Dilek Güner

Adenovirus-mediated overexpression of p14(ARF) induces p53 and Bax-independent apoptosis.

The human INK4a gene locus encodes two structurally unrelated tumor suppressor proteins, p16INK4a and p14ARF, which are frequently inactivated in human cancer. Whereas p16INK4a acts through engagement of the Rb-cdk4/6-cyclin D pathway, both the pro-apoptotic and cell cycle-regulatory functions of p14ARF were shown to be primarily dependent on the presence of functional p53. Recent reports have also implicated p14ARF in p53-independent mechanisms of cell cycle regulation and apoptosis induction, respectively. To further explore the pro-apoptotic function of p14ARF in relation to functional cellular p53, we constructed a replication-deficient adenoviral vector for overexpression of p14ARF (Ad-p14ARF). As expected, Ad-p14ARF efficiently induced apoptosis in p53/Rb wild-type U-2OS osteosarcoma cells at low multiplicities of infection. Interestingly, Ad-p14ARF also induced apoptosis in both p53-deleted SAOS-2 osteosarcoma cells and HCT116 colon cancer cells with a bi-allelic knock-out of p53 (HCT116-p53−/−). Similarly, adenovirus-mediated overexpression of p14ARF induced apoptosis in p53/Bax-mutated DU145 prostate cancer cells as well as in HCT116 cells devoid of functional Bax (HCT116-Bax−/−). Restoration of Bax expression by retroviral gene transfer in DU145 cells did not further enhance p14ARF-triggered cell death. Infection with Ad-p14ARF induced activation of mitochondrial permeability shift transition, caspase activation and apoptotic DNA fragmentation irrespective of the presence or absence of either Bax or functional cellular p53. Nevertheless, overexpression of the anti-apoptotic Bcl-2 homolog Bcl-xL markedly inhibited p14ARF-induced apoptosis. This may indicate that p14ARF triggers a so far unknown activator of mitochondrial apoptosis which can be inhibited by Bcl-2 but which acts either independently or downstream of Bax. Taken together, this report demonstrates the participation of signaling pathways apart from the p53/Mdm-2 rheostat and Bax in p14ARF-mediated apoptosis.

Multidomain Bcl-2 homolog Bax but not Bak mediates synergistic induction of apoptosis by TRAIL and 5-FU through the mitochondrial apoptosis pathway

The death ligand TRAIL synergizes with DNA-damaging therapies such as chemotherapeutic drugs or ionizing irradiation. Here, we show that the synergism of TRAIL and 5-fluorouracil (5-FU) and cross-sensitization between TRAIL and 5-FU for induction of apoptosis, entirely depend on Bax proficiency in human DU145 and HCT116 carcinoma cells. DU145 prostate carcinoma cells that have lost Bax protein expression due to mutation fail to release cytochrome c and to activate caspase-3 and -9 when exposed to TRAIL and 5-FU. In contrast, TRAIL sensitized for 5-FU-induced apoptosis and vice versa upon reconstitution of Bax expression. Isobolographic analyses of ED50 doses for 5-FU at increasing TRAIL concentrations showed a clear synergism of TRAIL and 5-FU in Bax-expressing cells. In contrast, the effect was merely additive in DU145 cells lacking Bax. Notably, both DU145 and HCT116 Bax-deficient cells still express Bak. This indicates that Bak is not sufficient to mediate cross-sensitization and synergism between 5-FU and TRAIL. Stable overexpression of Bak in DU145 sensitized for epirubicin-induced apoptosis but failed to confer synergy between TRAIL and 5-FU. Moreover, we show by the use of EGFP-tagged Bax and Bak that TRAIL and 5-FU synergistically trigger oligomerization and clustering of Bax but not Bak. These data clearly establish distinct roles for Bax and Bak in linking the TRAIL death receptor pathway to the mitochondrial apoptosis signaling cascade and delineate a higher degree of specificity in signaling for cell death by multidomain Bcl-2 homologs.

Induction of p21CIP/WAF-1 and G2 arrest by ionizing irradiation impedes caspase-3-mediated apoptosis in human carcinoma cells

There is an ongoing controversy regarding the relevance of apoptosis induction by ionizing irradiation as compared with other end points including transient or permanent cell cycle arrest of damaged cells. Here, we show that such permanent cell cycle arrest and apoptosis represent two sides of the same coin. MCF-7 cells fail to express procaspase-3, which results in resistance to apoptosis induced by anticancer drugs. Conversely, restoration of procaspase-3 sensitizes MCF-7 cells to chemotherapeutics including epirubicine, etoposide and taxol. In contrast, irradiation does not trigger apoptotic cell death but results in prolonged arrest in the G2 phase of the cell division cycle regardless of procaspase-3 expression. This suggested that the propensity of MCF-7 cells to arrest at the G2 checkpoint results in resistance to apoptosis upon γ-irradiation. This G2 arrest was associated with upregulation of p21CIP/WAF-1. Inhibition of DNA-damage-induced stress kinases and p21CIP/WAF-1 expression by caffeine abrogated G2 arrest and induced apoptosis of the irradiated cells in a caspase-3-dependent manner. Inhibition of cell cycle progression by adenoviral expression of the cyclin dependent kinase inhibitor p21CIP/WAF-1 prevented apoptosis upon caffeine treatment indicating that cell cycle progression, that is, G2-release, is required for induction of apoptosis. Likewise, cells homozygously deleted for p21CIP/WAF-1 (HCT116 p21−/−) display enhanced irradiation-induced apoptosis via a caspase-3-dependent mechanism. These data indicate that the disruption of G2 checkpoint control overcomes cell cycle arrest and resistance to γ-irradiation-induced cell death. Thus, DNA damage may trigger a permanent G2 arrest as an initial inactivation step of tumor cells where the phenomenon of apoptosis is hidden unless cell cycle arrest is overcome. The efficient induction of apoptosis upon G2 release thereby depends on the propensity to activate the key executioner caspase-3. This finding is of crucial importance for the understanding of molecular steps underlying the efficacy of ionizing radiation to delete tumor cells.

Multigene analysis of Rb pathway and apoptosis control in esophageal squamous cell carcinoma identifies patients with good prognosis

Deregulation of cell‐cycle G1‐restriction point control by disruption of Rb‐pathway components is a frequent event in cancer. In concert with the inactivation of cell death pathways, such events not only contribute to tumor development but also determine the intrinsic and acquired resistance to cancer therapy and, ultimately, disease prognosis. We previously observed that the cyclin‐dependent kinase inhibitor p16INK4a and the proapoptotic Bcl‐2 homolog Bax are positive prognostic factors and identify patients with good prognosis in esophageal squamous cell carcinoma (SCC). In the present study, we therefore extend our analysis to additional genes controlling the G1 restriction point and apoptosis, respectively. This retrospective analysis was performed in a cohort of 53 patients undergoing surgery for esophageal SCC with curative intent, i.e., R0 resection. Protein expression profiles of cyclin D1, p16INK4a, Rb, p21CIP/WAF‐1, p53, Bax and Bcl‐2 were analyzed by immunohistochemistry and compared to p53 mutational status, as determined by SSCP‐PCR of exons 5–8. Loss of p16INK4a, Rb, p21CIP/WAF‐1 or Bax and overexpression of cyclin D1 were associated individually with shorter overall survival, while Bcl‐2 expression and p53 mutation were not of prognostic relevance. The longest survival was observed in a subgroup of patients whose tumors bore a combination of favorite genotypes, i.e., low cyclin D1 and high Rb, p21CIP/WAF‐1, p16INK4a and Bax protein expression. These results show that multigene analyses based on limited sets of functionally linked genes reliably identify patients with good vs. poor prognosis.

Molecular ordering of hypoxia-induced apoptosis: critical involvement of the mitochondrial death pathway in a FADD/caspase-8 independent manner

Dys-regulated growth and improper angiogenesis commonly lead to areas of hypoxia in human tumors. Hypoxia is known to be associated with a worse outcome since a lack of oxygen interferes with the efficacy of chemotherapy or radiotherapy. In parallel, hypoxia-induced apoptosis may also impose a selection pressure favoring growth of more resistant tumor cells. However, the mechanisms of hypoxia-induced apoptosis and the relative contribution of intrinsic and extrinsic apoptotic pathways are not understood. Therefore, Jurkat cell lines with defined defects in the extrinsic or intrinsic signaling cascades were used to evaluate the role of either pathway for induction of apoptosis under hypoxic conditions. Jurkat cells were incubated in hypoxia and the rate of apoptosis induction was determined by Western blotting, fluorescence microscopy and flow cytometry. Hypoxia-induced apoptosis was not affected by lack of caspase-8 or FADD, whereas overexpression of Bcl-2 or expression of dominant-negative caspase-9 mutant rendered the cells resistant to hypoxia-induced apoptosis. These results suggest that hypoxia-induced apoptosis mainly relies on intrinsic, mitochondrial pathways, whereas extrinsic pathways have no significant implications in this process. Thus, in human tumors, hypoxia will mainly lead to the selection of hypoxia-resistant cells with defects in mitochondrial apoptosis signaling pathways.

Cyclic exposure to hypoxia and reoxygenation selects for tumor cells with defects in mitochondrial apoptotic pathways

The negative influence of hypoxia on the outcome of malignant tumors may be caused by direct oxygen effects, and potentially, the selection of resistant tumor cells under repetitive hypoxia. To evaluate whether cyclic hypoxia selects for resistant cells and to analyze the underlying mechanisms, the influence of cyclic hypoxia on intracellular death pathways was determined in tumor cells. It could be demonstrated that cyclic hypoxia selects for cells with increased resistance against hypoxia‐induced apoptosis. These cells exhibited a cross‐resistance against paradigmatic triggers of mitochondrial apoptotic pathways (ionizing radiation/etoposide). In contrast, TRAIL‐receptor mediated apoptosis remained unaffected. Thus, cyclic hypoxia selects for cells with defects of the mitochondrial rather than receptor‐mediated pathways. Selection of p53‐defective cells has been described as a consequence of cyclic hypoxia; therefore, we evaluated the impact of hypoxic selection on activation of p21 and downstream mediators of p53‐dependent apoptosis. p53 function and protein levels of key mediators of mitochondrial apoptosis remained unaffected by hypoxic selection. However, radiation‐induced conformational changes of Bax were reduced after cyclic hypoxia. In summary, it could be demonstrated that hypoxic stress confers a selection pressure on mitochondrial apoptotic pathways and, consecutively, to an increased resistance toward mitochondrial death triggers.

Bak functionally complements for loss of Bax during p14ARF-induced mitochondrial apoptosis in human cancer cells.

In contrast to the initial notion that the biological activity of p14ARF strictly depends on a functional mdm-2/p53 signaling axis, we recently demonstrated that p14ARF mediates apoptosis in a p53/Bax-independent manner. Here, we show that p14ARF induces breakdown of the mitochondrial membrane potential and cytochrome c release before triggering caspase-9- and caspase-3/7-like activities in p53/Bax-deficient DU145 prostate cancer cells expressing wild-type Bak. Re-expression of Bax in these cells failed to further enhance p14ARF-induced apoptosis, suggesting that p14ARF-induced apoptosis primarily depends on Bak but not Bax in these cells. To further define the role of Bak and Bax in p14ARF-induced mitochondrial apoptosis, we employed short interference RNA for the knockdown of bak in isogeneic, p53 wild-type HCT116 colon cancer cells either proficient or deficient for Bax. There, combined loss of Bax and Bak attenuated p14ARF-induced apoptosis whereas single loss of Bax or Bak was only marginally effective, as in the case of DU145. Notably, HCT116 cells deficient for Bax and Bak failed to release cytochrome c and showed attenuated activation of caspase-9 (LEHDase) and caspase-3/caspase-7 (DEVDase) upon p14ARF expression. These data indicate that p14ARF triggers apoptosis via a Bax/Bak-dependent pathway in p53-proficient HCT116, whereas Bax is dispensable in p53-deficient DU145 cells. Nevertheless, a substantial proportion of p14ARF-induced cell death proceeds in a Bax/Bak-independent manner. This is also the case for inhibition of clonogenic growth that occurs, at least in part, through an entirely Bax/Bak-independent mechanism.

Loss of p21 disrupts p14ARF-induced G1 cell cycle arrest but augments p14ARF-induced apoptosis in human carcinoma cells

The human INK4a locus encodes two structurally unrelated tumor suppressor proteins, p16INK4a and p14ARF (p19ARF in the mouse), which are frequently inactivated in human cancer. Both the proapoptotic and cell cycle-regulatory functions of p14ARF were initially proposed to be strictly dependent on a functional p53/mdm-2 tumor suppressor pathway. However, a number of recent reports have implicated p53-independent mechanisms in the regulation of cell cycle arrest and apoptosis induction by p14ARF. Here, we show that the G1 cell cycle arrest induced by p14ARF entirely depends on both p53 and p21 in human HCT116 and DU145 carcinoma cells. In contrast, neither loss of p53 nor p21 impaired apoptosis induction by p14ARF as evidenced by nuclear DNA fragmentation, phosphatidyl serine exposure, and caspase activation, which included caspase-3/7- and caspase-9-like activities. However, lack of functional p21 resulted in the accumulation of cells in G2/M phase of the cell cycle and markedly enhanced p14ARF-induced apoptosis that was, nevertheless, efficiently inhibited by the cell permeable broad-spectrum caspase inhibitor zVAD-fmk (valyl-alanyl-aspartyl-(O)-methyl)-fluoromethylketone). Thus, loss of cell cycle restriction point control in the absence of p21 may interfere with p14ARF-induced apoptosis. Finally, these data indicate that the signaling events required for G1 cell cycle arrest and apoptosis induction by p14ARF dissociate upstream of p53.

Array-based comparative gene expression analysis of tumor cells with increased apoptosis resistance after hypoxic selection.

Tumor hypoxia is an adverse prognostic factor. In a recent study, we could demonstrate that cyclic hypoxia selects for hypoxia-tolerant tumor cells, which are cross-resistant to other stimuli of mitochondrial death pathways. In contrast, sensitivity of the cells to death-receptor ligands was mainly not affected. The aim of the present study was to further elucidate cellular changes induced by cyclic hypoxia and to identify alterations in gene expression pattern upon hypoxic selection by means of DNA-microarray analysis. Our data reveal that cyclic hypoxia resulted in the selection of cells with resistance to doxorubicine and radiation. Furthermore, hypoxic selection was accompanied by constitutive changes of the gene expression pattern with downregulation of 156 and upregulation of 82 genes. Most of the differentially regulated genes were involved in cellular responses to hypoxia and reoxygenation. While many of the genes that were downregulated upon hypoxic selection represent genes that are usually upregulated by acute hypoxia, the genes that were upregulated represent genes that are involved in stress resistance and anti-apoptotic signalling. Most importantly, hypoxic selection was not associated with changes of single apoptosis relevant genes, but with alterations in gene expression levels of a wide variety of genes indicating a more complex adaptation process.

Disruption of cell death signaling in cancer: impact on disease prognosis and response to therapy.

Disruption of cell cycle and apoptosis signaling pathways are key events during tumorigenesis, tumor progression and development of resistance against anticancer therapies. Thus, the analysis of functional alterations within these signaling cascades is of utmost importance for the understanding of resistance mechanisms, clinical outcome and risk-adapted treatment strategies. Key signaling pathways involved in the treatment resistance include the p53/p14ARF signaling complex and the mitochondrial apoptosis machinery. Apart from the direct genetic events, these signaling cascades are subject to epigenetic modulations implied by the tumor microenvironment.