Comparative analysis of global gene expression profiles between diabetic rat wounds treated with vacuum‐assisted closure therapy, moist wound healing or gauze under suction

Abstract

How differential gene expression affects wound healing is not well understood. In this study, Zucker diabetic fatty (fa/fa) male inbred rats were used to investigate gene expression during wound healing in an impaired wound‐healing model. Whole genome microarray surveys were used to gain insight into the biological pathways and healing processes in acute excisional wounds treated with vacuum‐assisted closure (V.A.C.^®) Therapy, moist wound healing (MWH) or gauze under suction (GUS). Global gene expression analyses after 2 days of healing indicated major differences with respect to both number of genes showing fold changes and pathway regulation between the three different wound treatments. Statistical analysis of expression profiles indicated that 5072 genes showed a >1·6‐fold change with V.A.C. Therapy compared with 3601 genes with MWH and 3952 genes with GUS. Pathways and related genes associated with the early phases of wound healing diverged between treatment groups. For example, pathways involving angiogenesis, cytoskeletal regulation and inflammation were associated with elevated gene expression following V.A.C. Therapy. This study is the first to assess wound healing by whole genome interrogation in a diabetic rat model treated with different healing modalities.

Introduction

Diabetes affects 246 million people worldwide and is expected to impact 380 million by 2025 (1). In the USA alone, more than 20·8 million people are known to have diabetes (2), and it is one of the most common causes of non traumatic lower extremity amputations because of non healing wound complications in Europe and the USA 3, 4, 5. In 2002, about 82 000 non traumatic, lower limb amputations were performed on US diabetic patients secondary to wound complications (6). Foot ulceration is generally associated with peripheral neuropathy and develops in about 15% of patients with diabetes 4, 7, 8. Because diabetic wounds are complex, proper wound bed preparation and early intervention are essential to successful healing (9). Research indicates that 67% of diabetic foot ulcers are unhealed after 20 weeks of care using standard wound‐healing methods (10). Vacuum‐assisted closure (V.A.C.^® KCI, San Antonio, TX, USA) Therapy is widely used in the treatment of diabetic foot ulcers and has been clinically proven to prepare the wound bed for closure in both acute and chronic wounds 7, 8, 11, 12, 13, 14. Gene signature changes during V.A.C. Therapy are unknown, and a wound‐healing gene signature has yet to be defined.

Wound healing may be classified into four overlapping phases; haemostasis, inflammation, proliferation and remodelling (15). Haemostasis starts at the moment the tissue is injured and when blood moves into the site of injury. The inflammation phase follows haemostasis whereby neutrophils and macrophages appear, initiating the phagocytotic processes with increased secretion of growth factors and inflammatory cytokines, including tumour necrosis factor alpha (TNF‐α) and interleukin (IL)‐6 15, 16, 17. Neutrophils also may help activate fibroblasts and epithelial cells 16, 18. During the proliferation phase, fibroblasts migrate into the wound. They deposit new extracellular matrix, stimulate protease inhibitor activity, promote angiogenesis and release cytokines such as interleukins, fibroblast growth factor and TNF‐α. In the remodelling phase, the wound becomes re‐epithelised, the extracellular matrix becomes cross‐linked and the healed wound becomes less vascular.

Microarray studies during the various phases of wound healing may provide insight into gene expression patterns, leading to the various events of the wound‐healing cascade. Of interest are the early steps in wound healing such as the inflammatory phase, given that quick resolution will result in healing, whereas a delay or inability to resolve the inflammatory response could lead to chronic ulceration as seen in diabetic patients (16). Chronic wounds with prolonged or non resolving inflammation phases are associated with increased levels of proteases such as matrix metallopeptidases, which destroy components of the extracellular matrix and damage growth factors and receptors needed for healing (16). Assessing trends in resolution of the inflammatory phase by way of upregulation or downregulation of particular genes could lead to the creation of improved treatment modalities.

This study used whole genome survey microarrays and TaqMan^® quantitative real‐time polymerase chain reaction (PCR) technologies to track changes in gene expression profiles during the early phase of healing in Zucker diabetic fatty (ZDF) rats. We compared gene expression over the first 2 days of healing using two different negative pressure wound therapy modalities and moist wound healing (MWH) as a control. The 48‐hour time point was chosen to examine gene expression as representative of the first dressing change during V.A.C. Therapy. This is the first study to interrogate the entire genome by assessing differential gene expression during the early phase of wound healing using three different wound treatment modalities: V.A.C. Therapy, MWH or gauze under suction (GUS).

Methods

Animals

Male ZDF rats were obtained from Charles River Laboratories (Wilmington, MA) at 17–22 weeks of age. Male rats in this inbred strain develop non insulin‐dependent type 2 diabetes and are characterised by impaired wound healing (19). Blood glucose levels were monitored using a glucometer (LifeScan, Inc/Johnson & Johnson, Milpitas, CA) to verify that the animals were diabetic (glucose levels 339 mg/dl–600 mg/dl). Animals were housed individually, maintained at 22–24°C with a 12‐h light/dark cycle and allowed food and water ad libitum. All animal experiments were approved by the University of Texas Health Science Center Institutional Animal Use and Care Committee. Six animals were treated with V.A.C. Therapy (V.A.C.), six with Tegaderm™ (MWH) and four with GUS.

Wound creation and area

The day of wound creation was designated day 0. The hair on the dorsolateral back was removed, and one circular 3‐cm diameter full‐thickness wound was created on the mid‐dorsum of each animal. Wound area was calculated using the Visitrak grid system (Smith & Nephew, Inc, Largo, FL) at wound creation and end of treatment on day 2.

Treatment of wounds

Wounds treated with V.A.C. Therapy were fitted with a polyurethane dressing (V.A.C. GranuFoam^® Dressing; KCI, San Antonio, TX) and covered with a polyurethane drape (V.A.C.^® Drape; KCI) and a connector tubing (T.R.A.C.^® Pad; KCI). The proximal end of the tubing was connected to a swivel mounted on the top of the cage to allow the animals’ free mobility and access to food and water. The swivel was in turn connected to a V.A.C. Therapy ATS^® System (KCI), set to deliver continuous subatmospheric pressure of −125 mmHg, the manufacturer’s recommended clinical pressure setting.

MWH‐treated wounds received Tegaderm Transparent Film Dressing (3M Health Care, St Paul, MN). This dressing was used as a baseline, untreated control. The same style of connector tubing was placed on top of the MWH dressing so that all animals received a comparable tethering device. This was performed to control for stress to the back of the animals because of the added weight of the tethering device.

For GUS, 3 × 3 inch Aquaphor™ Gauze (Smith & Nephew, Inc) was cut to 3 cm diameter and placed directly in the wound bed. The tip of a Jackson‐Pratt™ (Cardinal Health, Dublin, OH) drain was trimmed, and the end was placed in the wound over the Aquaphor. A piece of double‐sided adhesive hydrogel was applied to the periwound tissue to anchor the drain and to keep the dressing assembly from repositioning. The drain was covered with approximately four layers of sterile gauze and then covered with an Opsite™ (Smith & Nephew, Inc) dressing to create a seal. The dressing assembly was connected to a swivel as previously mentioned, and the swivel was connected to a Versatile 1™ pump (BlueSky, Carlsbad, CA). The pump was set at continuous suction of −75 mmHg, the manufacturer’s recommended clinical pressure setting used for this therapy, and a handheld manometer (Omega, Stamford, CT) was used to monitor Versatile 1 pressure.

To control pain, all animals received approximately 0·04 mg/kg buprenorphine twice daily as determined by the attending veterinarian during the first 24–48 hours post wounding. After all experiments were completed, rats were euthanatized by exsanguination.

Tissue samples

On day of wound creation, a portion of the non wounded skin was quickly placed into RNAlater^® (Ambion, Austin, TX) and stored at −20°C for subsequent preparation of total RNA. At the conclusion of the experiment on day 2, wounded tissue was quickly removed and stored in RNAlater at −20°C. Total RNA was isolated from tissue using TRIzol^® (Invitrogen, Carlsbad, CA) with modifications to remove DNA using RNeasy columns and a DNase I Kit (Qiagen, Valencia, CA). RNA was stored at −80°C in nuclease‐free H₂O (Qiagen). Quality and quantity of RNA were determined using an Experion Automated Electrophoresis System (Bio‐Rad, Hercules, CA), and the same samples were divided into aliquots for microarray and TaqMan quantitative real‐time PCR analysis (Applied Biosystems, Foster City, CA).

Microarray

In this study, data from a total of 32 microarrays were compared. The V.A.C. and MWH groups had six biological replicates, while the GUS group had four biological replicates. Gene expression profiles were generated using the Applied Biosystems Rat Genome Survey Microarray. Each microarray contains approximately 28 000 features that include a set of about 1000 controls. Each microarray uses 60‐mer oligonucleotide probes (26 857) designed against 27 088 genes covering 43 508 transcripts.

Nucleic acid labelling and raw microarray data generation were carried out at the Vanderbilt Microarray Shared Resource (http://array.mc.vanderbilt.edu/). Prior to amplification and labelling, the quality and quantity of total RNA isolated from tissues were determined using the Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA). Briefly, 750 ng of total RNA was used to transcribe DIG (Digoxigenin)‐labelled copy RNA (cRNA) using the Applied Biosystems NanoAmp™ RT‐IVT Kit. Ten micrograms of DIG‐labelled cRNA was hybridised onto the rat microarrays. Microarray processing, chemiluminescence detection, imaging, auto gridding and image analysis were performed according to manufacturer’s protocols and the AB1700 Chemiluminescent Microarray Analyzer software v. 1.0.3. Raw data were analysed using the ABarray data analysis package (http://bioconductor.org/packages/1.9/bioc/html/ABarray.html). Probe signal intensities across microarrays were normalised using the quantile method (20). Features with signal/noise values ≥3 and quality flag values <5000 were considered detected and were compared by t‐test using a fold change ≥1·6, a Benjamini and Hochberg false discovery rate of <0.05 (21) and/or a P value of <0·05. Lists of differentially expressed genes were then classified using the PANTHER™ database (http:www.pantherdb.org) 22, 23. Fold change values were calculated by dividing day 2 (wound) values by day 0 (unwounded) values.

Validation using TaqMan‐based quantitative real‐time PCR gene expression assays

Validation of the microarray data was performed by quantitative real‐time PCR. Approximately 1 μg of total RNA of each sample per 100 μl reaction was used to generate complementary DNA (cDNA) using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems). The TaqMan assays for the FAM™‐labelled gene of interest was duplexed with the VIC^®‐labelled 18S RNA endogenous control assay using a 7500 Fast Real‐Time PCR System (Applied Biosystems). For each sample, three technical replicates per gene were run in a 96‐well format plate. On each plate, a no‐template control was also run in triplicate. Relative quantification analysis was performed using SDS software v. 1.3.1 (Applied Biosystems).

Results

Global gene expression

Microarray‐generated global gene expression profiles detected an average of 14 559 probes in V.A.C.‐treated samples. This compares with averages of 14 543 and 13 681 probes in MWH and GUS, respectively. Genes based on expression levels (1·0‐fold to 4·0‐fold change) showed that expression levels within a treatment group are consistent; gene expression levels for V.A.C. samples group together tightly, and this was also observed in the other treatment groups as well using quantile normalised expression and hierarchical clustering (data not shown).

Using a P value cut off of ≤0·05, 5072 genes were found to be differentially expressed (either upregulated or downregulated) with a fold change ≥1·6 over the 2‐day treatment with V.A.C. Therapy (Table 1). In comparison, 3601 genes were significantly differentially expressed for wounds treated with MWH and 3952 genes were significantly differentially expressed in wounds treated with GUS.

Table 1.

Number of genes up (↑) and down (↓) regulated (P < 0·05) in the three wound‐healing treatment groups based on fold change (FC)

Global gene expression among treatment groups
V.A.C.®		MWH		GUS
↓ FC	Number of genes	↓ FC	Number of genes	↓ FC	Number of genes
1·6–10	1707	1·6–10	1176	1·6–10	1184
10–50	358	10–50	251	10–50	271
50–100	72	50–100	57	50–100	60
>100	78	>100	64	>100	95
Total	2215	Total	1548	Total	1610
↑ FC	Number of genes	↑ FC	Number of genes	↑ FC	Number of genes
1·6–10	2683	1·6–10	1988	1·6–10	2334
10–50	163	10–50	62	10–50	8
50–100	9	50–100	1	50–100	0
>100	2	>100	2	>100	0
Total	2857	Total	2053	Total	2342

GUS, gauze under suction; MWH, moist wound healing; V.A.C., vacuum‐assisted closure Therapy.

There were approximately 28% and 41% more genes (1·6‐fold to over 100‐fold change) upregulated and downregulated following V.A.C. than GUS treatment and MWH treatment, respectively. Of the 2857 upregulated genes during the first 2 days of V.A.C. Therapy, 479 were upregulated more than two times for V.A.C. than following GUS and 249 were upregulated more than two times for V.A.C. than following MWH. Conversely, of the 2342 genes upregulated following GUS treatment, only 13 were upregulated more than two times following GUS than following V.A.C. Of the 2053 genes upregulated during MWH, only 14 were upregulated more than two times following MWH than following V.A.C.

PANTHER was used to classify the pathways represented by the aforementioned genes. Genes in the following pathways were found to be significantly overrepresented in all three treatment groups, with V.A.C. having the most genes showing differential expression (summarised in Table 2): inflammation mediated by chemokine and cytokine signalling (V.A.C. 224 genes, MWH 176 genes and GUS 213 genes), integrin signalling (V.A.C. 177, MWH 125 and GUS 155), B‐cell activation (V.A.C. 62, MWH 47 and GUS 52), interleukin signalling (V.A.C. 100, MWH 68 and GUS 77), platelet‐derived growth factor (PDGF) signalling (V.A.C. 129, MWH 90 and GUS 92), cytoskeletal regulation by Rho guanosine triphosphatase (Rho GTPase) (V.A.C. 82, MWH 58 and GUS 70) and angiogenesis (V.A.C. 153, MWH 111 and GUS 120). There were also pathways where one or more treatment groups, either MWH or GUS, were not associated with significant gene overrepresentation. Those included pathways of oxidative stress response [V.A.C. 49, MWH (not significant) and GUS 38], G‐protein signalling [V.A.C. 90, MWH 69 and GUS (not significant)], vascular endothelial growth factor (VEGF) signalling [V.A.C. 53, MWH (not significant) and GUS 48], fibroblast growth factor (FGF) signalling [V.A.C. 74, MWH (not significant) and GUS 59], pentose phosphate pathway [V.A.C. 16, MWH (not significant) and GUS (not significant)] and apoptosis signalling [V.A.C. 85, MWH (not significant) and GUS 72].

Table 2.

Thirteen pathways of interest to wound healing expressed as number of genes observed over genes expected (obs/exp) in PANTHER™ and showing significant gene overrepresentation following 2 days of treatment with either vacuum‐assisted closure Therapy (V.A.C.^®), moist wound healing (MWH) or gauze under suction (GUS)

Differential expression of genes within pathways
Pathways of upregulated genes	Number of genes (obs/exp)
	V.A.C.	MWH	GUS
Inflammation mediated by chemokine and cytokine signalling pathway	244/105	176/75	213/82
Integrin signalling pathway	177/76	125/55	155/60
Angiogenesis	153/79	111/57	120/62
PDGF signalling pathway	129/62	90/45	92/49
Interleukin signalling pathway	100/57	68/41	77/44
G‐protein signalling	90/61	69/44	72/47 †
Apoptosis signalling	85/50	54/36 †	72/39
Cytoskeletal regulation by Rho GTPase	82/37	58/27	70/29
FGF signalling	74/46	53/33 †	59/36
B‐cell activation	62/30	47/22	52/24
VEGF signalling	53/28	35/20 †	48/22
Oxidative stress response	49/25	31/18 †	38/19
Pentose phosphate pathway	16/5	12/4 †	12/4 †

FGF, fibroblast growth factor; PDGF, platelet‐derived growth factor; Rho GTPase, Rho guanosine triphosphatase; VEGF, vascular endothelial growth factor.

Pathways with fold change greater or equal to 1·6 and P ≤ 0·05 were used for analysis.

^† Ratios that were not statistically significant (P > 0·05) are represented.

Twelve genes of interest, based upon high fold change and including selections from the overrepresented pathways previously mentioned, were examined more closely for potential development of wound‐healing signatures. These genes along with the signalling pathway represented and associated fold change are presented in Table 3. There is a higher fold change for the majority of the genes in the V.A.C. Therapy group of both upregulated and downregulated genes than in the other two treatment groups.

Table 3.

Genes that were identified for use in future screening for wound‐healing studies

Gene name	Signalling pathway	Fold change
		V.A.C.® M	V.A.C. q	MWH M	MWH q	GUS M	GUS q
Interleukin 6	Cytokine signalling/interleukin signalling	80·8	478·9	29·8	208·1	7·9	28·6
Chemokine (C‐C motif) ligand 7	Cytokine signalling	50·3	127·3	16·0	66·5	6·7	15·6
Tissue inhibitor of metalloproteinase 1	Metalloprotease inhibitor/proteolysis	39·3	159·9	35·8	170·5	5·1	31·1
Integrin alpha M	Integrin signalling	25·5	18·9	14·8	14·3	5·4	6·1
Suppressor of cytokine signalling 3	Inflammation/interferon gamma signalling	24·6	31·0	11·1	11·7	5·9	8·6
Matrix metallopeptidase 8	Plasminogen‐activating cascade	18·5	131·0	4·2	44·5	7·6	22·8
Macrophage inflammatory protein‐1 alpha receptor gene	Inflammation	15·2	34·0	6·2	19·5	3·8	7·9
Toll‐like receptor 2	Toll receptor pathway	15·1	63·8	4·1	37·1	5·4	13·0
Tumour necrosis factor receptor superfamily member 1b	Cytokine/immunity and defence	14·4	14·2	6·1	9·7	3·6	6·6
Heat shock protein 70	Apoptosis signalling pathway	2·9	6·5	2·6	5·5	1·8	4·8
Calmodulin‐like 3	G‐protein signalling	−208·8	−662·9	−126·9	−292·9	−38·1	−1265·2
Keratin complex 1 acidic gene 14	Intermediate filament/cell structure	−707·9	−6652·6	−314·4	−6716·1	−14·3	−17·8

GUS, gauze under suction; MWH, moist wound healing; V.A.C., vacuum‐assisted closure Therapy.

Fold changes for the genes were calculated for each treatment group form microarray (denoted by the M after the treatment group) and quantitative real‐time polymerase chain reaction data (denoted by the q after the treatment group). Ten upregulated and two downregulated genes (denoted by negative numbers) are shown. Genes were selected based on either high fold change or fold changes in one treatment group that were at least two times higher than those for other treatment groups. Differentially expressed genes with a P < 0·05 were considered significant.

Global gene relationships

When genes were evaluated for unique expressors, it was found that V.A.C. was associated with differential expression of 1180 unique genes not expressed in other treatment groups; 654 genes were unique to MWH and 1611 genes were unique to GUS (P < 0·05) (Figure 1). Analysis of relationships between groups for common expressors showed that 864 genes were common to both V.A.C.‐treated wounds and MWH‐treated wounds, 199 genes were common to both MWH‐treated wounds and GUS‐treated wounds and 1290 genes were common to both V.A.C.‐treated wounds and GUS‐treated wounds. There were also 2519 genes found to be common wound‐healing gene signatures because these genes were common between all three treatment groups.

Figure 1.

Venn diagram of overlapping resulting total gene lists (P < 0·05) by interactions between vacuum‐assisted closure Therapy (V.A.C.^®), moist wound healing (MWH) and gauze under suction (GUS).

Validation of microarray results

Microarray results were validated by quantitative real‐time PCR analysis of 27 individual genes. These genes were chosen by placing genes into a matrix to best represent all combinations in fold change bins from low (1·0–1·2), medium (1·2–4) and high (>4) and in expression levels from low, medium and high.

There was a 90% correlation between microarray and quantitative real‐time PCR results. This percentage may be somewhat low because of the fact that certain selected TaqMan inventoried gene assays might not have covered all possible gene transcripts for the gene of interest.

Wound‐healing phenotype

Wound areas were measured on all animals over the first 2 days of healing. V.A.C.‐treated animals had a significantly higher percentage wound closure than did the other treatment groups (P < 0·05; Figure 2). Means percent wound closure over the first 2 days of treatment were as follows: V.A.C. 14·0 ± 1·6%, MWH 5·4 ± 3·5% and GUS −4·3 ± 2·6%. The GUS group wounds showed no significant closure during the first 2 days of therapy; on the contrary, wound sizes increased from day 0 to day 2, which is believed to be a direct effect of the therapy in this model.

Figure 2.

Wound area measurements between the three different wound‐healing modalities of vacuum‐assisted closure Therapy (V.A.C.^®), moist wound healing (MWH) and gauze under suction (GUS). Percent change represents the average closure rate ± standard deviation. V.A.C. had significantly higher decrease in wound area than the other two modalities, whereas GUS caused an increase in wound area. V.A.C. was statistically significant (P < 0·05) over MWH and GUS as represented by the asterisk.

Discussion

Pathways

Analysing gene expression patterns are crucial to understanding the biology of wound healing, especially in the very early stages where it is not clearly understood why some wounds heal, while others become chronic. In this study, the impact of three treatment modalities (V.A.C., MWH and GUS) on gene expression during the early phases of wound healing in diabetic ZDF rats was assessed. The analysis showed that a substantial number of genes were differentially expressed over the first 2 days of healing with V.A.C. Therapy. Differentially expressed genes following V.A.C. Therapy mapped to pathways differently than for the other two modalities tested. Indeed, there were more pathways with significantly overrepresented gene numbers following V.A.C. Therapy than for GUS or MWH. Overrepresentation of differentially expressed genes in pathways indicates that the treatment is having a specific effect on certain pathways rather than arbitrarily affecting all pathways in the same manner. The overrepresented pathways are of known importance to wound healing. For example, inflammation, the second phase of wound healing, is characterised initially by the infiltration of neutrophils and later by macrophage phagocytosis of neutrophils (24) and the release of PDGF. PDGF, along with other factors, leads the wound from the inflammation phase and into the proliferation phase of wound healing 25, 26. As shown in Table 2, V.A.C. Therapy resulted in a significant overrepresentation (both upregulated and downregulated genes) of the inflammation pathway. This included inflammation mediated by chemokine and cytokine signalling pathways, interleukin signalling and PDGF signalling. Interleukins are a class of cytokines and are important in the early stages of inflammation. While inflammation is necessary for proper wound healing, resolution of the inflammatory response in a timely manner is important to successful healing. Failure to resolve inflammation in a timely fashion is associated with wound chronicity (27). The inflammation pathway genes overrepresented included both inflammatory and anti‐inflammatory genes. For example, suppressor of cytokine signalling 3 and 4 and IL‐10 that modulate inflammation were all upregulated.

The third, or proliferative, phase of wound healing typically occurs from day 1 to day 30 and often peaks at day 3 following wounding in humans and is characterised by the migration of fibroblasts into the wound (15). Once in the wound, fibroblasts proliferate and are also involved in the production of extracellular matrix to help fill the wound with granulation tissue (15). Cell movement, including migration, is associated with several different pathways including Rho GTPase‐ and integrin‐mediated pathways. In tissues treated with V.A.C. Therapy for 2 days, the cytoskeletal regulation by the Rho GTPase pathway was significantly overrepresented. V.A.C. Therapy‐treated wounds differentially expressed 82 genes, MWH‐treated wounds differentially expressed 58 genes and GUS‐treated wounds differentially expressed 70 genes. This pathway affects cytoskeletal elements such as actin and microtubules, which enable cells to move and change shape. Also important during the proliferative phase of wound healing is integrin signalling. Integrins provide a mechanical connection between matrix components and are involved in the transmission of a variety of cell communication signals important in pathways such as angiogenesis and cell migration 28, 29. As mechanotransducers, integrins help sense and transduce mechanical stresses, such as macrostrain and microstrain. Both strain types are imparted to tissues when negative pressure is manifolded through GranuFoam Dressing during V.A.C. Therapy (30).

Wound‐healing phenotype and genes

Of clinical importance is the fact that wound area over the first 2 days of treatment differed significantly between treatment groups. The greatest reduction in wound area occurred in animals treated with V.A.C. Therapy, while animals treated with GUS experienced an increase in wound area. The significant increase in size in wounds from GUS‐treated animals may be because of the fact that GUS has a negative impact on the early stages of wound healing in diabetic rats.

A goal of this study was to identify a wound‐healing gene signature. This is of interest because of its potential as a prognostic indicator of successful wound healing. When looking at gene relationships, it is not necessarily the number of genes that are most important but rather how the patterns of uniquely expressed genes within a treatment group, along with differential fold changes of shared genes, correlate to a strong clinically relevant response. In this study, this response is wound healing as assessed by wound area. In V.A.C. Therapy‐treated animals, there were a large number of uniquely expressed genes, as well as shared genes, with higher fold changes. This gene signature correlated with a significantly greater decrease in wound area (Figure 2). On the contrary, the uniquely expressed genes following GUS were associated with an increase in wound size. For the V.A.C.‐treated group, there must be a unique gene signature, which caused the wound area to decrease significantly by day 2 of healing.

The 12 genes potentially important to wound healing were chosen from pathways shown in Table 2, based on their higher fold changes in V.A.C. Therapy compared with the other two groups. The gene targets selected were IL‐6, chemokine ligand 7, tissue inhibitor of metalloprotease 1 (TIMP1), integrin alpha M (Itgam), suppressor of cytokine signalling 3 (Socs3), matrix metalloproteinase‐8 (MMP‐8), macrophage inflammatory protein‐1 alpha receptor gene (MIP1), toll‐like receptor 1 (TLR1), tumour necrosis factor receptor superfamily member 1b, heat shock protein 70 (Hsp70), calmodulin‐like 3 and keratin complex 1 acidic gene 14 (Table 3). Each of these genes belongs to pathways important in successful wound healing.

Under subatmospheric pressure, the GranuFoam Dressing used in V.A.C. Therapy imparts both macrostrain and microstrain to tissue (30). These micromechanical forces have been associated with cellular responses such as cell signalling. Recent research supports the theory that cells sense changes in their environment through expression of integrins such as Itgam (31). As shown in Table 3, following V.A.C. Therapy, Itgam exhibited a 72·3% higher fold change than following MWH and a 372·2% higher fold change than following GUS.

IL‐6, MIP1, TLR1 and Socs3 are representative of various inflammatory pathways (as shown in Table 3) and are all associated with successful wound healing 32, 33, 34, 35. IL‐6 is involved in growth and differentiation of cell types. It has been shown that IL‐6 knock out mice display significantly delayed cutaneous wound healing (32). It has also been shown that toll‐like receptor pathways can be activated by endogenous inflammatory stimuli (36). A function of toll receptors is one of epithelial homoeostasis and protection from epithelial injury. They may also directly induce the expression of factors including heat shock proteins, IL‐6 and tumour necrosis factor, which are involved in tissue repair and in protection of tissue from injury (33). Socs3 is a key modulator of cytokine signalling by proteins that attenuate signal transduction (35). In this study, the Socs3 gene was upregulated in V.A.C. Therapy‐treated tissues 2·2 times higher than that for MWH‐treated tissues and 4·2 times higher than that for GUS‐treated tissues. It has previously been observed that an induction of Socs3 early upon skin injury provides a decrease in inflammatory potency of rapidly induced cytokines at the wound site 35, 37. This corroborates previous proteomic results that indicated that V.A.C. Therapy may function in part through modulation of the inflammatory response 38, 39. This modulation of the inflammatory response at the genomic level may then be one of the factors that led to the significant decrease in wound area seen in this study.

V.A.C. Therapy promotes granulation tissue formation in both acute and chronic wounds, including complex diabetic foot wounds 8, 11, 40, 41. The remaining members of the 12 genes selected for validation and gene signatures may be involved in the production of this granulation tissue. Granulation tissue is mainly composed of extracellular matrix, endothelial cells and fibroblasts. Induction of Hsp70, a molecular chaperone, has been associated with the development of thermotolerance and protection against various stresses including hypoxia and ischaemia (42). Cells within the wound bed should contain significant amounts of inducible Hsp70 in an effort to maintain proper function within the healing wound (43) because cells must be maintained and protected from wound environment stresses during the early granulation phase.

MMP‐8 is the predominant collagenase present in the early phases of acute, healing wounds. Excessive collagenolytic activity has been associated with reduced levels of TIMP1 in chronic wounds (44). In this study, levels of MMP‐8 and TIMP1 mRNA were both greatly increased following 2 days of treatment with V.A.C. Therapy. Following 2 days of treatment, TIMP1/MMP‐8 ratios were 2·2 for V.A.C., 8·5 for MWH and 0·7 for GUS (Table 3). While extracellular matrix degradation is important for wound healing, excessive degradation by MMP‐8 is associated with poor healing outcomes (45). The high fold change increase observed for TIMP1 mRNA may be important for keeping wound proteolytic activity from becoming detrimental to the wound‐healing process.

Conclusions

Overall, more genes were upregulated and healing pathways overrepresented in tissues that had been treated with V.A.C. Therapy than with the other two treatment modalities tested. Wound area was also decreased significantly in V.A.C. Therapy‐treated animals. To facilitate a better understanding of wound‐healing gene signatures, 12 genes were selected as potential markers of successful wound healing. Future studies will assess levels of IL‐6, Itgam, Socs3 and Hsp70 mRNA over the first 7 days of healing by quantitative PCR and corresponding protein levels will also be evaluated. Additional time points and testing will provide a more concrete wound‐healing signature. Further evaluation of the unique genes differentially expressed during V.A.C. Therapy and finding selected gene targets may provide diagnostic insight into the early evaluation of treatment success and provide explanations of how V.A.C. Therapy creates an environment that promotes wound healing.