This article is mentioned in:

Introduction

Somatic mutations in the tyrosine kinase (TK) domain of epidermal growth factor receptor (EGFR) gene have been reported in patients with non-small-cell lung cancer (NSCLC). Certain mutations in the EGFR gene, such as leucine-to-arginine substitution at amino acid position 858 (L858R) in exon 21 or deletions in exon 19, are highly correlated with sensitivity to EGFR-TK inhibitors (TKIs) (1, 2). The EGFR-TKIs gefitinib and erlotinib are effective in the treatment of EGFR-mutant NSCLC; however, there are also cases of EGFR-mutant NSCLCs exhibiting resistance to EGFR-TKI treatment. EGFR-TKIs have been shown to achieve a response in ∼ 80% NSCLC patients with EGFR mutations, indicating that ∼ 20% of NSCLC patients with EGFR mutation are unresponsive to this treatment (3).

A threonine-to-methionine substitution at amino acid position 790 (T790M) in exon 20 was reportedly associated with acquired resistance to EGFR-TKIs (4, 5). In addition, hepatocyte growth factor receptor (HGFR; MET) gene amplification was reportedly associated with acquired resistance to EGFR-TKIs (6), while hepatocyte growth factor (HGF)-mediated MET activation was reported as the mechanism underlying EGFR-TKI resistance in lung cancer with EGFR-activating mutations (7). However, these studies were not pertaining to resistance, but rather investigating acquired resistance to EGFR-TKIs.

It was recently reported that pretreatment of NSCLC with T790M shortens the duration of response to EGFR-TKIs (9–12). However, over the last few years, we have observed long progression-free survival (PFS) in patients with T790M.

In this study, we aimed to investigate the pattern of EGFR mutations in NSCLC that affects sensitivity to EGFR-TKIs, determine the cause of shortened EGFR-TKI response duration and determine the correlation between resistance to EGFR-TKIs and phosphorylated MET or HGF expression.

Materials and methods

Patients and specimens

We investigated 44 surgically resected NSCLCs between 2001 and 2012. The specimens were obtained from patients treated with gefitinib as 1st-, 2nd-, or 3rd-line therapy for postoperative recurrent NSCLC.

The NSCLCs were histologically classified as 43 adenocarcinomas and 1 large-cell neuroendocrine carcinoma. The patients included 19 men and 25 women, aged 27–78 years (mean age, 63.0 years).

Immunostaining

We detected EGFR mutations in matching formalin-fixed, paraffin-embedded tissue samples using the cycleave PCR technique (SRL Inc., Tokyo, Japan). We used an anti-MET rabbit monoclonal antibody (clone SP44; cat no. 518-108830; Ventana Medical Systems, Inc., Tucson, AZ, USA) for MET staining and a goat polyclonal anti-human HGF antibody (cat no. 36073; LifeSpan BioSciences, Inc., Seattle, WA, USA) at a 1:40 dilution for HGF staining. Immunostaining for MET and HGF was performed using the Ventana System (Ventana Medical Systems, Inc, Harvard, MA, USA).

Type of response to gefitinib

We assessed the progression of non-responders to gefitinib treatment over a 2-month period. Intermediate responders included patients who responded (exhibiting at least stable disease) to gefitinib for 3–11 months. Long-term responders included patients who responded to gefitinib therapy for >12 months.

Results

EGFR mutations

The 44 NSCLC specimens included 43 adenocarcinomas and one large-cell neuroendocrine carcinoma. There was 1 patient with an exon 18 point mutation, 23 with an exon 19 deletion, 2 with an exon 20 point mutation, 16 with an exon 21 point mutation and 2 with both exon 20 and 21 point mutations (Table I).

Table I. Patient characteristics and response to gefitinib.

Table I.

Patient characteristics and response to gefitinib.

Patients	Gender	Age (years)	Histological type	Exon	Lobectomy	Responsea
1	M	46	AdenoCa	18	Total	2
2	F	73	AdenoCa	19	Total	2
3	F	59	AdenoCa	19	Total	3
4	F	71	AdenoCa	19	Total	3
5	F	54	AdenoCa	19	Total	1
6	M	63	AdenoCa	19	Total	2
7	F	59	AdenoCa	19	Total	2
8	M	78	AdenoCa	19	Partial	2
9	F	68	AdenoCa	19	Partial	2
10	F	71	AdenoCa	19	Total	2
11	F	73	AdenoCa	19	Total	2
12	M	66	AdenoCa	19	Total	2
13	F	51	AdenoCa	19	Total	2
14	M	61	AdenoCa	19	Total	2
15	M	47	AdenoCa	19	Total	2
16	M	60	AdenoCa	19	Total	3
17	F	54	AdenoCa	19	Total	3
18	M	79	AdenoCa	19	Total	2
19	F	27	AdenoCa	19	Total	2
20	M	75	AdenoCa	19	Total	2
21	F	61	AdenoCa	19	Total	3
22	F	77	AdenoCa	19	Total	3
23	F	56	AdenoCa	19	Total	3
24	F	55	AdenoCa	19	Total	3
25	F	55	AdenoCa	20	Total	1
26	M	72	AdenoCa	20	Total	1
27	F	71	AdenoCa	21	Total	2
28	F	66	AdenoCa	21	Total	3
29	M	69	AdenoCa	20,21	Total	3
30	M	64	AdenoCa	20,21	Total	3
31	M	72	AdenoCa	21	Total	3
32	F	60	AdenoCa	21	Total	1
33	F	78	Large-cellCa	21	Total	2
34	M	57	AdenoCa	21	Total	2
35	M	78	AdenoCa	21	Total	2
36	F	76	AdenoCa	21	Total	2
37	M	65	AdenoCa	21	Total	3
38	F	60	AdenoCa	21	Total	2
39	M	39	AdenoCa	21	Total	2
40	F	57	AdenoCa	21	Total	2
41	M	70	AdenoCa	21	Total	2
42	M	42	AdenoCa	21	Total	3
43	F	62	AdenoCa	21	Total	3
44	F	73	AdenoCa	21	Total	2

a 1, No response; 2, i ntermediate respon se; 3, l ong-term respon se . M, male; F, female; Ca, carcinoma.

Association of EGFR mutations with response to gefitinib

There were 4 non-responders, including the 2 patients with exon 20 mutation, 25 intermediate responders (including 10 patients under ongoing treatment) and 15 long-term responders (2 of whom are under ongoing treatment), including the 2 patients with both exon 20 and 21 mutations (Table II).

Table II. Type of EGFR mutation and response to gefitinib.

Table II.

Type of EGFR mutation and response to gefitinib.

	Type of mutation
Type of response	Exon 18	Exon 19	Exon 20	Exon 21
Non-responders	0	1	2	1
Intermediate responders	1	14	0	10
Long-term responders	0	8	2a	7a

a Including 2 patients with both exon 20 and 21 mutations. EGFR, epidermal growth factro receptor.

Immunostaining results

We investigated MET and HGF immunostaining in 4 non-responders, 3 of whom were MET-positive and HGF-negative, whereas 1 patient was MET-negative and HGF-positive (Fig. 1).

Figure 1. Immunohistochemical staining of non-small-cell lung cancer specimens. (A) The immunohistochemical staining for hepatocyte growth factor receptor (HGFR; MET) exhibited strong reactivity in the cell membranes of specimens no. 5, 25 and 26, whereas there was no reactivity with MET in specimen no. 32. (B) The immunohistochemical staining for HGF exhibited strong reactivity in the cell membranes of specimen no. 32, whereas there was no reactivity with HGF in specimens no. 5, 25, and 26.

Discussion

Previous studies reported that the causes of acquired resistance to EGFR-TKIs in patients with EGFR mutations are a second mutation (T790M), MET amplification, or HGF-mediated MET activation. In those studies, ∼ 50% of the cases with resistance to EGFR-TKIs exhibited a second mutation and ∼ 20% were due to MET amplification (8). Our results were similar to those of previous studies, where EGFR-TKI therapy was the initial treatment. However, in our study, patients with NSCLC and exon 20 mutation responded to gefitinib in the presence of an additional EGFR mutation. In particular, 2 cases (29 and 30) in this study were treated with gefitinib for 14 and 21 months, respectively. Our results were better in terms of PFS compared to those previously reported (2–13 months) (9–12).

Inukai et al reported that a small fraction of T790M-positive tumor cells at the beginning of treatment may lead to clinical gefitinib resistance as a result of the selective proliferation of T790M mutant cells (9). We therefore considered that the growth speed of T790M -positive cells and the number of T790M cells prior to EGFR-TKI treatment regulation were important for predicting PFS in patients with NSCLC and EGFR mutations. We considered that the T790M cell number was more important, rather than the T790M cell growth speed, as the latter is low (13). However, there is no established clinical method to quantitatively measure the number of T790M cells. Therefore, we must make a prediction based on the sensitivity of EGFR mutation testing in patients with NSCLC. The sensitivity of direct sequencing was previously found to be ∼ 25%, that of cycleave PCR was ∼ 5% and that of Scorpion ARMS was 1% (14).

The PFS of NSCLC patients, as assessed by direct sequencing in a study by Wu et al was 2 months (11). However, the PFS of NSCLC patients assessed using cycleave PCR in our study was 17.5 months. We attributed the longer PFS in our study to the detection of fewer T790M cells in our patients using more sensitive cycleave PCR prior to EGFR-TKI treatment.

Consequently, our findings suggest that NSCLC patients may be long-term responders if a double mutation is identified using a highly sensitive method, such as cycleave PCR or Scorpion ARMS.

Our data and previous reports taken together, indicate that NSCLC with exon 20 mutation will respond to gefitinib treatment in the presence of an additional EGFR mutation. However, further investigations are required to determine the mechanism underlying our findings.

References