Journal of Plastic, Reconstructive & Aesthetic Surgery (2012) 65, 219e227

Fat grafting accelerates revascularisation anddecreases fibrosis following thermal injury*

Steven M. Sultan a, Jason S. Barr a, Parag Butala a, Edward H. Davidson b,Andrew L. Weinstein a, Denis Knobel a, Pierre B. Saadeh a,Stephen M. Warren a, Sydney R. Coleman a, Alexes Hazen a,*

a Institute of Reconstructive Plastic Surgery, New York University Medical Center, 560 First Avenue, TCH-169, New York,NY 10017, USAbUniversity of Pittsburgh Division of Plastic Surgery, University of Pittsburgh Medical Center, Pittsburgh, PA, USA

Received 21 April 2011; accepted 24 August 2011

KEYWORDSBurn;Fat graft;Stem cell

* Presented in part at: iFats Annual Ce Washington D.C., USA, October 201* Corresponding author. Tel.: þ1 212E-mail address: alexes.hazen@nyu

1748-6815/$-seefrontmatterª2011Bridoi:10.1016/j.bjps.2011.08.046

Summary Background: Fat grafting has been shown clinically to improve the quality of burnscars. To date, no study has explored the mechanism of this effect. We aimed to do so bycombining our murine model of fat grafting with a previously described murine model ofthermal injury.Methods: Wild-type FVB mice (nZ 20) were anaesthetised, shaved and depilitated. Brass rodswere heated to 100 �C in a hot water bath before being applied to the dorsum of the mice for10 s, yielding a full-thickness injury. Following a 2-week recovery period, the mice underwentDoppler scanning before being fat/sham grafted with 1.5 cc of human fat/saline. Half weresacrificed 4 weeks following grafting, and half were sacrificed 8 weeks following grafting. Bothgroups underwent repeat Doppler scanning immediately prior to sacrifice. Burn scar sampleswere taken following sacrifice at both time points for protein quantification, CD31 stainingand Picrosirius red staining.Results: Doppler scanning demonstrated significantly greater flux in fat-grafted animals thansaline-grafted animals at 4 weeks (fat Z 305 � 15.77 mV, saline Z 242 � 15.83 mV;pZ 0.026). Enzyme-linked immunosorbent assay (ELISA) analysis in fat-grafted animalsdemonstrated significant increase in vasculogenic proteins at 4 weeks (vascular endothelialgrowth factor (VEGF): fatZ 74.3 � 4.39 ng mle1, salineZ 34.3 � 5.23 ng mle1; pZ 0.004)(stromal cel l-derived factor-1 (SDF-1): fat Z 51.8 � 1.23 ng ml e1, sal ine graf-tedZ 10.2� 3.22 ngmle1; p< 0.001) and significant decreases in fibrotic markers at 8 weeks(transforming growth factor-ß1(TGF-ß): salineZ 9.30 � 0.93, fatZ 4.63 � 0.38 ng mle1;p Z 0.002) (matrix metal lopeptidase 9 (MMP9): sal ine Z 13.05 � 1.21 ng ml e1,

onference e Dallas, TX, USA, October 2010, Northeastern Society of Plastic Surgeons Annual Meeting0.263 8745; fax: þ1 212 263 2138.

mc.org (A. Hazen).

tishAssociationofPlastic,ReconstructiveandAestheticSurgeons.PublishedbyElsevierLtd.All rightsreserved.

220 S.M. Sultan et al.

fatZ 6.83� 1.39 ng mle1; pZ 0.010). CD31 staining demonstrated significantly up-regulatedvascularity at 4 weeks in fat-grafted animals (fatZ 30.8� 3.39 vessels per high power field(hpf), salineZ 20.0� 0.91 vessels per high power field (hpf); pZ 0.029). Sirius red stainingdemonstrated significantly reduced scar index in fat-grafted animals at 8 weeks(fatZ 0.69� 0.10, salineZ 2.03� 0.53; pZ 0.046).Conclusions: Fat grafting resulted in more rapid revascularisation at the burn site as measuredby laser Doppler flow, CD31 staining and chemical markers of angiogenesis. In turn, this re-sulted in decreased fibrosis as measured by Sirius red staining and chemical markers.ª 2011 British Association of Plastic, Reconstructive and Aesthetic Surgeons. Published byElsevier Ltd. All rights reserved.

According to the American Burn Association, there areapproximately half a million burns requiring medicalattention in the United States each year.1 Early burnprogression and inadequate systems of measurementcomplicate assessment of burn depth.2 Many deep partial-thickness burns are therefore treated without surgicalintervention. Left to heal without excision and skin graft-ing, deep burns often result in fibrotic scars characterisedby abnormal colour, texture, thickness and pliability.3

When untreated burns are located in cosmetically sensi-tive areas, significant deformity can result. A number oftopical and minimally invasive techniques have been usedin an attempt to improve the quality and appearance ofburn scars. These treatment options include topical siliconegels, pressure dressings, corticosteroids and, most recently,autologous fat grafting.4e7

Over the last decade, the tissue engineering communityhas come to recognise the regenerative potential ofadipose tissue.8 In turn, plastic surgeons have translatedthis revelation into clinical applications. Autologous fatgrafts have been used in a number of settings to improvethe quality of overlying tissues.9e11 In one study, clinicalimprovement of hypertrophic burn scars was notedfollowing autologous fat grafting beneath the burn site.7 Todate, no study has examined the effect of fat grafting onburns in a controlled experiment. Employing our model ofmurine fat grafting in conjunction with a previouslydescribed murine model of thermal injury, we aimed todetermine the mechanism by which fat grafting may effectchange within a burn scar.12,13 To this end, markers of neo-vascularisation and fibrosis were compared in fat- andsaline-treated mice at two time points following thermalinjury.

Figure 1 Study timeline.

Materials and methods

Mice

This study was approved by the NYU Medical Center Insti-tutional Animal Care and Use Committee (IACUC #070911-02). The study was designed as follows (Figure 1): Ten-week-old wild type FVB mice (nZ 20) (Jackson Laborato-ries, Bar Harbor, ME, USA) were anesthetised using a murineanaesthetic cocktail consisting of ketamine, xylazine andacepromazine (0.1 ml/10 g intra-peritoneal (IP)). Theirdorsal skin was shaved using an electric clipper followed by

a depilatory agent to remove all remaining hair. Theanimals were then divided evenly into two groups: the firstgroup consisted of animals that underwent subcutaneousfat grafting (nZ 10) 2 weeks following thermal injury,while the second group consisted of animals that under-went subcutaneous saline grafting (nZ 10) 2 weeksfollowing thermal injury.

Laser Doppler scanning

All animals were subjected to laser Doppler scanning atthree separate points in the study: immediately prior tothermal injury, 2 weeks following thermal injury (immedi-ately prior to fat or saline grafting) and immediately priorto wound bed harvest (at 4 or 8 weeks following grafting).Once anaesthetised and shaved, the animals were orientedvertically within the field of the laser Doppler scanner(MoorLDI2-IR; Moor Instruments, Devon, UK) and scanned. A0.25-cm2 area at the centre of the burn wound was isolatedand the flux (mV), a measure of blood flow, was calculatedwithin this region of interest (MoorLDI MeasurementSoftware).

Thermal injury

Anaesthetised and shaved animals were subjected toa single burn wound on their dorsum. A 10-mm brass hexcap screw (Grainger Industrial Supply, Lake Forest, IL, USA)was heated to 100 �C in a hot water bath before beingapplied to the dorsal skin of the anaesthetised animals for10 s. This application was found in preliminary studies toproduce a consistent full-thickness burn. The animals wereallowed to recover under a heat lamp and were keptthereafter in a supervised animal facility with murine foodpellets and water ad libitum.

Fat grafting in burn injury 221

Fat grafting

Fat or sham grafting was carried out 2 weeks followingthermal injury using our previously described murine modelof fat grafting.12 This time point was chosen because a firmeschar formed at the burn site within 2 weeks followinginjury. The mice were anaesthetised before the posteriorpart of their dorsal skin was shaved and prepared in anaseptic fashion. A small incision was made in the posteriormidline of the dorsum. Discarded human fat (IRB #H12756-01b) was processed using the Coleman technique.14 Thehighest density fat was partitioned into 1-cc syringes usinga luer lock to luer lock transfer system (Part No. LL-LL12;Mentor, Santa Barbara, CA, USA). The fat was then infil-trated beneath the dorsal skin of the mouse using a Style-1Coleman Infiltration Cannula (Part No. COL-17; Mentor,Santa Barbara, CA, USA). Each animal was diffusely infil-trated with 1.5 cc of fat. The incision was then closed using4/0 nylon sutures. Sham-grafted animals were treated in anidentical fashion, but 1.5 cc of sterile saline were injectedin place of fat. The animals were allowed to recover undera heat lamp before being returned to the supervised animalfacility.

Serial photography

Mice were photographed serially to evaluate progression ofthermal injury or other apparent changes following fat/sham grafting.

Tissue harvest

Dorsal skin was harvested for analysis at 4 weeks and 8weeks following grafting (nZ 5 from each treatment groupat each time point). The mice were sacrificed by CO2

narcosis and cervical dislocation. A central 1.5-cm2 portiondorsal skin was then excised from the burn scar site.Remaining portions of the fat graft or other subcutaneoustissue adherent to the skin were removed. Each sample wasthen divided evenly into three pieces designated for poly-merase chain reaction (PCR) arrays, enzyme-linked immu-nosorbent assay (ELISA) assays and histologic analysis,respectively. Following homogenisation of tissue samples(Polytron tissue homogeniser; Kinematica, Bohemia, NY,USA) nucleic acid and protein extraction were performed(AllPrep DNA/RNA/Protein Mini Kitl; Qiagen, Valencia, CA,USA). Following extractions, RNA and DNA were quantifiedwith a Nanodrop-1000 Spectrophotometer (Thermo FisherScientific Inc., Waltham, MA, USA). Protein was quantifiedusing a Pierce 660 nm Protein Assay (Thermo Fisher Scien-tific, Rockford, IL, USA) and spectrophotometer. Skinsamples intended for histology were fixed in formalin for24 h before being transferred to 70% ethanol. Thesesamples were then embedded in paraffin, sectioned andmounted.

ELISA protein quantification

ELISAs (R&D Systems, Minneapolis, MN, USA) for vascularendothelial growth factor (VEGF), stromal cell-derivedfactor-1 (SDF-1), transforming growth factor-b1 (TGF-b1)

and matrix metallopeptidase 9 (MMP9) were carried out onsamples from both groups at both time points (nZ 5 in eachgroup at each time point). Briefly, samples and controlswere loaded into antibody-coated microplates along with50 ml of assay diluent. The plate was then incubated for 2 hbefore 100 ml of conjugate was added to each well. Theplates were incubated for a further 2 h before 100 ml ofsubstrate solution was added. The plate was then incu-bated in the dark for 30 min after which 100 ml of stopsolution was added. The absorbance was then read usinga SpectraMax 340 (Molecular Devices, Sunnyvale, CA, USA).Finally, the measured absorbance was converted to proteinconcentration using SoftMax Pro Data Analysis Software(Molecular Devices, Sunnyvale, CA, USA).

PCR arrays

Reverse transcription was performed using the QuantiTectReverse Transcription kit (Qiagen). Expression of mouseBax, Bcl-2, HIF-1, VEGF, SDF-1, Col1a1, MMP9, Smad3, TGF-b1 and tissue inhibitor of metalloproteinases (TIMP-1)messenger RNAs (mRNAs) was measured in samples fromboth groups at both time points (nZ 5 in each group at eachtime point) by the real-time quantitative RT-PCR methodusing SYBR green master mix (Qiagen) and the ABI Prism7900HT Sequence Detection System (Applied Biosystems,Foster City, CA, USA) (Table 1 e primer sequences). Rela-tive mRNA levels were determined by the comparativethreshold cycle (2�DDCT) method. The expression of allmRNAs was normalised to that of 18s mRNA.

CD31 staining

CD31 staining was carried out as a measure of vascularityusing a previously described protocol.15 Five previouslyembedded and sectioned slides from each group at eachtime point were deparaffinised before being incubated in3% H2O2 for 10 min. They were then washed in wash bufferfor 5 min. The slides were then blocked with 150 ml ofblocking solution for 1 h at room temperature before 150 mlof CD-31 antibody (Abcam, Cambridge, MA, USA) was addedin a 1:50 dilution and allowed to incubate on the slidesovernight at 4 �C. The primary antibody was then removedand the slides were washed in wash buffer before 150 ml ofsecondary antibody (Vector Labs, Burlingame, CA, USA) wasadded to each of the sections. They were then incubatedfor 30 min at room temperature. Following a 30-min incu-bation in 150 ml of ABC Reagent at room temperature, thesections were washed in wash buffer before 150 ml of DABreagent was added to each. The DAB reagent was allowedto incubate until there was a visible colour change(approximately 2 min). The slides were then dehydrated inthree changes of 100% ethanol before being mounted.

Following staining, slides were examined under themicroscope at high power using red light excitation todemonstrate 40,6-diamino-2-phenylindole (DAPI) staining ofcell nuclei. Slides were then viewed under green lightexcitation to demonstrate CD31-positive cells. DAPI andCD31 images were acquired of multiple unique fields beforecomposite images were created by superimposing them.

Table 1 PCR primer sequences.

Protein Forward Reverse

VEGF GTAACGATGAAGCCCTGGAGTG TGAGAGGTCTGGTTCCCGAAACSDF-1 GTCTAAGCAGCGATGGGTTC GAATAAGAAAGCACACGCTGCHIF-1 GATGAGTTCTGAACGTCGAAAAGAAAAGT GAAGTTTTCTCACACACAAATAACTGATGGTCBcl-2 CATCTTCTCCTTCCAGCCT ATCTCCCTGTTGACGCTCTBax ATGCGTCCACCAAGAAGCTGAG CCCCAGTTGAAGTTGCCATCAGTGF-ß TATTTGGAGCCTGGACACAC CTTGCGACCCACGTAGTAGAMMP9 GGGAAGGCTCTGCTGTTCAGC TCTAGAGACTTGCACTGCACGCol1A1 TGTCCCAACCCCCAAAGAC CCCTCGACTCCTACATCTTCSmad3 GCAGCAAATTCCTGGTTGTT TTTCGTCCAGTCTCCCAACTTIMP-1 CGCAGATATCCGGTACGCCTA CACAAGCCTGGATTCCGTGG18s GAGAAACGGCTACCACATCC GGACACTCAGCTAAGAGCATCG

222 S.M. Sultan et al.

Finally, the number of CD31 positive vessels was quantifiedin each composite high-powered field.

Picrosirius red staining

Picrosirius red staining was carried out as a measure ofcollagen organisation. Five previously embedded andsectioned blank slides from each group at each time pointwere deparaffinised and hydrated. They were then stainedin Weigert’s haematoxylin for 8 min before being rinsed intap water for 10 min. This was followed by staining in pic-rosirius red (Sigma Aldrich, St. Louis, MO, USA) for 1 hbefore being washed in two changes of acidified water. Theslides were then dehydrated in three changes of 100%ethanol before being mounted in a resinous medium.16

Following staining, slides were viewed under themicroscope using a polarising filter (Olympus USA, CenterValley, PA, USA). Representative images were chosen atrandom from each slide and captured. Colours within theyellow-green spectrum and the red-orange spectrum wereselected and quantified (Nikon NIS-Elements BR software;Nikon Instruments Inc., Melville, NY, USA). When viewedunder polarised light, organised collagen stained with Siriusred appears yellow-green. Fibrotic, disorganised collagenappears red-orange. As previously described, the ratio ofthe area that falls within the red-orange spectrum to thearea within the yellow-green spectrum is the Scar Index, aneffective measure of collagen disorganisation.17 A ScarIndex was obtained from each of the images by the abovemethod.

Statistics

Results were compared using paired t-tests to determinesignificance between groups. P-values less than 0.05 wereconsidered significant.

Results

Serial photography

Serial photography (Figure 2) demonstrated improvementsin colouration and texture in the burn scars of fat-graftedanimals when compared to saline-grafted controls.

Laser Doppler scanning

Laser Doppler scanning (Figure 3) 2 weeks followingthermal injury demonstrated significantly decreased fluxthrough the zone of injury when compared with uninjuredcontrol mouse skin (control fluxZ 392� 7.25 mV, burnfluxZ 189� 10.26 mV; p< 0.001). Four weeks followingfat and saline grafting (6 weeks following thermalinjury), fat-grafted animals demonstrated significantlygreater flux through the burn wound than saline-treatedanimals (fat graftedZ 305� 15.77 mV, saline graftedZ242� 15.83 mV; pZ 0.026). Eight weeks following fat andsaline grafting (10 weeks following thermal injury), fat-grafted animals continued to demonstrate greater flux inthe burn wound bed when compared with the saline-grafted animals, though this difference was no longerstatistically significant (fat graftedZ 327� 6.87 mV, salinegraftedZ 303� 7.19 mV; pZ 0.054).

ELISA protein quantification

Concentrations of both VEGF and SDF-1 were significantlyelevated in fat-grafted animals 4 weeks followinggrafting when compared with saline-grafted animalsat the same time point (VEGF: fat graftedZ 74.3� 4.39ng mle1, saline graftedZ 34.3� 5.23 ng mle1; pZ 0.004)(SDF-1: fat graftedZ 51.8� 1.23 ng mle1, saline graftedZ 10.2� 3.22 ng mle1; p< 0.001) (Figure 4(A)). Eightweeks following grafting both VEGF and SDF-1remained significantly elevated in fat-grafted animalswhen compared with saline-grafted animals (VEGF: fatgraftedZ 12.4� 1.22 ng mle1, saline graftedZ 1.2� 0.29ng mle1; p< 0.001) (SDF-1: fat graftedZ 2.3� 0.31ng mle1, saline graftedZ 1.4� 0.19 ng mle1; pZ 0.035).

There was no significant difference in the concentrationof TGF-b when comparing fat- and saline-grafted animals4 weeks following grafting (fat graftedZ 1.29� 0.95ngmle1, saline graftedZ 2.31� 0.06 ngmle1; pZ 0.34).The same was true when comparing the concentrationof MMP9 in fat- and saline-grafted animals 4 weeksfollowing grafting (fat graftedZ 8.03� 2.31 ngmle1, salinegraftedZ 6.61� 2.03 ngmle1; pZ 0.66). Eight weeksfollowing grafting, the concentration of both TGF-ß andMMP9 were significantly increased in saline-grafted animals

Figure 3 Laser Doppler scanning: blood flow was assessed atthe burn site using a laser doppler scanner. A flow void waspresent prior to fat and saline grafting (note: dashed lineindicates average flux measured immediately prior to grafting,188� 10.26 mV). This void gradually resolved in both groupsover the course of the experiment. Flow at the burn site wassignificantly increased in fat grafted animals at four weekswhen compared to saline grafted controls (fat graf-tedZ 305� 15.77 mV, saline graftedZ 242� 15.83;pZ 0.026).

Figure 2 Serial photography: serial photography demonstrates improvements in colouration and texture in the burn scars of fatgrafted animals when compared to saline grafted controls.

Fat grafting in burn injury 223

when compared with fat-grafted animals (TGF-b: salinegraftedZ 9.30� 0.93 ngmle1, fat graftedZ 4.63� 0.38;pZ 0.002) (MMP9: saline graftedZ 13.05� 1.21 ngmle1,fat graftedZ 6.83� 1.39 ngmle1; pZ 0.010) (Figure 4(B)).

PCR arrays

PCR for vasculogenic markers demonstrated up-regulationin fat-grafted animals 4 weeks following grafting. This up-regulation was statistically significant in both VEGF andSDF-1 (VEGF saline:fatZ 1:3.78, pZ 0.02; SDF-1 saline:-fatZ 1:2.34, pZ 0.04; hypoxia-inducible factor-1 (HIF-1)saline:fatZ 1:1.89, pZ 0.06). PCR for the pro-apoptoticfactor BAX demonstrated decreased expression in fat-grafted animals 4 weeks following grafting (BAX saline:-fatZ 1:0.72, pZ 0.069). Moreover, PCR for Bcl-2, an anti-apoptotic factor, demonstrated significant up-regulation infat-grafted animals at the same time point (Bcl-2 saline:-fatZ 1:5.99, pZ 0.007) (Figure 5(A)). There was nosignificant difference in these vascular and apoptoticmarkers 8 weeks following fat and saline grafting (data notshown).

PCR for fibrotic factors Col1a1, MMP9, Smad3, TGF-b andTIMP-1 demonstrated no significant differences in mRNAexpression between fat- and saline-grafted animals 4 weeksfollowing grafting (data not shown). Eight weeks followinggrafting there was, with the exception of Smad3, significantdown-regulation detected in all cases (Col1a1 saline:-fatZ 1:0.38, p< 0.01; MMP9 saline:fatZ 1:0.58, pZ 0.03;

Figure 4 ELISA protein quantification: (A) Vasculogenicmarkers VEGF and SDF were both significantly upregulatedin fat grafted animals four weeks following graftingwhen compared to saline grafted controls (VEGF:fatZ 74.3� 4.39 ng/ml, salineZ 34.3� 5.23; pZ 0.004)(SDF-1: fatZ 51.8� 1.23 ng/ml, salineZ 10.2� 3.22; p< 0.001).(B) Fibrotic markers MMP9 and TGF-ß were both significantlydownregulated in fat grafted animals eight weeks followinggrafting when compare to saline grafted controls (MMP9: sali-neZ 13.05� 1.21 ng/ml, fatZ 6.83� 1.39; pZ 0.010)(TGF-ß:salineZ 9.30� 0.93 ng/ml, fatZ 4.63� 0.38; pZ 0.002).

Figure 5 PCR arrays: (A) PCR for vasculogenic markersdemonstrated upregulation in fat grafted animals four weeksfollowing grafting (VEGF saline:fatZ 1:3.78, pZ 0.02; SDF-1saline:fatZ 1:2.34, pZ 0.04; HIF-1 saline:fatZ 1:1.89,pZ 0.06). PCR for proapoptotic factor BAX demonstrateddecreased expression in fat grafted animals four weeksfollowing grafting (BAX saline:fatZ 1:0.72, pZ 0.069). PCR forBcl-2, an anti-apoptotic factor, demonstrated significantupregulation in fat grafted animals at the same time point (Bcl-2 saline:fatZ 1:5.99, pZ 0.007). (B) PCR for fibrotic factorsCol1a1, MMP9, Smad3, TGF-ß, and TIMP-1 eight weeksfollowing grafting demonstrated downregulation in fat graftedanimals when compared to saline grafted controls (Col1a1saline:fatZ 1:0.38, p< 0.01; MMP9 saline:fatZ 1:0.58.pZ 0.03; Smad3 saline:fatZ 1:0.81, pZ 0.27; TGF-ß saline:-fatZ 1:0.23, p< 0.01; TIMP-1 saline:fatZ 1:0.12, p< 0.01).

224 S.M. Sultan et al.

Smad3 saline:fatZ 1:0.81, pZ 0.27; TGF-b saline:-fatZ 1:0.23, p< 0.01; TIMP-1 saline:fatZ 1:0.12,p< 0.01) (Figure 5(B)).

Histology

CD31 staining at 4 weeks following grafting demonstratedsignificantly greater vascular density surrounding the burnwound in fat-grafted animals when compared with saline-treated animals (fat graftedZ 30.8� 3.39 vessels per hpf,saline graftedZ 20.0� 0.91 vessels per hpf; pZ 0.029)(Figure 6(A)). Eight weeks following grafting, vasculardensity in the tissue surrounding the burn wound wasagain found to be significantly greater in fat-graftedanimals when compared with saline-grafted animals(fat graftedZ 18.4� 1.8 vessels per hpf, salinegraftedZ 12.3� 1.4 vessels per hpf; pZ 0.038).

Picrosirius red staining at 4 weeks following grafting didnot demonstrate a significant difference in fibrosis betweenfat-grafted animals and saline-grafted animals (a scar index

was not obtained at this time point as collagen bundleswere not sufficiently formed within the scar to generatea scar index by the above-described protocol). Eight weeksfollowing grafting, however, burn scars harvested from fat-grafted animals demonstrated a significantly lower scarindex than burn scars harvested from saline-grafted animals(fat graftedZ 0.69� 0.10, saline graftedZ 2.03� 0.53;pZ 0.046) (Figure 6(B)).

Discussion

Applying our murine model of fat grafting in the settingof thermal injury resulted in accelerated revascularisa-tion of burn scar tissue and, ultimately, decreasedfibrosis. Serial photography demonstrated improvement inthe texture and extent of burn scars grafted with fat in

Figure 6 Histology: (A) CD31 staining four weeks following grafting demonstrated a significantly increased number of vessels perhigh powered field at the periphery of the burn site in fat grafted animals when compared to saline grafted controls(fatZ 30.8� 3.39 vessels/hpf, salineZ 20.0� 0.91; pZ 0.029). (Note: white arrows indicate examples of representative CD31þvessels). (B) Sirius red staining eight weeks following grafting demonstrated more organised collagen at the burn site fat graftedanimals when compared to saline grafted controls (note: yellow/green, organized collagen; red/orange, disorganized collagen).

Fat grafting in burn injury 225

place of saline (Figure 2). Revascularisation of the burnscar, as measured using laser Doppler scanning (Figure 3)and CD31 staining (Figure 6(A)), was significantly moreadvanced in fat-grafted animals 4 weeks followinggrafting. This finding was paralleled by pro-vasculogenicchanges in gene expression found on ELISA (Figure 4(A))and pro-vasculogenic/anti-apoptotic trends in PCR arrays(Figure 5(A)) (e.g., a 3.78-fold increase in expression ofVEGF mRNA and a 5.99-fold increase in expression of Bcl-2mRNA).

Eight weeks following grafting, saline-grafted animalswere revascularised nearly to the same extent as the fat-grafted animals, as demonstrated by laser Doppler scanning(Figure 3). Early revascularisation in fat-grafted animalsproved important, however, in the prevention of fibrosisat later time points. Eight weeks following grafting, fat-grafted animals demonstrated significantly less fibrosisthan saline-grafted controls (Figure 6(B)). Significantlydecreased markers of fibrosis were found on both ELISA(Figure 4(B)) and PCR arrays (Figure 5(B)) as well, corrob-orating this histologic finding (e.g., a 77% reduction inexpression of TGF-ß mRNA).

To understand the clinical significance of early revascu-larisation in a burnwound, it is essential to first consider burnpathophysiology. Following a burn, thrombosis in themicrovasculature causes ischaemia in an area known as the

zone of stasis.18 In response, endothelial progenitor cells(EPCs) are mobilised from the bone marrow.19 EPCs areknown to home to areas of ischaemia and differentiate intoendothelial cells that are instrumental in blood vesselformation.20 The concentration of EPCs in circulation peaksaround 24 h following a burn and drops off significantlyafterward, however, returning to basal levels within 72 h.19 Ithas been shown as well that burns of increasing severityresult in delayed and diminished release of EPCs into circu-lation.13 As a result, many burn wounds revascularise slowlyand, in turn, mature hypertrophic burn scars demonstratefunctional changes in their microcirculation.13,21,22

Microcirculatory anomalies play an important role in scararchitecture as low oxygen tension induces an up-regulation of TGF-ß1. This factor is an important regu-lator of collagen production and, when unchecked,fibrosis.23,24 Moreover, hypoxia alters the way in which TGF-ß1 interacts with the MMPs, a family of proteins thatregulate the remodelling of collagen.25,26 It has thereforebeen suggested that any intervention that improves thevascularity and, by extension, oxygenation of a burn scarwould similarly improve the scar’s quality and texture.13

Early revascularisation of a burn scar may thereforeprotect the wound bed from up-regulation of TGF-b1 andresultant derangements in collagen synthesis at the earlystage of scar formation.

226 S.M. Sultan et al.

This study has several limitations that should be noted:first, further experiments will be required to pinpoint theexact mechanism by which autologous fat grafting accel-erates revascularisation in burn wounds. The potential foradipose-derived stem cells to act as endothelial progenitorcells and promote neo-vascularisation has been describedpreviously, however.27 Second, a larger animal model (e.g.,Red Duroc porcine skin model or rabbit ear model) will berequired to more extensively explore the phenotypic effectof fat grafting in thermal injury. Finally, it should be notedas well that the use of wild type FVB mice (in place ofimmunocompromised animals) does not lead to an immunereaction against the human fat grafts. This phenomenon isnot well understood and further studies are being designedto address it.

Despite the limitations of this study, however, themurine model chosen was sufficient to demonstrate thepositive effect of fat grafting on markers of vascularity andfibrosis following full-thickness thermal injury. This studytherefore speaks to the potential of fat grafting as an earlyand minimally invasive intervention for severe burns left toheal by secondary intention or, potentially, for partialthickness burns that may not warrant excision and skingrafting, but do appear to hold the potential for hypertro-phic scarring. In these cases, early revascularisation mayyield significant improvement in the quality of the eventualburn scar. Fat grafting may prove valuable as an adjunct toskin grafting in cases of severe burns as well.

Conclusions

Infiltration of processed human lipoaspirate beneathsubacute burns in mice resulted in significant changes in thehealing wound, most notably accelerated revascularisation.In turn, this led to down-regulation of fibrotic pathwayswith resultant improvements in scar quality. Although thesefindings must be re-examined in larger animal models, fatgrafting may in fact be applicable in the treatment ofthermal injuries in the future.

Funding

This project received funding from the National Endow-ment for Plastic Surgery. Sydney Coleman receives royaltiesand is a paid consultant for Mentor. He is a paid consultantfor the Armed Forces Institute of Regenerative Medicine.No financial support or benefits have been received by anyother co-author, by any member of our immediate family orany individual or entity with whom or with which we havea significant relationship from any commercial source whichis related or indirectly related to the scientific work re-ported on in this article.

References

1. American Burn Association, Burn incidence factsheet. Avail-able from: http://www.ameriburn.org/resources_factsheet.php [accessed 9/16/2010].

2. Grunwald TB, Garner WL. Acute burns. Plast Reconstr Surg2008;121:311ee9e.

3. Sullivan T, Smith J, Kermode J, Courtemanche DJ. Rating theburn scar. J Burn Care Rehabil 1990;11:250e60.

4. Shih R, Waltzman J, Evans GR. Review of over-the-countertopical scar treatment products. Plast Reconstr Surg 2007;119:1091e5.

5. Acikel C, Ulkur E, Guler MM. Treatment of burn scar depig-mentation by carbon dioxide laser-assisted dermabrasion andthin skin grafting. Plast Reconstr Surg 2000;105:1973e8.

6. Griffith BH, Monroe CW, McKinney P. A follow-up study on thetreatment of keloids with triamcinolone acetonide. PlastReconstr Surg 1970;46:145e50.

7. Klinger M, Marazzi M, Vigo D, Torre M. Fat injection for cases ofsevere burn outcomes: a new perspective of scar remodelingand reduction. Aesth Plast Surg 2008;32:465e9.

8. Zuk PA, Zhu M, Ashjian P, et al. Human adipose tissue isa source of multipotent stem cells. Mol Bio Cell. 2002;13:4279e95.

9. De Ugarte DA, Ashjian PH, Elbarbary A, Hedrick MH. Future offat as raw material for tissue regeneration. Ann Plast Surg2003;50:215e9.

10. Rigotti G, Marchi A, Galie M, et al. Clinical treatment ofradiotherapy tissue damage by lipoaspirate transplant: a heal-ing process mediated by adipose-derived adult stem cells.Plast Reconstr Surg 2007;119:1409e22.

11. Serra-Renom JM, Munoz-Olmo JL, Serra-Mestre JM. Fat graftingin postmastectomy breast reconstruction with expanders andprostheses in patients who have received radiotherapy:formation of new subcutaneous tissue. Plast Reconstr Surg2010;125:12e8.

12. Thanik VD, Chang CC, Lerman OZ, et al. A murine model forstudying diffusely injected human fat. Plast Reconstr Surg2009;124:74e81.

13. Zhang X, Wei X, Liu L, et al. Association of increasing burnseverity in mice with delayed mobilization of circulatingangiogenic cells. Arch Surg 2010;145:259e66.

14. Coleman SR. Facial recontouring with lipostructure. Clin PlastSurg 1997;24:347e67.

15. Huss WJ, Hanrahan CF, Barrios RJ, Simons JW, Greenberg NM.Angiogenesis and prostate cancer: identification of a molecularprogression switch. Cancer Res 2001;61:2736e43.

16. Junqueira LCU, Bignolas G, Brentani RR. Picrosirius stainingplus polarization microscopy, a specific method for collagendetection in tissue sections. Histochem J 1979;11:447e55.

17. Flanders KC, Major CD, Arabshahi A, et al. Interference withtransforming growth factor-beta/Smad3 signaling results inaccelerated healing of wounds in previously irradiated skin.Am J Pathol 2003;163:2246e57.

18. Hettiaratchy S, Dziewulski P. ABC of burns: pathopysiology andtypes of burns. BMJ 2004;328:1427e9.

19. Fox A, Smythe J, Fisher N, et al. Mobilization of endothelialprogenitor cells into the circulation in burned patients. Br JSurg 2008;95:244e51.

20. Crosby JR, Kaminski WE, Schatteman G, et al. Endothelialcells of hematopoietic origin make a significant contribu-tion to adult blood vessel formation. Circ Res. 2000;87:728e30.

21. Sloan DF, Brown RD, Wells CH, Hilton JG. Tissue gases inhypertrophic burn scars. Plast Reconstr Surg 1978;61:431e6.

22. Clark JA, Leung KS, Cheng JC, Leung PC. The hypertrophic scarand microcirculation properties. Burns 1996;22:447e50.

23. Falanga V, Zhou L, Yufit T. Low oxygen tension stimulatescollagen synthesis and COL1A1 transcription through the actionof TGF-ß1. J Cell Physiol 2002;191:42e50.

24. Border WA, Noble NA. Transforming growth factor ß in tissuefibrosis. N Engl J Med 1994;10:1286e92.

25. Saed GM, Zhang W, Diamond MP. Effect of hypoxia onstimulatory effect of TGF-ß1 on MMP-2 and MMP-9

Fat grafting in burn injury 227

activities in mouse fibroblasts. J Soc Gynecol Investig 2000;7:348e54.

26. Rananti L, Kahari VM. Matrix metalloproteinases in woundrepair. Int J Mol Med 2000;6:391e407.

27. Miranville A, Heeschen C, Sengenes C, Curat CA, Busse R,Bouloumie A. Improvement of postnatal neovascularization byhuman adipose tissue-derived stem cells. Circulation 2004;110:349e55.