Tel.: +48-61-8546 577 Fax: +48-61-8546-573
Cancer Cell Biology
Oxidative stress-dependent increase in ICAM-1 expression promotes adhesion of colorectal and pancreatic cancers to the senescent peritoneal mesothelium
Article first published online: 10 NOV 2009
Copyright © 2009 UICC
International Journal of Cancer
Volume 127, Issue 2, pages 293–303, 15 July 2010
How to Cite
Ksia̧żek, K., Mikuła-Pietrasik, J., Catar, R., Dworacki, G., Winckiewicz, M., Frydrychowicz, M., Dragun, D., Staniszewski, R., Jörres, A. and Witowski, J. (2010), Oxidative stress-dependent increase in ICAM-1 expression promotes adhesion of colorectal and pancreatic cancers to the senescent peritoneal mesothelium. Int. J. Cancer, 127: 293–303. doi: 10.1002/ijc.25036
- Issue published online: 20 MAY 2010
- Article first published online: 10 NOV 2009
- Accepted manuscript online: 10 NOV 2009 12:00AM EST
- Manuscript Accepted: 30 SEP 2009
- Manuscript Received: 26 MAR 2009
- Polish Ministry of Science and Higher Education. Grant Number: N401 094 31/2181
- cell senescence;
- mesothelial cells;
- oxidative stress
Intercellular adhesion molecule-1 (ICAM-1) has been implicated in adhesion of colorectal and pancreatic cancer cells (of the SW480 and PSN-1 line, respectively) to the peritoneal mesothelium. It has been demonstrated that ICAM-1 expression increases with senescence in some cell types, however, the significance of this phenomenon in the context of malignant dissemination remains elusive. In this report we show that the adherence of SW480 and PSN-1 cells to senescent human omentum-derived mesothelial cells (HOMCs) in vitro is greater than to early-passage cells and that the effect is mediated by ICAM-1. Senescent HOMCs display increased expression of ICAM-1 mRNA and cell surface protein. The development of this phenotype is related to increased oxidative stress in senescent cells. The augmented ICAM-1 expression in HOMCs can be reduced by culturing cells with antioxidants; in contrast, exposure of HOMCs to an oxidant, t-BHP, leads to cellular senescence and increased ICAM-1 expression. The effect is partly mediated by activation of p38 MAPK and AP-1 signaling pathways. Finally, culture of HOMCs in the presence of a strong antioxidant, PBN, significantly reduces the senescence-associated increase in SW480 and PSN-1 cancer cell binding. These results indicate that increased oxidative stress and increased expression of ICAM-1 in senescent HOMCs may facilitate peritoneal adhesion of selected colorectal and pancreatic cancers.
There is mounting evidence that the accumulation of senescent cells in tissues may contribute to the exponential rise of cancer incidence with age.1 It has been demonstrated that by secreting inflammatory cytokines and growth factors and by remodeling the extracellular matrix, senescent cells promote the growth and angiogenic activity of premalignant or malignant cells.2–7 We have recently observed that senescent mesothelial cells intensify the peritoneal adherence of ovarian cancer cells,8 which is thought to be a key step in dissemination of cancer cells and progression of the disease.9
Peritoneal carcinomatosis is generally viewed as heralding terminal disease. It is typically associated with cancer of the ovary, but occurs also in colorectal, gastric and pancreatic malignancies. Recent years have significantly increased our understanding of molecular mechanisms underlying the peritoneal spread of malignancies.10 It is now clear that adhesion of cancer cells to the peritoneal mesothelium is mediated largely by adhesion molecules whose pattern of expression and involvement differs between the cancer types. Dissemination of ovarian cancer cells appears to be primarily related to interactions between their β-integrins and mesothelial cell-derived extracellular matrix molecules.11–14
Other complexes postulated to participate in the process include hyaluronan-CD44,15 L1-neuropilin-1,16 CXCL12-CXCR417, 18 and MUC16-mesothelin.19 Peritoneal spreading of certain tumors has also been linked to mesothelial intercellular adhesion molecule-1 (ICAM-1).20 ICAM-1 (CD54) is a transmembrane glycoprotein expressed constitutively on many cell types, including peritoneal mesothelial cells.21, 22 Mesothelial expression of ICAM-1 may be further upregulated by proinflammatory cytokines and mediate polymorphonuclear leukocyte adhesion during peritonitis.22, 23 It has also been shown that ICAM-1 may support mesothelial adhesion of hepatoma, colorectal and pancreatic cancer cells.24–26 Interestingly, while leukocytes bind to mesothelial ICAM-1 through their integrin counter-receptors (CD11/CD18),22 the adherence of tumor cells is probably mediated by a different ligand of ICAM-1, sialophorin (CD43).26 The role of ICAM-1 in tumor progression is not restricted to the mesothelium, as interactions with ICAM-1 on vascular and lymphatic endothelial cells have been found to promote adhesion of small-cell lung carcinoma27 and breast adenocarcinoma.28 Furthermore, ICAM-1 may be expressed by cancer cells themselves and contribute to their invasiveness.29–31
We have earlier demonstrated that increased adherence of ovarian cancer to the senescent mesothelium is related to a rise in fibronectin production by senescent cells.8 In the present study we have set out to examine whether the cancer adhesion-promoting activity of senescent mesothelial cells occurs also toward other cancers known to metastasize to the peritoneal cavity and whether such an activity might be mediated by mesothelial ICAM-1. We have therefore chosen to investigate colorectal and pancreatic carcinomas whose adhesion to the mesothelium has been shown to be mediated primarily through ICAM-1.25 Here we demonstrate that the attachment of selected cancer cells to the senescent mesothelium is increased and that the effect is primarily mediated by increased expression of ICAM-1 on mesothelial cells. Furthermore, we identify senescence-associated oxidative stress and activation of p38 MAPK and AP-1 signaling pathways as a leading cause of augmented ICAM-1 expression.
Material and Methods
Unless otherwise stated, all chemicals were purchased from Sigma-Aldrich (St. Louis, MO). All tissue culture plastics were from Nunc (Roskilde, Denmark). Monoclonal anti-CD44 (Clone IM7) was from Pharmingen BD Biosciences (San Diego, CA). Monoclonal anti-ICAM-1 (Clone BBIG-I1), anti-β1 integrin (Clone P5D2), anti-CD43 (Clone 290111) and all appropriate isotype-matched IgG controls were obtained from R&D Systems (Wiesbaden, Germany). SB202190, a specific p38 MAPK inhibitor was from Calbiochem (San Diego, CA), MG-132 that neutralizes NF-κB by blocking IκBα degradation was from Biomol (Plymouth Meeting, PA), and 3-aminobenzamide (3-AB), an inhibitor of poly(ADP-ribosyl)-ation and of oxidant-induced c-fos expression and AP-1 binding activity, was purchased from Sigma.
Human omentum-derived peritoneal mesothelial cells (HOMCs) were isolated, as described in detail elsewhere.32 Specimens of the omentum were obtained from consenting patients undergoing elective abdominal surgery. The donors had no evidence of peritoneal inflammation and/or malignancy. Cells were propagated in medium M199 supplemented with L-glutamine (2 mM), penicillin (100 U ml−1), streptomycin (100 μg ml−1), hydrocortisone (0.4 μg ml−1), and 10% (v/v) fetal bovine serum (FBS). Colorectal cancer cell line SW480 was purchased from the American Type Culture Collection (Rockville, MD) and maintained in DMEM with L-glutamine (2 mM) and 10% FBS. Pancreatic cancer cell line PSN-1 was purchased from the European Collection of Cell Cultures (Porton Down, UK) and propagated in RPMI-1640 with 10% FBS.
Replicative HOMC senescence
HOMCs were serially passaged at 4–5-day intervals until their capacity to divide exhausted. Cells from passages 1–2 were treated as “early-passage” cells, while cultures that failed to increase in number during 4 weeks and stained >70% for senescence-associated β-galactosidase (SA-β-Gal) were considered as senescent.33 In some experiments HOMCs were cultured to senescence in the presence of a spin-trap reactive oxygen species (ROS) scavenger, N-tert-butyl-alpha-phenylnitrone (PBN), at the concentration of 800 μM.34
Oxidative stress-induced HOMC senescence
To induce senescence, HOMCs were repeatedly exposed to sub-lethal doses of hydrophobic organic oxidant, tert-butyl hydroperoxide (t-BHP).35 In brief, subconfluent early-passage cells were treated with 30 μM t-BHP for 1 hr daily for 7 days. Between the exposures cells were maintained in standard medium. After 7 days of such treatment cells were allowed to recover in standard growth medium for 5 days, during which they developed the senescence phenotype, as evidenced by hypertrophic morphology and extensive positive staining for SA-β-Gal.8 In selected experiments early-passage cells were treated acutely with 30 μM t-BHP for up to 2 hr.
Adherence of cancer cells to HOMCs was determined as described previously.8 Briefly, HOMCs were plated in flat-bottom 96-well plates (105 cells/well) and left to settle overnight. Cancer cells were detached by trypsinization, washed with phosphate buffered saline (PBS), and probed with 5 μM solution of calcein-AM (Molecular Probes, Invitrogen, Eugene, OR) for 30 min at 37°C. Calcein-labeled cells were then washed with medium to remove the free dye and added (4 × 104 cells/well) on top of mesothelial cells to adhere. Preliminary experiments determined that maximal adherence occurred within 30 and 45 min for PSN-1 cells and SW480 cells, respectively. These incubation times were used in all subsequent experiments. After incubation, total fluorescence in each well was recorded in a spectrofluorimeter (Perkin-Elmer, Turku, Finland) using 485 and 535 nm wavelengths for excitation and emission, respectively. The nonadherent cells were removed by gentle aspiration and washing, and the measurement of fluorescence was repeated to estimate the percentage of cells adhered. In selected experiments, cancer cells were incubated with anti-CD44, anti-β1 integrin or control antibodies for 30 min at 37°C prior to the adhesion assay. Alternatively, HOMCs were treated with either anti-ICAM-1 or control IgG for 30 min at 37°C prior to the addition of cancer cells.
Conditioned media (CM) were collected from both early-passage and senescent cells. Briefly, subconfluent HOMCs were washed thoroughly and then incubated in serum-free medium for 72 hrs. The samples of CM collected were centrifuged, filtered through a 0.2-μm pore size filter to remove any cellular debris, and then stored in aliquots at −80°C until required. The cells in culture flasks were detached and counted. Since the number of cells in early-passage cultures that gave rise to CM was greater than that in senescent cultures, appropriate volumes of serum-free medium were added to samples of CM to normalize for a difference in cell number.7 To exclude the possibility, that the added medium provided fresh portion of some components, it was preincubated in the culture flasks containing no cells in a manner corresponding exactly to that used for CM harvesting.
Expression of CD43 on cancer cells was assessed by flow cytometry. Briefly, freshly passaged SW480 and PSN-1 cells were transferred to M199 medium containing 10% FBS and cultured until ∼70% confluent. Then the cells were incubated in serum-free medium for 48 hrs and exposed to CM from early-passage and senescent HOMCs for the next 48 hrs. After the incubation, the cells were harvested, washed twice in cold PBS, divided into 105 cell aliquots, and incubated for 45 min on ice with either a monoclonal anti-human CD43 antibody or the isotype control antibody. After washing with cold PBS, cells were incubated for 45 min on ice with the anti-mouse fluorescein-conjugated secondary antibody. Cells were then washed and analyzed with the FACSCanto™ flow cytometer (Becton Dickinson, Franklin Lakes, NJ). The data obtained were analyzed with PC-Lysis software (Becton Dickinson).
Expression of mesothelial cell ICAM-1 was measured by cell-bound immunoassay. In brief, HOMCs in 96-well plates were washed with PBS, fixed with 4% formaldehyde (pH 7.4) and blocked with 1% bovine serum albumin in PBS for 2 hr at 4°C. Fixed cells were first incubated at 4°C with anti-human ICAM-1 (1:1,000) and then at 37°C with peroxidase-coupled goat anti-mouse IgG (1:5,000) for 1 hr. After gentle and extensive washing, cells were treated with a peroxidase substrate solution (Pharmingen, BD Biosciences) until the color developed. The reaction was terminated by the addition of 2 N H2SO4, and the absorbance was recorded at 450 nm. The results obtained were normalized for the number of cells that were counted in representative wells using the Bürker chamber.
Total RNA was extracted with the TRIzol® Reagent (Invitrogen, Carlsbad, CA), purified and reverse transcribed into cDNA with random hexamer primers, as previously described.36 Conventional PCR amplification was performed on the Mastercycler Gradient 5331 thermocycler (Eppendorf, Hamburg, Germany) using Platinium Taq DNA Polymerase (Invitrogen). Specific oligonucleotide primers were synthesized by TIB MolBiol SyntheseLabor (Berlin, Germany). Primer sequences were as follows: ICAM-1 forward: 5′-GTG ACATGCAGCACCTCCTG-3′, ICAM-1 reverse: 5′-TCCATG GTGATCTCTCCTCA-3′,37 and α-actin forward: 5′-GGAG CAATGATCTTGATCTT-3′; α-actin reverse: 5′-CCTTCCT GGGCATGGAGTCCT-3′.38 The reaction was carried out either for 24 cycles with annealing temperature of 52°C for ICAM-1 or for 30 cycles with annealing temperature of 55°C for α-actin. PCR products were separated by electrophoresis in ethidium bromide-stained 3% agarose.
Real-time PCR was performed in the Applied Biosystems 7500 Fast Real-Time PCR system (Applied Biosystems, Foster City, CA). The reaction was carried out in 20-μl reaction volumes containing 2 μl of cDNA, GAPDH and ICAM-1 primers (0.5 μM each) and 10 μl of Power SYBR Green PCR Master Mix (Applied Biosystems). Primer sequences were as follows: ICAM-1 (GenBank BC015969.2) forward (5′-GGCTGGAGCTGTTTGAGAAC-3′), reverse (5′-ACTGTGG GGTTCAACCTTCTG-3′); GAPDH (GenBank AB062273.1) forward (5′-CATCACCATCTTCCAGGAGCG-3′), reverse (5′-TGACCTTGCCCACAGCCTTG-3′). Reaction conditions included a 2-min initial activation step at 50°C and a 10-min denaturing step at 95°C, followed by 40 cycles of 15 sec at 95°C, and 1 min at 60°C. Specificity of the reaction was verified by performing the melting curve analysis at the end of each series of assays. The relative amount of ICAM-1 transcript was calculated by the cycle threshold method using the Applied Biosystems 7500 System v.1.2.3 software and normalized for the endogenous reference (GAPDH).
Measurement of ROS
ROS production was assessed using 2′,7′-dichlorodihydrofluorescein diacetate (H2DCFDA) probe. One hundred thousand cells were incubated in the presence of 5 μM H2DCFDA (Molecular Probes, Invitrogen) for 45 min at 37°C. The fluorescence intensity in cell lysates was monitored in a spectrofluorymeter (Perkin-Elmer, Turku, Finland) with excitation at 485 nm and emission at 535 nm. The results were expressed as relative light units (RLU) per 105 cells.
Measurement of NF-κB, AP-1 and p38 MAPK activity
Activation of the transcription factors NF-κB (p50 subunit) and AP-1 (c-Jun subunit) was examined using TransAMTM Kits (Active Motif, Rixensart, Belgium). Nuclear extracts (2.5 μg per sample) were prepared and incubated with immobilized NF-κB or AP-1 consensus binding sites, according to manufacturer's instructions. The presence of bound complexes was detected with the use of specific antibodies and quantified by colorimetric reaction. Competitive studies with wild type and mutated oligonucleotides were performed to confirm specificity of the reactions.
The level of activated (phosphorylated) and total p38 Mitogen Activated Protein Kinase (MAPK) was quantified using fast activated cell-based ELISA (FACETM) p38 Kit (Active Motif) as per manufacturer's instructions. The results were expressed as a ratio of phosphorylated to total p38 MAPK activity.
To assess the role of selected pathways in regulation of ICAM-1 expression, HOMCs were pretreated with either SB202190, MG-132 or 3-AB that act as inhibitors of signaling through p38 MAPK, NF-κB and AP-1, respectively. The doses of inhibitors were chosen from the literature39–41 and pilot experiments ensured they did not affect cell viability, as measured by trypan blue exclusion test.
Statistical analysis was performed using GraphPad Prism™ 4.00 software (GraphPad Software, San Diego, USA). The data were compared with the paired t-test or repeated measures analysis of variance, as appropriate. Results are expressed as means ± SD. The indicated number of experiments (n) refers to the experiments performed with HOMCs from different donors. Differences with a p-value <0.05 were considered significant.
Senescent HOMCs promote adhesion of SW480 and PSN-1 cells
Under the assay conditions the adhesion of SW480 and PSN-1 cells to early-passage HOMCs was 46% ± 4% and 53% ± 6%, respectively, of the total number of cells seeded. The adherence of both cell types to senescent HOMCs was marginally but significantly and consistently increased (Fig. 1). When expressed as a percentage of binding to early-passage cells, the adhesion of SW480 and PSN-1 cells to senescent HOMCs was greater by 16% ± 11% and 27% ± 13%, respectively.
Adhesion of SW480 and PSN-1 cells to senescent HOMCs is largely mediated by ICAM-1
Since β1 integrins and CD44 are thought to be central to cancer cell adhesion to the mesothelium,13, 15 we began by comparing the contribution of these pathways to that mediated by ICAM-1. To this end the adhesion assays were performed after neutralization of either β1 integrins and CD44 on cancer cells or ICAM-1 on HOMCs. Preincubation of SW480 cells with an anti-β1 integrin antibody had no effect on their adhesion to HOMCs (Fig. 2a). Anti-CD44 antibodies could reduce SW480 adhesion by up to 30% (Fig. 2b). The application of a specific anti-ICAM-1 antibody decreased SW480 adhesion in a dose-dependent manner so that at the highest dose of the antibody tested, the magnitude of inhibition was 46% ± 5% (p < 0.001) (Fig. 2c).
The adhesion of PSN-1 cells to HOMCs was reduced maximally by 19% ± 6%, 30% ± 6% and 41% ± 7% with anti-β1 integrin, anti-CD44 and anti-ICAM-1, respectively (Figs. 2d–2f). In all above experiments the specificity of the inhibition observed was confirmed by the lack of such an effect in response to isotype-matched IgG controls.
We then went on to determine whether the contribution of the 3 pathways to cancer cell adhesion was similar in early-passage and senescent HOMCs. In comparison with early-passage cells, the β1 integrin-mediated adherence to senescent HOMCs remained fairly insignificant for SW480 cells and increased only marginally for PSN-1 cells (Fig. 3a). The relative contribution of the CD44-mediated pathway did not differ between early-passage and senescent HOMCs for both cancer cell types (Fig. 3b). In contrast, the analysis of the inhibition exerted by anti-ICAM-1 antibodies in early-passage and senescent HOMCs revealed that the ICAM-1-dependent pathway contributed significantly more to binding of SW480 and PSN-1 cells to the senescent mesothelium than to young HOMCs (Fig. 3c).
Taken together, these results pointed to a predominant role of ICAM-1 in mediating SW480 and PSN-1 cell adhesion to the senescent mesothelium.
CD43 expression on SW480 and PSN-1 cells is only slightly affected by senescent HOMCs
To examine whether senescent HOMCs could have modulated the adherence of cancer cells by altering their expression of an ICAM-1 ligand, CD43, SW480 and PSN-1 cells were exposed to conditioned media from HOMCs and analyzed by flow cytometry. These experiments showed that, in comparison with cancer cells incubated with CM from early passage HOMCs, the exposure to CM from senescent HOMCs resulted in a slightly increased expression of CD43 on cancer cells. This increase, however, was too small and inconsistent to reach statistical significance. Mean fluorescence intensities (MFI) corresponding to CD43 expression on SW480 cells were 772 ± 153 and 931 ± 113 for cells treated with CM from early-passage and senescent HOMCs, respectively (n = 3). For PSN-1 cells the respective MFI values were 1,354 ± 145 and 1,584 ± 101 (n = 3).
Expression of ICAM-1 increases during HPMC senescence in vitro
Since changes in CD43 expression did not seem to convincingly account for increased adhesion of cancer cells, we focused on the expression of ICAM-1 on HOMCs. As assessed by cell-bound immunoassay, the expression of cell surface ICAM-1 on senescent HOMCs was ∼3-fold greater compared with early-passage cells (Fig. 4a). This effect was accompanied by a rise in the expression of ICAM-1 mRNA (Fig. 4b). When analyzed with quantitative PCR, this increase was found to be more than 23-fold above the expression level in early-passage HOMCs (Fig. 4c).
Oxidative stress contributes to increased ICAM-1 expression in senescent HOMCs
Since oxidative stress may trigger ICAM-1 expression40 and senescent HOMCs generate more ROS,42 we wanted to determine whether ROS might be responsible for an increase in ICAM-1 expression in senescent HOMCs. To this end, cells were cultured to senescence in the presence or absence of a ROS scavenger, PBN. Exposure to PBN led to a decrease (by 34% ± 8%) in senescence-associated ROS production (Fig. 5a). This effect was accompanied by a reduction in cell surface and mRNA ICAM-1 expression almost to the levels seen in early-passage cells (Figs. 5b and 5c).
To show directly that ROS induce ICAM-1 in senescent HOMCs, cellular senescence was induced by a potent oxidant, t-BHP.35 HOMCs treated with t-BHP displayed significantly increased expression of both ICAM-1 mRNA and ICAM-1 cell surface protein. When exposure to t-BHP was followed by the treatment with PBN, the elevated ICAM-1 expression was reduced to control levels (Figs. 5d and 5e).
Oxidative stress stimulates ICAM-1 expression through p38 MAPK and AP-1 pathways
To identify signaling pathways involved in ROS-induced ICAM-1 overexpression in senescent HOMCs, the activation of selected cellular mediators was assessed following an exposure to t-BHP. The treatment with t-BHP resulted in activation of p38 MAPK in a time-dependent manner. Increased phosphorylation of p38 MAPK was detected 1 hr after stimulation, peaked at 2 hrs, and remained elevated in cells that became senescent as a result of repeated exposure to t-BHP for 7 days (Fig. 6a). Under equal conditions nuclear binding of c-Jun increased rapidly within 1 hr but then decreased and in cells treated repeatedly with t-BHP it was not different from that in control cells (Fig. 6b). In contrast, neither acute nor repeated exposure to t-BHP affected nuclear binding of the NF-κB p50 subunit (Fig. 6c).
When acute exposure to t-BHP was preceded by a preincubation with either SB202190 or 3-AB, that inhibit p38 MAPK and AP-1, respectively, the expression of cell surface ICAM-1 and ICAM-1 mRNA was significantly reduced. On the other hand, preexposure of HOMCs to MG-132 that inhibits NF-κB activation did not have a significant effect on ICAM-1 protein and mRNA expression (Figs. 6d and 6e).
Reduction of oxidative stress decreases adhesion of SW480 and PSN-1 cells to senescent HOMCs
As the treatment with PBN ameliorated oxidative stress-induced increase in the expression of ICAM-1 in senescent HOMCs, we wanted to see whether it might also decrease the adhesion of cancer cells. The adherence of either SW480 or PSN-1 cells to HOMCs grown to senescence in the presence of PBN was significantly reduced and approximated the level in early-passage HOMCs (Fig. 7).
Cellular senescence is thought to be an antagonistically pleiotropic mechanism that protects young organisms against cancer, but exerts deleterious effects later in life.43 These effects may include loss of tissue integrity and function with age, which may eventually promote late-life cancers. We have recently shown that increased production of fibronectin is that feature of the senescent phenotype that mediates increased adhesion of ovarian cancer cells to the senescent mesothelium.8 The present study demonstrates that this is not the only pathway whereby senescent HOMCs might contribute to peritoneal carcinomatosis.
We found that adherence of PSN-1 and SW480 cancer cells to senescent HOMCs was greater than to their younger counterparts. These cancer cell lines have been reported to adhere to the mesothelium largely through an ICAM-1-dependent mechanism.25 Indeed, we could confirm that a specific ICAM-1-blocking antibody significantly reduced the attachment of either cancers to HOMCs. Furthermore, the comparison of inhibitory effects of the antibody in young and senescent HOMCs showed that the contribution of ICAM-1-mediated adhesion was significantly greater in senescent HOMCs. In contrast, the adherence of SW480 and PSN-1 cancers mediated by β1 integrins and CD44 was less pronounced and did not increase significantly with HOMC senescence.
The involvement of the ICAM-1 pathway could be related to a minor extent to altered expression of CD43 on cancer cells. It has been demonstrated that the blockade of this ICAM-1 ligand with anti-CD43 antibodies reduced adhesion of selected cancers to the mesothelium.26 We observed that CM from senescent HOMCs slightly increased CD43 expression on SW480 and PSN-1 cells. One may hypothesize that the effect was mediated by increased release of as yet unidentified factors by senescent HOMCs. In this respect, it has been demonstrated that the expression of CD43 on human leukemic and skin mast cells could be modulated by retinoic acid.44 The additional difficulty in an interpretation of our experiments could be that CM media from early-passage and senescent HOMCs were generated by different numbers of cells. Although the volume of CM was normalized for cell number, as described by others,7 one cannot be sure that the release of hypothetical CD43 modulators into CM occurred linearly with HOMCs senescence. Therefore the mechanisms underlying the potential modulation of CD43 expression by senescent HOMCs require further studies.
As the role of CD43 could be regarded as uncertain, we looked into the mechanism by which cellular senescence could change the expression of ICAM-1 in HOMCs. Senescent HOMCs were found to have increased ICAM-1 expression at both mRNA and protein levels. This appears to be a more general feature of cellular senescence as a similar effect was observed in senescent fibroblasts,45 endothelial cells46 and smooth muscle cells.47
Since oxidative stress may strongly impinge on ICAM-1 expression48 and senescence of HOMCs is characterized by the dysfunction of mitochondria and increased generation of ROS,42 we considered oxidative stress as a likely cause of increased ICAM-1 expression. Indeed, we found that the attenuation of oxidative stress by ROS scavenger, PBN, resulted in a significantly decreased ICAM-1 expression. In contrast, the expression of ICAM-1 increased markedly when HOMCs were exposed to a strong oxidant, t-BHP, that was also shown to induce senescence of HOMCs.49 These findings are in keeping with earlier studies on endothelial cells and keratinocytes, which demonstrated an increase in ICAM-1 expression in response to hydrogen peroxide40, 50, 51 and a decrease in ICAM-1 expression following the treatment with antioxidants.50, 52
Expression of ICAM-1 is regulated in a complex, cell type- and stimulus-specific manner with ROS involved at different levels.48 As signaling through the NF-κB pathway is frequently associated with oxidative stress, we have assessed nuclear binding of the NF-κB p50 subunit following the exposure of HOMCs to t-BHP. Surprisingly, we found no activation of NF-κB in response to such a regimen. Accordingly, we observed no effect of the NF-κB inhibitor, MG-132, on ICAM-1 expression. These data indicate that the upregulation of ICAM-1 in HOMCs rendered senescent by exposure to oxidants is independent of NF-κB. A similar lack of NF-κB involvement in senescence-associated ICAM-1 overexpression was reported in endothelial cells40 and fibroblasts.45
On the other hand, exposure to t-BHP activated both p38 MAPK and c-Jun/AP-1, which are known to act as down-stream mediators of ROS and regulate ICAM-1 expression in other cell types.48 Indeed, the blockade of either p38 MAPK or AP-1 led to a marked decrease in ICAM-1 expression in t-BHP-treated HOMCs. The ICAM-1 promoter contains several binding sites for AP-1,53 whose activity can be regulated by p38 MAPK.54 Our results confirm the importance of p38 MAPK signaling for the induction of ICAM-1, an effect observed earlier in endothelial cells.55 However, the precise mechanism of this regulation in senescent HOMCs remains to be determined. Interestingly, oxidative stress led to sustained activation of p38 MAPK, but only a transient increase in c-Jun/AP-1 activity. Yet, the inhibition of the AP-1 pathway diminished ICAM-1 expression to control levels. Regulation of AP-1 activity occurs at many levels including transcription of genes encoding AP-1 subunits, stabilization of the mRNA transcribed, post-translational processing and interactions with other transcription factors.56 In this respect, it has been demonstrated that Ets transcription factors cooperate with AP-1 to form the oxidative stress responsive element of the ICAM-1 promoter.40 Moreover, it has been reported that p53, a key mediator of replicative senescence, may directly contribute to increased ICAM-1 expression in senescent fibroblasts.45 We have recently demonstrated that senescence of HOMCs is accompanied by a rise in the expression of p21 protein, a major down-stream target of p53.8
Having observed that ICAM-1 overexpression in senescent HOMCs was mediated by oxidative stress, we applied a potent antioxidant PBN to assess whether it would decrease the ICAM-1-mediated adhesion of cancer cells. We found that following such a treatment the generation of ROS by senescent HOMCs declined and that the increased adherence of pancreatic and colorectal cancer cells decreased to a level observed in early-passage HOMCs. These results suggest that alleviation of oxidative stress could be a potential therapeutic avenue to reduce the peritoneal dissemination of cancer cells, at least those attaching via mesothelial cell ICAM-1. In fact, malignant environment is known to produce considerable oxidative stress57 that may play a significant role in the development of both primary and metastatic tumors.58 In this respect it has been demonstrated that serum levels of ROS in patients with colorectal cancers correlated with size of the tumor.59 Moreover, it has been shown that by reducing the activity of mitogen-activated protein kinase phosphatase-3, ROS may increase tumorigenicity and chemoresistance of ovarian cancer cells in mice.60 Furthermore, it has been demonstrated that intraperitoneal administration of a catalase conjugate in mice decreased adhesion of melanoma cells to the peritoneum, which coincided with reduced ICAM-1 expression in the omentum.61
The pattern of intraperitoneal dissemination of cancer cells depends on the anatomic location of the primary tumor, its type of growth, the site-specific profile of gene expression62 and established pathways of ascitic flow.63 Results of the current study indicate that the accumulation of senescent HOMCs in the peritoneal cavity may favor the adherence of selected cancers to the peritoneum and contribute to their peritoneal spread.
The authors declare no competing financial interests.
- 52Ionizing radiation induces, via generation of reactive oxygen intermediates, intercellular adhesion molecule-1 (ICAM-1) gene transcription and NF kappa B-like binding activity in the ICAM-1 transcriptional regulatory region. Free Radic Res 1997; 27: 127–42., , , , , , .