A Newly Established Cuproptosis-Related Long Noncoding RNA Signature for Predicting Prognosis and the Immune Microenvironment in Stomach Adenocarcinoma

doi:10.21203/rs.3.rs-2835696/v1

Download PDF

Article

A Newly Established Cuproptosis-Related Long Noncoding RNA Signature for Predicting Prognosis and the Immune Microenvironment in Stomach Adenocarcinoma

https://doi.org/10.21203/rs.3.rs-2835696/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Cuproptosis is a form of cell death that is considered as a new direction for future cancer therapy. It has been less reported in stomach adenocarcinoma (STAD). Download transcriptomic data, genomic mutation data and clinical data for STAD from the TCGA website. We identified 430 cuproptosis-related long noncoding RNA (LncRNAs). Cox models screened 8 cuproptosis-related lncRNAs (AC005050.3, AC016394.2, AL606970.1, BX890604.1, C5orf66, HAGLR, LINC01094, TDRKH- AS1), and the risk model were constructed. Based on the median risk score, STAD patients were divided into a high-risk group and a low-risk group. kaplan-Meier survival curves showed that the low-risk group had a higher survival rate than the high-risk group. Operating characteristic curve and principal component analysis showed that the model had better predictive power. The tumor mutational burden was higher in the low-risk group, suggesting greater sensitivity to immunotherapy, and subsequently screening for sensitive drugs in different risk groups. Finally, qPT-PCR results showed that eight lncRNAs were highly expressed in STAD cells. In conclusion, these 8 LncRNAs may be potential biomarkers for predicting the prognosis of STAD patients and assessing the effect of immunotherapy and drug sensitivity in the future.

Stomach adenocarcinoma (STAD) is the most common histologic type of gastric cancer, ranking fifth by frequency of occurrence and third among the deadliest cancers. Due to its advanced clinical presentation, genetic heterogeneity, and chemotherapy resistance mechanisms, stomach adenocarcinoma has a high mortality rate¹. As a result, strategies to prevent and treat gastric cancer are actively pursued, including surgical resection, targeted drug therapy, and other options, but the 5-year survival rate for patients with advanced gastric adenocarcinoma remains low². It is therefore crucial to find reliable biological signatures for predicting the prognosis of STAD patients and to develop effective treatment options based on these signatures.

Tsvetkov et al. proposed a mechanism for cuproptosis in March of this year[22].Cuproptosis is a new type of cell death that is copper dependent, regulated and different from other known cell death regulatory mechanisms. Notably, cancer cells have a higher demand for copper than normal cells; Some cancers express a lot of thioacylated mitochondrial proteins and show high intensity of respiration^4,5. Based on this novel mechanism of cell death, some scholars have recently started to study it in different tumors^6–8.

Long non-coding RNAs (LncRNAs) are RNAs that do not contain protein-coding sequences and range in length from 200 nucleotides to several thousand bases⁹. LncRNAs may play a crucial role in gene regulation, as well as being promising diagnostic and prognostic markers for tumors^10,11. LncRNA expression has been demonstrated to be an effective prognostic marker for a variety of cancers in numerous studies^12–14. However, the relationship between cuproptosis-related LncRNAs and prognosis of stomach adenocarcinoma patients has not yet been explored. To provide a new basis for future diagnosis and treatment of stomach adenocarcinoma patients, this study investigated prognosis-related cuproptosis LncRNA signatures.

In this study, we downloaded the transcriptomic data, clinicopathological data, survival data and tumor somatic mutation data of STAD from the official website of TCGA. STAD patients were randomized into training and validation groups. We identified cuproptosis-related prognostic signatures and developed a prognostic model. To select appropriate treatment strategies for gastric adenocarcinoma in the future, a predictive model was evaluated for its predictive ability and clinical value. Finally, the model was shown to be capable of predicting immunotherapy and drug sensitivity in STAD patients.

Data Collection

We download transcriptomic data, clinicopathological data, survival data, and tumor somatic mutation data of stomach adenocarcinoma from the TCGA-STAD database ¹⁵(https://portal.gdc.cancer.gov/). The transcriptomic data for 407 tissues were downloaded (32 normal samples and 375 tumor samples). Clinicopathological characteristics such as age, sex, TNM stage, clinical stage, and grade were collected. Annotations of LncRNAs were obtained from the GENCODE website (https://www.gencodegenes.org/). Also, cuproptosis-related gene were obtained by referring to previous literature³.

Identification of Cuproptosis-Related LncRNA

In R, the "Limma" package was used to analyze the coexpression of cuproptosis-related genes with lncRNAs. The conditions of |Cor|>0.4 and P < 0.001 were met, indicating an association. Our Sankey diagram was created using the R packages "ggplot2", "ggalluvial", and "dply" based on the results of the coexpression analysis.

Construction of the Cuproptosis-Related LncRNA prediction model

The STAD patients were randomized into training and validation groups using the "caret" package in R. Univariate Cox proportional risk regression was performed for each cuproptosis-related lncRNAs with survival data using the "survival" R package. To avoid overfitting, we used the "glmnet" package in R software to perform the least absolute shrinkage and selection operator (LASSO)-penalized Cox regression¹⁶. The optimal and minimum criteria for the penalty (λ) using 10 times cross-validation were selected. In order to identify the cuproptosis-related lncRNAs prognostic signature, multiple stepwise Cox regression analyses were performed. Subsequently, a risk scoring model is constructed according to the following equation:

$$Risk score=\sum _{i=1}^{n}coe{f}_{i}*{Expr}_{i}$$

Wherewhere $coe{f}_{i}$ is the coefficient and ${Expr}_{i}$ is the expression value of the corresponding cuproptosis-related lncRNAs. Risk score for each patient according to our established LASSO prognostic model.

Validation of Risk Models

Based on the median risk score and the corresponding coefficient of the training group, the patients in the training group, validation group and entire cohorts were categorized into high- and low-risk groups. In the training group, validation group, and entire cohort, Kaplan–Meier analysis with a log-rank test was used to compare overall survival (OS) between the high-risk and low-risk groups. Cuproptosis-related lncRNA prognostic signatures were evaluated using the area under the curve (AUC) and the subject operating characteristic curve (ROC). We validated the risk model using principal component analysis (PCA) and visualized the results using "scatterplot3D" in R. Progression-free survival (PFS) analysis was performed using the "survival" and "Survminer" packages in R. Using the R packages "rms," "ply," “survival,” and “pec”, we calculated the C-index to predict risk model accuracy.

Establishing and Evaluating a Nomogram

We used univariate and multivariate Cox regression to assess the independent prognostic effect of the risk model. Based on the univariate and multivariate Cox regression results, we used the "rms", "regplot", and "survival" packages in R to construct nomograms to predict the OS (1-, 3-, and 5-year) of STAD patients. Calibration curves were used to assess nomogram accuracy.

Assessing the Relationship Between the Risk score and Clinicopathological Characteristics

To validate the relationship between the model and clinicopathological characteristics, we analyzed the OS rates of patients in high- and low-risk groups with different clinical characteristics (e.g., age, gender, and clinical stage), and the results were expressed as Kaplan‒Meier curves.

Pathway Enrichment Analysis and Gene set Enrichment Analysis

R's "Limma" package was used to identify differentially expressed genes (DEGs) between high-risk and low-risk groups with log₂ |fold change| >1 and false discovery rate < 0.05¹⁷. Using the "clusterProfiler", "org.Hs.eg.db" and "enrichplot" packages in R, we explored the Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) database pathways to elucidate molecular functions and key signaling pathways.

Estimation of Intratumoral Immune Cell Infiltration and Immunotherapy

To explore the relationship between risk score and immune cell infiltration in the prognostic model, the single-sample gene set enrichment analysis (ssGSEA) algorithms function package in the R software genome variation analysis package was used to evaluate the infiltration and function of tumor-infiltrating immune cells¹⁸. We drew the related heat map. Application of the tumor immune dysfunction and exclusion (TIDE) algorithm to predict response to immunotherapy based on simulation of tumor immune escape mechanisms (http://tide.dfci.harvard.edu)¹⁹. As a result, we examined the effects of immunotherapy based on the TIDE algorithm in high-risk and low-risk patients.

Drug Sensitivity Assessment

IC50 represented the semiinhibitory concentration of the measured antagonist. Through the use of the R package “pRRophetic” and its dependencies including “car, ridge preprocessCore, genefilter and sva”, we calculated the IC50 of the chemotherapeutic drugs to determine the efficacy of cuproptosis-related prognostic signature in STAD clinical treatment²⁰. There were 138 drugs included, including dasatinib and tipifarnib. The differences in the IC50 of chemotherapeutic drugs between the high-risk and low-risk groups were compared using the Wilcoxon sign rank test. R's "ggplot2" package was used to create the boxplots that represented the results.

Calculation of Tumor Mutation Burden Scores

Using the "maftools" package in R, we analyzed the mutation data from the TCGA-STAD database. Waterfall diagrams were used to show the relationship between the risk score and tumor mutation burden scores (TMB) in STAD patients. In the waterfall diagram, we show the 15 genes with the highest mutation frequency.

Cell culture and qRT-PCR

Human normal gastric mucosal epithelial cells GES-1 and human gastric cancer cells (GC) HCG-27, SGC-7901, and MKN-45 were purchased from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). Cells were cultured in RPMI-1640 medium (Gibco, New York, USA) with antibiotics (1% penicillin/streptomycin, Gibco, New York, USA) and 10% FBS (Gibco, New York, USA). The cultures were incubated in a sterile constant temperature incubator containing 5% CO2 at 37°C and 70–80% humidity. Referring to the protocol of Li et al[16]. Total RNA was extracted with TRIzol reagent (Thermo Fisher Scientific, USA), and 1000 ng of RNA was reverse transcribed into cDNA by a reverse transcription system (Takara, Japan). Then, mRNA expression levels were detected according to the SYBR Green Master Mix (Thermo Fisher Scientific, USA) protocol on an ABI Prism 7500 assay system (Applied Biosystems, Inc., USA). 18S RNA was used as an internal reference. The expression of each transcript was then calculated using the 2-∆∆Ct method. The qRT‒PCR primers are listed in Table 1.

Table 1

qRT‒PCR primers
Gene symbol	Forward (5'-3')	Reverse (5'-3')
AC005050.3	GTGCGGGATTTAGGGTTTGC	TGAGTCCTGGTGGAAAAGGC
AC016394.2	ACCCGAAGGAAGACTCCTCT	ATCCTGGTGTGCAGAATGGG
AL606970.1	CTCCCCTCCACCATGAAACC	AGTAACTGACACAAGGCCCG
BX890604.1	AAGACACCTCGGAAGGCG	AGCCTCGACCCTGTAAGAGA
C5orf66	CTTCAGCATCAGTCTGCCTTTG	GGGTCGAGAGGAATCTTGGG
HAGLR	TAAGAAGGTTCTGGGCGGGA	CACTTCGTGTCATCGGGGAT
LINC01094	GCCCAGGAGCATTTACCAGT	GGAATGAGCCCCTCAGACAC
TDRKH − AS1	ACTCAGTGACCATCGGGAGA	TGTGATGGAAGCTAGTGCGG
18sRNA	GTAACCCGTTGAACCCCATT	CCATCCAATCGGTAGTAGCG

Statistical Analysis

All analyses were performed using R v.4.1.2 (http://www. R-project.org). The "limma" package in R was used to combine transcriptional data and somatic mutation data from TCGA-STAD. The Pearson correlation test was used to analyze the correlation between cuproptosis-related genes and cuproptosis-related lncRNAs. The "limma" package in R was used to screen for differentially expressed genes in the cuproptosis-related prognostic signature. The "survival" and "survminer" packages in R were used for Cox regression and survival analysis. Hazard ratios were calculated using Cox proportional risk regression models for univariate and multifactorial analyses. GO terms and KEGG pathways were analyzed using the "clusterProfiler" package in R. Heatmaps for cluster analysis were drawn using the "Pheatmap" R package. Differences in quantitative data between the two groups were compared using the Wilcoxon rank-sum test. Differences in quantitative data between multiple groups were compared by one-way ANOVA. The categorical information between the two groups was compared using the chi-square test. Statistical significance was defined as a p value of 0.05.

Identification of Cuproptosis-related LncRNA in STAD Patients

Using the GENCODE database, we identified a total of 16,876 LncRNAs in the TCGA-STAD dataset. We referenced a total of 19 cuproptosis-related genes collected in the literature³. A total of 430 cuproptosis-related LncRNAs were screened by Pearson correlation test. In Fig. 1A, Sankey diagrams illustrate the relationship between cuproptosis-related genes and LncRNAs. Next, univariate Cox regression analysis was used to explore cuproptosis-related LncRNAs (p < 0.05), and we screened a total of 18 cuproptosis-related LncRNAs (Fig. 1B). We divided 407 STAD patients into training and validation groups, and the corresponding clinical traits of the two groups is shown in Table 2. There were no differences between the training and validation groups in clinical traits.

Table 2

Clinical characteristics of STAD patients in the TCGA database
Variable	Entire cohort (n = 371)	Training cohort (n = 186)	Validation cohort (n = 185)	p value
Age
≤ 65	163(43.94%)	84(45.16%)	79(42.7%)	0.7438
༞65	205(55.26%)	101(54.3%)	104(56.22%)	-
unknow	3(0.81%)	1(0.54%)	2(1.08%)	-
Gender
FEMALE	133(35.85%)	70(37.63%)	63(34.05%)	0.5414
MALE	238(64.15%)	116(62.37%)	122(65.95%)	-
Grade				0.1437
G1	10(2.7%)	8(4.3%)	2(1.08%)	-
G2	134(36.12%)	66(35.48%)	68(36.76%)	-
G3	218(58.76%)	105(56.45%)	113(61.08%)	-
unknow	9(2.43%)	7(3.76%)	2(1.08%)
Stage				0.266
Stage I	50(13.48%)	24(12.9%)	26(14.05%)	-
Stage II	111(29.92%)	50(26.88%)	61(32.97%)	-
Stage III	149(40.16%)	77(41.4%)	72(38.92%)	-
Stage IV	38(10.24%)	24(12.9%)	14(7.57%)	-
unknow	23(6.2%)	11(5.91%)	12(6.49%)	-
T				0.6726
T1	18(4.85%)	8(4.3%)	10(5.41%)	-
T2	78(21.02%)	40(21.51%)	38(20.54%)	-
T3	167(45.01%)	80(43.01%)	87(47.03%)	-
T4	100(26.95%)	55(29.57%)	45(24.32%)	-
unknow	8(2.16%)	3(1.61%)	5(2.7%)	-
M				0.3983
M0	328(88.41%)	161(86.56%)	167(90.27%)	-
M1	25(6.74%)	15(8.06%)	10(5.41%)	-
unknow	18(4.85%)	10(5.38%)	8(4.32%)	-
N				0.6169
N0	108(29.11%)	48(25.81%)	60(32.43%)	-
N1	97(26.15%)	51(27.42%)	46(24.86%)	-
N2	74(19.95%)	39(20.97%)	35(18.92%)	-
N3	74(19.95%)	36(19.35%)	38(20.54%)	-
unknow	18(4.85%)	12(6.45%)	6(3.24%)	-

Construction and Validation of a Risk Score Model for STAD Patients Based on the Cuproptosis-related LncRNA Prognostic Signature

The 18 cuproptosis-related lncRNAs from the previous step were further screened by LASSO Cox regression analysis, resulting in 15 lncRNAs screened in total. The trajectory changes in regression coefficients of lncRNAs were identified, as were the cross-validation results of model construction (Fig. 2a, b). Eight lncRNA signatures linked to cuproptosis were screened using multiple stepwise Cox regression analysis, and risk score models were built using these signatures. To assess the correlation between cuproptosis-related genes and prognostic signatures, we plotted a heatmap (Fig. 2c). Risk score = (0.612310337796917* Expression _AC005050.3) + (1.66740693091285* Expression _C5orf66) + (0.398655774690665* Expression _HAGLR) + (-0.485279771224916* Expression _TDRKH−AS1) + (0.549647873930273* Expression _LINC01094) + (-0.525617348023318* Expression _BX890604.1) + (-0.680536241533385* Expression _AC016394.2) + (-0.483203374377389* Expression _AL606970.1). Based on the median risk score, we divided STAD patients into high-risk and low-risk groups and analyzed the OS of patients in the high- and low-risk groups in the training and validation cohorts using the Kaplan‒Meier method. In the high-risk group, OS was significantly lower than in the low-risk group (P < 0.05; Fig. 2d, e).

Figure 3 shows a positive correlation between the risk score and the number of deaths and a negative correlation with patient survival time. Figure 3a-c shows the relationship between the risk score and patient survival status. We see an increasing number of patient deaths as the risk score increases. The expression levels of eight cuproptosis-related prognostic signatures in the different risk groups were plotted using heatmaps (Fig. 3d-f). PCA also revealed significant differences between the high- and low-risk groups. Based on the risk model constructed with cuproptosis-related lncRNAs, we observed that STAD patients were clearly divided into two groups (Fig. 4a-d).

Cuproptosis-related LncRNA Prognostic Signatures are Independent Prognostic Factors in STAD Patients

We used univariate and multivariate Cox regression analyses to assess the prognostic value of the risk model. In univariate and multivariate Cox regression analyses, age, stage, and risk score were prognostically relevant independent risk factors (P < 0.05; Fig. 5a, b). Next, we further evaluated the accuracy of the risk score, age, sex, tumor grading and staging model in predicting patient prognosis using ROC and C-index curves. As shown in Fig. 5c and 5d, the risk score model was more accurate than the other clinical prediction models, indicating that this risk score model was more predictive of patient prognosis and was more correlated with patient prognosis. Subsequently, we further evaluated the accuracy of the model for predicting 1-, 3-, and 5-year survival in STAD patients using TimeROC curves. As shown in Fig. 5e, the AUC was 0.721 at 1 year, 0.645 at 3 years, and 0.713 at 5 years, which indicated that the model was accurate. In addition, PFS showed a statistically significant difference in progression-free survival between the high-risk and low-risk groups (P < 0.05; Fig. 5f), which further revealed that the model has good diagnostic value for predicting patient prognosis.

Nomograms have become a common predictive tool for analyzing patients' clinical outcomes. Based on the risk score and some clinicopathological features, we produced a nomogram (Fig. 6a). We found that age and risk score were significantly associated with patient prognosis (P < 0.001; Fig. 6a). In the subsequent calibration plots, the calibration plots matched well with the nomogram predictions (Fig. 6b).

Relationship Between Prognostic Signature and Clinicopathological Features in Stomach Adenocarcinoma

Using the above univariate and multivariate ANOVAs, we found that age and tumor stage were associated with patient prognosis. To explore the clinical value of the risk score model, we explored the relationship between the risk score model and age and tumor stage. The results indicated that there were significant differences between the distribution of risk score and age and tumor stage. In terms of age, there was a statistically significant difference between patients aged ≤ 65 years and > 65 years (p < 0.05; Fig. 7a, b); in terms of tumor stage, there was a statistically significant difference between stage I-II and stage III-IV (p < 0.05; Fig. 7c, d).

Functional Enrichment Analysis

We used the "Limma" package in R to screen the DEGs in the high-risk and low-risk groups for GO and KEGG enrichment analysis. There were 374 DEGs identified. Based on GO analysis, DEGs were mainly related to muscle system process, muscle contraction, and regulation of muscle system process in the category of biological process. In the category of cellular components, DEGs were mainly focused on collagen − containing extracellular matrix, immunoglobulin complex and contractile fiber. In the molecular function category, DEGs were mainly involved in glycosaminoglycan binding, heparin binding, and extracellular matrix structural constituent (Fig. 8a, c, e). KEGG enrichment analysis revealed that DEGs were enriched in vascular smooth muscle contraction, dilated cardiomyopathy, focal adhesion, and hypertrophic cardiomyopathy (Fig. 8b, d, f).

Evaluation of Immune Cell Infiltration and Immunotherapy in Stomach Adenocarcinoma

As a first step, we examined the association between cuproptosis-related lncRNA prognostic signatures and immune cell infiltration. We performed ssGESA analysis on TCGA-STAD data and plotted the results as a heatmap. As a result of a correlation analysis between immune cell populations and related functions, Type_II_IFN_Reponse, Parainflammation, Type_I_IFN_Reponse, APC_co_inhibition, T_cell_co-inhibition, Check-point, T_cell_co-stimulation, Cytolytic_activity, Inflammation-promoting, HLA, APC_co_ stimulation, and CCR were abundantly expressed in the high-risk group (Fig. 9a). This suggests that cuproptosis-related prognostic signatures are associated with immune cell infiltration. To assess the effects of immunotherapy in the high- and low-risk groups, we also used the TIDE score. Figure 9b shows that the TIDE scores were significantly higher in the high-risk group than in the low-risk group, indicating that the high-risk group was more susceptible to immune escape and weaker to immunotherapy.

Tumor Mutational Burden of the Cuproptosis-related lncRNA Prognostic Signature in Stomach Adenocarcinoma

We calculated TMB scores in different risk groups using STAD somatic mutation data from TCGA. Compared to the high-risk group, the low-risk group had a higher mutation burden (Fig. 9c). The mutation rate in the low-risk group was 91.28%, whereas in the high-risk group, it was 88.6%, with the top three mutated genes being TTN, TP53, and MUC16 (Fig. 9f, g). Analyzing the patient mutation data, we found that TMB was higher in the low-risk group than in the high-risk group, suggesting that the low-risk group population is more likely to benefit from immunotherapy. Subsequently, we classified STAD patients into "high-TMB" and "low-TMB" groups according to the median cutoff point and performed survival analysis. As shown in Fig. 9d, patients with "high TMB" also had a greater survival rate. Next, we performed a combined survival analysis based on tumor mutation load and risk score. The results showed that low-TMB patients with low risk had the highest survival probability, while high-TMB patients with low risk had the lowest survival probability (Fig. 9e).

Drug Sensitivity

To evaluate the use of the cuproptosis-related prognostic signature in the treatment of STAD, we explored the relationship between the risk score and drug IC50 in the treatment of STAD. We compared the sensitivity of 138 anticancer drugs in the high- and low-risk groups using the "pRRophetic" package in R. A total of 23 drugs were found to have different sensitivities in the high- and low-risk groups. The IC50 of 16 drugs in the low-risk group was low (Fig. 10), and the positive correlation between risk score and IC50 is shown in more detail in Additional file 1. The IC50 of seven drugs in the high-risk group was lower (Fig. 11), and the negative correlation between risk score and IC50 is described in more detail in Additional file 2. As a result, the low-risk group may be more sensitive to drug therapy, and seven drugs represented by dasatinib may play a role in the future treatment of advanced STAD.

Expression of Cuproptosis-related lncRNA Prognostic Signature in GC

To further investigate the expression of eight cuproptosis-related prognostic markers, we observed the differences in the expression of eight lncRNAs in gastric cancer cell lines (MKN-45, SGC-7901 and HCG-27) and human gastric epithelial cells (GES-1). According to quantitative real-time polymerase chain reaction (qRT‒PCR) analysis, eight cuproptosis-related prognostic markers were significantly differentially expressed in gastric cancer cells and GES-1 cells (P < 0.05) (Fig. 12). In gastric cancer cells, all cuproptosis-related prognostic markers except for LINC01094 were significantly increased. The expression of LINC01094 was decreased in MKN-45 cells compared with normal gastric epithelial cells and increased in SGC-7901 and HCG-27 cells compared with normal gastric epithelial cells. Different species of gastric cancer cells may be responsible for this difference. In conclusion, the expression profiles of the eight cuproptosis-related prognostic markers were generally consistent with previous analyses, suggesting a possible important role in stomach adenocarcinoma.

With the continuous development of endoscopic technology and imaging techniques in recent years, the survival and treatment of patients with early stomach adenocarcinoma have been greatly improved. However, patients with advanced tumor stages and extensive tumor metastasis still face a poor prognosis^21,22. Cuproptosis may be an effective treatment for cancer in the future. Furthermore, some LncRNAs have been found to influence gastric cancer development and treatment^23,24. The interaction between cuproptosis and LncRNAs, especially in gastric adenocarcinoma, remains poorly understood. Therefore, this study focuses on the potential interactions between cuproptosis and LncRNAs and explores potential prognostic markers.

In this study, we first identified 19 cuproptosis-related genes and 430 cuproptosis-related LncRNAs. Then, 18 cuproptosis-related LncRNAs were further screened using univariate Cox regression analysis. From the 18 LncRNAs, 8 LncRNAs were further selected using LASSO regression and Cox regression as cuproptosis-related LncRNA prognostic signatures, and risk models were constructed. Based on the risk model, we explored survival, gene enrichment analysis, immunotherapy and drug sensitivity in different risk groups.

Through LASSO and multivariate Cox regression analysis, we identified 8 signatures associated with gastric adenocarcinoma prognosis, namely AC005050.3, AC016394.2, AL606970.1, BX890604.1, C5orf66, HAGLR, LINC01094, and TDRKH-AS1. Scholars have studied and reported some of these markers in recent years. It is possible to use C5orf66 as a prognostic marker for gastric adenocarcinoma and cervical cancer, since its expression is correlated with overall survival ^25,26. Cheng et al. ²⁷ identified 13 LncRNA, including C5orf66, as prognostic markers for predicting disease-free survival in patients with gastric cancer. A study by Hu et al. ¹⁰found that blocking HAGLR inhibited gastric cancer cell proliferation and sensitized gastric cancer cells to 5-fluorouracil (5-FU). LINC01094 was found to be associated with the necroptosis pathway in gastric cancer and could be used as a prognostic marker for necroptosis ²⁸. Jiao et al. ²⁹ found that TDRKH-AS1 was upregulated in most colorectal cancer patients and that TDRKH-AS1 targeted the β-catenin in the Wnt signaling pathway to exert its carcinogenic activity. Despite this, few studies have been conducted on AC005050.3, AC016394.2, AL606970.1, and BX890604.1, so further studies on their mechanisms of action are needed in the future.

Using prognostic signatures associated with cuproptosis, we developed a risk model and evaluated its clinical utility. Based on Kaplan-Meier analysis, low-risk patients had a longer OS than high-risk patients. The risk score was found to be an independent prognostic factor in both univariate and multivariate Cox analyses. The ROC curves suggest that the model has good accuracy for predicting 1-, 3-, and 5-year survival. According to the C-index, the model has a higher predictive accuracy than other clinical prediction models. Further differences between high risk and low risk groups can be seen in the PCA plots and PFS charts below. We constructed a hybrid nomogram by combining risk score, gender, age, tumor TNM stage, grade and stage to assess patient survival at 1, 3 and 5 years. In addition, we performed GO and KEGG analyses in order to further explore the potential biological mechanisms of the eight cuproptosis-related LncRNA prognostic signatures.We explored the key signaling pathways of eight cuproptosis-related LncRNA prognostic signatures.Based on GO and KEGG analyses, signatures are differentially expressed primarily in vascular smooth muscle contraction, focal adhesion, dilated cardiomyopathy, and hypertrophic cardiomyopathy.

We scored TIDE scores, and found that high-risk groups had higher TIDE scores than low-risk groups. When we compared the tumor mutation burden between high-risk and low-risk groups, we found that the low-risk group had a higher mutation burden. In light of the above, it is clear that low-risk patients will benefit more from immunotherapy. Finally, we performed drug sensitivity analysis and found that dasatinib was more sensitive in the high-risk group. Dasatinib is currently more studied in gastrointestinal mesenchymal tumors and less reported in gastric adenocarcinoma³⁰, which may provide a novel therapeutic strategy for the treatment of gastric adenocarcinoma in the future.

Finally, we analyzed the screened eight cuproptosis-related prognostic markers in an in vitro assay. According to our analysis, most cuproptosis-related prognostic markers are significantly elevated in gastric cancer, which is consistent with our results. The expression levels of LINC01094 differed in different gastric cancer cell lines, which may be due to intercellular differences. Thus, further in vivo and in vitro studies are needed to validate LINC01094's efficacy in stomach adenocarcinoma.

Cuproptosis is a new form of apoptosis and some scholars believe it may play a role in future cancer therapy. Numerous articles have also discussed the role of LncRNA in gastric adenocarcinoma. However, the link between cuproptosis and LncRNA currently remains to be further explored. In conclusion, this study provides new insights into the prognosis and treatment of stomach adenocarcinoma from the perspective of cuproptosis. Eight cuproptosis-related LncRNA prognostic signatures were identified, which could provide a reference for future clinical treatment. However, we have only validated them in TCGA, and a large number of preclinical studies are still needed for further validation in the future. There is also a need for further in vivo and in vitro validation.

In conclusion, we identified eight risk models for cuproptosis-related LncRNAs that can accurately predict the prognosis of STAD patients. In this way, we provide a new basis for individualized immunotherapy and treatment in the future for patients with STAD. These 8 cuproptosis-related LncRNAs may become potential targets for the future treatment of STAD.

Acknowledgements

This work was supported by the Xuzhou Science and Technology Program (KC21185). Yaodong Xiao and Yiheng Shi were responsible for collecting and analyzing data related to stomach adenocarcinoma and were the main contributors to this manuscript. Zhang Cheng is responsible for the in vitro experiments of this manuscript. Bowen Li was responsible for revising the grammar of this manuscript. All authors read and approved the final manuscript.

Ethical Approval

Not applicable.

Competing interests

Not applicable.

Authors' contributions

Yaodong Xiao and Yiheng Shi were responsible for collecting and analyzing data related to stomach adenocarcinoma and were the main contributors to this manuscript. Zhang Cheng is responsible for the in vitro experiments of this manuscript. Bowen Li was responsible for revising the grammar of this manuscript. All authors read and approved the final manuscript.

Funding

This work was supported by the Xuzhou Science and Technology Program (KC21185).

data availability

The datasets generated and/or analysed during the current study are available in the The Cancer Genome Atlas (TCGA) repository, (https://portal.gdc.cancer.gov/, accessed 03 May 2022.).

Competing interests

The authors declare that there is no conflict of interest regarding the publication of this paper.

Marin, J. J. G. et al., Molecular Bases of Mechanisms Accounting for Drug Resistance in Gastric Adenocarcinoma. CANCERS 12 2116 (2020).
Sexton, R. E., Al, H. M., Diab, M. & Azmi, A. S., Gastric cancer: a comprehensive review of current and future treatment strategies. CANCER METAST REV 39 1179 (2020).
Tsvetkov, P. et al., Copper induces cell death by targeting lipoylated TCA cycle proteins. Science (American Association for the Advancement of Science) 375 1254 (2022).
Oliveri, V., Selective Targeting of Cancer Cells by Copper Ionophores: An Overview. FRONT MOL BIOSCI 9 841814 (2022).
Tsvetkov, P. et al., Copper induces cell death by targeting lipoylated TCA cycle proteins. Science (American Association for the Advancement of Science) 375 1254 (2022).
Bian, Z., Fan, R. & Xie, L., A Novel Cuproptosis-Related Prognostic Gene Signature and Validation of Differential Expression in Clear Cell Renal Cell Carcinoma. GENES-BASEL 13 851 (2022).
Han, J., Hu, Y., Liu, S., Jiang, J. & Wang, H., A Newly Established Cuproptosis-Associated Long Non-Coding RNA Signature for Predicting Prognosis and Indicating Immune Microenvironment Features in Soft Tissue Sarcoma. J ONCOL 2022 1 (2022).
Yang, L. et al., Cuproptosis-Related lncRNAs are Biomarkers of Prognosis and Immune Microenvironment in Head and Neck Squamous Cell Carcinoma. FRONT GENET 13 (2022).
McFadden, E. J. & Hargrove, A. E., Biochemical Methods To Investigate lncRNA and the Influence of lncRNA:Protein Complexes on Chromatin. BIOCHEMISTRY-US 55 1615 (2016).
Hu, J., Huang, L., Ding, Q., Lv, J. & Chen, Z., Long noncoding RNA HAGLR sponges miR‐338‐3p to promote 5‐Fu resistance in gastric cancer through targeting the LDHA‐glycolysis pathway. CELL BIOL INT 46 173 (2022).
Wang, J. et al., Role of long non-coding RNA H19 in therapy resistance of digestive system cancers. MOL MED 27 (2021).
Kim, S., Lim, K., Yang, S. & Joo, J., Long non-coding RNAs in brain tumors: roles and potential as therapeutic targets. J HEMATOL ONCOL 14 (2021).
Wang, M. et al., Long noncoding RNA LINC00336 inhibits ferroptosis in lung cancer by functioning as a competing endogenous RNA. Cell Death & Differentiation 26 2329 (2019).
Xiao, S., Liu, X., Yuan, L. & Wang, F., A Ferroptosis-Related lncRNAs Signature Predicts Prognosis and Therapeutic Response of Gastric Cancer. FRONT CELL DEV BIOL 9 (2021).
Hutter, C. & Zenklusen, J. C., The Cancer Genome Atlas: Creating Lasting Value beyond Its Data. CELL 173 283 (2018).
Friedman, J., Hastie, T. & Tibshirani, R., Regularization Paths for Generalized Linear Models via Coordinate Descent. J STAT SOFTW 33 1 (2010).
Fu, J. et al., Large-scale public data reuse to model immunotherapy response and resistance. GENOME MED 12 (2020).
Subramanian, A., Kuehn, H., Gould, J., Tamayo, P. & Mesirov, J. P., GSEA-P: a desktop application for Gene Set Enrichment Analysis. BIOINFORMATICS 23 3251 (2007).
Jiang, P. et al., Signatures of T cell dysfunction and exclusion predict cancer immunotherapy response. NAT MED 24 1550 (2018).
Keenan, T. E., Burke, K. P. & Van Allen, E. M., Genomic correlates of response to immune checkpoint blockade. NAT MED 25 389 (2019).
Joshi, S. S. & Badgwell, B. D., Current treatment and recent progress in gastric cancer. CA-CANCER J CLIN 71 264 (2021).
Seeneevassen, L. et al., Gastric Cancer: Advances in Carcinogenesis Research and New Therapeutic Strategies. International Journal of Molecular Sciences 22 3418 (2021).
Zhu, T. et al., Long non-coding RNA ZFAS1 promotes the expression of EPAS1 in gastric cardia adenocarcinoma. J ADV RES 28 7 (2021).
Zhu, T. et al., Long non-coding RNA ZFAS1 promotes the expression of EPAS1 in gastric cardia adenocarcinoma. J ADV RES 28 7 (2021).
Luo, W., Wang, M., Liu, J., Cui, X. & Wang, H., Identification of a six lncRNAs signature as novel diagnostic biomarkers for cervical cancer. J CELL PHYSIOL 235 993 (2020).
Razavi, H. & Katanforosh, A., Identification of novel key regulatory lncRNAs in gastric adenocarcinoma. BMC GENOMICS 23 (2022).
Cheng, C., Wang, Q., Zhu, M., Liu, K. & Zhang, Z., Integrated analysis reveals potential long non-coding RNA biomarkers and their potential biological functions for disease free survival in gastric cancer patients. CANCER CELL INT 19 (2019).
Luo, L. et al., A Necroptosis-Related lncRNA-Based Signature to Predict Prognosis and Probe Molecular Characteristics of Stomach Adenocarcinoma. FRONT GENET 13 (2022).
Jiao, Y. et al., Long Non-coding RNA TDRKH-AS1 Promotes Colorectal Cancer Cell Proliferation and Invasion Through the β-Catenin Activated Wnt Signaling Pathway. FRONT ONCOL 10 (2020).
Zeng, C. et al., Spectrum of activity of dasatinib against mutant KIT kinases associated with drug-sensitive and drug-resistant gastrointestinal stromal tumors. GASTRIC CANCER 23 837 (2020).

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

A Newly Established Cuproptosis-Related Long Noncoding RNA Signature for Predicting Prognosis and the Immune Microenvironment in Stomach Adenocarcinoma

Status:

Version 1

Abstract

Figures

Introduction

Materials and Methods

Data Collection

Identification of Cuproptosis-Related LncRNA

Construction of the Cuproptosis-Related LncRNA prediction model

Validation of Risk Models

Establishing and Evaluating a Nomogram

Assessing the Relationship Between the Risk score and Clinicopathological Characteristics

Pathway Enrichment Analysis and Gene set Enrichment Analysis

Estimation of Intratumoral Immune Cell Infiltration and Immunotherapy

Drug Sensitivity Assessment

Calculation of Tumor Mutation Burden Scores

Cell culture and qRT-PCR

Statistical Analysis

RESULTS

Identification of Cuproptosis-related LncRNA in STAD Patients

Cuproptosis-related LncRNA Prognostic Signatures are Independent Prognostic Factors in STAD Patients

Relationship Between Prognostic Signature and Clinicopathological Features in Stomach Adenocarcinoma

Functional Enrichment Analysis

Evaluation of Immune Cell Infiltration and Immunotherapy in Stomach Adenocarcinoma

Tumor Mutational Burden of the Cuproptosis-related lncRNA Prognostic Signature in Stomach Adenocarcinoma

Drug Sensitivity

Expression of Cuproptosis-related lncRNA Prognostic Signature in GC

DISCUSSION

Conclusions

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1