ABSTRACT

In plastic and reconstructive surgery there is an increasing demand for malleable implants to repair soft tissue congenital defects, or those resulting from aging, traumatic injury and tumour resection. However, currently available methods present a number of limitations such as volume loss over time and eventual resorption of the graft. Tissue engineering techniques provide promising therapeutic solutions to these inconveniences through development of engineered equivalents that best imitate adipose tissue, both structurally and functionally. Here we review the latest achievements in the human adipose tissue engineering field, with a focus on its regenerative potential for a number of clinical applications.

1. Introduction

Adipose tissue is a highly vascularized connective tissue, ubiquitous throughout the human body. It is responsible for energy storage and release of a number of adipokines that may act in an endocrine or paracrine fashion [1]. It is also a highly plastic tissue than can increase energy depots by hypertrophic growth and hyperplasic expansion of adipose stromal cells (ASCs) [2]. It contains many cell types including adipocytes, ASCs (a term that refers to the tissue-resident mesenchymal stromal cells (MSCs) and adipose progenitor cell populations), endothelial cells (ECs), fibroblasts, macrophages and leukocytes.

The human adipose tissue can be found either in the white adipose tissue (WAT) or the brown adipose tissue (BAT) form. WAT is the most abundant form in humans and it accumulates energy in the form of triacylglycerols. WAT is composed of large spherical or ovoid cells 60 - 80 μm in diameter, characterized by a single lipid droplet that accounts for about 90% of the cell volume. The lipid droplet provokes peripheral displacement of the nucleus and appearance of a variable number of round or elongated mitochondria with short and randomly oriented cristae [3-5] . Unlike WAT, BAT is characterized by polygonal shaped cells, 30 - 50 μm in size, with complex internal structure of mitochondria which confers to these cells their characteristic brown coloration, granular appearance of the cytoplasm and the near absence of cytoplasmic membranes [5]. Brown fat cells present the ability to dissipate energy by producing heat to ensure body temperature regulation, rather than storing it as triglycerides [3]. Whether such thermogenic properties also play a role in body weight regulation and metabolic disorders is still being debated [6-10] .

Tissue engineering is the interdisciplinary field where materials, cells, growth factors and other bioactive molecules are combined together to make transplantable constructs, the final goal being to promote repair and regeneration of damaged tissue [11]. Loss or damage of adipose tissue needs repair and regenerative approaches not only for the cosmetical impact of absent tissue but also for the emotional well being of the patients. In some cases damaged adipose tissue may even impair functionality of the affected organ. The first reported fat tissue repair was performed in 1893 by Neuber, who transplanted a fat graft to fill soft tissue depressions on the face [12]. In the following years several authors tried to refine the technique with variable results [13,14]. In 1956, Peer et al. transplanted autologous subcutaneous adipose tissue from the abdomen into the face and breast of a patient and stated that large grafts eventually lead to adiponecrotic cyst formation and resorption [15]. In 1999, Patrick et al. reported the generation of the first in vitro fat construct using poly(lactic-co-glycolic) (PLGA) scaffolds and rat preadipocytes [16], but its long-term maintenance remained elusive at 12 month follow-up [17]. Since then, several attempts have been made for soft-tissue repair and regeneration with different strategies but the medical need is still considered as “unmet” in terms of safety and efficacy [18], long-term sustainability, and optimal aesthetical results of the grafted tissue [19].

Tissue engineering strategies thus hold great promise to hopefully offer a permanent solution for adipose tissue repair and regeneration in the near future. All the strategies discussed in this review present different advantages and disadvantages and the choice of one or the other must be carefully studied in a case by case basis (Table 1).

2. Autologous Fat Transfer

Autologous fat transfer is the technique most used clinically for fat tissue repair. It consists of the removal of the fat from one area of the body to be reinserted into a separate location, where it will fill subcutaneous tissue loss. It is a minimally invasive and safe procedure that allows patients to benefit from autologous tissue transplantation with minimal risk of immune rejection or transmission of viral pathogens.

Its applications range from severe conditions such as facial lipoatrophy [20-23] , Parry-Romberg syndrome [24] and schleroderma [25,26] , cranio maxillo-facial deformities [27] , mastectomy or breast volume augmentation [28-30] or chronic ulceration [31]; to cosmetic applications such as rejuvenation [32-34] , gluteal [35] and labia majora augmentation [36] or phalloplasty [37] (Table 2).

However, autologous fat transfer rarely achieves longterm volume augmentation due to the limited survival of mature adipocytes after the liposuction trauma and the poor revascularization that leads to cyst formation and an eventual resorption of the graft [38-40] . Even if some authors have observed up to 8 years permanence of the graft in some conditions [18], most of the studies evidence varying degrees of resorption over time with only a small percentage (30%) of the graft presenting long-term survival beyond the first year; success being attributed at least partly to the presence of ASCs within the graft [41,42] . As a result, complications such as contour irregularity; lumpiness and unpredictability of the graft take and final shape make the grafting technique inefficient and unsuccessful.

In an attempt to overcome low revascularization, vascularized flaps have been used. They consist of skin and

fat removed together with blood vessels and transferred to the desired area. However the use of flaps is also limited by donor site defect, infection, unstability and wound healing disorders [43]. The latest surgical approach for autologous fat transfer is the lipofilling technique, which consists of the injection of microportions of autologous fat obtained from non-traumatic vacuum aspiration. In combination with cells isolated from the adipose stromal vascular fraction, this technique has proved to be effective, safe and superior to conventional lipoinjection [42,44] . Some authors report that even the sole injection of in vitro differentiated autologous adipocytes is an effective strategy to achieve soft tissue augmentation [45,46] .

Despite the comprehensive efforts of a number of fully dedicated clinicians, a standard procedure for harvesting and processing of cells, injection and reinjection site has not been developed as of yet. All of the above are key elements to be considered for a successful autologous fat transfer, although many other factors might also affect the final outcome [18,47,48] .

3. Adipose Tissue Engineering

Adipose tissue engineering is recently gaining significant importance due to the increasing need for clinical soft tissue filler procedures and the known impact of ASCs and their secretions in wound healing [49,50] . Currently, most adipose tissue engineering approaches include two different strategies: in situ adipogenesis for small volume loss, and 3D in vitro tissue engineering for large adipose tissue defects. Both strategies usually imply the use of living cells and/or biomolecules and/or biocompatible scaffolds.

3.1. Cells

3.1.1. MSCs and Adipocytes

The ideal cell population for any tissue engineering strategy should present the following characteristics: 1) Autologous or at least non-immunogenic; 2) Available in sufficient quantities or easy to propagate in vitro with no loss of differentiation potential; and 3) Easy to harvest with minimally invasive procedures [51].

One of the barriers that hampered the in vitro fat tissue development was the impossibility to maintain the longterm culture of mature adipocytes ex vivo [52-55] . Mature adipocytes are prone to mechanical damage during fat tissue manipulation and also highly susceptible to ischemic death. Moreover, they are terminally differentiated and thus unable to proliferate. These characteristics make them unsuitable for regenerative purposes [56]. Recently we developed a new culture method for adipocytes using a plasma hydrogel scaffold. We were able to maintain an in vitro adipocyte culture with minimal medium change for 2 years, a considerable improvement over previous reports [57].

The use of ASC and bone marrow MSCs (BM-MSC) has been reported in several studies and patents for adipose tissue engineering [58,59] . The Mesenchymal and Tissue Stem Cell Committee of the International Society for Cellular Therapy (ISCT) proposed minimal criteria for defining MSCs. They must be plastic-adherent, demonstrate the expression of the surface markers CD105, CD73, and CD90 and show minimal or null expression of CD45, CD34, CD14 or CD11b, CD79a or CD19 and HLA Class II proteins. Finally, MSCs must demonstrate the ability to differentiate to osteogenic, adipogenic and chondrogenic lineages. Mesenchymal stromal cells are preferred over mature adipocytes due to their capacity to proliferate in vitro while maintaining their multilineage differentiation potential [60] to bone [61] , cartilage, fat [62] and vessels [63]. Besides, they possess angiogenic and anti-inflammatory properties [64] and an interesting secretome with highly variable quantities of proteins that may play important roles in regenerative environment, such as hepatocyte growth factor (HGF) and vascular endothelial growth factor (VEGF) [65,66] .

A number of studies reported different strategies to develop 3D adipose tissue constructs using ASC or BMMSC as the cellular starting material: Vermette et al. and Vallée et al. used the so called “self-assembly strategy” [67] to recreate three-dimensional adipose substitutes in vitro devoid of exogenous biomaterials [68,69] . After 21 - 28 days’ culture, stimulation of extracellular matrix (ECM) secretion by ascorbic acid and induction of adipogenesis, cells show a robust differentiation potential. Monfort et al. embedded ASC and BM-MSC within fibrin scaffolds and proved similar usefulness for adipocytic differentiation of this material, leading to formation of an adipose construct after 21 days’ culture [57]. Other authors have used diverse scaffolds such as collagen, gelatin or silk-fibroin [70-72] .

3.1.2. Endothelial Cells (EC)

A second main obstacle for in vitro fat tissue development, as yet unresolved, is the low in vivo revascularization of large volume adipose tissue grafts. Lack of vascularization leads to grafted tissue shrinkage, fibrosis and oil cyst formation. Implanted tissues need days or weeks to develop new blood vessels and during this process an insufficient supply of nutrients and oxygen may compromise the viability of the graft [73,74] . To create functional vascularization, an intact endothelial layer covering the inner side of branched vessel structures with a maximum distance between capillaries of approximately 200 μm is needed [75]. In order to achieve this goal, different approaches have been proposed: 1) Endothelial cell-based strategies to form new vessels; 2) Scaffoldbased strategies focused on the generation of biologically or synthetically-derived vessel systems [75,76] ; 3) In vivo integration of vascular pedicles [77]. The latter needs several surgical interventions and morbidity and invasiveness of the procedure are major concerns to be considered for risk/benefit analyses.

In vitro prevascularization consists of seeding the scaffolds with EC alone or in co-culture with other cell types such as BM-MSCs, ASCs, or fibroblasts in order to build 3D structures that anastomose with the host vasculature when grafted, thus improving vascularization of the construct [59,78-80] . Moreover it has also been shown that adipose tissue-derived ECs secrete soluble adipogenic factors [81,82] . Similarly, ASCs secrete factors during their adipogenic differentiation process that stimulate EC growth and motility in vitro as well as angiogenesis in vivo [83,84] . In this regard, a co-culture model including ASCs and ECs within 3D silk scaffolds showed the formation of endothelial tubes concomitant with the adipogenic differentiation of the ASCs [85].

3.2. Biomolecules

3.2.1. Adipogenic Biomolecules

A central issue in the preconditioning of MSCs for adipose tissue regeneration is the adipogenic induction before transplantation. In vitro adipogenic differentiation is characterized by growth arrest previous to the induction and expression of the adipogenic genes [86]. However, a cell-cell contact seems to be essential to direct terminal differentiation into mature adipocytes  ADDIN EN.CITE  ADDIN EN.CITE.DATA [62] . Studies in which adipogenic induction is performed prior to implantation of either ASCs or BM-MSCs demonstrate a clear advantage for in vivo adipose tissue development [70,87-89] . In the last year, >20 articles have been published for the in vitro adipogenic differentiation of mouse and human MSCs but generally, a 3-component cocktail is used. This cocktail comprises insulin, glucocorticoids (e.g. dexamethasone) and isobutyl methylxanthine (IBMX), a phosphodiesterase inhibitor that increases the levels of intracellular cAMP. Other molecules frequently found in adipogenic protocols are peroxisome proliferator-activated receptor gamma (PPARγ) agonists such thiazolidinediones (e.g. rosiglitazone, troglitazone) and indomethacin. The final concentrations of the different inductors in the media is highly variable, ranging from 0.2 to 12,000 nM for Insulin, from 250 to 1000 nM for Dexamethasone, and from 1 to 200 μM for indomethacin [90].

3.2.2. Neovasculogenic and Angiogenic Biomolecules

Addition of angiogenic growth factors to engineered constructs is one of the strategies to improve neovasculogenesis and angiogenesis within the grafts. VEGF has long been established as the prime angiogenic molecule during development, adult physiology and pathology [91]. Hypoxic conditions induce the expression of angiogenic factors such as VEGF and leptin. VEGF is the most important endothelial cell growth factor necessary to initiate the formation of immature vessels by either vasculogenesis or angiogenic sprouting [92].

The hormone leptin is a 16-kDa protein predominantly synthesized and secreted by adipocytes. Besides its main roles in the orchestration of the nutritional state and as a sensor of the availability of endogenous energy resources [93] , several studies have confirmed that leptinmediated autoand paracrine stimulation of human fibroblasts [94] and keratinocytes [95,96] accelerate wound healing [97,98] and angiogenesis [99-102] . Thus, an important role of adipokines in regenerative conditions is generally acknowledged [50,103] . Leptin also modulates the expression of VEGF and plays a role in matrix remodelling, facilitating the angiogenic process and thus promoting neovascularization of the microenvironment [104] .

In numerous studies basic fibroblast growth factor (bFGF) has been supplied to the media as a potent angiogenic factor, leading to vascularised adipose tissue [105-110] . Basic FGF induces an angiogenic response via a direct effect on endothelial cells, synergistically with VEGF [111,112] . Ordered orchestration by controlled release of the concentration of different biomolecules in the 3D constructs would permit successful neovasculogenesis and angiogenesis and, eventually, full and permanent take of the grafts [113-119] .

In the last few years, new molecules have emerged as important regulators of angiogenesis and neovasculogenesis and their application for tissue engineering purposes is being considered [120-122] .

3.3. Materials

Besides the appropriateness of cells and biomolecules, biomaterials are central components of tissue engineering strategies and play a pivotal role in the final outcome of the grafts [123].

The scaffolds are usually classified in two categories: porous or “hard” and injectable or “soft” scaffolds. The selection of one or the other type of scaffold depends on the strategy: hard scaffolds are used for predefined-shape 3D in vitro adipose tissue constructs (usually for large adipose tissue volume restoration) and soft matrices are used for in situ adipogenesis (for small volume loss).

3.3.1. Hard Scaffolds for 3D in Vitro Adipose Tissue Engineering

The ideal hard scaffolds should meet a long list of selection criteria: 1) They should support cell attachment, migration, cell-cell interactions, cell proliferation and differentiation; 2) They must be biocompatible; 3) They should also be biodegradable at a controlled rate -ideally one to match the rate of neotissue growthto facilitate the integration of engineered tissue into the surrounding host tissue; 4) They must provide structural support for cells and neotissue formed in the scaffold during the initial stages post-implantation; 5) They should present interconnected pores to facilitate vessel growth and nutrient transport; 6) They should have versatile processing options to alter structure and morphology related to defect-specific needs [124,125] .

Materials may be classified into natural and synthetic polymers. Among natural polymers hyaluronan [126] , fibrin [127], collagen [128], gelatin [71] , silk fibroin [72], chitosan-silk fibroin [129], alginate [130], decellularized placenta [131], extracellular matrix from adipose tissue [132-136] and combinations of them [137] have been very widely studied [51,138,139] .

The synthetic polymers include polyethyleneglycol (PEG) [140,141] , polytetrafluoroethylene [142], polyethylenene terephthalate [143], polyglycolic acid (PGA) [144] , polylactic acid (PLA) and copolymers (PLGA) [16,109] , polycaprolactone, polyurethanes, poly(ortho ester) and poly (anhydrides) [11,51,86] . Frequently, combinations are used such as heparin-conjugated fibrin hydrogels [145] and PEG-cross-linked heparin [146] among many others [147, 148] . Synthetic materials possess several drawbacks when compared to natural ones, such as the absence of intrinsic surface ligands for cell attachment and a potential impact of their degradation products on cell function [125]. Lately filamentous nanofibers are gaining importance, since their small diameter closely matches that of extracellular matrix fibers, allowing them to be used as biomimetic scaffolds [149-152] .

Scaffold prevascularization strategies offer great promise to the field of adipose tissue engineering [79] . Prevascularization of matrices with the co-culture of endothelial cells and fibroblasts increased the ability of the constructs to create anastomoses with the host vasculature [153]. Once the capillary network was developed in vitro, the graft was implanted into the recipient. Natural scaffolding materials have extensively been used in the vascularisation of tissue engineering constructs including collagen, chitosan [154-156] , decellularized extracellular matrix [157], and silk fibroin-fibrin [158,159] . The two most commonly used synthetic materials used for vascularization are PEG [160] and PLGA [160,161] . Although it would be highly desirable to control the kinetics of protein bioavailability through matrix release [162], clinical translation of systems capable of mimicking such naturally occurring processes is not foreseeable in the near future [163,164] .

3.3.2. Soft Scaffolds for in Situ Neoadipogenesis

Soft scaffolds permit the in situ generation of new adipose tissue. It may be achieved with the co-injection of growth factors, cells, tissue fillers or a combination of them. The ideal injectable material for soft tissue augmentation has possibly not yet been discovered, although some of them seem to be less risky than others [165]. In situ adipogenesis is a very appealing option to minimize scarring, risk of infections and eliminate the need for surgical interventions, thus reducing patient discomfort and intervention-related complications. Moreover, soft materials adapt to the shape of the cavity in which they are placed. However, they are only suitable for small volume replacements. Depending on their resorption rate, fillers are classified in three categories: resorbable within months, within years, and permanent fillers [166] .

As seen for the hard scaffolds, soft scaffolds may also be classified as for their natural or synthetic origin. Derivatives of hydroxylapatite, PLA, collagen or hyaluronic acid are the most popular products used for in situ adipogenesis purposes. However, they only serve as temporary fillers and repeated injections are required to maintain the desired volume, due to the lack of a regenerative component [167] . Collagen has been on the market for the last two decades, and several collagenous products have been FDA-approved [168]. Alginate [169] or fibrin hydrogels have been extensively studied with promising results [89,170,171] . Other notable materials include polymethylmethacrylate, polyacrylamide and dextran [40].

Injectable micro and nanospheres are commonly used in combination with cells and growth factors for regenerative purposes [172,173] : PLGA microparticles carrying ASCs [174], BM-MSCs [175] or preadipocytes [176] are the most studied examples for adipose tissue engineering. Nevertheless, and due to their synthetic origin, their degradation products may cause pH variations that in turn lead to protein denaturing and impaired cell functionality [40]. Natural biopolymers in combination with cells such as collagenous microbeads with ASCs [177], chitosan microspheres with ASCs [178] or alginate microbeads [179] are alternatives with a safer degradation profile.

Finally, latest advances in biomaterial technology include new approaches such as 1) Temperature-sensitive hydrogels that are liquid at room temprature and gelify at 37˚C [180,181] ; 2) ECM injectable powders shown to promote neoadipogenesis and angiogenesis in vivo [182] , thus improving the results obtained with the autologous fat injection; 3) ECM microcarriers [183]; and 4) A photoactivated PEGHA hydrogel filler that crosslinks in situ by a shining array of LEDS emitting light at a wavelength of 520 nm. This filler can be massaged into the desired shape in situ before gelation through photoactivation [184] .

4. Concluding Remarks

For the numerous adipose tissue engineering strategies here discussed, the most challenging task is repair and regeneration of large tissue defects. All the adipose tissue engineering studies conducted with cells have shown a positive effect when compared to a non-cell control, and thus cellular strategies are prevailing in the field. Cells may act in a paracrine manner, triggering a host response that stimulates the recruitment of endogenous stem cells to the site and promoting their differentiation along the required lineage [66,185] . At present most of the authors use MSCs for their adipose tissue engineering studies. ASCs and BM-MSCs are ideal starting populations for tissue engineering applications [51]. However, other authors suggest embryonic stem cells or germ cells as ideally suited for in vitro regenerative studies [186,187] . Risk/benefit balance of such pluripotent cell use should also be carefully considered.

Unlike other tissues, adipose tissue continuously undergoes expansion and regression and a constant adjustment of its capillary network is necessary. Therefore graft vascularization is a fundamental issue to be addressed when designing novel regenerative strategies. Although some authors have reported hypoxia-mediated angiogenesis in vivo [188], prevascularization of the grafts seems necessary for successful long-term outcome.

Despite the wide range of materials that have been intensively studied, every scaffold presents strengths and limitations and the ideal matrix for adipose tissue engineering surely has to be invented. Naturally-derived materials present many issues regarding pathogen transmission, or rapid volume lost. Synthetic materials often cause immunogenic reactions and integrate poorly with the host tissue. Acellular dermal matrices have been studied with promising results [136,189] . Several such dermal substitutes have already been approved by the FDA and are currently used in the clinics [190].

The in vitro development of 3D adipose tissue is also an invaluable tool for ex vivo studies on physiological disorders, metabolic diseases and drug testing. Monfort et al. speculated that BAT might be generated in their system due to the presence of the laminar cristae in the mitochondria reminiscent of that seen in human BAT [57, 191] . Although further evidence must be gathered to support this speculative hypothesis, the development of the BAT phenotype in vitro would be an interesting breakthrough for the study of its associated resistance to obesity and related metabolic disorders. Importantly, and regarding eventual grafting into patients, VEGF expression is much greater in BAT than in WAT [192,193] . This single fact makes BAT tissue highly attractive for regenerative purposes due to its intrinsic angiogenic properties.

The in vitro generation of 3D adipose tissue constructs led to the eventual development of complete trilayered engineered skin equivalents where the functional interaction of the adipose hypodermal layer with dermis and epidermis in vitro was evidenced [57,194] . Besides their usefulness for in vitro studies and drug testing, the application of trilayered skin constructs for in vivo grafting of full-thickness skin and its pharmaceutical development under GMP regulation could be considered [195].

In conclusion, adipose tissue engineering is rapidly advancing with the multidisciplinary collaborative work of stem cell biologists, biomaterial scientists, biomedical engineers and clinicians. It can be envisaged that these efforts will open innovative avenues for the development of a clinically relevant construct to support soft tissue repair and regeneration in the short term.

5. Acknowledgements

This work was financed by grants provided by Ministerio de Ciencia e Innovación [Proyecto Inv. Salud (PI10/02871) and INNPACTO (IPT-300000-2010-17) programs], Diputación Foral de Gipuzkoa (OF 53/2011 and OF 94/2008) and Gobierno Vasco (Dpto. de Industria SAIOTEK program-SAIO10-PE10BF01). A.I. was supported by the “Programa I3SNS” (CES09/015) from Instituto de Salud Carlos III and by Osakidetza-Servicio Vasco de Salud (Spain).

REFERENCES

NOTES

^*Corresponding author.