*3.3. Clinical E*ff*ectiveness of Skin Substitutes*

## 3.3.1. Effectiveness of Acellular Dermal Regeneration Templates

Dermal substitutes must guarantee correct regeneration of the dermis compartment [13]. After implantation, a dermal substitute has to allow a fast recruitment of endothelial cells in order to guarantee the correct take of a thin STSG. As a long-term aim, the skin substitute should guide a correct regeneration avoiding the formation of scar tissue. It has been recognized that an ideal dermal substitute must satisfy the following requirements [38]:


To date it is difficult to make a comparison between the efficacy of the different commercially available dermal substitutes because of the lack of long-term follow-up studies in which different dermal substitutes have been studied in parallel. Even where different clinical studies exist for each of the skin substitutes, it is difficult to make an objective comparison due to different factors. Notably, different studies use different methods for the evaluation of the success of the healing. Indeed, different scales exist to evaluate the features of the scar after implantation: Vancouver Scar Scale, Hamilton Burn Scar Rating, Patient and Observer Scar Assessment Scale, Manchester Scar Scale and Visual Analog Scale [121–124]. Moreover, the studies are conducted starting from different type of wounds (burns, ulcers, trauma, etc.). Finally, the experience of the surgeon plays a crucial role in the success of the take of both dermal regeneration template and STSG. This is highlighted by analyzing the results on the take of grafts coming from multicenter and single center studies: the former possesses higher degrees of uncertainty than the latter [124].

Nevertheless, it is possible to extrapolate some general indications showing the benefits and downsides of the different dermal substitutes. By focusing attention on partial- to full-thickness wounds, the gold standard treatment is the application of a partial- or full-thickness STSG. The application of an STSG is an autografting procedure and possesses two main issues: (i) the use of an STSG is an invasive technique because after the removal of large parts of healthy tissues, other damage is induced; and (ii) if the total burn surface is higher than 25% of the total body area, the donor site cannot provide enough tissue to cover the wounded area. For this reason, the majority of the clinical trials have been conducted by comparing the efficacy of a dermis substitute to the efficacy of the gold standard treatment. Among the acellular dermal substitutes, Integra®, Matriderm® and Alloderm® have been subjected to a systematic review [38] thanks to the availability of long-term follow-up studies starting from similar wounds (partial- or full-thickness wounds due to thermal burns). The comparison between the dermal substitutes and the control groups has been performed by analyzing primary outcomes (graft take, wound infection, scar quality) and secondary outcomes (donor site morbidity, convalescence of the patient, need to re-graft). Compared with an STSG, only Integra® showed a significantly lower take (*p* < 0.001). Concerning the infection rate, a value of 85% of infection in patients having a total burn surface of 45% and treated with Integra® has been reported [125,126]. However, other studies reported no difference in infection percentage between Integra® and the control group [26,126,127]. A 6% rate of infection has been detected in a study where Alloderm® was used [127]. Matriderm® did not reveal any difference of infection rate compared

with the controls [128]. Concerning scar quality, data only showed better scar quality for Integra® compared to that of an STSG. For all other models, no significant differences were detected in scar quality compared with the control groups. A relevant increase in donor site healing (lower donor site morbidity) has been detected in all studies [38]. This is because by using a dermal substitute to fill the wound bed, the thickness of the STSG can be reduced to 0.006 of an inch, compared to 0.01 of an inch for gold standard treatments.

With the exception of Integra®, few data concerning the mortality and length of stay of patients have been found [38]. No significant differences in rates of re-graft were reported between the dermal substitutes and control groups. Other studies have compared five dermal substitutes (involving Integra®, Matriderm® and Hyalomatrix®) in a preclinical pig model [110]. No difference was found at the end of the survey (six months after implantation) in terms of scar quality and healing properties. A difference was found in the evolution of the healing process. Integra® took more time for fibroblast and blood vessel influx and was still present, as a fragmented matrix, at the end of the survey. Hyalomatrix® showed a lower degradation rate than Matriderm®, but both had completely disappeared at the end of the survey. The quality of the scar assessed with the Vancouver Scar Scale was similar. A long-term follow-up revealed both histological and clinical outcomes in a randomized study using Integra® and Nevelia® [103]. Nevelia® is a porous degradable matrix of about 2 mm thickness made of stabilized native collagen type 1 from calf hides and a silicone sheet of about 200 μm thickness mechanically reinforced with a polyester fabric, and does not contain any chondroitin-6-sulphate GAG. Collagen is purified from calf hides from animals younger than nine months sourced from safe countries. Collagen is then cross-linked with a very low percentage of glutaraldehyde. Clinical results were evaluated through the healing time, Manchester Scar Scale (MSS) and Visual Analog Scale (VAS) up to three years. The differences in healing time between groups, pain and self-estimation was not statistically significant up to a one-year follow-up. Nevelia® showed early regenerative properties in terms of epidermal proliferation and dermal renewal at three weeks, compared with Integra®. Furthermore, Nevelia® showed a more evident angiogenesis vs. Integra®, evaluated as α-SMA immunohistochemistry. Differences in the MSS score were statistically significant at three years follow up in favor of Nevelia® group (*p* = 0.001), together with clinical outcomes. Histological and immunohistochemistry data showed that Nevelia® allows faster neo-angiogenesis and tissue regeneration with neo-formed tissue architecture closer to the physiology of the skin. This data confirms the importance of both vascularization and neodermis influx within the dermal substitute. Indeed, it is well accepted that the faster the vascularization, the higher the take of the STSG. Integra® contains a fraction of chondroitin-6-sulphate GAG that had two side effects: firstly, such a kind of GAG slows the influx of endothelial cells; secondly, the presence of such a high hydrophilic component may retain lesion inflammatory exudates. Furthermore, the high degree of cross-linking in Integra® masks both the adhesion site for fibroblasts and the proteolytic degradation sites, hampering the recipient cells' influx and the degradation of such exogenous ECM. Finally, it seems that rapid vascularization and presence of endogenous dermal cells (fibroblast and vascular network) may play a crucial role in the effectiveness of the dermal substitute.

#### 3.3.2. Effectiveness of Cellularized Skin Substitutes

The introduction of cellularized skin substitutes in the treatment of full-thickness wounds has been necessary in order to overcome some limitations of the acellular dermis. In particular, the presence of dermic cell lines (fibroblast or endothelial cells) serves to reduce the time for neodermis influx after implant. By using acellular dermal substitutes, surgeons have to wait at least four weeks for the implantation of the STSG in order to allow fibroblast and endothelial cells' influx. Without such cellular components, the take of the STSG is not possible. Moreover, the larger the delay before the STSG is implanted, the higher the probability of infection. Thus, the success of the regeneration process is strictly related to both "colonization" time and the quality of the neodermis formed in the porosity of the acellular dermis. A cell populated scaffold introduced in the wound bed aims at reducing such

colonization time as a primary task. We can highlight additional requirements that engineered skin substitutes have to possess compared with acellular substitutes:


To date, the use of tissue engineered products for the treatment of burns, although it has received broad scientific success, seems to be not convenient in terms of commercial outcomes. For this reason, many companies have focused their attention on the commercialization of engineered tissues for the treatment of chronic diseases: venous valve insufficiency, arterial diseases, diabetes, vasculitis, skin malignancies and blistering diseases. Regardless of the origin of the wound, the major aim is to reestablish the physiological conditions of the dermis layer. The cells that act to maintain the dermis are the fibroblasts [108], which synthesize and assemble collagen, elastin, and proteoglycans. When inflammation occurs, fibroblasts migrate to the wound site, attracted by bFGF and TGF-β secreted by inflammatory cells and platelets, where the fibroblasts are stimulated to replicate, migrate into the wound, and secrete IGF-1, bFGF, TGF-β, platelet-derived growth factor, and KGF, enabling fibroblast–keratinocyte interaction. When the wound is chronic, the continuous inflammation state induces a premature and stress-induced cellular senescence of the fibroblasts. Moreover, a decreased proliferative potential, impaired capacity to react to growth factors and abnormal protein production is observed. When the percentage of senescent cells in the defect is greater than 15%, wounds are described as hard to heal [14,15,129,130]. The use of Dermagraft® for the treatment of non-healing ulcers has been demonstrated in different clinical studies [130–132]. Summarizing the results, it is possible to establish that the percentage of healing using Dermagraft® ranged from 50% to 71.4%. The complete closure of the wound can be obtained only for ulcers with 12 months of duration or less.

The most popular bi-layered skin substitute (containing fibroblasts and keratinocytes) is Apligraf®, which has been studied since 1999. It has been demonstrated that use of such a bi-layered engineered skin is an effective treatment for ulcers of greater than one-year duration, with a percentage of wound closure of about 50%. Moreover, the number of osteomyelitis and lower-limb amputations were less frequent in the Apligraf® group. The advantage of dermal–epidermal substitutes is the presence of the living epidermal layer that avoids the use of gauzes and two-step procedures. This reduces the risk of infections and improves the healing process due to the presence of fibroblast–keratinocyte cross-talk, which plays a crucial role in the healing process of deep wounds. One of the limitations of the aforementioned categories of skin grafts is the presence of allogenic cell lines, which contain an associated biological risk. This limitation is being overcame by developing models containing autologous cell lines. Two models that use autologous fibroblasts and keratinocytes are denovoDerm® and denovoSkin®, which are under clinical trials at University Children's Hospital of Zurich [71,133–135]. Finally, to deeply investigate on both FDA status and the effectiveness of current skin substitutes the reader can refer to "FDA" and "clinicaltrial" databases by searching for the desired skin model [136,137].

#### **4. Advanced Bioengineered Skin Equivalents: A Future Perspective**

#### *4.1. Pre–Vascularization of Dermis Substitutes*

The treatment and the evolution of deep wounds due to thermal burns is schematized in Figure 2A-D. After the debridement of the wound, the bed is filled with a DRT supporting an artificial layer of silicone-based epidermis. After a period of four weeks, the epidermal layer is detached and an autologous STSG is applied. In a clinical study that used Integra® as the DRT, 20 patients presenting deep wounds were treated using the procedure described in Figure 2A-D. The evolution of the wound was analyzed by means of histology, immunocytochemistry and the Vancouver Scar Scale [20]. It was observed that the vascularization of the DRT played a crucial role in the take of the STSG. For instance, if the STSG was applied after two or three weeks, the take rate was very low. On the contrary, if the STSG was applied after the fourth week, the take increased up to 95%. Histological and immunostaining analyses demonstrated that at two weeks the vascularization of DRT was poor but increased four weeks after implantation. These data suggest that vascularization of the DRT and the take of the STSG are strictly related [20].

Other relevant findings concern the evolution of the dermis compartment over the time. Weekly histological investigation revealed that influx of exudates and host fibroblasts occurred during weeks one–two. At three weeks, the influx of endothelial cells and the synthesis of immature extracellular matrix components by fibroblasts began. During week four, the formation of a capillary network (Figure 2E) was observed. After the application of the STSG, the wound continued its evolution: at week six a well-organized capillary network was observed, but the dermis–epidermis interface presented no rete ridge profile; at month three, a layer of endogenous collagen network was observed underneath the STSG; after two months, the wound was completely repaired but the neo-tissue was different from the healthy skin. Finally, the complete substitution of the initial DRT with the neodermis occurred at two years post implantation. Even though the patients recovered partial mobility of the damaged parts, it was observed that the repaired zone showed an impaired pigmentation, the mechanical properties between healthy and repaired sites were different, and the organization of the collagen network of the neodermis was different than that of the collagen in the healthy dermis. Finally, neither elastin nor adnexa were present, and differentiation of fibroblasts in myofibroblasts was observed. On the basis of such findings two main issues affecting the DRT emerge: (i) the lack of vascularization [22,23]; and (ii) the limited capability in inducing regeneration instead of repairing processes [135]. The take of the STSG has huge implications related to the repairing process, patient mortality, and morbidity and healthcare costs. Indeed, a low take percentage increases the number of re-grafts and the risk of infection by causing either the death of the patient or an increase of hospitalization time in case of morbidity. To increase the take of STSGs, new emerging strategies involve the use of pre-vascularized DRTs [22,23,33,34,36]. By seeding a DRT with adipose tissue-derived microvasculature fragments, a faster vascularization after implants was observed [138]. Complete reperfusion of the DRT occurred at day six. The percentage of the take was high if the STSG was applied just after day six, indicating that reperfusion rather than simple vascularization played a crucial role in the take. These data suggest that pre-vascularization of the DRT can contribute to shortening the timeframe needed for the application of an STSG. On the other hand, a one-step surgery, which may decrease the number of surgical operations, cannot be performed yet. To do this, not only vascularization, but also fast reperfusion should be promoted.

#### *4.2. Engineered Skin Composed of Fibroblast-Assembled Extracellular Matrix*

The lack of vascularization at the moment of implantation has been recognized as the main issue affecting the take of the STSG. No studies have been performed yet on the role that the extracellular matrix comprising the DRT may play on both vascularization and longtime dermal remodeling [17]. The dermis compartment of the totality of the skin substitutes (either cellularized or acellular) are composed by exogenous extracellular materials, i.e., not assembled by the fibroblasts of the patient. This should represent the limitation of the currently available tissue engineering skins. Indeed, exogenous matrices, even though of natural origins, cannot fully replicate the complexity of the living dermis. This may ultimately compromise the repository and regulatory role that the native cell-assembled extracellular matrix plays [16–18]. Such a mismatch between an exogenous material and the living dermis may be responsible for the impaired repair process at both cellular

and extracellular levels. Firstly, because the repository and regulatory role of the native ECM is depressed, the growth factors secreted by fibroblasts are not correctly presented to other cell types (e.g., keratinocytes and endothelial cells) neither in space nor time, generating a possible "mistake" in cell–cell signaling [16,17]. This could explain both the delay in the vascularization time and the delayed formation of the rete ridge profile at dermal epidermal interfaces [20,26,72]. As confirmation of this, in vitro tissue engineered skin made by exogenous natural hydrogels (i.e., collagen, fibrin, etc.) presents a flat dermal–epidermal interface. On the contrary, if epithelial cells are grown on a fibroblast-assembled ECM, it is possible to observe a rete ridge profile with spontaneous formation of epithelial invagination and follicular-like structures (Figure 3C) [72], which are typical of the physiologic dermal–epidermal cross-talk mediated by the extracellular matrix [72]. The lack of endogenous ECM-mediated signaling may also explain the absence of both cutaneous adnexa and nerve endings in repaired deep wounds [20,26,129]. Secondly, when fibroblasts colonize the inner porosity of the DRT, they produce an immature extracellular matrix with a degree of assembly much lower than the degree of assembly of the surrounding healthy dermis. Such an immature protein network is not able to withstand the traction forces of the fibroblasts [74,110], generating a different architecture of the collagen fibers in the wound compared to the healthy dermis [15,19,21,77]. Macroscopically, these phenomena generate a portion of the cutis possessing different mechanical properties, different pigmentation, absence of sensing properties and high contracture, provoking both severe functional and aesthetic concerns.

To overcome such limitations, a tissue engineering strategy to produce a human dermis substitute composed of a fibroblast-assembled extracellular matrix has been developed [73,75]. The innovative idea of such strategy is to let human fibroblasts producing their own ECM in vitro. This process provides the possibility of modulating the properties of the cell-synthesized ECM, in order to obtain a final dermis having both composition and assembly degree of the collagen network relatively similar to those present in vivo. Moreover, no exogenous materials are present. This bottom-up tissue engineering strategy starts with the fabrication of dermal building blocks obtained [81] by seeding human fibroblasts in porous gelatin microspheres (Figure 3A). It has been demonstrated that by optimizing the culture conditions, the fibroblasts can produce their own extracellular matrix. Such building blocks, named Dermal-μTissues, were subsequently molded and packed in maturation chambers where both cell–cell and ECM–ECM interactions took place, leading to the formation of a continuum, up to 2 mm thick, made of an endogenous dermis containing fibroblasts and gelatin microspheres. By modulating the stiffness and the degradation rate of the gelatin microspheres and by engineering the dynamic culture conditions (Figure 3), it was possible to obtain fine control over the maturation status and assembly of both collagen and elastin networks [74,91]. During the duration of the process (approximately five weeks), gelatin microspheres were degraded by protease digestion and the final tissue, named EndoDermis, was completely made up of fibroblasts embedded in their own extracellular matrix (Figure 3A). Interestingly, the collagen network was characterized by a stiffness and degree of assembly similar to that featuring the human skin. In the ECM elastin, hyaluronic acid, fibronectin and elastin were also present (Figure 3B-F). In order to produce a pre-vascularized endogenous human dermis model, human umbilical vein endothelial cells (HUVECs) were seeded on the EndoDermis and it was allowed to form an interconnected capillary network [34] that occurred within three weeks (Figure 3D, E). At the best of our knowledge, other than a capillary network, such an engineered DRT is the first model completely formed by a fibroblast-assembled extracellular matrix [34]. After subcutaneous implant in a nude mouse model, fibroblasts and their own ECM (the neodermis) were already present and well-assembled. Thus, no additional time is required for fibroblast influx and neodermis formation. The only phenomenon required is the anastomosis and perfusion of the engineered capillary network. This was shown to occur within seven days of implantation (Figure 3H). Although further investigations are currently being conducted of a more representative wound model, such data are encouraging. In addition to vascularization, which has been recognized as a critical issue [22,23,25,33,34] affecting the effectiveness of a DRT, the described tissue engineered strategy

allows the fabrication of a DRT composed of a native extracellular matrix starting from a small number of fibroblasts derived from the patient. In this way, the risks associated with the allogenic nature of the cells and the impaired ECM assembly during wound closure, can be drastically reduced. According to this idea, the formation of severe scars can be reduced.

#### **5. Discussion and Conclusions**

Dermal regeneration templates and tissue engineered skin [40] has been reviewed in the light of their effectiveness in guiding the closure of deep wounds toward a regeneration process rather than a repair process [38]. By analyzing the literature, no strategies are currently available that are able to completely restore the whole functionality of the reconstructed part, including pigmentation, mechanical properties, adnexal structures and sensing properties. In other words, the formation of severe scarring still represents a concern in the field of skin reconstruction. Many advances have been made to limit the use of thick STSGs. Indeed, using a last generation DRT, the thickness of STSGs can be reduced to 0.006 of an inch, compared to 0.01 of an inch for gold standard treatments. Furthermore, the take of STSGs has been improved by promoting the vascularization of the DRT [22,23,25]. It has been observed that a DRT composed of non-cross-linked matrices promotes the invasion of fibroblasts and endothelial cells, increasing the take of the STSG compared to the case of cross-linked matrices [9]. On the other hand, cross-linked matrices are able to better withstand contracture during the neodermis remodeling process, due to their superior stiffness. To hasten neodermis growth after implantation, tissue engineering strategies aim at cellularizing the dermis compartment prior to implantation with either fibroblasts or endothelial cells. Pre-vascularization has been shown to improve the take of the STSG and the reperfusion of the dermal bed. This aids the oxygenation of the zone and the removal of waste. The presence of fibroblasts serves to shorten the migration time of fibroblasts from the recipient and to also promote the synthesis of the neo-ECM in the wound bed [26]. Nevertheless, once assembled, the final ECM is still far from its physiologic condition. By analyzing the composition of the bioengineered skin models, it is possible to highlight that they are characterized by a common denominator: cells are always embedded in exogenous matrices (i.e., not synthesized by fibroblasts). This can represent an issue, since native ECMs possess a specific arrangement of moieties, which regulates the cross-talk between fibroblasts and keratinocytes that ultimately leads to the formation of skin adnexa and skin appendages [16–18]. The in vitro fabrication of human bioengineered dermis composed of a fibroblast-assembled ECM incorporating a vascularized network may provide a means of overcoming the scarring process. In this regard, it has been shown that a tissue engineering strategy allows the control of the assembly of the ECM produced by fibroblasts. By modulating the process variables, it is possible to produce an engineered dermis possessing composition, organization, and signal presentation capabilities relatively similar to the native dermis [19,34,35,72,74,81]. This may led to different benefits in the scenario of skin regeneration: (i) To date, the neodermis takes at least two years to form; by introducing the reconstructed patients' dermis in the wound bed, possessing the final architecture and composition at the moment of the implant, no further time will be required for neodermis formation; (ii) by controlling in vitro the degree of assembly of the ECM, it will possible to decrease the differences between repaired and healthy tissues; and (iii) the spatial organization and the functionality of ECM components is not compromised by external factors (e.g., chemicals or physical treatments) and communications among all cell types (e.g., fibroblasts, keratinocytes, nerve endings and macrophages) are correctly orchestrated.

**Author Contributions:** F.U. collected the majority of the data from the literature, performed a critical analysis of data retrieved from both literature and FDA-CRINICAL TRIAL database, then it wrote the manuscript with inputs from all authors (C.C., G.I. and P.A.N.). C.C. collected data from bottom up tissue engineering skin models and produced histological images and bioengineered skin samples described in the in the Figure 3. He participated to critical discussions during the editing phase. G.I. worked on the arrangement of paragraphs togheter with F.U. and contributed to the development of figures and participated to critical discussions. P.A.N. was the supervisor of the project; he gave inputs about the organization of the manuscript, he gave a crucial contribution concerning the arrangement of discussion in the scientific framework.

**Acknowledgments:** The authors would like to acknowledge Alessia Larocca, for her contribution in the preparation of the Graphical Abstract.

**Conflicts of Interest:** The authors declare no conflict of interest.
