The epidermal growth factor receptor (EGFR) was the first extracellular receptor identified as a tyrosine kinase and the first target of selective small molecule tyrosine kinase inhibitors. One of these, gefitinib, was the second kinase inhibitor to be approved by the FDA, in 2003, for the treatment of non-small-cell lung cancer (NSCLC); a second EGFR inhibitor, erlotinib, was approved the following year. (The Bcr-Abl inhibitor, imatinib (Gleevec), was approved for breast cancer in 2001.) Activating EGFR mutations were discovered in a subgroup of NSCLC patients from the initial gefitinib clinical trials. These mutations, which render the tyrosine kinase constitutively active in the absence of ligand, occur in about 50% of lung adenocarcinomas from Asian patients and about 10% for non-Asians. Though discovery of the activating mutations initiated a new paradigm of personalized therapy for NSCLC and vastly improved patient response rates, most patients eventually relapse. In the majority of cases acquired resistance is caused by the T790M secondary mutation, which renders EGFR insensitive to gefitinib and other first line inhibitors. For the last several years, researchers have been intensively focused on how to overcome acquired resistance to first generation EGFR inhibitors.
One of the most promising and extensively studied approaches for overcoming acquired resistance to drugs targeting EGFR has been the development of irreversible, covalent-binding, inhibitors that are active against EGFR T790M. These “second generation” inhibitors, which include afatinib, neratinib and dacomitinib, have a different molecular mechanism of action than the first generation ATP-competitive inhibitors and their covalent binding results in more effective and durable repression of EGFR kinase activity. However, despite promising results in EGFR-mutated cell lines harboring the T790M mutation, none of the second generation have shown significant efficacy in NSCLC patients. This may be because they inhibit WT and T790M EGFR kinase activity with similar potency, and achieving a high enough dosage to overcome resistance causes unacceptable toxicity. Current efforts are focused on development of inhibitors that are specific for T790M versus wild type EGFR, some of which advanced to clinical trials.
A group of investigators led by Elisa Giovannetti at the VU University Medical Center in Amsterdam and at Cancer Pharmacology Lab, University of Pisa, recently evaluated a promising T790M-selective EGFR inhibitor, CNX-2006, to assess its selectivity and identify potential mechanisms of acquired resistance. They tested the inhibitor in a panel of more than 20 cell lines expressing WT and various mutant EGFR proteins, and assessed EGFR autophosphorylation status as a measure of EGFR kinase activity. They found that, like first line drugs such as erlotinib and gefitinib, CNX-2006 inhibited WT and activated EGFR mutant kinases. However, unlike the first line inhibitors, it also inhibited EGFR carrying the T790M acquired resistant mutation, either alone or in combination with an activating mutation. In fact CNX-2006 was 10-fold more potent for inhibition of activated EGFR-T790M than for the WT receptor. CNX-2006 also inhibited phosphorylation of proteins downstream of EGFR activation.
The selectivity of CNX-2006 for activated EGFR T790M versus the wild type receptor was even more pronounced based on its antiproliferative activity. In cell lines with a wild type receptor, gefitinib, erlotinib and CNX-2006 all inhibited growth with low micromolar potencies (IC50). As expected, cells carrying EGFR with an activating mutation were sensitive to all of the EGFR inhibitors at much lower concentrations of 5 to 50 nanomolar, reflecting their “addiction” to EGFR-driven growth. However, in cells carrying both an EGFR activating mutation and T790M, sensitivity to the first generation inhibitors was lost, whereas CNX-2006 still blocked proliferation with low nanomolar potency (8 – 72 nM). This suggested that CNX-2006 might have sufficient selectivity to overcome acquired resistance conferred by the T790M mutation without causing excessive toxicity from inhibition of WT EGFR.
The in vitro results by the researchers were confirmed in a mouse model. Administration of CNX-2006 caused a drastic reduction in the growth of subcutaneous tumors from cells carrying activated EGFR with the T790M mutation, and rapid tumor growth ensued after the drug was withdrawn. Also consistent with the in vitro results, CNX-2006 selectively blocked autophosphorylation of activated EGFR T790M in the implanted tumor cells, but had no visible effect on the WT receptor from lung tissue.
Dr. Giovannetti’s team next developed an in vitro acquired resistance model for CNX-2006 by treating activated EGFR-T790M cells with increasing doses of the drug for several months. This resulted in cells which were 60-fold less sensitive to the antiproliferative effects of CNX-2006. Interestingly, these cells were also much less sensitive to the second generation inhibitors afatinib and dacomitinib.
In exploring the mechanism for the CNX-2006 resistance, the investigators looked for recurrent mutations in EGFR and other candidate genes such as KRAS and BRAF, but found none. Activated EGFR-T790M was the only target alteration found, and CNX-2006 still potently inhibited its kinase activity in the resistant cells. Nor did they find any changes in the expression level of genes reported to be affected by other third generation EGFR inhibitors or in genes involved with epithelial-mesenchymal transition, which has been implicated in lung cancer. They did however observe decreased expression of two genes related to cell adhesion, e-cadherin and β-catenin, and surprisingly found that the resistant cells were actually less motile and invasive than the parental CNX-2006 sensitive cells.
Their clue as to the potential mechanism of CNX-2006 resistance came out of an analysis of activated kinases using a kinase substrate peptide array. This analysis indicated that of 24 kinases that showed at least a 2-fold increase or decrease in activity relative to the parental cells, 17 were linked in some way with the NF-κB pathway. Overexpression of NF-κB protein and its precursor as well as increased phosphorylation of a downstream target confirmed its involvement in the development of resistance to CNX-2006. The resistant cells were essentially dependent of NF-κB for survival, with only 1% remaining viable after genetic knockdown, compared with 60% viability for the parental cells. Moreover, activation of NF-κB by phosphorylation was unaffected in the resistant cells, whereas it was completely inhibited in the parental cells. Consistent with these results, 1 μM CNX-2006 had no effect on expression of an NF-κB dependent reporter gene in the resistant cells, whereas it was inhibited up to 50% in the parental cells.
Lastly, Dr. Giovannetti’s team showed that the CNX-2006 resistant cells were more sensitive than the parental cells to pharmacological inhibition of the NF-κB pathway. They found that the CNX-2006 resistant cells were 3- to 5-fold more sensitive to the antiproliferative effects of drugs that interfere at various points upstream of NF-κB activation. And combining CNX-2006 with the NF-κB pathway inhibitors decreased the viability of resistant cells synergistically.
Third generation EGFR inhibitors that are specific for activated EGFR T790M found in drug resistant tumors are just beginning to be tested in the clinic. Dr. Giovannetti and her team have clearly shown that NF-κB activation may replace oncogene signaling in lung cancer when the drug resistant EGFR T790M form of the receptor is effectively shut down by these highly selective inhibitors. Importantly, their demonstration that inhibiting NF-κB activity reduces the viability of such super-resistant cells and provides some hope for a therapeutic solution.
– Bob Lowery