High frequency of epidermal growth factor receptor mutations in non small cell lung cancers in vietnam

Methods: 181 paraffin -embedded tumor tissues from grade IIIB/IV NSCLC patients were screened for mutations in exons 18 to 21 of the EGFR gene, using both the direct sequencing and the S

HIGH FREQUENCY OF EPIDERMAL GROWTH FACTOR

RECEPTOR MUTATIONS IN NON - SMALL CELL LUNG CANCERS

IN VIETNAM Nguyen Ha Minh 1 , Tran Huy Thinh 1,2 , Tran Van Khanh 2

1 Dept of Biochemistry, Hanoi Medical University 2

Center for Gene & Protein Research, Hanoi Medical University

Mutations in the tyrosine kinase (TK) domain of the epidermal growth factor receptor (EGFR) gene impact a patient’s response to targeted therapy in non - small cell lung cancer (NSCLC) Detection of these mutations plays an important role in therapeutic decision making in NSCLC Methods: 181 paraffin -embedded tumor tissues from grade IIIB/IV NSCLC patients were screened for mutations in exons 18 to 21

of the EGFR gene, using both the direct sequencing and the Scorpion ARMS method Mutations in the kinase domain (exon 18 to exon 21) of the EGFR gene were identified in 106 of the 181 NSCLC specimens (58.6%) The following mutations were the two common mutation types found: in-frame deletions in exon 19 (found in 48.1% of the patients in our study with EGFR mutations) and single missense mutations L858R in exon 21 (found in 40.6% of the patients in our study with EGFR mutations) Two tumors carried double mutations and four tumors showed new mutations The EGFR TK mutations were more frequent in never smokers than ever smokers (74.4% versus 44.2%), in females than males (75.0% versus 45.5%) and in adenocarcinoma NSCLC versus other histologic subtypes (65.9% versus 20.7%) In conclusion: EGFR TK mutations can be found in many patients with advanced NSCLC in Vietnam Direct sequencing and the Scorpion ARMS method could be combined to screen for these mutations.

Keywords: Epidermal growth factor receptor, Non-small cell lung cancer, EGFR mutation

Corresponding author: Tran Van Khanh, Center for Gene

and Protein Research, Hanoi Medical University

E-mail: tranvankhanh@hmu.edu.vn

Received: 02 November 2016

Accepted: 10 December 2016

I INTRODUCTION

Despite improvements in diagnostic and

therapeutic approaches to treatment, lung

cancer remains the major cause of

cancer-related deaths worldwide [1] Non - small cell

lung cancer (NSCLC), the major form of lung

cancer, is classified into three histologic

sub-types: adenocarcinoma, squamous cell

carci-noma and large cell carcicarci-noma NSCLC is

also characterized by the accumulation of

mul-tiple genetic alterations, including those that

result in the activation of oncogenes and the

inactivation of tumor suppressor genes [2; 3] Epidermal growth factor receptor (EGFR) is a tyrosine kinase receptor Activation of the downstream pathways of EGFR leads to cell proliferation, differentiation, migration and ad-hesion, protection from apoptosis, enhanced survival, and gene transcription [4] Targeting the EGFR gene is an appealing strategy for the treatment of NSCLC, as the EGFR gene has been found to be expressed, sometimes strongly, in NSCLC tumors [5] Researchers have found that EGFR mutations are strong determinants of tumor response to EGFR tyrosine kinase (EGFR TK) inhibitors [5 – 8] These mutations cluster in a wide spectrum from exon 18 to exon 21 within the EGFR gene Numerous mutations, including deletion,

insertion, duplication and substitution

mutations, are classified into EGFR TK

inhibitor-sensitizing or - resistant mutations

The two most common NSCLC-associated

EGFR mutations are the 15 - bp nucleotide

in-frame deletion in exon 19 (E746_A750del)

and the point mutation that replaces leucine

with arginine at codon 858 in exon 21 (L858R)

[5; 9] Previous research has shown that rates

of EGFR TK mutations among NSCLC

patients varies according to patients’ ethnic

backgrounds Asian NSCLC patients have

been found to carry higher numbers of

muta-tions than those in Europe and the United

Sta-tes In addition, several clinical characteristics

have been found to be associated with overall

better clinical response, i.e adenocarcinoma,

female gender and a history of non - smoking

[7; 10]

As efforts to identify clinically relevant

mutations intensify, mutation testing of tumors

will become more routine Direct sequencing is

the classic method for the detection of

genomic mutations and is still widely

em-ployed to uncover ‘‘new’’ mutations However,

this technique has several drawbacks in the

clinical setting, where the focus is on the

detection of known recurrent mutations that

have clinical relevance Direct sequencing

involves multiple steps, so several days are

required to obtain a result after tissue

acquisi-tion More importantly, the sensitivity of direct

sequencing is suboptimal for clinical tumor

samples Mutant DNA needs to comprise over

25% of the total DNA to be easily detected

Thus, mutation detection by direct sequencing

could lead to ‘‘false - negative’’ results

An-other method to detect genomic mutations is

the Scorpion - ARMS method Scorpion

prim-ers are used in a fluorescence - based way for the specific detection of PCR products [11] Scorpion primers can be used in combination with the Amplified Refractory Mutation System (ARMS) to enable the detection of single -base mutations and to enhance specificity [11; 12] The Scorpion - ARMS method is highly sensitive and fast when it comes to the detection of known mutations [13] This study sought to investigate the occurrence of the EGFR TK mutation in Vietnamese NSCLC patients by simultaneously employing the direct sequencing and the Scorpion ARMS methods

We also sought to clarify the relationship between EGFR mutation status and the clinical characteristics of NSCLC in Vietnamese patients

II SUBJECTS AND METHODS

1 Subjects

Tumor tissues and DNA extraction

In total, 181 tumor tissue biopsies were taken from the Departments of Pathology in five hospitals across Northern Vietnam: the National Cancer Hospital, the Hanoi Cancer Hospital, Bach Mai Hospital, Huu Nghi Hospital and the Hanoi Medical University Hospital These tissue biopsies were formalin-fixed and paraffin - embedded Patient data was collected from medical records, and included: smoking status, age, pathological diagnosis, and clinical stage This data was reviewed by oncologists at the hospitals where tumor tissue biopsies were obtained

All samples underwent a haematoxylin and eosin pathology review to confirm the diagno-sis of NSCLC and the presence of tumor in the samples The NSCLC samples were

macro-dissected by scraping only the tumor area that

had been selected by a pathologist Genomic

DNA was extracted and purified from 20μm

thin sections using the phenol/Chloroform

method [14] The concentration and purity of

the extracted DNA were determined by

Nano-Drop ND - 1000 (NanoNano-Drop Technologies,

Rockland, DE) The extracted DNA was

stocked at -20oC until use

2 Methods

PCR amplification and direct

sequenc-ing

Exons 18 to 21 of the EGFR gene were

amplified using four specific pairs of primers,

and the PCR amplicons were subjected to

direct sequencing (the sequences of the

primers and the amplification procedures are

shown in Supplemental Tables 1 and 2)

Sequencing reactions were done with the

same primers for PCR amplification and ABI

BigDye Terminator kit v3.1 (Applied

Biosys-tems, Foster City, CA), according to the

manu-facturer’s instructions Sequencing reactions

were electrophoresed on an ABI3700 genetic

analyzer Sequence variations were

determi-ned using Seqscape software (Applied

Biosys-tems) with the EGFR reference sequence

(NM_005228.3, National Center for

Biotechno-logy Information)

Scorpion ARMS analysis

An EGFR Scorpion Kit (Qiagen

Manches-ter Ltd., UK), which combines the ARMS and

Scorpion methods to detect 29 EGFR TK

mu-tations in real-time PCR reactions, was used

to conduct the Scorpion-ARMS analysis The

scorpion primers were designed and

synthe-sized by Qiagen All reactions were done in 25

-µL volumes using 5 µL of template DNA, 19.5

µL of primer mix and 0.5 µL of Taq poly-merase Real-time PCR was carried out using Stratagen MX 3000P (Life Technologies, Thermo Fisher Scientific Inc.) under the follow-ing conditions: initial denaturation at 95oC for 4 minutes, 40 cycles of 95oC for 30 seconds, and finally 60 seconds at 60oC with fluores-cence reading (set to FAM (fluorescein) that allows optical excitation at 480 nm and meas-urement at 520 nm) at the end of each cycle The cycle threshold (Ct) was defined as the cycle at the highest peak of the second deriva-tive curve, which represented the point of maximum curvature of the growth curve The

∆Ct value was calculated as ∆Ct = sample Ct – control Ct, after ensuring that the sample and control Ct values were from the same sample The ∆Ct value was then compared with the cut-off value of the mutation, which was obtained from the rQiagen’s instructions

A positive result was defined as a ∆Ct value less than or equal to the cut-off value

Statistical analysis

Chi-square tests and Fisher’s exact tests (when there were fewer than five expected counts in the contingency table) were used to assess the relationship between the presence

of EGFR mutations in patients with NSCLC and other patient characteristics, including age, gender, tumor histology and smoking status All statistical tests were two sided and

a p value of less than 05 was considered

sta-tistically significant Statistical analyses were done using the SPSS software package ver-sion 21.0

3 Research ethics

This research was approved by the ethics committee of Hanoi Medical University, decree

No.161 - HMUIRB, signed February 15th, 2014.

III RESULTS

1 Extracted DNA from paraffin -

embed-ded tissues

181 patients were enrolled between

Janu-ary 02, 2012 and December 30, 2013 Tumor

tissues were collected from all participants

before the initiation of gefitinib or erlotinib

treatment Genomic DNA was extracted in all

181 samples at a range of purity from 1.7 to

2.0 and a concentration from 1,213 to 4,380

ng/µL

2 The relationship between EGFR

muta-tions and patients’ clinical characteristics

The 181 NSCLC tumor specimens included

152 adenocarcinomas (84.0%), 21 squamous

cell carcinomas (11.6%), and 8 large cell

car-cinomas (4.4%) 101 patients were male

pa-tients (55.8%) and 80 were female papa-tients

(44.2%) The median age of all patients was

62 years old (range from 36 to 82 years old),

with 61.3% of participants (111 of 181) being

under 65 years old In terms of smoking

his-tory, 47.5% patients (86 of 181) were never

smokers

Mutations in the kinase domain (exon 18 to

exon 21) of the EGFR gene were identified in

106 tumor tissues, giving an overall mutation

rate of 58.6% (106 of 181) Adenocarcinomas

showed mutations more often than any other

histology Moreover, 100 out of 152

adenocar-cinomas (65.9%) showed mutations, versus

six out of 29 samples (20.7%) for all other

samples (p = 004) When all histologies were

analysed together, women also showed a

significantly higher frequency of EGFR

muta-tions compared with men Moreover, 60 out of

80 women showed EGFR mutations (75.0%) versus 46 out of 101 men (45.5%) f (p < 0001) When the subgroup of adenocar-cinomas was analysed according to gender, women continued to show a higher mutation rate (56 out of 70 women (80.0%) versus 40 out of 82 men (48.8%) (p < 0001) Smoking status was also closely associated with EGFR mutation rate Among patients with EGFR mutations (n = 106), the rate of never smokers was 1.7 times higher than the rate for ever smokers 64 out of 86 never smokers showed EGFR mutations (74.4%) versus 42 out of 95 ever smokers (44.2%) (p < 0001) This difference was also observed when smoking

adenocarcinomas 57 out of 75 never smokers

mutations (76.0%) versus 39 out of 77 ever

(p < 0001))

However, no significant difference in the rate of mutation was found with regards to age For patients over 65 years old, 44 out of

70 patients displayed EGFR mutations (62.9%), and for those patients under 65 years old, 62 out of 111 displayed EGFR mutations (55.9%) (p= 352) We also found no signifi-cant difference in the rate of mutations with regards to age among patients with adenocar-cinomas Moreover, 39 out of 56 patients over

65 years old with adenocarcinoma displayed EGFR mutations (69.6%) versus 57 out of 96 patients under 65 years old with adenocarci-noma (59.4%) (p= 206) Patient characteris-tics and EGFR mutation rates are summarized

in table 1 The relationship between EGFR mutation status and patient characteristics are graphed in figure 1

Table 1 Characteristics of patients according to EGFR mutation status

Screened (n = 181) With EGFR mutation

Gender

Age

Smoking status

Histology subtypes

With no EGFR mutation

Figure 1 EGFR mutation status and patient characteristics

EGFR mutation spectrum of direct

sequencing compared to Scorpion ARMS

Exon 18 to exon 21 of the EGFR gene

were screened in all 181 tumors Mutations

were found in all four exons, but mainly

clustered in exons 19 and 21 (Fig 2) In exon

19, the only mutation found was a 15 -

nucleo-tide deletion from nucleonucleo-tides 2481 to 2495,

which resulted in elimination of codons 746 to

750 This type of in - frame deletion, called an

LREA deletion, was found in 51 of 106 tumors

with EGFR mutations (48.1%)

In exon 21, two types of missense

muta-tions were observed A c.2819T > G mutation

resulted in a change to L858R in 43 tumors

(40.6%) The other missense mutation was

L861Q, which occurred in 4 tumors In

addi-tion, one tumor contained a new 23 -

nucleo-tide deletion, spanning from nucleonucleo-tide 2499

to 2521 (c.2499_2521del mutation); and one tumor carried a 3-nucleotide ACA insertion at the position between nucleotide 2554 and

2555, leading to a threonine insertion into the position between valine (V851) and lysine (K852) (V851_K852insT mutation) (Fig 3)

In exon 18, the two forms of mutation seen were missense mutation G719S (c.2155G>A) and the small deletion c.2137delA Each of these mutations was found once In exon 20, three missense mutations and one small deletion were found A concurrent S768I and V769L mutation was found in one tumor The T790M mutation occurred in two tumors, one

of which had an L858R mutation in exon 21 Deletion of two adenines at the position of nucleotide 2373 caused the non-in-frame deletion c.2373_2374delAA (Fig 3)

Figure 2 Incidence of EGFR mutations in 181 NSCLC patients

Although most tumors had a single mutation in the kinase domain of the EGFR gene, a combi-nation of two types of mutation were seen in two patients, one of which contained the T790M (exon 20) and L858R (exon 21) mutations and the other of which carried the S768I and V769L mutations (both in exon 20)

The EGFR mutation frequency detected by

direct sequencing and the Scorpion ARMS

methods were 44.2% (80 out of 181 patients)

and 55.8% (101 out of 181 patients),

respec-tively The Scorpion ARMS method found

more tumors containing LREA deletion

muta-tions and single missense mutamuta-tions than the

direct sequencing method Moreover, the

Scorpion - ARMS method found 51 tumors

containing LREA deletion mutations, versus

45 tumors with LREA deletion mutations found

by the direct sequencing method; the Scorpion

-ARMS method also discovered 49 tumors

with single missense mutations, compared

with 29 tumors with single missense mutations

discovered via direct sequencing

Direct sequencing, however, discovered four novel or unknown mutations from four tumors, including c.2137delA in exon 18,

c.2554/2555insACA (V851_K852insT) in exon

21 In addition, the direct sequencing method detected four mutations in two double-missense mutations, while the Scorpion-ARMS found only three Thus, the combina-tion of the two methods elevated the overall mutation rate of this study to 58.6% (106 out

of 181 patients) The frequency and the mo-lecular features of EGFR mutations detected

by direct sequencing and the Scorpion ARMS method are shown in table 2

Figure 3 Novel mutations in the EGFR gene in non - small cell lung cancer

(A) A deletion of one adenine at the position of nucleotide 2137 in exon 18 caused a c.2137delA mutation (B) A deletion of two adenines at the position between nucleotides 2373 and 2374 in exon 20 led to a c.2373_2374delAA mutation (C) An insertion of three nucleotides, Adenine - Cytosine - Adenine, at the position between nucleotides 2554 and 2555 created a c.2554/2555insACA mutation This is an in - frame insertion mutation with amino acid change V851_K852insT (D) A deletion of 23 nucleotides at the position between nucleotides 2499 and

2521 in exon 19 caused a c.2499_2521del23 mutation The mutations in (A), (B) and (D) resulted

in alterations in encoded amino acids The arrow shows the position of the mutation.

Table 2 Molecular features of 106 lung cancer patients with EGFR mutations

Sequencing

Scor-pion ARMS

Both Site, nucleotide

change

Amino acid change

Site, nucleotide change

Amino acid change

Exon 19,

Exon 20,

Exon 21,

Exon 21,

2554/2555insACA

V851_K852ins

IV DISCUSSION

The current study reports the results of a

molecular analysis of the EGFR gene in 181

Vietnamese NSCLC patients before the

initiation of gefitinib or erlotinib therapy

Among the 181 patients, 106 patients (58.6%)

were found to carry EGFR mutations in their

NSCLC tumors The rate of EGFR mutations

found in this study was much higher than that

previously reported for Caucasian NSCLC

patients [7; 15; 16] but only somewhat higher

than that observed in East Asian patients The

frequency of EGFR mutations was reported to

be 39% in Japan, 34.6% in Korea, 38.6% in

Taiwan and 57.4% in Thailand [17 - 20] One

possible explanation for the higher mutation

rate in this study lies in the distribution of

histological subgroups While the expected

proportion of adenocarcinomas in routine

studies is between 50% to 70% depending on

literature, 84% of the participants in this

study had adenocarcinoma In a manner

con-sistent with previous reports, the EGFR

muta-tion was tightly associated with this pathologic

subtype Since the adenocarcinomas included

in this study were more likely to exhibit an

frequency of the EGFR mutation may have

been overestimated Similarly high mutation

rates have been observed in previous studies

adenocarcinoma [21; 22]

Another factor with great influence on the

EGFR mutation rate was the geographical

origin of patients included in this study For

quite some time, it has been known that East

Asian patients with NSCLC have a higher

inci-dence of EGFR mutations compared with

other ethnic groups [17 – 22] Since we only recruited Vietnamese patients for this study, the mutation rate here was both higher than that reported in studies with non - Asian pa-tients and also differed from frequencies re-ported in studies conducted in other Asian countries

The method that we used to determine the EGFR mutation rate is also an important factor impacting our results In this study, the rate of mutations in tumor DNA detected by the Scor-pion-ARMS was compared with that deter-mined by the direct sequencing method, which

is the current standard DNA from tumor sam-ples always consisted of a mixture of mutant DNA and wild - type DNA because the EGFR mutation status was heterogeneous in all sam-ples and the complete removal of normal cells, such as normal epithelial cells and inflamma-tory cells, from tumor specimens is difficult Since parallel tumor tissue investigations were done on all specimens using both direct sequencing and the Scorpion - ARMS method

- a recognized advantage in the present study

- the maximum number of EGFR mutations were detected (106 cases in total) This is in contrast to the 101 cases that would have been detected using Scorpion - ARMS alone,

or the 80 cases that would have been detected using direct sequencing alone

Elli-son et al reported an overall EGFR mutation

rate of 12.6%, higher than the single Scorpion ARMS method or sequencing (8.4% and 7.9%, respectively) [23] We can speculate that the high sensitivity and specificity of the EGFR Scorpion kit allowed us to detect more mutations and to confirm uncertain results that

we obtained from direct sequencing alone Figure 4 shows an example of how the Scorpion ARMS method was used to confirm

uncertain results Although Shi et al, in the

PIONEER study, reported a high frequency of

EGFR mutations in seven Asian countries

using the Scorpion-ARMS method, they did

not discover any new or unknown mutations

[22] In this study, four novel mutations

(deletion and insertion mutations) and one

V769L mutation were determined by using

direct sequencing alone This is the first time

in Vietnam that an EGFR mutation analysis

was carried out by both direct sequencing and

the Scorpion-ARMS method One limitation of

the EGFR Scorpion kit is that while it is able to

detect mutations targeted by the designed

Scorpion primers, it misses those EGFR

mutations that are not at these unique sites

but are instead clustered around the

ATP-binding site in exon 18, 19, 20 and 21 [5 – 9]

In addition, the likelihood of discovering an

EGFR mutation depends on various technical

factors such as the tissue extraction method,

the qualification of the tumor tissue, and the

DNA purification procedure

In this study, we found that NSCLC

pa-tients with EGFR mutations were more likely

to be non - smokers, women, and/or have

tumors of the adenocarcinoma subtype These

numerous previous studies [7, 24] The high

mutation rates in these subgroups suggest

that it is possible to screen for potential

NSCLC patients who are likely to carry EGFR

mutation(s) Such screening could play an

important role in selecting patients for targeted

therapeutic decision-making in NSCLC

Independent of the total rate of mutations,

the frequency of EGFR mutations in exon 19

(51 out of 106 patients with EGFR mutations,

48.1%) and exon 21 (49 out of 106 patients

with EGFR mutations, 46.3%) found in this study are comparable with rates found in of numerous previous studies [19; 21; 23; 25] The LREA mutation in exon 19 occurred at a higher rate than the L858R mutation in exon

21, corresponding to what past research has shown [5] We also discovered new mutations, including deletions in exons 18, 20 and 21 and

an insertion in exon 21 (listed in Table 2) These genetic changes have not been reported

in the literature Thus, their roles for EGFR tyrosine kinase inhibitory therapy are unclear Moreover, the new mutations that we found

in this study are complex Two tumors carried double mutations, including both T790M + L858R and S768I+V769L This finding would suggest that the cancer cells in these patients' tumor specimens might be accumulating multi-ple molecular mutations in the EGFR gene In terms of the impact of EGFR mutations on the responsiveness of tyrosine kinase inhibitors, sensitizing mutations were more frequent than resistant mutations in this study (99 sensitizing mutations out of 106 tumors, 93.4%, versus 3 resistant mutations out of 106, 2.8%), which is similar to what has been found in previous studies [5; 8] Two of three tumors containing resistant mutations carried double mutations T790M, S768I and V769L were reported as resistant mutations, while L858R is classified

as a sensitizing mutation Although EGFR double mutations have been reported in previous studies, where the frequency has varied from 13% to 17% [19; 26], they have not previously been studied in Vietnamese NSCLC patients [22] T790M + L858R is a complex double mutation that leads to a combination of sensitization and resistance In

a study by Tam et al., the in vitro gefitinib