Human stem cells prevent flap necrosis in preclinical animal models: A systematic review

Background and Aim: Adipose-derived mesenchymal stem cells (ADSCs) have been proven effective to prevent distal skin flap necrosis in preclinical models. However, to appropriately translate these findings to clinical trials, the effect of ADSC of human origin (hADSC) needs to be evaluated. We hypothesize that hADSC treatment is as effective as animal ADSC treatment at preventing distal skin flap necrosis in animal flap models. Methods: Three databases were inquired on August 17, 2020, to evaluate the necrotic flap area after using hADSCs in animal models of ischemic flaps. No publication status or dates were considered. Studies were included if they used hADSCs, measured the surviving or necrotic skin area of flaps, used animal models, and were in English. Studies were excluded if they did not use cells of human origin. The flap survival or necrotic area, perfusion, capillary density, vascular endothelial growth factor secretion and HIF-1α expression were extracted. Results: Ten studies met inclusion criteria. The mean absolute risk reduction (ARR) in necrotic skin area was 22.37% (95% confidence interval [CI] 16.98-27.76%, P<0.05) for flaps treated with animal ADSCs and 18.04% (95% CI 2.74-33.33%, P<0.05) for flaps treated with hADSCs. The difference between mean ARRs was not statistically significant (4.33%, 95% CI – 34.47-43.13%, P>0.05). Conclusion: Human ADSCs prevent skin flap necrosis to the same degree as animal ADSCs in rodent and rabbit flap models. Relevance for Patients: This review found that adipose-derived stem cells of human origin are equally effective at reducing the risk of surgical flap necrosis in preclinical models of small animals as autologous animal cells. The findings in this review should encourage researchers to use human adipose-derived stem cells in animal models of ischemic flaps to accelerate their translation into clinical trials and, eventually, surgical practice. The low immunogenicity of these cells should be leveraged to gain insight into the effects of the products that will be ultimately administered to patients. Furthermore, human adipose-derived stem cells’ pro-angiogenic mechanism of action sets this therapy as a promising preventive measure for flap necrosis.


Introduction
Flaps are routinely used in plastic surgery to cover tissue defects. Although rare, irreversible ischemia leading to necrosis can occur in the distal portion of random pattern skin flaps or the random portion of pedicled and free flaps [1]. The unpredictable vascular support of random pattern skin flaps needs consideration when defining their size and shape since inappropriate length-to-width ratios may predispose to ischemia and necrosis [2,3]. Free flap loss still occurs in 5.1-7.7% of cases, with rates almost doubling DOI: http://dx.doi.org/10.18053/jctres.08.202202.006 in post-mastectomy skin flaps [4]. Furthermore, post-mastectomy skin flap treatment represents an extra expense of up to $7000 per patient [5][6][7]. Therefore, serious esthetic, functional, and economic repercussions can develop when skin necrosis occurs in flaps used in these sensitive areas [8,9].
Ischemic preconditioning is a process that aims to improve flap survival by exerting brief episodes of ischemia to induce tissue tolerance and lessen necrosis and apoptosis [10][11][12]. Ischemic preconditioning methods include surgical procedures or pharmacological therapies. Surgical procedures have the disadvantage of needing an additional surgical intervention before the final flap harvest [3,13]. On the other hand, pharmacological interventions are limited by adverse effects, high costs, the difficulty of the administration method, and the drugs' unavailability [8].
Growth factors involved in angiogenesis, such as vascular endothelial growth factor (VEGF), have also been used as preconditioning therapeutics. While systemic and viral-mediated gene therapy can expose subjects to serious side effects [14][15][16][17], other existing delivery methods are limited by their short halflife, instability, local side effects, and appropriate dosage [18]. Adipose-derived mesenchymal stem cells (ADSCs) have shown promising results in the treatment of ischemic diseases through the secretion of VEGF and other pro-angiogenic cytokines [19]. Locally transplanted ADSCs can, therefore, continuously secrete VEGF to enhance neovascularization and improve ischemia. However, most of the studies use ADSCs of animal origin.
The architecture of human adipose tissue differs significantly from that of mice, likely affecting their functional properties [20]. In vitro studies have shown protein expression differences between hADSCs and animal ADSCs [21,22]. Furthermore, the differentiation capacity of hADSCs and animal ADSCs differs [23]. The evidence, therefore, implies that the results observed using ADSC of animal origin might not be entirely translatable to humans due to inherent molecular differences or differences in the mechanism of action [21,24,25].
Most preclinical studies using ADSC as a pre-operative treatment for skin flap ischemia and necrosis in animals have used either autologous or allogeneic ADSC, showing a decreased necrotic skin area in random pattern and pedicled flap models [26]. Although it is thought that the immune response in nonimmunosuppressed animals might skew the outcomes when using human-derived cell products [27], a recent comparison between porcine ADSC and hADSCs showed that both cell types had an equally strong capacity to reduce the proliferation of alloreactive splenocytes in mixed lymphocyte reaction assays [28].
Based on the previous background information, this review aims to show the efficacy of hADSCs at preventing distal skin flap necrosis in animal models of random pattern, pedicled, or free flaps. We hypothesize that hADSCs treatment is as effective as animal ADSCs treatment at preventing distal skin flap necrosis by enhancing angiogenesis in response to VEGF, as measured by the percentage of necrotic skin area. Furthermore, understanding the clinical effect of hADSCs in animal models is crucial for translational medicine since the cell products that will ultimately be used in patients will be of human origin.

Methods
Studies were identified by searching PubMed, CINAHL, and EMBASE databases from inception to present. The search was conducted on August 17, 2020. MeSH terms for "mesenchymal stem cells," "adipose-derived mesenchymal stem cells," "bone marrow mesenchymal stem cells," and "flap" were used. Search terms were arranged as follows: (stem cell*, mesenchymal OR mesenchymal stem cell* OR bone marrow mesenchymal stem cell* OR bone marrow stromal cell* OR adipose-derived mesenchymal stem cell* OR adipose-derived mesenchymal stem cell* OR mesenchymal stem cell*, adipose-derived OR mesenchymal stem cell*, adipose-derived OR adipose tissue-derived mesenchymal stem cell* OR adipose tissue-derived mesenchymal stem cell*) AND (surgical flap OR flap, surgical OR island flaps OR pedicled flap OR flap, pedicled).
Studies were included if they (1) used hADSCs, (2) measured the survival or necrotic skin area of flaps, (3) used animal models, and (4) were in English. The primary endpoint was to show that groups treated with hADSCs had a significantly lower percentage of necrotic flap skin compared to those not receiving this therapy within 14 days of cell administration. The secondary endpoints were the following: (1) To show that groups treated with hADSCs had a significantly increased flap perfusion compared to those not receiving this therapy within 14 days of cell administration; (2) to show significantly increased VEGF secretion and HIF-1α expression in hADSCs either in vitro compared to control culture media or ex vivo compared to those groups not receiving this therapy within 14 days of cell administration; and (3) to show significantly increased capillary density in the groups treated with hADSCs compared to those not receiving this therapy within 14 days of cell administration.
No specific publication status was considered. The study selection process, along with the reasons for exclusion, is detailed in Figure 1. Eligibility assessment and data extraction were performed by one reviewer (FRA), following the PRISMA guidelines. The risk of bias of included studies was assessed using the ROBINS-I tool of the Cochrane Library for non-randomized studies. A summary and a graph were created using RevMan 5.3 (Cochrane Collaboration), which allows for bias stratification in several domains (Figures 2 and 3).

Results
Out of 149 studies, 10 fulfilled the inclusion criteria. Studies assessing pedicled flaps used either a long thoracic artery [29] (Supplementary Figure 1) or a superficial inferior epigastric artery pedicled flap (Supplementary Figure 2) [30]. Studies assessing random pattern skin flaps used modified McFarlane flaps in their animal models ( Supplementary Figures 3 and 4). The McFarlane flap was introduced in 1965 as the first standardized surgical flap technique [31]. It was initially a cranially based flap positioned between the lower angles of the scapulae measuring 10×4 cm and yielding a length-to-width ratio of 2.5:1 [31]. Gong et al. [32] followed a similar surgical technique in a rabbit model. The included studies are summarized in Table 1.

VEGF levels and HIF-1α expression
Six out of 10 studies measured either VEGF, HIF-1α, or both ( Table 2) [30,32,33,[35][36][37]. All these studies found increased VEGF levels or HIF-1α expression. However, only two studies provided specific values to their measurements [30,32]. Although three studies graphically showed elevated levels and expression of VEGF or HIF-1α in either the culture supernatant or the ex vivo analysis of the flaps, they did not provide a statistical comparison of the hADSC-treated groups and the untreated control groups [35][36][37].

Comparison between hADSCs and animal ADSCs
In addition to extracting the necrotic or surviving skin areas observed in animal models of random pattern and pedicled skin flaps treated with hADSCs, we extracted these data from studies that used ADSCs of animal origin to prevent skin flap necrosis ( Table 3). If the studies provided the surviving skin areas, these values were subtracted from the total area to obtain the necrotic skin area. The studies that did not provide specific values for these data were not included in the calculation. Therefore, based on the available data, the use of animal ADSCs was associated with a decrease in skin necrotic area of 22 Figure 4). The ARR difference in skin necrotic area between animal ADSCs and hADSCs was not statistically significant (difference in risk: 0.0433 [4.33%]; 95% CI -0.3447-0.4313; P>0.05).

Discussion
Preconditioning aims to increase a flap's surviving length [40]. The first preconditioning method proposed for flap surgeries was surgical delay, consisting in the partial interruption of a flap's blood flow before transfer. However, the need for an additional intervention, increased patient risk, and increased health-care costs made this approach unsuitable for clinical practice [40,41]. Ischemic preconditioning, which followed surgical delay, consists of applying a brief period of ischemia and reperfusion to the flap, increasing its resistance to reperfusion injury [40]. However, this approach was never fully adopted for the same reasons as surgical delay [40]. A different approach, remote ischemic preconditioning (rIPC), showed positive results in preclinical models and prevented endothelial dysfunction in humans [42]. However, a recent randomized clinical trial failed to show improved free flap outcomes [43].
More recently, preconditioning has also been achieved in preclinical models by inducing hyperthermia or hypothermia in the region of interest or using pharmacological agents, growth factors, and mechanical stress. Many studies evaluating these preconditioning approaches have been performed in preclinical animal models, with few published clinical trials. These interventions have decreased necrosis in preclinical models compared to control groups [8,40,44,45]. A recent systematic review on thermal preconditioning by Kankam et al. only found three clinical trials, randomized and non-randomized, showing a lower incidence of flap necrosis and surgical reintervention in those patients using hyperthermic preconditioning compared to sham controls [46][47][48][49].
The pharmacological agents used for flap preconditioning are therapies with known dose-dependent side effects. For example, although nitric oxide donors have proven effective for flap preconditioning in preclinical models, they can lead to a dosedependent drop in blood pressure [50]. Furthermore, growth factors are not exempt from obstacles to their use. These molecules are limited by a short half-life, rapid diffusion from the delivery site, and low cost-effectiveness [51]. On the contrary, ADSCs' effects after administration are long sustained [52], providing a substantial advantage over other therapeutics.
Although initially, this review aimed to analyze the efficacy of both hBMSCs and hADSCs on skin flap necrosis prevention, there were no data on the former's use, and thus, the focus turned to hADSCs solely. This is most likely because hADSCs are easier to extract, have shorter replication times, secrete a higher number of cytokines, and yield a more consistent number when harvested from patients of different ages compared to hBMSCs [53]. An increased percentage of healthy skin was noted in eight out of 10 studies, confirming our working hypothesis. However, the fact that only five of those studies described a statistical analysis highlights the need for further studies with more rigorous methods.
The cumulative evidence in previous reviews shows that ADSCs increase skin flap survival through increased growth factor secretion, with a certain degree of ADSC endothelial differentiation [26]. However, these findings are based mostly on the use of animal ADSCs. Formal analyses of the differences between hADSCs and animal ADSCs are scarce in the literature. Nahar et al. recently found that 92% of the proteins expressed by hADSCs and mouse ADSCs were similar [21]. The clinical repercussion of this finding is still unknown. Understanding the clinical effect of hADSCs in animal models is crucial for translational medicine, since the cell products that will ultimately be used in patients will be of human origin.
Although animal ADSCs were associated with a higher reduction in flap skin necrosis than hADSCs, these values were not significantly different (P>0.05). Therefore, hAdMSC treatment is associated with a similar reduction in skin flap necrotic area compared to autologous or allogeneic animal ADSCs in preclinical animal models. Out of the six studies using hADSCs that were used to calculate the ARR in flap skin necrotic area, only two used immunosuppressed animals. Although hADSCs do not generate a substantial immunogenic reaction in vitro [28,54], it is unclear if the heterogeneity in the state of the immunologic systems of animal models influenced the results. Only three of     the studies analyzed in this systematic review present information regarding the immunological reaction after xenogeneic stem cell transplantation in immunocompetent hosts. Gong et al. stated that there was no evident macroscopic reaction (e.g., erythema and fever) in the animals treated with hADSCs [32]. In addition, the number of CD3+ cells and the CD4/CD8 ratio in pathology slides of treated and control groups were not statistically different (P>0.05) [32]. Furthermore, IFN-γ, IL-2, IL-4, and IL-10 levels were also not statistically different between groups (P>0.05) [32]. groups; however, no statistical analysis was done [38]. Finally, Feng et al. found that TNF-α, IFN-γ, and IL-6 levels were lower in groups treated with a low cell dose compared to controls [30].
In this same study, groups receiving a high dose of stem cells showed levels similar to those in control groups [30]. Although these studies point to an absence of a substantial inflammatory reaction and in some cases, a decrease in pro-inflammatory cytokines, a conclusion cannot be drawn at this time due to lack of information. Further studies using hADSCs to prevent skin flap necrosis should measure the immunologic reaction, both in vivo and ex vivo, after cell transplantation. Although some studies did not find differences between the hADSC-treated groups and the controls [38,39], the absence of a statistically significant difference between the hADSC-treated groups and the best performing groups of each study imply that improvements in the methodologies (e.g., increased number of animal models or transplanted cell number adjustments) could yield conclusive results. Some studies' primary outcome was to study the addition of hypoxia preconditioning methods (e.g., lowlevel light therapy [LLLT] or remote ischemic preconditioning), either on the cells or the skin, on flap survival [35][36][37]39]. The fact that these studies found that using these methods increased flap survival, proangiogenic cytokine secretion, and capillary density to a higher degree than hADSCs alone (P<0.05) is an important finding that should be further studied. The results of the included studies suggest improved small vessel vascularity using hADSCs. The results of Park et al. [35,37] imply that LLLT, applied to flaps transplanted with either monolayer hADSCs or hADSC spheroids, could enhance these cells' secretory capacity and survival, thereby increasing the capillary number and flap perfusion. However, it should be noted that LLLT-treated hADSC spheroids had increased levels of endothelial markers [37]. The excessive use of hypoxia preconditioning methods might compromise hADSCs to the vascular endothelial lineage before transplantation.
In vitro comparisons of hADSCs and rat ADSCs have shown a better endothelial differentiation potential of the latter when cultured in commercial endothelial differentiation methods [55]. When evaluated ex vivo, most of the included studies found that hADSCs differentiated to endothelial cells to a low degree [29,30,[32][33][34][35][36][37]. One study calculated that hADSC differentiation to endothelial cells contributed to 15.4% of the flaps' neovascularization [32]. However, one study did not find human cells in the flap even though it found an increased number of vessels in the hADSC-treated group, concluding that this increase was due to paracrine effect [38]. The contribution of hADSCs to flap neovascularization should be quantified in further studies.
Although the studies had favorable results, the lack of protocol standardization might pose a substantial bias. In 2014, Lee et al. [34] compared the effectiveness of different hADSC delivery routes to improve the viability of ischemic flaps and found that their application with a collagen sponge provided the best results. In 2020, Feng et al. [30] proved that intra-arterial delivery of hADSC through the femoral artery is also an efficient method to prevent flap ischemia and necrosis. Studies comparing these two methods are required to elucidate the delivery route that shows the best therapeutic efficacy.
Studying the best route of administration and the influence of the immunologic response should be further analyzed since only two studies evaluated these factors [34,38]. However, the study of hADSCs to prevent flap necrosis should first focus on elucidating the general contribution of the direct cell differentiation to endothelium and the paracrine effect on neovascularization. This might further derive in comparative studies between hADSCs and their cell products.
Finally, most studies evaluated random pattern skin flap necrosis, with few focusing on ischemia/reperfusion injuries. Ischemia/reperfusion injury is an acute process characterized by mitochondrial damage and cell death due to an abrupt increase in reactive oxygen species and other inflammatory mediators after an ischemic tissue regains perfusion [56,57]. This being an acute pathologic process, stem cells of any type might not be particularly suitable for treating it since their effects are more gradual and sustained. Therefore, pathologies requiring increased perfusion and gradual evolution are more suitable for stem cell treatment. Soft-tissue defect reconstruction and wound healing fit these requirements, and thus, stem cell research in plastic surgery has mainly focused on those processes [58,59].

Conclusion
The effect of hADSCs in flap viability improvement is being increasingly studied. The results provided in this review show that hADSCs prevent flap necrosis to the same extent as animal ADSCs in rodent and rabbit models of random pattern skin flaps and pedicled flaps.

Limitations
This study has several limitations. Since only studies published in English were included in this review, some studies may have been missed. Other limitations include the scarcity of studies reporting on this topic, the potential bias of misinterpreting data and results, and the study selection process, the latter being a potential source of bias common to systematic reviews. Specific for this systematic review, the absence of high-quality data, and the relative absence of information regarding the animal's immunologic response to human cells preclude us from analyzing if the immunologic status of the animal model should be considered for these studies. The major use of one flap technique (i.e., McFarlane flap) poses a risk of bias when extrapolating these findings to other types of flaps (e.g., pedicle flaps). Finally, the lack of a reason for using cells of human origin in the studies also poses a substantial risk of bias.