Prognostic features and comprehensive genomic analysis of NF1 mutations in EGFR mutant lung cancer patients

Abstract

Objective

NF1 is a tumor suppressor gene that encodes the neurofibromin protein and negatively regulates Ras signaling. This study was aimed to investigate the molecular, clinical characteristics, and prognostic features of NF1 gene in EGFR mutant lung cancer patients.

Method

The next-generation sequencing (NGS) was used to analyze the data from lung cancer patients in the Guangdong Lung Cancer Institute (GLCI) from June 2016 to December 2020.

Results

Somatic NF1 mutations were present in 4.2% (135/3220) of Chinese lung cancer patients. NF1 mutations where clearly enriched in older (p < 0.001), male (p < 0.001), and smoking (p < 0.001) patients. Patients with NF1 mutations were more likely to have TP53 (p = 0.003), BRAF (p = 0.001) and RASA1 (p = 0.026) mutations and mutually exclusive with EGFR mutations (p = 0.006). TP53 mutation had worsen prognosis in cases of NF1 mutant (p = 0.026) or EGFR/NF1 co-mutant (p = 0.031) lung adenocarcinomas (LUAD) patients. There was no effect on overall survival (OS) in LUAD patients with and without NF1 mutations, even in LUAD driver-gene negative patients. NF1/EGFR co-mutation patients had a longer OS than a single mutation of either the EGFR gene (median OS: 47.7 m vs. 30.2 m, hazard ratio [95% CI], 0.47 [0.30–0.74], p = 0.004) or NF1 gene (47.7 m vs. 19.0 m, 0.44 [0.27–0.73], p = 0.003). Furthermore, NF1 mutations significantly prolonged OS in EGFR mutant/TP53 wild-type LUAD patients (106.5 m vs. 25.5 m, 0.28 [0.13–0.59], p = 0.003) but not in patients with EGFR/TP53 co-mutations (36.8 m vs. 30.2 m, 0.70 [0.39–1.26], p = 0.280).

Conclusion

Our results indicated NF1 mutations served as a good prognostic factor in EGFR mutant/TP53 wild-type lung cancer patients in this single-center study. TP53 mutation was obviously enriched in NF1 mutant patients and had shorter OS.

---

What's new?

The molecular, clinical characteristics and prognostic features of NF1 gene in EGFR mutant lung cancer patients have not been extensively explored. We analyzed NF1 mutations in a large-scale cohort of Chinese lung cancer patients. We reported that The frequency of NF1 mutations was significantly lower in the Chinese population than western populations. NF1 mutant tumors could define a specific population with distinct clinical and molecular profile. NF1 mutations were no effect on overall survival in lung cancer patients, but maybe a potential biomarker for good prognosis to EGFR mutant/TP53 wild-type lung adenocarcinoma patients. TP53 mutation was obviously enriched in cases of NF1 mutant patients and had shorter overall survival. The study would provide more information for exploring biomarker of lung cancer.

---

INTRODUCTION

Molecular targeted therapies have dramatically improved treatment for patients whose tumors harbor somatically activated oncogenes, such as epidermal growth factor receptor (EGFR) mutations, anaplastic lymphoma kinase (ALK) rearrangements, and other mutations.^1,2 Additional molecular targets and drug resistance mechanisms using targeted therapy in lung cancer must be investigated.

The NF1 gene is a tumor pathogenic suppressor gene located on chromosome 17q11.2. Mutation inactivation was initially found in patients with the common inherited tumor predisposition syndrome neurofibromatosis type I (NF1).^3,4 The NF1 gene is one of the largest genes in the human genome; it encodes a Ras GTPase-activating protein (neurofibromin) composed of 2839 amino acids.^5,6 The neurofibromin protein inhibits tumor growth by negatively regulating the RAS proto-oncogene. The inactivation of NF1 function has an important role in carcinogenesis that involves the hyperactivation of wild-type RAS proteins and dysregulation of the RAS/MAPK pathway, in which GTP-bound RAS activates the RAF–MEK–ERK signaling cascade to control proliferation.⁷ Activated RAS-GTP also stimulates PI3K/AKT signaling, which protects cells from apoptosis. Acquired somatic NF1 mutations have been identified in various sporadic malignancies that were not associated with NF1,⁸ including lung cancer,⁹ ovarian cancer,¹⁰ breast cancer,¹¹ and acute myeloid leukemia.¹² The Cancer Genome Atlas (TCGA) Research Network reported that NF1 gene mutations are related to the development of lung cancer. Whole-exon sequencing showed that the mutation rate of the NF1 gene in lung cancer was 11%; it was enriched in samples otherwise lacking oncogene mutations.¹³ Moreover, NF1 mutations are not just considered as lung cancer driver genes. In lung cancer models, resistance to EGFR targeted therapy was mediated by NF1 expression, and blocking MEK restored the response.¹⁴

Deleterious NF1 mutations have been previously reported in non-small-cell lung cancer (NSCLC), but the genomic landscape of this molecular subgroup remains poorly characterized in the lung cancer population; the clinical outcomes of NF1 mutant NSCLC also remain unknown. In the present study, we retrospectively analyzed the clinicopathological characteristics, molecular profiles, and prognostic features of NF1 mutations in EGFR mutant lung cancer patients.

TECHNOLOGY AND METHODS Patients

From June 2016 to December 2020, 3230 samples from lung cancer patients from Guangdong Lung Cancer Institute (GLCI) were screened, analyzed, and sequenced for NF1 mutations. Clinical data from these patients were retrospectively collected through a review of their electronic medical records. Factors included in the analysis were sex, age, date of diagnosis, smoking status at diagnosis, pathological type, and clinical stage at the time of diagnosis. The date of death was determined from regular follow-up in the electronic medical records.

Study design

This study involved 3230 samples from lung cancer patients (Figure 1). All specimens were subjected to NF1 mutation analysis by next-generation sequencing (NGS). For patients with multiple screenings, the results of initial tissue samples were prioritized, followed by blood, pleural effusion, and cerebrospinal fluid. The incidences of NF1 mutations and the distributions of NF1 mutation types were analyzed; the clinical characteristics and concurrent alterations of NF1 mutations were evaluated in NSCLC patients. Additionally, 118 patients harboring NF1 mutations and 236 NF1-negative patients (1:2 matched according to sex, age, date of diagnosis, smoking status at diagnosis, pathological type, and clinical stage at the time of diagnosis) were analyzed for survival analysis and gene correlation analysis. The genes of interest in the study included NF1, TP53, EGFR, RB1, KEAP1, ALK, RET, KRAS, BRAF, CDKN2A, PTEN, ROS1, PIK3CA, STK11, NRAS, and RASA1.

FIGURE 1. Flow chart of the study

Targeted sequencing

Tissue DNA was extracted using the QIAamp DNA FFPE Tissue Kit (Qiagen); cell-free DNA in the plasma, pleural effusion, and cerebrospinal fluid was extracted using the QIAamp Circulating Nucleic Acid Kit (Qiagen). Purified DNA was hybridized with oligonucleotide baits; targeted capture was performed using a panel of 520 cancer-related genes (OncoScreen Plus; Burning Rock Biotech., Ltd) or a panel of 425 cancer-related genes (Nanjing Geneseeq Technology., Ltd.). The probes of the two company kits covered the NF1 gene sequence. Samples were sequenced using the Illumina HiSeq 4000 platform or the NovaSeq 6000 platform (Illumina) with 2 × 150 base pair cycles at target sequencing depths of 1000× for tissue DNA samples and 10,000× for cell-free DNA samples. Paired white blood cells were sequenced to filter out clonal hematopoiesis-related mutations. The processing of sequencing data was performed using optimized bioinformatics pipelines to analyze various cancer-related somatic mutations at the DNA level, including point mutations, insertions/deletions, copy number variations, and gene rearrangements. Protocols from previous publication were followed for both experimental procedures and sequence data analysis.^15,16 Table S1 showed the sequencing coverage and quality statistics of the next generation sequencing generated of NF1 mutant patients in this study.

Statistical analysis

Analysis of associations between NF1 gene mutations and clinical factors was performed using the chi-squared test or Fisher's exact test using the Statistical Package for the Social Sciences Software (SPSS) version 25.0. Survival analysis was performed using graphPad prism Software version 9.0. Overall survival (OS) was measured from the date of pathological diagnosis of lung cancer to the date of death or last follow-up, with a cut-off date of August 31, 2021. Kaplan–Meier survival curves were generated to estimate OS in different genomic groups.

RESULTS Clinicopathological characteristics of patients with NF1 mutations

The clinical and pathological characteristics of Chinese lung cancer patients with NF1 mutations and wild-type NF1 were compared (Table 1). Inactivating somatic NF1 mutations were present in 4.0% (108/2696) of the LUAD cases and 6.5% (18/279) of the lung squamous cell carcinomas (LUSC) cases. The median of the NF1 mutant and wild-type patients were 64 and 59 years old, respectively. Patients with NF1 mutations were more likely to be elderly (63.7% vs. 42.3%, p < 0.001). The proportion of men was higher in the NF1 mutant patients than in patients without NF1 mutations (72.6% vs. 57.1%, p < 0.001). 52.6% of NF1 mutation patients were smokers, compared with 33.0% in the NF1 wild-type subgroup (p < 0.001). There were no significant differences in terms of pathological type or stage.

TABLE 1 Summary of demographic and clinicopathological characteristics of lung cancer patients with defined NF1 mutations

[Table omitted. See PDF]

Note: Never-smokers were patients who smoked less than 100 cigarettes in their lifetime.

Abbreviations: NF1, Neurofibromin; SCLC, small cell lung cancer.

^*Comparison between adenocarcinoma and squamous; Statistical analysis excluded “unknown” subgroup.

Distribution of NF1 mutation types

All 160 NF1 mutant sites were detected in our 135 NF1 mutant lung cancer patients. Among these patients, missense mutations (53.8%), nonsense mutations (25.0%), and splice site mutations (10.6%) mainly occurred in the NF1 gene (Figure 2A,B); the single base substitutions C > T and C > A were significantly prominent (Figure 2C). 11.9% (16/135) of patients had more than one NF1 mutant sites; one patient harbored seven NF1 mutations (Figure 2D). There were no hotspot mutations in the NF1 genes; almost all of the 160 NF1 mutant sites occurred only once and were spread throughout all exons of the NF1 gene (Figure 2E).

FIGURE 2. Distribution of all detected 160 NF1 mutant types in our cohort of 135 lung cancer patients. (A–C) Variant classification, variant type, and single nucleotide polymorphism (SNP) class; (D) NF1 variant number per patient; (E) The package “maftools” in R (version 4.0.2) was used to identify NF1 variant sites in the amino acid sequence in our database.

Genetic co-mutation of NF1

NF1 mutations frequently occurred with other oncogenic mutations in our study. The frequency of concurrent mutations and the mutation landscape of selected genes are presented in Figure 3. Among the 135 samples with NF1 mutations, co-mutations were found in the following genes of interest: TP53, 103 (76.3%); EGFR, 36 (26.7%); RB1, 21 (15.6%); KEAP1, 21 (15.6%); KRAS, 13 (9.6%); BRAF, 11 (8.2%); CDKN2A, 10 (7.4%); PTEN, 9 (6.7%); PIK3CA, 8 (5.9%); ALK fusion, 7 (5.2%); RET fusion, 6 (4.4%); STK11, 6 (4.4%); ROS1 fusion, 4 (3.0%); and NRAS, 2 (1.5%). Figure S1 shows an overview of the genomic alterations in >8% of the co-mutant genes in 135 NF1 mutant samples: TP53, EGFR, LRP1B, DPYD, NQO1, RB1, KEAP1, SPTA1, MTHFR, ALK, FAT1, EP300, XRCC1, RET, KRAS, ARID1A, SMARCA4, CTNNB1, and BRAF.

FIGURE 3. Genomic landscape of known oncogenes and suppressor genes in 135 NF1 mutant patients. (A) Frequencies of selected known oncogenes and suppressor genes. (B) The waterfall function of GenVisR package in R (version: 4.0.3) was used to visualize the mosaic plot of the mutation landscape of selected genes. *Percentages of ALK, ROS1, and RET genes indicate fusion subtypes. TIS, tissue; FFPE, formalin-fixed paraffin-embedded tissue; PLA, plasma; PE, pleural effusion; CSF, cerebrospinal fluid.

Gene correlations in the NF1 mutation population were compared between 118 patients with the NF1 mutant NSCLC and 236 patients with NF1 wild-type NSCLC (Table 2). NF1 mutant patients were more likely to have TP53 (p = 0.003), BRAF (p = 0.001) and RASA1 (p = 0.026) mutations; NF1 mutations were mutually exclusive with EGFR mutations (p = 0.006). KRAS mutations were equally distributed in the NF1 mutant and wild-type groups (p = 0.615). Other co-mutations also had similar frequencies between the two cohorts, including ALK (5.9% vs. 11.0%, p = 0.121), RB1 (13.6% vs. 7.6%, p = 0.074), PTEN (6.8% vs. 3.8%, p = 0.219), PIK3CA (5.9% vs. 6.8%, p = 0.760), CDKN2A (8.5% vs. 6.8%, p = 0.564), KEAP1 (15.3% vs. 8.5%, p = 0.052), STK11 (3.4% vs. 6.8%, p = 0.193), RET (4.2% vs 2.1%, p = 0.311), ROS1 (3.4% in both, p = 1.000), and NRAS (1.7% vs. 0%, p = 0.110).

TABLE 2 Gene correlations compared between 118 NF1 mutant cases and 236 non-NF1 mutant cases in NSCLC patients

[Table omitted. See PDF]

^*p-value calculated using Fisher's exact test.

Prognostic features of NF1 mutant patients

The median overall survival of 354 lung cancer patients enrolled in the study was 27.0 m (Figure S3A). Different NF1 mutant types had no effect on OS (Figure S2A); there was no effect on OS between one and more than one NF1 mutant sites (Figure S2B). NF1 mutations had no effect on OS in all patients (Figure S3B); they also had no effect on median OS (mOS) in cases of LUAD (mOS: 33.5 m vs. 26.5 m, hazard ratio [HR] = 0.87, 95% CI: 0.64–1.16, p = 0.345; Figure 4A) and cases of LUSC (Figure S3C), respectively. Furthermore, NF1 mutations had no effect on OS in LUAD cases who were driver gene-negative (without EGFR mutations and ALK fusion) (mOS: 19.0 m vs. 23.6 m, HR = 1.05, 95% CI: 0.70–1.58, p = 0.810; Figure 4B). In LUAD cases, NF1/EGFR co-mutant patients had significantly longer OS than NF1 mutant/EGFR wild-type patients (mOS: 47.7 m vs. 19.0 m, HR = 0.44, 95% CI: 0.27–0.73, p = 0.003). EGFR/NF1 co-mutant patients also had significantly longer OS than EGFR mutant/NF1 wild-type patients (mOS: 47.7 m vs. 30.2 m, HR = 0.47, 95%CI: 0.30–0.74, p = 0.004) (Figure 4C,D). NF1 mutations had no effect on OS in LUAD patients with EGFR/TP53 co-mutations (mOS: 36.8 m vs. 30.2 m, HR = 0.70, 95% CI: 0.39–1.26, p = 0.280) (Figure 4E). However, NF1 mutations had a significant effect on OS in LUAD patients with EGFR mutant/TP53 wild-type (mOS: 106.5 m vs. 25.5 m, HR = 0.28, 95% CI: 0.13–0.59, p = 0.003; Figure 4F). Table 3 showed the related clinical and genetic information in 12 cases of NF1 mutant LUAD patients with EGFR mutant/TP53 wild-type. Since there were fewer LUSC patients with NF1 mutation (n = 15), subgroup analysis cannot be performed in LUSC patients.

FIGURE 4. Overall survival in each subgroup of our patients. (A) Overall survival compared between NF1+ and NF1- cases in LUAD patients; (B) Overall survival compared between NF1+ and NF1- cases in LUAD patients without EGFR mutations and ALK fusion; (C) Overall survival compared between NF1 + EGFR+ and NF1 + EGFR- cases in LUAD patients; (D) Overall survival compared between EGFR + NF1+ and EGFR + NF1- cases in LUAD patients; (E) Overall survival compared between NF1+ and NF1- cases in EGFR + TP53+ LUAD patients; (F) Overall survival compared between NF1+ and NF1- cases in EGFR + TP53- LUAD patients. (G) Overall survival compared between TP53+ and TP53- cases in NF1+ LUAD patients. (H) Overall survival compared between TP53+ and TP53- cases in EGFR + NF1+ LUAD patients. mOS, median overall survival; HR, hazard ratio; +, mutation; −, wild-type.

TABLE 3 The clinical and related genetic information in 12 cases of LUAD patient with NF1 mutant/EGFR mutant/TP53 wild-type

No	Sex	Age	Smoking	Stage	OS (m)	Status	KRAS	ALK	RET	NRAS	ROS1	BRAF	PIK3CA	RB1	CDKNA	KEAP1	STK11	PTEN	Therapy
1	F	64	No	IA	199.2	death	−	−	−	−	−	−	−	−	−	−	−	−	Surg; Targ; Chem
2	M	64	Yes	IB	89.0	death	−	−	−	−	−	−	−	−	−	−	−	−	Surg; Targ
3	M	65	No	IIA	78.6	live	−	−	+	−	−	−	−	−	−	−	−	−	Surg; Targ; Chem
4	F	68	No	IIIB	27.0	live	−	−	−	−	−	−	−	−	−	+	−	−	Targ; Chem
5	M	71	No	IVA	106.5	death	−	−	−	−	−	−	−	+	−	−	−	−	Targ; Chem
6	F	65	No	IVA	40.5	live	−	−	−	−	−	−	+	−	−	−	−	−	Targ
7	M	59	No	IVA	29.0	death	−	−	−	−	−	−	−	−	−	−	−	−	Targ
8	F	68	No	IVA	25.9	live	−	−	−	−	−	−	−	−	−	−	−	−	Targ
9	F	59	No	IVA	22.4	live	−	−	−	−	−	−	−	−	−	−	−	−	Targ; Chem
10	F	63	No	IVB	107.5	live	−	−	−	−	−	−	−	−	−	−	−	−	Targ; Chem
11	M	39	Yes	IVB	37.4	death	+	−	−	−	−	−	−	−	−	−	−	−	Targ
12	M	67	No	IVB	37.0	live	−	−	−	−	−	−	−	−	−	−	−	−	Targ; Chem

Abbreviations: Chem, chemotherapy; F, female; LUAD, lung adenocarcinoma; M, male; Surg, surgery; Targ, targeted therapy for EGFR mutation.

TP53 mutation had worsen prognosis in cases of NF1 mutant (mOS: 73.1 m vs. 20.8 m, HR = 0.54, 95% CI: 0.32–0.91, p = 0.026) or EGFR/NF1 co-mutant (mOS: 106.5 m vs. 36.8 m, HR = 0.36, 95% CI: 0.14–0.97, p = 0.031) LUAD patients (Figure 4G,H). Genes with a co-mutant rate of more than 8% in NF1 mutant patients were enrolled for survival analysis (Figure S4). NF1/CTNNB1 co-mutant had better effect on OS in NSCLC patients (p = 0.027). LRP1B, RB1, KEAP1, FAT1, KRAS, ARID1A, and SMARCA4 mutation had no effect in cases of NF1 mutant NSCLC patients, respectively.

Prognostic features of NF1 mutant patients in the TCGA database

NF1 mutations had no effect on OS in LUAD patients in the TCGA database (mOS: 31.2 m vs. 46.7 m, HR = 1.62, 95% CI: 0.72–3.63, p = 0.156) (Figure S5A). NF1 mutations also had no effect on OS in LUAD patients without EGFR mutations and ALK fusion (mOS: 31.2 m vs. 76.2 m, HR = 1.90, 95% CI: 0.80–4.53, p = 0.062; Figure S5B). Unfortunately, there was only one patient with NF1/EGFR co-mutation, and there were no NF1 mutations in patients with EGFR mutant/TP53 wild-type adenocarcinoma.

DISCUSSION

Genetic and molecular profiling of NSCLC has led to the discovery of actionable oncogenic driver alterations, which has revolutionized treatment for lung cancer. There are increasing efforts to investigate the oncogenic events and drug resistance mechanisms of targeted therapy behind the “driver-related gene-negative” cohort. In lung cancer, NF1 gene mutations occur in approximately 7.0%–11.8% cases of LUAD^{7–9,13,17,18} and 10.3%–12.0% cases of LUSC in Western population.⁸ In our data set, somatic NF1 mutations were present in 4.0% (108/2696) cases of LUAD and 6.5% cases (18/279) of LUSC. The incidence of NF1 mutations was significantly lower in Asia than in Western populations. This difference indicates that ethnicity may be an important determining factor, similar to findings regarding EGFR and KRAS mutations. Our data showed that the clinicopathological features of lung cancer patients with NF1 mutations were clearly enriched in older (p < 0.001), male (p < 0.001), and smoking (p < 0.001) patients. Tlemsani et al. also reported that most patients with NF1 alterations were men and smokers with LUAD.¹⁹ The occurrence of NF1 mutations was not associated with the pathological stage (p = 0.563) in our study.

NF1 mutations occurred with other oncogenic mutations in our study. As shown in Table 2, NF1 mutant patients exhibited concurrent TP53, BRAF, or RASA1 mutations; NF1 mutations were mutually exclusive with EGFR mutations in 118 NF1 mutant NSCLC patients and 236 NF1 wild-type NSCLC patients. Although NF1 negatively regulates the RAS proto-oncogene, no difference was observed in terms of KRAS mutations between the NF1 mutant and wild-type cohorts. Redig et al. reported that NF1 mutations occur more frequently with TP53 mutations and other oncogenic alterations in lung cancer patients.¹⁸ Mice that carry linked germline mutations in NF1 and TP53 develop malignant peripheral nerve sheath tumors.^20,21 In the present study, TP53 mutations were obviously enriched in the NF1 mutant lung cancer patients (103/135, 76.3%), with 71.3% (77/108) in LUAD cases and 100% (18/18) in LUSC cases. Our study showed TP53 mutation had worsen prognosis in cases of NF1 mutant or EGFR/NF1 co-mutant LUAD patients. Thus far, most studies^18,19,22 concerning the genetic correlations of NF1 mutations have been descriptive; they have not involved statistical analysis. The present study clearly demonstrated a relationship between NF1 mutations and various oncogenes and suppressor genes of lung cancer.

NF1 mutations were reported as potential biomarkers for poor prognosis in pancreatic ductal adenocarcinoma,²³ colorectal cancer.²⁴ In our study, 118 NF1 mutant patients and 236 NF1 wild-type patients had enrolled for survival analysis. Our results showed that NF1 mutations had no direct effect on OS in LUAD cases or LUSC cases, even in lung cancer patients without driver genes (EGFR mutations and ALK fusion). However, EGFR/NF1 co-mutations caused a significantly longer OS than a single mutation of either the EGFR or NF1 gene in NSCLC patients. Thus, NF1 mutations served as a good prognostic factor in patients with EGFR mutation, particularly in patients with EGFR mutant/TP53 wild-type patients. At present, limited data are available concerning NF1 mutant patient prognosis. The prognostic features of NF1/EGFR co-mutations have rarely been reported. Pan et al. reported that NF1 mutant patients had distinctive survival features; they may not benefit from EGFR tyrosine kinase inhibitor treatment.²² However, we considered that evidence to be marginally useful because there was only one patient. Furthermore, our survival analysis results were verified after comparison with the TCGA database. The OS of TCGA patients with NF1 mutations (n = 25) indicated a trend of poor prognosis. Nevertheless, no conclusions could be drawn because of the statistically insignificant OS benefit; subgroup analysis cannot be performed because there are few samples in TCGA database. This year, Negrao et al. from the Anderson Cancer Center²⁵ reported that NF1 mutant/high tumor mutational burden (TMB) NSCLC patients had longer progression-free survival than NF1 wild-type/high TMB NSCLC patients in a large-scale study of PD-L1 immune checkpoint inhibitor therapy. The NF1 mutant NSCLC showed robust sensitivity to immune checkpoint inhibitor therapy, compared with NF1 wild-type NSCLC. Therefore, NF1 mutations should not always be considered a poor biomarker in lung cancer, similar to mutations in the KRAS gene. The study by Negrao et al. and the present study may provide a new perspective concerning NF1 mutations.

It's worth mentioning that 14 patients had RASA1 mutations among 938 tested lung cancer patients (1.5%) in this study; RASA1 gene was detected only by the panel of 520 cancer-related genes. However, there were 5 cases of RASA1 mutations among 59 NF1 mutation patients (8.5%), which indicated that the occurrence of RASA1 mutations is obviously related to the occurrence of NF1 mutations in lung cancer patients. All 5 RASA1/NF1 co-mutation patients had TP53 mutations; all patients lacked other known driver genes containing EGFR, KRAS, ALK, ROS1, or RET. Hayashi and colleagues^26,27 demonstrated that targeting downstream MAPK signaling with MEK inhibition in vitro was significantly more potent in NSCLC cells with RASA1/NF1 co-mutations than a single mutation of either RASA1 gene or NF1 gene. In our study, the OS in three adenocarcinoma patients with RASA1/NF1/TP53 co-mutations was 1 month (Stage IVA), 3 months (Stage IVB), and 31 months (Stage IIIB). One adenosquamous carcinoma patient (Stage IVA) has survived for 13 months of follow-up, while one case was lost to follow-up. Survival analysis could not be completed because of the small number of RASA1/NF1 patients in our study. Luo J et al. reported the NF1, TP53, and RB1 triple combined knockout zebrafish displayed severe developmental disruption.²⁸ Here, we observed 14 cases of patients harboring NF1, TP53, and RB1 co-mutations; RB1/NF1/TP53 co-mutant patients showed no effect on OS (Figure S4).

There were several limitations in this study. First, this was a retrospective analysis. Second, this study only presents the results of survival analysis and fails to clarify the mechanism. At present, the mechanism of NF1 mutations in lung cancer initiation and progression has not been explored. A preclinical study showed that NF1 mutation-induced RAS-MAPK signaling activation could be effectively inhibited by MEK inhibitor therapy in NF1 patients.^7,29 The effectiveness of MEK inhibitors in NSCLC patients harboring NF1 mutations has not yet been elucidated. An ongoing phase II trial examining the use of MEK inhibitor (trametinib) in patients with metastatic or locally advanced NSCLC harboring NF1 mutations aims to answer this question (NCT03232892).³⁰

CONCLUSION

This study was the largest comprehensive analysis for NF1 gene in East Asia lung cancer patients. The frequency of NF1 mutations in the Chinese population is significantly lower than that in western populations. NF1 mutations were more common in older, male, and smoking lung cancer patients, and could define a specific population with a distinct clinical profile. NF1 mutant patients enriched TP53, BRAF, and RASA1 mutations, conversely, mutually exclusive with EGFR mutations. This study was more clearly showed the effect of NF1 mutations on EGFR mutant lung cancer patients. NF1 mutations served as a good prognostic factor in EGFR mutant/TP53 wild-type (not EGFR/TP53 co-mutation) lung cancer patients in this single-center study. NF1 mutations should not always be considered a poor biomarker in lung cancer. Additionally, TP53 had worsen prognosis in cases of NF1 mutant or EGFR/NF1 co-mutant LUAD patients.

AUTHOR CONTRIBUTIONS

Conceptualization was performed by Hong-xia Tian and Yi-long Wu; methodology was performed by Hong-xia Tian, Zhi-hong Chen and Guang-Ling Jie; formal analysis was performed by Hong-xia Tian, Guang-Ling Jie and Hong-hong Yan; investigation was performed by Hong-xia Tian, Si-pei Wu and Zhen Wang; resources was performed by Yi-long Wu; data curation was performed by Zhi-hong Chen, Guang-Ling Jie and Dan-xia Lu; supervision was performed by Xu-chao Zhang and Hong-hong Yan; validation was performed by Yi-long Wu and Xu-chao Zhang; writing-original draft preparation was performed by Hong-xia Tian; writing-review and editing was performed by Hong-xia Tian, Shui-lian Zhang and Guang-Ling Jie.

ACKNOWLEDGMENTS

This work was supported by Guangdong Provincial Key Laboratory of Lung Cancer Translational Medicine (Grant No. 2017B030314120); Supporting Fund for Young Scientist to NSFC Application; Science and Technology Program of Guangzhou, China (Grant No. 201804010477 & 201904010028).

CONFLICT OF INTEREST

The authors declare no conflicts of interest.

ETHICAL APPROVAL STATEMENT

The study protocol was approved by the Research Ethics Committee of Guangdong Provincial People's Hospital (No. KY-Z-2020-260-03). All patients provided written informed consent to participate.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are openly available in TCGA at www.tcga.org. The data that supports the findings of this study are available from the corresponding author upon reasonable request.