This article is mentioned in:

Introduction

Precision molecular targeted agents in non-small cell lung cancer (NSCLC) have improved survival of patients harboring driver gene mutations. Epidermal growth factor receptor (EGFR)-tyrosine kinase inhibitor (TKI) improves progression-free survival (PFS) of NSCLC patients with EGFR mutations compared with traditional platinum-based doublet chemotherapy (1,2). Furthermore, osimertinib, a third-generation EGFR-TKI, is promising as first-line treatment for EGFR mutant NSCLC (3,4). Although good responses to EGFR-TKI therapy have been shown, tumor cells can acquire resistance through several methods, in particular, secondary gene mutations that cause structural changes in the ATP binding site of the EGFR tyrosine kinase domain. EGFR T790M mutation occurs in almost half of patients following first- or second-generation EGFR-TKI therapy (5), and EGFR C797S mutation is the most common mechanism of acquired resistance against third-generation EGFR-TKIs (6).

Approximately 0.4–8% of NSCLC patients harboring de novo or germline T790M mutations are resistant to first- or second-generation EGFR-TKIs (7). However, the frequency of EGFR C797S gene mutation remains unclear. To the best of our knowledge, only one case of an NSCLC patient harboring concurrent C797S and L858R mutations prior to receiving EGFR-TKI treatment has been reported (8).

Several other mechanisms of resistance against all generation EGFR-TKIs have been identified including tertiary gene mutations other than EGFR C797S mutation (9–11), activation of bypass signaling by gene amplification (e.g., ERBB2 (12) and MET (13,14), driver gene mutations (e.g., RAS, RAF and PIK3CA) (3,15), gene alteration in cell cycle genes (14), and transformation to mesenchyme, small cell carcinoma (SCC), or squamous cell carcinoma (SqCC) (2,16,17). These described EGFR-TKI resistance mechanisms may also be expressed during the pre-TKI NSCLC state (6,15) and can be a challenge for cancer treatment of NSCLC patients with EGFR mutations.

In this retrospective study, we assessed potential resistance against third-generation EGFR-TKI therapy, such as EGFR C797S mutation, and gene amplification in EGFR-TKI-naïve surgical specimens from patients harboring known EGFR mutations.

Materials and methods

Patient selection

Consecutive patients who underwent initial lung resection or surgical tumor biopsy in Fukushima Medical University Hospital and were diagnosed with NSCLC harboring a known EGFR gene activating mutation (e.g., exon 19 deletion, L858R, T790M, S768I, G719X and L861Q) at the time samples were collected, and whose specimens were available for gene examination described below, were included in this study. Patients who had received systemic treatment or irradiation therapy before surgery were excluded.

Ethics statement

This study was conducted with approval of the ethics board at Fukushima Medical University (approval no. 2955). Human rights and welfare of participants were protected in accordance with the Declaration of Helsinki, and written informed consent was obtained from participants.

Preparation of genomic DNA

Tumor DNA was extracted from macro-dissected tumor tissue of formalin-fixed paraffin-embedded surgical specimens using the QIAamp DNA FFPE Tissue kit (Qiagen) according to the manufacturer's instructions. Tumor DNA quantity was assessed using the Qubit dsDNA HS Assay kit (Thermo Fisher Scientific, Waltham, MA, USA) and Qubit 2.0 Fluorometer (Thermo Fisher Scientific).

Peptide nucleic acid-locked nucleic acid PCR clamp method

Premix Ex Taq (Takara Bio Inc.), 200 nM of primer, 100 nM of mutation LNA probe, 100 nM of total probe, 250–2,500 nM of clamp probe, and sample DNA were mixed in a total reaction volume of 25 µl and analyzed as described previously (18). Real-time PCR was performed in 50 cycles using Light Cycler480 II (Roche; denaturation: 5 sec at 95°C; annealing and extension: 30 sec at 62°C). EGFR C797S and T790M mutations were judged using LightCycler 480 software (Roche).

Multiplex ligation-dependent probe amplification (MLPA) Assay

Gene amplification was analyzed according to the standard protocol for MLPA (19) using SALSA MLPA Probemix P175 Tumor Gain (MRC-Holland). ERBB2, MET, EGFR, ALK, BRAF, FGFR1, MYC, RET, CCND1, CCND2, CDK4, CDK6, MDM2 and MDM4 gene amplification was assessed. Fragments of PCR products were analyzed using the 3130×l Genetic Analyzer (Thermo Fisher Scientific) and amplification was judged by Coffalyser Data analysis software (MRC-Holland).

Statistical analysis

Statistical analysis was performed using SPSS 21.0 (IBM; SPSS). Continuous variables were compared by two-tailed t-tests or one-way ANOVA, and categorical variables were compared by the chi-squared test or Fisher's exact test. Multivariate analyses using a binary logistic regression model were performed to evaluate independent predictors, and hazard ratio (HR) and confidence interval (CI) were calculated. PFS was estimated using the Kaplan-Meier method, and survival curves were compared using log-rank tests. P-values of less than 0.05 were considered statistically significant. Some of the statistical analysis was performed by AC Medical Inc. (Tokyo, Japan).

Results

Patient characteristics

Between January 2007 and December 2015, we identified 248 patients, of which 246 patients were eligible for this study and included in the analyses. Of the 232 patients who had undergone complete or microscopically incomplete resection, recurrence was found in 49 patients. Surgery with incomplete resection or tumor biopsy was performed in 14 patients. Of the 63 advanced or recurrent NSCLC patients, 54 patients received EGFR-TKI therapy (Fig. 1). Demographic data and clinicopathological characteristics of patients are presented in Table I: Age at surgery or biopsy ranged from 37 to 90 years (median age: 67 years). There were 159 (64.6%) women, 166 (67.4%) never smokers, 240 (97.6%) patients with adenocarcinoma, and 193 (78.4%) patients with pathological stage I (Table I).

Figure 1. Flowchart of the patient selection process.

Table I. Clinicopathologic, genetic and histologic characteristics of patients.

Table I.

Clinicopathologic, genetic and histologic characteristics of patients.

A, Clinicopathologic characteristic
Characteristic	Total (n=246)	Received EGFR-TKI therapy(n=44)
Sex
Male	87 (35.4)	14 (31.8)
Age (y)
Median (range)	67.0 (37–90)	65.5 (37–82)
Smoking status
Never	166 (67.4)	30 (68.2)
Former	67 (27.2)	11 (25.0)
Current	13 (5.3)	3 (6.8)
Pathological subtypes
Adeno	240 (97.6)	42 (95.5)
AdSq	2 (0.8)	2 (4.5)
Sq	4 (1.6)	0 (0.0)
p-Stage at initial surgery
I	193 (78.4)
II	16 (6.5)
III	24 (9.7)
IV	12 (4.9)
B, EGFR mutation
Characteristic	Total (n=246)	Received EGFR-TKI therapy(n=44)
C797S	0 (0.0)	0 (0.0)
Ex19del	86 (35.0)	21 (47.7)
Ex19del+T790M	1 (0.4)	0 (0.0)
L858R	134 (54.5)	20 (45.5)
L858R+G719S	1 (0.4)	1 (2.3)
L858R+T790M	4 (1.6)	0 (0.0)
T790M	7 (2.8)	0 (0.0)
G719A	3 (1.2)	0 (0.0)
G719C	1 (0.4)	0 (0.0)
L861Q	1 (0.4)	1 (2.3)
G719A+S768I	2 (0.8)	0 (0.0)
G719A+L861Q	2 (0.8)	1 (2.3)
G719C+L861Q	1 (0.4)	0 (0.0)
C, Gene amplification
Characteristic	Total (n=246)	Received EGFR-TKI therapy (n=44)
ERBB2	1 (0.4)	1 (2.3)
MET	0 (0.0)	0 (0.0)
EGFR	5 (2.0)	3 (6.8)
ALK	0 (0.0)	0 (0.0)
RET	0 (0.0)	0 (0.0)
BRAF	0 (0.0)	0 (0.0)
FGFR1	1 (0.4)	0 (0.0)
MYC	2 (0.8)	1 (2.3)
CCND1	2 (0.8)	2 (4.5)
CCND2	1 (0.4)	1 (2.3)
CDK4	15 (6.0)	4 (9.1)
MDM2	17 (6.8)	6 (13.6)
MDM4	2 (0.8)	1 (2.3)

[i] Data are presented as n (%) unless otherwise indicated. AdSq, adenosquamous; Sq, squamous; EGFR, epidermal growth factor receptor; TKI, tyrosine kinase inhibitor.

EGFR C797S and T790M mutations

No patients harbored concurrent C797S mutation with known EGFR mutations. T790M mutation was identified in 12 patients (4.9%); 5 patients had the deletion in exon19 or L858R and 7 patients had T790M mutation alone (Table I).

Gene amplification

EGFR gene amplification was found in five patients. All patients with EGFR gene amplification had advanced disease or recurrence (9.8 vs. 0.0%, P<0.01). ERBB2 gene amplification was found in only one patient (0.4%) who developed recurrence. No patients harbored MET gene amplification. MDM2 gene amplification was found in 17 patients (7.1%) and CDK4 gene amplification in 14 patients (5.8%); duplication of both MDM2 and CDK4 was observed in 11 patients (4.6%). MDM2 gene amplification was significantly associated with tumor recurrence (16.7 vs. 5.1%, P=0.032). Patients with tumors with any gene amplification had significantly more advanced disease or developed recurrence compared with patients without recurrence after surgery (33.3 vs. 8.9%, P=0.002) (Fig. 2).

Figure 2. Correlation between gene amplification and disease-free survival. EGFR gene amplification was found more frequently in advanced or recurrent cohort than cohort without recurrence. There was no MET gene amplification and ERBB2 gene amplification (0.4%) was indicated in the study cohort. MDM2 gene amplification was associated with the tumor in a recurrent or advanced state. EGFR, epidermal growth factor receptor.

We next evaluated associations between clinicopathological parameters and/or gene amplification and progress of first- and second-generation EGFR-TKI treatment. Fifty-four patients had received EGFR-TKI therapy, and 44 patients underwent gene amplification analysis and were followed up. Patients who developed central nervous system metastasis (HR, 2.259, 95% CI, 1.150–4.436) and had MDM2 gene amplification (HR, 3.405, 95% CI, 1.209–9.591) exhibited significantly shorter PFS (Figs. 3,4). Patients who had EGFR gene amplification (HR, 0.656, 95% CI, 0.195–2.210), ERBB2 gene amplification (HR, 0.801, 95% CI, incalculable), CDK4 gene amplification (HR, 0.194, 95% CI, 0.660–7.802) (Figs. 3,4), CCND1 gene amplification (HR, 2.823 95% CI, 0.653–12.203), CCND2 gene amplification (HR, 0.801, 95% CI, incalculable), CDK4 gene amplification (HR, 0.194, 95% CI, 0.660–7.802), and MDM4 gene amplification (HR, 0.465, 95% CI, 0.063–3.441) (data not shown) showed no significant difference in PFS (Figs. 3,4).

Figure 3. Correlation between gene amplification and progression-free survival for first- and second-generation EGFR-TKI therapy. EGFR, epidermal growth factor receptor; TKI, tyrosine kinase inhibitor; OS, overall survival; DFS, disease free survival.

Figure 4. Survival curves for progression-free survival (PFS) in accordance with patients' genetic or clinical features. (A) Unadjusted survival curve and (B) adjusted survival curves for MDM2 amplification, (C) CDK4 amplification, and (D) EGFR amplification. MDM2 gene amplification was associated with shorter progression-free survival for first- or second-generation EGFR-TKI therapy significantly. EGFR, epidermal growth factor receptor; TKI, tyrosine kinase inhibitor.

Discussion

We herein showed that in our study cohort no EGFR-TKI-naïve NSCLC patients harbored concurrent EGFR C797S mutation with known EGFR gene mutation. Amplification of several genes was found before EGFR-TKI therapy, and MDM2 gene amplification was associated with resistance to first-generation EGFR-TKIs.

C797S mutation, which leads to acquired resistance to third-generation EGFR TKIs, occurs in 5.3–40.0% of NSCLC patients following osimertinib treatment (2,3,6,11,20) and is the most common resistance mechanism against osimertinib. No concurrent C797S mutation was found in this study, and, to our knowledge, only one case of de novo somatic L858R and C797S mutations has been reported (8). C797S mutation is an important challenge for EGFR mutant NSCLC therapy. A recent report suggests that therapeutic efficacy is dependent on allelic context of common EGFR mutations, C797S and T790M. If T790M and C797S mutations occur on separate alleles, then combination therapy comprising first- and third-generation EGFR-TKIs can restore EGFR inhibition in vitro (21). Unfortunately, these mutations almost always occur on the same allele (22). Conversely, combination therapy composed of osimertinib, bevacizumab, and brigatinib is effective for NSCLC with T790M and C797S mutations occurring on the same allele (23). Combination therapy comprising an allosteric inhibitor and cetuximab was also shown to be effective in a murine model of NSCLC driven by EGFR L858R, T790M, and/or C797S mutation (24). Furthermore, other tertiary mutations of EGFR [L792X (11), G796D (9), L718Q, and L844V (5)] in the EGFR tyrosine kinase domain were reported, and novel therapies for tumors with these mutations are required.

Other resistance mechanisms against third-generation EGFR-TKIs involving activation of signaling pathways have been reported. Lin et al (2) and Ramalingam et al (3) carried out plasma-based analyses after development of osimertinib resistance in NSCLC. The following gene alterations were noted: EGFR C797S mutation in 5.2–16.7%, KRAS mutation in 2.6–7.7%, BRAF mutation in 7.7%, PIK3CA mutation in 2.6%, MEK mutation in 2.6%, JAK2 mutation in 2.6%, MET amplification in 2.6–50%, KRAS amplification in 2.6%, ERBB2 insertion in 2.6%, transformation to SCC in 9.1%, and transformation to SqCC in 4.5% of patients (2,3). Furthermore, analysis of matched samples of pre- and post-administration of first- or second-generation EGFR-TKIs showed acquired resistance in addition to T790M mutation in lung adenocarcinoma, MET amplification in 8%, ERBB2 amplification in 5%, and EGFR amplification in 16% of patients (15). Transformation to SCC was noted in 2.6–5.0% of patients (15), and another report demonstrated SqCC rarely occurred (17).

The genetic alterations described above occurred in some pretreatment EGFR mutant NSCLC tumors and patients with these mutations can exhibit resistance to osimertinib therapy. Concurrent gene alterations in EGFR-TKI-naïve, EGFR mutant NSCLC have been reported. Yu et al, conducted next-generation sequencing analysis of tissue samples and revealed EGFR mutant NSCLC had concurrent alteration of TP53 mutation in 60%, RB1 mutation in 10%, PIK3CA mutation in 12%, CTNNB1 mutation in 18.9%, EGFR amplification in 22%, MDM2 mutation in 12%, CDK4 mutation in 10%, ERBB2 amplification in 8.4%, and MET amplification in 2% of patients before EGFR-TKI treatment (15).

Gene amplification of ERBB2, MET, MDM2 or CDK4 alone leads to poor prognosis for NSCLC regardless of EGFR gene mutation (25–27). Furthermore, Le et al (11) and Blakely et al (14) reported that cell cycle gene alteration, such as CDK4 or CDK6, shortens PFS following osimertinib therapy. Patients with EGFR mutant NSCLC accompanied by TP53 mutation, ERBB2 amplification, or MET amplification before first- or second-generation EGFR-TKI treatment exhibited a shorter time to progression and also showed resistance to third-generation EGFR-TKIs (12,13,28).

In this study, MDM2 gene amplification was shown to be associated with shorter PFS for first- or second-generation EGFR-TKIs. MDM2 is a proto-oncogene that is often coexpressed with CDK4 in liposarcoma (29). MDM2 binds to TP53 to downregulate transcription, leading to proteasome degradation by ubiquitination (30).

The limitations of this study are as follows: First, we did not evaluate EGFR mutations associated with acquired resistance to EGFR-TKIs other than C797S (e.g., L792X, G796X and L718Q), exon 20 insertion, mutation in driver genes KRAS, BRAF and PIK3CA, or transformation to mesenchyme or other pathological subtypes. Furthermore, the study cohort was limited to NSCLC patients harboring known EGFR gene mutations. Therefore, whether C797S mutation alone could be an oncogenic mechanism remains unclear. Second, this study did not evaluate single nucleotide polymorphisms or germline gene mutations. As only seven patients had EGFR T790M mutation, they could have germline mutations (7) rather than somatic mutations.

Finally, because our survival analysis was conducted in a small cohort, it is possible that amplification of genes other than MDM2, such as MDM4 and ERBB2 evaluated in this study, could be biomarkers for EGFR-mutant NSCLC.

In conclusion, no concurrent EGFR-C797S with known EGFR mutant NSCLC was identified in our study cohort. This result confirms the efficacy of third-generation EGFR-TKIs; however, it is important to remain aware of how genetic alterations can affect EGFR-TKI responses. Therefore, further studies in a large cohort are required to completely elucidate resistance mechanisms against third-generation EGFR-TKIs.

Acknowledgements

The authors would like to thank Ms. Mie Otsuki, Ms. Eiko Otomo and Ms. Yukiko Kikuta (Department of Chest Surgery, Fukushima Medical University) for their technical supports

Funding

This work was supported by grant from AstraZeneca. K.K. (Osaka, Japan; grant no. ESR-16-12328).

Availability of data and materials

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Authors' contributions

TY, SM and HS designed the study. TY, HM, HT, MW, YO, TI, MF, NO, YM, TH, JO, MHo, MHi and YS collected and analyzed the data. TY and HS wrote and revised the manuscript. All the authors approved the final manuscript.

Ethics approval and consent to participate

The current study was conducted with approval of the ethics board at Fukushima Medical University (approval no. 2955). Human rights and welfare of participants were protected in accordance with the Declaration of Helsinki, and written informed consent was obtained from participants.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

References