Data from 157 patients with rectal cancer who received permanent colostomy between March 2008 and September 2013 at Beijing Chao-Yang Hospital (Beijing, China) were reviewed retrospectively. All patients underwent abdominal perineal resections. Abdominal CT scans were performed every 6 to 12 months to evaluate the development of PH during the follow-up stage. PH was identified in a total of 55 cases. Information concerning age, sex, BMI (kg/m²), diabetes, hypertension, radiation history, length of hospital stay (days), aperture size and expression of types I and III procollagen mRNA from each patient were collected during outpatient visits in the hospital. Informed consent was provided by each patient and all experiments were authorized by the Ethics Committee of Beijing Chao-Yang Hospital. The dermal tissues and matched normal tissues were obtained from 55 patients with PH for the reverse transcription quantitative polymerase chain reaction (RT-qPCR) and western blot analysis.

Briefly, fibroblasts were taken from tissues 1-mm distance away from the center of the skin samples. The tissues were placed on the upper wall of the culture bottle, to which was added Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% FBS (both Gibco; Thermo Fisher Scientific, Inc.), 100 U/ml penicillin and 100 mg/ml streptomycin in a 90% humidified incubator with 5% CO₂ at 37°C. After 4 h, the culture bottle was turned over carefully and the tissues were immersed into the medium. Primary fibroblast cells were formed and subjected to cell identification following tissue culture for 3-14 days. The cells were amplified, passaged weekly at 80-90% confluence and used at early (between 2nd and 4th) passages (P) to avoid in vitro replicative-induced ageing. At P2 or P3, the cells were seeded at 5,000 cells/cm² for all the experiments, unless otherwise stated.

The fibroblasts were fixed in 4% paraformaldehyde for 20 min at room temperature and extracted in 0.5% Triton X-100 for 10 min. Following washing in PBS, the samples were incubated with rabbit anti-human vimentin monoclonal antibody (cat. no. AX10005; 1:200; Abgent, Inc.) for 1 h at room temperature and then with fluorescein isothiocyanate-conjugated donkey anti-rabbit IgG (cat. no. R37119; 1:100; Molecular Probes; Thermo Fisher Scientific, Inc.). Nuclei were stained with DAPI (5 μg/ml; Abgent, Inc.) for 15 min in the dark and images were observed with a fluorescence DM5000 B microscope (magnification, ×200; Leica Microsystems, Inc.).

Total RNA from PH tissues or tissues without PH (WPH) and skin fibroblasts was extracted by TRIzol^® (Thermo Fisher Scientific, Inc.). TRIzol^® reagent and chloroform were added to the samples and mixed for 5 min. The samples were then centrifuged at 2,000 x g for 10 min at room temperature to recover the supernatant. Next, the supernatant was incubated with an equal volume of isopropyl alcohol at 0°C for 5 min, followed by centrifugation at 12,000 x g at 4°C for 10 min. Following removal of the supernatant, the 75% ethanol was added to wash the precipitPate and the RNA was eluted with nuclease-free water. The purity and content for the reverse transcription were determined using NanoDrop 2000 spectrophotometer (NanoDrop Technologies; Thermo Fisher Scientific, Inc.). The cDNA was obtained by RNA with mixture in PrimeScript™ 1st Strand cDNA Synthesis kit (Takara Biotechnology, Co., Ltd.). The reactions were conducted using the following primers: α1 (I) procollagen forward, 5′-GTTCGTCCTTCTCAG GGTAG-3′; α1 (I) procollagen reverse, 5′-TTGTCGTAGCAGGGTTCTTT-3′; α1 (III) procollagen forward, 5′-CGAGGTAACAGAGGTGAAAGA-3′; α1 (III) procollagen reverse, 5′-AACCCAGTATTCTCCACTCTT-3′; β-actin forward, 5′-GGTTACCTCCCATCAGCT-3′; and β-actin reverse: 5′-CAGTGTCCGGAAATCTCC-3′, using a LightCycler system (Roche Diagnostics) using the under the following thermocycler conditions: 94°C for 4 min, then 40 cycles at 94°C for 45 sec, 56°C for 45 sec and 72°C for 2 min. The results were calculated using the 2^−ΔΔCq method (23).

Cells (~5×10³) were maintained on 96-well plates for the measurement of proliferation using a Cell Counting Kit-8 (CCK-8; Dojindo Molecular Technologies, Inc.). Following culture for 12, 24 and 48 h, the cells were incubated with CCK-8 solution (10 μl) at 37°C for 2 h. Absorbance was read at 450 nm using an iMark plate reader (Bio-Rad Laboratories, Inc.).

A total of 2 μg small interfering RNA (siRNA; 5′-UCAACCAGACCACCUUAU AdT dT-3′; Shanghai GenePharma Co., Ltd.) was used to silence TIMP-1 expression. A negative control (NC; cat. no. A06001; Shanghai GenePharma Co., Ltd.) was also included. The cells (4×10⁵) were seeded in 6-well plates. After culture for 24 h, the medium was replaced by Opti-MEM (Invitrogen; Thermo Fisher Scientific, Inc.). The fibroblasts were transfected with TIMP-1 siRNA using Lipofectamine^® 2000 (Invitrogen; Thermo Fisher Scientific, Inc.). According to the manufacturer's protocol. The transfection efficiency was assessed using western blot analysis. The subsequent experiments were performed 24 h after transfection.

Cell migratory ability was measured using specific wound assay chambers (ibidi GmbH). Fibroblasts (1×10⁶) were seeded onto 24-well plates. Following attachment of the cells, a wound between the fibroblasts was created using a sterile pipette tip (10 μl). Then, the fibroblasts were washed and incubated for 24 h. The cell migration from 0 to 24 h was observed under a phase-contrast inverted microscope (magnification, ×200; Olympus Corporation) with a TUCSEN camera. The cell migration was measured with Wimasis Image Analysis (Onimagin Technologies SCA) at 0 and 24 h.

The fibroblast invasion capabilities were determined using a Matrigel-coated Transwell assay (Corning Incorporated). Following serum starvation overnight, the fibroblasts were added into the upper chambers in DMEM containing 1% FBS. Concomitantly, the lower chambers were filled with 500 ml DMEM containing 20% FBS. The fibroblasts were incubated in an incubator for 24 h at 37°C. The non-invading fibroblasts blocked by the Matrigel were removed from the upper surface using a wet cotton swab. Following rinsing with PBS, the bottom surface of the filter was fixed with 95% ethanol at room temperature for 10 min and stained with 0.1% crystal violet at room temperature for 5 min. Invasion was detected by counting the stained cells under a light microscope (magnification, ×200; Olympus Corporation).

The fibroblasts and PH tissues or tissues WPH and were lysed in lysis buffer (RIPA; Beyotime Institute of Biotechnology) and centrifuged at 2,000 × g for 10 min at room temperature for supernatant recovery. The protein concentrations were determined using a BCA kit (Beyotime Institute of Biotechnology). Following separation on 10% SDS-PAGE, the proteins (20 μg/lane) were transferred onto nitrocellulose membranes (EMD Millipore) and blocked with 5% non-fat milk for 1 h at room temperature. Then the membranes were incubated at 4°C overnight with anti-MMP-2 (cat. no. 40994; 1:1,000; Cell Signaling Technology, Inc.), anti-MMP-9 (cat. no. sc-21733; 1:500; Santa Cruz Biotechnology, Inc.), anti-TIMP-1 (cat. no. sc-365905; 1:1,000; Santa Cruz Biotechnology, Inc.), anti-collagen I (cat. no. sc-293182; 1:1,000; Santa Cruz Biotechnology, Inc.), anti-collagen III (cat. no. sc-514601; 1:1,000; Santa Cruz Biotechnology, Inc.) and anti-β-actin (cat. no. 4970; 1:1,000; Cell Signaling Technology, Inc.) followed by incubation with horseradish peroxidase (HRP)-conjugated goat anti-rabbit IgG (cat. no. 10285-1-AP; 1:2,000, ProteinTech Group, Inc.) and HRP-goat anti-mouse IgG H&L (cat. no. ab205719; 1:2,000; Abcam) secondary antibodies at room temperature for 1 h. The proteins were visualized using an ECL kit (Amersham; GE Healthcare). The density of the blots was measured using the Quantity One software version 2.4 (Bio-Rad Laboratories, Inc.).

Differences between groups for categorical data were analyzed by the χ² test and Fisher's exact test and continuous data were analyzed using a Student's t-test. The cumulative incidence rate of PH was calculated by the Kaplan-Meier estimate analysis with log-rank test. Correlations among the risk factors were assessed using logistic regression analysis with a 95% confidence interval (CI) for the inclusion of additional prediction into the model. The clinical features were used as independent variables, while the presence or absence of PH was treated as a dependent variable. The differences between multiple groups were assessed with one-way analysis of variance and Tukey's post hoc test. P<0.05 was considered to indicate a statistically significant difference. SPSS 16.0 software (SPSS Inc.) was used for the analysis and the results were represented as mean ± standard deviation.

A total of 157 patients, including 55 patients with PH and 102 patients without PH, were reviewed. Among them, several parameters (age, history of diabetes, hypertension, radiation history, length of hospital stay and α1 (I) procollagen expression) exhibited no marked association with PH occurrence. However, the sex, BMI, aperture size and α1 (III) procollagen expression levels in the patients with PH were closely associated with the PH (Table I). The hernias occurred at between 7-57 months post-surgery, with an average of 23 months.

Among the 55 pairs of tissues, no marked difference in the α1 (I) procollagen expression was observed between the dermal tissues of patients with PH and that in tissues from patients without PH (Fig. 1A). By contrast, the α1 (III) procollagen level in dermal tissues of patients with PH was markedly increased compared with those in the tissues from patients without PH (P<0.05; Fig. 1B). The protein levels of collagen I were irregular while the collagen III levels were notably increased in the dermal tissues of patients with PH (Fig. 1C). Furthermore, the ratio of collagen I/III was clearly decreased in the dermal tissues of patients with PH (Fig. 1D).

According to the Kaplan-Meier analysis, the 5-year cumulative incidence of PH was calculated based on the clinical parameters. The incidence rate in female patients (50.5%) was significantly increased compared with that of male patients (25.4%; P<0.001; Fig. 2A). Patients with increased α1 (III) procollagen expression (68.9%) exhibited a higher incidence rate compared with those with decreased α1 (III) procollagen levels (26.7%) (P<0.001; Fig. 2B). The 5-year cumulative incidence rate of patients with a BMI >24.7 kg/m² was 53.1%, which was remarkably increased compared with that of patients with a BMI <24.7 kg/m² (25.4%) (P=0.005; Fig. 2C). In addition, PH incidence rate of patients with an aperture size >3.3 cm (41.3%) had no significant difference compared with that of patients with an aperture size <3.3 cm (31.5%) (P=0.119; Fig. 2D).

In the logistic analysis, female sex [odds ratio, 2.59; 95% confidence interval (CI), 1.107-6.059; P=0.028], aperture size (Odds ratio, 2.247; 95% CI, 1.280-3.946; P=0.005) and α1 (III) procollagen (Odds ratio, 0.109; 95% CI, 0.045-0.265; P<0.001) were independent risk factors for PH formation, however, BMI was not associated with PH incidence (Table II).

Positive vimentin expression was observed in the fibroblasts derived from tissues from patients with PH and without PH. The cells exhibited long shuttle or flat star shapes (Fig. 3A and B).

The wound closure of the fibroblasts from tissues from patients without PH (control) reached ~80% by 24 h, while only ~50% of the area was denuded by the fibroblasts from PH dermal tissues (Fig. 4A). The invasion rate in the normal fibroblast group was evidently enhanced, while it was notably inhibited in the hernia fibroblast group (Fig. 4B). The data demonstrated that the migration and invasion rates of fibroblasts were significantly suppressed in patients with PH (P<0.05; Fig. 4C and D). In addition, the fibroblast viability was also decreased in patients with PH, compared with patients without PH (Fig. 4E).

To explore the potential effects of collagen synthesis and the ECM on the PH, the protein levels of collagen I, collagen III, MMP-2, MMP-9 and TIMP-1 were detected (Fig. 5A). The collagen I expression levels in the fibroblasts from tissues from patients without PH were not significantly different from those in the fibroblasts from PH dermal tissues. However, among the fibroblasts from PH dermal tissues, the levels of collagen III and TIMP-1 were markedly increased while those of MMP-2 and MMP-9 were evidently inhibited in comparison with the non-PH fibroblasts (P<0.05; Fig. 5B).

For the fibroblasts from non-PH tissues, the cell-free area in the control-NC group at 24 h was larger compared with that in control-siTIMP-1 group. Similarly, among the fibroblasts from PH dermal tissues, the cell-free area in the hernia-NC group at 24 h was larger compared with that in the hernia-siTIMP-1 group (Fig. 6A). Silencing TIMP-1 notably increased the mean number of normal skin fibroblasts migrating into the bottom chamber. In addition, the number of fibroblasts from PH dermal tissues was significantly increased by silencing TIMP-1, compared with that in hernia-NC group (Fig. 6B). It was also identified that silencing TIMP-1 markedly enhanced the migratory and invasive abilities of fibroblasts (P<0.05; Fig. 6C and D).

Furthermore, the protein levels of the ECM-associated molecules were detected (Fig. 7A). The TIMP-1 expression was significantly suppressed in the groups transfected with siTIMP-1, indicating a successful transfection. It was identified that TIMP-1 silencing markedly inhibited the collagen III expression levels. In addition, the results also suggested that the levels of MMP-2 and MMP-9 were evidently increased by silencing TIMP-1, which indirectly indicated the negative correlation between MMP-2 or MMP-9 and collagen III (P<0.05; Fig. 7B).