The Science of Autologous Fat Grafting: Views on Current and Future Approaches to Neoadipogenesis

Abstract

Learning Objectives: The reader is presumed to have a broad understanding of plastic surgical procedures and concepts. After studying this article, the participant should be able to:

1. Describe the current clinical applications and limitations of autologous fat grafting.

2. Identify the important physiological steps and molecular pathways of neoadipogenesis.

3. Cite current in vitro and in vivo models for the analysis of fat grafting techniques.

Physicians may earn 1 AMA PRA Category 1 credit by successfully completing the examination based on material covered in this article. The examination begins on page 322. ASAPS members can also complete this CME examination online by logging on to the ASAPS Members-Only website (http://www.surgery.org/members) and clicking on “Clinical Education” in the menu bar.

Autologous fat transplantation has become a well established and frequently applied method of soft tissue augmentation for both cosmetic and reconstructive indications. There is no consensus, however, about the best fat grafting technique, nor is there reproducible data regarding its durability. The most significant drawback to autologous fat grafting remains its largely unpredictable rate of resorption. A thorough understanding of the developmental biology and molecular regulation of adipogenesis and adipocyte survival is critical to optimizing the fat grafting technique. Consequently, numerous in vitro and in vivo studies on fat graft viability have recently been undertaken. Here, we discuss the latest advances in the basic science of adipogenesis, adipocyte viability, and its clinical application to fat grafting, arguing that the data produced by in vitro and in vivo studies still fail to produce a clear picture of the required components for successful, consistent, and durable fat transplantation; however, it is undetermined if this lack of clarity may simply be a lack of systematic scientific data acquisition or if these findings truly reflect the biology of neoadipogenesis. As a first step in strengthening autologous fat grafting scientific data collection, we recommend that a collective, multidisciplinary, multicenter effort be undertaken to establish in vitro and in vivo models of neoadipogenesis that are clearly reproducible from one investigator to another. With the implementation of systematic scientific approaches to the study of neoadipogenesis, we anticipate the future of autologous fat transplantation for correction of soft tissue volume loss to be extremely promising.

Soft tissue volume loss acquired through aging, trauma, or congenital malformation is a common and sometimes devastating problem that is frequently presented to plastic surgeons. Attempts at restoring soft tissue volume have consisted of either filling the defect with local or free tissue flaps or with autologous or alloplastic filler injections. Recent trends toward minimizing invasive procedures for mild to moderate defects have made autologous and alloplastic injectable fillers an especially attractive option for many patients and physicians.

The ideal injectable agent for soft tissue augmentation must fulfill certain criteria. It must be easy to use, biocompatible, inexpensive, and lack toxicity. Furthermore, this agent should fill the tissue envelope with a natural feel and appearance for optimal cosmesis, and it must produce consistent and reproducible results. Autologous fat transfer, first introduced by Neuer in 1893¹ and subsequently modified by Bruning in 1911² for use in injectable form, has become a well established method of soft tissue augmentation for both cosmetic and reconstructive indications. Fat naturally fulfills many of the characteristics required of a soft tissue filler. It is autologous, nontoxic, biocompatible, easily available in most patients, and potentially removable and long lasting. The obvious suitability of fat for soft tissue augmentation and the introduction of lipoplasty techniques into the United States in the 1980s popularized the injection of the lipoplasty byproduct, a semi-liquid fat, for the restoration of soft tissue deficits.^3,4 However, the panacea effect faded when reports of variable fat survival and unacceptable failure rates surfaced in the literature.⁵ Fat grafting remained shrouded in a negative light throughout the 1990s despite many subsequent positive fat grafting reports by plastic surgeons who were able to achieve success with refinement of their techniques.^6–8 Unfortunately, no consensus was ever achieved about the best technique for fat grafting, longevity data, or other clinical parameters, and many attempts at optimizing these data were abandoned with the introduction of a new generation of alloplastic commercial fillers in the mid-2000s which were thought to be superior to their collagen-based predecessors.⁹ Hyaluronic acid—and calcium hydroxylapatite–based fillers not only provided plastic surgeons with a myriad of soft tissue filler options, but also with an established method of application and a high rate of reproducibility and availability.^10–13 These alloplastic agents can also be associated with significant flaws including allergenicity, relatively short duration of correction, high cost, and potentially unnatural texture and filling of the soft tissue envelope.¹⁴

While there are clearly appropriate and ideal applications for alloplastic commercial fillers, a resurgence of interest in fat grafting has been observed in the past several years, making it once again a common procedure for plastic surgeons.^15,16 A recent national consensus survey queried 508 surgeons about the trends in techniques for harvest, preparation, application, and perception of short- and long-term results of fat grafting.¹⁷ The survey concluded that while autologous fat transfer is a relatively common procedure, few surgeons perform it in high volume or with identical techniques. Furthermore, 84% of patients seem to be pleased with the short-term results and 80% are pleased with long-term results. Yet our knowledge regarding fat graft behavior after transplantation remains tenuous. A thorough understanding of the developmental biology and molecular regulation of adipogenesis and adipocyte survival is critical to optimizing the technique of fat grafting. Consequently, numerous in vitro and in vivo studies on fat graft viability have been recently undertaken and the data generated will be reviewed here.

Adipose Tissue: Structure and Physiology

Adipose tissue (fat) is composed of a major lipid-filled cell type, the adipocyte (fat cell) that is surrounded by stromal vascular cells (SVCs) such as fibroblasts, immune cells, collagen fibers, and blood vessels. The extracellular matrix (ECM) interconnects adipocytes and forms the fat lobules in adipose tissue. There are 2 general histologic types of adipose tissue, brown fat and white fat. In humans, brown adipose tissue is predominantly found during the neonatal period and is responsible for generating thermogenesis from triglycerides. Brown fat does not play a significant role in adult metabolism in humans. For the purposes of this review, we will limit our discussion to white fat.

White adipose tissue, composed of adipocytes with a single large lipid inclusion and a large peripherally-located nucleus, represents the predominant type of fat in humans. It is involved in a variety of physiological roles including the storage of energy-rich triglycerides, cushioning of vital structures and organs, metabolic homeostasis, immune regulation, reproduction, and angiogenesis.^18–21 The imbalance of adipose tissue resulting in either too much fat, such as in generalized obesity, or too little fat, such as in genetic or acquired lipodystrophies and aging, is an increasingly prevalent problem worldwide. Such disorders are associated with physiologic derangements in insulin metabolism, triglyceride and cholesterol stores, as well as generalized insults to end organs involved in these pathways. The increased recognition of these problems has highlighted the need for a more complete understanding of adipose biology.

Adipose tissue influences metabolic homeostasis by producing an assortment of hormones, cytokines, growth factors, and other peptides. While adipose tissue has historically been considered a semi-inert tissue, we now know the opposite to be true. The factors secreted by adipose tissue are involved in a broad spectrum of physiological and molecular pathways, including lipid and steroid metabolism, growth factor, binding protein, cytokine signal transduction, vasoactivity, eicosanoid activity, alternative complement system, extracellular matrix, and others. These molecules, loosely termed adipokines, exert their effects in an endocrine, paracrine, and autocrine manner. While a comprehensive discussion on adipocyte biology is beyond the scope of this paper, some of the major adipokines involved in obesity have recently received media attention and are briefly worth mentioning.^18–21 Leptin and adiponectin, which are increased and decreased in obesity, respectively, function by regulating appetite and energy expenditure. Tumor necrosis factor–alpha (TNF-α) and interleukins-8 (IL-8) and -6 (IL-6) are all increased in obesity and are proinflammatory cytokines that promote insulin resistance and lipolysis. In addition, type 1 plasminogen activator inhibitor (PAI-1) is increased in obesity and functions to promote thrombosis by inhibiting fibrinolysis, thus acting as the main endogenous regulator of fibrinolysis.

Molecular Pathways and Developmental Regulation of Adipogenesis

Adipocyte differentiation involves the transformation of mesodermal stem cells into adipoblasts, preadipocytes and, finally, mature lipid-synthesizing, lipid-storing white or brown adipocytes. Differentiation involves the coordinated expression of specific genes involved in producing the adipocyte phenotype.²² Transcriptional regulation of adipocyte differentiation during development is largely controlled by peroxisome proliferator-activated receptor-γ (PPAR-γ), which is the factor currently thought to be the most specific to adipogenic differentiation. When this receptor becomes activated by its agonist ligand in fibroblasts, a given cell begins to differentiate into an adipocyte via morphologic changes, lipid accumulation, and activation of genes distinctively expressed by adipocytes. Additional support during differentiation comes from CCAAT/enhancer-binding proteins (C/EBPs), with C/EBP-β and C/EBP-δ specifically stimulating PPAR-γ expression during early differentiation and C/EBP-α having a similar role later in the pathway. Recently, lipoprotein lipase (LPL), specific Krüppel-like factors (KLF5, KLF15, and KLF2), early growth response 2 (Krox20), and early B-cell (O/E-1) factors have also been implicated in adipocyte differentiation.^23–27 Not surprisingly, both insulin and insulin-like growth factor–1 (IGF-1) are known to be required for adipocyte differentiation. In contrast, preadipocyte factor–1 (pref-1), whose role is in the maintenance of the preadipocyte phenotype, is decreased during adipocyte differentiation. Late markers of differentiation, which are factors produced once adipocytes mature, include adipsin, angiotensinogen II, acyl-coenzyme A (Co-A)–binding protein (ACBP), leptin, and two fatty acid–binding proteins (FABP) known as adipocyte lipid binding protein (aP2) and keratinocyte lipid-binding protein. Finally, as preadipocytes differentiate into adipocytes, they lose their morphologic similarities to fibroblasts via decreased expression of collagen types I and III and increased production of collagen type IV, laminin, entactin, and glycosaminoglycans. In fact, the inhibition of collagen synthesis during this process actually prevents preadipocyte differentiation.²⁸ Clearly, adipose tissue is a highly complex physiological constituent whose development requires a meticulously orchestrated set of events. Therefore, it is not surprising that preliminary scientific attempts at neoadipogenesis have failed to produce consistent and conclusive results.

Models of Neoadipogenesis

Cell-Based In Vitro Tissue Engineering

Tissue engineering that involves combining adipocytes or preadipocytes with a myriad of chemical compounds and matrices or scaffolds to generate neoadipose tissue is an exciting approach to the study of adipogenesis that is currently being investigated by several groups.²⁹ One model combined preadipocytes with poly(lactic-co-glycolic acid) (PLGA), a biodegradable and biocompatible copolymer scaffold.^30,31 Preliminary results of this model were promising, but data examined at 5 months and later revealed a complete loss of both adipose tissue and scaffolding. Another study combined preadipocytes with either a hyaluronic acid sponge (HYAFF 11) or a collagen sponge matrix.^32,33 While HYAF 11 was found to be better than collagen in supporting earlier and more robust preadipocyte differentiation and replication, it remains unclear whether either of these models can produce long-lasting (>3 months) or permanent neoadipose tissue.^34,35 Reports of a novel bioabsorbable co-poly(ester amide) nonwoven matrix based on e-caprolactam, adipic acid, and 1,4-butanediol that was seeded with preadipocytes demonstrated good adherence, proliferation, and differentiation of preadipocytes.^36,37 Likewise, these promising preliminary results remain to be verified in animal models.

Another study examined 4 variations of the injectable photopolymerizable poly(ethylene glycol) matrix with respect to preadipocyte viability, adhesion and proliferation.³⁸ The poly(ethylene glycol) diacrylate scaffold alone, which is not biodegradable, performed the worst, with preadipocyte death and no proliferation observed within 1 week. The addition of adhesion sites, incorporating the laminin-binding peptide sequence YIGSR to the scaffold, produced longer preadipocyte viability but no proliferation. The third variation, a modified scaffold, which included the peptide sequence LGPA, permitting polymer degradation by cell-secreted collagense, but no adhesion peptide, produced early preadipocyte proliferation followed by death. The final scaffold variation which included both LGPA and YIGSR performed the best, enabling both preadipocyte adherence and proliferation. Though limited in longevity, this study was instrumental in demonstrating the importance of molecular modification of biocompatible matrices.

While the aforementioned models focused on scaffolding material as the variable of choice, other studies examined the interaction of preadipocytes with other cell types and different tissue culture media. Studies of different preadipocyte culture media revealed the superiority of DMEM/F12 over OPTIMEM.³⁹ The addition of fetal calf serum (FCS) to the medium was found to improve cellular proliferation, while human serum (hS) was found to induce substantial differentiation. It was further demonstrated that fibronectin-coated culture dishes not only increase preadipocyte viability but also differentiation into mature adipocytes. When preadipocytes were subjected to defined hypoxic conditions, endothelial cells were found to release a soluble factor that helped to sustain the viability of preadipocytes, confirming an intimate link between successful adipogenesis and angiogenesis.⁴⁰ Another model system using photocured, styrenated, gelatin-based microspheres (SGMs) with different drug release rates of immobilized angiogenic and adipogenic factors demonstrated that at 4 weeks, the triglyceride content in the injection site for the group that received basic fibroblast growth factor (bFGF)-, insulin-, and IGF-I–immobilized SGMs was significantly higher than that for the group that received insulin- and IGF-I–immobilized SGMs.⁴¹ These data confirm that neoadipogenesis requires both angiogenesis and differentiation of preadipocytes into adipocytes, and that these two biologic events may be induced by a single well designed biomatrix.

The previously described models all used preadipocytes as the cell of choice for fat engineering. In an attempt to explore alternative options, one group used bone marrow–derived mesenchymal stem cells (MSCs) as a cell source and combined them with custom-made PLGA scaffolds in the presence or absence of bFGF.⁴² This model not only demonstrated successful adipocyte differentiation, but also bFGF-dependent activity of glycerol-3-phosphate dehydrogenase, a marker of adipogenesis, as well as the expression of adipocyte-specific genes such as peroxisome proliferator activated receptor-gamma2 (PPARgamma2) and glucose transporter-4 (GLUT4). This study verified that an alternative cell type, in combination with bFGF, may represent a complementary approach to adipose tissue engineering.

In vitro models that have attempted to produce the ideal regenerative-based soft tissue filler in the form of adipose tissue have significantly advanced our understanding of neoadipogenesis. We have learned from these studies that neoadipogenesis is a complex process that requires several steps. First, viable preadipocytes (or other adipocyte precursor cells) must be obtained. These cells should be combined with a biocompatible and bioabsorbable matrix that contains preadipocyte adhesion sites for cellular adhesion. Subsequently, precursor cell differentiation into adipocytes must occur, a step which can be optimized by combining the scaffold with angiogenic and adipogenic factors. Finally, in order to promote viability during these steps as well as longevity of the neoadipose tissue, the entire process must take place in an environment favorable to angiogenesis. Nevertheless, despite these advances, a consensus on the exact optimal parameters for engineering adipose tissue remains to be established, and many of these in vitro models have yet to be tested in animal systems.⁴³

Animal Models

The success of in vitro adipose tissue engineering models paved the way for animal models of neoadipogenesis. Data from animal model studies have recently begun to emerge, bringing plastic surgeons one step closer to the clinical application of these techniques. The requirement for concurrent angiogenesis during neoadipogenesis was demonstrated in previously discussed in vitro models. This observation was corroborated by multiple animal studies, including one of the earliest, the 3T3-F442A model of adipose tissue development, in which 3T3-F442A preadipocytes were implanted subcutaneously into nude mice.⁴⁴ The implanted cells developed into highly vascularized fat pads, over 14 to 21 days, which were morphologically similar to normal subcutaneous adipose tissue. Neoangiogenesis was observed as early as 5 days after cell implantation and the expression of endothelial cell markers and adipogenesis markers increased in parallel during fat pad development. Finally, this study suggested that the neovasculature was initiated by sprouting from larger, host-derived blood vessels rather than from endothelial progenitor cells. Analogous results were observed from a nude mouse model of epididymal fat grafting in which angiogenesis was inhibited by the antiangiogenesis factor TNP-470, resulting in graft volume reduction and impairment of adipogenic gene expression.⁴⁵ When endothelial progenitor cells (EPCs), which are known to incorporate into active sites of angiogenesis and augment collateral vessel growth in ischemic tissues, were added to transplanted fat grafts in nude mice, a significant increase in graft volume, fat quality, and capillary density was observed.¹ A similarly positive effect on neoadipogenesis quality was observed when the proangiogenic cytokine IL-8 was added to the fat grafts before injection into nude mice.⁴⁷ Further evidence for the importance of neoangiogenesis in neoadipogenesis was produced by a study in which nude mice transplanted with 1 ml of human free fat harvested by suction-assisted lipectomy were treated with adenovirus-mediated vascular endothelial growth factor (Ad-VEGF).⁴⁸ At 15 weeks posttransplantation, mice treated with the proangiogenic factor Ad-VEGF demonstrated enhanced fat survival and higher quality fat grafts when compared with those not treated with the vector.

As in the in vitro studies, animal models also examined the efficacy of different biomatrices in generating neoadipogenesis. In one study, gelatin microspheres containing bFGF that were prepared for the controlled release of bFGF, suspended with human preadipocytes, and incorporated into a collagen sponge scaffold were subcutaneously implanted into the back of nude mice.⁴⁹ Significant adipose tissue formation composed of human matured adipocytes was observed at 6 weeks when the combination of gelatin microspheres containing 1 μg per site of bFGF and 1.0 × 10⁵ cells/site of preadipocytes with the collagen sponge was used. Another nude mouse model tested intradermally injected gelatin spheres that had either been preseeded with human fibroblasts or preadipocytes or left unseeded.⁵⁰ After 56 days, adipose tissue generation with developing neoangiogenesis was elicited by all 3 gelatin sphere experiments. However, the effect of spheres preseeded with preadipocytes surpassed the effect of spheres preseeded with fibroblasts, which in turn surpassed the effect of unseeded spheres. The results of this study surprisingly imply that macroporous gelatin spheres alone could be used for the correction of minor soft tissue defects. In a complementary study, a collagen sponge scaffold incorporated with gelatin microspheres with different water contents for the controlled release of bFGF was injected with syngeneic rat preadipocytes (1 × 10⁵ cells/site) into a defect of the rat fat pad.⁵¹ By 4 weeks, neoadipogenesis with concurrent angiogenesis was observed with the microspheres containing 1.0 μg of bFGF. Remarkably, as was seen in the previous study, the addition of rat syngeneic preadipocytes did not promote adipogenesis, suggesting that the collagen scaffold when combined with an appropriate controlled release bFGF was able to achieve the in situ formation of adipose tissue even without preadipocytes.

A study of different concentrations of human preadipocytes in fibrin injected into athymic nude mice examined long-term adipose tissue formation (up to 9 months) without the use of an additional scaffold.⁵² The most robust and stable long-term neoadipogenesis was observed when 30 million preadipocytes were implanted. In contradistinction to other studies in which the acellular scaffolding was able to generate neoadipogenesis, this report demonstrates that long-term stable adipose tissue can be engineered in vivo by simple injection of human preadipocytes using fibrin as a carrier material without a biologic scaffold.

Another study used the pig as an animal model and examined hyaluronan gels mixed with autologous undifferentiated preadipocytes.⁵³ Pig preadipocytes were isolated from intraabdominal pig fat by collagenase digestion, plated on fibronectin-coated culture dishes in Dulbecco's Modified Eagle Medium/Ham's F12 combined with 10% pig serum, expanded, and mixed with twp types of hyaluronan gel with varying degrees of amidation of the carboxyl groups (HYADD3 and HYADD4) and injected into pig ears. At 6 weeks, islets of mature adipocytes and vessels embedded in fat tissue surrounded by gel were observed in the HYADD3 gel, but not in the HYADD4 gel. When HYAFF 11 sponges were seeded with human preadipocytes with or without the coating of glycosaminoglycan hyaluronic acid and implanted into nude athymic mice, there was good penetration of the cells into the matrix, but little or no adipose tissue formation at any time points tested.⁵⁴

The differentiation of preadipocytes or other precursor cells into adipocytes remains a challenging step that has yet to be refined by in vitro or in vivo models of neoadipogenesis. An investigation of whether the implantation of adipogenic-differentiated preadipocytes enhances the adipose tissue formation when compared with implantation of undifferentiated preadipocytes in subcutaneous spaces of athymic mice revealed that adipogenic- differentiated preadipocytes resulted in more extensive adipogenesis than the implantation of undifferentiated preadipocytes.⁵⁵ As expected, this study further confirmed that bFGF enhances not only neovascularization in newly formed adipose tissues, but also neoadipogenesis.

The mechanical effect of dome-shaped scaffold support structures was examined in the subcutaneous pockets of athymic mice.⁵⁶ Biodegradable synthetic polymers fabricated from polyglycolic acid fiber-based matrices with polyL-lactic acid were combined with human preadipocytes in a fibrin matrix. Six weeks after implantation, regeneration of adipose tissues was observed in the group in which preadipocytes were injected into the space under the support structures, suggesting that adipose tissues may be augmented with volume conservation.

While the scope of this review prevents a discussion of all neoadipogenesis animal model studies, we have presented here data representative of the studies currently underway and hereby draw several important conclusions from the animal data. First, neovascularization is critical for optimizing the quantity and quality of neoadipogenesis. When combined with injected preadipocytes, bFGF acts as a strong molecular promoter of adipogenesis. The injection of preadipocytes that have been differentiated in culture media may produce more robust neoadipose tissue than the injection of undifferentiated preadipocytes. And finally, somewhat surprisingly, certain matrices, including gelatin microspheres and dome-shaped polyglycolic acid fiber-based matrices with polyL-lactic acid, are capable of inducing (minor) amounts of neoadipose tissue sans cellular supplementation. In contradistinction, long-term stable adipose tissue can be engineered by injection of human preadipocytes using fibrin as a carrier material without a biologic scaffold.

Refinement of Fat Harvesting, Preparation, and Grafting Techniques

In the last 20 years, although several different techniques of lipoinjection have been developed, a standard procedure has not been adopted by all practitioners. Variables that remain to be settled include (1) the ideal cannula and technique for harvesting; (2) the best way of processing the fat to ensure maximal take and viability of the graft; and (3) the best technique for reinjecting the fat. Investigations into the optimal technique of fat harvesting, preparation, and grafting techniques have utilized both in vitro and in vivo models of neoadipogenesis. Common harvesting approaches include syringe aspiration and lipoaspiration.^8,57,58 Once the fat is harvested, it is often prepared for injection via one of several methods, including washing with physiologic buffers, centrifugation for separation of cells from debris, decantation, or by concentrating it using cotton towels or other absorbent media.^6,59–63 The fat is subsequently injected into the subcutaneous tissues with an assortment of delivery methods using either sharp or blunt needles. A recent literature review of the techniques of autologous fat transfer concluded that while scientific data fails to provide definitive evidence of fat survival among various techniques, there is a modest preference in favor of the following approach: harvesting abdominal fat with a blunt cannula technique followed by centrifugation without washing or addition of growth factors and immediate injection of small amounts of fat by means of multiple passes.⁶⁴

The lack of consensus on the best fat grafting technique stems in large part from equivocal results obtained when multiple methodologies were compared. For example, one study evaluated 2 different harvesting and 6 different preparation techniques in the severe combined immunodeficient mouse model.⁶⁵ Fat was harvested either via 10-cc syringe aspiration or lipoplasty using the standard Coleman cannula and prepared with one of 6 methods: (1) centrifugation alone; (2) washing with Lactated Ringer's solution alone; (3) washing with normal saline alone; (4) washing with Lactated Ringer's solution and centrifugation; (5) washing with normal saline and centrifugation; and (6) no treatment. Samples were injected into the flanks of severe combined immunodeficient mice and analyzed at 12 weeks via the XTT assay. No differences in cell viability were found between the 2 harvesting techniques or among the 6 preparation methods, with all samples demonstrating 40% to 50% viability at 12 weeks. Another study examined 2 different methods of fat harvesting, lipoplasty cannula versus a blunt needle adapted to a fine-needle aspiration apparatus, as well as 2 techniques of tissue processing, decantation versus cotton towel drying.⁶⁶ Samples were processed for culture, and adipocytes and preadipocytes were plated in culture medium and expanded in vitro. Fat viability was found to be better when fat was harvested by fine needle aspiration and the highest number of viable preadipocytes obtained when these harvests were dried on a cotton towel. Unfortunately, this study was limited to an in vitro approach and its clinical applicability thus markedly limited. When another group compared samples prepared either via centrifugation or cotton towel drying in a nude mouse model, no differences in fat weight or volume were observed, yet the towel separated samples demonstrated decreased fibrosis on histologic analysis.⁶⁷ While the clinical implication of these findings remains to be established, these studies suggest that cotton towel drying may be a more delicate method of fat graft preparation that could translate into improved fat graft quality.

A novel mouse model was recently developed in an attempt to optimize and standardize data acquisition on fat graft harvest and preparation.⁶⁸ In this model, mouse inguinal fat pads are used as donor grafts and subjected to different harvest and preparatory techniques. Viability and purity were highest after excisional harvest versus blunt or needle harvest, and additionally improved after saline wash or centrifugation. Interestingly, grafts treated with an initial collagenase digestion followed by an idealized centrifugation regimen and a single wash step consistently demonstrated 96% viability and 93% purity irrespective of harvest technique. The effect of different harvesting methods was examined in combination with the effect of the free radical scavenger coenzyme Q₁₀ on adipocyte viability in another recent study.⁶⁹ Fat grafts were harvested from 6 patients using either the Coleman syringe technique or the lipoaspiration technique. Half the samples were treated with coenzyme Q₁₀ and all examined by immunoenzymatic, biochemical, and morphological methods. Lipoaspiration was found to decrease adipocyte viability while the treatment with coenzyme Q₁₀ was found to reduce and even inhibit adipocyte apoptosis the cultured tissues. While these studies present intriguing suggestions for modifications in fat preparation, it remains to be determined whether these in vitro data will translate into clinically significant improvements.

Another controversial variable in the technique of fat grafting is the ideal donor site. When harvests from 4 commonly used donor sites (abdominal fat, thigh fat, flank fat, or knee fat) from 5 subjects were evaluated, no immediate differences in adipocyte viability were observed.⁷⁰ These results were corroborated by another study in which no differences in fat viability were observed when fat was harvested from 3 donor sites from one subject (thigh, abdomen, and breast) and injected into a nude mouse model.⁷¹ These equivocal results remain unchallenged and thus at present there are no indications that one donor site is superior to another.

A final question that remains to be addressed is that of fat graft preservation using different storage techniques. In an attempt to answer this question, one study examined aspirated fat preserved either at room temperature for 1, 2, 4, and 24 hours (n = 10 each), at 4°C for 1, 2, and 3 days (n = 14 each), or at −80°C for 1 month (n = 3).⁷² Although scanning electron microscopic assays demonstrated no anatomic changes among the different preservation methods, decreases in adipose stem cell isolation yields were observed with both room temperature storage and cryopreservation at almost all time points tested, and after 24 hours of preservation at 4°C. In contradiction, when the impact of cryopreservation on adipogenic differentiated preadipocytes was examined by comparison of freeze-thawed and fresh preadipocytes in vitro and in vivo, there were no volumetric, molecular, or morphologic differences observed, suggesting that cryopreservation is a feasible method of fat graft preservation.⁷³ Clearly, additional studies are required for a more definitive resolution of the fat graft storage question.

Conclusions

Limitations of Current Data and Study Techniques

We have reviewed here the latest advances in the basic science of neoadipogenesis, adipocyte viability, and its clinical application to fat grafting. As demonstrated, data produced by in vitro and in vivo studies fail to produce a cohesive image of the required components for successful, consistent, and durable fat transplantation. This is in part because of the methodologic flaws from which most of the studies suffer. For example, there are no universally accepted in vitro or in vivo models of neoadipogenesis. In contrast, most of the analyses are carried out using remarkably different tissue culture systems and mouse models. Furthermore, the assays, donor and recipient sites used to examine fat graft, viability, volume, and quality are equally diverse. For every result produced by a given study that was reproduced by a second study, there are many more results which are contradictory or inconclusive. Unfortunately, it remains unclear whether this is related to a lack of systematic scientific data acquisition, or whether these findings truly reflect the biology of neoadipogenesis.

Future Directions

As a first step in strengthening scientific data collection on autologous fat grafting, we recommend that a collective, multidisciplinary, multicenter effort be undertaken to establish in vitro and in vivo models of neoadipogenesis that are clearly reproducible from one investigator to another. In vitro culture systems must use the same reagents, culture media, duration of differentiation, etc. Similarly, in vivo mouse models must be carefully systematized with respect to genotype, age, sex, zone of transplant injection, and duration of experiments. For example, is 12 weeks the ideal amount of time to assess fat grafting in a mouse that has a life span of 9 to 12 months? Experiments with these systems should be analyzed with equivalent assays and use sufficiently large numbers of mice or tissue samples for adequate statistical power to distinguish minor experimental variables. The models should simulate the molecular and physiological aspects of adipogenesis as closely as possible and must be sufficiently user-friendly so as to be embraced by both national and international communities. Furthermore, they should be designed so that data from the in vitro model can be converted to and easily tested in vivo, and subsequently in human trials. Once in place, these models should first be used to test the basic variables that have plagued the fat grafting community for so many years. These include the most effective type of harvest, preparation, and injection techniques, the most effective long-term storage of harvested fat, etc.

Subsequently, many questions will remain to be answered regarding autologous fat grafting, some of which have been previously tackled, and others that will require elucidation in the future. These should likewise be addressed in a systematic manner. First, the lipogenic activity of donor site (abdomen, thigh, knee, flank, etc.) and donor age and weight (underweight, normal, or overweight) should be established, using a large number of donors. Then, a range of recipient sites (nasolabial folds, periorbital zones, hands, breast, posttraumatic sites, etc.) should be studied because motion, scarring, previous trauma, radiation, and other factors surely play a role in successful transplantation. These variables should then be cross tested against one another in a methodical manner in order to assess possible interactions among them. For example, is abdominal fat better at reconstructing a radiation-induced defect? Does fat from an obese donor have more lipogenic activity and thus last longer in a mobile recipient such as the hands? Does fat from an older underweight donor have lower viability, requiring multiple frequent applications? The answer to these questions would permit the generation of practice guidelines for specific clinical scenarios. Once these general parameters are established, secondary issues can be examined. The benefit of growth factors, cytokines, or other molecular agents should be elucidated, and likewise, the addition of matrices or scaffolds.

Without a doubt, clinical studies will be required to confirm data generated from in vitro and animal model studies and, further, to address questions that may not be amenable to laboratory systems. Such studies could include longitudinal investigations on long-term fat transplant viability, the need for overcorrection based on recipient site or type of defect, volumetric assessment using three-dimensional cameras or magnetic resonance imaging, and the effects of smoking or other disease states on graft take. Finally, as novel molecular data and innovative technologies become available, they should be applied to the study of neoadipogenesis in order to encourage the evolution of fat transplantation. For example, one of the most exciting novel areas of investigation in cosmetic breast augmentation and breast reconstruction is the use of vacuum-based soft-tissue envelope expanders such as the BRAVA external breast tissue expander system (BRAVA LLC, Coconut Grove, FL).^74,75 These systems, which are thought to stimulate soft-tissue envelope expansion as well as local neovascularization, have been hypothesized to be a perfect complement to fat grafting, because the newly expanded soft tissue envelope would serve as a much more hospitable environment to fresh fat grafts. This combination would be particularly important for the reconstruction of defects secondary to radiation, trauma, burns, other scarring, or long-term contracture, which can be among the most difficult defects to treat for plastic surgeons. Clearly, with the implementation of systematic scientific approaches to the study of neoadipogenesis, the future of autologous fat transplantation for correction of soft tissue volume loss promises to be dazzling.

Disclosures

The authors have no disclosures with respect to the contents of this article.