Purified Viable Fat Suspended in Matrigel Improves Volume Longevity

Abstract

Background: Autologous fat is an excellent soft tissue filler with cosmetic and reconstructive utility. However, graft longevity is unpredictable.

Objective: This study sought to evaluate the effect on in vivo fat graft performance of contemporary adipocyte tissue engineering techniques that have not previously been applied to mature fat cells due to the difficulty of their purification and their high metabolic demand.

Methods: Using a recently reported protocol, the adipocyte viability and purity of lipo-harvested fat were optimized. Before graft administration, these purified cells were suspended in GFR-Matrigel (BDBiosciences), a basement membrane protein matrix known to improve early angiogenesis. It was posited that by suspending the purified cells in this resorbable matrix, the high metabolic demand of these cells would be met and graft performance could be improved. The in vivo longevity of these tissue engineered fat grafts was tested in a murine model in which each subject received posterior subcutaneous injections of three types of fat graft: unpurified fat after lipo-harvest alone; fat harvested in identical fashion, but purified and suspended in GFR-Matrigel; and a control of GFR-Matrigel alone. Graft volumes and quantitative histologic characteristics were examined at 1 week, 1 month, and 3 months.

Results: At 3 months, purified fal/GFR Matrigel grafts showed superior fat volume maintenance (80.2% versus 29.7% for unpurified grafts [P < .05]) and adipocyte cellular longevity (70.1% versus 45.6% [P < .001). Unpurified grafts were largely replaced by fibrosis at 3 months (96.5% [95% CI 0.90–0.970]), despite starting with three times as many viable adipocytes as purified grafts. A correlation was noted between the poor performance of unpurified grafts and a disproportionate presence of early inflammation in fat grafts prepared without purification techniques.

Conclusions: A preparatory regimen consisting of a preadministration purification followed by cellular suspension in a resorbable protein matrix may ultimately improve the predictability and longevity of autologous fat grafts.

Autologous fat grafting was first introduced by Neuber to the German Surgical Society in 1893, and was used for the purposes of soft tissue augmentation in 1911 by Bruning. The advent of liposuction techniques in the 1980s simplified harvest and led to widespread clinical use. Autologous fat has the potential for being the ideal soft tissue filler because it possesses a natural consistency, is easy and safe to harvest, has no hypersensitivity or foreign body reaction and is readily available. Nonetheless, fat grafting is limited by unpredictable survival, with longevity reported between 3 months and 8 years.¹–¹⁶ This variability in outcome may stem in part from a lack of clinical procedural standardization.

Harvest methods vary greatly among clinicians, and comparative studies between harvest methods have also been inconsistent. Graft preparatory techniques after harvest are equally variable.

Graft three-dimensional structure appears important, but its precise role remains undefined. Grafts composed of surgically excised fat has been shown to maintain its volume better than suction-assisted fat grafts (42.2% vs 31.6%).¹⁷ Kononas et al showed similar results when they evaluated the fat “pearl” versus fat “cell” graft and reported greater long-term augmentation with pearl technique at 12 weeks in a rabbit model. The high metabolic demands of adipocytes, as well as their fragile nature, have also been cited as limiting factors in the survival of fat grafts.

Current tissue engineering research has focused less on mature adipocytes and more on their precursors because of their stability, low metabolic rate, and potential for growth. Preadipocytes, when combined with resorbable protein matrixes or plated on scaffolds have been shown in the short term to develop into fat pads.¹⁸,¹⁹ This growth maintenance has been limited by the longevity of the matrix in vivo, however.²⁰ The clinical utility of preadipocytes is currently limited by a time-consuming and expensive process of cellular isolation. Mature adipocytes are potentially advantageous because of their abundance, ease of harvest, and known volume. But these cells have not been fully evaluated with current tissue engineering technologies.

We sought to improve fat graft outcomes by applying these newer tissue engineering concepts to adult adipocytes. A protocol for the reliable isolation of a pure suspension of viable adipocytes was recently described, producing purity and viability of 94% and 97%, respectively (Figure 1).²⁰

To improve initial graft revascularization, and thus satisfy the high metabolic demands of these cells, suspension of the pure single cell suspension in a resorbable protein matrix was performed. GFR Matrigel (Laminin, Collagen IV, Heparan Sulfate), (BDBiosciences, San Jose, CA) was chosen because of its ability to mix as a liquid with cells at room temperature, but to form a suspension gel at body temperature thus providing a resorbable three-dimensional structure in vivo; additionally, Matrigel is known for its potential for hastening revascularization.²¹

Materials and Methods

Twelve genetically identical, age- and sex-matched mice were used as subjects. Fat was harvested and prepared in a sterile manner from mice as described by Piasecki et al.²⁰ Each subject received 3 grafts of equal volume but different composition placed subcutaneously in one of 3 standardized locations on the mouse's back (Figure 2). Graft types included the following: (1) Unpurified fat: lipo-harvested fat alone, not subjected to the purification regimen; (2) Purified Fat/GFR Matrigel: fat harvested in identical fashion to the unpurified fat, but then subjected to the purification regimen, and then suspended in a 1:3 volume ratio with GFR Matrigel; (3) GFR Matrigel alone: as a control. A 10-mL syringe with 18-gauge needle and standardized pressure was used to harvest fat. Half of this fat was prepared per the UW protocol as outlined by Piasecki et al.²⁰ A total of 33 μL of purified fat for each graft was suspended in 67 μL of GFR Matrigel evenly at 20°C, then made gelatinous by continuous rotation at 37°C for 15 minutes. Viability, purity, and absolute number of viable cells were calculated for clinical control fat and UW/Matrigel grafts.

Surgical subjects were anesthetized with subcutaneous fentanyl and intraperitoneal etomidate. Posterior trunks were shaved atraumatically, and surgical sites were prepped and draped sterile. Midline marks were made with a surgical marker: starting just cranial to the confluence of the scapulas and then in sequence at 1-cm intervals caudal to this point. With a sterile technique, 100-μL aliquots of fat graft were placed in the subcutaneous plane directly under these marks (Figure 2). Unpurified fat, purified fat/GFR Matrigel, and GFR Matrigel alone grafts were randomized to different sites in each mouse such that any advantageous or deleterious effects related to anatomic placement would be statistically removed.

Volume measurements through external measurements were taken at 1 week, 1 month, and 3 months. Histologic sections were taken at 3 months, and slides were prepared with Hematoxylin and Eosin. These slides were analyzed at original magnification × 20 by use of Axiovision software (AxioVision, Carl Zeiss AG, Bernried, Germany) to calculate percentage of graft volume occupied by fibrosis, intact fat cells, and Matrigel. Additionally, the number of blood vessels in each graft was calculated on histologic section at original magnification × 20.

Results

Preoperative counts of viable adipocytes showed that purified/GFR Matrigel grafts started with one-third as many total viable cells as unpurified grafts (55 × 10⁴ cells/graft [95% CI {49-61}] versus 195 × 10⁴ cells/graft [95% CI {150-240}]; [Figure 3]). Despite this, total external graft volumes were not statistically significant between these two groups at 3 months (P > .05) (Figure 4). Interestingly, unpurified grafts showed a statistical initial increase in volume at 10 days, which correlated histologically with high levels of inflammatory cell count and necrosis (Figure 5). GFR-Matrigel control grafts were completely resorbed over 2 to 3 months (Figure 6), a process also observed histologically, involving the GFR-Matrigel that had been mixed with the purified fat cells in the purified/GFR-Matrigel grafts. When volumes were plotted as a percentage of the volume of fat per graft at the time of graft placement, purified/GFR-Matrigel grafts showed a statistically significant 80.8% (95% CI [0.58-1.050]) maintenance of fat volume at 3 months, versus 20.5% (95% CI [0-0.44]) for unpurified grafts (Figure 7). Quantitative histologic study (Figure 8) revealed that purified/GFR-Matrigel grafts were 70.2% (95% CI [0.58-0.72]) composed of intact fat cells at 3 months, versus only 4.6% (95% CI [0.02-0.07]) for unpurified grafts (Figure 9). With regard to tissue fibrosis, the reverse was true (Figure 10): unpurified grafts were 96.2% (95% CI [0.94-0.98]) composed of fibrotic material at 3 months, compared with 22% (95% CI [0.21-0.23]) for purified/GFR-Matrigel grafts (Figure 11). As mentioned above, GFR-Matrigel was largely resorbed (Figure 12), occupying only 7% (95% CI [0.04-0.11]) of purified/GFR-Matrigel grafts at 3 months (Figure 13). Interestingly, GFR-Matrigel controls showed no sign of tissue ingrowth or vascularization during their linear resorption over 6 weeks (Figure 6). A disproportionate presence of early inflammation and necrosis was observed histologically in unpurified grafts versus purified/GFR-Matrigel grafts at 10 days (Figure 14; see also Figure 5), which correlated with the increase in external graft volume observed at 10 days (Figure 4).

Discussion

This study should be viewed as a preliminary report because it was performed in a small animal model, not a human one. Because the subjects are mice, there is no reliable way to correlate mouse graft longevity with longevity in human beings, and we caution readers from doing this. For instance, the duration of follow-up was chosen with respect to the lifespan of the mice—3 months in a mouse is different than 3 months in a human being. Although we stand by our results as significant, future studies in human beings are necessary and would most certainly need longer follow-up to reflect any true clinical relevance.

With respect to the results of the study, however, we believe they are noteworthy because purified fat/GFR-Matrigel grafts started with one third the volume of fat as their unpurified fat graft counterparts and maintained 80% of this starting volume after the suspension matrix had resorbed; unpurified fat grafts maintained only 20% of their fat volume despite starting with three times as many viable fat cells. Moreover, the volume that was maintained in the unpurified fat grafts at 3 months was largely (96%) composed of fibrosis. Also noteworthy is the fact that the tissue-engineered grafts appeared to achieve what their predecessor preadipocyte grafts previously described in the literature have not—they maintained their fat volume even after the loss of their original suspension matrix (Figure 6).

Clearly, the purified fat cells mixed with GFR-Matrigel appeared to improve fat volume maintenance and the relative percentage of intact fat cells present at 3 months. What is not clear is how much of this effect (if any) was caused by the initial presence of GFR-Matrigel, the purification process, or both. For instance, if Matrigel is capable of enhancing tissue ingrowth from surrounding mesenchymal elements/fat progenitor cells, the presence of the purified fat cells may be a superfluous. Indeed, growth factor–rich Matrigel mixed with FGF and connected to a pedicled blood supply has been shown to produce de novo fat pads in vivo.²² However, the Matrigel used in this study was “growth factor reduced” containing 10-fold lower concentrations of growth factors than the established in vitro thresholds for adipocyte growth and differentiation, suggesting that GFR-Matrigel should not have an intrinsic ability to develop in vivo into a fat pad by itself.²³ Additionally, the relatively unceremonious resorption over 2 to 3 months of the GFR-Matrigel control grafts in this study, with no sign histologically of tissue or vascular ingrowth of any kind, argues against the suspension matrix playing the only role in the observed improved performance of these grafts.

Because the purified grafts (GFR-Matrigel mixed with purified fat cells) did show intense vascular ingrowth, it is most likely that both factors (the presence of fat cells, and the presence of the matrix) played a role. In the cellular biology literature, Matrigel has been shown to corral endogenous cytokines secreted by the cells that are suspended in it, effectively increasing their local concentration.²⁴ Perhaps endogenously secreted growth factors and angiogenic factors from the suspended adipocytes were more efficiently preserved in this “bioreactor,” improving vascular ingrowth and thus cellular survival. Unfortunately, growth factor concentrations were not assessed in this study.

The differential presence of vascular ingrowth in purified/GFR-Matrigel versus unpurified grafts was also profound (Figure 15). Another potential role of the matrix may have been the provision of a three-dimensional scaffold for cytokine/nutrient exchange and/or ingrowth of blood vessels. Although Matrigel itself may not be an ideal delivery matrix for fat cells (given its historical use in oncology research), any matrix chosen will likely need to persist long enough for these processes to occur. The ideal length of matrix longevity is yet to be determined.

Certainly, the role of cellular purification is debatable, given the fact that the performance of unpurified fat mixed with GFR-Matrigel was not assessed (out of a desire to have a control graft prepared in a more conventional fashion as a form of benchmark to compare outcomes to). We chose the study arms as described because we wanted to evaluate experimental grafts with respect to current clinical practices (ie, unpurified fat—current clinical practices are not to mix fat with Matrigel). Moreover, unpurified fat is not homogenous and would not mix evenly throughout Matrigel, adding experimental bias and lack of control—the purified fat cells evenly disperse in Matrigel in a consistent and reproducible manner making it more appropriate to compare results between groups. If these discrepancies could be worked out, future studies with different study arms would be interesting.

Of additional significance is the fact that the unpurified grafts—despite starting with three times as many viable adipocytes—showed statistically significant volume loss compared with purified grafts and on histologic sections were shown to be almost entirely replaced by fibrotic material at 3 months. These observations clearly suggest a revised understanding of what happens to fat grafts prepared in a conventional fashion—that in some clinical settings, the long-term volume of tissue provided by fat grafts may not necessarily be caused by fat, but rather from scar tissue. These findings, in combination with the discovery that these grafts were associated with a statistically significant presence of early inflammation and necrosis at 10 days, stress the potential important role of “all the other stuff” that accompanies the live fat cells within a standard fat graft. An intense inflammatory reaction produced by this dead debris may not only lead to fibrosis, but also may lead to the death of initially viable cells, and would argue for the beneficial effects of a purification regimen. It also helps explain to some degree why fat grafts may be so variable; specifically, if harvested fat is placed in a syringe and layers itself to some degree within the syringe, the content of graft placed in different graft aliquots may have different percentages of viable cells, dead cells, and cell debris, ultimately leading to varying degrees of inflammation, cellular survival, and ultimately graft performance.

It is prudent here to also make mention of the plane of administration. In this study, the fat grafts were placed in the subcutaneous layer. Most clinicians would agree that subcutaneous fat is not as vascularized as other tissues, prompting some surgeons to inject smaller-volume fat grafts into muscle or dermis. Certainly if possible, administering grafts into a more vascular bed is preferred. However, the most vexing clinical challenges that face plastic surgeons are those that require larger-volume subcutaneous administration (facial augmentation, breast reshaping, and liposuction contour irregularities). As such, we sought to study graft survival capabilities in this more difficult and challenging location; because if research can better characterize how fat grafts behave/survive in a subcutaneous layer, we will be in a better position to affect positive change for bigger clinical challenges in the future.

Readers accustomed to reading clinical studies might criticize these conclusions on the basis of the number of subjects used. Certainly, any scientific conclusion is afforded more weight with a larger sample size. However, in this study the number of mice was chosen to generate the appropriate statistical power of 0.80 and was validated by the statistical significance demonstrated (P < .05). If hypothetically, no statistical significance had been shown, it could be argued that in truth of fact, there was a difference between treatment groups, but that this difference was not shown because of the small sample size. This was not the case with this study since P values reflected statistical significance with very little variance within groups—increasing the sample size would likely have simply made small P values even smaller. Certainly more subjects would lend even more weight to the conclusions—but because statistical significance was reached (P < .05), the conclusions are still statistically and thus scientifically valid. In summary, by applying cell isolation and tissue engineering concepts to mature adipocytes, insight into the physiological behavior of autologous fat grafts—and avenues for improved performance—were established. This article evaluated for the first time the role of resorbable suspension matrixes in the context of clinical fat grafting (ie, with adult fat cells, not preadipocytes). This is important because prior studies have shown mixed results with stem cell precursors, but none has evaluated mature cells—and mature cells are the ones clinicians are using every day. Moreover, the results shed light on the negative effect of inflammation in grafts prepared to the current clinical standard and the absence of this in purified fat grafts. A clinically practical technique for creating a pure single cell suspension of viable adipocytes that is reproducible, regardless of tissue harvest method was developed, addressing the cellular fragility that has previously limited precise cellular purification. Additionally, proinflammatory (and potentially cytotoxic) cellular debris and dead cells were removed by this regimen. By suspending these purified cells in a resorbable protein matrix, graft structure, and early revascularization may have been optimized, satisfying the high metabolic demands of these cells. Grafts prepared in this fashion showed greater in vivo fat volume longevity, less inflammation, and greater adipocyte survival than controls. Additionally, histologic analysis suggested that volume maintenance of grafts prepared in a conventional fashion was composed of fibrotic material, not surviving fat cells, implying an important role for preadministration purification. Because these findings were made in a basic science mouse model, they are not ready for “prime time” human use. Only human research can do this. However, the conclusions do suggest that with minimal resources and time, the possibility exists that fat graft outcomes can be improved by applying current tissue engineering concepts to mature cells. Future research into the importance of matrix resorption characteristics, viscosity/pore size, and growth factor concentrations are necessary before the ideal suspension matrix for tissue engineered mature fat grafts is found and these techniques reach clinical fruition.

Conclusion

Free mature fat grafts are currently unpredictable with regard to volume maintenance. Adipocyte tissue engineering strategies have focused on precursor cells and have been unable to show long-term persistence of adipose tissue after the degradation of suspension matrix material. In this study, the application of a preadministration purification regimen to harvested fat, followed by suspension in GFR-Matrigel was associated with improved fat volume maintenance and histological cellular longevity, even after the resorption of the suspension matrix. Despite starting with nearly threefold as many viable adipocytes, unpurified grafts were composed of statistically less intact fat at 3 months, largely replaced by fibrosis; slower vascular ingrowth and the potential deleterious effects of a disproportionate presence of early inflammation may have been contributing factors. Future research into the relative role of matrix characteristics on in vivo performance, the effect of growth/angiogenic factors, and ultimately correlate studies in larger animal models are warranted.

Disclosures

This project was funded entirely by divisional resources from the University of Wisconsin Hospital and Clinics, Division of Plastic Surgery. None of the authors has a financial or commercial interest in any products, devices, or materials described that could pose or create a conflict of interest.

References

Additional Readings