Disulfidptosis and pentose phosphate pathway-associated prognosis signature guides immunotherapy for lung adenocarcinoma patients

doi:10.21203/rs.3.rs-5051024/v1

Download PDF

Research Article

Disulfidptosis and pentose phosphate pathway-associated prognosis signature guides immunotherapy for lung adenocarcinoma patients

https://doi.org/10.21203/rs.3.rs-5051024/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Background

Disulfidptosis is a form of cell death, where generation of nicotinamide adenine dinucleotide phosphate (NADPH) through the pentose phosphate pathway (PPP) play an important role. The discovery of disulfidptosis provides new in-sights into lung adenocarcinoma (LUAD) therapeutics.

Research Design and Methods:

Disulfidptosis regulators (DSRs) was used to identify subgroups. Meanwhile, WGCNA and single-cell analysis were performed to identify genes related to disulfidptosis and PPP (DPRGs). To determine the risk signature, clinical features were analyzed, as well as prognostic pre-dictive ability, tumor immune microenvironment (TIME), immunotherapeutic response and drug sensitivity. Finally, the results were experimentally verified in vitro and vivo.

Results

We identified two DSR and DPRG clusters associated with distinct immune profiles involved in regulating different biological processes. The risk signature was effective in assessing LUAD prognosis in patients. It showed a strong correlation with TIME and could predict the immunotherapy response. After LRRC61 knockdown, the proliferation, migration and anti-apoptotic ability of LUAD cells were significantly reduced. Moreover, the xenograft tumors showed tumour growth was promoted when overexpressing LRRC61.

Conclusion

We analyzed DSRs and DPRGs in LUAD and developed an evaluation system that assesses the risk and guides the clinical application of drugs, including chemotherapeutic and immunotherapeutic agents.

Disulfidptosis

Lung adenocarcinoma

Pentose phosphate pathway

Tumor immune microenvironment

Prognosis

Immunotherapy

Immunotherapy has demonstrated sustained clinical relevance in lung adenocarcinoma (LUAD)¹, and the clinical use of various immune checkpoint inhibitors (ICIs) in advanced LUAD has been approved². Despite reported efficacy, many patients receiving immunotherapy continue to have poor prognoses. Therefore, it is essential to thoroughly investigate the intrinsic regulatory mechanisms of the tumor immune microenvironment (TIME) in patients with LUAD³.

Cell death is a physiological process that maintains biological development and internal environmental homeostasis. Targeting cell death-related pathways to kill cancer cells is a major direction in cancer therapy⁴. Recently, Gan et al. discovered and identified a new type of cell death, disulfidptosis, which offers new possibilities for cancer therapy⁵. Tumor cells typically require a high expression of solute carrier family 7 member 11 (SLC7A11) to uptake more cystine for glutathione synthesis to balance the oxidative effects arising from their highly active metabolism. However, ingested cystine must be reduced to cysteine promptly to eliminate toxicity⁶. This process requires the consumption of large amounts of nicotinamide adenine dinucleotide phosphate (NADPH) generated via the pentose phosphate pathway (PPP). Otherwise, excessive accumulation of cystine may cause tumor cell death. When cells with high SLC7A11 levels lack glucose, disulfide accumulates abnormally, causing disulfide stress, which leads to the contraction of actin filaments, destruction of the cytoskeletal structure, and consequently cell death⁷. SLC7A11 is overexpressed in patients with LUAD and is positively correlated with tumor progression⁸. Therefore, we sought to explore the role of disulfidptosis in LUAD, particularly in relation to TIME.

The PPP, which is closely related to disulfidptosis, is a major glucose catabolism pathway that differs from glycolysis⁹. The PPP is essential for cancer cells to meet their anabolic needs and respond to oxidative damage¹⁰. Additionally, activated immune cells rely on the PPP to regulate the immune response¹¹. However, despite the well-known role of the PPP in promoting tumor progression and immune regulation, there is a lack of systematic studies on its effects in LUAD. Therefore, exploring the effects of the PPP on the TIME in LUAD is worth considering, as it may help develop effective immunotherapeutic strategies.

In this study, we analyzed the characteristics of disulfide regulators (DSRs) in LUAD patients using data from public databases and performed a consensus clustering analysis based on DSR expression. By performing WGCNA and single-cell analysis, we identified genes associated with disulfidptosis and PPP, respectively, and identified DSRs and PPP-related genes (DPRGs). Using DPRGs to construct a risk signature, we systematically analyzed clinical prognosis, TIME, and response to immunotherapy in LUAD patients. We hypothesized that the risk signature would effectively differentiate LUAD patients, allowing for more individualized and effective therapeutic strategies.

2.1 Transcriptome dataset collection

We obtained transcriptome data and clinical information for patients with LUAD from the TCGA-LUAD cohort in The Cancer Genome Atlas (TCGA, https://portal.gdc.cancer.gov/), as well as from datasets GSE68645 and GSE26939 in the Gene Expression Omnibus database (GEO, https://www.ncbi.nlm.nih.gov/geo/). To eliminate batch effects on data from different databases, we used the sva R package. We retrieved a gene set containing nine DSRs from the report by Gan et al. (5). Additionally, we obtained thirty pentose phosphate pathway-related genes (PRGs) from the Reactome database (https://reactome.org/, R-HSA-71336) and the Kyoto Encyclopedia of Genes and Genomes (KEGG, https://www.genome.jp/kegg/) database (has00030).

2.2 Weighted Gene Coexpression Network Analysis (WGCNA)

In this study, we utilized WGCNA to identify genes associated with disulfidptosis and PPP. Outliers were removed, and soft thresholds were selected based on the scale-free topology criterion to construct an expression network. Gene modules related to disulfidptosis and PPP were then identified. For the analysis, a scale-free topological fit index of 1–20 was used, with a screening criterion of 0.9, and a minimum of 30 genes per module was set.

2.3 Single-cell Analysis and AUCell

Raw data for single-cell analysis were downloaded from GSE149655. The raw count matrices were merged and converted into an expression matrix. The following pre-processing steps were applied: samples were excluded if fewer than three cells expressed the genes, if fewer than 200 genes occurred in a single cell, or if the cells contained too many mitochondrial genes and an abnormal number of expressed genes. Next, the data were normalized to enable practical comparison between different cell samples. Principal component analysis (PCA) was then used to reduce the dimensionality of the data and reveal the principal factors of the different subgroups. Clustering operations were performed using the Leiden algorithm, and the resolution was set to 0.25. Cell types were annotated, and marker genes were compared across taxa based on known cell markersTo identify the most significant genes associated with disulfidptosis and PPP, we employed the AUCell R software package, which assigned a disulfidptosis activity fraction and PPP activity fraction to each cell. The area under the curve (AUC) scores of the DRGs and PRGs were calculated to determine their enrichment levels in the expressed genes of each cell. Subsequently, cells were classified based on their median AUC score.

2.4 Consensus clustering analysis

Consensus clustering was performed using the expression levels of the nine DSRs. The overlapping genes between differentially expressed genes (DEGs) among the DSR clusters, gene modules identified by WGCNA, and DEGs obtained from the scRNA-Seq analysis, were selected as DPRGs for further analysis. Using univariate Cox regression analysis, we screened the DPRGs that were associated with overall survival (OS). Subsequently, clustering analysis based on these OS-related DPRGs yielded two DPRG clusters.

2.5 Analysis of infiltrating immune cell population

The abundance of infiltrating immune cells was quantified using multiple software packages, including the Estimation of STromal and Immune cells in MAlignant Tumor tissues Expression algorithm¹², single-sample gene set enrichment analysis (ssGSEA), xCELL¹³, TIMER¹⁴, quanTIseq¹⁵, MCP-counter¹⁶, EPIC and CIBERSORT¹⁷.

2.6 Enrichment analysis

The dataset c2.cp.kegg.v2022.1.Hs.symbols was obtained from MsigDB (https://www.gsea-msigdb.org/gsea/msigdb) and used for differential functional identification through the GSVA R package. GSEA 4.3.2 software was used to perform gene set enrichment analysis (GSEA) of the curated gene sets.

2.7 Construction of the risk signatures

By performing the least absolute shrinkage and selection operator (LASSO) Cox regression on the OS-related-DPRGs, we constructed a risk signature containing seven genes. The risk score was determined using the formula: Σ(coefficient_k*expression_k), where coefficient_k and expression_k are the risk coefficient and expression of each gene, respectively. After calculating the risk scores for each sample, based on the median score, the patients were divided into low-risk and high-risk groups.

2.8 Evaluation of the risk signatures

The prognostic significance of the risk signature was assessed using receiver-operating characteristic (ROC) curve analysis, Cox regression, and Kaplan–Meier survival analysis, and the concordance index was calculated for both the training and test sets. Nomograms were generated using the rms R package, and their accuracy was evaluated using cumulative hazard and calibration curves. To assess the benefits of the intervention with nomograms and risk scores for LUAD patients, Decision Curve Analysis (DCA) was performed.

2.9 Analyses of immune checkpoints expression and immunotherapy response in different risk signature subgroups

The limma R package was used to examine variations in immune checkpoint expression between the subgroups. Additionally, we assessed the responses of patients in different clusters and signature subgroups to immunotherapy according to the Tumor Immune Dysfunction and Exclusion (TIDE) score¹⁸ as well as The Cancer Immunome Atlas (https://tcia.at/home).

2.10 Examination of the drug sensitivity

The OncoPredict R package¹⁹ was used to screen for sensitive drugs between different risk signature subgroups.

2.11 Pancancer analysis of LRRC61

We assessed the differential expression of LRRC61 and its prognostic impact on multiple cancers. In addition, we performed a correlation analysis of LRRC61 expression with the degree of immune cell infiltration and immune checkpoint expression.

2.12 Cell culture and tissue collection

All cell lines were purchased from the Institute of Biochemistry and Cell Biology (Shanghai, China) and were maintained in Dulbecco's modified Eagle's medium (Gibco, Life Technologies, CA, USA) supplemented with 10% fetal bovine serum (FBS) (Hyclone, Logan, UT, USA) and 100 µg/mL penicillin-streptomycin (Hyclone).

Eight LUAD and matching paracancerous non-tumor tissue samples were obtained during routine surgery at the First Affiliated Hospital of Nanjing Medical University, and the specimens were stored at − 80°C until further testing.

2.13 Reverse transcription-quantitative PCR

Total RNA was isolated from tissues using TRIzol reagent (Takara Bio, Shiga, Japan) and reverse-transcribed to cDNA using a PrimeScriptTM RT reagent Kit (Takara). SYBR® Premix Ex Taq (Takara) was used for qRT-PCR analysis. Detailed primer sequences are listed in Supplementary Table S1.

2.14 Immunohistochemistry Staining

Slices of tissue were blocked and subjected to suitable concentrations of primary antibody (Sigma-Aldrich, #HPA019355) at 4°C. Biotinylated sections were treated for 30 min with the secondary antibody (Dako, Denmark). Finally, the slices were stained with a Dako diaminobenzidine (DAB) kit and counterstained with hematoxylin (Baso, Zhuhai, China).

2.15 RNA interference

For the knockdown of LRRC61, specific small interference RNAs (siRNAs) and scrambled siRNA (si-NC) were transfected into A549 and H1299 cells using LipofectamineTM 3000 (Invitrogen, Carlsbad, CA, USA). Target RNAs were measured 48 h after transfection using qRT-PCR. The siRNA sequences are listed in Supplementary Table S1.

2.16 Cell proliferation assay

LUAD cells were planted in 96-well plates, followed by the addition of 20 µL CCK8 solution to each well. At 1, 2, 3, 4, and 5 days, the absorbance at 450 nm (OD450) was determined using a microplate reader (Thermo Fisher, Multiskan 51119000, MA, USA).

2.17 Colony formation

The cells were maintained in 6-well plates for 2 weeks. Cell colonies were stained with crystal violet solution for 1.5 h after being fixed with 4% formaldehyde for 10 min. The mean colony counts were calculated.

2.18 Transwell assay

For the migration assay, 1 × 105 cells were plated in the top chamber of the Transwell chamber, and 10% FBS culture medium was added to the lower chamber. The cells that migrated across the membrane were photographed and counted under a microscope (Nikon, Ts2R, Japan).

2.19 Apoptosis assay

Cells were seeded into 6-well plates. Then we stained cells for 15 min with the annexin V-FITC/PI Apoptosis Detection Kit (Beyotime Institute of Biotechnology, Jiangsu, China). Samples were analyzed using FACSCanto TM II flow cytometry (Becton, Dickinson and Company).

2.20 Animal studies

Nanjing Medical University’s Committee on Ethics for Animal Experiments ap-proved the experimental procedures (IACUC-1706007). We used 6–8-week-old male BALB/c nude mice (SPF grade) from the Nanjing Medical University Animal Center. H1299 (1 × 106) cells were injected into the right side of each mouse. Every 3 days, the tumour volume was measured. Six mice were randomly divided into the control and LRRC61 groups. The mice were euthanized at a right time. Tumor width and length were recorded, and tumor volume cal-culated using the formula: [width 2 × length]/2. In order to euthanize the mice, an appropriate amount of mice were put in a euthanasia box, then carbon dioxide was introduced. When the signs of life disappear completely, make sure that there is no breathing and no heartbeat, we remove the body. This approach meets the requirements of animal ethics.

2.21 Statistical analysis

The Wilcoxon rank test and Kruskal–Wallis test were used to comparing two independent groups and three or more independent groups, respectively. The differences in patient survival were evaluated using Kaplan-Meier and log-rank tests. Spearman's rank correlation analysis was used for the correlation analysis. The R software (version 4.2.3) and GraphPad Prism 9 were used for all statistical analyses.

3.1 Characteristics and prognostic value of DSRs in LUAD

After removing the batch effect of GSE68545 and TCGA-LUAD (Figs. 1A and 1B), we analyzed the nine DSRs using the combined data. Their chromosomal distribution and CNVs are displayed in Fig. 1C and 1D, respectively. Among them, RAC1, SLC2A1, NCKAP1, SLC3A2, ABI2, and BRK1 exhibited significant copy number amplification, while WASF2, CYFIP1, and SLC7A11 showed copy number deletions. Moreover, the expression levels of DSRs were markedly different between LUAD tumors and healthy tissues (Fig. 1E). The network diagram in Fig. 1F illustrates the interactions and prognostic values of the DSRs, where only BRK1 played a favorable prognostic role, and all other DSRs were prognostic risk factors. We also conducted Kaplan-Meier survival analysis of individual DSRs separately, and the outcomes were consistent with the network diagram (Fig. 1G). Additionally, SLC7A11, SLC3A2, NCKAP1, RAC1, and SLC2A1 were identified as independent prognostic factors in patients with LUAD (Fig. 1H).

3.2 Identification of distinct clusters and their functions based on DSR expression

To investigate whether DSR expression was associated with clinical characteristics in LUAD, we performed a consensus clustering analysis based on the nine DSRs. We identified two DSR clusters (Fig. 2A, Table S2), and Kaplan–Meier survival analysis showed that patients in cluster B had a higher likelihood of survival (Fig. 2B). The DSR clusters effectively differentiated patients with LUAD, as shown by PCA, t-SNE, and UMAP analysis (Figs. 2C–2E). The expression of DSRs and associated LUAD clinical characteristics in the two DSR clusters is depicted in Fig. 2F. Differential expression analysis revealed that all DSRs except RAC1 were more highly expressed in cluster A (Fig. 2G). We also observed variations in immune infiltration levels between the two DSR clusters, with cluster A having higher ratios of CD56dim neutrophils, type 2 T helper cells, and natural killer cells, while the majority of the remaining immune cells were significantly more abundant in cluster B (Fig. 2H). Finally, we identified 3519 DEGs between the two DSR clusters for further analysis.

To investigate the biological roles of the DSRs in the different DSR clusters, we performed GSEA (Figs. 2I and 2J) and GSVA (Figs. 2K and 2L). Consistent with the results of the immune infiltration analysis, cluster B showed a stronger correlation with the immune response, which could explain the better survival outcome of LUAD patients in cluster B. On the other hand, cluster A was associated with the metabolism of various compounds involved in biochemical processes, including the disulfidptosis process (highlighted in blue) and the pentose phosphate pathway (highlighted in green), which is closely related to disulfidptosis.

3.3 Identification of gene modules highly associated with disulfidptosis and pentose phosphate pathways by WGCNA

To identify genes associated with DSR and PPP, we performed WGCNA. We chose 5 as the soft threshold based on the scale-free topology fit index to generate a proximity matrix (Fig. 3A). After module cutting with a dynamic tree and merging similar gene modules, 12 gene modules were identified and assigned various color designations (Fig. 3B) which were then correlated with DSR and PPP (Fig. 3C). Among them, the black (Fig. 3D), pink (Fig. 3E), salmon (Fig. 3F), and turquoise modules (Fig. 3G) had a strong correlation with PPP, while the midnight blue (Fig. 3H), salmon (Fig. 3I), and turquoise (Fig. 3J) modules showed a strong correlation with DSR. A total of 6305 genes from these modules were selected for further analysis.

3.4 Identification of genes highly associated with disulfidptosis and pentose phosphate pathways by single-cell analysis

We conducted a single-cell analysis to investigate genes associated with DSR and PPP. To ensure sample validity, we removed doublets and filtered cells based on mitochondrial gene ratio (Figure S1A). The number of tagged transcripts was positively correlated with the number of detected genes (Figure S1B), and we corrected for cell cycle heterogeneity (Figure S1C). After quality control, we removed batch effects (Fig. 4A), and we used t-SNE dimensionality reduction to classify cells into 28 clusters (Fig. 4B) that were validated using cell markers (Fig. 4C). Gene-specific expression for different cell types is shown in Figure S1D. We used the AUCell R package to separately score the enrichment of DSRs and PRGs (Table S3) in each cell (Figs. 4D and 4F) and classified cells into low and high groups based on the median DSR-AUC score/PPP-AUC score (Figs. 4E and 4G). B cells, mononuclear phagocytes, fibroblasts, and epithelial cells showed higher DSR-AUC scores, whereas B cells and mononuclear phagocytes had higher PPP-AUC scores. We identified 1983 DEGs between the low and high DSR-AUC groups and 1846 DEGs between the low and high PPP-AUC groups for further analysis. GSVA revealed that in addition to being associated with disulfidptosis and PPP (marked in yellow), DSRs and PRGs may participate in pathways such as oxidative phosphorylation, ferroptosis, and protein export (marked in green) that differentiate the low and high AUC groups (Figs. 4H and 4I).

3.5 Construction and assessment of risk signature based on OS-related DPRGs

Based on the previous analysis, we identified 2050 DSRs and PRGs (DPRGs) (Figure S2A), out of which 302 OS-related DPRGs were selected (p < 0.0001) using univariate COX regression analysis (Table S4). Functional enrichment analysis of these OS-related DPRGs revealed their association with PPP and platinum drug resistance (Figure S2B). We performed a consensus clustering analysis of LUAD patients in the study cohort using these OS-related DPRGs, resulting in two DPRG clusters (Figure S2C). The DPRG clusters effectively stratified LUAD patients (Figure S2D-S2F), with patients in cluster B demonstrating higher survival rates, as evidenced by Kaplan–Meier survival analysis (Figure S2G). Additionally, most DSRs were expressed at lower levels in DPRG cluster B (Figure S2H).

Finally, we performed LASSO on the OS-related DPRGs to establish corresponding risk signatures (Figs. 5A and 5B). The risk signature consisted of seven genes (RTN4, ACHE, ABAT, EGLN1, FAM117A, LRRC61, and ADM), and their correlations with DSRs and PRGs are shown in Figure S3A and Figure S3B, respectively. All seven risk model genes can be independently presented as prognostic indicators in patients with LUAD, as demonstrated in Fig. 5C. To generate risk scores for patients in the TCGA-BLCA and GSE68465 cohorts, we utilized the expression and coefficient of each candidate gene in the risk signatures (Table S5). We evenly divided the patient population into training and test subsets to analyze the accuracy of the signatures.

To validate the risk signature, we conducted Kaplan–Meier OS analysis (Figs. 5D–5F). For all categories, survival was significantly greater in the low-risk group. The risk signature exhibited good predictive ability at 1-, 3-, and 5-year time points, according to time-dependent ROC curves (Figs. 5G–5I). The relationships among the DSR cluster, risk score, and LUAD patient survival status are summarized in a Sankey plot in Fig. 5J. We then examined the distribution and variations in risk scores between the DSR and DPRG clusters (Figures S3C, S3D). Furthermore, there were notable differences in survival status, gender, T-stage, N-stage, and TNM stage between the low-risk and high-risk groups of LUAD patients (Fig. 5K), and the distribution statistics are shown in Figs. 5L–5P.

3.6 Value for prognosis of the risk signature

In Fig. 6A, it is evident that the risk score acts as an independent prognostic factor, and the concordance index values indicate that the risk signature exhibits the best predictive ability compared to other clinical characteristics (Fig. 6B). We also created a nomogram that plots the prognostic value of the risk signature along with clinical features to quantify their predictive abilities in patients with LUAD (Fig. 6C). To assess the precision of their predictive abilities, we conducted cumulative hazard and calibration assessments (Figs. 6D, 6E). Additionally, decision curve analysis at 1-, 3-, and 5-year time points demonstrated that interventions based on the nomogram and risk score resulted in greater benefits for LUAD patients than those based on other clinical traits (Figs. 6F–6H). Finally, we conducted KEGG (Fig. 6I) and GO (Fig. 6J) functional enrichment analysis using GSEA, which suggested that low risk was due to the immune response, while the high risk was associated with cell cycle, PPP, actin cytoskeleton regulation, cancer pathways, and disulfide oxidoreductase. This indicates a significant link between the risk model we developed and disulfidptosis.

3.7 Correlations of the risk signature scores with TIME and immunotherapy response

Given the functional enrichment results that suggest a correlation between the risk score and the immune response, we were particularly interested in exploring the interaction between the risk score and the tumor immune microenvironment (TIME). Figure 7A and Figure S4A illustrate the relationship between the risk score and the infiltration of immune cells, and the degree of immune cell infiltration for each gene in the risk signature is plotted in Fig. 7B. There were variations in immune checkpoint expression levels between the low-risk and high-risk groups (Fig. 7C), and the candidate genes for the risk signature showed different correlations with immune checkpoints (Fig. 7D). The high-risk group exhibited higher expression levels of immunosuppressive factors regarding cytokines (Figure S4B). There were observable differences in immune cell quantity and tumor purity (ESTIMATE score) between the low-risk and high-risk groups (Fig. 7E). Furthermore, there was a difference in TIDE scores for the efficacy of immunotherapy between the two groups (Fig. 7F). While the immunophenoscore score for the CTLA4–PD1 + immunophenotype did not differ significantly between the two groups, all other groups showed differences in immunotherapy response (Figs. 7G–7J). Additionally, the combined analysis of risk signatures and immune checkpoints could more effectively differentiate the prognosis of LUAD patients (Figure S5). Finally, we performed a drug sensitivity analysis of the DPRG risk signature. As the high-risk group was insensitive to clinically used chemotherapeutic agents (Figure S6A), we screened for potential therapeutic agents that might benefit them (Figure S6B).

3.8 Function enrichment, expression and prognosis analysis of LRRC61

We discovered that LRRC61, one of the candidate genes for the risk signature, is involved in disulfide formation and is associated with PPP and cancer pathways (Figs. 8A–8D). Due to the limited research on LRRC61 to date, we aimed to investigate its characteristics in cancer. The expression of LRRC61 was found to be high in numerous tumor tissues (Fig. 8E) and can be utilized as an independent prognostic factor (Fig. 8F). Specifically, high expression of LRRC61 was correlated with LUAD progression (Fig. 8G) and poor overall survival (Figs. 8H–8J). In a pan-cancer analysis, LRRC61 was also linked with immune infiltration (Figure S7) and immune checkpoints (Figure S8).

3.9 Role of LRRC61 as an oncogene of LUAD cells in vitro

We validated LRRC61 expression in eight pairs of LUAD samples and their corresponding noncancerous lung samples using qRT-PCR (Fig. 9A) and IHC staining (Fig. 9B). We observed that LRRC61 was upregulated in LUAD tissues and cells (Fig. 9C), which is consistent with the RNA-seq results. Subsequently, we successfully silenced LRRC61 in A549 and H1299 cells using specific siRNAs. LRRC61 knockdown inhibited the proliferation (Figs. 9D – 9F), migration (Fig. 9G), and apoptosis resistance (Fig. 9H) of A549 and H1299 cells. These results indicate that LRRC61 functions as an oncogene during the development of LUAD tumors. We established subcutaneous tumor formation models by 6–8-week-old male BALB/c nude mice (SPF grade) to evaluate the effects of LRRC61 in vivo. Images shows the xenograft tumors of the LRRC61 group were bigger than that of the control group. Moreover, tumor volume and tumor weight were measured to evaluate nude mice.

Recent research has identified a new form of cell death induced by disulfide stress called disulfidptosis ²⁰. In the presence of glucose starvation or NADPH depletion, excessive accumulation of disulfide molecules in cells with high SLC7A11 expression causes disulfide stress. Aberrant disulfide bonds across the actin cytoskeleton led to actin cytoskeletal collapse and cell death. This study also confirmed that NADPH is essential for providing reducing power to break down abnormally accumulated disulfide bonds, thereby inhibiting disulfidptosis ²⁰. Therefore, NADPH, which is mainly derived from PPP, plays an essential role in both preventing disulfidptosis and reversing disulfide stress. In this study, we included PPP-related genes (PRGs) and analyzed disulfidptosis in LUAD using an integrated approach.

In addition to SLC7A11, SLC3A2 (which encodes the SLC7A11 chaperone)^21,22, and SLC2A1 (glucose transporter), Gan et al. demonstrated that the Rac-WAVE regulatory complex (WRC)-mediated lamellipodia formation promotes disulfidptosis, while deletion of either component of the WRC could inhibit disulfidptosis. Therefore, we selected RAC1 and components of the WRC (NCKAP1, WASF2, CYFIP1, ABI2, and BRK1) as DSRs. All nine DSRs were associated with the prognosis of LUAD except WASF2. Among them, five (SLC7A11, SLC3A2, NCKAP1, RAC1, and SLC2A1) could be considered independent prognostic risk factors.

In this study, the patient cohort was divided into two distinct DSR clusters based on DSR expression, with patients in cluster A showing higher expression and worse overall prognosis, while patients in cluster B exhibited higher immune infiltration, which is known to have a positive effect on prognosis^23–25. Enrichment function comparison of the two DSR clusters revealed that cluster B was associated with favorable immune responses, while cluster A was involved in metabolic pathways related to disulfidptosis and PPP. The analysis based on single-cell data also showed that PRGs were significantly involved in the disulfidptosis process.

After conducting the aforementioned analysis, we obtained the DPRGs and utilized them for secondary clustering analysis. The resulting DPRG clusters could effectively distinguish between LUAD patients and provide a dependable foundation for further analysis. The functional annotation of DPRGs demonstrated their involvement in regulating the cell cycle, immune response, and platinum drug resistance, in addition to PPP. Given that platinum drugs are frequently employed in LUAD treatment regimens, it is clinically relevant to employ DPRGs for assessing varying susceptibilities to platinum drugs in patients with LUAD.^26–28.

The risk signatures constructed using DPRGs are highly effective in predicting the prognosis of LUAD, and LUAD patients with low-risk scores still benefit from immune response modulation. Despite the benefits of immunotherapy, most patients experience disease progression²⁹. Therefore, there is an urgent need to optimize immunotherapy regimens³⁰. Multi-targeted combination immunotherapy can significantly improve immunotherapy efficacy and has been reported in a variety of tumors including lung cancer, pancreatic cancer and colorectal cancer^31–33. Relationship between cancer immune response and mechanisms of resistance to immunotherapy has been under investigation for a long time. combination therapies (e.g., immunotherapy with chemotherapy, radiation therapy and targeted therapy), and discuss combination therapies approved by the US Food and Drug Administration demonstrated benefits to patients. Many targeted therapies such as targeting cytokines and other soluble immunoregulatory factors, ACT, virotherapy, innate immune modifiers and cancer vaccines³⁴, as well as combination therapies that exploit alternative immune targets and other therapeutic modalities were studied³². We found that the risk signature related to DPRGs could be implemented to guide immunotherapy planning, given that patients with high-risk scores expressed stimulatory immunological checkpoints much less than those with low-risk scores did, and as a result, responded to immunotherapy less effectively.We believe that our study will also bring benefits to oncology patients, especially LUAD patients.

TIME is involved in tumorigeneses, progressions, and metastases³⁵. The intra-tumor immune landscape is a critical factor influencing patient survival and response to immunotherapy³⁶. TIME plays a significant role in the progression of lung cancers^37,38 and is linked to prognosis³⁹. Multiple algorithms to assess the abundance of immune infiltration showed that risk score was significantly positively correlated with neutrophil and negatively correlated with T cells and B cells, consistent with reported^40,41. Tumor-associated neutrophils enhance the proliferation of tumor cells through pathways such as neutrophil extracellular traps⁴². This result suggests that risk score may weaken immune response and promote tumor development by regulating TIME. In addition to being associated with immune cells and immune checkpoints, the expression of other critical molecules in the TIME regulatory network, such as cytokines, chemokines, and growth factors⁴³, also differed significantly between the low-risk and high-risk groups. Therefore, the DPRG-related risk signature may have significant clinical value in regulating the TIME network and preventing immune escape.

LRRC61, a functionally enriched component gene in the risk signature, is associated with disulfidptosis and PPP. It is highly expressed in several cancer types and demonstrates prognostic value, including in LUAD. In vitro experiments showed that LRRC61 knockdown successfully reduced the ability of LUAD cells to proliferate, invade, migrate, and prevent apoptosis. Therefore, the mechanism by which LRRC61 exerts oncogenic effects in LUAD through the regulation of disulfidptosis and PPP deserves further investigation.

This study has several limitations. The study subjects were enrolled based on the limited resources available in public databases, and the retrospective analysis may be subject to selection bias, which affects accuracy. Moreover, clinical information on patients with LUAD in public databases is not comprehensive, which may have led to the omission of additional characteristics associated with disulfidptosis and PPP.

Through the analysis of bulk RNA-seq and scRNA-seq, we identified DPRGs. By constructing a DPRG-related risk signature, we revealed the relevance of disulfidptosis and PPP in LUAD to the clinical characteristics, prognosis, TIME, and immunotherapeutic response of patients, and provided insight into the underlying regulatory mechanisms. These promising predictive tools may help the development of more effective and personalized treatment strategies for patients with LUAD.

Patents:

None

Ethics approval and consent to participate:

IACUC-2210031

Consent for publication:

Not applicable

Competing interests:

The authors declare that they have no competing interests.

Funding:

This study was supported by grants from the National Natural Science Foundation of China (Grant Nos. 82172889), and the Clinical Frontier Technology of Jiangsu Provincial Department of Science and Technology (BE2020783).

Author Contribution

Zhou HY and Xiao Y wrote the main manuscript and completed most of the experiments. Wu J and Wang XZ provided ideas and reviewed data. All authors reviewed the manuscript.

Data Availability

Data is provided within the manuscript or supplementary information files.

Guo X, et al. Global characterization of T cells in non-small-cell lung cancer by single-cell sequencing. Nat Med. 2018;24:978–85. 10.1038/s41591-018-0045-3.
Yang L, et al. Single-cell transcriptome analysis revealed a suppressive tumor immune microenvironment in EGFR mutant lung adenocarcinoma. J Immunother Cancer. 2022;10. 10.1136/jitc-2021-003534.
Ramos-Casals M, et al. Immune-related adverse events of checkpoint inhibitors. Nat reviews Disease primers. 2020;6. 10.1038/s41572-020-0160-6.
Mou Y, et al. Ferroptosis, a new form of cell death: opportunities and challenges in cancer. J Hematol Oncol. 2019;12. 10.1186/s13045-019-0720-y.
Liu X, et al. Actin cytoskeleton vulnerability to disulfide stress mediates disulfidptosis. Nat Cell Biol. 2023;25:404–14. 10.1038/s41556-023-01091-2.
Liu X, et al. Cystine transporter regulation of pentose phosphate pathway dependency and disulfide stress exposes a targetable metabolic vulnerability in cancer. Nat Cell Biol. 2020;22:476–86. 10.1038/s41556-020-0496-x.
Hu K, et al. Suppression of the SLC7A11/glutathione axis causes synthetic lethality in KRAS-mutant lung adenocarcinoma. J Clin Investig. 2020;130:1752–66. 10.1172/JCI124049.
Jiang P, Du W, Wu M. Regulation of the pentose phosphate pathway in cancer. Protein Cell. 2014;5:592–602. 10.1007/s13238-014-0082-8.
Patra KC, Hay N. The pentose phosphate pathway and cancer. Trends Biochem Sci. 2014;39:347–54. 10.1016/j.tibs.2014.06.005.
Ma H, et al. c-Src facilitates tumorigenesis by phosphorylating and activating G6PD. Oncogene. 2021;40:2567–80. 10.1038/s41388-021-01673-0.
Ghergurovich JM, et al. A small molecule G6PD inhibitor reveals immune dependence on pentose phosphate pathway. Nat Chem Biol. 2020;16:731–9. 10.1038/s41589-020-0533-x.
Yoshihara K, et al. Inferring tumour purity and stromal and immune cell admixture from expression data. Nat Commun. 2013;4:2612. 10.1038/ncomms3612.
Aran D, Hu Z, Butte AJ. xCell: digitally portraying the tissue cellular heterogeneity landscape. Genome Biol. 2017;18:220. 10.1186/s13059-017-1349-1.
Li T, et al. A Web Server for Comprehensive Analysis of Tumor-Infiltrating Immune Cells. Cancer Res. 2017;77:e108–10. 10.1158/0008-5472.CAN-17-0307.
Finotello F, et al. Molecular and pharmacological modulators of the tumor immune contexture revealed by deconvolution of RNA-seq data. Genome Med. 2019;11:34. 10.1186/s13073-019-0638-6.
Becht E, et al. Estimating the population abundance of tissue-infiltrating immune and stromal cell populations using gene expression. Genome Biol. 2016;17:218. 10.1186/s13059-016-1070-5.
Newman AM, et al. Robust enumeration of cell subsets from tissue expression profiles. Nat Methods. 2015;12:453–7. 10.1038/nmeth.3337.
Jiang P, et al. Signatures of T cell dysfunction and exclusion predict cancer immunotherapy response. Nat Med. 2018;24:1550–8. 10.1038/s41591-018-0136-1.
Maeser D, Gruener RF, Huang RS. oncoPredict: an R package for predicting in vivo or cancer patient drug response and biomarkers from cell line screening data. Brief Bioinform. 2021;22. 10.1093/bib/bbab260.
(!!!. INVALID CITATION !!!).
Koppula P, Zhuang L, Gan B. Cystine transporter SLC7A11/xCT in cancer: ferroptosis, nutrient dependency, and cancer therapy. Protein Cell. 2021;12:599–620. 10.1007/s13238-020-00789-5.
Kunda P, Craig G, Dominguez V, Baum B, Abi. Sra1, and Kette control the stability and localization of SCAR/WAVE to regulate the formation of actin-based protrusions. Curr biology: CB. 2003;13:1867–75. 10.1016/j.cub.2003.10.005.
Leonetti A, et al. Molecular basis and rationale for combining immune checkpoint inhibitors with chemotherapy in non-small cell lung cancer. Drug Resist updates: reviews commentaries Antimicrob anticancer Chemother. 2019;46:100644. 10.1016/j.drup.2019.100644.
Abbasi J. Potential New Immune-Checkpoint Inhibitor Partner for Lung Cancer. JAMA. 2021;326:901. 10.1001/jama.2021.14765.
Sun J, et al. Characterization of immune landscape in papillary thyroid cancer reveals distinct tumor immunogenicity and implications for immunotherapy. Oncoimmunology. 2021;10:e1964189. 10.1080/2162402X.2021.1964189.
Rudin CM, et al. Pembrolizumab or Placebo Plus Etoposide and Platinum as First-Line Therapy for Extensive-Stage Small-Cell Lung Cancer: Randomized, Double-Blind, Phase III KEYNOTE-604 Study. J Clin oncology: official J Am Soc Clin Oncol. 2020;38:2369–79. 10.1200/JCO.20.00793.
Gadgeel S, et al. Updated Analysis From KEYNOTE-189: Pembrolizumab or Placebo Plus Pemetrexed and Platinum for Previously Untreated Metastatic Nonsquamous Non-Small-Cell Lung Cancer. J Clin oncology: official J Am Soc Clin Oncol. 2020;38:1505–17. 10.1200/JCO.19.03136.
Jassem J, et al. Updated Overall Survival Analysis From IMpower110: Atezolizumab Versus Platinum-Based Chemotherapy in Treatment-Naive Programmed Death-Ligand 1-Selected NSCLC. J Thorac oncology: official publication Int Association Study Lung Cancer. 2021;16:1872–82. 10.1016/j.jtho.2021.06.019.
Passaro A, Brahmer J, Antonia S, Mok T, Peters S. Managing Resistance to Immune Checkpoint Inhibitors in Lung Cancer: Treatment and Novel Strategies. J Clin oncology: official J Am Soc Clin Oncol. 2022;40:598–610. 10.1200/JCO.21.01845.
Shi R, et al. APOBEC-mediated mutagenesis is a favorable predictor of prognosis and immunotherapy for bladder cancer patients: evidence from pan-cancer analysis and multiple databases. Theranostics. 2022;12:4181–99. 10.7150/thno.73235.
Lahiri A, et al. Lung cancer immunotherapy: progress, pitfalls, and promises. Mol Cancer. 2023;22:40. 10.1186/s12943-023-01740-y.
Zhu S, et al. Combination strategies to maximize the benefits of cancer immunotherapy. J Hematol Oncol. 2021;14. 10.1186/s13045-021-01164-5.
Agarwal P, Le DT, Boland PM. Immunotherapy in colorectal cancer. Adv Cancer Res. 2021;151:137–96. 10.1016/bs.acr.2021.03.002.
Lorentzen CL, Haanen JB, Met O, Svane IM. Clinical advances and ongoing trials on mRNA vaccines for cancer treatment. Lancet Oncol. 2022;23:e450–8. 10.1016/S1470-2045(22)00372-2.
Sun S, et al. Development and validation of an immune-related prognostic signature in lung adenocarcinoma. Cancer Med. 2020;9:5960–75. 10.1002/cam4.3240.
Zhu M, et al. Clinical Significance and Immunometabolism Landscapes of a Novel Recurrence-Associated Lipid Metabolism Signature In Early-Stage Lung Adenocarcinoma: A Comprehensive Analysis. Front Immunol. 2022;13:783495. 10.3389/fimmu.2022.783495.
Sorin M, et al. Single-cell spatial landscapes of the lung tumour immune microenvironment. Nature. 2023;614:548–54. 10.1038/s41586-022-05672-3.
Zuo S, Wei M, Wang S, Dong J, Wei J. Pan-Cancer Analysis of Immune Cell Infiltration Identifies a Prognostic Immune-Cell Characteristic Score (ICCS) in Lung Adenocarcinoma. Front Immunol. 2020;11:1218. 10.3389/fimmu.2020.01218.
Shinohara S, et al. New evaluation of the tumor immune microenvironment of non-small cell lung cancer and its association with prognosis. J Immunother Cancer. 2022;10. 10.1136/jitc-2021-003765.
Linde IL et al. Neutrophil-activating therapy for the treatment of cancer. Cancer cell 41, 356–372 e310, 10.1016/j.ccell.2023.01.002 (2023).
Paijens ST, Vledder A, de Bruyn M, Nijman HW. Tumor-infiltrating lymphocytes in the immunotherapy era. Cell Mol Immunol. 2021;18:842–59. 10.1038/s41423-020-00565-9.
Xiao Y et al. Cathepsin C promotes breast cancer lung metastasis by modulating neutrophil infiltration and neutrophil extracellular trap formation. Cancer cell 39, 423–437 e427, 10.1016/j.ccell.2020.12.012 (2021).
Gajewski TF, Schreiber H, Fu YX. Innate and adaptive immune cells in the tumor microenvironment. Nat Immunol. 2013;14:1014–22. 10.1038/ni.2703.

Table 1 is available in the Supplementary Files section.

No competing interests reported.

Download PDF

Editorial decision: Revision requested
10 Sep, 2024
Editor assigned by journal
09 Sep, 2024
Submission checks completed at journal
09 Sep, 2024
First submitted to journal
08 Sep, 2024

You are reading this latest preprint version

Disulfidptosis and pentose phosphate pathway-associated prognosis signature guides immunotherapy for lung adenocarcinoma patients

Status:

Version 1

Abstract

Background

Research Design and Methods:

Results

Conclusion

Figures

1. Introduction

2. Materials and Methods

2.1 Transcriptome dataset collection

2.2 Weighted Gene Coexpression Network Analysis (WGCNA)

2.3 Single-cell Analysis and AUCell

2.4 Consensus clustering analysis

2.5 Analysis of infiltrating immune cell population

2.6 Enrichment analysis

2.7 Construction of the risk signatures

2.8 Evaluation of the risk signatures

2.9 Analyses of immune checkpoints expression and immunotherapy response in different risk signature subgroups

2.10 Examination of the drug sensitivity

2.11 Pancancer analysis of LRRC61

2.12 Cell culture and tissue collection

2.13 Reverse transcription-quantitative PCR

2.14 Immunohistochemistry Staining

2.15 RNA interference

2.16 Cell proliferation assay

2.17 Colony formation

2.18 Transwell assay

2.19 Apoptosis assay

2.20 Animal studies

2.21 Statistical analysis

3. Results

3.1 Characteristics and prognostic value of DSRs in LUAD

3.2 Identification of distinct clusters and their functions based on DSR expression

3.3 Identification of gene modules highly associated with disulfidptosis and pentose phosphate pathways by WGCNA

3.4 Identification of genes highly associated with disulfidptosis and pentose phosphate pathways by single-cell analysis

3.5 Construction and assessment of risk signature based on OS-related DPRGs

3.6 Value for prognosis of the risk signature

3.7 Correlations of the risk signature scores with TIME and immunotherapy response

3.8 Function enrichment, expression and prognosis analysis of LRRC61

3.9 Role of LRRC61 as an oncogene of LUAD cells in vitro

4. Discussion

5. Conclusions

Declarations

Patents:

Ethics approval and consent to participate:

Consent for publication:

Competing interests:

Funding:

Author Contribution

Data Availability

References

Tables

Additional Declarations

Supplementary Files

Status:

Version 1