A universal immunohistochemical kit, Elite ABC kit, was purchased from Vector Laboratories (Peterborough, UK). The total RNA isolation reagent was purchased from Sigma-Aldrich (Dorset, UK) and reverse transcription kits (iScript) were obtained from Bio-Rad Laboratories (Hemel Hempstead, UK).

Skin biopsies were obtained from patients attending the University Hospital of Wales (UHW, Cardiff, UK) wound healing clinic, as described previously (23,24). Ethical approval was granted by the Local Research Ethics Committee. Written informed consent was obtained from each patient.

Biopsies from 14 patients with chronic leg ulcers were used during the study. Venous disease was diagnosed by duplex ultrasonography and all the wounds were present for ≥6 months, with no evidence of healing occurring 6 weeks before biopsy. The wounds had a minimum area of 4 cm² prior to biopsy and had no clinical indications of infection. Using an aseptic technique, 6-mm punch biopsies were removed, following the application of local anesthetic (1% lidocaine), from the wound margin, incorporating epidermis and dermis at the wound edge with adjacent granulation tissue.

Single wedge biopsies were obtained from 10 patients with acute surgical wounds subsequent to undergoing excision of pilonoidal disease. These wounds were judged to be clinically noninfected. The biopsies were obtained from the edge of the healing wound within 6 weeks from the surgical excision.

To provide a comparison to wound tissue, normal skin tissue was also examined. Under local anesthetic, 3-mm punch biopsies were removed from the inner aspect of the upper arm of 10 healthy volunteers working within the Wound Healing Research Unit (School of Medicine, Cardiff University, Cardiff, Wales).

Frozen sections from wound tissues were first fixed in an acetone/methanol solution and rehydrated in wash buffer (MenaPath Autowash buffer; A. Menarini Diagnostics, Berkshire, UK) prior to placing the samples into a wash buffer solution containing 10% horse serum to aid in the blocking of non-specific antigen binding. An avidin/biotin complex (ABC) immunohistochemistry (IHC) kit (Vector Laboratories, Nottingham, UK) was used in accordance with the manufacturer's protocol. A polyclonal antibody to TEM-8, previously generated (9), was used and diluted in a buffer that contained 1% horse serum and 0.1% Tween-20 at 1:40 dilution. After a 1-h incubation period with the primary antibody, the slides were washed 4 times in a washing buffer and a universal biotinylated secondary antibody was added for 30 min. Following washing, avidin and biotin were added through the addition of the ABC complex. A DAB colour developing system was used to indirectly detect protein staining. Sections were dehydrated through a series of graded alcohols, cleared in xylene, mounted and evaluated on an Olympus microscope equipped with a digital camera (Olympus, Southend-on-Sea, UK).

The HaCaT human keratinocyte cell line was purchased from the German Cancer Research Institute (Heidelberg, Germany). Cells were maintained in Dulbecco's modified Eagle's medium supplemented with penicillin, streptomycin and 10% fetal calf serum (PAA Laboratories Ltd., Somerset, UK). The cells were incubated at 37°C, 5% CO₂ and 95% humidity.

Cells were grown to confluence in a 25-cm² flask prior to RNA extraction using the total RNA isolation reagent in accordance with the manufacturer's protocol. Sample RNA was quantified using a spectrophotometer (WPA UV 1101; Biotech Photometer, Cambridge, UK) and standardized to a concentration of 500 ng prior to being used as a template to reverse transcribe cDNA using an iScript cDNA synthesis kit (Bio-Rad Laboratories). Following cDNA synthesis, samples were probed using glyceraldehyde 3-phosphate dehydrogenase (GAPDH) primers to check the cDNA quality and confirm uniform sample cDNA levels, together with those specific for TEM-8 transcript (Table I) as previously reported (9,11).

Conventional RT-PCR primers were designed using Beacon designer Software (Beacon Designer, Palo Alto, CA, USA), to allow amplification of regions that have no overlap with known genes and span at least one intron. Primers were synthesized by Invitrogen (Paisley, UK). Conditions for conventional RT-PCR to amplify transcripts of TEM-8 were: 94°C for 40 sec, 54°C for 30 sec, 72°C for 50 sec and a final extension phase of 10 min for 34–36 cycles, with GAPDH used as the reference gene. The PCR products were separated on a 0.8% agarose gel and stained with ethidium bromide prior to examination under UV light.

RT-qPCR was used to determine the transcript expression levels of TEM-8 in wound tissues. This methodology has been reported previously (25). Briefly, the iCycler IQ system (Bio-Rad, Camberley, UK) was used to quantify the transcript level of TEM-8 within each of the clinical samples. Results are provided as number of transcripts per microlitre and are based on an internal standard run and amplified in conjunction with the samples. Normalisation of samples was achieved through comparison of sample GAPDH levels. The Amplifluor system (Intergen Inc., New York, NY, USA) was used in conjunction with a universal probe (UniPrimer), which recognised a specific sequence (z sequence), incorporated into the primers. RT-qPCR conditions were as follows: 95°C for 15 sec, 54°C for 20 sec and 60°C for 60 sec.

Hammerhead ribozyme transgenes, specifically targeted to TEM-8 transcripts were constructed based on the secondary structure of TEM-8 mRNA, as described previously (10) (Table I). Following the design and synthesis by Invitrogen, two separate ribozymes, targeting different regions of the TEM-8 transcript, were cloned into a mammalian pEF6/His TOPO vector (Invitrogen) and transfected into HaCaT cells, as previously reported (10). Following a period of blasticidin selection (5 µg/ml), the cells were maintained in media containing 0.5 µg/ml blasticidin. This process allowed the generation of stably transfected HaCaT cells containing the ribozyme transgene and expressing reduced TEM-8 levels (HaCaT^{ΔTEM-8 rib1} and HaCaT^{ΔTEM-8 rib3}) and control HaCaT cells transfected with a closed pEF6 plasmid (HaCaT^pEF6). Suppression of TEM-8 expression in HaCaT^{ΔTEM-8 rib1} and HaCaT^{ΔTEM-8 rib3} cells was verified, in comparison to HaCaT^pEF6 controls, using RT-PCR.

The effect of TEM-8 suppression on HaCaT cell growth rates was assessed using an in vitro growth assay. Cells were seeded at a density of 3,000 cells/well in 96-well plates. Triplicate plates were set up and incubated for 3- and 5-day periods before analyses. Following incubation, the plates were fixed in 4% formaldehyde (v/v) and stained with 0.5% (w/v) crystal violet and were subsequently treated with 10% acetic acid (v/v), prior to colorimetric detection of cell density by spectrophotometric analysis at 540 nm using a Bio-Tek ELx800 mutliplate reader (Bio-Tek Instruments Inc., Winnoski, VT, USA).

The ECIS 9600 system (Applied Biophysics Inc., Troy, NJ, USA) was used to detect and track HaCaT cell migration, as described previously (26,27). Briefly, cells were simultaneously plated in ECIS 8W10 arrays and incubated until a confluent monolayer had formed over the array electrodes. This monolayer was subsequently wounded electrically by applying 6V for a 30-sec time-period to create a simultaneous physical break in the cell monolayer of equal dimensions. The rate of change in resistance as cells migrated back onto the electrode, was subsequently monitored and measured using the ECIS software provided. Prior to use, ECIS arrays were treated with L-cysteine solution for 40 min followed by several washes with complete medium to remove any particles present on the electrode.

The SigmaPlot 11 statistical package (Systat Software, San Jose, CA, USA) were used to identify statistical differences between the test groups using one-way analysis of variance (ANOVA) or ANOVA on RANKS tests. P<0.05 was considered to indicate a statistically significant difference. All the in vitro functional assays were repeated a minimum of three times.

TEM-8 is minimally expressed in all layers of the epidermis of normal skin. However, in acute wounds, expression of TEM-8 increases within the cytoplasm of keratinocytes and endothelial cells, consistent with its putative role in angiogenesis. Expression was also increased in chronic wounds compared to normal skin, but is lower compared to the acute wounds (Fig. 1A–C).

Consistent with the immunohistochemical analyses, RT-qPCR detection and normalisation of TEM-8 levels also revealed higher TEM-8 expression in acute (median, 3.964; IQR, 12.049-0.0445) and chronic (median, 1.755; IQR, 7.426-0.144) wounds and again, extremely low levels were observed in normal skin (median, 0.381; IQR, 1.996-0.0629), although no significant differences were observed within the group (P>0.05) (Fig. 1D).

To determine the effect of reduced TEM-8 expression on keratinocyte cell growth, HaCaT keratinocytes were transfected with ribozyme transgenes specifically targeted to the TEM-8 transcript to reduce TEM-8 transcript levels within HaCaT cells, as assessed by RT-PCR (Fig. 2A).

In HaCaT keratinocytes in which TEM-8 expression was reduced, a decrease was also observed in the in vitro growth rate of the HaCaT cells (Fig. 2B). The growth rate of HaCaT^{ΔTEM-8 rib1} and HaCaT^{ΔTEM-8 rib3} cells over a 5-day incubation period was significantly lower compared to the HaCaT^pEF6 control cells (P=0.004 and P=0.014, respectively). No significant differences between the groups were observed over a 3-day incubation period (P>0.05).

As TEM-8 expression was reduced in non-healing wounds compared to acute wounds and suppression of TEM-8 expression also reduced cell growth, the effects of TEM-8 knockdown on keratinocyte migration directly were investigated. An ECIS assay was carried out to examine the effects of TEM-8 suppression on HaCaT migration following electrical wounding (Fig. 3). Substantial differences between the migratory rates of HaCaT^pEF6, HaCaT^{ΔTEM-8 rib1} and HaCaT^{ΔTEM-8 rib3} were observed as the cell lines responded to re-colonise the wounded area of the array. HaCaT^pEF6 cells migrated at a steady rate following wounding as indicated by the increase in resistance recorded across the array. In contrast to this, HaCaT^{ΔTEM-8 rib1} and HaCaT^{ΔTEM-8 rib3} cells showed a reduced capacity to migrate into the wound and recover the monolayer over the experimental time (Fig. 3A). Three-dimensional analysis further demonstrated this trend, showing changes in resistance across a range of tested frequencies and time for HaCaT^pEF6 (Fig. 3B), HaCaT^{ΔTEM-8 rib1} (Fig. 3C) and HaCaT^{ΔTEM-8 rib3} (Fig. 3D).