Decellularized xenografts in regenerative medicine: From processing to clinical application

Decellularized xenografts are an inherent component of regenerative medicine. Their preserved structure, mechanical integrity and biofunctional composition have well established them in reparative medicine for a diverse range of clinical indications. Nonetheless, their performance is highly influenced by their source (ie species, age, tissue) and processing (ie decellularization, crosslinking, sterilization and preservation), which govern their final characteristics and determine their success or failure for a specific clinical target. In this review, we provide an overview of the different sources and processing methods used in decellularized xenografts fabrication and discuss their effect on the clinical performance of commercially available decellularized xenografts.

the lack of mechanical stability. 14 All these limitations can be potentially addressed by the appropriate selection of a clinical indication specific biomaterial.
An ideal biomaterial for tissue engineering applications should guarantee cytocompatibility, maintain appropriate/desired cellular functions and phenotype for the specific application, induce tissue growth and provide mechanical support until it is absorbed and replaced by natural extracellular matrix (ECM). Synthetic biomaterials can be tailored to obtain desired topographical, mechanical, chemical and morphological properties 15,16 ; however, they do not support cell attachment and bioactivity due to the lack of functional domains/cell recognition sites and often induce foreign-body response and acute inflammation. [16][17][18] On the other hand, natural biomaterials present biological compatibility and functionality due to their cell recognition motifs that promote cell adhesion, proliferation and differentiation and advances in chemistry through provision of elegant crosslinking systems offer control over mechanical stability and biodegradation. 19,20 However, natural biomaterials are of inconsistent composition and high variability as a function of source or batch. 21-24 Independently on whether the biomaterial is natural or synthetic in origin and despite the significant strides in engineering, currently available scaffold fabrication technologies poorly imitate the in vivo architecture, mechanical properties and compositional complexity of native tissues (Figure 1). Considering that decellularized grafts closely imitate the biophysical, biochemical and biological milieu of the tissue to be replaced, they can overcome all the aforementioned limitations of natural and synthetic scaffolds and ultimately provide functional reparative therapies, as long as issues associated with immune rejection and availability, in the case of allografts, are addressed.
Undeniably, tissue grafts are an inherent part of tissue engineering and regenerative medicine with numerous products being clinically available for a diverse range of clinical indications (Table 1).
Herein, we discuss advancements and limitations in xenograft development and how processing steps (eg decellularization, crosslinking, sterilization) affect their properties and differentiate success from failure in their clinical applications.
F I G U R E 1 Histology and immunohistochemistry analyses of a collagen-based biomaterial and three xenografts clearly illustrate the superior biofunctionality of the latter, as judged by high levels of compositional and structural biomimicry. Scale bars: 200 µm

| PRO CE SS ING OF TISSUE G R AF TS
Each processing step in the developmental cycle of a tissue graft can influence its mechanical, chemical and biological features, determining the success or failure of the implant. [25][26][27] It is therefore an active field of development, as evidenced by numerous registered processing protocols (eg Tutoplast ® (RTI Biologics), 28 BioCleanse ® (RTI Surgical), 29 dCELL ® Process (Tissue Regenix), 30 Tecnoss ® (Tecnoss ® ) 25 ) that is also well-regulated (ie FDA provides guidance on medical devices containing materials from animal sources, 31 any product related on xenotransplantation in humans 32 and specific documentation for registering newly developed materials of animal origin 33 ). The general steps necessary to manufacture a tissue graft and the associated quality control checkpoints are sequentially summarized in Figure 2. In this section, we provide an overview of these processing steps.

| Donor and tissue selection
Porcine and bovine tissues are primarily used in biomedical field, although studies have been carried out using also equine, 34,35 ovine, 36 caprine, 37 kangaroo, 38 buffalo 39 and ostrich. 40 The properties of the graft depend not only on the species, but also on the breed, age 41 and tissue section 42,43 from where the graft is collected. Among all animal sources, the pig is preferred due to its abundant availability, similar size to human tissues, relative low cost of breeding and extended knowledge of its physiology. [44][45][46][47][48] Bovine tissues, although have shown similarities to human tissues, 48,49 in general, their size in adult animals is not appropriate for use in humans and breeding associated expenditure significantly increases the value of goods, creating reimbursement issues. Advancements in molecular biology and genetic edition have allowed for the development of genetically modified animals as source of organs or tissues. Most of these studies are carried out in domestic pigs to prevent the immune rejection of the grafts, 50 where site-specific nucleases are employed to prevent the presence of the Gal epitope in the donor cells by inactivating the α-1,3-galactosyltransferase enzyme. 50,51 Several studies have demonstrated safety and efficacy in pre-clinical models for skin, 52,53 liver, 54 cornea 55 and kidney 56 between genetically modified pigs and non-human primates (in combination with immunosuppressive drugs) and clinical trials are expected in the coming years.
The age of the animal can also influence the characteristics and properties of the tissue graft. For example, the level of crosslinks is age-dependent 57 and influences, among others, the thermal stability and mechanical properties. In addition, the age influences cellbinding sites, 58,59 impacting on cellular behaviour and phenotype in vitro and in vivo. 60,61 For instance, in pig pancreas islets xenotransplantation for the treatment of diabetes, islets from adult pigs present higher resistance to in vivo degradation and higher neovascularization potential due to the presence of a mature ECM. 41 Small intestinal submucosa (SIS) has also been shown to present different mechanical, structural and biological characteristics, as well as M2 macrophage immune response and remodelling potential as a function of the stage of maturity of the pigs. 62,63 Screening of the donor is also necessary before harvesting a graft. Although screening of human patients is relatively easy due to availability of medical records, 64 this safety net is not necessarily available in animal-derived grafts that harbour high risks of infection of multiple pathogens, 65,66 but not so much of viral contaminants. 67 The FDA has stablished guidelines on infectious diseases in xenotransplantation to prevent and control interspecies disease transmission with full instructions and precautions that should be carried out during animal breeding and tissue harvesting. 68

| Decellularization, crosslinking, sterilization and preservation
Once the xenograft donor and tissue source have been chosen, its processing follows a sequential order that includes decellularization, crosslinking (optional), sterilization and preservation, using a variety of techniques and agents ( Table 2). Decellularization of tissue grafts should have minimal effect on the integrity, microstructure, composition and biological activity of the ECM, while removing all cellular material and reducing antigens that could trigger immune response. 69 Cellular remnants contain domains that are recognized as foreign matter and trigger immune response 70,71 and, although ECM components are highly preserved among species, 72-76 decellularized ECM components can also elicit immune response 77 that can induce macrophage polarization to M1 or M2 phenotypes. 70 Classically, the presence of cells 78 and/or cellular material 71 promotes M1 inflammatory response, 79,80 whereas effectively decellularized scaffolds are related to M2 phenotype. [80][81][82] However, recent studies suggest that decellularized scaffolds promote a combined M1/M2 macrophage phenotype, involving adaptative immunity 83,84 and triggering remodelling. A typical decellularization process involves the lysis of cellular matter with physical means or chemical agents, followed by separation of cellular matter from the ECM with enzymes and finally removal of cell matter and debris with detergents. 85 Crosslinking is a process which ends with an interconnection between molecules. Although it occurs in vivo as a post-translational modification of proteins via enzymatic and non-enzymatic mechanisms, the native crosslinking of tissue grafts may be insufficient and previous decellularization process may compromise ECM's mechanical properties and stability upon implantation. Therefore, exogenous crosslinking can be used to increase mechanical properties and the reabsorption time in vivo. 86 However, crosslinking decreases the number of available recognition cues for cell attachment and degradation products can elicit cytotoxicity 87,88 and calcification, 89 particularly those elicited by chemical agents. 90,91 This has motivated research into natural agents, [92][93][94][95] bearing always in mind that the ideal crosslinker should be economical, effective and with minimal side effects.

TA B L E 1 (Continued)
F I G U R E 2 Sequential processing of any tissue xenograft from any source (exemplified with pig cardiac tissue), along with quality control check points The effects that the preservation and the duration of storage have on a tissue graft are commonly overlooked and not specified in the protocols; however, they can affect the structure and therefore the properties of a decellularized graft. 105 The most extended techniques for the preservation of acellular tissue grafts are freeze drying and cryopreservation. Freeze drying results in stable materials that can be further sterilized with physical irradiation methods or ETO. However, during the process crystal nucleation occurs, which may damage the ECM structure, thus, parameters such as temperature and cooling speed should be closely monitored and appropriately optimized. 105,106 Lyoprotectants that protect the tissue from the crystals growth can be used, although they may also affect the ECM structure and its biomechanical properties. 107 Cryopreservation is a cooling process in wet state in the presence of cryoprotectants. Cryopreservation has been shown to preserve the functionality of tissue grafts, 108,109 but cryoprotectants may induce a cytotoxic side effect. 110 In the processing of a tissue graft, once each step has been finalized, the assessment of its efficacy and effects on the material have to be assessed. Efficacy of decellularization is generally assessed through histology ( Figure 1) and DNA quantification, where 50 DNA ng/mg dry tissue is considered a safe threshold. 111 The degree of crosslinking can be calculated by quantifying the free amines or denaturation temperature. 112 Counting of colony-forming units (CFU)/ ml after sterilization can be employed to calculate the reduction on the number of viable microorganisms. 113 Also, effects of the processing steps on the ECM structure and the mechanical properties of the grafts must be analysed, followed by classic in vitro and in vivo compatibility assays and specific assays for the specific future application of the tissue graft.

| XENOG R AF TS IN CLINI C AL INDIC ATIONS
Many xenografts are commercially available for various clinical indications. Table 3 summarizes in vitro, in vivo and clinical data that have been obtained to-date with commercially available xenografts.
In this section, we discuss advances and shortfalls of xenografts per clinical indication.

| Soft tissue
Porcine dermis is extensively used in hernia and abdominal wall repair. 114  Clinical data: Safety for aortic valve replacement in 686 patients reported in a prospective non-randomized multicentre study, although bleeding rates increased in the long-term follow-up Bio-Oss ® [197] - [200,201] In vitro: Promoted secretion of VEGF by periodontal ligament cells, but in a lower extent than other xenogeneic bone grafts from porcine and equine origin Clinical data: Increased osteogenesis and width of the alveolar process alone or in combination with autogenous tissue, with a high (96%) survival of dental implants  [198] [199] In vitro: Promoted VEGF production of periodontal ligament cells and angiogenesis in endothelial cells in a higher extent than a bovine bone xenograft In vivo: Lower inflammatory reaction in rats muscle implantation than an alloplastic material.

386]
In vitro: Presented higher mechanical properties and resistance to enzymatic degradation than other tissue sources and non-crosslinked matrices. Supported the adhesion and invasion of MSCs, fibroblast, tenocytes, etc, but at a lower extent than other non-crosslinked matrices, cytotoxic events related to crosslinking. Elicited an inflammatory response as per cytokine production by macrophages in vitro In vivo: Subcutaneous models showed poor cell invasion, remodelling and neovascularization, higher inflammation and immune response that non-crosslinked matrices. Longer times of resorption than other non-crosslinked matrices (up to 24 wk in rats). Hernia models in rodents and rabbits showed optimal mechanical properties performance, although poor resorption and fibrotic tissue formation. Similar data were obtained in tendon and vascular patch models. Skin regeneration models in rats revealed a very modest performance due to poor resorption and epithelization Clinical data: Scattered results observed in pelvic wall repair (ie positive patient satisfaction and graft performance compared to synthetic substitutes after 1-year follow-up versus relation to complications and recurrence) and in hernia repair of contaminated fields (ie positive outcomes and no recurrence versus 50%-88% recurrence and 37% rate of infection). Treatment of massive tears in rotator cuff showed improvement in pain relief and shoulder functionality in small studies (5 and 10 patients). Breast reconstruction small case series showed an acceptable outcome in skin-sparing mastectomy regarding patient satisfaction scores and absence of complications. In eyelid repair, it was related to a higher rate of complications than other xenografts and materials provoked by its stiffness and low resorption [ [177][178][179]182] In vitro: Elastic modulus comprised between human rotator cuff values, but lower strength and strain. Lower mechanical properties than dermal derived grafts in both tensile tests and suture pull-out test. Lower cytocompatibility with tenocytes than non-crosslinked dermal grafts.
In vivo: Related to a M2 polarization and remodelling in an abdominal wall model in mice, although other studies in rabbits and mice also report an inflammatory reaction to the material due to ineffective decellularization, which can be decreased with the use of autologous cells. In tendon repair models (eg rotator cuff in rabbits and lambs), showed an initial inflammatory response and a later complete resorption, but without recovery of mechanical properties Clinical trials: Negative results were obtained in clinical trials; the material did not improve the healing and mechanical functionality of massive nor moderate rotator cuff tears, with a high rate of complications like inflammation and immune reaction [ 116-119, 122,161-166] In vitro: High rate of structure preservation and similar resistance to enzymatic degradation than native tissue, although lower resistance than other dermal crosslinked xenografts. Presence of soluble factors able to promote cell proliferation in a higher extent than other commercially crosslinked xenografts (dermis) and allografts (dermis).

SJM Pericardial Patch
Lower activation and production of inflammatory cytokines by mononuclear cells elicited than crosslinked dermal xenografts. Resistance to bacterial penetration due to compact structure In vivo: Lower inflammatory response and faster degradation and remodelling than crosslinked dermal xenografts in subcutaneous models in rodents and pigs. In hernia models in rats, it has shown a positive performance regarding integration, low inflammatory response and the mechanical properties of the hernia repair, which was confirmed in an abdominal wall repair in monkeys In vivo: Encapsulation of the tissue was observed in an abdominal wall repair in rats, although without an inflammatory reaction elicited by other crosslinked xenografts compared. It has been effectively used as a MSCs delivery system in an infarcted heart model in mice, improving remodelling and angiogenesis TutoBone ® -- [190] Clinical data: A 10-year retrospective study including 556 patients reported inferior results to autografts in terms of cervical fusion, but similar to other alternative options and at a lower cost Tutomesh ® - [392,398] [172] In vivo: Lower collagen deposition and mechanical properties in a rabbit ventral hernia repair than a dermal xenograft. In a rat Achilles tendon repair, it showed capability to integrate and the absence of host response, but with no functional nor mechanical assessment.
Clinical data: Breast reconstruction in 24 patients supported the safety of the material and technical efficiency, although post-operative seroma formation was reported as risk Veritas [130,388] [388] [171] In vitro: Studies on biochemical and biophysical properties have shown low ECM structure preservation, low thermal stability and low resistance to enzymatic degradation In vivo: Moderate inflammation in a full-thickness abdominal defect in monkeys (higher than a dermal xenograft but lower to crosslinked xenografts), but a fast resorption Clinical data: Retrospective multicentre study on 54 patients showed low rate of complications, at a similar or lower level than those observed in allogeneic dermal matrix products [ 182] In vivo: In a ventral hernia rabbit model, similar results in integration when compared to other two non-crosslinked dermal xenografts, although it did not show the optimal performance regarding collagen deposition and mechanical properties Clinical data: Higher pain relief and functionality recovery in rotator cuff massive tears than those treated with porcine SIS or equine pericardium in a study including 22 patients XenMatrix™ [130,388] [388] -In vitro: Lower preservation of ECM structure and resistance to collagenase degradation than other dermal xenografts and native tissue, and higher enzymatic degradability than crosslinked xenografts In vivo: Severe inflammatory response by the host in an abdominal wall defect in monkeys

TA B L E 3 (Continued)
also considered as a suitable and safe material for hernia repair in clinics, 126,127 matching the performance of synthetic meshes. 128 However, it has not shown suitability when employed as abdominal wall reinforcement in challenging scenarios, offering poor mechanical resilience ending in the discomfort of the patient and complications. 129 This could be related to its lower mechanical properties and resistance to enzymatic degradation, when compared to other tissue sources like dermis or crosslinked xenografts. 130 Xenografts are also extensively used in wound treatment and skin replacement. 131 are also customarily employed in wound healing, where they have shown improved and accelerated healing of ulcerous wounds [147][148][149] and have showed similar or superior performance to biomaterials 150,151 and allografts 152 and at a substantial lower cost. 153 Also, ovine forestomach, 154 equine pericardium 155 and even fish skin 156 have shown promising results in small clinical trials, but further investigation is required to demonstrate safety and efficacy.
Despite all these positive patient outcomes, all studies agree that randomized clinical trials with larger number of patients are needed to further support the use of xenografts in soft tissue repair.

| Tendon, bone and dentistry
Commercially available xenografts in tendon regeneration are used as augmentation systems. Tendon augmentation with porcine SIS has been related to failure in clinical trials, [177][178][179] [185][186][187] possibly attributed to efficient decellularization and/or crosslinking protocols and selecting a tissue with appropriate composition and mechanical resilience. In any case, overall, no definitive tendon augmentation device seems to be available. Although xenografts play a crucial role in irreparable defects augmentation, in moderate to large injuries, they have not achieved a satisfactory outcome.
Bone xenografts provide an optimal microenvironment for cell adhesion, proliferation and infiltration in vitro and de novo bone generation in vivo. This osteoinductive effect is related to the preserved micro-and macro-structure of the decellularized bone together with the partial preservation of ECM components. 188 Despite their high availability, low cost and good mechanical and osteoinductive properties, only a few bone xenografts are available, which have shown limited positive clinical results, 189,190 and as a result, they are rarely employed in orthopaedics. 191 Nonetheless, advances in materials sciences and tissue engineering have resulted in the development of composite materials combining xenogeneic mineral matrix, synthetic and/or natural polymers, which have been tested positively in clinical trials. [192][193][194] In dentistry, xenografts are used as bone-filling materials, with data to-date showing superiority in clinical outcomes even over autologous treatments. 195 These materials are normally deproteinized with high temperature processing, maintaining the mineral component and microstructure of the bone. 196 Bovine, porcine and equine in origin grafts have shown positive results in vitro, 197 in pre-clinical models 198 and in clinical setting. [199][200][201] Porcine non-crosslinked dermis has also been used successfully as augmentation/contention system in clinical trials, 202,203 largely attributed to its integration with the surrounding soft tissue, which prevents the second operation that synthetic materials require for removal. 204

| Cardiovascular
In cardiac graft replacement or patching, decellularized porcine SIS is one of the most employed material, 205 but only few clinical studies can be considered in a class III relevance, and the reported class IV correspond to case series or small trials. 206,207 In addition, high heterogenicity in responses has been observed, together with considerable complications, such as inflammatory response 208 and/or graft failures. 209,210 These complications were often related to high pressure conditions, which could be indicative of the low performance due to insufficient mechanical properties of the source of tissue, as observed in vitro. 205 Therefore, crosslinked materials with enhanced mechanical properties, like bovine pericardium, that include processing steps preventing crosslinker-related complications (eg calcification 211 ) have been developed, which have shown positive short-term results in paediatric cardiovascular applications, 211-213 but long-term assessment is required.
Vascular replacement xenografts (eg crosslinked bovine vein, bovine ureter) have mainly resulted in failure. [214][215][216] These poor outcomes were related to inappropriate processing, which resulted in insufficient mechanical properties to support the pressure of the circulatory system 217 and low antigen removal. 218 Having said that, recent reports on bovine artery crosslinked with starch dialdehyde (as opposed to GTA) have shown positive results as haemodialysis access 219 or as lower extremity bypass. 220 However, their implantation is not a generalized procedure, as their performance has not improved synthetic grafting. Further efforts in the processing technology and recellularization of the grafts are needed to improve the current outcomes. 221 In the field of valve replacement, the use of the stented valves is a common procedure, which utilizes metal/polymeric stents and processed animal tissue (eg bovine or porcine cardiac tissue processed with glutaraldehyde and anti-calcification solvent

| CON CLUS I ON S AND FUTURE PER S PEC TIVE S
The low availability of autologous and allogeneic tissues, coupled with advances in decellularization, crosslinking, sterilization and preservation, have made numerous, primarily, porcine and bovine xenografts commercially and clinically available for a diverse range of tissue engineering and regenerative medicine applications. In general, non-crosslinked grafts (in particular, grafts from young animals) demonstrate better resorption for tissues that do not require mechanical resilience. Crosslinking, although significantly improves mechanical properties, is often associated with immune response and calcification, imposing the need for development and assessment of new methods. As immune rejection remains the major concern of xenografts in clinical practice, genetically engineered pigs with reduced immunogenicity could become the ideal source for xenografts in the years to come. Overall, although pre-clinical and clinical studies have demonstrated, in most cases, safety and efficacy, the field urgently requires randomized double-blinded clinical trials to safely conclude on the potential of a specific xenograft for a specific clinical indication. Considering the, unmatched by human-made devices, physicochemical and biological similarity of xenografts to human tissues, we believe that xenografts will continue gaining pace in modern regenerative medicine.

CO N FLI C T O F I NTE R E S T
https://www.hhs.gov/guida nce/docum ent/fda-anima l-produ ctsdatab ase-data-entry -form 34. Mulder G, Lee DK. A retrospective clinical review of extracellular matrices for tissue reconstruction: equine pericardium as a biological covering to assist with wound closure.