First-generation EGFR-TKIs were designed to inhibit ATP binding to wild-type EGFR tyrosine kinase via reversible competitive binding. By contrast, second-generation EGFR-TKIs, such as afatinib, neratinib, and dacomitinib, were designed to covalently bind to ERBB receptor family members and irreversibly block their enzymatic activity. Second-generation EGFR-TKIs inhibit tyrosine kinase activity in EGFR/ERBB1 and also in HER2/ERBB2 and HER4/ERBB4, which contain an electrophilic group capable of a Michael addition to conserved cysteine residues within the catalytic domains of EGFR (Cys797), HER2 (Cys805), and HER4 (Cys803). The precise role of HER2 and HER4 in lung cancer remains unclear; the conformation of HER2 resembles a ligand-activated state and may favor pathogenic EGFR signaling after formation of a heterodimer with EGFR [50]. The K_d of afatinib for wild-type EGFR, exon19 EGFR, and L858R EGFR is 0.25 nM, 0.11 nM, and 0.2 nM, respectively (Table 10.2) [51]. Afatinib shows comparable ability to inhibit EGFR tyrosine kinases with uncommon mutations, such as G719X or L861Q. Notably, the K_d of afatinib for EGFR with both L858R and T790M is 1.1 nM, and antitumor activity was estimated for EGFR acquire T790M by these enzymatic analyses.

Cell-based assays showed that afatinib suppresses EGFR tyrosine kinase activity for a longer time than the reversible first-generation EGFR-TKIs; this effect is due to the irreversible suppression of kinase activity that continues until the cancer cells synthesize new EGFR [52]. The effective concentration of afatinib is one to two orders of magnitude below those needed for inhibition of colony formation by erlotinib in soft agar assays of tumor cells harboring EGFR mutations [53]. Although in vitro assays have shown the potency of afatinib (IC₅₀ of 9–10 nM) against cancer cells carrying the EGFR L858R/T790M double mutation [54], a xenograft model showed that afatinib alone did not exhibit enough therapeutic efficacy against tumor cells harboring the exon20 T790M mutation. An additional anti-EGFR mAb, cetuximab, was required to overcome EGFR-TKI resistance in cells with T790M [55].

Two phase III trials, LUX-Lung 3 and LUX-Lung 6, have been conducted to evaluate the efficacy of first-line afatinib on overall survival of patients who have advanced lung adenocarcinoma harboring EGFR-activating mutations [56, 57, 58]. It is noteworthy that overall survival was significantly longer for patients with EGFR exon19 del in the afatinib group than in the chemotherapy group in both trials (Fig. 10.6). By contrast, there were no significant differences between treatments for patients with EGFR L858R tumors. Four randomized phase III studies conducted to evaluate first-generation EGFR-TKIs, such as gefitinib and erlotinib, for advanced lung cancer patients harboring EGFR-activating mutations could not demonstrate superiority of EGFR-TKI to platinum doublet chemotherapy in overall survival of patients with either exon19 del or L858R; in these studies, most patients who were assigned to the chemotherapy group received first-generation EGFR-TKIs as a second-line treatment. It is likely that EGFR-TKI and platinum doublet chemotherapy are mutually non-cross resistant and that the sequence of the two treatments makes no differences to antitumor efficacy. In contrast, in LUX-Lung 3 and LUX-Lung 6, most patients who were assigned to chemotherapy received first-generation EGFR-TKIs but not second-generation EGFR-TKIs as a second-line treatment. Kato et al. analyzed the Japanese subgroup in LUX-Lung 3 and demonstrated that first-line afatinib produced a consistent and significant improvement in overall survival results in patients with exon19 del but not in patients with L858R, even though most patients (93.5 %) received subsequent first-generation EGFR-TKI therapy [59]. Thus, LUX-Lung 3 and LUX-Lung 6 indicated a possible superiority of afatinib to first-generation EGFR-TKIs in EGFR exon19 del patients and showed for the first time biologically significant differences between two common activating EGFR mutations, exon19 del and L858R.

It has been established that wild-type EGFR is primarily driven by ligand-binding-induced extracellular domain dimerization. After extracellular dimerization, a critical step in EGFR activation is the formation of an asymmetric dimer of kinase domains, in which the C-terminal lobe of one kinase domain (donor) and the N-terminal lobe of another kinase domain (acceptor) associate (Fig. 10.7a). EGFR with L858R preferentially assumes the acceptor role and requires a wild-type EGFR donor for superacceptor activity (Fig. 10.7b) [60, 61]. By contrast, EGFR with an exon19 del is active as either acceptor or donor and is oncogenic even in the absence of dimerization (Fig. 10.7c) [62]. Overall, the oncogenic signaling of exon19 del is dimerization independent and that of L858R is dependent on heterodimer formation with wild-type EGFR. It is still uncertain why afatinib and first-generation EGFR-TKI show different antitumor activities for lung cancer patients harboring exon19 del EGFR. It is likely that EGFR with exon19 del has a higher oncogenic potential and requires stronger inhibition by EGFR-TKI for optimal antitumor efficacy.

Approximately 10 % of lung cancer patients with mutated EGFR harbor uncommon mutations, such as G719X, L861Q, or rare mutations. However, there is a paucity of data regarding the sensitivity of these tumors to EGFR-TKI. Yang et al. [[63] ]collected data from LUX-Lung 2, LUX-Lung 3, and LUX-Lung 6 and conducted a post hoc analysis to assess antitumor efficacy of afatinib for uncommon mutations. As estimated from the affinity data for the inhibition of enzymatic activity of mutated EGFR, afatinib demonstrated significant antitumor activity against lung cancer with uncommon EGFR mutations, especially G719X, L861Q, and S768I.

Although afatinib showed antitumor activity in cell-based assays, it had to be combined in vivo with anti-EGFR mAb to exhibit antitumor efficacy against cancer harboring EGFR T790M [64, 65]. One of the reasons why the in vitro antitumor activity could not be translated into in vivo antitumor efficacy is that a plasma concentration sufficient to inhibit EGFR T790M could not be obtained because nonselective inhibition of wild-type EGFR by afatinib resulted in intolerable adverse events.

The most common grade 3 or 4 adverse events related to afatinib treatment were rash or acne, diarrhea, paronychia, and stomatitis or mucositis. Compared to first-generation EGFR-TKIs, afatinib induced grade 3 or 4 skin rashes, acne, stomatitis, and diarrhea more frequently because of its higher affinity to wild-type EGFR. The severity of these common adverse events depends on afatinib plasma concentration. By contrast, the frequency of interstitial lung disease in patients treated with afatinib is similar to that in patients treated with first-generation EGFR-TKIs [56, 57]. Grade 3 or 4 transaminase elevation was less frequent in patients treated with afatinib than in patients treated with gefitinib, because afatinib is not metabolized by cytochrome enzymes in the liver.

Although second-generation EGFR-TKIs show stronger activity for inhibiting mutated EGFR tyrosine kinases by forming irreversible covalent bonds, the nonselective inhibition of wild-type EGFR tyrosine kinase results in adverse events such as skin rashes, acne, stomatitis, and diarrhea. Furthermore, the acquisition of a T790M mutation lowers the affinity of EGFR-TKIs for mutant EGFR, and the EGFR-TKI concentrations needed to inhibit tyrosine kinase activity are not achievable due to toxicity related to nonselective inhibition of wild-type EGFR. Third-generation EGFR-TKIs are designed to inhibit EGFR tyrosine kinase harboring activating mutations and T790M through the formation of irreversible bonds while sparing the activity of wild-type EGFR tyrosine kinase. WZ4002 was the first agent to be published. Rociletinib (CO1686), which is closely related to WZ4002, is in clinical trials. Osimertinib (AZD9291) and HM61713 are other third-generation EGFR-TKIs that have progressed to clinical trials [66].