How iron death, metabolic reprogramming and immune infiltration are 1+1+1 greater than 3

What I share with you today is an article published in Briefings in Bioinformatics (IF: 11.
6215), which is mainly based on the construction of a prognostic classification model of pancreatic ductal adenocarcinoma based on iron death-related lncRNA.
The idea is clear and the process is simple
.

There are more prognostic markers of gene pairs, so change your mind and see how the model of lncRNA pairs is constructed
.

In order to better give back to new and old customers and believers, we will give priority to some new ideas in the future.
Welcome to scan the QR code to fill out the form to add friends and give priority to the tips.
Ferroptosis-related lncRNA pairs to predict the clinical outcome and molecular characteristics of pancreatic ductal adenocarcinoma is based on lncRNA to predict the prognostic outcome and molecular characteristics of pancreatic ductal adenocarcinoma.
Iron death (ferroptosis) is a new type of cell death discovered in recent years, which is characterized by excessive accumulation of lipid peroxides and reactive oxygen species (ROS)
.

Since this process relies on iron, it is called iron death
.

Iron death is not only related to the occurrence and development of many diseases, but also key proteins in related signaling pathways can become targets of drugs
.

Therefore, iron death may be related to the therapeutic effects of chemotherapy and immunotherapy for a variety of cancers including pancreatic ductal adenocarcinoma (PDAC)
.

Many studies have noted that lncRNA can regulate the biological behavior of cancer cells by combining with DNA, RNA and proteins
.

However, few studies have reported the role of lncRNA in the process of iron death and the function of lncRNA related to iron death
.

The main purpose of this study is to identify lncRNAs related to iron death through bioinformatics methods, and construct a prognostic model of PDAC patients based on the size and order relationship between iron death-related lncRNAs, and verify the characteristics of immune infiltration, drug sensitivity and other aspects The classification reliability of the model (Figure 1)
.

Figure 1.
Flow chart 1.
Data (1) PDAC patient gene expression, lncRNA expression, mutation and prognosis data in TCGA
.

(2) Sequencing data from the self-test of Fudan University Shanghai Cancer Center (FUSCC)
.

(3) Sequencing data from pancreatic cancer patients (E-MTAB-6134) in ArrayExpress
.

(4) Iron death related genes (FRGs) from FerrDb (http://
.

2.
Identify iron death-related lncRNAs (FRLs) based on bioinformatics and experimental verification methods.
Based on Pearson correlation analysis to identify the correlation between 265 FRGs and 14 806 lncRNAs in TCGA-PDAC, only those with r> 0.
5 and P <0.
0001 are retained lncRNA, as iron death-related lncRNAs (FRLs)
.

Among 786 FRLs
.

There are 132 FRLs differentially expressed in cancer tissues and normal tissues
.

Some lncRNAs are highly correlated with the expression of the three core iron death-related regulatory factors ACSL4, SLC7A11 and GPX4 (Figure 2A)
.

In the pancreatic cancer samples of the FUSCC cohort or E-MTAB-6134 cohort, researchers can also observe that the expression of SLCO4A1-AS1 is significantly up-regulated in tumor tissues, and is positively correlated with the expression of SLC7A11 (Figure 2B-D)
.

The expression of SLCO4A1-AS1 was detected in six common pancreatic cancer cell lines, and it was found that SLCO4A1-AS had the highest expression in panc-1 cells (Figure 2E)
.

Knock out the expression of SLCO4A1-AS1 in panc-1 cells, and its reduction has no significant effect on the proliferation of cancer cells (Figure 2F-G)
.

Considering that the expression of SLCO4A1-AS1 is positively correlated with the expression of SLC7A11, the researchers further evaluated whether the decreased expression of SLCO4A1-AS1 would affect the sensitivity to iron death
.

The results showed that lncRNA SLCO4A1-AS1 gene knockdown significantly up-regulated the sensitivity of pancreatic cancer cells to erastin- and RSL-3 induced iron death, indicating that lncRNA SLCO4A1-AS1 may be a new molecule that mediates iron death resistance (Figure 2H-I )
.

The decrease of lncRNA SLCO4A1-AS1 significantly increased the concentration of malondialdehyde (MDA) (Figure 2J-K)
.

In addition, when SLCO4A1-AS1 was knocked out, the expression of SLC7A11 was significantly reduced (Figure 2L)
.

Figure 2.
Identifying iron death-related lncRNA based on bioinformatics and experimental verification methods 3.
Constructing an FRL pair model (FRLM) for predicting OS in PDAC patients Based on the size and order relationship of 132 FRL expressions, constructing 0 or 1 composed of iron death-related lncRNA Matrix
.

When lncRNA A>lncRNA B is expressed, it is defined as 1, otherwise it is defined as 0
.

Based on single factor Cox regression to screen lncRNA pairs related to survival time, and based on Lasso regression to calculate the coefficient of each lncRNA pair in FRLM, a prognostic model consisting of 14 pairs of lncRNAs was finally identified
.

According to the FRLM risk score, the ROC curve is drawn to evaluate the accuracy of the patient's OS prediction
.

The optimal threshold is defined as the risk score that can reflect the maximum AUC value (Figure 3A-B)
.

Patients were divided into two groups according to the optimal threshold of FRLM risk (Figure 3C).
More deaths were observed in the high-risk group, suggesting that the increased risk of FRLM in PDAC patients is associated with a poor prognosis (Figure 3D)
.

The Kaplan Meier curve shows that patients with a higher risk of FRLM have a shorter OS (Figure 3E)
.

There was no difference in age, gender, etc.
between FRLM risk groups, but more liver metastases were found in high-risk patients (Figure 3F-G)
.

The predictive ability of traditional clinical parameters is significantly weaker than FRLM risk score (Figure 3H)
.

Multivariate Cox regression also showed that FRLM risk score is an independent factor predicting OS in PDAC patients (Figure 3I)
.

The results in the FUSCC data also show that FRLM can accurately distinguish PDAC patients with different prognosis (Figure 3J)
.

The 1-year and 2-year AUCs of FRLM in FUSCC were 0.
82 and 0.
64, respectively (Figure 3K)
.

Figure 3.
Construction of an FRL pair model (FRLM) for predicting OS in PDAC patients.
4.
Differentially expressed genes reveal that FRLM risk in PDAC is related to metabolic reprogramming.
A total of 726 differentially expressed genes were identified between the FRLM high-risk group and the low-risk group ( Figure 4A)
.

GO enrichment analysis showed that most genes were enriched in metabolic reprogramming processes such as glycolysis and NADH production
.

KEGG analysis also showed that these genes are mainly involved in metabolic pathways, such as glycolysis/gluconeogenesis and n-sugar biosynthesis (Figure 4B)
.

The researchers further tested the metabolic reprogramming process associated with FRLM risk in PDAC samples and identified 12 differentially activated metabolic pathways, of which 5 pathways were up-regulated in the FRLM low-risk group, and 7 pathways were up-regulated in the FRLM high-risk group, such as Pyrimidine metabolism (Figure 4C)
.

The researchers further tested the correlation between the differentially activated metabolic pathways in the FRLM high-risk group, and the results showed that the amino acid metabolism, sugar metabolism and other metabolic pathway activities in this subtype were highly correlated, showing the interaction in the reorganized metabolic network (Figure 4D)
.

Figure 4.
Differentially expressed genes reveal that FRLM risk in PDAC is associated with metabolic reprogramming.
5.
FRLM risk and immune infiltration.
Researchers further studied the differences in molecular characteristics between FRLM high-risk groups and low-risk groups
.

The researchers first compared the FRLM classification with other PDAC subtype classification methods, and the results showed that the percentage of PDAC subtypes was consistent with the FRLM-based clustering results (Figure 5A-C), indicating that FRLM can produce independent classifications
.

The researchers further analyzed the correlation between FRLM risk and immune cell infiltration (Figure 5D) and found that FRLM risk was negatively correlated with stromal and microenvironment scores, indicating that the increased risk of FRLM was associated with the decrease of infiltrating stromal cells in the tumor (Figure 5E- F)
.

FRLM risk is negatively correlated with multiple anti-tumor infiltrating cells such as CD8 + T cells, natural killer (NK) cells and activated dendritic cells (Figure 5D)
.

In addition, the researchers also discovered that CTLA4, a classic immune checkpoint gene, was up-regulated in the FRLM low-risk group (Figure 5G)
.

The number of CD8+ T cell infiltration decreases with the increase of FRLM risk score, suggesting that CD8+ T cells are relatively enriched in FRLM low-risk samples, but the overexpression of inhibitory receptors (such as CTLA4) inhibits the cytotoxic function of CD8+ T cells
.

CTLA4 inhibitors such as ipilimumab and tremelimumab may be effective treatments for these patients
.

In addition, increased bone marrow-derived suppressor cell infiltration was positively correlated with FRLM risk score (Figure 5H)
.

Figure 5.
FRLM risk and immune infiltration 6.
The expression and mutation characteristics of FRLM risk and iron death-related regulatory factors.
In the FRLM high-risk group, a total of 12 anti-iron death regulatory factors were up-regulated, and in the FRLM low-risk group, there were 3 iron The death promoter was up-regulated (Figure 6A)
.

Further based on GSVA enrichment analysis, by integrating 128 FRGs that inhibit iron death, the anti-iron death index (FRI) was calculated, and the samples were divided into high FRI group and low FRI group (Figure 6B)
.

The risk of FRLM was significantly increased in the high FRI group (Figure 6C), indicating that the inherent resistance to iron death in patients with a high FRLM risk score may be one of the reasons for the poor prognosis of PDAC
.

In addition, lower FRI is associated with longer OS
.

Compare the mutations of FRLM high-risk group and low-risk group
.

The results showed that samples with higher FRLM risk tended to have more mutation events, including KRAS mutations (Figure 6D); in the group with lower FRLM risk, the frequency of gene mutations was significantly reduced (Figure 6D)
.

The mutation frequency of the four driving genes of PDAC differed between the two groups
.

Among the samples with lower FRLM risk scores, TP53 rather than KRAS was the most mutated gene, which may help explain the better prognosis of these PDAC patients
.

Figure 6.
Molecular characteristics of FRLM risk 7.
Pancreatic cancer subtypes with different FRLM risk and anticancer drug sensitivity are suitable for different clinical decisions
.

Therefore, the researchers compared the sensitivity of the high-risk group and the low-risk group to 30 common anticancer drugs to determine potential pancreatic cancer treatment models
.

The researchers first predicted the IC50 values ​​of common anticancer drugs in PDAC patients in TCGA data based on the pRRophetic package
.

The results showed that the IC50 of the two anti-tumor drugs approved by the FDA, imatinib and axitinib, was higher in patients with a higher risk of FRLM, indicating that the reduction in FRLM risk was accompanied by the treatment of imatinib and axitinib.
Increased sensitivity (Figure 7)
.

Therefore, low-risk patients may benefit from treatment with the above two drugs
.

Figure 7.
FRLM Risk and Anticancer Drug Sensitivity Today’s content is this, integrating the three hot research directions of iron death, metabolic reprogramming, and immune infiltration to perfectly achieve the ultimate goal of 1+1+1>3
.

However, the editor still found a little problem when reading the article.
First, the proofreading was not serious enough.
In P6, Figure 3K and Figure 3J were wrongly written as Figure 2K and Figure 2J
.

The second is the original text without quoting data E-MTAB-6134
.

Everyone must pay attention to these small details when writing.
Although flaws cannot be covered up, it is definitely better to be perfect! The editor is not easy to write, please contact the good-talking editor butler at 15510012760 (same number on WeChat) for questions related to Shengxin: 18501230653 (same number on WeChat) Welcome to follow Shengxin human transcriptome | methylation | resequencing | single cell | m6A | Multi-omics cytoscape | limma | WGCNA | Water bear worm legend | Linux electrophoresis | PCR | A brief history of sequencing | Karyotype | NIPT | Basic experimental genes | 2019-nCoV | Enrichment analysis | Joint analysis | Microenvironmental plague chase Fierce | Summary of Ideas | Scholars | Research | Withdrawal | PhD Reading | Work