Patients

All clinical investigations were conducted in accordance with accepted ethical and humane practices as stated by the Declaration of Helsinki principles. Written consent was obtained from all patients.

The patient whose skin was examined in the present study suffered from an accidental whole-body exposure to X rays associated with a major cutaneous radiation syndrome characterized by an extensive moist epidermitis without necrosis. The radiological cutaneous syndrome involved the whole posterior side of the thorax from the waist up to the end point of the scapulae. Based on clinical signs and according to the medical staff in charge of the patient, three different concentric skin areas were defined. At the periphery of the skin lesion, a first area was defined as healthy. This area was exposed to X-ray doses lower than 5.2 Gy. Concentric and enclosed in this area, a second area characterized by “normal healing” tendency was defined. This inflamed area was exposed to X-ray doses ranging from 5.2 to 18.9 Gy. The third area was the center of the lesion. This area, i.e. the “pathological wound healing area”, was the most exposed area (>25 Gy) and was characterized by an extensive moist epidermitis without healing tendency. Dose mapping was obtained according to the biological dosimetry procedure as described in our previous paper (8). The whole lesion including the three areas was excised surgically at day 88 after irradiation, and a skin graft was performed.

Excess surgical tissue from resected abdominal skin from three unirradiated patients (Caucasian men 35–45 years old) was obtained and used as controls.

Tissue Sampling

Additional small biopsies were taken from the excised skin of both central patients and the patient suffering from the radiological cutaneous syndrome (Fig. 1B). In both cases, some of the tissue samples were fixed in formalin and embedded in paraffin, while others were snap-frozen in liquid nitrogen and stored at −80°C.

Skin Histopathology

A primary histological examination of a 5-μm tissue section was performed by staining with hematoxylin and eosin (H&E) after fixation in 5% formaldehyde.

Skin RNA Extraction and Quality Control

Whole frozen tissue samples (epidermis and dermis) were crushed into powder under liquid nitrogen. Samples were then homogenized in 4 M guanidium thiocyanate using 1.2 gauss syringes. Total RNAs were purified according to the technique of Chomczynski and Sacchi after phenol/ chloroform extraction. The RNA was treated with RNase-free DNase (0.5 U/μl) to remove contaminating genomic DNA. The success of the DNA digestion, in both quantity and quality (28S/18S ratio), of the RNAs extracted was determined for each sample using an Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA).

Skin mRNA Expression Analysis using cDNA Arrays

The gene expression analysis was performed through hybrid selection of radiolabeled cDNA on high-density arrays of membrane-bound cDNAs. The Atlas Human 1.2 (1176 genes + 9 housekeeping genes) cDNA expression array (Clontech Laboratories, Palo Alto, CA) was used for this experiment. PolyA+ RNA enrichment was obtained after column chromatography by using a streptavidin magnetic bead preparation according to the Atlas Pure total RNA labeling system techniques (Clontech Laboratories). The cDNA was synthesized as described in the Atlas cDNA Expression Array kit and 370 kBq/μl [α-³²P]dATP was incorporated into synthesized cDNA. To purify the labeled cDNA from labeled ³²P nucleotides and small cDNA fragments, each sample was purified using column chromatography. The radioactivity of the probe was verified by scintillation counting (Beckman Coulter, Fullerton, CA). The cDNA probes were hybridized overnight (a 20-h incubation period was optimized to obtain an acceptable hybridization signal with high signal/background ratio and to avoid any signal covering a neighboring area) in the Atlas array with continuous agitation at 68°C. After the hybridized membrane was washed according to the Atlas cDNA expression array protocol, the Atlas array was removed from the container and wrapped in plastic wrap. The Atlas arrays were in contact with a phosphorimaging screen for 20 h at room temperature and revealed by using a Typhoon 9400 phosphorimager (Amersham Biosciences Europe, Freiburg, Germany). Comparisons between the different samples were made using BD Atlasimage 2.7 software (Clontech Laboratories). Four housekeeping genes were selected (phospholipase-2, gyceraldehyde-3 phosphate dehydrogenase, ribosomal protein L13a and tyrosine 3 monooxygenase/tryptophan 5 monooxygenase activation protein), and only the irradiated:control ratios above 2 and below 0.5 were taken into account.

Skin Differential Gene Expression as Confirmed by Real-Time PCR

To provide independent confirmation of the array data, real-time RT-PCR was performed on selected transcripts. Briefly, 2 μg of skin extract total RNA was reverse transcribed with Superscript II reverse transcriptase (Life Technology-Invitrogen, Cergy Pontoise, France) with 200 U of Superscript reverse transcriptase in a 20-μl reaction containing 1× Superscript buffer, 1 mM dNTP, 20 ng random hexamer, 10 mM DTT and 20 U of RNase inhibitor. This mixture was incubated for 50 min at 42°C and 10 min at 70°C. For TNFA and IL1B, primers from the manufacturer were used to amplify first-strand cDNA through 40 PCR cycles with TaqMan (Applied Biosystems, Foster City, CA). The PCR amplification of the other genes used the Syber PCR master Mix; the primer sequences purchased from Life Technology-Invitrogen (Cergy Pontoise, France), were as follows: MnSOD, 5′-GGT-GGTCATATCAATCATAG-3′ (forward); 5′-AGTGGAATAAGGTTTGT-TGT-3′ (reverse), Cu/ZnSOD, 5′-CAGTGCAGGTCCTCACTTTA-3′ (forward); 5′-CCTGTCTTTGTACTTTCTTC-3′ (reverse), catalase, 5′-TTA-ATCCATTCGATCTCACC-3′ (forward); 5′-GGCGGTGAGTGTCAGGAT-AG-3′ (reverse). Samples were subjected to 40 cycles of amplification at 95°C for 15 s, followed by 60°C for 1 min using ABI PRISM 7500 detection system (Applied Biosystems). Both water and genomic DNA controls were included to ensure specificity. PCR fluorescent signals were normalized to the signal obtained from the housekeeping gene 18s for each sample. All data were examined for integrity by analysis of the amplification plot and dissociation curves. Relative mRNA quantification was performed using the comparative ΔΔC_T method. Relative quantification in the skin biopsy was equal to 2^−ΔΔCT (10). ΔΔC_T is defined as the difference between the mean ΔC_T (exposed skin) and the mean ΔC_T (healthy skin). ΔC_T is defined as the difference between the mean C_T (antioxidant enzyme) and C_T (18s), which is the endogenous control in our experiment.

Skin Immunostaining

Five-micrometer sections were cut using a cryostat (OTF, Bright, England) at −30°C. Frozen sections were fixed with either paraformaldehyde (0.5%, 30 min) or acetone (4°C, 10 min). Sections were then permeabilized with Triton X-100 (0.1%, 20 min), washed in PBS, and saturated with PBS/BSA solution (2%). The sections were then incubated overnight (4°C) with two different antibodies [CD 3 T cell (DAKO, Trappe, France), MPO (Novocastra Laboratory, Newcastle upon Tyne, UK)]. Immunostaining was performed using a Nexes IHC automated immunostainer (Ventana, Illkirch, France). The sections were fixed [glutaraldehyde (0.05%, 4 min, 37°C)], incubated with goat anti-mouse antibody and then incubated with streptavidin-alkaline phosphatase, Fast Red and naphthanol. Sections were counterstained with hematoxylin. The sections were mounted in PBS-glycerol (50%) and examined with an epifluorescence microscope. A semi-quantitative approach using an image analyzer Visioscan (Biocom, Paris, France) was chosen. Ten different fields were measured on each slide, and six different slides were analyzed for each region.

Measurement of Skin Antioxidant Defenses

Tissue samples stored at −80°C were homogenized in pH 7.5 phosphate buffer with an ultra-turrax (IKA Labortechnick, Munchen, Germany) at 4°C. The skin homogenate was then centrifuged, and the supernatant was evaluated for protein content using the Coomassie protein assay. The decomposition of hydrogen peroxide catalyzed by catalase contained in the skin homogenate supernatant was followed spectrophotometrically at 240 nm on a Uvikon 840 (Biotek-Kontron, Saint Quentin en Yvelines, France). The H₂O₂ concentration was adjusted spectrophotometrically at 240 nm. The activity was measured from the slope of the line as micromoles of H₂O₂ decomposed per minute per milligrams of protein. Glutathione peroxidase activity was measured by a standard coupled assay procedure. The reaction mixture consisted of phosphate buffer (0.5 M, pH 7.0), 3.12 mM EDTA, 12.5 mM KCN, 0.005 mM glutathione, 0.86 mM NADPH, H⁺ and 0.25 IU ml⁻¹ glutathione reductase. Erythrocyte hemolysate was added to the above mixture and was incubated for 15 min at 30°C before initiation of the reaction by addition of 100 μl tert butyl hydroperoxide. Absorption at 340 nm was recorded for 10 min on a Uvikon 840. The activity was calculated from the slope as micromoles of NADPH oxidized per minute. The total SOD was quantified in the skin homogenate supernatant by its capacity to inhibit the reduction of nitro blue tetrazolium (0.012 μM) by superoxide, which was produced by a xanthine (0.02 mM)/xanthine oxidase (0.1775 IU/ml) system at 30°C. The appearance of reduced NBT was monitored at 560 nm on a Uvikon 840. SOD prevents the reduction of NBT by superoxide and decreases the slope of the curve for the appearance of reduced NBT proportionally. The supernatant of the skin homogenate was diluted 1/10 to ensure a decrease in the slope between 20 and 30%. To determine the MnSOD activity, the assay was carried out in the presence of 1 mmol/liter potassium cyanide. The Cu/ZnSOD activity was obtained by subtracting the value for MnSOD from that of the total SOD.

Skin Cytokine Content

Tissue samples stored at −80°C were homogenized in phosphate buffer containing a protease inhibitor cocktail using an Ultra Turrax at 4°C. The skin homogenate was then centrifuged and the protein content of the supernatant was evaluated using a Coomassie protein assay. IL1B, IL6, IL8, TNFA, IL1RA and IL10 levels were measured in the supernatant by specific ELISA according the manufacturer's recommendation (R&D Systems, Abingdon, UK). The cytokine content was determined by specific ELISA according to the manufacturer's recommendations

Statistical Methods

All results are expressed as means ± the standard errors of the means (SEM) for duplicate skin biopsies from three different healthy patients and for duplicates of three different skin biopsies from the irradiated patient. Significance tests used an ANOVA associated with Dunn's or Dunnett's tests. Significance was set at P < 0.05.

Ionizing Radiation Induces Specific Morphological Changes as Revealed by the Histological Examination

The examination of H&E sections (Fig. 1C) revealed common histological features of moist desquamation, depending on the position within the skin lesion. Marked epidermolysis associated with a loss of adhesion of the epidermis to the basal layer and microvascular destruction was observed in the center of the lesion. The lack of dermal necrosis, as noted after examination of H&E-stained sections, was confirmed by BAX immunohistochemical staining (Fig. 1C). Also, the epidermal hyperproliferative response measured using Ki67, the perivascular inflammatory cell infiltration, and the extracellular matrix protein deposition, i.e. collagen, revealed a healing process surrounding the lesion (Fig. 1C).

Ionizing Radiation Induces a Specific Set of Genes as Revealed by the cDNA Array Technique

The intensity of the arrays from the healthy skin (pool of three different human skin biopsies) was averaged and compared with the intensity of arrays from two different irradiated areas. The first area sampled was located in the “healthy area” (position G5; see the Patients and Methods section and Fig. 1B) and the second was located in the “pathological healing area” (position G33; see Patient and Methods section and Fig. 1B). The differentially expressed genes represent 11% (145/1314) of the genes spotted on the membrane, including mainly genes involved in the control of the cellular redox status (Fig. 2) and in the control of the inflammatory response (Fig. 3).

Collapse of Skin Antioxidant Status during Pathological Wound Healing after Radiation Exposure

Interrelationship of Skin Redox Modulation and Skin Inflammatory Response to Ionizing Radiation

Radiation-Induced Modulation of Intracellular and Nuclear Signaling

Involvement of classical intracellular redox-sensitive signaling was observed using the array technique. It was interesting to note the modulation of gene expression of the homeobox family in the skin after radiation exposure. Expression of the genes encoding HOXA5, B5 and B7 was up-regulated in both normal and pathological healing skin. A 25-fold increase was noted for HOXB5 for normal healing skin. Increases of 26- and 35-fold were observed for PBX1 and HOXA5, respectively, in pathological skin (Table 1).