Identification and Validation of Cuproptosis-Related Genes in Prognosis and immune microenvironment in colon cancer patients

doi:10.21203/rs.3.rs-3331236/v1

Download PDF

Research Article

Identification and Validation of Cuproptosis-Related Genes in Prognosis and immune microenvironment in colon cancer patients

https://doi.org/10.21203/rs.3.rs-3331236/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Recently, the copper-induced regulated cell death form that depends on mitochondrial respiration was identified as cuproptosis. However, the potential roles of cuproptosis-related genes(CRGs) with tumor prognosis and tumor microenvironment(TME) in colon cancer(CC) patients remain unclear. From four independent databases, we characterized the set of CRGs in 1124 CC samples from the fields of genetic and transcriptional, then evaluated their expression patterns. We identified two CRGclusters, and found distinguishing clinicopathological features, prognosis, and TME cell infiltrating characteristics correlated with CRGs expression. Moreover, a predictable risk score for overall-survival (OS) was established and its predictive capability in CC patients was validated. Subsequently, the nomogram was constructed to improve the risk score's clinical features applicability. The high-risk level, characterized by an increased proportion of mutation burden, microsatellite instability-high (MSI-H), and immune cell infiltration, indicates the potential for immunotherapy response. Our analysis of CRGs in CC demonstrated their potential roles in the tumor-immune-stromal microenvironment, clinicopathological features, and prognosis. These findings may improve our understanding of CRGs in CC and pave a new path for the assessment of prognosis and the development of more effective immunotherapy strategies.

cuproptosis

colon cancer

tumor microenvironment

microsatellite instability

overall-survival

Recently a new form of programmed cell death, the copper induced cell death, was defined as “cuproptosis1”. It is characterized by aggregation of these copper-bound, lipoylated mitochondrial proteins which are the components of the tricarboxylic acid (TCA) cycle, and subsequent Fe-S cluster protein loss2. Researches showed that cuproptosis mechanism differs from other forms of regulated cell death(RCD), whereas the morphology of copper death has not as characteristically described as ferroptosis and pyroptosis1. However, plenty of evidence have demonstrated that copper induced cell death could be regulated by a range of genes1-4, defined as cuproptosis-related genes(CRGs).

Cuproptosis can be detected in many kinds of solid tumors, such as non-small cell lung cancer, breast cancer, and ovarian cancer1. The cell validity after the treatment of elesclomol was depending on the expression level of FDX1(an essential factor of lipoylation)5. Furthermore, overexpressing SLC31A1(a copper importer) showed a higher cellular sensitivity to copper ions at physiological concentrations in human embryonic kidney 293T and ABC1 cells5. Whether these cuproptosis-related genes have the potential to modulate cuproptosis levels in colon cancer(CC) cells and whether there is any crosstalk between cuproptosis and the tumor microenvironment(TME) needs to be further detected. The TME includes more than tumor cells, fibroblasts, endothelial cells, immune and inflammatory cells, extracellular matrix elements, and diffusible cytokines and chemokines secreted from cancer and stromal cells together constitute the TME6. The extensive and complex crosstalk between tumor cells and nonmalignant cells and stroma that could affect cancer development, progression and adjuvant therapy resistance7, 8. Therefore, TME cell infiltration mediated by multiple cupropotosis-related genes needs to be comprehensively discussed, which may provide important clues to gain insight into the underlying mechanisms of CC tumorigenesis, development and predict prognosis and adjuvant therapy response.

Data sources

Figure 1 shows the flow diagram of this study. Using The Cancer Genome Atlas (TCGA) (https://portal.gdc.cancer.gov/) and the Gene Expression Omnibus (GEO) (https://www.ncbi.nlm.nih.gov/geo/) databases, TCGA-COAD cohorts and GEO CC cohorts (GSE395829, GSE1753610, and GSE3883211) and were obtained for subsequent analyses. The fragments per kilobase million(FPKM) of gene expression matrix along with the clinicopathological features and survival situation of CC patients were downloaded. Afterwards, the R software(version 4.1.3) was used to perform data preprocessing with associated packages. Statistical significance was set at p < 0.05.

Mutation Landscape, Gene Ontology, and Kyoto Encyclopedia of Genes and Genomes Analysis

After obtaining the single nucleotide variation (SNV) and copy number variation (CNV) data of CC from the TCGA database, we then explore their mutation landscape using R with the “maftools” package. In Gene Ontology (GO) analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways analysis, the “ggplot2” package was utilized to perform this analysis.

Construction of the CRGcluster, genetype, risk score and nomogram scoring

system

This research assessed the expression patterns of cupropotosis-related genes and obtained a comprehensive view of the landscape of intratumoral immune by two versatile computational methods, CIBERSORT and ESTIMATE12. First of all, 1097 patients with sufficient clinical information were divided into two discriminate CRGclusters based on CRGs expression patterns, utilizing “ConsensusClusterPlus” package. Afterwards, patients were then classified into 2 gene subtypes according to cuproptosis-related differentially expressed genes (DEGs). Moreover, the differential OS outcomes among 2 distinct gene subtypes were evaluated and presented by Kaplan–Meier curves, relying on the “survival” and “survminer” packages. Consequently, the CC patients were 1:1 randomly categorized into training (n = 549) and testing (n = 548) sets, then the former was used to establish the cuproptosis-related prognostic risk score. Cox analysis to establish a prognostic risk score in the training set. The risk score was calculated as below:

Risk score = Σ(Expi * coefi)

Based on cuproptosis-related prognostic genes, the Lasso Cox regression “glmnet” package13 was utilized to select an optimal model based on cross-validation. Furthermore, we take the clinical features, including age, sex, and TNM stage, into consideration, established a nomogram system, aimed to predict overall-survival (OS). In the nomogram system, each variable is listed separately, with a corresponding number of points assigned to a given magnitude of the variable14. Then, the cumulative point score for all the variables is matched to a scale of 1-, 3-, and 5-year survival outcomes. Calibration plots15 were performed to compare the predictive value of the nomogram, between the predicted 1-, 3-, and 5-year survival conclusion and the virtually follow-up outcomes. This predictive nomogram system, following tested by the test group, was testified to be capable of predicting patient outcomes.

Stratification analyses of the risk levels

For further understood clinical significance and the therapy effect predicting capability of the risk levels, the immune status, microsatellite instability(MSI), and drug sensitivity differences between 2 risk levels were evaluated and revealed statistical significance. To assess the proportions of tumor infiltrating immune cells(TIICs) in the TME, CIBERSORT was utilized to quantify the abundance of 23 infiltrating immune cells in heterogeneous samples in the 2 risk levels, also the correlations between the 23 TIICs and 4 genes involved in the risk score model. We then evaluated the difference in the expression patterns of immune checkpoints between 2 risk scores, results shown by a boxplot. Subsequently, for testifying the therapeutic effects differences between the 2 risk levels, the semi-inhibitory concentration (IC50) values of adjuvant therapy drugs was calculated by the “pRRophetic” package16. Differential susceptibility of drugs associated with cuproptosis and colon cancer treatment were presented in boxplots.

CRGs Genetic and transcriptional alterations in CC

We created a catalog of 19 CRGs including NFE2L2, NLRP3, ATP7B, ATP7A, SLC31A1, FDX1, LIAS, LIPT1, LIPT2, DLD, DLAT, PDHA1, PDHB, MTF1, GLS, CDKN2A, DBT, GCSH and DLST1, 17, 18. Somatic mutations of CRGs were explored in 447 colon cancer samples from TCGA-COAD. 83 samples exhibited mutations of CRGs, with a mutation frequency of 18.57%(Fig. 2a). Above all, the NLRP3 showed the highest mutation frequencies followed by ATP7A. In contrast, the essential factor for cuproptosis, FDX1 and copper-importer SLC31A1 exhibited no mutations. Then, the number variations(CNVs) of CRGs was investigated, as is shown in Fig. 2b, CNV alterations of CRGs in CC were common. The copper exporter ATP7B whereas MTF1 and NLRP3 were focused on the amplification in copy number, while others had more deletion frequency. The CNV alteration mutations on chromosomes are displayed in Fig. 2c. Consequently, most of the CRGs were identified to have expression variation between CC samples and adjacent normal tissue samples(Fig. 2d). Among those differentially expressed CRGs, 7 genes (ATP7B, LIPT1, LIPT2, PDHA1, GLS, CDKN2A and GCSH) were upregulated, while 6 genes (NLRP3, FDX1, DLD, MTF1, DBT and DLST) were downregulated in CC. Of note, CRGs are previously published for their collaboration in copper induced cell death1. Hence, the correlation between these 19 CRGs in CC were evaluated(Table S1, Fig. 2e). Many of CRGs showed apparently correlated with each other. These relationships prompted that the cooperation between the CRGs may promote the progress of copper-induced cell death. As is shown in the scatter plot of Fig. 2e, GLS and NFE2L2 performed the highest correlation. We also analyzed the relationship between independent CRGs expression status and OS of patients, 15/19 CRGs are demonstrated to correlate with prognosis(Fig. S1). Figure 2f showed that CRGs and immune cells are broadly related. Among them, neutrophils and NLRP3 presented the most prominent positive correlation.

Classification of coproptosis clusters in CC

For further analyse the expression patterns of the prognostic value of CRGs, 1097 patients from TCGA-COAD and GEO(GSE39582, GSE17536, and GSE38832) databases with sufficient clinical information were involved in subsequent studies(Table S2). The full-scale of CRGs interactions, associations, and prognostic value in CC patients were demonstrated in a network map(Fig. 3a). Then, we performed a multivariate Cox regression analysis on CRGs, 3 of which (DLD, PDHA1 and PDHB) were identified as independent predictive factors (Table 1). According to the expression patterns of CRGs, patients with CC were categorized by the “ConsensusClusterPlus” package. Our findings showed that k = 2 appeared to be the selectable result for sorting the entire cohort into CRGcluster A (n = 530) and B (n = 594) (Fig. 3b, S2). Using PCA analysis showing the result of significant differences between the two clusters of CRGs transcription level (Fig. 3c). Survival curves revealed a longer OS time in patients in CRGcluster A than that in patients in CRGcluster B (Fig. 3d). In Fig. 3e, the heatmap showing the clinicopathological features differences and CRGs expression variation in CC patients in different CRGclusters was performed, which revealed that cluster B was related to higher lymph nodes positive rate (p < 0.05) compared to those in cluster A(Table S8).

Table 1

Multivariate Cox regression analysis of 3 CRGs associated with CC patients OS.
Gene Symbol		HR	95.0% CI	pvalue
DLD		0.736197	0.591526–0.916251	0.006079
PDHA1	0.792511		0.644698–0.974214	0.027244
PDHB	0.720232		0.531829–0.975377	0.033914

Relationship between CRGs clusters in pathways and immune characteristics in CC

The GSVA result showed that CRGcluster A was significantly enriched in mitochondrial activity and TCA cycle(tricarboxylic acid cycle) metabolism(Fig. 4a-b and Table S3-4). To further comprehend the role of CRGs in the TME of CC, we assessed the correlations between the two CRGclusters and 23 infiltrating immune cells by utilizing the CIBERSORT algorithm(Table S5). The infiltration of most immune cells between the two CRGclusters has shown significant differences(Fig. 4c). However, the CRGcluster B with a poor prognosis has shown a significantly higher proportion of immune cell infiltration. Including some tumor-killing cells such as activated CD8 + T cells and NK cells19. The TME score(stromal score, immune score, and estimate score) of the two CRGcluster was also identified using the “ESTIMATE” package. AS the TME score shows, higher stromal scores or immune scores represented relatively higher contents of stromal cells or immune cells in the TME, while the ESTIMATEScore, indicating the aggregation of stromal or immune scores in the TME, is used to indirectly reflect tumor purity. Similarly, the result in Fig. 4d showed higher TME scores in CRGcluster B patients.

Classification of gene subtypes based on DEGs

In order to explore the possible prognostic predictive potential of each CRGcluster gene expression pattern(Fig. S3), the “limma” package was utilized to identify 136 CRGcluster-related DEGs, followed by the functional enrichment analysis(Fig. 5a-b). We observed that those CRGcluster-related genes were significantly enriched in biological processes that were correlated with metabolism and copper response(Fig. 5a-b). At the aim of further validating this prognostic predictive potential, we used a consensus clustering algorithm to divide patients into 2 gene subtypes I and II based on the screened-out 85 genes from CRGcluster-related DEGs which related to OS time by Cox regression(Table S6 and Fig. 5c). The Kaplan-Meier curves showed that patients with gene subtype II had the relatively lower OS (log-rank test, p < 0.001; Fig. 5d). Through the result of CRGs expression(Fig. 5e), we found that many CRGs showed differences in 2 gene subtypes, indicating subtype II has a lower level of cuproptosis. In Fig. 6 we noticed that CRGcluster-related DEGs have complex connections and interactions in protein level.

Construction and validation of the prognostic risk level

Consequently, based on the subtypes of CRGcluster-related DEGs, the risk level was established. Figure 7a demonstrates the distribution of patients in the two CRGclusters, two relative gene subtypes, and two risk levels. Firstly, we randomly classify the patients into training (n = 549) and testing (n = 548) sets at a ratio of 1:1 using the “caret” package. Next, 10 OS-associated genes from 85 CRGcluster-related prognostic DEGs(Table S7) which remained according to the minimum partial likelihood deviance were precisely selected by LASSO regression analysis(Fig. S4). Those 10 selected genes were used to perform multivariate Cox regression analysis. We then selected candidates from 10 OS-associated genes based on the Akaike information criterion (AIC)20 value to finally obtain four genes (CFTR, HOXC6, MSLN and CKMT2), including two high-risk genes(HOXC6 and MSLN) and two low-risk genes(CKMT2 and CFTR). According to the results of the multivariate Cox regression analysis, the risk score was constructed as follows:

Risk score = (− 0.185* expression of CFTR) + (− 0.0807*expression of CKMT2) + (0.09779* expression of HOXC6) +(0.07347* expression of MSLN).

Whether there is a significant difference in risk_score between CRGclusters and gene subtypes was testified and presented by boxplot in Fig. 5b-c. We observed that the risk level of CRGcluster B and gene subtype II is higher than it in subtype I (Fig. 7b-c). Patients with a risk_score lower than the median were categorized into the low-risk level group (n = 275), whereas those whose risk score was greater than the median were placed in the high-risk level group (n = 274). Figure 7d-e showing the distribution of the risk_score revealed that survival times decreased along with the increase in risk scores. Comparing with the patients with high risk scores, the Kaplan–Meier survival curves revealed that patients with low scores had a relatively higher overall survival rate (log-rank test, p < 0.001; Fig. 7f). In addition, the 1-, 3-, and 5-year survival rates of risk_score were represented by AUC values of 0.709, 0.677 and 0.632, respectively (Fig. 7g). Figure 7h showed the 4 model performing genes expression level in 2 risk levels. To validate the stability and reliability of the risk_score, we performed risk_scores into the testing set and aggregate database (TCGA-COAD and GEO, n = 1097). Fig. S5–S6 indicates that the risk_score had an excellent ability to predict the survival of CC patients. Between the high- and low-risk levels, a chi-square test of clinical information found that the high-risk group had a higher proportion of elderly patients and a relatively late stage (Table S9). We further combine the risk levels and clinical information established the nomogram in order to predict OS rates in 1-, 3-, and 5-year in CC patients (Fig. 7i). Moreover, in both the training and testing sets, comparing to an ideal model, the proposed nomogram had a relatively good performance, which was suggested by the subsequent calibration plots showed in Fig. 7j. Figure 7k showed that in the training set, the 1-, 3-, and 5-year AUC values of the nomogram were 0.833, 0.774, and 0.740, indicating that the prediction power of the nomogram is dependable, and it has certain clinical practicability. The subsequent calibration plots and ROC suggested that the proposed nomogram had a similar performance in the training, testing and all sets compared to an ideal model (Fig. S7).

TME and immune checkpoints were assessed between the high- and low-risk levels

CIBERSORT algorithm was utilized to assess the association between risk score and the abundance of immune cells infiltration. As Fig. 8a the scatter diagrams showing the risk score was positively correlated with M1 macrophages, CD8 + T cells, follicular helper T cells, gamma delta T cells, activated NK cells, neutrophils and negatively correlated with resting NK cells, plasma cells, activated CD4 + T cells, resting CD4 + T cells. In addition, the correlation between the 4 genes in the proposed model and the infiltration of immune cells was also analyzed. A significant correlation of immune cells with those 4 genes in the proposed model was revealed, excepting mast cells, B cells, dendritic cells and eosinophils (Fig. 8b). Interestingly as the results we observed in Fig. 3, a low risk score was also closely associated with a relatively low immune score, whereas a high risk score was associated with a high immune score (Fig. 8c). To supplement, we investigated the variations between immune checkpoints gene expression and risk levels. Figure 8d showed that 22 of 32 immune checkpoints were differentially expressed in the two risk levels, however Current targets for immune checkpoint inhibitors21 such as PD-1, PD-L1, and CTLA-4 showed no difference.

Assessment of MSI and TMB in 2 risk levels

The mutated genes ranking in the high- and low-risk levels were basically the same, APC, TP53, TTN, KRAS, PIK3CA, SYNE1, MUC16, FAT4, ZFHX4, and RYR2 comes to the top 10 (Fig. 9a-b). Patients with a high risk score had markedly lower frequencies of APC and TP53 mutations compared to those patients with a low risk score. However, the exact opposite was observed regarding the mutation levels of other genes on the plot. Both of evidence demonstrate that Colon cancer patients with high microsatellite instability (MSI-H) are more sensitive to immunotherapy22, 23. Compared with low-risk level, the high-risk level showed a higher proportion of MSI-H patients, meanwhile MSI-H group showed a higher risk stratification (Fig. 9c-d). We then performed survival analysis in the MSI and MSS groups to evaluate the influence of MSI status on OS in patients with CC. Although there is no statistical difference. However, in the long-term prognosis, those two groups began to show a separation tendency (p = 0.785; Figure S8). Accumulative evidence shows that patients with a high TMB may benefit from immunotherapy due to their higher numbers of neoantigens24. A positive association between the risk score and the TMB was illustrated by Spearman correlation analysis (p ＜ 0.01; Fig. 9e). By analyzing the mutation data from the TCGA-COAD cohort, higher TMB in the high risk score group than that in the low-risk group was demonstrated in Fig. 9f, implying that the high-risk group might benefit better from immunotherapy. Between high- and low-risk levels, we have tried to select adjuvant therapy drugs currently used to treat colon cancer patients or associated with cuproptosis, then evaluate the susceptibilities. Interestingly, we found that the patients in the low risk score group had a higher IC50 value for Elesclomol(Fig. 9g). Other drugs, although showing differences, however, are not currently used to treat colon cancer(Fig. S9).

The transcriptome information and clinical data of patients that have been diagnosed with CC were totally collected from the Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) large-scale databases. Assessing and analyzing this large-scale database allowed us to have an overall view of the genetic landscape, identify new biomarkers, screen therapeutic strategies and the predict prognosis of these cancers. Moreover, we may find further directions for experimental verification through this results25.

Cuproptosis is a newly revealed regulated cell death(RCD) form that is distinct from other RCD forms1, the role of cuproptosis in tumor prognosis prediction remains unknown. Researches have shown that cuproptosis could be regulated by a series of genes24, 26. In our current study on colon cancer, we first evaluated the expression variation of cuproptosis-related genes and found that 7 out of 19 CRGs were upregulated, while 6 out of 19 genes were downregulated as compared to normal tissues. In fact, the highly expressed regulatory genes in the tumor tissues mostly acted as negative regulators. While those downregulated genes are mostly positively cuproptosis related genes. That is, the tumor group maybe had lower levels of copper-induced cell death than the normal group. The CC patients were then classified into two CRGclusters according to the expression pattern of CRGs. By analyzing that, we found that CC patients expressing higher positively regulated CRGs accompany by longer survival periods than those with lower positively regulated CRGs expression. Therefore, we speculated that cuproptosis-related gene signatures are potential indicators for the prognosis of colon cancer. Around this CRGcluster A and B, we further compared the pathway activity and immune infiltration of them. The GSVA of GO and KEGG database showed that CRGcluster A was significantly enriched in mitochondrial activity and TCA cycle(tricarboxylic acid cycle) metabolism, which may also indicate that CRGcluster A has a relatively higher level of cuproptosis. However, interestingly the immune infiltration abundance comparison showed that CRGcluster with higher immune cell infiltration and immune scores accompanied by unfavorable survival outcomes, which is contrary to previous studies27, 28. Some hypotheses were discussed below.

We identified 2 gene subtypes depending on the DEGs related to CRGclusters. Our findings showed significant differences in survival curves, clinicopathological features and CRGs expression in 2 gene subtypes, indicating that CRGs might serve as a predictor for evaluating the clinical outcome and an auxiliary forecaster for the immunotherapy response of CC patients. Therefore, an effective and reliable prognostic risk scoring system was then constructed and its predictive ability was testified. Finally, we established a quantitative nomogram by integrating the risk score and clinical features, which diametrically improved the feasibility and practicality of the risk score. Prognostic models are suitable for predicting the stratification of colon cancer patients, will help to better understand the molecular mechanism of CC and provide new hints for targeted therapy. To clarify the specific role of CRGs risk stratification in the above goals, analyses about clinicopathological characteristics, prognosis, mutation, TMB, immune checkpoints, MSI, TMB, and drug susceptibility between patients with low- and high-risk scores were performed.

Generally, the prognosis of CC is better in patients with higher levels of tumor-infiltrating lymphocytes, tumor neoantigens, and checkpoints29, 30. However, our findings, whether it is in CRGclusters or risk levels, the group with more immune infiltration and ESTIMATE score showed poor OS rate. CD8 + tumor-infiltrating lymphocytes were found to be related to chemosensitive response and better prognosis and negative PD-L1 expression31. Previous studies have reported the association between PD-L1 and cytotoxic CD8 + TILs for cancer patients32. Our result in 2 risk levels showed no significant differences between PD-1 and PD-L1. While activated CD8 + T cells have no significant correlation with the risk score. Whereas the gamma delta T cells, T. helper cells and T cell follicular helper are positively correlated with risk scores. Gamma delta (γδ) T cells can effectively recognize and kill colorectal cancer (CRC) cells, thereby suppressing tumor progression via multiple mechanisms. They also have abilities to exert a protumor effect via secreting interleukin-17 (IL-17)33. γδ T cells have been selected as potential immunocytes for antitumor treatment because of their significant cytotoxic activity34. Also, in colorectal cancer patients, the higher expression level of follicular helper T cell (Tfh)-associated genes the better prognosis was demonstrated by Kuske and coworkers35. In patients with advanced stages of CRC, the capability of stimulating CD8 + Tcells and producing IL-21 of Tfh cells was reduced greatly. In addition, the capacity to express IL-21 of Tfh cells could be further suppressed by the treatment with autologous PD-L1-expressing tumor cells36. Moreover, gut microbiota may also have the ability to enhance Tfh cells anti-tumor immunity in colorectal cancer37. Although there are ample number of evidence showed that some CRGs may positively regulate immunity activities in many solid tumors and indicate a better prognosis38-40. Regrettably, those are mainly theoretical speculations that urgently needed sufficient in vitro and in vivo experiments to testify the real phenomenon and mechanism in organisms. However, our results in the subtypes of cuproptosis and risk levels higher immune cell infiltration rate didn’t show a higher OS rate. This may indicate that CRGs may regulate immune cell activity through other interacting pathways. Recently, research on lung cancer demonstrated that the impairment of mitochondrial metabolism induced by intermittent hypoxia increases the ROS-dependent stabilization of HIF-1α/LAP complex in NSCLC, producing chemo-immuno-resistance41. It further explained that intermittent hypoxia impaired the electron transport chain and reduced O2 consumption, increasing mitochondrial ROS that favor the stabilization of C/EBP-β mRNA mediated by HIF-1α41. Cuproptosis, as a mitochondrial respiration-dependent regulatory cell death, may be related to similar mechanisms. More interesting, Colorectal cancer differs from other tumors' gut microbiota microenvironment along with CRGs may be capable to regulate immune cell activity42. In the naturally occurring environment of colon cancer, gut bacteria may have a more pronounced effect on immune cell infiltration than colon cancer cells. Our team is further investigating this area.

Unfortunately, this study had several limitations that need further improvement. Primarily, all analyses were based merely on public databases, and all samples in our study were enrolled retrospectively. Therefore, an inevitable selection bias may have influenced the results. Furthermore, though the prognostic model we built has been tested by our data set, the public database content is not enough to support us to complete the high-quality independent external data set test. Accordingly, large-scale prospective studies along with in vitro and in vivo tests are still needed for further verification.

Our prediction model based on cuproptosis classification has acceptable practicability and predictive performance. In our research, we also revealed a complex regulatory correlation between CRGs and TME, immune status, TMB and therapeutic sensitivity. Our findings showed the prominence of the irreplaceable clinical significance of CRGs, also providing new ideas for assisting personalized therapy strategies for colon cancer patients.

Data availability statement

The datasets analyzed in this study can be found in the online repositories. TCGA-COAD(http://www.cancer.gov/tcga) and GEO (https://www.ncbi.nlm.nih.gov/geo/. GSE39582/GSE17536/GSE38832).

Author contributions

J.Z. and W.M. contributed equally to the work. J.Z. and W.M. designed the study; J.Z. and W.M. analyzed the data and drafted initial manuscript; J.Z. acquired the data and searched the literatures; Y.Z. reviewed and edited the manuscript. All authors read and approved the final manuscript

Competing interests

The authors declare no competing interests.

TSVETKOV P, COY S (2022) PETROVA B, et al. Copper induces cell death by targeting lipoylated TCA cycle proteins [J], vol 375. Science (New York, NY), pp 1254–1261. 6586
LI SR, BU L L CAIL (2022) Cuproptosis: lipoylated TCA cycle proteins-mediated novel cell death pathway [J]. Signal Transduct Target therapy 7(1):158
LI J, CHEN S, LIAO Y et al (2022) Arecoline Is Associated With Inhibition of Cuproptosis and Proliferation of Cancer-Associated Fibroblasts in Oral Squamous Cell Carcinoma: A Potential Mechanism for Tumor Metastasis [J]. Frontiers in oncology, 12(925743
XU S, LIU D, CHANG T et al (2022) Cuproptosis-Associated lncRNA Establishes New Prognostic Profile and Predicts Immunotherapy Response in Clear Cell Renal Cell Carcinoma [J]. Frontiers in genetics, 13(938259
WANG Y, ZHANG L (2022) ZHOU F. Cuproptosis: a new form of programmed cell death [J]. Cellular & molecular immunology
BRAUMüLLER H, MAUERER B, BERLIN C et al (2022) Senescent Tumor Cells in the Peritoneal Carcinomatosis Drive Immunosenescence in the Tumor Microenvironment [J]. Frontiers in immunology, 13(908449
LEE SS, CHEAH Y K (2019). The Interplay between MicroRNAs and Cellular Components of Tumour Microenvironment (TME) on Non-Small-Cell Lung Cancer (NSCLC) Progression [J]. J Immunol Res, 2019(3046379.
JANSSEN J B E, MEDEMA J P, GOOTJES E C et al (2022) Mutant RAS and the tumor microenvironment as dual therapeutic targets for advanced colorectal cancer [J]. Cancer treatment reviews, 109(102433
MARISA L, DE REYNIèS A, DUVAL A et al (2013) Gene expression classification of colon cancer into molecular subtypes: characterization, validation, and prognostic value [J]. PLoS Med 10(5):e1001453
CHEN MS, LO Y H, CHEN X et al (2019) Growth Factor-Independent 1 Is a Tumor Suppressor Gene in Colorectal Cancer [J]. Mol cancer research: MCR 17(3):697–708
TRIPATHI M K, DEANE N G, ZHU J et al (2014) Nuclear factor of activated T-cell activity is associated with metastatic capacity in colon cancer [J]. Cancer Res 74(23):6947–6957
CHEN B, KHODADOUST M S, LIU C L et al (2018) Profiling Tumor Infiltrating Immune Cells with CIBERSORT [J]. Methods in molecular biology. Clifton, NJ). 1711(243 – 59.
TANG Z, SHEN Y, ZHANG X et al (2017) The spike-and-slab lasso Cox model for survival prediction and associated genes detection [J]. Bioinf (Oxford England) 33(18):2799–2807
BALACHANDRAN V P GONENM (2015) Nomograms in oncology: more than meets the eye [J]. Lancet Oncol 16(4):e173–e180
AUSTIN PC, HARRELL F E JR (2020) VAN KLAVEREN D. Graphical calibration curves and the integrated calibration index (ICI) for survival models [J]. Stat Med 39(21):2714–2742
CHEN Q, TANG P, HUANG H et al (2022) Establishment of a circular RNA regulatory stemness-related gene pair signature for predicting prognosis and therapeutic response in colorectal cancer [J]. Frontiers in immunology, 13(934124
ZHANG X, WANG S, TANG K et al (2022) Cu(2+) Embedded Three-Dimensional Covalent Organic Framework for Multiple ROS-Based Cancer Immunotherapy [J], vol 14. ACS applied materials & interfaces, pp 30618–30625. 27
DING F, LI F, TANG D et al (2022) Restoration of the Immunogenicity of Tumor Cells for Enhanced Cancer Therapy via Nanoparticle-Mediated Copper Chaperone Inhibition [J]. 61(31):e202203546 Angewandte Chemie (International ed in English)
YANG C, LIU Y, HU Y et al (2022) Myc inhibition tips the immune balance to promote antitumor immunity [J]. Cellular & molecular immunology
ANYASODOR A E, NWOSE E U, BWITITI P T et al (2022) Cost-effectiveness of community diabetes screening: Application of Akaike information criterion in rural communities of Nigeria [J]. Frontiers in public health, 10(932631
GHISONI E, WICKY A, BOUCHAAB H et al (1990) Late-onset and long-lasting immune-related adverse events from immune checkpoint-inhibitors: An overlooked aspect in immunotherapy [J]. European journal of cancer (Oxford, England: 2021, 149(153 – 64
GANESH K, STADLER Z K, CERCEK A et al (2019) Immunotherapy in colorectal cancer: rationale, challenges and potential [J]. Nat reviews Gastroenterol Hepatol 16(6):361–375
ZHANG X, YANG R, WU T et al (2022) Efficacy and Safety of Neoadjuvant Monoimmunotherapy With PD-1 Inhibitor for dMMR/MSI⁃H Locally Advanced Colorectal Cancer: A Single-Center Real-World Study [J]. Frontiers in immunology, 13(913483
FENG L, LIU Z, LI C et al (2022) Development and validation of a radiopathomics model to predict pathological complete response to neoadjuvant chemoradiotherapy in locally advanced rectal cancer: a multicentre observational study [J]. Lancet Digit health 4(1):e8–e17
LV H, LIU X et al (2022) ZENG X,. Comprehensive Analysis of Cuproptosis-Related Genes in Immune Infiltration and Prognosis in Melanoma [J]. Frontiers in pharmacology, 13(930041
DUAN W J, HE RR (2022) Cuproptosis: copper-induced regulated cell death [J]. Sci China Life Sci 65(8):1680–1682
YUE Y, ZHANG Q, SUN Z (2021) CX3CR1 Acts as a Protective Biomarker in the Tumor Microenvironment of Colorectal Cancer [J]. Frontiers in immunology, 12(758040
GU W X, CHEN Y (2022) WANG W. Immune Infiltrates of m5C RNA Methylation-Related LncRNAs in Uterine Corpus Endometrial Carcinoma [J]. J Oncol, 2022(1531474.
RAMLEE S, HULSE D, BERNATOWICZ K et al (2022) Radiomic Signatures Associated with CD8(+) Tumour-Infiltrating Lymphocytes: A Systematic Review and Quality Assessment Study [J]. Cancers, 14(15)
KUMAR N, PAPILLON-CAVANAGH S, TANG H et al (2022) A multi-omic single cell sequencing approach to develop a CD8 T cell specific gene signature for anti-PD1 response in solid tumors [J]. International journal of cancer
CHEN TW, HUNG W Z, CHIANG S F et al (2022) Dual inhibition of TGFβ signaling and CSF1/CSF1R reprograms tumor-infiltrating macrophages and improves response to chemotherapy via suppressing PD-L1 [J]. Cancer letters, 543(215795
ZHANG P, MA Y, LV C et al (2016) Upregulation of programmed cell death ligand 1 promotes resistance response in non-small-cell lung cancer patients treated with neo-adjuvant chemotherapy [J]. Cancer Sci 107(11):1563–1571
MA R, YUAN D, GUO Y et al (2020) Immune Effects of γδ T Cells in Colorectal Cancer: A Review [J]. Front Immunol 11:1600
POGGI A (2014) ZOCCHI M R. γδ T Lymphocytes as a First Line of Immune Defense: Old and New Ways of Antigen Recognition and Implications for Cancer Immunotherapy [J]. Frontiers in immunology, 5(575
KUSKE M, HAIST M, JUNG T et al (2022) Immunomodulatory Properties of Immune Checkpoint Inhibitors-More than Boosting T-Cell Responses? [J]. Cancers, 14(7)
SHI W, SUN DONGL (2018) Follicular helper T cells promote the effector functions of CD8(+) T cells via the provision of IL-21, which is downregulated due to PD-1/PD-L1-mediated suppression in colorectal cancer [J]. Exp Cell Res 372(1):35–42
OVERACRE-DELGOFFE A E, BUMGARNER H J, CILLO A R et al (2021) Microbiota-specific T follicular helper cells drive tertiary lymphoid structures and anti-tumor immunity against colorectal cancer [J]. Immunity 54(12):2812–24e4
ZHOU Y, SHU Q, FU Z et al (2022) A novel risk model based on cuproptosis-related lncRNAs predicted prognosis and indicated immune microenvironment landscape of patients with cutaneous melanoma [J]. Frontiers in genetics, 13(959456
CHEN Y. Identification and Validation of Cuproptosis-Related Prognostic Signature and Associated Regulatory Axis in Uterine Corpus Endometrial Carcinoma [J]. Frontiers in genetics, (2022) 13(912037
LU Y, PAN Q, GAO W et al (2022) Reversal of cisplatin chemotherapy resistance by glutathione-resistant copper-based nanomedicine via cuproptosis [J]. J Mater Chem B,
SALAROGLIO I C, BELISARIO D C, AKMAN M et al (2022) Mitochondrial ROS drive resistance to chemotherapy and immune-killing in hypoxic non-small cell lung cancer [J]. J experimental Clin cancer research: CR 41(1):243
GUO H, GIBSON S A, TING J P Y (2020) Gut microbiota, NLR proteins, and intestinal homeostasis [J]. J Exp Med, 217(10)

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Identification and Validation of Cuproptosis-Related Genes in Prognosis and immune microenvironment in colon cancer patients

Status:

Version 1

Abstract

Figures

Introduction

Materials and methods

Data sources

Mutation Landscape, Gene Ontology, and Kyoto Encyclopedia of Genes and Genomes Analysis

Construction of the CRGcluster, genetype, risk score and nomogram scoring

system

Stratification analyses of the risk levels

Result

CRGs Genetic and transcriptional alterations in CC

Classification of coproptosis clusters in CC

Relationship between CRGs clusters in pathways and immune characteristics in CC

Classification of gene subtypes based on DEGs

Construction and validation of the prognostic risk level

TME and immune checkpoints were assessed between the high- and low-risk levels

Assessment of MSI and TMB in 2 risk levels

Discussion

Conclusion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1