Integrating bulk, single-cell, and spatial transcriptomics to identify and functionally validate novel targets to enhance immunotherapy in NSCLC

Cao, Kui; Wei, Shenshui; Ma, Tianjiao; Yang, Xinxin; Wang, Yuning; He, Xiangrong; Lu, Mengdi; Bai, Yuwen; Qi, Cuicui; Zhang, Luquan; Li, Lijuan; Meng, Hongxue; Ma, Jianqun; Zhu, Jinhong

doi:10.1038/s41698-025-00893-x

Download PDF

Article
Open access
Published: 16 April 2025

Integrating bulk, single-cell, and spatial transcriptomics to identify and functionally validate novel targets to enhance immunotherapy in NSCLC

Kui Cao¹^na1,
Shenshui Wei²^na1,
Tianjiao Ma³^na1,
Xinxin Yang⁴^na1,
Yuning Wang²,
Xiangrong He²,
Mengdi Lu¹,
Yuwen Bai¹,
Cuicui Qi¹,
Luquan Zhang¹,
Lijuan Li¹,
Hongxue Meng⁴,
Jianqun Ma¹ &
…
Jinhong Zhu^2,5

npj Precision Oncology volume 9, Article number: 112 (2025) Cite this article

3411 Accesses
2 Citations
1 Altmetric
Metrics details

Subjects

Abstract

Programmed cell deaths (PCDs) are crucial for tumor progression. By analyzing 18 PCDs, we generated a robust multigene signature, Combined Cell Death Index (CCDI), comprising necroptosis and autophagy genes for non-small cell lung cancer (NSCLC). The CCDI accurately stratified patients by survival prognosis and predicted immunotherapy responses. We validated CCDI and prioritized CCDI genes using five single-cell RNA sequencing and two spatial transcriptomics datasets. CCDI positively correlates with tumor malignancy, invasiveness, and immunotherapy resistance. Four necroptosis genes (PTGES3, MYO6, CCT6A, and CTSH) may affect cancer cell evolution. In vitro, CTSH overexpression or PTGES3 knockdown inhibited NSCLC cell proliferation and migration while inducing necroptosis with necrosome formation. Moreover, we observed diminished CTSH, heightened PTGES3, and low necroptosis activity in 12 pairs of NSCLC tumors and normal tissues. CTSH overexpression or PTGES3 knockdown induced necroptosis and improved anti-PD1 therapy efficiency in syngeneic cancer mouse models. These findings indicate necroptosis genes as potential therapeutic targets in cancer treatments.

Identifying key circulating tumor DNA parameters for predicting clinical outcomes in metastatic non-squamous non-small cell lung cancer after first-line chemoimmunotherapy

Article Open access 10 August 2024

Molecular determinants of response to PD-L1 blockade across tumor types

Article Open access 25 June 2021

Gene expression profiles of the original tumors influence the generation of PDX models of lung squamous cell carcinoma

Article 25 January 2021

Introduction

Non-small cell lung cancer (NSCLC), constituting about 85% of lung cancers, is a major cause of cancer mortality. Its high mortality is largely due to late diagnosis, tumor aggressiveness, and treatment resistance¹. The 5-year survival rate for NSCLC is about 25%, dropping to 5–10% in advanced stages². Immunotherapy is promising in treating NSCLC, yet reliable biomarkers to predict response are lacking³. Effective markers can tailor treatments to individuals, enhancing efficacy and reducing toxicity.

Multigene predictive models surpass single genes in forecasting prognosis and drug responses in cancers⁴. These models could reveal biological effects and prognostic information missed by single-gene analyses. Numerous multigene signature models have been developed over the years, with some already implemented in clinical practice, such as the application of Oncotype DX in breast cancer, a 21-gene assay^5,6. However, NSCLC multigene panels are still in early development.

Programmed cell death (PCD) is a regulated process of controlled self-destruction. PCDs have long been associated with tumor development, metastasis, immunogenicity, and therapy response, driving the development of many prognostic signatures^7,8,9,10. Eighteen forms of PCD have been identified, including intrinsic¹¹ and extrinsic apoptosis¹², necroptosis¹³, cuproptosis¹⁴, oxeiptosis¹⁵, ferroptosis¹⁶, alkaliptosis¹⁷, pyroptosis¹⁸, parthanatos¹⁹, autophagic cell death (ACD)²⁰, lysosome-dependent cell death (LCD)²¹, disulfidptosis²², immunogenic cell death (ICD)²³, entosis¹³, anoikis²⁴, NETosis²⁵, methuosis²⁶, and paraptosis²⁷ (Table 1). We collected PCD-related genes to develop and compare various prognostic models. Ultimately, the model integrating 9 ACD and 11 necroptosis genes demonstrated superior predictive capabilities. Autophagy is a crucial cellular process that degrades and recycles components to support cell survival. Necroptosis is a programmed necrotic cell death²⁸. Autophagy and necroptosis are intricately linked to tumor immunity, affecting tumor development and the immune response²⁹. Thus, we investigated the robustness of CCDI in predicting response to immunotherapy, validated across seven ICI clinical trials in various cancers^{30,31,32,33,34,35,36}.

Table 1 Detailed information for 18 PCD genes

Full size table

Unlike bulk RNA-seq, which averages gene expression from mixed cells, single-cell RNA sequencing (scRNA-seq) detects RNA from individual cells, revealing diverse cell populations in a tumor. Additionally, spatial transcriptomics (ST) integrates gene expression analysis with spatial context, mapping gene profiles to specific tissue locations. We then identified key CCDI genes using scRNA-seq and ST datasets. We analyzed five public scRNA-seq data comparing normal tissue with primary tumors, primary tumors with metastatic tumors, and metastatic tumors with lymph node and brain metastases, as well as immunotherapy responders vs. nonresponders. We tracked cell differentiation and developmental trajectories, dynamically monitoring CCDI gene expression across different cell states and transitions in the tumor microenvironment (TME). Moreover, using ST data, we mapped CCDI gene expression to tissue architecture, visualizing gene expression patterns and cell distribution in situ. We identified CCDI genes CTSH and PTGES3 as key regulators of cell differentiation and fate. Finally, we analyzed the two genes’ functions in vitro and their impact on NSCLC sensitivity to anti-PD-1 antibodies in a mouse model.

Results

Establishing and validating models for 18 PCDs and CCDI

As shown in the graphical summary (Fig. 1a and Supplementary Fig. 1), we first established and evaluated prognostic models for each PCD form using univariable and multivariable Cox regression analyses (Supplementary Fig. 2a–r). To create combined models, we selected five signatures (ACD, necroptosis, LCD, ICD, and oxeiptosis) with AUCs exceeding 0.7 at four or more time points over 5 years, ensuring robust predictive power and stability across long-term dynamic cancer progression intervals (Fig. 1b). The model integrating ACD and necroptosis achieved the highest predictive accuracy (AUC = 0.800) (Fig. 1c), which was termed the combined cell death index (CCDI) for subsequent analysis. CCDI assigned each patient a risk score (range: 0–1) based on gene expression and weights, with a cutoff of >0.34 indicating high risk while ≤0.34 indicating low risk. A higher CCDI risk score is associated with a worse patient prognosis. Henceforth, ‘CCDI’ refers to ‘CCDI risk score,’ and ‘ACD’ and ‘necroptosis’ are used similarly. The time-dependent ROC curves demonstrated that CCDI consistently showed robust performance in predicting survival at 1, 2, 3, 4, and 5 years (Fig. 1d). Kaplan–Meier survival curves for the TCGA-LUAD cohort indicated that CCDI better stratifies survival outcomes than individual ACD and necroptosis signatures (Fig. 1e). UMAP analysis confirmed CCDI’s effectiveness in distinguishing patients with favorable and poor prognoses (Fig. 1f). CCDI contains 9 ACD (RAC1, PRKCD, MBTPS2, EIF4G1, EIF2AK3, CCR2, BNIP3L, ATIC, ATG12) and 11 necroptosis genes (TP53I3, PTGES3, PLEC, PKP2, MYO6, GSK3A, FLNC, CTSH, CD40LG, CCT6A, AURKA) (Supplementary Fig. 3). The CCDI model outperformed ACD and necroptosis models in prognosis prediction, validated in five NSCLC datasets (Supplementary Fig. 4a–c).

**Fig. 1: Establishment and validation of 18 cell death pattern models and CCDI models.**

Predicting responses to ICIs in NSCLC patients using CCDI

Given ACD and necroptosis’s link to tumor immunogenicity and immunotherapy²⁹, we evaluated CCDI’s ability to predict ICI response. We first examined the association between CCDI and various immune-related genes out of 47 immune checkpoints or regulatory factors (Supplementary Fig. 5a, b). The immunophenotype score (IPS) evaluates immunotherapy sensitivity, with higher scores indicating greater sensitivity³⁷. The low-risk group had higher IPS and greater responsiveness to anti-PD-1/L1/L2 therapies than the high-risk group (Fig. 2a). Third, the ImmuCellAI also indicated that the low-risk group might have better ICI responses (Fig. 2b and Supplementary Fig. 5c–g). By adopting three NSCLC ICI trials with known therapy responses as training sets, the SubMap algorithm verified that CCDI could successfully predict ICI response in the TCGA-LUAD cohort (Fig. 2c and Supplementary Fig. 5h–l). Finally, we substantiated the CCDI’s capacity to predict immunotherapy response with seven additional clinical trials with ICIs (Fig. 2d–l). In the IMvigor210 dataset (n = 310, metastatic urothelial carcinoma treated with anti-PD-L1)³⁰, CCDI significantly stratified patients by survival (Fig. 2d–f). Moreover, low-risk patients tended to respond to ICIs than high-risk counterparts across melanoma³¹, renal cell carcinoma³², esophageal adenocarcinoma (EAC)³³, and NSCLC^34,35,36, with lower CCDI in responders than nonresponders (Fig. 2g–l), indicating CCDI’s potential for selecting ICI candidates across cancers.

**Fig. 2: CCDI predicts response to ICI in NSCLC patients.**

Differences in tumor-immune microenvironment between the high and low-risk groups

When applying ESTIMATE in the TCGA-LUAD dataset, we observed higher immune and stromal scores in the low-risk group (Supplementary Fig. 6a, b), while CCDI was negatively correlated with these scores (Supplementary Fig. 6c, d). The MCP counter identified significant differences in NK, B, stromal cells, and neutrophils between the two groups (Supplementary Fig. 6e) and the correlation between CCDI and these cells (Supplementary Fig. 6f). Reduced NK cells in the high-risk group might reflect impaired antitumor immunity. Moreover, CIBERSORT-ABS, EPIC, CIBERSORT, QUANTISEQ, TIMER, and XCELL confirmed different immune cell infiltration patterns between the groups and correlations between risk scores and immune cell subgroups (Supplementary Fig. 6g, h).

Dynamic alterations of CCDI and its genes during lung cancer cell state transition

In a LUAD scRNA-seq dataset by Philip Bischoff et al. (n = 24)³⁸ (Fig. 3a and Supplementary Fig. 7a), CopyKAT, evaluating epithelial cells’ single-cell CNV, was used to distinguish malignant (aneuploid) from non-malignant (diploid) cells (Fig. 3b). Malignant cells were further clustered into C1–C3 subgroups, with the CCDI scores increasing (Fig. 3c). GSVA revealed increased activity in tumor-promoting (EMT, cell cycle, angiogenesis, Wnt) and decreased activity in tumor-suppressive pathways (e.g., necroptosis, apoptosis, and p53) in C3 vs. C1 cells (Fig. 3d). Differential expression analysis revealed upregulation of CCDI genes CCT6A and PTGES3 in C2 and C3, whereas CTSH was upregulated in C1 and downregulated in C2 and C3 (Fig. 3e). Slingshot’s trajectory analysis revealed dynamic changes in PTGES3, CCT6A, CTSH, and MYO6 expression during malignant epithelial cell progression (Fig. 3f, g). Consistent results were observed in another LUAD scRNA-seq dataset (GSE123902, n = 17) (Fig. 3h, i and Supplementary Fig. 7b–g). Therefore, these CCDI genes may regulate lung cancer cell evolution.

**Fig. 3: Relationship between CCDI and tumor malignancy at the single-cell level of LUAD using scRNA-seq datasets.**

Association of lung cancer metastasis with CCDI and constituent genes

The scRNA-seq data (GSE131907) involved biopsies from 44 patients with lung cancer in situ (tLung), lung cancer with metastasis (tL/B) and lymph node metastasis (mLN), and brain metastases (mBrain) (Fig. 4a, b and Supplementary Fig. 8)³⁹. GSVA indicated increased activity in some tumor-related pathways (orange rectangle) and elevated CCDI in tumor cells of the tL/B and mLN groups, while immune responses progressively decreased (Fig. 4c). High-risk patients were more likely to develop lymph node and brain metastasis and progress to stage IV tumors than the low-risk group (Fig. 4d, e), suggesting CCDI’s potential in distinguishing early- and late-stage lung cancer. The tL/B and mLN tumor cells with the highest CCDI were further clustered into C1 to C4 subpopulations (Fig. 4f). CCDI and tumor-related pathway activity (orange rectangle) increased from C1 to C4 (Fig. 4g). Slingshot analysis mapped a pseudotemporal transition from C1 to C4 (Fig. 4h), aligning with CCDI gene expression variations (Fig. 4i). Moreover, another DSP dataset (GSE200563) confirmed increased tumor pathway activity and CCDI in mLN (Fig. 4a, j, k), suggesting a correlation between CCDI and the metastatic transition of cancer cells.

**Fig. 4: The relationship between CCDI and distant metastasis in LUAD using scRNA-seq and DSP datasets.**

Spatial distribution of CCDI and its genes during the LUAD invasion

We analyzed the LUAD scRNA-seq dataset (GSE189357) comprising 88,411 cells from in situ (AIS), minimally invasive (MIA), and invasive adenocarcinoma (IAC), with two samples per tumor state⁴⁰ (Fig. 5a and Supplementary Fig. 9a, b). The same tissues also underwent Visium spatial analysis (GSE189487). GSVA indicated increasing tumor malignancy and CCDI levels from AIS to MIA to IAC (Fig. 5b). CCDI negatively correlated with mast cells, fibroblasts, neutrophils, and DCs (Supplementary Fig. 9c), along with correlations for the four CCDI genes (Supplementary Fig. 9d). Given the importance of cell spatial distribution in cancer malignancy^41,42, we used cell type markers from matched scRNA-seq data to annotate the 17,022 Visium spots, mapping cells to their tissue locations (Fig. 5c). Spatial gene expression analysis showed that as LUAD progressed from AIS to IAC, PTGES3 and CCDI levels increased, while CTSH levels decreased in tumor tissue structures (Fig. 5d). We isolated and analyzed C1–C3 tumor cell subgroups from an IAC sample, and the original study identified C3 as highly malignant⁴⁰. PTGES3 expression increased from C1 to C3 (Fig. 5e, f), suggesting its role in regulating the transition from the less malignant C1 and C2 to the highly malignant C3. Overall, CCDI is closely related to LUAD progression.

**Fig. 5: Spatial distribution of CCDI in cancer areas and its relationship with LUAD progression.**

Dynamics of CCDI in NSCLC treated with PD-1 blockade and chemotherapy

We retrieved scRNA-seq data from 8 LUAD patients undergoing chemotherapy combined with PD-1 blockade, involving 45,408 single cells (GSE207422, n = 8) (Fig. 6a and Supplementary Fig. 10a, b). The GSVA unveiled more active immune-related pathways and lower CCDI in the major pathological response (MPR) than in the non-MPR (NMPR) group, potentially explaining their differential responses to chemoimmunotherapy (Fig. 6b). The ROC curve showed that CCDI effectively assessed chemoimmunotherapy efficacy (Fig. 6c). Differential analysis also reveals MYO6 upregulation in the MPR group and elevated CCT6A in the NMPR group (Fig. 6d). T-cell analysis revealed that the MPR group had more NK cells and less CD4⁺ T-cell exhaustion (Tex) than the NMPR group (Fig. 6e and Supplementary Fig. 10c). CCDI was negatively correlated with NK cells and positively with CD4⁺ Tex cells (Fig. 6f, g). Subgroup analysis of myeloid cells and B cells indicated increased plasma cells and decreased M1 macrophages and cDC1 cells in the MPR group, with CCDI negatively correlated with these cell types (Fig. 6h–l and Supplementary Fig. 10d, e). This suggests that variations in tumor-immune cell infiltration may impact chemoimmunotherapy efficacy, with CCDI potentially predicting this efficacy. Using ‘Copykat’ and ‘inferCNV’, we found that tumor cells have higher CNV than normal lung cells, with CCDI positively correlated with CNV (Fig. 6m–p and Supplementary Fig. 10f, g). This may partly explain the association between CCDI and poor prognosis, likely reflecting increased tumor malignancy and reduced treatment sensitivity.

**Fig. 6: Relationship between CCDI and response to chemotherapy combined with anti-PD-1 therapy at the single-cell level.**

Functional analysis of CTSH and PTGES3

We analyzed PTGES3 and CTSH levels in indicated cell lines (Fig. 7a) and confirmed CTSH overexpression and PTGES3 knockdown in NSCLC cells (Fig. 7b, c). CTSH overexpression or PTGES3 depletion inhibited the proliferation and migration of NSCLC cells (Fig. 7d–g). Such gene manipulations also caused membrane disruption and cell death, reversible by NSA and RIPA-56 (Fig. 7h–k). Moreover, the same genetic manipulations could activate the kinases like Receptor-Interacting Protein Kinase 1 (RIPK1)/RIPK3/pseudokinase Mixed Lineage Kinase Domain-Like protein (MLKL) necroptotic pathway, which was reversed in a concentration-dependent manner by necroptosis inhibitors NSA and RIPA-56 (Fig. 7l–o). Necrosome, indicated by RIPK1-RIPK3 complex assembly, is a hallmark of necroptosis⁴¹. Co-IP results show enhanced necrosome formation in NSCLC cells with CTSH overexpression or PTGES3 knockdown (Fig. 7p, q). Additionally, we detected reduced CTSH and increased PTGES3 in tumors, correlating with decreased RIPK3 and MLKL expression, indicating reduced necroptosis (Fig. 8a–i). Necroptosis has immunogenic potential⁴². In vivo study indicated that Ctsh overexpression and Ptges3 knockdown inhibited tumor growth, induced necroptosis, and enhanced ICI therapy efficacy, as shown by necroptosis makers, tumor growth curves, and weight measurements (Fig. 8j–r). IHC analysis showed that overexpressing CTSH and silencing PTGES3 genes significantly increased CD8⁺ cytotoxic T cells and CD4⁺ helper T cells while reducing FOXP3⁺ regulatory T cells in tumor tissues compared to controls. These effects were further enhanced with PD-1 antibody treatment, suggesting a synergistic boost in antitumor immunity. Thus, regulating Ptges3 and Ctsh could inhibit tumor growth and modulate the TME to boost immunotherapy (Fig. 9a–d).

**Fig. 7: Impacts of CTSH and PTGES3 on NSCLC cell proliferation, migration, and necroptosis.**

**Fig. 8: Clinical validation and in vivo study of CTSH and PTGES3 functions.**

**Fig. 9: Immunohistochemistry (IHC) staining was used to detect immune cells in the tumor.**

Discussion

This CCDI, including necroptosis and ACD genes, remarkably improves patient classification for survival prognoses and ICI therapy predictions, validated through multiple bulk RNA-seq, scRNA-seq, and spatial transcriptomics datasets. Key CCDI genes, MYO6, CCT6A, CTSH, and PTGES3, are related to tumor cell differentiation and fate. Furthermore, CTSH and PTGES3 regulate necroptosis and boost the efficacy of anti-PD-1 therapy when manipulated to induce necroptosis.

Recent efforts have developed predictive models using single PCD signature genes, but their success in predicting cancer prognosis and drug response remains moderate^{9,43,44,45,46,47}. A single PCD type-based model is inadequate to capture tumor features due to the redundancy and crosstalk in cell death pathways⁴⁸. Therefore, predicting models incorporating multiple PCD types have merged lately^49,50,51,52. The combined multi-cell death forms model offers a comprehensive view of disease complexity and adapts better to patient heterogeneity. The current CCDI with an AUC of 0.8 surpassed many existing LUAD prognostic signatures, such as the aneuploidy-based gene signature⁵³, vitamin metabolism clustering-related gene signature⁵⁴, glycosyltransferase gene signature⁵⁵, immune-related lncRNA pair signature⁵⁶, m⁶A RNA methylation regulators⁵⁷, and signature derived from diverse cell death patterns⁵⁸.

Despite the promising performance of ICIs, effective biomarkers for patient selection remain lacking. Biomarkers, such as PD-L1 expression, microsatellite status (MSI-H), and tumor mutational burden (TMB), face limitations and lack uniform testing protocols⁵⁹. Given the selective efficacy and frequent serious toxicities⁶⁰, accurate biomarkers are crucial for selecting suitable patients for ICIs. The CCDI’s ability to predict responses to ICIs was validated via seven ICI cancer trials, including NSCLC, EAC, melanoma, urothelial carcinoma, and renal cell carcinoma^{30,31,32,33,34,35,36}. CCDI shows great potential for survival prediction and personalized ICI therapies.

The robustness of CCDI is further underscored by its validation across multiple high-dimensional datasets, including scRNA-seq and spatial transcriptomics. scRNA-seq provides precise insights by analyzing individual cell gene expressions and leading to a more accurate understanding of biological processes and disease mechanisms^61,62,63. We included four LUAD scRNA-seq datasets to validate CCDI involving normal tissue, non-metastatic and metastatic primary tumors, corresponding lymph node and brain metastases, tumors at different invasion and metastasis stages, and tumors undergoing immunotherapy^38,39,64. Consistently, we found that CCDI increased as cancer cells progressed into more malignant phenotypes. scRNA-seq analysis showed elevated CCDI levels in tumor cells from metastatic lung cancer and its metastases (lymph nodes and brain) compared to non-metastatic lung cancer. Moreover, cancer cells responsive to ICIs exhibit lower CCDI than their resistant counterparts. Further analysis of immune cell subsets revealed that CCDI was inversely related to NK cells, M1 macrophages, cDC1, and plasma cells but directly related to CD4⁺ Tex cells, implying a potential link between CCDI and an immunosuppressive TME. Using scRNA-seq data, we validated that CCDI can predict tumor cell malignancy and responsiveness to ICI therapy.

Unlike scRNA-seq, spatial transcriptomics retains the tissue architecture and permits the analysis of how gene expression varies across different tumor regions, enabling the study of intratumoral heterogeneity in situ. Using two spatial transcriptomics datasets with tumor tissues at different stages of invasion and metastasis^40,65, we confirmed that CCDI increases with tumor malignancy and invasiveness in the spatial structures. Overall, validation across bulk, scRNA-seq, and spatial transcriptomics enhances the CCDI’s reliability in predicting patient outcomes and guiding therapeutic decisions.

To comprehend the discrepant ICI responses among the two risk groups, we compared the TIME of the two groups and observed differences in IPS, immune scores, stromal scores, and immune cell infiltration patterns. These parameters are all known to be associated with immunotherapy efficacy^37,50,66. The CCDI comprised genes related to necroptosis and autophagy, suggesting that differences in necroptosis and autophagy activities may influence TIME. Including necroptosis and autophagy-related genes in the CCDI underscores the complex nature of tumor cell death mechanisms and their combined effect on cancer progression and treatment response. Necroptosis and autophagy/ACD are crucial in cancer biology and immunogenicity. ACD can occur under severe nutrient deprivation or specific cancer treatments. Autophagy may boost immune activation by aiding in the processing and presentation of tumor antigens, helping the immune system recognize and target tumor cells^67,68. Conversely, in specific contexts, autophagy can suppress immune responses by regulating inflammation or inhibiting the production of specific cytokines. Mitophagy removes damaged mitochondria, preventing the release of cytosolic mitochondrial DNA that can activate the cGAS-STING pathway and IFN-I secretion, thereby enhancing T-cell cross-priming and boosting the immune response⁶⁹. Necroptosis is triggered by various stimuli, including death receptor activation and pattern recognition receptors (PRRs)⁷⁰, featured by necrosome form from RIPK1 and RIPK3 interaction and cell membrane rupture. The resulting release of intracellular contents, including DAMPs (e.g., HMGB1, ATP)^71,72, which boost antigen uptake by phagocytes and trigger dendritic cells (DC) maturation, causing efficient CD8⁺ T-cell cross-priming and enhanced antitumor immunity⁷².

During tumor progression, cancer cells often advance to more malignant states due to genetic and epigenetic changes, microenvironmental influences, and clonal selection^73,74,75. It is indispensable to understand the mechanisms underpinning the heterogeneity and plasticity of cancer cells. scRNA-seq can identify rare cell populations, elucidate cellular interactions, and track dynamic gene expression changes⁷⁶. Trajectory analysis is crucial for identifying genes driving critical cell state transitions, providing key insights into tumor progression and heterogeneity⁷⁷. We identified that four necroptosis-related CCDI genes (PTGES3, MYO6, CCT6A, and CTSH) exhibited dynamic expression changes across different tumor cell subpopulations. This strategy aids in deciphering tumor evolution and discovering new biomarkers and potential therapeutic targets related to PCDs. Similarly, we previously found that necroptosis genes HOOK and CUL4A changed with ESCC cell state transitions and verified them as independent predictors of ESCC prognosis⁵⁰. Using single-cell and functional analyses, Kong et al. revealed that PLK1 drove an immunosuppressive TME in LUAD⁷⁸. Furthermore, validation with ST datasets strengthens the findings by preserving the spatial context of gene expression.

ST data show that CTSH and PTGES3 vary remarkably across tumor progression and invasion stages. In vitro analyses demonstrated that CTSH overexpression or PTGES3 knockdown inhibited the proliferation and migration of NSCLC cells while inducing necroptosis. Prostaglandin E Synthase 3 (PTGES3) is an enzyme converting prostaglandin endoperoxide H2 (PGH2) into prostaglandin E2 (PGE2). Bioinformatics analysis demonstrates elevated PTGES3 levels in various tumors^79,80,81, with its overexpression in LUAD linked to poor prognosis and increased immune cell infiltration⁷⁹. Drugs like gedunin, genistein, and diethylstilbestrol could target PTGES3 to inhibit breast cancer cell proliferation⁸⁰. In acute respiratory distress syndrome, the PTGES3/p23 could activate RIPK3/MLKL, driving endothelial cell necroptosis⁸². Cathepsin H (CTSH) is a lysosomal cysteine proteinase essential for breaking down lysosomal proteins. Our results align with previous studies showing decreased CTSH levels in lung tumors compared to non-tumorous lung tissue, suggesting CTSH may act as a tumor suppressor in LUAD^83,84. We also found diminished CTSH and heightened PTGES3 in tumors, along with decreased RIPK3 and MLKL expression, implying lower necroptosis activity. Moreover, the present study demonstrated that altering CTSH and PTGES3 to trigger necroptosis significantly improves the efficacy of anti-PD-1 therapy and enhances antitumor immunnity. ICI’s effectiveness is restricted by primary and acquired resistance, originating from genetic, epigenetic, and metabolic changes that hinder T-cell movement to the tumor, disrupt immune cell function, and impair antigen presentation⁸⁵. ICI therapy’s success heavily depends on the presence of intratumoral CD8⁺ T cells. However, ICI alone often fails to significantly increase CD8⁺ TIL numbers⁸⁶. Inflammatory PCDs (e.g., pyroptosis and necroptosis) release DAMPs and cytokines, promoting DCs to engulf dying cells, attracting and activating CD8⁺ T cells. Therefore, ICIs combined with therapies inducing inflammatory cell death may work synergistically^87,88. Necroptosis influences the TME by modulating immune cell infiltration and activity, consequently enhancing antitumor immune, potentially making cancer cells more susceptible to immunotherapy^42,89. Overall, the induction of necroptosis may be a promising strategy to enhance anti-PD-1 efficacy.

The limitations of the study should be addressed. First, since many publicly stored high-throughput datasets are not accessible, we included clinical trials with heterogeneous treatments, such as immunotherapy alone, chemotherapy combined with immunotherapy, and neoadjuvant chemoimmunotherapy. Second, although our results demonstrate that CCDI shows clinical significance across these datasets, this study is exploratory. Intensive validations are required before its clinical translation, especially in large, well-designed prospective studies. Third, while we acknowledge the significance of known predictors of ICI response, such as PD-L1 expression and the presence of oncogenic drivers (e.g., EGFR and ALK alterations), we chose not to correct the CCDI model for these factors but consider CCDI as a complementary and independent model. Finally, the CCDI model was constructed using bulk RNA sequencing data. We further validate the clinical relevance of CCDI in tumor subpopulations using publicly available scRNA-sequencing datasets. However, accurately quantifying the proportion of CCDI gene expression from non-tumor cells in TME is currently unfeasible. Although this study provides critical insights into tumor-immune interactions, it is inherently limited by its preclinical focus. Future work will explore CCDI’s clinical utility through multicenter cohort studies and integrative analyses with established biomarkers, such as PD-L1 and oncogenic driver status.

In conclusion, CCDI is a comprehensive tool for prognostic and ICI therapeutic prediction. Identifying CTSH and PTEGS3 as key factors in tumor progression and therapy response underscores their potential as novel therapeutic targets, paving the way for more personalized and effective treatments and improving cancer prognosis.

Methods

Data source

Twenty RNA sequencing (RNA-seq) datasets and clinical information were retrieved from several public big data repositories, including the CODE OCEAN (https://codeocean.com/), the Gene Expression Omnibus (GEO) (https://www.ncbi.nlm.nih.gov/geo), the Cancer Genome Atlas (TCGA) (https://www.cancer.gov/tcga/), and the International Cancer Genome Consortium (ICGC) (https://dcc.icgc.org/). Thirteen of them are bulk RNA-seq datasets: TCGA-LUAD (n = 569), ICGC-LUAD (n = 425), GSE31210 (n = 226), GSE72094 (n = 442), GSE50081 (n = 181), GSE37745 (n = 196), GSE78220 (n = 28), GSE67501 (n = 11), GSE190265 (n = 34), GSE126044 (n = 16), GSE135222 (n = 27), GSE165252 (n = 77), and the IMvigor210 cohort (n = 310)³⁰ (Table 2). Additionally, five scRNA-seq datasets were included: GSE123902 (n = 17), GSE189357 (n = 6), GSE131907 (n = 58), GSE207422 (n = 8), and Philip Bischoff’s study (n = 24)³⁸. Furthermore, the study incorporated one digital spatial profiling (DSP) dataset (GSE200563, n = 109) and a Visium spatial transcriptomic (ST) dataset (GSE189487, n = 6) (Table 2).

Table 2 Detailed information on datasets included in this study

Full size table

The signature gene sets and sources for 18 PCD are detailed in Supplementary Table 1, including intrinsic apoptosis (n = 500), extrinsic apoptosis (n = 500), disulfidptosis (n = 10), necroptosis (n = 500), cuproptosis (n = 17), oxeiptosis (n = 146), ferroptosis (n = 283), alkaliptosis (n = 73), pyroptosis (n = 387), parthanatos (n = 23), ACD (n = 232), LCD (n = 194), ICD (n = 500), entosis (n = 23), anoikis (n = 338), NETosis (n = 24), methuosis (n = 8), and paraptosis (n = 23).

Construction and validation of the CCDI model

TCGA-LUAD dataset served as the training set for developing and optimizing the predictive models. Initially, univariate Cox analysis identified significant prognostic genes from PCD signature genes based on overall survival (OS). Multivariate Cox regression then constructed prognostic models for each type of PCD using these gene coefficients by ‘survminer’ (v.0.4.9) and ‘survival’ (v.3.5-8) R packages.

To construct the multivariate Cox regression model, we used the ‘stepAIC’ function from the ‘MASS’ (v.7.3-60.2) R package with ‘bidirectional’ stepwise selection and Akaike Information Criterion (AIC) minimization to determine the optimal model. This approach ensures the robustness and reliability of the prognostic model⁹⁰. We first constructed and compared prognostic models from individual PCDs and assessed their discriminating accuracy using the Receiver Operating Characteristic (ROC) approach. Both ROC (‘pROC’ R package, v.1.18.5) and time-dependent ROC (‘survivalROC’ R package, v.1.0.3.1) curves at 1, 2, 3, 4, and 5 years were plotted, calculating the Area Under the Curve (AUC). The time-dependent AUCs evaluated each model’s predictive power over different time horizons, providing insight into the models’ effectiveness in distinguishing survival status at specific time points. Higher AUC values indicate better performance. A threshold of AUC > 0.7 was adopted. We considered PCD models with AUC > 0.7 across four or five intervals to be stable and effective. The resulting five models with consistently high performance were used further to build combined models. Next, a series of combination models, comprising 2–5 multigene signatures, were constructed using logistic regression. Among these, one model, integrating ACD and necroptosis, achieved the lowest AIC value and the highest AUC, thus being selected for further investigation and named after CCDI⁹⁰. The CCDI risk score was calculated using the formula: Risk score \({\rm{CCDI}}=\frac{\exp ({{\rm{\beta }}}_{0}+{{\rm{\beta }}}_{1}{{\rm{\chi }}}_{{ACD}}+{{\rm{\beta }}}_{2}{{\rm{\chi }}}_{{necroptosis}})}{1+\exp ({{\rm{\beta }}}_{0}+{{\rm{\beta }}}_{1}{{\rm{\chi }}}_{{ACD}}+{{\rm{\beta }}}_{2}{{\rm{\chi }}}_{{necroptosis}})}\) for each LUAD patient. In this context, \({{\rm{\beta }}}_{0}\) acts as the intercept, with \({{\rm{\beta }}}_{1}\) and \({{\rm{\beta }}}_{2}\) as the logistic regression coefficients. The risk scores, χ_ACD and χ_necroptosis, are derived from “\({{\rm{\chi }}}_{{\rm{ACD}}}={\sum }_{i=1}^{n}{\beta }_{i}{x}_{i}\)” and “\({{\rm{\chi }}}_{{\rm{Necroptosis}}}={\sum }_{j=1}^{n}{\beta }_{j}{x}_{j}\)”, respectively. The minimum P-value method was applied to the Kaplan–Meier curve to determine the CCDI score’s cutoff value, which stratified LUAD patients into high and low-risk groups. The ‘umap’ (v.0.2.10.0) R package was adopted to visualize the distribution disparities in risk scores between high- and low-risk groups as defined by the ACD, necroptosis, and CCDI models. A 1000-fold bootstrap validation was conducted internally for the 18 PCD models and CCDI. Additionally, external validations were performed across five independent datasets (ICGC-LUAD, GSE31210, GSE72094, GSE50081, and GSE37745).

Assessing the ability of the CCDI model’ ability to predict ICI responses

Several approaches were used to verify CCDI’s capacity to predict immune checkpoint inhibitor (ICI) responses. First, a total of reported 47 immune checkpoint (ICP) genes were used⁹¹. Second, Immunophenoscore (IPS) is a computational metric generated to evaluate tumor immunogenicity by integrating effector cells, immune suppressor cells, major histocompatibility complex (MHC) molecules (antigen processing), and checkpoint/immune modulators. IPS was developed to quantify the capability of a tumor to elicit an immune response and predict the potential effectiveness of ICIs. The IPS scale spans from 0 to 10, indicating a range from minimal to maximal immunogenicity. IPS for LUAD patients was retrieved from The Cancer Immunome Atlas (TCIA) (https://tcia.at/home)³⁷. Third, the Immune Cell Abundance Identifier (ImmuCellAI, http://bioinfo.life.hust.edu.cn/web/ImmuCellAI/) is a web-based computational tool that estimates the abundance of 24 immune cell types in bulk RNA sequencing (RNA-seq) and microarray data. ImmuCellAI predicts patient responses to ICIs by assuming ICI therapy effectiveness is influenced by specific immune cell types⁹². Fourth, the Subclass Mapping (SubMap) algorithm compares gene expression profiles from tumor samples with those of patients who have responded to ICIs. By identifying overlaps between new patient cohorts and known responders, SubMap predicts potential responders to similar therapies⁹³. The submap algorithm was used to compare gene expression profiles between individual training datasets and the validation datasets, including TCGA-LUAD, ICGC-LUAD, GSE31210, GSE72094, GSE50081, and GSE37745, to find similarities or “consensus” between them. This evaluation aimed to determine the accuracy of CCDI in predicting responses to ICIs. Three independent clinical trials involving ICIs for the treatment of NSCLC were used as training sets, including GSE126044 (n = 16)³⁴, GSE135222 (n = 27)³⁵, and GSE190265 (n = 34)³⁶.

Finally, we applied CCDI to seven independent immunotherapy trials to assess differences in immunotherapy responses between high- and low-risk groups, including the IMvigor210 dataset (n = 310, metastatic urothelial carcinoma, anti-PD-L1 therapy)³⁰, GSE78220 (n = 28, melanoma, anti-PD-1 therapy)³¹, GSE67501 (n = 11, renal cell carcinoma, anti-PD-1 therapy)³², GSE165252 (n = 77, esophageal adenocarcinoma, anti-PD-L1 therapy)³³, GSE126044 (n = 16, NSCLC,anti-PD-1 therapy)³⁴, GSE135222 (n = 27, NSCLC, anti-PD-1/PD-L1 therapy)³⁵, and GSE190265 (n = 34, NSCLC, chemotherapy and anti-PD-1/PD-L1 therapy)³⁶.

scRNA‑seq data and ST data processing

The raw gene expression matrix obtained from scRNA-seq was transformed into a Seurat object using the ‘Seurat’ (v.4.4.0) R package⁹⁴. The ‘CreateSeuratObject’ function from the package generated a Seurat object by taking the gene expression matrix as input and organizing it into a structured format suitable for downstream analysis. The Seurat object allows for the efficient handling of scRNA-seq data, enabling advanced functionalities such as quality control, normalization, clustering, and visualization. Cells exhibiting less than 500 genes per cell, more than 2500 genes per cell, or a mitochondrial gene content exceeding 20% were classified as low quality and removed from the analysis. Batch effects were addressed through the application of the ‘RunHarmony’ function from the ‘harmony’ (v.1.2.1) R package⁹⁵. Dimensionality reduction was undertaken using ‘Seurat’ R package’s ‘RunPCA’, ‘UMAP’, and ‘TSNE’ functions. The “FindClusters” function was used to perform unsupervised clustering (Louvain algorithm), which automatically identifies clusters by optimizing modularity values based on the cell neighborhood network. The final subgroups were determined by integrating dimensionality reduction visualization, stability evaluation, and the expression of specific marker genes. In cluster annotation, reference data for cell annotation were retrieved and integrated from CellMarker (http://xteam.xbio.top/CellMarker/), SingleR (v.2.6.0), and published literature^96,97.

Zhu et al. performed both scRNA-seq and Visium spatial transcriptomic analysis on the same LUAD tissues, and the resulting data were deposited as a scRNA-seq dataset (GSE189357) and a Visium ST dataset (GSE189487)⁴⁰. The ST datasets were imported into Seurat using the ‘Load10X_Spatial’ function. Batch effect correction was implemented utilizing the ‘SCTransform’ function within the ‘Seurat’ R package. Subsequent steps in dimensionality reduction followed a processing workflow akin to that applied to scRNA-seq data. We analyzed and integrated scRNA-seq and spatial transcriptomic data of LUAD of these public LUAD datasets by comparing gene expression profiles from single cells with those in specific spatial regions to infer cell type locations within the tissue. For the cellular-type deconvolution analysis of spatial transcriptomic data, an analysis pipeline based on Robust cellular-type deconvolution (RCTD) was employed to analyze reference data and Visium spatial transcriptomic data⁹⁸. The doublet mode was set to full mode. Then, we mapped scRNA-seq data back to the original tissue context, visualizing the expression and distribution patterns of targeted genes and CCDI. The CCDI score for each cell or spot was calculated by extracting the expression matrix of the model-related genes and applying it to the risk formula of the model.

Cell trajectory prediction

A trajectory analysis of scRNA‑seq data was conducted using “slingshot” (v.2.12.0) R package to delineate the developmental trajectory of cells and determine the lineage differentiation of distinct cell populations⁹⁹. For the ST data, cellular evolutionary trajectories were deduced using the R package ‘Monocle’ (v.2.32.0). This input data was a gene-cell matrix at the UMI count scale, derived from the Seurat subset, which was processed with default parameters after downscaling and cell sorting¹⁰⁰.

Discrimination of malignant and non-malignant epithelial cells

We identified malignant and non-malignant epithelial cells by using the “copykat” (v.1.1.0)¹⁰¹ and “infercnv” (v.1.20.0)¹⁰² R packages. Specifically, “copycat” was run with default settings to identify polyploid and diploid cells. In “infercnv,” cells from the TME were used as normal reference cells according to the “Using 10× data” guidelines, and their copy number variation (CNV) estimates were used to set the baseline. A gene sorting file was generated from the human CRCh38 assembly, incorporating the chromosomal start and end positions for each gene. In the CNV inference analysis, the minimum average read count for each gene in the reference cells was set to 0.1, and the output results underwent denoising to ensure the accurate prediction of CNV levels.

Identifying differentially expressed genes

The Seurat “FindAllMarkers” function, identifying markers of clusters in scRNA-seq data, was employed to detect differentially expressed genes (DEGs) between different groups of cells by clustering algorithms. Wilcoxon rank-sum tests were used to determine the statistical significance. Genes with an absolute log2 fold change (log2FC) ≥ 0.58 and a false discovery rate (FDR) < 0.05 were considered differential within the cellular subgroups by default.

Gene set variation analysis (GSVA)

Gene enrichment analysis was performed using curated gene sets from the Molecular Signatures Database (MSigDB)¹⁰³. GSVA was implemented through the “GSVA” (v.1.52.3) R package, utilizing gene expression profiles to assess the activity scores for each cell type or sample¹⁰⁴. The differences in activity scores were quantified to elucidate variations among different immune cell subtypes, signaling pathways, biological states, and processes across distinct cell clusters or samples.

Evaluation of tumor-immune microenvironment

The “ESTIMATE” (v.1.0.13) R package (Estimation of STromal and Immune cells in MAlignant Tumor tissues using Expression data) was utilized to evaluate differences in stromal scores and immune scores between high and low-risk groups¹⁰³. The MCP counter method¹⁰⁵ and CIBERSORT¹⁰⁵ analysis were used to assess immune cell infiltration differences between high- and low-risk groups. These differences were further explored using five methods from the TIMER 2.0 website, including TIMER¹⁰⁶, XCell¹⁰⁷, EPIC¹⁰⁸, QUANTISEQ¹⁰⁹, and CIBERSORT-ABS¹⁰. The comparison between the high-risk and low-risk groups was conducted using the Wilcoxon test to evaluate significant differences in immune scores, stromal scores, and immune cell infiltration levels between the two groups. The pie chart was drawn based on the relative proportions of each cell type in the high-risk and low-risk groups. The denominator was defined as the total infiltration value of the specific cell type across both groups (i.e., the sum of the infiltration values in the high-risk and low-risk groups), and the relative proportions for the high-risk and low-risk groups were then calculated accordingly. Pearson correlation analysis was conducted to assess the correlations between risk scores and immune scores, as well as their association with immune-infiltrating cells.

Cell culture and reagents

Human lung bronchial epithelial cells (BEAS-2B, cat# CBP60577) were obtained from COBIOER Biosciences Co., Ltd. (Nanjing, China). NSCLC cell lines, including A549 (cat# SCSP-503), PC-9 (cat# SCSP-5085), HCC827 (cat# SCSP-538), NCI-H1299 (cat# SCSP-589), and H1650 (cat# SCSP-592), were sourced from the Cell Bank of the Chinese Academy of Science (Shanghai, China). Murine Lewis lung Carcinoma 1 (LLC1/CL-0140) cells were acquired from Procell Life Science & Technology Co., Ltd. (Wuhan, China). BEAS-2B and LLC1 cells were cultured in DMEM medium (PM150210, Procell, China), while other cell lines in RPMI-1640 medium (PM150110, Procell, China), both supplemented with 10% fetal bovine serum (164210, Procell, China) and 1% penicillin-streptomycin solution (PB180120, Procell, China). Cells were incubated at 37 °C in 5% CO₂ with humidity. RIPA-56 (HY-101032, MedChemExpress, Monmouth Junction, NJ, USA), a RIPK3 inhibitor, and necrosulfonamide (NSA) (HY-100573, MedChemExpress, Monmouth Junction, NJ, USA), an MLKL inhibitor, were used to treat A549 and H1299 cells for 24 h to inhibit necroptosis.

Lentivirus infection and plasmid transfection

All the lentiviral vectors and plasmids were obtained from GeneChem Biotechnology Co., Ltd. (Shanghai, China). Regarding human gene manipulation, the CTSH plasmids were used to express CTSH ectopically. The PTGES3 knockdown plasmid, constructed with three unique shRNA sequences: 9893-1 (5′-CGAAGGGACTATGTCTTCATT-3′), 9894-2 (5′-TGTCTCGGAGGAAGTGATAAT-3′), and 9895-1 (5′-GAAGACAGTAAGGATGTTAAT-3′), and the shControl (shCtrl) (CMV-enhancer-MCS-3FLAG-SV40-Puromycin) were used for the loss-of-function study. When the cells reached 70% to 80% confluence, the plasmids (50 nM) were transfected into the human NSCLC cells in 6-well plates using jetPRIME® transfection reagent (101000027, Polyplus Transfection Inc., New York, NY, USA). After 24 h, the expression levels of CTSH and PTGES3 were measured by qRT-PCR and Western blot analysis.

Lentiviral vectors containing the full-length cDNA of mouse Ctsh and an empty vector (Ubi-MCS-3FLAG-SV40-puromycin) were used to overexpress Ctsh in LLC1 cells. We also obtained the shControl (shCtrl) (hU6-MCS-CMV-puromycin) and lentiviral vectors carrying three different shRNA sequences to target the mouse Ptges3 gene: 132643-1 (5′-CGTTTCTCTGAGATGATGGAT-3′), 132644-2 (5′-GCTGTAATTTCATGTATTTAC-3′), and 132645-1 (5′-TACTGGAATGGAGTTTGGTTT-3′). LLC1 cells were cultured in 6-well plates. When cells reached 70% confluence, each well received 8 µl of lentivirus carrying either the Ctsh cDNA or shPtges3 sequences, followed by a 24-h incubation. Stable cell lines were established by adding a puromycin-containing medium (LabGeni, Beijing, China). Protein expression levels of CTSH and PTGES3 in LLC1 cells were then assessed by Western blot analysis.

Quantitative reverse transcription PCR (qRT-PCR)

Total RNA was extracted using TRIzol reagent (Thermo Fisher Scientific, Waltham, MA, USA), and its concentration was determined with a NanoDrop1000 spectrophotometer (NanoDrop, Madison, WI, USA). One μg of RNA was reverse-transcribed into cDNA using the SureScript™ First-Strand cDNA Synthesis Kit (Genecopoeia, Guangzhou, China). qPCR was conducted in a 20 μL reaction volume on a qTOWER 3 Real-Time PCR Thermal Cycler (Analytik, Jena, Germany) containing 10 μL of 2× BeyoFast™ SYBR Green qPCR Mix (Beyotime, Shanghai, China), 2 μL of primers, 2 μL of cDNA templates, and 6 μL of ddH₂O. The primers used in this study were as follows: CTSH (Forward primer: 5′-ACTGGTCCCTACCCACCTTC-3′, Reverse primer: 5′-ACAACTGAGGCTGCAAGGAC-3′), PTGES3 (Forward primer: 5′-ACCGGAGAGAAAAAGCGGAG-3′, Reverse primer: 5′-AGAAGCAGGCTGCATTGTGA-3′), RIPK3 (Forward primer: 5′-AAAATGCCCGAGCCTTACCA-3′, Reverse primer: 5′-TGTCTCTCAGCCCCCTGATT-3′), MLKL (Forward primer: 5′-TTTCCCTCAGGAGGCGGT-3′, Reverse primer: 5′-TAACTTCTCAGAGGGCACGC-3′), and GAPDH (Forward primer: 5′-ACAGCCTCAAGATCATCAGC-3′, Reverse primer: 5′-GGTCATGAGTCCTTCCACGAT-3′). Gene expression levels in cells were normalized to GAPDH. The relative gene expression levels were calculated using the formula 2^(-ΔΔCt).

Western blot analysis

Cells were lysed with RIPA buffer (P0013B, Beyotime, China) on ice for 20 min. After centrifugation, the supernatant was collected, and protein concentration was determined using a BCA Protein Assay Kit (P0012S, Beyotime, China). We standardized protein concentration by adding the appropriate amount of loading buffer (P0015, Beyotime, China) to each sample. Equal volumes of protein samples were separated by SDS-PAGE (PG112, Epizyme Biotech, Shanghai, China) and transferred to PVDF membranes (ISEQ00010, Millipore, USA). After blocking, the membranes were incubated with the following primary antibodies under specified conditions: CTSH (1:1000, Cat#10315-1-AP, RRID: AB_2087534, Proteintech, Wuhan, China), PTGES3 (1:1000, Cat#15216-1-AP, RRID: AB_2284772, Proteintech, Wuhan, China), MLKL (1:1000, Cat.#: 380559, Zen Biotechnology Co., Ltd., Chengdu, China), Phospho-MLKL (Ser358) (1:1250, Cat# 382136, Zen Biotechnology Co., Ltd., Chengdu, China), RIPK3 (1:1000, Cat# AF4808, RRID: AB_2844793, Affinity Biosciences, China), Phospho-RIPK3 (1:5000, Cat# 93654, RRID: AB_2800206, Cell Signaling Technology, China), RIPK1 (1:1000, Cat.# 3493, RRID: AB_2305314, Cell Signaling Technology, China), and GAPDH (1:2000, Cat.# 60004-1-Ig, RRID: AB_2107436, Proteintech, Wuhan, China). The following secondary antibodies were used to amplify the signal of the target protein: HRP-conjugated Affinipure Goat Anti-Mouse IgG (H+L) (1:2000, Cat#SA00001-1, RRID: AB_2722565) and HRP-conjugated Affinipure Goat Anti-Rabbit IgG (H+L) (1:2000, Cat#SA00001-2, RRID: AB_2722564, Proteintech, China).

Co-immunoprecipitation (Co-IP)

Cells were lysed on ice using IP lysis buffer (P0013, Beyotime, China) supplemented with PMSF protease inhibitor (ST505, Beyotime, China). Lysates were spun at 13,500 rpm for 5 min to obtain supernatants for the next stage; meanwhile, magnetic agarose beads coated with protein A/G were allowed to conjugate with anti-RIPK1 antibody (Cat# 3493, RRID: AB_2305314, Cell Signaling Technology, China) or IgG (Cat# AC005, ABclonal, Wuhan, China). The resulting complexes were incubated with supernatants overnight at 4 °C. Next day, the immunoprecipitants were washed thoroughly three times, eluted with 1× loading buffer (P0015A, Beyotime, China), and boiled at 99 °C for 5 min. Magnetic beads were removed using a magnetic stand, and the samples were subjected to Western blot analysis. The detection of target proteins was performed using antibodies specifically raised against RIPK1 (Cat# 3493, RRID: AB_2305314, Cell Signaling Technology, China), RIPK3 (Cat# AF4808, RRID: AB_2844793, Affinity Biosciences, China), MLKL (Cat# 380559, Zen Biotechnology Co., Ltd., Chengdu, China), or GAPDH (Cat# 60004-1-Ig, RRID: AB_2107436, Proteintech, Wuhan, China).

Proliferation and migration assay

The Dojindo Cell Counting Kit-8 (CCK-8, GK10001, GlpBio, USA) assay was used to assess cell proliferation, while wound healing and Transwell assays were adopted to evaluate cell migration capabilities. For the CCK-8 assay, 5000 cells were introduced into each well of a 96-well plate (3599, Corning, Tewksbury, MA, USA). Once the cells had fully adhered, cell viability was measured at 0, 24, 48, and 72 h using the CCK-8 assay. Moreover, cells were plated into 6-well plates for the wound healing assay. When the cells reached 95% confluence, a scratch was made to create a wound in the cell monolayer using a pipette tip. The cells were maintained in RPMI-1640 medium (PM150110, Procell, China). Images were captured at 0 and 24 h to measure the scratch gap. During the Transwell migration assay, cells at a density of 3 × 10⁴ per well in serum-free suspension were introduced into the upper chambers of 8 μm Transwell inserts (3422, Corning, Tewksbury, MA, USA). Medium with 10% FBS (164210, Procell, Wuhan, China) was added to the lower chambers. After 24 h, migrated cells were fixed with 4% paraformaldehyde, stained with 0.1% crystal violet, and counted using a microscope.

LDH release assay

Cells were seeded into a 96-well cell culture plate (3599, Corning, Tewksbury, MA, USA) at an appropriate density. Upon cells reaching 80% confluence, lactate dehydrogenase (LDH) levels of the culture medium were measured using an LDH Cytotoxicity Assay Kit (C0016, Beyotime, China) in accordance with the manufacturer’s protocol.

In vivo treatments

Female C57BL/6 mice aged between four and six weeks were purchased from Huachuang Sino (Jiangsu, China). Housing and handling procedures followed protocols approved by the Institutional Review Board at Harbin Medical University. A total of 1 × 10⁶ murine LLC1 cells with Ctsh overexpression or Ptges3 knockdown and control cells were subcutaneously implanted into the right flank of mice. Mouse anti-PD-1 antibody (10 mg/kg, BE0146, BioXcell, Lebanon, NH, USA) or IgG (A7050, Beyotime, China) was intraperitoneally administrated three times a week, beginning on the seventh day after cell implantation. Tumor dimensions were measured with calipers, and the volume was determined using the formula: 0.5 × length × width × height. At the end of the experiment, mice were euthanized using CO₂ inhalation followed by cervical dislocation to ensure humane sacrifice in accordance with institutional animal care guidelines.

All research procedures were approved by the Ethics Committee of Harbin Medical University (No. HMUIRB2024009). Written informed consent was obtained from each patient prior to sample collection. Research involving humans has been performed in accordance with the Declaration of Helsinki.

Immunohistochemistry staining (IHC)

The synegenic mouse tumor specimens were fixed, embedded in paraffin, and sectioned into 4 µm thick slices using a Leica microtome. After standard deparaffinization, antigen retrieval, and blocking of nonspecific binding sites, the tissue sections were incubated overnight at 4 °C with primary antibodies against CK (Cat# 10164-2-AP, RRID: AB_2134679, Proteintech), NK (Cat# AG2387, Beyotime), CD11B (Cat# AF6396, Beyotime), CD3 (Cat# 17617-1-AP, RRID: AB_1939430, Proteintech), CD4 (Cat# AG1393, Beyotime), CD8 (Cat# AG1411, Beyotime), and FOXP3 (Cat# 22228-1-AP, RRID: AB_11182376, Proteintech). Next day, HRP-conjugated secondary antibodies (Cat# RGAR011, RRID: AB_3094534, Proteintech) were applied for detection. DAB chromogen (50:1, Novus Biologicals, Centennial, CO, USA) was used for the visualization, followed by hematoxylin counterstaining for nuclear contrast. Sections without the primary antibody were used as negative controls. The stained tumor tissue sections were imaged under a high-resolution microscope, and quantitative analysis was performed using ImageJ (version 1.43) and QuPath (version 0.5.1) software. Representative tumor areas were selected for analysis. The percentage of positive cells was then calculated relative to the total cell count.

Statistics

All statistical analyses were performed using R (version 4.4.0, specific R packages are mentioned in the “Methods” section) and GraphPad Prism (version 9.0). Continuous variables between two groups with equal variances and normal distribution were analyzed using the Student’s t-test; otherwise, the Wilcoxon rank-sum test was used. For comparisons among three or more groups with equal variances and normal distribution, one-way analysis of variance (ANOVA) was used; otherwise, the Kruskal–Wallis test was applied. The Chi-square test was used to assess differences in proportions across multiple groups. All experiments were repeated at least three times, and results for continuous variables are presented as mean ± standard deviation (SD). Unless otherwise stated, P < 0.05 was considered statistically significant.

Data availability

Public data used in this work can be acquired from (TCGA, https://www.cancer.gov/tcga/), the International Cancer Genome Consortium (ICGC, https://dcc.icgc.org/), CODE OCEAN (https://codeocean.com/), and Gene Expression Omnibus (GEO, http://www.ncbi.nlm.nih.gov/geo/). The raw experimental data supporting the conclusions of this article will be made available by the corresponding author.

Code availability

The code involved in the research is reasonably available from the corresponding author.

Abbreviations

ACD:: Autophagic cell death
AIC:: Akaike Information Criterion
APCs:: Antigen-presenting cells
ATG:: Autophagy-related genes
AUC:: Area Under the Curve
cDC1:: Conventional dendritic cell type 1
CCDI:: Combined Cell Death Index
CCK-8:: Cell Counting Kit-8
cDNA:: Complementary DNA
CNV:: Copy number variation
CR:: Complete response
DCs:: Dendritic cells
DEGs:: Differentially expressed genes
DMEM:: Dulbecco’s Modified Eagle Medium
DSP:: Digital Spatial Profiling
DAMPs:: Damage-associated molecular patterns
EAC:: Esophageal adenocarcinoma
ESCC:: Esophageal squamous cell carcinoma
ESTIMATE:: Estimation of STromal and Immune cells in MAlignant Tumor tissues using Expression data
FDR:: False discovery rate
GEO:: Gene Expression Omnibus
GSVA:: Gene Set Variation Analysis
IAC:: Invasive adenocarcinoma
ICGC:: International Cancer Genome Consortium
ICD:: Immunogenic cell death
ICIs:: Immune checkpoint inhibitors
IHC:: Immunohistochemistry
IPS:: Immunophenoscore
IPR:: Immune Phenotype Repertoire
LDH:: Lactate Dehydrogenase
LCD:: Lysosome-dependent cell death
LLC1:: Lewis Lung Carcinoma 1
LUAD:: Lung Adenocarcinoma
MHC:: Major histocompatibility complex
MIA:: Minimally invasive adenocarcinoma
MIP:: Minimum inhibitory power
MLKL:: Mixed Lineage Kinase Domain-Like protein
MMP:: Matrix metalloproteinase
MPR:: Major pathological response
mRNA:: Messenger RNA
MSI-H:: Microsatellite instability-high
MSigDB:: Molecular Signatures Database
NMPR:: Non-major pathological response
NK:: Natural killer
NSCLC:: Non-small cell lung cancer
PBMCs:: Peripheral blood mononuclear cells
PCD:: Programmed cell death
PD-1:: Programmed Death 1
PD-L1:: Programmed Death-Ligand 1
PGH2:: Prostaglandin endoperoxide H2
PGE2:: Prostaglandin E2
PR:: Partial response
PRRs:: Pattern recognition receptors
PVDF:: Polyvinylidene difluoride;
RIPA:: Radioimmunoprecipitation Assay
RIPK1:: Receptor-Interacting Protein Kinase 1
RIPK3:: Receptor-Interacting Protein Kinase 3
ROC:: Receiver Operating Characteristic
RCTD:: Robust cellular-type deconvolution
scRNA-seq:: Single-cell RNA sequencing
SD:: Standard deviation
ST:: Spatial transcriptomics
TCGA:: The Cancer Genome Atlas
TCIA:: The Cancer Immunome Atlas
TEX:: T cell exhaustion
TIME:: Tumor immune microenvironment
TME:: Tumor microenvironment
TMB:: Tumor mutational burden
UMAP:: Uniform Manifold Approximation and Projection
UMI:: Unique molecular identifiers

References

Rotow, J. & Bivona, T. G. Understanding and targeting resistance mechanisms in NSCLC. Nat. Rev. Cancer 17, 637–658 (2017).
Article CAS PubMed Google Scholar
Bray, F. et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J. Clin. 74, 229–263 (2024).
Article PubMed Google Scholar
Mountzios, G. et al. Immune-checkpoint inhibition for resectable non-small-cell lung cancer - opportunities and challenges. Nat. Rev. Clin. Oncol. 20, 664–677 (2023).
Article PubMed Google Scholar
Nicolini, A., Ferrari, P. & Duffy, M. J. Prognostic and predictive biomarkers in breast cancer: Past, present and future. Semin. Cancer Biol. 52, 56–73 (2018).
Article CAS PubMed Google Scholar
Cronin, M. et al. Analytical validation of the Oncotype DX genomic diagnostic test for recurrence prognosis and therapeutic response prediction in node-negative, estrogen receptor-positive breast cancer. Clin. Chem. 53, 1084–1091 (2007).
Article CAS PubMed Google Scholar
Albanell, J. et al. Prospective transGEICAM study of the impact of the 21-gene Recurrence Score assay and traditional clinicopathological factors on adjuvant clinical decision making in women with estrogen receptor-positive (ER+) node-negative breast cancer. Ann. Oncol. 23, 625–631 (2012).
Article CAS PubMed Google Scholar
Yang, Z. et al. Ferroptosis-related lncRNAs: Distinguishing heterogeneity of the tumour microenvironment and predicting immunotherapy response in bladder cancer. Heliyon 10, e32018 (2024).
Article CAS PubMed PubMed Central Google Scholar
Zhao, H., Huang, Y., Tong, G., Wu, W. & Ren, Y. Identification of a novel oxidative stress- and anoikis-related prognostic signature and its immune landscape analysis in non-small cell lung cancer. Int J. Mol. Sci. 24, 16188 (2023).
Article CAS PubMed PubMed Central Google Scholar
Ling, X. et al. Development of a cuproptosis-related signature for prognosis prediction in lung adenocarcinoma based on WGCNA. Transl. Lung Cancer Res. 12, 754–769 (2023).
Article CAS PubMed PubMed Central Google Scholar
Newman, A. M. et al. Robust enumeration of cell subsets from tissue expression profiles. Nat. Methods 12, 453–457 (2015).
Article CAS PubMed PubMed Central Google Scholar
Lindsten, T. et al. The combined functions of proapoptotic Bcl-2 family members bak and bax are essential for normal development of multiple tissues. Mol. Cell. 6, 1389–1399 (2000).
Article CAS PubMed PubMed Central Google Scholar
Ashkenazi, A. & Dixit, V. M. Death receptors: signaling and modulation. Science 281, 1305–1308 (1998).
Article CAS PubMed Google Scholar
Overholtzer, M. et al. A nonapoptotic cell death process, entosis, that occurs by cell-in-cell invasion. Cell 131, 966–979 (2007).
Article CAS PubMed Google Scholar
Tsvetkov, P. et al. Copper induces cell death by targeting lipoylated TCA cycle proteins. Science 375, 1254–1261 (2022).
Article CAS PubMed PubMed Central Google Scholar
Holze, C. et al. Oxeiptosis, a ROS-induced caspase-independent apoptosis-like cell-death pathway. Nat. Immunol. 19, 130–140 (2018).
Article CAS PubMed Google Scholar
Dixon, S. J. et al. Ferroptosis: an iron-dependent form of nonapoptotic cell death. Cell 149, 1060–1072 (2012).
Article CAS PubMed PubMed Central Google Scholar
Tang, D., Kang, R., Berghe, T. V., Vandenabeele, P. & Kroemer, G. The molecular machinery of regulated cell death. Cell Res. 29, 347–364 (2019).
Article CAS PubMed PubMed Central Google Scholar
Shi, J., Gao, W. & Shao, F. Pyroptosis: gasdermin-mediated programmed necrotic cell death. Trends Biochem Sci. 42, 245–254 (2017).
Article CAS PubMed Google Scholar
Wang, Y. et al. Poly(ADP-ribose) (PAR) binding to apoptosis-inducing factor is critical for PAR polymerase-1-dependent cell death (parthanatos). Sci. Signal. 4, ra20 (2011).
Article PubMed PubMed Central Google Scholar
Kroemer, G. & Levine, B. Autophagic cell death: the story of a misnomer. Nat. Rev. Mol. Cell Biol. 9, 1004–1010 (2008).
Article CAS PubMed PubMed Central Google Scholar
Aits, S. & Jaattela, M. Lysosomal cell death at a glance. J. Cell Sci. 126, 1905–1912 (2013).
Article CAS PubMed Google Scholar
Zheng, T., Liu, Q., Xing, F., Zeng, C. & Wang, W. Disulfidptosis: a new form of programmed cell death. J. Exp. Clin. Cancer Res. 42, 137 (2023).
Article CAS PubMed PubMed Central Google Scholar
Kroemer, G., Galluzzi, L., Kepp, O. & Zitvogel, L. Immunogenic cell death in cancer therapy. Annu. Rev. Immunol. 31, 51–72 (2013).
Article CAS PubMed Google Scholar
Taddei, M. L., Giannoni, E., Fiaschi, T. & Chiarugi, P. Anoikis: an emerging hallmark in health and diseases. J. Pathol. 226, 380–393 (2012).
Article CAS PubMed Google Scholar
Papayannopoulos, V. Neutrophil extracellular traps in immunity and disease. Nat. Rev. Immunol. 18, 134–147 (2018).
Article CAS PubMed Google Scholar
Maltese, W. A. & Overmeyer, J. H. Non-apoptotic cell death associated with perturbations of macropinocytosis. Front. Physiol. 6, 38 (2015).
Article PubMed PubMed Central Google Scholar
Sperandio, S., de Belle, I. & Bredesen DE An alternative, nonapoptotic form of programmed cell death. Proc. Natl Acad. Sci. USA 97, 14376–14381 (2000).
Article CAS PubMed PubMed Central Google Scholar
Yan, J., Wan, P., Choksi, S. & Liu, Z. G. Necroptosis and tumor progression. Trends Cancer 8, 21–27 (2022).
Article CAS PubMed Google Scholar
Gao, W., Wang, X., Zhou, Y., Wang, X. & Yu, Y. Autophagy, ferroptosis, pyroptosis, and necroptosis in tumor immunotherapy. Signal Transduct. Target. Ther. 7, 196 (2022).
Article PubMed PubMed Central Google Scholar
Necchi, A. et al. Atezolizumab in platinum-treated locally advanced or metastatic urothelial carcinoma: post-progression outcomes from the phase II IMvigor210 study. Ann. Oncol. 28, 3044–3050 (2017).
Article CAS PubMed PubMed Central Google Scholar
Hugo, W. et al. Genomic and transcriptomic features of response to anti-PD-1 therapy in metastatic melanoma. Cell 165, 35–44 (2016).
Article CAS PubMed PubMed Central Google Scholar
Ascierto, M. L. et al. The intratumoral balance between metabolic and immunologic gene expression is associated with anti-PD-1 response in patients with renal cell carcinoma. Cancer Immunol. Res. 4, 726–733 (2016).
Article CAS PubMed PubMed Central Google Scholar
van den Ende, T. et al. Neoadjuvant chemoradiotherapy combined with atezolizumab for resectable esophageal adenocarcinoma: a single-arm phase II feasibility trial (PERFECT). Clin. Cancer Res. 27, 3351–3359 (2021).
Article PubMed Google Scholar
Cho, J. W. et al. Genome-wide identification of differentially methylated promoters and enhancers associated with response to anti-PD-1 therapy in non-small cell lung cancer. Exp. Mol. Med. 52, 1550–1563 (2020).
Article CAS PubMed PubMed Central Google Scholar
Jung, H. et al. DNA methylation loss promotes immune evasion of tumours with high mutation and copy number load. Nat. Commun. 10, 4278 (2019).
Article PubMed PubMed Central Google Scholar
Limagne, E. et al. MEK inhibition overcomes chemoimmunotherapy resistance by inducing CXCL10 in cancer cells. Cancer Cell. 40, 136–52.e12 (2022).
Article CAS PubMed Google Scholar
Charoentong, P. et al. Pan-cancer immunogenomic analyses reveal genotype-immunophenotype relationships and predictors of response to checkpoint blockade. Cell Rep. 18, 248–262 (2017).
Article CAS PubMed Google Scholar
Bischoff, P. et al. Single-cell RNA sequencing reveals distinct tumor microenvironmental patterns in lung adenocarcinoma. Oncogene 40, 6748–6758 (2021).
Article CAS PubMed PubMed Central Google Scholar
Kim, N. et al. Single-cell RNA sequencing demonstrates the molecular and cellular reprogramming of metastatic lung adenocarcinoma. Nat. Commun. 11, 2285 (2020).
Article CAS PubMed PubMed Central Google Scholar
Zhu, J. et al. Delineating the dynamic evolution from preneoplasia to invasive lung adenocarcinoma by integrating single-cell RNA sequencing and spatial transcriptomics. Exp. Mol. Med. 54, 2060–2076 (2022).
Article CAS PubMed PubMed Central Google Scholar
Pasparakis, M. & Vandenabeele, P. Necroptosis and its role in inflammation. Nature 517, 311–320 (2015).
Article CAS PubMed Google Scholar
Aaes, T. L. et al. Vaccination with necroptotic cancer cells induces efficient anti-tumor immunity. Cell Rep. 15, 274–287 (2016).
Article CAS PubMed Google Scholar
Ba, Y. et al. Pyroptosis-derived long noncoding RNA profiles reveal a novel signature for evaluating the prognosis of patients with lung adenocarcinoma. JCO Precis Oncol. 8, e2300405 (2024).
Article PubMed PubMed Central Google Scholar
Mao, X., Huang, W., Xue, Q. & Zhang, X. Prognostic impact and immunotherapeutic implications of NETosis-related prognostic model in clear cell renal cell carcinoma. J. Cancer Res Clin. Oncol. 150, 278 (2024).
Article CAS PubMed PubMed Central Google Scholar
Zhang, K., Li, J. & Yuan, E. A necroptosis-related gene signature to predict prognosis and immune features in hepatocellular carcinoma. BMC Cancer 23, 660 (2023).
Article CAS PubMed PubMed Central Google Scholar
Li, C., Weng, J., Yang, L., Gong, H. & Liu, Z. Development of an anoikis-related gene signature and prognostic model for predicting the tumor microenvironment and response to immunotherapy in colorectal cancer. Front. Immunol. 15, 1378305 (2024).
Article CAS PubMed PubMed Central Google Scholar
Liu, Y. et al. Prognostic implications of autophagy-associated gene signatures in non-small cell lung cancer. Aging 11, 11440–11462 (2019).
Article CAS PubMed PubMed Central Google Scholar
Hanggi, K. & Ruffell, B. Cell death, therapeutics, and the immune response in cancer. Trends Cancer 9, 381–396 (2023).
Article CAS PubMed PubMed Central Google Scholar
Zou, Y. et al. Leveraging diverse cell-death patterns to predict the prognosis and drug sensitivity of triple-negative breast cancer patients after surgery. Int. J. Surg. 107, 106936 (2022).
Article PubMed Google Scholar
Cao, K. et al. Analysis of multiple programmed cell death-related prognostic genes and functional validations of necroptosis-associated genes in oesophageal squamous cell carcinoma. EBioMedicine 99, 104920 (2024).
Article CAS PubMed Google Scholar
Liu, X. et al. Panoramic analysis of cell death patterns reveals prognostic and immune profiles of head and neck squamous cell carcinoma. Am. J. Cancer Res. 14, 2584–2607 (2024).
Article PubMed PubMed Central Google Scholar
Wu, Z., Jin, M., Xin, P. & Zhang, H. Leveraging diverse cell-death related signature predicts the prognosis and immunotherapy response in renal clear cell carcinoma. Front. Immunol. 14, 1293729 (2023).
Article CAS PubMed PubMed Central Google Scholar
Wang, X. et al. Integrated bulk and single-cell RNA sequencing identifies an aneuploidy-based gene signature to predict sensitivity of lung adenocarcinoma to traditional chemotherapy drugs and patients’ prognosis. PeerJ 12, e17545 (2024).
Article PubMed PubMed Central Google Scholar
Chen, Y. et al. Identification of a novel prognostic signature based on vitamin metabolism clustering-related genes in lung adenocarcinoma. Transl. Lung Cancer Res. 13, 1084–1100 (2024).
Article CAS PubMed PubMed Central Google Scholar
Zhou, J. et al. Identification and validation of a glycosyltransferase gene signature as a novel prognostic model for lung adenocarcinoma. Heliyon 10, e29383 (2024).
Article CAS PubMed PubMed Central Google Scholar
Cao, K. et al. Prediction of prognosis and immunotherapy response with a robust immune-related lncRNA pair signature in lung adenocarcinoma. Cancer Immunol. Immunother. 71, 1295–1311 (2022).
Article CAS PubMed Google Scholar
Liu, Y. et al. Contributions and prognostic values of m(6) A RNA methylation regulators in non-small-cell lung cancer. J. Cell Physiol. 235, 6043–6057 (2020).
Article CAS PubMed Google Scholar
Wang, S. et al. Machine learning reveals diverse cell death patterns in lung adenocarcinoma prognosis and therapy. NPJ Precis. Oncol. 8, 49 (2024).
Article CAS PubMed PubMed Central Google Scholar
Holder, A. M. et al. Defining clinically useful biomarkers of immune checkpoint inhibitors in solid tumours. Nat. Rev. Cancer 24, 498–512 (2024).
Article CAS PubMed Google Scholar
Postow, M. A., Sidlow, R. & Hellmann, M. D. Immune-related adverse events associated with immune checkpoint blockade. N. Engl. J. Med. 378, 158–168 (2018).
Article CAS PubMed Google Scholar
Xue, Q. et al. Promising immunotherapeutic targets in lung cancer based on single-cell RNA sequencing. Front. Immunol. 14, 1148061 (2023).
Article CAS PubMed PubMed Central Google Scholar
Traynor, S. et al. Single-cell sequencing unveils extensive intratumoral heterogeneity of cancer/testis antigen expression in melanoma and lung cancer. J. Immunother. Cancer 12, e008759 (2024).
Article PubMed PubMed Central Google Scholar
De Zuani, M. et al. Single-cell and spatial transcriptomics analysis of non-small cell lung cancer. Nat. Commun. 15, 4388 (2024).
Article PubMed PubMed Central Google Scholar
Laughney, A. M. et al. Regenerative lineages and immune-mediated pruning in lung cancer metastasis. Nat. Med. 26, 259–269 (2020).
Article CAS PubMed PubMed Central Google Scholar
Zhang, Q. et al. The spatial transcriptomic landscape of non-small cell lung cancer brain metastasis. Nat. Commun. 13, 5983 (2022).
Article PubMed PubMed Central Google Scholar
Yoshihara, K. et al. Inferring tumour purity and stromal and immune cell admixture from expression data. Nat. Commun. 4, 2612 (2013).
Article PubMed Google Scholar
Hahn, T. & Akporiaye, E. T. alpha-TEA as a stimulator of tumor autophagy and enhancer of antigen cross-presentation. Autophagy 9, 429–431 (2013).
Article CAS PubMed PubMed Central Google Scholar
Li, Y. et al. The vitamin E analogue alpha-TEA stimulates tumor autophagy and enhances antigen cross-presentation. Cancer Res. 72, 3535–3545 (2012).
CAS PubMed PubMed Central Google Scholar
White, E., Lattime, E. C. & Guo, J. Y. Autophagy regulates stress responses, metabolism, and anticancer immunity. Trends Cancer 7, 778–789 (2021).
Article CAS PubMed PubMed Central Google Scholar
Zhang, T., Wang, Y., Inuzuka, H. & Wei, W. Necroptosis pathways in tumorigenesis. Semin. Cancer Biol. 86, 32–40 (2022).
Article CAS PubMed PubMed Central Google Scholar
Kaczmarek, A., Vandenabeele, P. & Krysko, D. V. Necroptosis: the release of damage-associated molecular patterns and its physiological relevance. Immunity 38, 209–223 (2013).
Article CAS PubMed Google Scholar
Ahmed, A. & Tait, S. W. G. Targeting immunogenic cell death in cancer. Mol. Oncol. 14, 2994–3006 (2020).
Article CAS PubMed PubMed Central Google Scholar
Hanahan, D. & Weinberg, R. A. Hallmarks of cancer: the next generation. Cell 144, 646–674 (2011).
Article CAS PubMed Google Scholar
Greaves, M. & Maley, C. C. Clonal evolution in cancer. Nature 481, 306–313 (2012).
Article CAS PubMed PubMed Central Google Scholar
Polyak, K. Tumor heterogeneity confounds and illuminates: a case for Darwinian tumor evolution. Nat. Med. 20, 344–346 (2014).
Article CAS PubMed Google Scholar
Rood, J. E., Maartens, A., Hupalowska, A., Teichmann, S. A. & Regev, A. Impact of the human cell atlas on medicine. Nat. Med. 28, 2486–2496 (2022).
Article CAS PubMed Google Scholar
Trapnell, C. et al. The dynamics and regulators of cell fate decisions are revealed by pseudotemporal ordering of single cells. Nat. Biotechnol. 32, 381–386 (2014).
Article CAS PubMed PubMed Central Google Scholar
Kong, Y. et al. Single-cell analysis identifies PLK1 as a driver of immunosuppressive tumor microenvironment in LUAD. PLoS Genet. 20, e1011309 (2024).
Article CAS PubMed PubMed Central Google Scholar
Gao, P. et al. High expression of PTGES3 is an independent predictive poor prognostic biomarker and correlates with immune infiltrates in lung adenocarcinoma. Int. Immunopharmacol. 110, 108954 (2022).
Article CAS PubMed Google Scholar
Yin, Q. et al. The integration of multidisciplinary approaches revealed PTGES3 as a novel drug target for breast cancer treatment. J. Transl. Med. 22, 84 (2024).
Article CAS PubMed PubMed Central Google Scholar
Wang, H. et al. Comprehensive pan-cancer analysis of PTGES3 and its prognostic role in hepatocellular carcinoma. Front. Oncol. 13, 1158490 (2023).
Article CAS PubMed PubMed Central Google Scholar
Yu, X. et al. A cytosolic heat shock protein 90 and co-chaperone p23 complex activates RIPK3/MLKL during necroptosis of endothelial cells in acute respiratory distress syndrome. J. Mol. Med. 98, 569–583 (2020).
Article CAS PubMed Google Scholar
Schweiger, A. et al. Cysteine proteinase cathepsin H in tumours and sera of lung cancer patients: relation to prognosis and cigarette smoking. Br. J. Cancer 82, 782–788 (2000).
Article CAS PubMed PubMed Central Google Scholar
Du, Z. et al. A prognostic model of drug tolerant persister-related genes in lung adenocarcinoma based on single cell and bulk RNA sequencing data. Heliyon 9, e20708 (2023).
Article CAS PubMed PubMed Central Google Scholar
Bell, H. N. & Zou, W. Beyond the barrier: unraveling the mechanisms of immunotherapy resistance. Annu.p Rev. Immunol. 42, 521–550 (2024).
Article CAS Google Scholar
Tumeh, P. C. et al. PD-1 blockade induces responses by inhibiting adaptive immune resistance. Nature 515, 568–571 (2014).
Article CAS PubMed PubMed Central Google Scholar
Rosenbaum, S. R., Wilski, N. A. & Aplin, A. E. Fueling the fire: inflammatory forms of cell death and implications for cancer immunotherapy. Cancer Discov. 11, 266–281 (2021).
Article CAS PubMed PubMed Central Google Scholar
Li, S. et al. Targeted therapy for non-small-cell lung cancer: New insights into regulated cell death combined with immunotherapy. Immunol. Rev. 321, 300–334 (2024).
Article CAS PubMed Google Scholar
Luo, R. et al. Necroptosis-dependent Immunogenicity of Cisplatin: Implications for Enhancing the Radiation-induced Abscopal Effect. Clin. Cancer Res. 29, 667–683 (2023).
Article CAS PubMed Google Scholar
Aho, K., Derryberry, D. & Peterson, T. Model selection for ecologists: the worldviews of AIC and BIC. Ecology 95, 631–636 (2014).
Article PubMed Google Scholar
Huang, X., Zhang, G., Tang, T. & Liang, T. Identification of tumor antigens and immune subtypes of pancreatic adenocarcinoma for mRNA vaccine development. Mol. Cancer 20, 44 (2021).
Article CAS PubMed PubMed Central Google Scholar
Miao, Y. R. et al. ImmuCellAI: a unique method for comprehensive T-cell subsets abundance prediction and its application in cancer immunotherapy. Adv. Sci.7, 1902880 (2020).
Article CAS Google Scholar
Hoshida, Y., Brunet, J. P., Tamayo, P., Golub, T. R. & Mesirov, J. P. Subclass mapping: identifying common subtypes in independent disease data sets. PLoS ONE 2, e1195 (2007).
Article PubMed PubMed Central Google Scholar
Butler, A., Hoffman, P., Smibert, P., Papalexi, E. & Satija, R. Integrating single-cell transcriptomic data across different conditions, technologies, and species. Nat. Biotechnol. 36, 411–420 (2018).
Article CAS PubMed PubMed Central Google Scholar
Korsunsky, I. et al. Fast, sensitive and accurate integration of single-cell data with harmony. Nat. Methods 16, 1289–1296 (2019).
Article CAS PubMed PubMed Central Google Scholar
Aran, D. et al. Reference-based analysis of lung single-cell sequencing reveals a transitional profibrotic macrophage. Nat. Immunol. 20, 163–172 (2019).
Article CAS PubMed PubMed Central Google Scholar
Liu, Z. et al. Progenitor-like exhausted SPRY1(+)CD8(+) T cells potentiate responsiveness to neoadjuvant PD-1 blockade in esophageal squamous cell carcinoma. Cancer Cell. 41, 1852–70.e9 (2023).
Article CAS PubMed Google Scholar
Cable, D. M. et al. Robust decomposition of cell type mixtures in spatial transcriptomics. Nat. Biotechnol. 40, 517–526 (2022).
Article CAS PubMed Google Scholar
Street, K. et al. Slingshot: cell lineage and pseudotime inference for single-cell transcriptomics. BMC Genomics 19, 477 (2018).
Article PubMed PubMed Central Google Scholar
Qiu, X. et al. Reversed graph embedding resolves complex single-cell trajectories. Nat. Methods 14, 979–982 (2017).
Article CAS PubMed PubMed Central Google Scholar
Gao, R. et al. Delineating copy number and clonal substructure in human tumors from single-cell transcriptomes. Nat. Biotechnol. 39, 599–608 (2021).
Article CAS PubMed PubMed Central Google Scholar
Li, B. et al. Cumulus provides cloud-based data analysis for large-scale single-cell and single-nucleus RNA-seq. Nat. Methods 17, 793–798 (2020).
Article CAS PubMed PubMed Central Google Scholar
Subramanian, A. et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc. Natl Acad. Sci. USA 102, 15545–15550 (2005).
Article CAS PubMed PubMed Central Google Scholar
Hanzelmann, S., Castelo, R. & Guinney, J. GSVA: gene set variation analysis for microarray and RNA-seq data. BMC Bioinform. 14, 7 (2013).
Article Google Scholar
Petitprez, F. et al. The murine Microenvironment Cell Population counter method to estimate abundance of tissue-infiltrating immune and stromal cell populations in murine samples using gene expression. Genome Med. 12, 86 (2020).
Article CAS PubMed PubMed Central Google Scholar
Li, B. et al. Comprehensive analyses of tumor immunity: implications for cancer immunotherapy. Genome Biol. 17, 174 (2016).
Article PubMed PubMed Central Google Scholar
Aran, D., Hu, Z. & Butte, A. J. xCell: digitally portraying the tissue cellular heterogeneity landscape. Genome Biol. 18, 220 (2017).
Article PubMed PubMed Central Google Scholar
Racle, J., de Jonge, K., Baumgaertner, P., Speiser, D. E. & Gfeller, D. Simultaneous enumeration of cancer and immune cell types from bulk tumor gene expression data. eLife 6, e26476 (2017).
Article PubMed PubMed Central Google Scholar
Finotello, F. et al. Molecular and pharmacological modulators of the tumor immune contexture revealed by deconvolution of RNA-seq data. Genome Med. 11, 34 (2019).
Article PubMed PubMed Central Google Scholar

Download references

Acknowledgements

This work was supported by the Natural Science Foundation of China [82172786 and 82072985], Heilongjiang Provincial Natural Science Foundation of China [ZD2024H003], Beijing Xisike Clinical Oncology Foundation (Y-2024AZ(EGFR)MS-0187), the National Cancer Center Climbing Fund of China [NCC201908B06], and the Climbing Program of Harbin Medical University Cancer Hospital [PDYS2024-02], Heilongjiang Province Innovation Base Award Project [JD2023SJ03], Nn10 Project of Harbin Medical University Cancer Hospital [Nn102024-05].

Author information

These authors contributed equally: Kui Cao, Shenshui Wei, Tianjiao Ma, Xinxin Yang.

Authors and Affiliations

Department of Thoracic Surgery, Harbin Medical University Cancer hospital, Harbin, Heilongjiang, China
Kui Cao, Mengdi Lu, Yuwen Bai, Cuicui Qi, Luquan Zhang, Lijuan Li & Jianqun Ma
Department of Clinical Laboratory, Harbin Medical University Cancer Hospital, Harbin, Heilongjiang, China
Shenshui Wei, Yuning Wang, Xiangrong He & Jinhong Zhu
Department of Cardiovascular Surgery, Sun Yat-sen Memorial Hospital, Sun Yat-sen University, Guangzhou, China
Tianjiao Ma
Department of Pathology, Harbin Medical University Cancer Hospital, Harbin, Heilongjiang, China
Xinxin Yang & Hongxue Meng
Biobank, Harbin Medical University Cancer Hospital, Harbin, Heilongjiang, China
Jinhong Zhu

Authors

Kui Cao
View author publications
Search author on:PubMed Google Scholar
Shenshui Wei
View author publications
Search author on:PubMed Google Scholar
Tianjiao Ma
View author publications
Search author on:PubMed Google Scholar
Xinxin Yang
View author publications
Search author on:PubMed Google Scholar
Yuning Wang
View author publications
Search author on:PubMed Google Scholar
Xiangrong He
View author publications
Search author on:PubMed Google Scholar
Mengdi Lu
View author publications
Search author on:PubMed Google Scholar
Yuwen Bai
View author publications
Search author on:PubMed Google Scholar
Cuicui Qi
View author publications
Search author on:PubMed Google Scholar
Luquan Zhang
View author publications
Search author on:PubMed Google Scholar
Lijuan Li
View author publications
Search author on:PubMed Google Scholar
Hongxue Meng
View author publications
Search author on:PubMed Google Scholar
Jianqun Ma
View author publications
Search author on:PubMed Google Scholar
Jinhong Zhu
View author publications
Search author on:PubMed Google Scholar

Contributions

All authors contributed to the study's conception and design. Data preparation, collection, and analysis were performed by Kui Cao, Shenshui Wei, Tianjiao Ma, Jinhong Zhu, Jianqun Ma, Hongxue Meng, Xinxin Yang, Mengdi Lu, Yuwen Bai, Yuning Wang, Cuicui Qi, Xiangrong He, Luquan Zhang, and Lijuan Li. Jinhong Zhu and Mengdi Lu accessed and verified the underlying data. Reconciliation of underlying data was solved by Jianqun Ma and Jinhong Zhu. The draft of the manuscript was written by Jinhong Zhu, Kui Cao, and Jianqun Ma. All authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Corresponding authors

Correspondence to Hongxue Meng, Jianqun Ma or Jinhong Zhu.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Cao, K., Wei, S., Ma, T. et al. Integrating bulk, single-cell, and spatial transcriptomics to identify and functionally validate novel targets to enhance immunotherapy in NSCLC. npj Precis. Onc. 9, 112 (2025). https://doi.org/10.1038/s41698-025-00893-x

Download citation

Received: 20 October 2024
Accepted: 31 March 2025
Published: 16 April 2025
DOI: https://doi.org/10.1038/s41698-025-00893-x

Subjects

Abstract

Similar content being viewed by others

Identifying key circulating tumor DNA parameters for predicting clinical outcomes in metastatic non-squamous non-small cell lung cancer after first-line chemoimmunotherapy

Molecular determinants of response to PD-L1 blockade across tumor types

Gene expression profiles of the original tumors influence the generation of PDX models of lung squamous cell carcinoma

Introduction

Results

Establishing and validating models for 18 PCDs and CCDI

Predicting responses to ICIs in NSCLC patients using CCDI

Differences in tumor-immune microenvironment between the high and low-risk groups

Dynamic alterations of CCDI and its genes during lung cancer cell state transition

Association of lung cancer metastasis with CCDI and constituent genes

Spatial distribution of CCDI and its genes during the LUAD invasion

Dynamics of CCDI in NSCLC treated with PD-1 blockade and chemotherapy

Functional analysis of CTSH and PTGES3

Discussion

Methods

Data source

Construction and validation of the CCDI model

Assessing the ability of the CCDI model’ ability to predict ICI responses

scRNA‑seq data and ST data processing

Cell trajectory prediction

Discrimination of malignant and non-malignant epithelial cells

Identifying differentially expressed genes

Gene set variation analysis (GSVA)

Evaluation of tumor-immune microenvironment

Cell culture and reagents

Lentivirus infection and plasmid transfection

Quantitative reverse transcription PCR (qRT-PCR)

Western blot analysis

Co-immunoprecipitation (Co-IP)

Proliferation and migration assay

LDH release assay

In vivo treatments

Immunohistochemistry staining (IHC)

Statistics

Data availability

Code availability

Abbreviations

References

Acknowledgements

Author information

Authors and Affiliations

Contributions

Corresponding authors

Ethics declarations

Competing interests

Additional information

Supplementary information

Supplementary information

Rights and permissions

About this article

Cite this article

Share this article

Search

Quick links