The Prevalence and Concurrent Pathogenic Mutations of KRAS G12C in Northeast Chinese Non-small-cell Lung Cancer Patients.

OBJECTIVE: KRAS mutation is one of important driver genes in non-small-cell lung cancer (NSCLC) and the patients with KRAS G12C mutations benefit from the inhibitor AMG510. However, the frequency, concurrent pathogenic mutations, and clinical characteristic of KRAS G12C is unknown in the NSCLC population of Northeast China.
METHODS: The retrospective analysis was derived from 431 NSCLC patients in Jilin Cancer Hospital between January 2018 and June 2019. The mutation frequency and concurrent mutations of KRAS G12C in tumor or peripheral blood was detected by next-generation sequencing (NGS).
RESULTS: The RAS mutant rate was observed in 10.7% (46/431) of this cohort. All RAS-driver cancers are caused by mutations in the KRAS isoform, while the NRAS and HRAS isoforms were not detected. Among KRAS-mutant patients, 42 (91.3%) showed exon 2 mutation in 12 codon and 13 codon. KRAS G12C showed a 4.6% (20/431) mutation rate in this cohort and the highest frequency (43.5%, 20/46) in KRAS-mutant-positive patients. There was no difference between tumor tissue and plasma in terms of either KRAS or KRAS G12C mutation. The most frequent co-occurrence mutations with KRAS G12C were TP53, followed by PTEN. Furthermore, KRAS G12C was exclusive with STK11 mutation. KRAS G12C mutation was associated with age, disease stage, and smoking status (P=0.024; P=0.02; P=0.006), smoking remained an independent factor for KRAS G12C mutation (P=0.037), and higher mutation frequency in patients older than 60, stage I-III, or smoking in NSCLC (P=0.0151, P=0.0343, P=0.0046, respectively).
CONCLUSION: KRAS mutation was the only isoforms of RAS family, of these 43.5% harbored the KRAS G12C subtype in northeastern Chinese NSCLC patients. KRAS G12C is associated with age, pathological stage and smoking status, more commonly harbored TP53/PTEN mutations, and providing more genome profile for targeted therapy in local clinical practice.

Introduction

Non-small-cell lung cancer (NSCLC) is the most common histological type of lung cancer, accounting for 80–85% of lung cancers and has become the most fatal cancer in the world.1 Recently, targeted therapy based on various driver oncogene variants (EGFR, ALK and ROS1, KRAS, MET, PIK3CA, RET, BRAF) has shown great antitumor activity; unfortunately, KRAS mutations had a more complicated mechanism in comparison with other driver genes such as EGFR, with poor prognosis and high risk of tumor recurrence.2 Although prevalent, no specific treatment has been successfully developed for these NSCLCs.

KRAS mutations are some of the most prevalent alterations, approximately 10% of Asian NSCLC patients and 7.5% of Chinese NSCLC patients harbor the KRAS mutation, with codon 12 and 13 mutations being the most frequent and the most common subtypes are G12C, G12V and G12D.3,4 KRAS is a mutant type of KRAS guanosine triphosphatase (GTPase), and an inhibitor targeting KRAS is a promising novel tumor-specific therapy for tumors driven by mutant proteins.5 Current studies on KRAS inhibitors and the mechanism of drug resistance have confirmed that patients with KRAS mutations benefit from the inhibitor AMG510,6 which has also been approved by the FDA as an orphan drug for NSCLC and colon cancer with KRAS mutation. KRAS can induce allosteric switch II pocket (s-iip) and take cys-12 as the specific covalent target of alleles, which were considered as potential drug targets.2 Now, KRAS mutation was verified by the NGS, various clinical parameters and genetic mutation have been proposed to predict the relevance with KRAS (such as sex, age, smoking, co-mutation gene). In the current study we aim to discover a more precise delineation of candidate target populations and distinctive KRAS co-mutation subtypes in the northeast Chinese population. We retrospectively investigated and evaluated the KRAS mutation in northeast Chinese NSCLC, and the association between clinical factors and KRAS mutation status.

Materials and Methods

Patients and Samples

Four hundred and thirty-one samples were collected from Jilin Cancer Hospital between January 2018 and June 2019, 268 cases were tested through eight gene panel, 81 cases by 168 gene panel and 82 matched cases using 520 gene panel, respectively (Figure 1). Clinic pathological data were collected from the electronic medical records in Jilin Cancer Hospital, and the factors included age, sex, and clinical stage, smoking history, brain metastasis, PS score and histology. All participants signed the informed consent agreement before participating in the study, the data were anonymized, the study was approved by the Clinical Research Ethics Committee of Jilin Cancer Hospital and was conducted in accordance with the Declaration of Helsinki.

Figure 1

Study flowchart.

DNA Extraction

DNA was extracted by DNA FFPE tissue kit (AmoyDx, China) and ctDNA extraction kit (QIAGEN, Germany) according to the manufacturer’s instructions. DNA concentration was quantified by Nanodrop 3000C and Qubit 4.0 (Thermo Fisher Scientific, Waltham, MA, USA).

Next-generation Sequencing Analysis

Library preparation was performed following manufacturer’s protocol (Burning Rock Biotech, Guangzhou, China). DNA Fragments (range: 200–400 bp) were purified by AMPure beads (Beckman Coulter, CA, USA), and captured with probe baits, hybrid selection with magnetic beads by RT-PCR amplification. Subsequently, DNA quality and size were assessed by high-sensitivity DNA assay. Indexed samples were sequenced on a MiSeq system (Beckman Coulter) with paired-end reads. The input of extracted DNA should be in the range of (30−200 ng). Sequencing platform was used by Illumina NextSeq 500 Sequencing Platform with tissue DNA (1000X) and cfDNA (20000X). All samples were analyzed by NGS targeted panel (Burning Rock Dx, China), which eight-gene panel covers well-known lung adenocarcinoma driver genes, 168 genes covers known lung cancer-related genes and 520 genes covers solid tumor-related genes. ().

Statistical Analysis

All data was performed by SPSS Statistics 19.0 software (IBM Corporation, Armonk, NY, USA). Fisher’s exact test was used to evaluate mutation differences and clinical factor between KRAS and KRAS. Logistic regression analysis was used to identify as independent factors for KRAS mutations. A P-value of <0.05 was considered statistically significant.

Results

Patient Population

Among 431 samples were those from tumor tissue 332 (77.04%), 99 (22.96%) plasma; 198 women (54.07%) and 233 men (45.93%), with a median age of 63 years (range: 34–86 years), respectively. Of the 431 patients, 263 (61.02%) were smokers, and 168 were nonsmokers. The histological characterization of tumors revealed that 370 samples were adenocarcinoma (85.85%), 61 were squamous cell carcinoma (14.15%). Of the 431 patients, characterization of the pathological stage showed 115 samples in stage I–III (26.68%), and 316 samples in stage IV (73.32%) (Table 1).

Table 1

Patient Characteristics

Characteristics	n (%)
Age (years)	63 (34–86)
Sex
Male	198 (45.93)
Female	233 (54.07)
Stage
I–III	115 (26.68)
IV	316 (73.32)
Smoking history
Yes	168 (38.98)
No	263 (61.02)
Brain metastasis
Yes	106 (24.59)
No	325 (75.41)
PS score
0–1	365 (84.68)
2–3	66 (15.32)
Histology
Adenocarcinoma	370 (85.85)
Squamous cell carcinoma	61 (14.15)

KRAS is the Most Common Mutation Type of KRAS in NSCLC

The RAS mutation rate was 10.7% (46/431), and KRAS was the only mutation subtype of RAS (NRAS, KRAS, HRAS). 42 (91.3%) indicated KRAS gene exon 2 mutation, 12 and 13 codon of KRAS gene mutations were detected, and KRAS showed the highest frequency, the total mutation rate of KRAS in NSCLC was 4.6% (20/431) and 43.5% (20/46) of KRAS mutant subtypes, followed by 17.4% (8/46) of KRAS, 8.7% (4/46) of KRAS, and 8.7% (4/46) of KRAS. The mutation frequency of other KRAS types was lower (Figure 2).

Figure 2

Mutation frequencies of KRAS subtypes.

KRAS Mutation Between Tumor Tissue and Plasma

We compared the KRAS mutation spectrums between tumor tissue and ctDNA derived from peripheral blood in this study. Collectively, 37 (11.14%) and 16 (4.81%) patients had KRAS and KRAS mutation spectrum in tumor tissue, nine (9.09%) and four (4.04%) patients in ctDNA, but no significant difference was found in the two sample types (P=0.711, P=1.000, Table 2), respectively.

Table 2

Mutation Frequencies of KRAS Subtypes Between Tumor Tissue and Plasma

Sample Type	KRAS		P	KRASG12C		P
	mut	wt		mut	wt
Tumor tissue	37	295	0.711	16	316	1.000
Plasma	9	90		4	95
Total	46	385		20	411

Co-occurring Genomic Alterations Between KRAS and Lung Cancer Pathogenic Gene

Lung cancer driver genes (include EGFR, RAS, ALK, ROS1, MET, RET BRAF, and HER-2) mutation samples were observed in 332 (77.3%) of 431 patients. Eight (40%) of 20 patients harbored only KRAS mutations, and 12 (60%) had multiple KRAS mutations, including eight (40%) KRAS patients had co-occurring driver oncogenes, was higher trend than KRAS with driver oncogenes mutations (6/26,23%), but no statistical significance (P=0.33), the most commonly co-occurring genomic alterations with KRAS were EGFR (10%, 2/20), ROS1 (10%, 2/20), MET (10%), HER2 (5%, 1/20), ALK (5%, 1/20), BRAF (5%, 1/20), and RET (0%), respectively (Figure 3, ). One hundred and sixty-three patients from 168 gene panel or 520 gene panel found that the KRAS gene is often accompanied by TP53 and PTEN mutation, the mutation rates were 50% (3/6) and 16.7% (1/6), respectively, but STK11 (0.0%, 0/6).

Figure 3

Driver genetic mutations spectrums identified by next-generation sequencing of 332 patients with NSCLC tumor tissue and plasma. Side bar represents the percentage of patients with driver gene mutation. Top bar represents the number of mutations per patient. Different types of mutations are denoted in different colors.

Age, Smoking History and Pathological Stage Associated with KRAS Mutation

The mutation rate of KRAS gene in smokers was higher than that in nonsmokers, 8.33% (14/168) vs 2.28% (6/263), P=0.0046). KRAS has a higher mutation rate in age (≥60 years) 15.2% (18/274) vs 1.27% (2/157); P=0.0151). KRAS mutation was associated with the pathological staging of the patients, 8.69% (10/115) vs 3.16% (10/316), P=0.0343), but was not associated with gender, brain metastasis, PS score, and histology (P=0.2515, P=0.4282, P=0.5266 and P=0.7526) (Table 3), to further identify the values of clinical factor on KRAS mutations, logistic regression analysis was included. In the univariate logistic analysis, age, smoker, clinical stage were identified as independent factors for KRAS mutations (OR=0.551, P=0.024; OR=5.449, P=0.02; OR=0.343, P=0.006). In the multivariate logistic model, smoker (OR=0.306, P=0.037) remained independent factors for KRAS (Table 4). Furthermore, we found that KRAS was dominant in male smokers (100%, 4/4)

Table 3

431 Correlation Analysis Between KRAS and Clinic Pathological Factors in Patients

	KRASG12C-mut n=20	KRASG12C-wt n=411	P-value
Sex
Male	12	186	0.2515
Female	8	225
Age
<60 year	2	155	0.0151*
≥60 year	18	256
Stage
I–III	10	105	0.0343*
IV	10	306
Smoking history
Yes	14	154	0.0046**
No	6	257
Brain metastasis
Yes	3	103	0.4282
No	17	308
PS score
0–1	16	349	0.5266
2–3	4	62
Histology
Adenocarcinoma	18	352	0.7526
Squamous cell carcinoma	2	59

Notes: *P-value <0.05; **P-value <0.01.

Abbreviations: mut, mutation; wt, wild type.

Table 4

Univariate and Multivariate Analysis of KRAS and Clinical Factor

		Univariate Analysis			Multivariate Analysis
		OR	95%CI	P-value	OR	95%CI	P-value
Sex				0.202			0.936
	Male	1			1
	Female	0.551	0.221–1.377		1.044	0.363–3.001
Age				0.024			0.076
	<60	1			1
	≥60	5.449	1.247–23.805		3.932	0.868–17.823
Stage				0.02			0.082
	I–III	1			1
	IV	0.343	0.139–0.847		0.415	0.154–1.118
Smoking history				0.006			0.037
	Yes	1			1
	No	0.257	0.097–0.682		0.306	0.101–0.929
Brain metastasis				0.315			0.871
	Yes	1			1
	No	1.895	0.544–6.598		0.892	0.226–3.516
PS score				0.553			0.704
	0–1	1			1
	2–3	1.407	0.455–4.350		1.256	0.388–4.066
Histology				0.588			0.617
	Adenocarcinoma	1			1
	Squamous cell carcinoma	0.663	0.15–2.932		0.677	0.147–3.116

Discussion

Previously reported RAS was detected in about 25–30% of tumors, several studies consistently reported that Westerners have a higher mutation rate than Asians (26% vs 11%).7 Another report similarly indicated 30% of RAS mutations in Western patients and 5–15% in the Asian population,8 which accounts for about 86% KRAS, 11% NRAS and 3% HRAS mutation of RAS-induced NSCLC, KRAS accounts for 90% of RAS gene mutations in lung adenocarcinoma and is the most common oncogene in NSCLC.9 Our data are consistent with recent studies, our results might indicate the current view that KRAS was the only RAS-mutant isoform, the mutation rate was 10.7% in 431 NSCLC patients, similar to the rates reported by Jia’s group and Liu’s group.10,11 Further studies showed that the KRAS mutation rate is 4.6% in lung cancer, and 43.5% in KRAS mutation for our study. It was similar to several studies in that the KRAS mutation frequency range is from 35% to 45% followed by KRAS and KRAS in KRAS mutant lung cancer, but a lower frequency reported by Liu’s group.9,11–15 One key finding of our study was that KRAS, including KRAS mutation of NSCLC reflected no difference in tissue and blood. Furthermore, this study also reveals the widespread existence of concomitant mutations in patients with KRAS mutant advanced NSCLC, especially driver gene mutations. The three predominant KRAS co-mutations were detected including EGFR-KRAS (10%), equal to ROS1-KRAS (10%) and MET-KRAS (10%). We found the four cases with EGFR-KRAS concomitant mutations in our cohort were all tested before EGFR-TKI treatment, thus partly ruling out the possibility that EGFR-KRAS co-mutation was related to EGFR-TKI resistance.16 Unfortunately, neither were the four cases derived from two separate tumor tissue. The incidence rate of EGFR-KRAS in the Chinese cohort might be likely ethnic-unique, based on the knowledge that the prevalence of EGFR mutation is higher in the Asian population.17 The co-occurrence of EGFR and KRAS was 0.92% (4/431) in our study, which was supported by Scheffler et al13 (1.2%). The four concomitant mutations were KRAS (n=2) co-occurring with either EGFR V1097I (n=1) or EGFR amplification (n=1) and KRAS (n=2) co-occurring with EGFR 19del (n=2). Although previous studies had reported that KRAS are mutually exclusive with mutations in EGFR and ALK in NSCLC,18,19 but coexisting EGFR and KRAS mutations have also been reported..20,21 (Zhu et al reported that three patients with coexisting EGFR and KRAS mutations were found in 206 patients (1.4%).22 We infer that genetic mutation status could be related with different races, sample numbers, as well as test methodology. Nevertheless, current data about KRAS co-occurring mutations in lung cancer is insufficient. Co-occurrence with TP53 or STK11 mutations is common in KRAS mutations.23,24 KRAS and TP53 co-mutations indicated that tumors harboring those mutations couldbe more responsive to immune checkpoint inhibition in lung cancer.25 Conversely, tumors harboring concurrent KRAS and STK11 mutations could be associated with an immunosuppressive microenvironment.26,27 Furthermore, the absence of PTEN promotes resistance to T cell-mediated immunotherapy.28 So we evaluated the mutation status of TP53, STK11 or PTEN in KRAS mutant patients, and it indicated that in the landscape of concurrent genetic alterations in patients with KRAS, the co-mutation rates were 50% and 16.7%, but KRAS was exclusive with STK11 mutation.

KRAS (c.34G>T) alteration is a transversion and KRAS transversion mutations (G→T or G→C) were more common in smokers, in contrast, transition mutations (G→A) were more common in never-smokers in lung adenocarcinomas (n=500).29 Our data showed that smokers more commonly harbored KRAS mutations than KRAS (70% vs 37.5%), which is consistent with reports by Liu et al and Dogan et al.11,30 Data showed that KRAS-mutant NSCLC is genetically complex, with a higher frequency of co-occurring mutations with TP53, STK11, MET and ERBB2 amplifications,29 however, no conclusions implied that the co-occurrence mutations were related to the transversion. In comparison to KRAS, KRAS showed higher mutation frequency in patients older than 60 years, and stage I–III. Our findings were supported by other studies.11,13,31

In summary, our study indicated that KRAS mutations were the most frequent mutant subtype of KRAS in northeast Chinese NSCLC patients and might be involved in the smoking, age, and clinical stage, especially we demonstrated a high frequency of KRAS concomitant TP53/PTEN/EGFR. In addition, no difference was observed between tissue and plasma in the KRAS subgroup of the northeast Chinese NSCLC patients. Our findings might contribute to distinct therapeutic guidance in NSCLC. More data should be collected and explored to address predictive and prognostic value of KRAS in future studies.