenogeneic tissue, reducing but not eliminating the alpha-Gal immunogenicity. Noteworthy, the immune response against these xenogeneic residues has been recognized, together with the presence of nucleic acids materials, as an additional factor leading to tissue mineralization [32]. The amount of available and reactive alpha-Gal epitope evaluated in EP patch resulted in our investigation to be 15% higher than in commercially bovine heart valve substitute [35]. In a hypothetical biocompatibility classification, this assessment puts the EP tissue well behind the bovine and porcine pericardium for high pressure cardiovascular and orthopaedic purposes. Conversely good results were obtained for venous conduits reconstruction [65], treatment of diabetic foot wounds [62] and tympanic perforation [66].

Conclusions
Commercially available decellularized SIS and GLU-fixed EP materials are currently employed for the treatment of several tissue damages sequelae in humans (sometime with questionable results), even in the absence of any publicly disclosed quantitative/qualitative assessment concerning the retention and the elimination of alpha-Gal xenogeneic epitopes. Likewise, the decellularization treatment is not monitored enough to grant the reduction of the calcific potential associated to the nucleic acid remnants retained within the matrix. Now a day there is a growing consensus about the strong relationship linking the failure of bioprosthetic grafts provoked by calcific occurrence and immunologic reactions to the lack of a reliable control tool providing quantitative information about the undesirable remainings within a tissue engineered scaffold before its clinical use. The effective removal or coverage of both cellular components (including the xenogeneic epitopes express in the cellular membrane residues) and nucleic acid materials debris should very likely become a benchmark for the future approvals of biomaterials intended for clinical use.