Effectiveness of Cuproptosis-related long non-coding RNAs in predicting prognosis and immune response in patients with lung squamous cell carcinoma

doi:10.21203/rs.3.rs-2977475/v1

Download PDF

Research Article

Effectiveness of Cuproptosis-related long non-coding RNAs in predicting prognosis and immune response in patients with lung squamous cell carcinoma

https://doi.org/10.21203/rs.3.rs-2977475/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background

Thirty percent of non-small cell lung malignancies is lung squamous cell carcinomas (LUSC). However, its prognostic indicators are not universally accepted as standard. Long non-coding RNAs (lncRNAs), which plays complex and precise regulatory roles in gene expression, are involved in various biological processes including tumor proliferation, apoptosis, invasion, and metastasis, and are often widely studied as prognostic indicators. During mitochondrial respiration, cuproptosis, a novel form of cell death occurs when tricarboxylic acid cycle’s (TCA) lipid acylated components bind directly to copper ions. Cuproptosis causes proteotoxic stress due to aggregated lipid acylated proteins and the downregulation of iron-sulfur cluster proteins, eventually causing cell death. This research dealt with exploring the cuproptosis-related lncRNAs function in predicting clinical prognosis and immunotherapy in patients with LUSC.

Methods

Clinical, genomic, and mutational data of LUSC patients were accessed at the Cancer Genome Atlas (TCGA). Subsequently, the mRNA-lncRNA co-expression network was visualized to screen cuproptosis-related lncRNAs. LASSO and Cox regression analysis was executed to establish lncRNA risk models to assess the LUSC patients’ prognostic risk. The stratification of patients into high-risk and low-risk groups was performed. In addition, ROC, survival, risk curves, nomogram, C-Index, independent prognostic analysis, and clinical subgroup model validation were used to assess prognostic value. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis, immune-related functional analysis, and tumor mutation burden (TMB) analysis were performed. Subsequently, the impact of immune escape and immunotherapy in high- and low-risk groups was judged by the TIDE score. Finally, potential drugs for LUSC were identified, and their sensitivities were calculated.

Results

The resulting data identified five cuproptosis-related lncRNAs as being capable of independently predicting the prognosis (AC010328.1, LINC01740, AL358613.2, MIR3945HG, and AC002467.1). In addition, the patient’s risk scores were quantified. The two risk groups depicted significant differences in OS and PFS with a better prognosis for the lower-risk patients in contrast with the high-risk patients. Prognostic analysis by Kaplan-Meier survival analysis, ROC curve, nomogram, and C-index indicated the capacity of lncRNAs that were linked to cuproptosis in accurately predicting the patient’s prognosis. Heatmap and nomogram depicted clear distribution of cuproptosis-related lncRNAs in high- and low-risk groups. Enrichment analysis indicated that the lncRNAs’ biological functions are involved in the metabolism of antitumor drugs. Additionally, the vast majority of immune functions were significantly active in the high-risk group, whereas the median survival time of high-TMB and low-risk patients was considerably longer than the other groups. According to TIDE analysis, the high-risk group patients had a high risk of immune escape and worse immunotherapy outcomes. Several drugs with higher sensitivity for high-risk LUSC patients were screened as well.

Conclusion

A model based on five cuproptosis-related genes was established to predict LUSC patients’ prognoses. The model’s reliability was evaluated in various aspects such as immune response. Overall, the findings of this study may offer new perspectives into the clinical management and immunotherapy of LUSC.

Cuproptosis

Lung squamous cell carcinoma

Immunosuppression

Prognostic risk model

Long non-coding RNA

Lung cancer, the number one malignancy in terms of morbidity and mortality, seriously affects people's health and quality of life. Lung squamous carcinoma (LUSC), one of the most prevalent pathological types after lung adenocarcinoma, is closely related to smoking. The main treatments for lung squamous carcinoma are surgery, chemotherapy, radiotherapy, targeted therapy, and immunotherapy. However, LUSC is less sensitive to radiotherapy and chemotherapy than small-cell undifferentiated carcinoma. Most patients with LUSC have already developed distant metastases when diagnosed due to early insidious symptoms and have lost the opportunity for surgery. In spite of the breakthroughs in the diagnosis and treatment of LUSC in clinical and experimental oncology, patient survival has not substantially improved. Therefore, detailed knowledge of the processes that are involved in the onset and progression of LUSC is necessary to improve the diagnosis, prevention, and treatment of this disease.

Advances in whole-genome sequencing technologies have conferred a deeper understanding of the transcriptome of organisms. It is now believed that more than 90% of non-coding RNAs (ncRNAs) in the human genome play an important biological role [1]. Non-coding RNAs can be divided into two categories depending on their length: small non-coding RNAs (ncRNA < 200 nt) as well as long non-coding RNAs (lncRNA > 200 nt) [2]. Individual genes and gene expression programs are modulated by various lncRNAs utilizing epigenetic regulation or alteration of the fundamental transcriptional machinery, despite the lack of their protein-coding potential [3]. Past intensive studies of lncRNAs have shown that any abnormality in the expression of lncRNAs causes the normal biological and pathological mechanisms to be dysregulated, resulting in the development of disease [4]. In particular, aberrant expression of lncRNAs promotes the development and progression of human cancers, particularly lung cancer [5]. lncRNAs perform different cellular functions, regulating essential processes including cellular growth, proliferation, migration, apoptosis, maintenance of stem cells, and epithelial-mesenchymal transition, as well as acting as signal transducers, scaffolds, and molecular decoys [5]. lncRNAs have been identified as major regulators of several malignancies and play key roles in various aspects of carcinogenesis, including proliferation, migration, invasion, metabolism, and immune escape (Table 1). Further research on lncRNAs’ function in LUSC is needed.

Excess or deficiency of copper, an essential cofactor for all organisms, results in impaired cellular function and leads to cell death [6]. Copper-dependent death occurs when the tricarboxylic acid (TCA) cycle’s lipoylated components bind directly to copper which ultimately results in proteotoxic stress due to lipoylated protein aggregation (LIPA) and a further reduction in iron-sulfur cluster proteins, and eventual cell death. This distinctive mode of cell death is known as cuproptosis [7, 8]. Although it is not clear how excess copper leads to cell death, its mechanism of cell death is different from that of iron death. Recently, disturbances in copper homeostasis can be observed in many malignancies and are linked to the growth and proliferation of tumors, angiogenesis, and metastasis [9]. Cui L [10] et al. developed a safe copper-depleting nanoparticle (CDN) and tested it against triple-negative breast cancer (TNBC). Their results found that CDN depleted intracellular copper and induced cell death in TNBC cells. In three TNBC mouse models, administration of CDN inhibited the growth of the tumor and significantly improved survival in mice. Tumor angiogenesis and growth have been demonstrated to be reduced by copper depletion in earlier animal models and clinical trials [11]. The exact process of cuproptosis, however, has not been elucidated yet. Therapeutic and clinical diagnostic significance has been depicted by cuproptosis-related lncRNAs in LUSC. This research identified cuproptosis-related lncRNAs by bioinformatics analysis. Additionally, their biological functions and their prognosis prediction roles in LUSC patients were investigated in depth.

Table 1

Functional characterization of related lncRNAs genes in various tumors
Tumor type	LncRNA type	Expression	Role	Functional role	Related genes and pathways	References
non-small cell lung cancer	lncRNA MIR22HG	Down-regulated	Tumor suppressor	cell proliferation, migration, and invasion	Ybx1, MET, and p21	[12]
Colorectal	lncRNA SH3PXD2A-AS1	up-regulated	Tumor Promoter	cell proliferation, migration, and invasion	P57 and KLF2	[13]
Hepatocellular carcinoma	lncRNA HULC	up-regulated	Tumor Promoter	cell proliferation, migration,	miRNA15a, PTEN	[14]
gastric cancer	lncRNAsMEG3	Down-regulated	Tumor suppressor	cell proliferation, migration,	p53, miRNA-21	[15, 16]
breast cancer	lncRNA HOTAIR	up-regulated	Tumor Promoter	cell growth, migration, invasion, and apoptosis	miRNA-20a-5p/HMGA2	[17]
glioblastoma	lncRNA MIR22HG	up-regulated	tumor promoter	cell proliferation and invasion	MiR-22-3p/miR-22-5p, SFRP2/PCDH15, ACIL6JTK and Wnt/β-catenin signaling pathway	[18]

1.1 Dataset and preprocessing

The LUSC patients’ gene expression data regarding the RNA-Seq data were accessed at the TCGA database. A total of 502 transcriptome data of LUSC patients were obtained after collation. Disease-related clinical and mutation information of LUSC patients was extracted from TCGA. Subsequently, mRNA and lncRNA in the sequenced expression matrix transcripts were distinguished based on gene annotation. In addition, the expression profiles of cuproptosis-related genes in TCGA-LUSC samples were extracted and subjected to co-expression correlation analysis with lncRNAs. Finally, potential cuproptosis-related lncRNAs were obtained per the correlation coefficient criteria |R|>0.4 and P < 0.001.

1.2 Construction of cuproptosis-related lncRNA prognostic model

In a ratio of 7:3, the individuals with LUSC were stratified randomly into training and test groups. Prognosis-related lncRNAs obtained from the training group by one-way Cox analysis were subjected to LASSO regression analysis per 10-fold cross-validation of 1000 times. Subsequently, the resulting data of the screening were incorporated into a multivariate Cox analysis to establish a prognostic risk score model for the training group. The risk score (RiskScore) was calculated as follows: RiskScore = lncRNA expression₁ × Coef₁ + lncRNA expression₂ × Coef₂ + ... + lncRNA expression_n × Coef_n (Coef: regression coefficient of lncRNA in multivariate Cox regression analysis; n: regression coefficient associated with the number of prognosis-related lncRNAs). Finally, each LUSC patient’s risk score, including those in the test group, was calculated according to the above formula.

1.3 Assessment of the prognostic risk model

The LUSC patients’ clinical features depicted the difference between the training and test groups through clinical statistical analysis to determine bias on grouping (chi-square test). Heat maps of the link between lncRNAs involved in model construction and cuproptosis-related genes were drawn. As per the risk score’s median value stratification of the patients into high-risk and low-risk groups was executed to assess the value of this risk score in prognostic prediction. The variation in patient survival between the two risk groups in the test and training groups was compared, and subsequently, overall survival (OS) and progression-free survival (PFS) were derived for individuals with LUSC. Patient risk profiles, survival status, and lncRNAs expression heat map were plotted according to risk scores. Univariate and multivariate independent prognostic analyses were carried out to explore the predictive independence and prognostic value of the prognostic model independent of other clinical characteristics. Time-dependent ROC curves and clinical characteristics ROC curves were plotted and combined with risk scores to calculate the area under each curve (AUC) and construct C-Index curves. Finally, validation of the model’s prediction concerning patient survival was performed.

1.4 Nomogram construction

Incorporating risk classes and clinical risk characteristics, a nomogram was established for the prediction of 1-, 3-, and, 5-year survival rates of LUSC patients. In addition, the plotted calibration curves depicted the difference between nomogram-predicted outcomes and actual outcomes.

1.5 Principal component analysis (PCA) and risk difference analysis

PCA was performed to observe the performance of lncRNAs utilized in the establishment of the model in distinguishing patients between the risk groups. The two risk groups were assessed concerning differentially expressed genes (DEGs) through risk difference analysis. In addition, logFC > 0 indicated that the gene was enhanced in the high-risk group, whereas logFC < 0 indicated that the gene was upregulated in the low-risk group. Significant DEGs were obtained based on the criteria of |logFC|>1 and FDR < 0.05.

1.6 GO functional and KEGG pathway enrichment analyses

The names of DEGs were extracted and converted to gene IDs with P-value < 0.05 and corrected p-value < 0.05 as filtering criteria. The two risk groups were examined through Gene ontology (GO) for functional enrichment analysis of DEGs. KEGG enrichment analysis was then carried out to examine the enrichment of DEGs in the pathways. Finally, GO and KEGG circle maps were drawn.

1.7 Analysis of immune-related functions and tumor mutation burden (TMB)

The immune-related function difference between the two risk groups was analyzed, where P < 0.05 suggested the statistical significance of the differences. Gene mutation data files were obtained for the risk groups, and the mutation frequencies of the top 15 genes with the highest mutation frequency in the LUSC patient sample were compared and visualized (waterfall plot). Additionally, the association of risk score with TMB was determined. Survival difference between high- and low-TMB groups was comparatively analyzed by TMB survival analysis. Finally, TMB and patient risk were subjected to a joint analysis to plot survival curves. P < 0.05 suggested statistical significance.

1.8 Immunotherapy and potential drug screening

TIDE, representing tumor immune dysfunction and rejection, allows assessment of the possibility for tumor immune escape in the gene expression profile of LUSC samples. Higher TIDE scores indicate increased potential for immune escape and worse outcomes after receiving immunotherapy. The TIDE scores of the LUSC samples were downloaded from http://tide.dfci.harvard.edu/ to assess the difference in TIDE scores between high- and low-risk groups, where P < 0.05 suggested statistical relevance of the difference. Afterward, potential therapeutic agents were screened. Finally, the drug sensitivity difference between the two risk groups and its correlations with risk scores were calculated.

1.9 Statistical analysis

Perl was used to process the text files. LUSC transcriptome data, clinical data, and TCGA mutation burden data were collated using Perl scripts. Statistical processing and image plotting were executed through R language (version 4.2.1) (Table 2).

Table 2

Summary of the R packages used in each procedure of this study
Process	Packages
Extraction and collation of cuproptosis-related gene expression matrices and cuproptosis-related lncRNA expression, as well as combined cuproptosis-related lncRNA expression and survival data	limma
Sankey diagram	ggalluvial、ggplot2
Prognostic model construction	survival、caret、glmnet、survminer、timeROC
Heat map of the correlation between lncRNAs involved in prognostic model construction and cuproptosis-related genes	limma、reshape2、tidyverse、ggplot2
Functions for defining survival analysis and participation in plotting survival curves	survival、survminer
Functions for defining risk curves and plotting of risk curves, survival state diagrams, and risk heat maps	pheatmap
Independent prognostic analysis	survival
ROC curve plotting	survival、survminer、timeROC
C-index calculation and curve plotting	survival、rms、pec
Nomogram plotting, nomogram risk scoring, and calibration curve generation	survival、regplot、rms
Model validation for clinical groupings	survival、survminer
Definition of PCA analysis function and PCA graph plotting	limma、scatterplot3d
Risk difference analysis	limma
GO and KEGG enrichment analyses	colorspace、stringi、ggplot2、circlize、RColorBrewer、ggpubr
ssGSEA analysis, definition of ssGSEA score correction function, and heat map visualization	limma、GSVA、GSEABase、pheatmap、reshape2
waterfall plot generation	maftools
Differential analysis of tumor mutation burden, immune escape, and immunotherapy	limma、ggpubr
Drug sensitivity prediction, correlation testing, and potential drug screening	limma、ggpubr、pRRophetic、ggplot2、

2.1 Cuproptosis-related lncRNA identification and establishment of a prognostic model

Sixteen cuproptosis-related genes were determined from TCGA-LUSC samples. Based on the |R|>0.4 and P < 0.001 criteria, the 16,874 lncRNAs were filtered out to obtain 295 Cuproptosis-related lncRNAs. Visualization of the co-expression relationships of the genes and lncRNAs linked with cuproptosis was executed by Sankey diagrams (Fig. 1), where the genes and lncRNAs association was determined through their linkage. The 295 cuproptosis-related lncRNAs obtained above were subjected to one-way Cox regression analysis, yielding ten Cuproptosis-related lncRNAs with prognostic relevance (Fig. 2A). The eight lncRNAs with optimal efficacy for the prognostic model with the minimal error were obtained by Lasso regression ten-fold cross-validation (Fig. 2B, 2C). Afterward, a multivariate Cox analysis was executed to screen five cuproptosis-related lncRNAs that had the potential to predict the prognosis independently. Risk score was calculated as follows: Risk score = (0.50905619755965×AC010328.1 expression) + (0.255153440600426×LINC01740 expression) + (-0.37749067013803×AL358613.2 expression) + (0.538100881107615× MIR3945HG expression) + (-0.372555984713978×AC002467.1 expression).

2.2 Assessment of the prognostic model and value validation

The clinically relevant features did not depict any statistical difference in the training and test groups (p > 0.05), indicating no bias in clinical characteristics in the TCGA-LUSC sample subgroup (Table 3). The correlation between the cuproptosis-related genes and cuproptosis-related lncRNAs involved in the prognostic model in the TCGA dataset was assessed (Fig. 3). Kaplan-Meier survival analysis of the training group, the test group, and the whole dataset depicted that the group with lesser risk had better prognosis regarding the LUSC patients in contrast to the group with greater risk, with the two depicting considerable variations in OS and PFS (Fig. 4A-D). According to the risk curve, the risk score’s median value was 1, which could differentiate between patients of high- and low-risk (Fig. 5A, 5D, 5G). As the risk value rises in the survival status plot, the number of LUSC fatalities increases, while the survival time decreases (Fig. 5B, 5E, 5H). According to the results of the heat map, increased expression of AC010328.1, LINC01740, and MIR3945HG in the high-risk group, and increased AL358613.2 and AC002467.1 in the low-risk group were detected (Fig. 5C, 5F, 5I). The aforementioned data was also documented in the training group, the test group, and the entire dataset. Univariate and multivariate independent prognostic analyses depicted that the prognostic model, staging, and age have the potential to be utilized in prognosis predicting independent factors (Fig. 6A-B). The 1-, 3-, and 5-year survival AUC values for LUSC patients under risk score reached above 0.6, demonstrating the superior sensitivity and specificity of the model in the prediction of LUSC patients’ survival (Fig. 7A). The prognostic model had higher sensitivity and accuracy compared with the traditional clinicopathological characteristics (Fig. 7B). In addition, the C-index value of the risk score was also the largest, indicating the high agreement of the prognostic model-predicted patient survival with the actual clinical observation (Fig. 7C).

Table 3

Clinical statistical analysis of the subgroups
Covariates	Type	Total	Test	Train	P-value
Age	≤ 65	189 (38.18%)	58 (39.19%)	131 (37.75%)	0.828
	༞65	300 (60.61%)	88 (59.46%)	212 (61.1%)
	Unknown	6 (1.21%)	2 (1.35%)	4 (1.15%)
Gender	Female	129 (26.06%)	43 (29.05%)	86 (24.78%)	0.3794
Gender	Male	366 (73.94%)	105 (70.95%)	261 (75.22%)
Stage	Stage I	242 (48.89%)	76 (51.35%)	166 (47.84%)	0.2613
	Stage II	159 (32.12%)	38 (25.68%)	121 (34.87%)
	Stage III	83 (16.77%)	29 (19.59%)	54 (15.56%)
	Stage IV	7 (1.41%)	2 (1.35%)	5 (1.44%)
	Unknown	4 (0.81%)	3 (2.03%)	1 (0.29%)
T	T1	114 (23.03%)	38 (25.68%)	76 (21.9%)	0.6376
	T2	288 (58.18%)	86 (58.11%)	202 (58.21%)
	T3	70 (14.14%)	17 (11.49%)	53 (15.27%)
	T4	23 (4.65%)	7 (4.73%)	16 (4.61%)
M	M0	407 (82.22%)	129 (87.16%)	278 (80.12%)	1
	M1	7 (1.41%)	2 (1.35%)	5 (1.44%)
	Unknown	81 (16.36%)	17 (11.49%)	64 (18.44%)
N	N0	316 (63.84%)	95 (64.19%)	221 (63.69%)	0.2314
	N1	128 (25.86%)	33 (22.3%)	95 (27.38%)
	N2	40 (8.08%)	17 (11.49%)	23 (6.63%)
	N3	5 (1.01%)	2 (1.35%)	3 (0.86%)
	Unknown	6 (1.21%)	1 (0.68%)	5 (1.44%)

2.3 Construction of the prognostic nomogram and PCA

Seven factors including risk class and clinicopathological characteristics were included to construct a nomogram for survival prediction. A LUSC patient (ID: TCGA-98-A53C) was used as an example to plot the nomogram (Fig. 8A), wherein the 1-, 3-, and 5-year predicted survival rates of the patient were similar to the observed overall survival rate (Fig. 8B). The model constructed based on lncRNAs was further validated by PCA to have good grouping ability (Fig. 8F), and the distribution of high- and low-risk groups obtained based upon all genes as well as cuproptosis-related genes and lncRNAs was relatively dispersed (Fig. 8C-E). These data suggest that the model constructed using lncRNAs can effectively differentiate between low- and high-risk groups.

2.4 Enrichment analysis and immune-related functional analysis

GO analysis demonstrated the enrichment of DEGs in the two risk groups at three levels: biological processes, cellular components, and molecular functions. Enrichment of most of the DEGs was detected in eicosanoid metabolic process, phagocytosis, humoral immune response, recognition, and other processes. They are distributed on the plasma membrane’s external side immunoglobulin complex, the anchored component of the membrane and other cellular components, performing oxidoreductase activity, carboxylic acid binding, monocarboxylic acid binding, acting on the CH-OH group of donors, and other functions (Fig. 9A, Table 4). KEGG analysis suggests the association of these DEGs to the effects on xenobiotics metabolism by cytochrome P450, drug metabolism-cytochrome P450, tyrosine metabolism, and other pathways, suggesting the involvement of these DEGs in drug metabolism (Fig. 9B, Table 5). According to immune-related function analysis, such functions were significantly active in the group with greater risk (P < 0.05), except for the inhibition of the tumor suppressor gene (APC-co-inhibition), which was not significantly different between the risk groups (Fig. 9C).

Table 4

GO analysis results
Category	GO ID	Term	P-values	Related genes
GO-BP	GO:0006959	humoral immune response	1.23E-07	21
	GO:0006690	icosanoid metabolic process	1.04E-06	12
	GO:0006910	phagocytosis, recognition	1.28E-06	11
	GO:0001523	retinoid metabolic process	1.56E-06	10
	GO:0008037	cell recognition	1.89E-06	16
	GO:0016101	diterpenoid metabolic process	2.17E-06	10
GO-CC	GO:0019814	immunoglobulin complex	8.07E-10	18
	GO:0009897	external side of plasma membrane	1.26E-07	26
	GO:0031225	anchored component of membrane	1.61E-06	14
	GO:0045177	apical part of cell	5.88E-05	20
	GO:0016324	apical plasma membrane	6.12E-05	18
	GO:0042599	lamellar body	0.000179	4
GO-MF	GO:0033293	monocarboxylic acid binding	1.21E-08	12
	GO:0031406	carboxylic acid binding	4.73E-08	16
	GO:0016614	oxidoreductase activity, acting on CH-OH group of donors	8.32E-07	13
	GO:0016616		1.99E-06	12
	GO:0038024	cargo receptor activity	7.87E-06	9
	GO:0032052	bile acid binding	2.49E-05	4

Table 5

KEGG analysis results
Database type	ID	Description	P-values	Count
KEGG	hsa00980	Metabolism of xenobiotics by cytochrome P450	2.64E-08	12
	hsa00350	Tyrosine metabolism	4.66E-06	7
	hsa00982	Drug metabolism - cytochrome P450	9.15E-06	9
	hsa04514	Cell adhesion molecules	4.95E-05	12
	hsa00590	Arachidonic acid metabolism	0.000162	7
	hsa00360	Phenylalanine metabolism	0.000205	4
	hsa05204	Chemical carcinogenesis - DNA adducts	0.000352	7
	hsa04640	Hematopoietic cell lineage	0.000638	8
	hsa00480	Glutathione metabolism	0.00077	6
	hsa00830	Retinol metabolism	0.00195	6

2.5 TMB analysis and potential drug screening

In contrast to the group with higher risk values, the less risky group depicted slightly more frequent mutation of most genes (98.41% vs 96.88%) (Fig. 10A). In addition, tumor mutation burden did not depict a considerable difference in the two risk groups (P = 0.83). Though, differences in survival rates were detected between the high- and low-TMB groups, with individuals with high TMB having a remarkably better prognosis than those with low TMB (P < 0.05) (Fig. 10B). The combined TMB and risk score depicted considerably longer median survival time in the high-TMB low-risk group (P < 0.05) (Fig. 10C). Additionally, the TIDE algorithm further revealed greater TIDE scores in the group with higher risk than in the less risky group, depicting a higher risk of immune escape and worse immunotherapeutic outcomes in patients in the high-risk group (P < 0.05) (Fig. 11A). Finally, all antitumor drugs known to be used in clinical settings were screened to analyze the sensitivity of these drugs treating LUSC. The IC50 values of quizartinib (AC220), dasatinib, phenformin, rapamycin, and sunitinib were reduced in the groups with greater risk, implying the high sensitivity of these drugs in high-risk LUSC patients (Fig. 11B-F).

Cuproptosis is receiving increasing attention as a new mode of cell death and is associated with tumor proliferation and growth, angiogenesis, and metastasis [9]. In addition, lncRNAs influence the development of cancer and its treatment through biological means [19]. LncRNAs play an important role in NSCLC (including LUSC) [20–22] and could act as potentially efficient molecular targets for NSCLC treatment (including LUSC). The mechanism of cuproptosis regulation in lung cancer is currently unknown due to few studies regarding cuproptosis and lncRNAs’ co-regulatory role in lung malignancy. In the present research, the copper toxicity-associated lncRNAs were systematically identified per the involvement of lncRNAs and cuproptosis-related mRNAs in LUSC.

cuproptosis-related genes in LUSC were accessed at the database TCGA, and these genes have been reported in current cuproptosis-related studies. Polishchuk EV [23] found that copper overload following ATP7B deletion triggers autophagy in hepatocytes, and activation of autophagy involves the Mammalian Target of Rapamycin (MTP) and TFEB signaling pathways. However, autophagy activation attenuates copper toxicity in ATP7B-deficient hepatocytes and tissues. KRAS mutations lead to abnormal activation of the RAS/MAPK pathway, which causes evasion of the immune response and alteration of the tumor microenvironment, whereas ATP7A is upregulated in KRAS-mutated colorectal cancers and protects cells from excessive copper toxicity [24]. These findings suggest a protective role of ATP7A against KRAS mutant cells.

cuproptosis-related lncRNAs were identified based on co-expression relationships of LUSC cuproptosis-related genes. Subsequently, five of these lncRNAs, including AC010328.1, LINC01740, AL358613.2, MIR3945HG, and AC002467.1, were identified as independent prognostic factors. Hence, a prognostic model was established based on these five lncRNAs. Wang Y [25] et al. established a LUSC-related lncRNA-miRNA-mRNA network by integrating gene and miRNA expression data of LUAD. They finally identified four predictive genes, including LINC01740, which may serve as potential prognostic markers. MIR3945HG has been identified as a new candidate diagnostic marker for tuberculosis. MIR3945HG is the biomarker with the highest diagnostic value for diagnosing LUSC and is strongly linked with survival in patients with LUSC [26]. Wang Y [27] found that MIR3945HG expression was downregulated both in bioinformatics and qRT-PCR validation. In the present study, however, MIR3945HG was expressed higher in the high-risk group. It can be speculated that the amplification region of the target gene, amplification efficiency, and reagents involved in qPCR all affect the final expression. In addition, the possible GC preference of RNA-Seq sequencing may also affect the final expression and cause differences in the results. Therefore, further research is needed for the validation of the reliability of these results. According to the PubMed search, published reports on AC010328.1, AL358613.2, and AC002467.1 are currently unavailable.

This research did not depict any remarkable differences in TMB in the two risk groups. However, the high- and low-TMB groups varied concerning survival, suggesting that TMB and prognosis of LUSC patients were correlated. Li L [28] et al. explored the association of TMB and the prognosis of 33 tumors with the resulting survival data’s subsequent classification per the TMB levels into low- and high-TMB groups. Their results, consistent with this research’s findings, implied that better outcomes could be associated with high TMB LUSC patients. There are no uniform criteria for high and low TMB cut-offs. In addition, a TMB threshold that effectively predicts immunotherapy response in LUSC has not been set up yet. The association between the response to immunotherapy and TMB levels requires detailed research. Therefore, the findings of the present study may need to be updated and validated based on future studies. This study predicted sensitive drugs indicated for patients with high-risk LUSC, including quizartinib (AC220), dasatinib, phenformin, rapamycin, and sunitinib. Quizartinib, a selective FLT3 inhibitor, allows for precise treatment and overcoming acquired resistance. In this trial, TP53 was the most altered gene in the LUSC samples (percentage in the high and low-risk groups: 76% each). Pop-Bica C [29] performed next-generation sequencing (NGS) experiments on whole blood samples from 32 lung cancer patients, revealing that FLT3 was the gene most frequently altered (more than 90%). Dasatinib is indicated for adult patients with all stages of chronic myeloid leukemia (CML) who are resistant or intolerant to treatment regimens including imatinib mesylate. In a recent study, dasatinib synergized with anti-PD-1 to inhibit tumor growth in an experimental model of NSCLC [30]. In addition, a clinical study has shown that metformin and other biguanides (buformin and phenformin) have reduced the risk of several cancer types, including colorectal, lung, pancreatic, breast, and liver cancer [31]. Guo Z [32] utilized ErbB2 overexpressing breast cancer cells and animal models to examine the anticancer potential of phenformin. His findings suggest that phenformin-mediated IGF1/IGF1R modulation could possibly function as an anti-cancer mechanism. These findings provide ideas for the development of phenylephrine as a therapeutic agent for LUSC.

This research dealt with constructing a LUSC prognostic model based on cuproptosis-related lncRNAs (AC010328.1, LINC01740, AL358613.2, MIR3945HG, and AC002467.1). In addition, the association between tumor mutation burden, immunotherapy, and drug sensitivity was observed by risk score grouping. Overall, the findings of this research provide novel knowledge toward the assessment of LUSC patients’ survival and expansion of clinical drug use.

AUC: Area under each curve

CDN: Copper-depleting nanoparticle

DEG: Differentially expressed gene

GO: Gene Ontology

KEGG: Kyoto Encyclopedia of Genes and Genomes

lncRNA: Long non-coding RNA

LIPA: Lipoylated protein aggregation

LUSC: Lung squamous cell carcinomas

ncRNA: non-coding RNA

NGS: next-generation sequencing

OS: Overall survival

PCA: Principal component analysis

PFS: Progression-free survival

TCA: Tricarboxylic acid cycle

TCGA: The Cancer Genome Atlas

TNBC: Triple-negative breast cancer

TMB: Tumor mutation burden

Availability of data and materials

The datasets associated with LUSC in this study were available in the public database TCGA GDC(https://portal.gdc.cancer.gov/).

Acknowledgements

The authors would like to thank all the reviewers who participated in the review and Bullet edits (http://www.bulletedits.cn/) for its linguistic assistance during the preparation of this manuscript. Thanks to all the contributors of data to The Cancer Genome Atlas (TCGA).

Author’s Contributions

Zhe Tian and Lilan Cen played a major role in analyzing the data, and wrote the manuscript; Haoming Hua Modified the articles and contributed to the manuscriptJue; Dong and Yulan Hung collected the data; Chunyan Qin and Junhua Deng performed the visualization operation of correlation analysis; Yujie Jiang designed the experiments and polished the manuscript.

Funding

This study was supported by a grant from the National Natural Science Foundation of China (Grant No. 81860021), Guangxi Natural Science Foundation (2021GXNSFAA325003), Baise Regional Multimorbidity Joint Special Plan (Baike Zi [2022] No.41), Guangxi University Young and Middle-aged Teachers Research Basic Ability Promotion Project (2023KY0573).

Availability of data and materials

The authors declare that all data supporting the findings of this study are available within the article. These data include gene expression and clinically relevant data of LUSC, which can be downloaded with TCGA (https:// portal.gdc.cancer.gov/).

Ethics approval and consent to participate

No ethical certifcate is required.

Consent for publication

Not applicable.

Competing interests

The authors declare that the research was conducted in the absence of any commercial or fnancial relationships that could be construed as a potential confict of interest. All authors revised the content and approved the fnal manuscript.

Author details

¹Department of Pulmonary and Critical Care Medicine, Affiliated Hospital of Youjiang Medical University for Nationalities, Baise, 533000, China. ²Department of Pathology, The Second People's Hospital of Bengbu, 233400, Anhui, China. ³Guangxi Academy of Medical Sciences, Department of Infectious Disease, the People's Hospital of Guangxi Zhuang Autonomous Region, Nanning, 530021, China.

Mercer TR, Dinger ME, Mattick JS. Long non-coding RNAs: insights into functions. Nat Rev Genet. 2009;109(3):155-9.
St Laurent G, Wahlestedt C, Kapranov P. The Landscape of long noncoding RNA classification. Trends Genet. 2015;31(5):239-51.
Peng W, Wang J, Shan B, et al. Diagnostic and Prognostic Potential of Circulating Long Non-Coding RNAs in Non Small Cell Lung Cancer. Cell Physiol Biochem. 2018;49(2):816-27.
Akhade VS, Pal D, Kanduri C. Long Noncoding RNA: Genome Organization and Mechanism of Action. Adv Exp Med Biol. 2017;1008:47-74.
Ricciuti B, Mencaroni C, Paglialunga L, et al. Long noncoding RNAs: new insights into non-small cell lung cancer biology, diagnosis and therapy. Med Oncol. 2016;33(2):18.
Arredondo M, Núñez MT. Iron and copper metabolism. Mol Aspects Med 2005;26(4-5):313-27.
Tsvetkov P, Coy S, Petrova B, et al. Copper induces cell death by targeting lipoylated TCA cycle proteins. Science. 2022;375(6586):1254-61.
Chang WH, Lin HH, Tsai IK, et al. Copper Centers in the Cryo-EM Structure of Particulate Methane Monooxygenase Reveal the Catalytic Machinery of Methane Oxidation. J Am Chem Soc. 2021;143(26):9922-32.
Oliveri V. Selective Targeting of Cancer Cells by Copper Ionophores: An Overview. Front Mol Biosci. 2022;9:841814.
Cui L, Gouw AM, LaGory EL, et al. Mitochondrial copper depletion suppresses triple-negative breast cancer in mice. Nat Biotechnol. 2021;39(3):357-67.
Chen G, Niu C, Yi J, et al. Novel Triapine Derivative Induces Copper-Dependent Cell Death in Hematopoietic Cancers. J Med Chem. 2019;62(6):3107-21.
Su W, Feng S, Chen X, et al. Silencing of Long Noncoding RNA MIR22HG Triggers Cell Survival/Death Signaling via Oncogenes YBX1, MET, and p21 in Lung Cancer. Cancer Res. 2018;78(12):3207-19.
Ma Z, Peng P, Zhou J, et al. Long Non-Coding RNA SH3PXD2A-AS1 Promotes Cell Progression Partly Through Epigenetic Silencing P57 and KLF2 in Colorectal Cancer. Cell Physiol Biochem. 2018;46(6):2197-214.
Xin X, Wu M, Meng Q, et al. Long noncoding RNA HULC accelerates liver cancer by inhibiting PTEN via autophagy cooperation to miR15a. Mol Cancer. 2018;17(1):94.
Wei GH, Wang X. lncRNA MEG3 inhibit proliferation and metastasis of gastric cancer via p53 signaling pathway. Eur Rev Med Pharmacol Sci. 2017;21(17):3850-6.
Dan J, Wang J, Wang Y, et al. LncRNA-MEG3 inhibits proliferation and metastasis by regulating miRNA-21 in gastric cancer. Biomed Pharmacother. 2018;99:931-8.
Zhao W, Geng D, Li S, Chen Z, Sun M. LncRNA HOTAIR influences cell growth, migration, invasion, and apoptosis via the miR-20a-5p/HMGA2 axis in breast cancer. Cancer Med. 2018;7(3):842-55.
Han M, Wang S, Fritah S, et al. Interfering with long non-coding RNA MIR22HG processing inhibits glioblastoma progression through suppression of Wnt/β-catenin signalling. Brain. 2020;143(2):512-30.
Tan YT, Lin JF, Li T, Li JJ, Xu RH, Ju HQ. LncRNA-mediated posttranslational modifications and reprogramming of energy metabolism in cancer. Cancer Commun (Lond). 2021;41(2):109-20.
Hua Q, Jin M, Mi B, et al. LINC01123, a c-Myc-activated long non-coding RNA, promotes proliferation and aerobic glycolysis of non-small cell lung cancer through miR-199a-5p/c-Myc axis. J Hematol Oncol. 2019;12(1):91.
Pan J, Fang S, Tian H, et al. lncRNA JPX/miR-33a-5p/Twist1 axis regulates tumorigenesis and metastasis of lung cancer by activating Wnt/β-catenin signaling. Mol Cancer. 2020;19(1):9.
Li J, Xia R, Liu T, Cai X, Geng G. LncRNA-ATB Promotes Lung Squamous Carcinoma Cell Proliferation, Migration, and Invasion by Targeting microRNA-590-5p/NF90 Axis. DNA Cell Biol. 2020;39(3):459-73.
Polishchuk EV, Merolla A, Lichtmannegger J, et al. Activation of Autophagy, Observed in Liver Tissues From Patients With Wilson Disease and From ATP7B-Deficient Animals, Protects Hepatocytes From Copper-Induced Apoptosis. Gastroenterology. 2019;156(4):1173-89.e5.
Aubert L, Nandagopal N, Steinhart Z, et al. Copper bioavailability is a KRAS-specific vulnerability in colorectal cancer. Nat Commun. 2020;11(1):3701.
Wang Y, He R, Ma L. Characterization of lncRNA-Associated ceRNA Network to Reveal Potential Prognostic Biomarkers in Lung Adenocarcinoma. Front Bioeng Biotechnol. 2020;8:266.
Chen W-J, Tang R-X, He R-Q, et al. Clinical roles of the aberrantly expressed lncRNAs in lung squamous cell carcinoma: a study based on RNA-sequencing and microarray data mining. Oncotarget. 2017;8(37):61282-304.
Wang Y, Fu J, Wang Z, Lv Z, Fan Z, Lei T. Screening key lncRNAs for human lung adenocarcinoma based on machine learning and weighted gene co-expression network analysis. Cancer Biomark. 2019;25(4):313-24.
Li L, Bai L, Lin H, et al. Multiomics analysis of tumor mutational burden across cancer types. Comput Struct Biotechnol J. 2021;19:5637-46.
Pop-Bica C, Ciocan CA, Braicu C, et al. Next-Generation Sequencing in Lung Cancer Patients: A Comparative Approach in NSCLC and SCLC Mutational Landscapes. J Pers Med. 2022;12(3):453.
Redin E, Garmendia I, Lozano T, et al. SRC family kinase (SFK) inhibitor dasatinib improves the antitumor activity of anti-PD-1 in NSCLC models by inhibiting Treg cell conversion and proliferation. J Immunother Cancer. 2021;9(3):e001496.
Buttrus RN, Haki K, Eynon MF, et al. Synthesis of biguanides and an investigation of their therapeutic effects on brain tumours. Neuro Oncol. 2018;20(Suppl 1):i22.
Guo Z, Zhao M, Howard EW, et al. Phenformin inhibits growth and epithelial-mesenchymal transition of ErbB2-overexpressing breast cancer cells through targeting the IGF1R pathway. Oncotarget. 2017;8(36):60342-57.

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Effectiveness of Cuproptosis-related long non-coding RNAs in predicting prognosis and immune response in patients with lung squamous cell carcinoma

Status:

Version 1

Abstract

Background

Methods

Results

Conclusion

Figures

Introduction

1 Material and methods

1.1 Dataset and preprocessing

1.2 Construction of cuproptosis-related lncRNA prognostic model

1.3 Assessment of the prognostic risk model

1.4 Nomogram construction

1.5 Principal component analysis (PCA) and risk difference analysis

1.6 GO functional and KEGG pathway enrichment analyses

1.7 Analysis of immune-related functions and tumor mutation burden (TMB)

1.8 Immunotherapy and potential drug screening

1.9 Statistical analysis

2 Results

2.1 Cuproptosis-related lncRNA identification and establishment of a prognostic model

2.2 Assessment of the prognostic model and value validation

2.3 Construction of the prognostic nomogram and PCA

2.4 Enrichment analysis and immune-related functional analysis

2.5 TMB analysis and potential drug screening

3 Discussion

4 Conclusion

Abbreviations

Declarations

References

Additional Declarations

Status:

Version 1