Responses To Treatment With EGFR Inhibitor And Molecular Mechanisms Of Intrinsic Resistance In Human Colon Cancer Cells

Background and purposeAberrant signaling through the epidermal growth factor receptor (EGFR) plays a major role in the progression and maintenance of the malignant phenotype and the receptor is therefore a rational anti-cancer target. A variety of approaches have been developed to specifically target the EGFR which include monoclonal antibodies and small molecule tyrosine kinase inhibitors (TKI), such as gefitinib (ZD1839, Iressa?). However, the recent clinical experience across a range of cancer types is revealing that despite gefitinib demonstrating objective responses in a minority of patients, there is majority which has no significant effects, suggesting high level of intrinsic or acquired resistance to such treatment and moreover, overexpression of the EGFR is clearly not the sole determinant of response to gefitinib therapy. Such adverse phenomena, which serve to limit the overall therapeutic impact of this new agent, implies the existence of a greater complexity involved in the regulation of EGFR signaling than was previously assumed.Colorectal cancer (CRC) is one of the major types of cancer worldwide, in terms of both morbidity and mortality. EGFR has been shown to be overexpressed and elevated activation in CRC patient populations but its predictive value of patients'prognosis and sensitivity to gefitinib remains unclear. Both in vitro and in vivo studies indicated that gefitinib had antitumor activity as a monotherapy in some CRC cell lines. Disappointingly, however, phase I/II clinical studies in patients with advanced CRC showed that gefitinib had negligible single agent activity or combination settings with fluorouracil or irinotecan synergistic interaction. As the field of this molecular therapeutics continues to evolve, a comprehensive understanding of resistance mechanisms may contribute to identify the subset of patients who could benefit from gefitinib therapy, and ultimately lead to refinements in our regimens to provide better clinical outcomes. In addition to EGFR family members, there are redundant other growth factor receptors on tumor cells membrane, such as insulin-like growth factor receptor-1 (IGFR-1), hepatocyte growth factor receptor (MET), which have similar extracellular domain structures and tyrosine kinase activity also, and share similar downstream pathways to EGFR. It is now evident that aberrant expressions of IGFR-1 or MET is additional molecular alteration occurring during CRC progression, associated with cell growth and differentiation closely. Indeed, existence of crosstalk among tyrosine kinase receptors caused by cell signaling transduction networks may partially responsible for resistance to anti-EGFR treatment.On binding its ligands, EGFR undergoes dimerization or heterodimerization, which activates the intrinsic protein tyrosine kinase via autophosphorylation within its cytoplasmic domain. The tyrosine-autophosphorylated region serves as a binding site for cytoplasmic messenger protein, which, in turn, leads to downstream cascade activation of major signaling pathways, including the PI3K/AKT and Ras/Raf/MAPK pathways related to cellular proliferation and survival. Gefitinib is a selective EGFR tyrosine kinase inhibitor that competes with ATP and binds at the ATP site on EGFR, thereby inhibiting EGFR phosphorylation and downstream signal transduction. However, evidence is accumulating which demonstrate that the ability of certain tumor cells to maintain downstream signaling activation in the presence of gefitinib could represent resistant phenotype, which escapes the antiproliferative activity of this new class of agent.In the present study, we analyzed the relationship between EGFR and its downstream signals baseline expression or activation and sensitivity to gefitinib treatment in vitro of 6 colon cancer cell lines, and observed the activation status of EGFR and its downstream signals in the cancer cells of different sensitivity types with gefitinib treatment. We also investigated IGFR-1 or MET expression and activation which was relevant to EGFR of all cancer cell lines tested, and thereby explored the probably existent correlation with gefitinib sensitivity, then uncovered the mechanism underlying resistance to gefitinib treatment and reversed the resistant phenotypes of cancer cells based on our findings.Methods1. Immunocytochemistry and RT-PCR were used to detect the protein and mRNA expression of EGFR in 6 colon cancer cell lines respectively. 2. In vitro drug sensitivity assay via MTT was to evaluate the growth inhibition of cancer cells treated with or without gefitinib at increasing concentrations in EGF-stimulated and nonstimulated environments, and 50% inhibition concentration (IC50) was represent to cellular sensitivity to gefitinib treatment.3. Western blot analysis was used to detect baseline expression or activation of EGFR and its downstream signals in the cancer cells and their activation status in the presence of gefitinib.4. Flow cytometry was used to analyze the effects on cell cycle or apoptosis of gefitinib treatment in colon cancer cells.5. Other tyrosine kinase receptors IGFR-1 and MET baseline expression and activation were detected by Western blot analysis and their activation status in the presence of gefitinib as well, and the mRNA expression of matched receptor ligand was detected by RT-PCR.6. IGFR-1 or MET activation was inhibited by matched tyrosine kinase inhibitor, and the effects on common downstream signaling activation were assessed by Western blot.7. Interactions of receptors and association with downstream signals were measured by immunoprecipitation and Western blot analysis.8. MTT assay and flow cytometry were used to evaluate the growth inhibitory effects on cancer cells after reversing resistant phenotype treatment.9. The activation status of glycogen syntheses kinase 3β(GSK-3β) was detected by Western blot analysis, which is a common downstream kinase of AKT and MAPK. The influence of GSK-3βinhibitor lithium chloride (LiCl) on TKI-induced cytostasis was assessed by MTT assay. Immunofluorescence staining of colon cancer cells for activated GSK-3βwas served to observe the intranuclear accumulation with laser confocal microscopy.10. CRC tumor and healthy adjacent tissue specimens were harvested, and EGFR and IGFR-1 / MET baseline expression or activation were measured with Western blot analysis to investigate the difference between tumors and matched normal tissues, and analyze the correlations with their activation and the relationships with clinicopathological parameters in CRC patients. Results1. EGFR protein and mRNA expressions of 6 colon cancer cell lines: Lovo, HCT116, HT29, LS174T and SW480 cells exhibited different amounts of EGFR protein and mRNA expressions, and Lovo cells harbored the highest level of expression, while SW620 had little expression.2. Of all 6 cancer cell lines, Lovo was most sensitive to gefitinib, and HT29 and SW480 had lower sensitivity than Lovo, while HCT116, LS174T and SW620 were most resistant: Under 1μmol/L gefitinib treatment, Lovo had remarkably higher cell growth inhibition (34% in 10% FBS, 37% in EGF) than other cell lines (P<0.05). Lovo had most sensitivity with an IC50<10μmol/L, and HT29 and SW480 had moderate sensitivity with 10μmol/L 100μmol/L. There was no significant effect on cell growth inhibition with or without EGF stimulation (P>0.05).3. Sensitivity to gefitinib of colon cancer cells maybe some correlated with EGFR baseline expression or activation, but they were not the sole determiner: Lovo had the highest level of EGFR expression and activation with the most sensitivity to gefitinib, suggesting they cloud be correlated with cytostatic effect, whereas HCT116 and LS174T which exhibited EGFR expression had similar resistant to gefitinib with SW620 which exhibited little EGFR expression, and HT29, SW480, HCT116 and LS174T expressed similar amounts of EGFR but displayed differences in gefitinib sensitivity, suggesting EGFR expression could not predict cytostatic effect. The baseline expression and activation of AKT and MAPK or PTEN expression had no significant differences between sensitive cell lines and resistant ones.4. Sensitivity to gefitinib of colon cancer cells could rely on the EGFR dependent cell growth and the coupling between EGFR and its downstream signals: First, with EGF-stimulated environment, EGFR activation were up-regulated in Lovo and HT29 cells (P<0.05), while slight upregulated in HCT116 cells, and AKT and MAPK activation were also up-regulated in Lovo and HT29 (P<0.05), while no significant changes in HCT116. These suggested that the EGFR/AKT and EGFR/MAPK axes were intact in Lovo and HT29, and cellular growth independent of EGFR would relate to resistance to gefitinib. Second, under gefitinib treatment, EGFR expression and activation in Lovo cells were down-regulated accompanied with the down-regulated activation of AKT and MAPK (P<0.05). EGFR activation in HT29 cells was also down-regulated (P<0.05), whereas activation of AKT and MAPK were not down-regulated (P>0.05). EGFR, AKT and MAPK activation had no significant changes in HCT116 cells. PTEN expressions in all 6 colon cancer cells remained unchanged. These suggested that the uncoupling between EGFR and its downstream signals would relate to lower sensitivity to gefitinib.5. Gefitinib induced cytostatic effects on colon cancer cells through the regulation of cell cycle and apoptosis: Under 10μmol/L gefitinib treatment, G1 phase proportion of Lovo cells was significantly increased from (59.49±1.74) % to (76.25±3.01) % (P<0.05), accordingly S and G2 phase proportions were decreased, and apoptosis ratio of Lovo cells was also significantly increased from (1.97±0.35) % to (42.68±4.18) % (P<0.05), while G1 phase proportion and apoptosis ratio of HT29 cells were slight increased, and no obviously changes in HCT116 cells.6. Elevated activation of IGFR-1βcorrelated with resistance to gefitinib of colon cancer cells: First, the baseline activation of IGFR-1βbut not its expression was lower in Lovo, HT29 and SW480 cells than that in HCT116, LS174T and SW620 cells (P<0.05). And the baseline expression or activation of MET did not correlate with gefitinib sensitivity. This indicated that elevated baseline activation of IGFR-1βwould relate to resistance phenotype. Second, under gefitinib treatment, the expression or activation of IGFR-1βremained unchanged in Lovo and HCT116 cells (P>0.05), and IGFR-1βexpression had no change in HT29 cells, but activation of IGFR-1βwas remarkably elevated (P<0.05). This indicated that elevated activation of IGFR-1βin the presence of geifitinib would relate to decreased sensitivity to gefitinib. Finally, in the presence of gefitinib the mRNA expression levels of IGF-II that one of the IGFR-1β'ligands was significantly increased in HT29 cells (P<0.05), whereas no changes in Lovo and HCT116 cells, revealing the evidence of elevated functional IGFR-1βmediated by IGF-II autocrine loop of HT29 cells.7. IGFR-1βtyrosine kinase inhibitor AG1024 could reduce the maintaining AKT and MAPK activity, and overcome resistance to gefitinib in HT29 and HCT116 cells: AG1024 did not exert any significant effects on EGFR, IGFR-1β, AKT and MAPK expression or activation in Lovo cells (P>0.05). AG1024 reduced activation on IGFR-1βbut not on AKT and MAPK in HT29 cells, while in combination with gefitinib exerted inhibition on AKT and MAPK activation (P<0.05). AG1024 reduced activation on IGFR-1βaccompanied with AKT and MAPK in HCT116 cells (P<0.05), and in combination with gefitinib did not make more effective than AG1024 alone (P>0.05). And this combination made HT29 cell proliferation remarkably decreased, cell cycle G1 phase arrest and apoptosis increased respectively (P<0.05), and these inhibitory effects could be induced in HCT116 by AG1024 without coinhibition of EGFR (P<0.05). These findings suggested that AG1024 could overcome resistance to gefitinib in colon cancer cells.8. The resulting crosstalk from EGFR/IGFR-1βheterodimerization and constitutively activating the common downstream signaling AKT and MAPK pathways might contribute to gefitinib resistance of colon cancer cells: EGFR of Lovo cells had no physical association with IGFR-1β, and reduced associations with AKT and MAPK in the presence of gefitinib (P<0.05). EGFR of HT29 cells had physical association with IGFR-1βin the presence of gefitinib, and reduced associations with AKT and MAPK in combination with AG1024 (P<0.05). EGFR of HCT116 cells had physical association with IGFR-1β, and reduced associations with AKT and MAPK in the presence of AG1024 (P<0.05).9. Gefitinib or AG1024-induced cytostasis was mediated by GSK-3βactivation: Increased activity of GSK-3βemerged in Lovo cells but no significant changes in other colon cancer cells treated with gefitinib. GSK-3βactivation of HT29 cells was obviously increased in combination with AG1024 (P<0.05), and similar effect on HCT116 cells treated with AG1024 (P<0.05). And this drug-induced cytostasis could be rescued by GSK-3βinhibition using LiCl (P<0.05). Active GSK-3βof colon cancer cells showed nuclear translocation in the presence of gefitinib and/or AG1024.10. 55.6%(10/18) CRC tumor specimens had higher levels of EGFR expression than 37.5%(3/18) did in normal tissue specimens from the same patients. 33.3%(6/18) tumor tissue had significant higher levels of p-EGFR (activation form of EGFR) than 5.6%(1/18) did in normal tissue (P<0.05). 44.4%(8/18) tumor tissue had significant higher levels of p-IGFR-1β(activation form of IGFR-1β) than 11.1%(2/18) did in normal tissue (P<0.05). EGFR expression was no correlation with p-EGFR (r = 0.228, P>0.05), and p-EGFR expression was also no correlation with p-IGFR-1β(r = -0.127, P>0.05). None of the clinicopathological parameters showed any association with EGFR and p-EGFR expression. (P>0.05), while p-IGFR-1βexpression significantly increased in patients with lymph node metastasis (P<0.05), and there was a same trend in patients of advanced stage.Conclusions1. Of all 6 colon cancer cell lines Lovo was most sensitive to gefitinib treatment, relatively, HT29 and SW480 had intermediate sensitivity, HCT116, LS174T and SW620 were most resistant.2. Sensitivity to gefitinib of colon cancer cells maybe had some correlation with baseline expression and activation of EGFR, the higher expression or activation showed more sensitive to gefitinib, but they were clearly not the sole determinant of response to such treatment.3. Cellular growth independent of EGFR would relate to gefitinib resistance in HCT116 cells, and uncoupling between EGFR and its downstream signals would relate to lower sensitivity to gefitinib in HT29 cells, and elevated activation of IGFR-1βmight be partially responsible for the independence and uncoupling.4. The mechanism underlying resistance to gefitinib treatment of colon cancer cells might be arised from activation of IGFR-1βsignaling pathway through the formation of EGFR/ IGFR-1βheterodimerization, and consequently constitutive activation of the common downstream signaling AKT and MAPK.5. AG1024, an IGFR-1βtyrosine kinase inhibitor combinated with gefitinib could reverse resistant phenotype in HT29 cells, and so did in HCT116 cells with AG1024 alone.6. Gefitinib or AG1024-induced cytostasis was mediated by GSK-3βactivation, suggesting that the responses of colon cancer cells to EGFR/ IGFR-1βblockade could be predicted early in the course of treatment by measuring the activation of GSK-3β.7. Activation of EGFR and IGFR-1βwere significantly increased in CRC tumor tissue, and the EGFR expression was not correlated with its activation, and increased activation of IGFR-1βmay favor the metastasis of CRC.