Exploring the Cause of Survival Disparities in EGFR-mutated Lung Cancer Subtypes: Unraveling Distinctive Genomic and Phenotypic Features of 19Del and L858R Mutation Subtypes

doi:10.21203/rs.3.rs-3332742/v1

Download PDF

Research Article

Exploring the Cause of Survival Disparities in EGFR-mutated Lung Cancer Subtypes: Unraveling Distinctive Genomic and Phenotypic Features of 19Del and L858R Mutation Subtypes

https://doi.org/10.21203/rs.3.rs-3332742/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Differences in the efficacy of epidermal growth factor receptor tyrosine kinase inhibitors (EGFR-TKI) have been observed between lung cancer patients with 19 exon deletion (19Del) and L858R mutation. We investigate the multi-omics information from the TCGA LUAD dataset and validate it using the GEO (GSE190139, GSE147377) and MSK datasets. Somatic loss-of-function alteration of RBM10 and altered Immune infiltration profile correlated with L858R decreased survival. Meanwhile, in the L858R mutation group, 9p21.3 loss and CDKN2B methylation, increased cell cycle-related gene expression, and the enrichment in cell cycle pathways were associated with poor survival. Comprehensive genomic and phenotypic analysis of the EGFR-mutated lung cancer subtypes reveals distinctive features of each subtype, laying the groundwork for subtype-specific treatment and care options for lung cancer patients.

Non-Small Cell Lung Cancer (NSCLC) poses a significant public health challenge due to its high incidence rate and persistently low 5-year survival rate of less than 30%¹. Among the NSCLC cases, adenocarcinoma stands out as the most prevalent histology, comprising about 55% of cases². Most lung adenocarcinomas are associated with somatic EGFR mutations, which are key drivers in NSCLC, with prevalence in Asians (about 50%) and Caucasians (about 15%)^1,2. EGFR-tyrosine kinase inhibitors (EGFR-TKIs) like gefitinib and erlotinib are primary treatments for advanced NSCLC. However, despite their promise, there are distinct disparities in treatment outcomes among patients with different EGFR mutation subtypes. Specifically, patients harboring the 19Del mutation demonstrate more favorable progression-free and overall survival compared to those with the L858R mutation³.

The differences between the two EGFR mutation subtypes, 19Del, and L858R, remain an ongoing debate. Some studies suggest that structural disparities may exhibit different binding affinities with TKIs, with post-resistance T790M being more frequent in 19Del than in L858R mutated patients ^4–6. Despite the advancements in genetic profiling through next-generation sequencing (NGS), the identification of reasons for EGFR-TKI efficacy disparities also has limitations. Some studies relied on mouse models instead of human cells⁷, and others conducted multi-omics analyses, just highlighting the intricate nature of EGFR 19Del and L858R mutant cells⁸.

In this investigation, a comprehensive analysis of multi-omics data was performed using the public dataset to identify the molecular features of L858R and 19Del lung cancers, which provide the foundation for the development of disease-specific therapeutic approaches and prognostic biomarkers for NSCLC.

L858R and 19Del sample preparation in TCGA LUAD Dataset

In this study, we present a reproducible procedure for conducting a comprehensive analysis of lung adenocarcinoma (LUAD) samples sourced from The Cancer Genome Atlas (TCGA)⁹ repository. The dataset encompasses a wide range of molecular features, including somatic mutation data(n = 563), copy number segment information (n = 532), gene expression profiles (n = 585), and DNA methylation patterns (n = 503). To investigate and discern the differences between the L858R and 19Del lung cancer subtypes, we filtered the samples with specific criteria (Table 1):

Identify 19Del samples based on chromosome start and end positions, referencing USCS transcripts.

Considerate the L858R amino acid change.

Exclude T790M samples, focusing solely on pre-EGFR-TKI treatment cases.

These criteria provide a solid foundation for the argument and hypothesis in this paper. The results obtained from our analysis are further validated by individual datasets from MSK and GEO datasets, reinforcing the robustness of our findings (Fig. 1).

Table 1

Overview of samples after regrouping and filtering.
Sample ID	Amino Acid Change	Start position	End position
TCGA-75-6207-01A	p.E746-A750del	55174772	55174786
TCGA-49-4501-01A	p.E746-A750del	55174773	55174787
TCGA-97-8171-01A	p.E746-A750del	55174772	55174786
TCGA-55-6981-01A	p.L747-T751del	55174775	55174789
TCGA-71-6725-01A	p.L747-A750del	55174775	55174786
TCGA-55-8206-01A	p.E746-A750del	55174773	55174787
TCGA-97-8547-01A	p.E746-A750del	55174772	55174786
TCGA-38-4628-01A	p.E746-A750del	55174772	55174786
TCGA-86-8280-01A	p.E746-A750del	55174772	55174786
TCGA-97-8552-01A	p.E746-A750del	55174773	55174787
TCGA-38-6178-01A	p.E746-A750del	55174772	55174786
TCGA-62-A46U-01A	p.E746-A750del	55174773	55174787
TCGA-50-6591-01A	p.K754-I759del	55174797	55174814
TCGA-17-Z032-01A	p.E746-A750del	55174772	55174786
TCGA-75-7025-01A	p.E746-A750del	55174772	55174786
TCGA-97-A4M6-01A	p.E746-A750del	55174772	55174786
TCGA-86-8074-01A	p.E746-A750del	55174772	55174786
TCGA-75-6212-01A	p.L747-E749del	55174773	55174781
TCGA-J2-8192-01A	p.E746-A750del	55174773	55174787
TCGA-62-8402-01A	p.E746-A750del	55174772	55174786
TCGA-MP-A4T6-01A	p.L747-E749del	55174773	55174781
TCGA-55-A57B-01A	p.L747-T751del	55174775	55174789
TCGA-44-2661-01A	p.E746-A750del	55174772	55174786
TCGA-97-A4M1-01A	p.L858R	55191822	55191822
TCGA-86-8075-01A	p.L858R	55191822	55191822
TCGA-86-8055-01A	p.L858R	55191822	55191822
TCGA-49-4490-01A	p.L858R	55191822	55191822
TCGA-97-A4M7-01A	p.L858R	55191822	55191822
TCGA-67-3770-01A	p.L858R	55191822	55191822
TCGA-67-6217-01A	p.L858R	55191822	55191822
TCGA-62-8394-01A	p.L858R	55191822	55191822
TCGA-86-8668-01A	p.L858R	55191822	55191822
TCGA-55-8506-01A	p.L858R	55191822	55191822
TCGA-MP-A4T9-01A	p.L858R	5519182	55191822
TCGA-38-4627-01A	p.L858R	55191822	55191822
TCGA-MP-A4SW-01A	p.L858R	55191822	55191822
TCGA-55-8096-01A	p.L858R	55191822	55191822
TCGA-67-3772-01A	p.L858R	55191822	55191822
TCGA-71-8520-01A	p.L858R	55191822	55191822
TCGA-50-5944-01A	p.L858R	55191822	55191822
TCGA-97-8177-01A	p.L858R	55191822	55191822
TCGA-17-Z047-01A	p.L858R	55191822	55191822
TCGA-91-6835-01A	p.L858R	55191822	55191822

Survival Analysis in cbioportal

Survival analysis was conducted to explore the association between gene expression levels and the survival outcomes in lung adenocarcinomas. We utilized the cbioportal platform (https://www.cbioportal.org/), a tool allowing us to investigate the mutation status of these genes specifically in samples with EGFR mutations, and correlate them with relevant clinical data. The significance threshold was established at p-value < 0.05, indicating the presence of statistically significant differences between the wild-type and mutated gene groups.

GISTIC2 differential analysis for CNV data

The Genomic Identification of Significant Targets in Cancer 2 (GISTIC2) algorithm was utilized to analyze copy number variations (CNVs) in the dataset. GISTIC2 estimates the baseline level of CNVs and identifies recurrent CNV alterations across the entire genome. The algorithm applies statistical tests to assess the alterations' significance. It calculates a G-score (-log (Probability | Background)) and q-value = 0.05 for each region, with 0.9 confidence level that each event represents a true driver alteration rather than a random occurrence¹⁰.

DESeq2 and GSEA analysis for gene expression data

DESeq2 was employed to compare RNA-seq data from experimental groups featuring duplicate samples¹¹. This tool fits a model based on the negative binomial distribution using counts, FPKM, or RPKM data extracted from an expression matrix. Gene Set Enrichment Analysis (GSEA)¹² evaluates the differential expression of genes by ranking fold change levels. The significance of the enrichment scores was assessed through a permutation-based approach. The gene set with q-values less than 0.05 was recognized as enriched gene sets.

ChAMP R package for DNA methylated data

The ChAMP, a Bioconductor R package, was employed in this study to analyze DNA methylation data from Illumina 450k arrays. ChAMP perform preprocessing steps, quality control, and normalization procedures, which facilitated the identification of differentially methylated regions and significant changes in DNA methylation patterns within and between experimental groups. The package's visualization capabilities also helped the interpretation and representation of our findings, enhancing the overall understanding of the epigenetic changes under investigation.

Correlations Between gene Expression and Immune Cells in TIMER

The TIMER (http://cistrome.org/TIMER/) determines the association between gene expression and immune infiltration¹⁵. TIMER uses gene expression profiles to infer the quantity of tumor-infiltrating immune cells using a previously established statistical deconvolution method¹³. It contains 10,897 TCGA samples from 32 cancer types. Specifical gene expression was correlated with the quantity of all six categories of immune cells as well as tumor purity: B cells, CD4 + T cells, CD8 + T cells, neutrophils, macrophages, and dendritic cells.

Somatic mutational alternations in EGFR 19Del and L858R with survival

Analysis of both EGFR-mutated types revealed RBM10 and PTEN as the only significantly mutated genes (SMGs) in L858R and 19Del, respectively (Fig. 2A; 29% and 3%, 1% and 14%, respectively). Further examination using odd ratios (OR) and significance levels between mutational data from both groups demonstrated RBM10 loss of function mutations enriched in the L858R group and PTEN mutations enriched in the 19Del group (Fig. 2B; OR = 14.534 and 0.065). Moreover, across the entire cohort, RBM10 was found to be correlated with more decreased survival in L858R than in 19Del (Fig. 2C; p = 0.0264, n = 17 and 73). Moreover, the clinical analysis revealed that patients in the L858R subtype were mostly found in stage IIIA, IIIB of lung cancer (Fig. 2D).

RBM10 expression is correlated with immune infiltration in the tumor microenvironment (TME) in L858R

Immune infiltration between L858R and 19Del in LUAD revealed immune cell composition patterns within the tumor microenvironment (TME)¹⁴. Most of immune markers reflect little differences between L858R and 19Del samples. However, some differences were noted, despite the low magnitude, in specific immune cell types, namely macrophage M1 T cells, NK cells, and eosinophil cells (Fig. 3A). While there is little differences between L858R and 19Del in immune signature, the immune infiltration results from TIMER demonstrated positive correlations between RBM10 expression and the key immune cell infiltration, including B cells (R = 0.101, P = 2.48E–02), CD8 + T cells (R = 0.111, P = 1.41E–02), CD4 + T cells (R = 0.232, P = 1.96E − 07), macrophages (R = 0.269, P = 1.31E–09), neutrophils (R = 0.03, P = 5.09E–01) while negatively correlated with dendritic cells (R = -0.282, P = 1.91E − 10) (Fig. 3B)¹⁵.

Copy number profiles in EGFR 19Del and L858R

The comparison of chromosomal copy number profiles was conducted between L858R and 19Del samples in TCGA LUAD. The L858R samples displayed significant loss of chromosome 9p21.3 and gain of 5p15.2 (Fig. 4A), while the 19Del samples demonstrated gains of chromosomes 7 and 12 (Fig. 4B). This analysis reflects differences in the chromosomal alterations associated with the two mutation subtypes, indicating distinct genomic characteristics between L858R and 19Del mutations in LUAD.

Gene expression profiles in EGFR 19Del and L858R

For gene expression data, we performed Gene Set Enrichment Analysis (GSEA). The analysis revealed the enrichment of pathways associated with cell cycle and cell proliferation, including sister chromatid segregation, E2F targets, and G2M checkpoints, in the L858R group (Fig. 5A). Furthermore, using data from the GSE190139 dataset, which includes RNA-seq data from 46 human surgically-resected NSCLCs, we observed significantly higher expression of cell cycle and proliferation-related genes such as MCM4, CD44, TOP2A, CDK1, and MKI67 in the L858R samples (Fig. 5B).

Genomic methylation profiles in EGFR 19Del and L858R

In the comparison of methylation profiles between L858R and 19Del samples, distinct patterns emerged. The CpG sites were found to be significantly enriched in opensea positions in L858R samples, whereas in the 19Del group, enrichment was observed in island positions (Fig. 6A; 6B). Additionally, we investigated the CDKN2B methylation levels in both L858R and 19Del subgroups, revealing that CDKN2B displayed a high level of methylation in L858R positions, potentially contributing to its loss of function in these specific samples (Fig. 6C).

The molecular distinctions between L858R and 19DEL patients have remained unclear. Using a multi-omics analysis, we attempted to uncover more valuable information on the differences between L858R and 19DEL patients. We discovered that somatic alterations of RBM10, 9p21.3 loss and CDKN2B were correlated with decreased survival in L858R mutations.

Firstly, RBM10 missense mutations have effects on LUAD cell proliferation and RNA splicing. Loss-of-function mutations in RBM10 lead to alterations in splicing patterns, affecting the expression of genes involved in critical cellular processes such as cell proliferation, apoptosis, and cell cycle regulation^16,17. Our study also highlights the role of RBM10 mutations in immune infiltration within the tumor microenvironment ^18,19. These molecular mechanisms are represented by the prevalence of advanced disease stages primarily within the L858R subset, indicating a potential role of RBM10 mutations in disease progression and their clinical relevance in later-stage lung cancer cases. The correlation between RBM10 and reduced survival rates in L858R suggests its potential as a prognostic indicator for adverse outcomes in lung adenocarcinoma (LUAD) linked to L858R mutation ^20,21.

Equally intriguing are the implications of 9p21.3 loss on tumor cell behavior and clinical outcomes in L858R patients. 9p21.3 loss contributes to tumor cell proliferation and cell cycle variation with poor prognostics in L858R mutated samples ²². Differences between L858R and 19Del were observed in somatic alterations within the 9p21.3 region. The loss of 9p21.3 holds clinical implications, being identified as a potential diagnostic and prognostic marker across diverse cancer types. Its absence has consistently correlated with tumor progression and adverse patient outcomes²³. Moreover, the potential influence of 9p21.3 loss on immune evasion through cell cycle pathways and long-distance transcriptional regulation due to the presence of an abundance of DNA regulatory elements^24–26. Importantly, it is suggested that the absence of 9p21.3 loss might contribute to immune evasion through altered cell cycle pathways. The outcomes of our differential gene expression and Gene Set Enrichment Analysis (GSEA) underscored distinctive molecular profiles between L858R and 19Del. Genes associated with cell proliferation and cell cycle pathways were notably enriched in L858R samples, indicative of heightened cellular growth and division. This aligns with the presence of 9p21.3 loss in L858R samples, thus strengthening the hypothesis that alterations in these pathways lead to the differential survival rates observed in patients²⁷. Furthermore, the data indicated high methylation levels in the CDKN2B gene, a gene situated within the 9p21.3 locus, within L858R samples. CDKN2B acts as a tumor suppressor. Its compromised function can precipitate unchecked tumor growth and instability in cellular regulation^28,29. This may explain the observed variations in patient survival between the L858R and 19Del mutations. The linkage between methylation patterns and gene function adds depth to our understanding of the epigenetic mechanisms at play. As a result, our study establishes the prominence of 9p21.3 loss and CDKN2B as potential diagnostic and prognostic markers. By analyzing their impact on cell cycle regulation, and tumor suppression, our findings provide a framework for understanding the interplay between genetic and epigenetic factors in cancer progression and patient outcomes, which could explain the reason of poor prognostic phenomenon in L858R subtype²⁴.

In conclusion, our study has shed light on the distinctive genomic and phenotypic features of EGFR-mutated lung cancer subtypes, specifically the 19-exon deletion (19Del) and L858R mutation subtypes. These differences have significant implications for the survival outcomes of patients and pave the way for subtype-specific treatment approaches. Through comprehensive analysis of multi-omics data from the TCGA LUAD dataset, complemented by validation using independent GEO and MSK datasets, we discovered that somatic loss-of-function alterations of RBM10, 9p21.3 loss and CDKN2B were correlated with decreased survival in L858R mutations, implicating its potential as a prognostic indicator. These findings underscore the importance of these genetic alterations in influencing tumor cell proliferation and cell cycle pathways, ultimately impacting patient outcomes.

Funding:

There are no funders of the research.

Competing Interests declaration:

There are no Competing Interests of the research.

Author Contribution:

Yongguang Cai lead the team and organizes the outline of the experiment. Jiayi Cai, Wei Lu, Haiyan Liang, and Sixian Chen prepared the text and figures. Yongguang Cai, Yongfeng Chen, Qiayi Zha, Yuanyuan Li, Shuiqiang Hong, Suli Zhou, and Yuan Lu collectively analyzed the omics data and explored the data.

Cunha Santos GD, Shepherd FA, Tsao MS (2011) Annu Rev Pathol 6:49–69
Tan CS, Gilligan D, Pacey S (2015) Lancet Oncol 16:e447–e459
Singh M, Jadhav HR (2018) Drug Discov Today 23:745–753
Wang ZF, Ren SX, Li W, Gao GH BMC Cancer, 10.1186/s12885-018-4075-5
Liang H, Pan Z, Wang W, Guo C, Chen D, Zhang J, Zhang Y, Tang S, He J, Liang W (2018) J Thorac Dis 10:2311–2320
Liang H, Li C, Zhao Y, Zhao S, Huang J, Cai X, Cheng B, Xiong S, Li J, Wang W, Zhu C, Li W, He J, Liang W (2020) Cancer Manag Res 12:8653–8662
Liang H, Li C, Zhao Y, Zhao S, Huang J, Cai X, Cheng B, Xiong S, Li J, Wang W, Zhu C, Li W, He J, Liang W (2020) Cancer Manag Res 12:8653–8662
Ochi N, Takeyama M, Miyake N, Fuchigami M, Yamane H, Fukazawa T, Nagasaki Y, Kawahara T, Nakanishi H, Takigawa N Exp Cell Res, 10.1016/j.yexcr.2023.113503
Koboldt DC, Fulton RS, McLellan MD, Schmidt H et al (2012) Nature 490:61–70
Mermel CH, Schumacher SE, Hill B, Meyerson ML, Beroukhim R, Getz G Genome Biol,, 10.1186/gb-2011-12-4-r41
Anders S, Huber W Genome Biol, 10.1186/gb-2010-11-10-r106
Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, Paulovich A, Pomeroy SL, Golub TR, Lander ES, Mesirov JP (2005) Gene set enrichment analysis: A knowledge-based approach for interpreting genome-wide expression profiles,
Li T, Fu J, Zeng Z, Cohen D, Li J, Chen Q, Li B, Liu XS (2020) Nucleic Acids Res 48:W509–W514
Ricketts CJ, De Cubas AA, Fan H (2018) C C SmithCell Rep 23:313–326e5
Yuan Q, Sun N, Zheng J, Wang Y, Yan X, Mai W, Liao Y, Chen X Front Oncol, 10.3389/fonc.2019.01502
Wang K, Bacon ML, Tessier JJ, Rintala-Maki ND, Tang V, Sutherland LC (2012) J Cell Death, 5, JCD.S9073.
Zhao J, Sun Y, Huang Y, Song F, Huang Z, Bao Y, Zuo J, Saffen D, Shao Z, Liu W, Wang Y Sci Rep, 10.1038/srep40488
Zeng Q, Sun S, Li Y, Li X, Li Z, Liang H Front Oncol, 10.3389/fonc.2019.01555
Shi S, Ye S, Mao J, Ru Y, Lu Y, Wu X, Xu M, Zhu T, Wang Y, Chen Y, Tang X, Xi Y (2020) Autoimmunity 53:210–217
Li T, Pang X, Wang J, Wang S, Guo Y, He N, Xing P, Li J Front Oncol, 10.3389/fonc.2021.591922
Wang C, Liang H, Lin C, Li F, Xie G, Qiao S, Shi X, Deng J, Zhao X, Wu K, Zhang X Int J Mol Sci, 10.3390/ijms20174251
Hernández J, Bechara E, Schlesinger D, Delgado J, Serrano L, Valcárcel J (2016) RNA Biol 13:466–472
Spiliopoulou P, Yang SYC, Bruce JP, Wang BX, Berman HK, Pugh TJ, Siu LL (2022) Trends Immunol 43:379–390
Eichenauer T, Simmendinger L, Kluth M, Chirico V, Luebke AM, Höflmayer D, Hinsch A, Jacobsen F, Hube-Magg C, Möller-Koop C, Dahlem R, Fisch M, Rink M, Riechardt S, Tsourlakis MC, Büscheck F, Bernreuther C, Clauditz T, Lebok P, Simon R, Sauter G, Wilczak W, Fraune C (2020) Urologic Oncology: Seminars and Original Investigations, 38, 605.e1-605.e8
Alentorn A, Dehais C, Ducray F, Carpentier C, Mokhtari K (2015) Allelic loss of 9p21.3 is a prognostic factor in 1p/19q codeleted anaplastic gliomas,
Eichenauer T, Simmendinger L, Kluth M, Chirico V (2020) Urologic Oncology: Seminars and Original Investigations 38:605e1–60605. .e8
Ghobadi N, Mehramiz M, ShahidSales S, Rezaei A, Brojerdi (2019) J Cell Physiol 234:5070–5076
Hustinx SR, Leoni LM, Yeo CJ, Brown PN, Goggins M, Kern SE, Hruban RH, Maitra A (2005) Mod Pathol 18:959–963
Xia Y, Liu Y, Yang C, Simeone DM, Sun TT, DeGraff DJ, Tang M, Zhang Y, Wu XR Nat Commun, 10.1038/s41467-021-22327-5

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Exploring the Cause of Survival Disparities in EGFR-mutated Lung Cancer Subtypes: Unraveling Distinctive Genomic and Phenotypic Features of 19Del and L858R Mutation Subtypes

Status:

Version 1

Abstract

Figures

Introduction

Materials and methods

L858R and 19Del sample preparation in TCGA LUAD Dataset

Survival Analysis in cbioportal

GISTIC2 differential analysis for CNV data

DESeq2 and GSEA analysis for gene expression data

ChAMP R package for DNA methylated data

Correlations Between gene Expression and Immune Cells in TIMER

Results

Somatic mutational alternations in EGFR 19Del and L858R with survival

RBM10 expression is correlated with immune infiltration in the tumor microenvironment (TME) in L858R

Copy number profiles in EGFR 19Del and L858R

Gene expression profiles in EGFR 19Del and L858R

Genomic methylation profiles in EGFR 19Del and L858R

Discussion

Conclusion

Declarations

References

Additional Declarations

Status:

Version 1