Repeated H2 O2 exposure drives cell cycle progression in an in vitro model of ulcerative colitis

Abstract

The production of hydrogen peroxide (H2 O2 ) drives tumourigenesis in ulcerative colitis (UC). Recently, we showed that H2 O2 activates DNA damage checkpoints in human colonic epithelial cells (HCEC) through c-Jun N-terminal Kinases (JNK) that induces p21(WAF1) . Moreover, caspases circumvented the G1/S and intra-S checkpoints, and cells accumulated in G2/M. The latter observation raised the question of whether repeated H2 O2 exposures alter JNK activation, thereby promoting a direct passage of cells from G2/M arrest to driven cell cycle progression. Here, we report that increased proliferation of repeatedly H2 O2 -exposed HCEC cells (C-cell cultures) was associated with (i) increased phospho-p46 JNK, (ii) decreased total JNK and phospho-p54 JNK and (iii) p21(WAF1) down-regulation. Altered JNK activation and p21(WAF1) down-regulation were accompanied by defects in maintaining G2/M and mitotic spindle checkpoints through adaptation, as well as by apoptosis resistance following H2 O2 exposure. This may cause increased proliferation of C-cell cultures, a defining initiating feature in the inflammation-carcinoma pathway in UC. We further suggest that dysregulated JNK activation is attributed to a non-apoptotic function of caspases, causing checkpoint adaptation in C-cell cultures. Additionally, loss of cell-contact inhibition and the overcoming of senescence, hallmarks of cancer, contributed to increased proliferation. Furthermore, there was evidence that p54 JNK inactivation is responsible for loss of cell-contact inhibition. We present a cellular model of UC and suggest a sinusoidal pattern of proliferation, which is triggered by H2 O2 -induced reactive oxygen species generation, involving an interplay between JNK activation/inactivation, p21(WAF1) , c-Fos, c-Jun/phospho-c-Jun, ATF2/phospho-ATF2, β-catenin/TCF4-signalling, c-Myc, CDK6 and Cyclin D2, leading to driven cell cycle progression.

Keywords: DNA damage checkpoint adaptation; cell cycle progression; hydrogen peroxide; ulcerative colitis.

Figure 1

An in vitro model of ulcerative colitis showing loss of cell-contact inhibition, increased proliferation, and overcoming of senescence. (A) Study designed to mimic acute and chronic inflammation via ROS using H₂O₂. The ROS exposure in acute inflammation was mimicked by single H₂O₂ treatment. As chronic UC is characterized by damage-regeneration periods, chronic inflammation was simulated by repetitive injury, exposing human colonic epithelial cells (HCEC) to repeated H₂O₂ treatment cycles (C)1-C10 with recovery phases in between. In this way, 10 H₂O₂-exposed cell cultures were generated and named C-cell cultures C1-C10. C1-C3 cells were generated in the first study and C4-C10 cells in this study. (B) Loss of cell-contact inhibition occurred in C3 cells and continued until C10 cells. Phase contrast micrographs are shown after 5 days of recovery, and arrows indicate loss of cell-contact inhibition, piling up, and thus foci formation. (C) Increased proliferation of C-cell cultures. C1-C10 cells and HCEC cells were cultivated, and cell numbers were counted after 7 days. Data indicate mean ± SD and were obtained from four individual measurements. (D) Cells were grown for 48 hrs, fixed and subsequently stained for ß-galactosidase activity (blue areas).

Figure 2

Altered JNK activation is associated with p21^WAF1 down-regulation in C-cell cultures. (A) Lysates from C-cell cultures and human colonic epithelial cells (HCEC) were immunoblotted with anti-JNK, -phospho-JNK, -p21^WAF1 and -β-actin antibodies. β-actin served as loading control, and fold expression relative to HCEC cells is given below the blots. P21^WAF1 immunoblotting of HCEC and C1-C3 cells is published in . (B) Immunohistochemical analysis of p21^WAF1 in normal colonic mucosa, in active UC and in UC in complete remission. Arrows indicate proliferative cells with marginal or no expression.

Figure 3

JNK inactivation and p21^WAF1 down-regulation as pathogenetic factors. (A) Phase contrast micrographs of human colonic epithelial cells (HCEC) and C3 cells 48 hrs after treatment with DMSO or 50 μM SP600125. Arrows indicate loss of cell-contact inhibition, piling up, and foci formation. (B) Lysates from HCEC and C3 cells 24 hrs after treatment with DMSO or 50 μM SP600125 were immunoblotted with anti-p21^WAF1. β-actin served as loading control, and fold expression relative to HCEC cells is given below the blots. (C) Cell numbers of HCEC cells treated with DMSO (▴), SP600125 (▪), and without treatment (•) after 2, 3, 4 and 7 days are shown. The data represent mean ± SD of four individual measurements. (D) Lysates from C3 cells 72 hrs after treatment with DMSO or 50 μM of the pan-caspase-inhibitor Z-VAD-FMK were immunoblotted with anti-phospho-JNK. β-actin served as loading control, and fold expression relative to HCEC cells is given below the blots.

Figure 4

H₂O₂-induced reactive oxygen species (ROS) generation in human colonic epithelial cells (HCEC). The ROS generation by HCEC, C5 and C10 cells, measured either with DCFH or Amplex Red, is shown. (A) Intracellular ROS generation by HCEC cells (•) and by HCEC cells treated with 200 μM H₂O₂ (•). The arrow indicates H₂O₂-induced ROS generation (oxidative stress). The data represent mean ± SD of twelve individual measurements. (B) Time course of the concentration of H₂O₂ in PBS medium without (•) and with HCEC cells (10 μg protein of ml; ▴). For comparison, the release of H₂O₂ into the medium is also shown (•). The data represent mean ± SD of three individual measurements. (C) X-fold increase in intracellular ROS relative to HCEC cells is shown for C5 (•) and C10 (□) cells. The data represent mean ± SD of 12 individual measurements. (D) Time course of the release of H₂O₂ from HCEC (•), C5 (•) and C10 cells (▪). The data represent mean ▾ SD of 12 individual measurements.

Figure 5

Involvement of oncogenic transcription factors in the H₂O₂-associated colitis model. (A) Lysates from C1-C10 cells and human colonic epithelial cells (HCEC) were immunoblotted with anti-c-Fos, -c-Jun, -phospho-c-Jun, -ATF2, -phospho-ATF2 and -β-actin antibodies. β-actin served as loading control, and fold expression relative to HCEC cells is given below the blots. c-Fos immunoblotting of HCEC and C1-C3 cells is published in . (B) Lysates from C1-C10 cells and HCEC cells were immunoblotted with anti-phospho-p38, -phospho-ERK1/2, -c-Myc, -β-catenin, -TCF4, -Sp1, -STAT3, -phospho-STAT3, and -β-actin antibodies. β-actin served as loading control, and fold expression relative to HCEC cells is given below the blots. c-Myc and β-catenin immunoblotting of HCEC and C1-C3 cells is published in .

Figure 6

Driven cell cycle progression, associated with overexpression of CDK6 and Cyclin D2, is caused by checkpoint adaptation and apoptosis resistance. (A) Lysates from C1-C10 cells and human colonic epithelial cells (HCEC) were immunoblotted with anti-CDK4, -CDK6, -Cyclin D2, -CDK2, -Cyclin E, -CDK1, -Cyclin B1, and -β-actin antibodies. β-actin served as loading control, and fold expression relative to HCEC cells is given below the blots. (B) RNA from C1-C10 cells and HCEC cells was transcribed into cDNA, and real-time PCR was conducted for p21^WAF1 mRNA and β2-Microglobulin mRNA expression. (C) H₂O₂-treated C10 cells showed S and G2/M arrest after 24 hrs and increased G1 cell population but no apoptosis induction (Pre-G1) after 48 and 72 hrs. The data are representative of three independent experiments.

Figure 7

Proposed molecular mechanisms underlying driven cell cycle progression in the in vitro model of ulcerative colitis. (A) Increased proliferation showed a sinusoidal pattern consisting of maxima 1 and 2. Permanent molecular events were displayed by an arrow above the proliferation curve. (B) We suggest a model involving an interplay of selective JNK inactivation and p21^WAF1 down-regulation, selective JNK activation, AP-1 components and β-catenin/TCF4-signalling. Two molecular events may finally lead to p21^WAF1 down-regulation: selective JNK inactivation and β-catenin/TCF4-dependent suppression of p21^WAF1.