Electrophoretic mobility shift assays (EMSAs) revealed that aspirin-induced I B degradation was accompanied by a dose-dependent specific increase in nuclear NF- B (p50/p65) DNA binding complexes (Fig. 1A , B ). The findings from EMSAs were further corroborated by immunocytochemistry. Before aspirin treatment, p65 was localized mainly in the cytoplasm, but after 24 h treatment with 10 mM aspirin, there was extensive nuclear staining for the protein (Fig. 1D ). These results establish that 24 h exposure to aspirin activates the NF- B pathway colorectal cancer cells.

3. Aspirin-induced I B degradation and NF- B nuclear translocation precede cell death
To investigate the possibility that NF- B nuclear translocation was a consequence of cell death, we studied the kinetics of the aspirin effects on NF- B signaling and apoptosis. Aspirin (10 mM) treatment induced complete degradation of I B after 2–5 h. Similarly, an increase in nuclear NF- B DNA binding was observed 2 h after treatment and persisted for > 16 h. In comparison, aspirin-induced apoptosis, determined by externalization of phosphatidylserine, was not detected until 16 h after treatment. These results show that apoptosis occurred after NF- B nuclear translocation, suggesting the possibility of causal relationship between aspirin-induced NF- B activation and subsequent cell death.

4. Inhibition of NF- B nuclear translocation inhibits aspirin-induced apoptosis
To definitively prove the relationship between NF- B nuclear translocation and aspirin-mediated growth inhibition, we generated stable transfectants of HRT18 and CT26 colon cancer cells that constitutively express the I B^S32/36-tag construct. Using Western blot analysis with the anti-tag antibody, two HRT18 (I B^S32/36h1, h28) and two CT26 (I B^S32/36ct3, ct4) clones were identified that expressed high levels of mutant protein. Expression of I B ^S32/36 resulted in substantial inhibition of aspirin-induced nuclear translocation of NF- B in all clones compared with respective parental cells (Fig. 2A ).If nuclear translocation of NF- B were contributing to apoptosis, then inhibition should protect against aspirin-induced cell death. Indeed, the number of I B^S32/36 expressing viable cells actually increased in the presence of 1 mM aspirin compared with a 32.4% and 55.8% reduction in viable cell counts in parental HRT18 and CT26 cells, respectively (Fig 2B ). Similarly, 5 mM aspirin had significantly (P<0.05) less effect on the viability of mutant I B expressing clones than on the viability of parental lines (Fig. 2B ). These data indicate that aspirin-induced apoptosis of colorectal cancer cells requires phosphorylation and degradation of I B and subsequent nuclear translocation of NF- B complexes.

5. Cell specificity of aspirin-induced I B degradation and apoptosis
Epidemiological studies indicate that nonsteroidal anti-inflammatory drug (NSAID) -mediated protection is relatively specific to colorectal tumors. Therefore, we wanted to determine whether cells that are not derived from colorectal tumors also responded to aspirin with similar effects on the NF- B pathway and cell viability. In all 4 (SW480, HRT-18, HCT116, CT26) colorectal cancer cell lines tested, 5 mM aspirin induced substantial I B degradation whereas 10 mM aspirin induced almost complete degradation. In contrast, 10 mM aspirin had no effect on I B in embryonic kidney (293) or in lung adenocarcinoma (A549) cells. Substantial growth inhibitory effects were observed in all colorectal cancer cell lines treated with 5 or 10 mM aspirin. Numbers of viable cells decreased by 2- to 6.7-fold after aspirin (5 mM) treatment. On the contrary, numbers of viable 293 cells increased in the presence of 5 mM aspirin while A549 cells only showed a 1.2-fold decrease in viability after treatment. These results suggest that aspirin-mediated I B degradation and apoptosis may be cell type specific.

6. Aspirin induces I B degradation in normal colonic mucosa and tumors from rectal cancer patients
To establish the potential clinical significance of our results, we investigated aspirin effects in clinical material. We treated biopsy samples of normal mucosa and tumor material from patients undergoing surgical resection for rectal cancer. I B degradation was observed in both the tumor and, to a lesser extent, in the normal mucosa, after 5 h treatment ex vivo with 10 and 20 mM aspirin. No change was observed in levels of control (Cu/ZnSOD) protein. These data confirm that clinical tumor biopsy material show the same I B response as observed in cell line experiments.

CONCLUSIONS

There is compelling evidence that NSAIDs have a protective effect against colorectal cancer. However, their detrimental side effects limit their potential use as chemopreventative agents. Therefore, there is a pressing need to understand the mechanisms by which NSAIDs exert their chemopreventative effects in order to allow development of safer alternatives.

We show here for the first time that aspirin mediates a reduction in cytoplasmic I B levels in colorectal cancer cells that is time and concentration dependent and due to phosphorylation and proteasome-mediated degradation of the protein. Using EMSAs and immunocytochemistry, we demonstrate that I B degradation results in nuclear translocation of p50/p65 NF- B complexes, confirming that aspirin stimulates the NF- B pathway. Although these data would appear to question studies showing that NSAIDs inhibit activation of NF- B through specific modulation of the I B kinase, previous studies examined only the very short-term effects of NSAIDs on activation of NF- B mediated by cytokines. We investigated the effects of aspirin alone on NF- B signaling, which is highly relevant to the anti-tumor activity of the agent.

The evidence of a link between NF- B translocation and apoptosis due to aspirin treatment, was initially implied by the correlation between I B degradation and reduced cell viability. Time course experiments indicated that the NF- B response preceded cell death, raising the possibility of a causal relationship. This was confirmed in cells we engineered to continuously express a dominant negative mutant I B (I B ^S32/36). These cells showed inhibition of both aspirin-induced NF- B nuclear translocation and apoptosis vs. their parental counterparts. We also considered whether the lack of apoptotic response to aspirin in the mutant I B -expressing clones might be due to their slower rate of growth compared with parental cells. However, this is unlikely since these cells grew at a similar rate to other colorectal cancer cell lines (SW480) in which aspirin induced substantial apoptosis. Thus, we conclude that it is NF- B nuclear translocation that mediates aspirin-induced apoptosis of colorectal cancer cells.

These novel findings of a prolonged effect of aspirin on NF- B signaling provide new insight into the mechanism of action of aspirin against colorectal cancer and will inform chemoprevention strategies.