Abstract Small cell lung cancer (SCLC) is still one of the most formidable challenges in oncology. In this study, we introduce an innovative risk scoring model rooted in cancer-associated fibroblast (CAF)-related functional genes, designed to predict patient prognosis and illuminate the microenvironment of SCLC. Through Kaplan–Meier survival analysis and receiver operating characteristic (ROC) curves, our model could effectively classify patients into high- and low-risk groups, with distinct survival outcomes and remarkable predictive accuracy, which has been evidenced by the AUC values. The low-risk patients showed a more active immune environment, characterized by more infiltration of dendritic cells, natural killer cells, and higher expression of immune co-stimulation molecules. On the contrary, high-risk patients displayed an enrichment of DNA repair and glycolysis pathways associated with tumor aggressiveness and treatment resistance. These results suggest that the risk model offers a nuanced view of response to immunotherapy that may guide the identification of patients who may benefit from immunotherapy. Moreover, we also verified the function of the key gene UBE2E2 by SCLC cell line experiments. Silencing UBE2E2 results in decreased cell proliferation and migration as well as increased apoptosis, which enhances its important role in SCLC biology. In summary, our study highlights the prognostic potential of the CAF-related functional gene risk model and its implications for predicting immune microenvironment status and guiding SCLC treatment strategies. Supplementary Information The online version contains supplementary material available at 10.1007/s12672-025-02781-z. Keywords: Small cell lung cancer, Tumor immune microenvironment, Cancer-associated fibroblast, Predictive model Introduction Small cell lung cancer (SCLC) is a highly aggressive cancer type, known by rapid progression and early metastasis [[46]1]. According to the latest data from the SEER database, the 5-year survival rates for SCLC patients vary significantly by stage: 33.3% for localized stage, 19% for regional stage, and only 3.9% for distant metastatic stage. Unfortunately, the majority of patients are diagnosed at an extensive stage with metastasis already present. In recent years, immunotherapy has provided a glimmer of hope for SCLC patients, yet only approximately 15% of patients experience long-term benefits from this treatment [[47]2, [48]3]. The incomplete understanding of the SCLC tumor microenvironment (TME), coupled with the lack of reliable prognostic biomarkers, makes risk stratification and the identification of precise therapeutic targets more challenging compared to other cancers [[49]4]. Advances in single-cell technology are gradually revealing the complex nature of SCLC TME. One of the most striking features is the significant heterogeneity in immune cell infiltration [[50]5]. Especially the already scarce CD8 + T cells, are found to be localized to the tumor periphery rather than the core in many SCLC cases, a phenomenon known as immune exclusion [[51]6]. This limited immune infiltration has largely contributed to the restricted efficacy of immunotherapy in SCLC. Recent studies have suggested that cancer-associated fibroblasts (CAFs) are essential for the development of this immune-excluded phenotype [[52]7]. As a key stromal component of TME, CAFs actively remodel the extracellular matrix (ECM), creating a dense fibrotic barrier that impedes immune cell infiltration while facilitating tumor invasion [[53]8]. Beyond their structural role, CAFs secrete pro-inflammatory cytokines such as IL-6, IL-11, and TGF-β, which promote tumor proliferation, epithelial-mesenchymal transition (EMT), and therapy resistance [[54]9]. Additionally, CAFs contribute to immune evasion by recruiting immunosuppressive cells like myeloid-derived suppressor cells (MDSCs) and regulatory T cells (Tregs), further dampening antitumor immune responses [[55]10]. While their immunosuppressive effects are well-documented in many other cancers, their specific contribution to SCLC remains less well understood. Latest spatial transcriptomics data have revealed the potential role of CAFs to drive intratumoral heterogeneity and plasticity in SCLC [[56]11], underscoring their relevance in the TME. Due to the complexity of SCLC TME and its profound impact on tumor behavior, a multidimensional approach to understanding cancer progression is necessary. Recent studies incorporating multiomics perspectives have significantly advanced our understanding of tumor-immunological interactions, providing new insights into tumor heterogeneity and immune modulation [[57]12]. Risk score models that integrate transcriptomic and immune data with clinical outcomes have emerged as powerful tools for prognostic stratification and have been successfully applied to various cancer types [[58]13–[59]15]. However, their application in SCLC remains limited, and comprehensive models integrating CAF-related functional genes are particularly scarce. Here we developed a novel four-gene risk model based on CAF-related functional genes aimed at predicting overall survival (OS) and TME status of SCLC patients, and confirmed the biological function of the key molecule UBE2E2 by cell experiments. This model can not only serve as a reliable prognostic tool, but also provide new insights into SCLC TME. Materials and methods Dataset sources We downloaded 14 samples with complete single-cell expression profiles (BioProject ID PRJCA006026) from the National Genomics Data Center [[60]5]. Additionally, the [61]GSE60052 dataset, which includes transcriptomic data of 48 patients was downloaded from the GEO database and annotated with the platform [62]GPL11154 [[63]16]. For further verification, we retrieved the transcriptomic clinical information of 77 SCLC samples of the George cohort from the cBioPortal database [[64]17]. Moreover, we integrated the TCGA database lung adenocarcinoma and lung squamous carcinoma data to the TCGA-NSCLC cohort, including 1,001 patients with available survival information [[65]18]. These datasets provided the foundation for single-cell exploration, model construction, and external validation. Single cell analysis This analysis aimed to identify major cell subtypes in the SCLC tumor microenvironment. We first employed the Seurat package to process the expression profiles, filtering based on UMI count and gene detection per cell. The data were standardized, normalized, and reduced in dimensionality using PCA and Harmony. The optimal number of principal components was selected based on the elbow plot, and cluster relationships were visualized using UMAP. For cell annotation, we manually identified cell types by querying the CellMarker and panglaodb databases, as well as relevant literature, to find the marker genes specific to the tissue [[66]19, [67]20]. Genes with |avg log2FC|> 1 and adjusted p-value < 0.05 were selected as unique markers for each cell subtype. CellChat analysis This analysis was conducted to investigate intercellular communication, especially the signaling roles of CAFs in SCLC. To investigate the intercellular communication network, we employed CellChat, a computational tool designed to analyze single-cell data. CellChat uses normalized single-cell gene expression profiles, combined with cell type annotations, to infer ligand-receptor interactions between cells, thereby constructing the intercellular communication network [[68]21]. This method quantifies the interaction strength (weights) and frequency (counts) between different cell types, systematically analyzing their interactions. By analyzing these interactions, we gained a deeper understanding of how different cell types communicate within the TME of SCLC, shedding light on their potential contributions to tumor development and progression. Model construction and prognosis This step aimed to build a CAF-related gene signature that could predict patient survival. We initially identified 5267 genes associated with CAF functions from the CAFrgDB, a comprehensive database of CAF-related genes [[69]22]. We developed the risk score model by Lasso regression, with the weight of each gene determined by its respective regression coefficient [[70]23]. This allowed us to calculate an individual risk score for each patient. Patients were then categorized into low and high risk based on the median risk score. Kaplan–Meier survival analyses were performed to compare OS between the two groups. Receiver operating characteristic (ROC) curve analysis was used to ensure its reliability as a potential clinical tool [[71]24]. Nomogram model construction To improve clinical applicability, we integrated risk scores and clinical variables into a visual prediction model. Nomograms can simplify complex prediction models by intuitively representing the relationships between different variables [[72]25]. In this study, each variable was assigned a score proportional to its regression coefficient, reflecting its weighted contribution to the final prediction. These individual scores were then summed to calculate a total score that was used to predict OS. This approach provides a friendly way to visualize how various factors collectively influence prognosis, making it easier for doctors to apply the model in practice. Immune cell infiltration analysis This analysis aimed to compare immune infiltration patterns between high- and low-risk patients. In order to unravel the intricate TME map of SCLC, we employed single-sample genomic enrichment analysis (ssGSEA), a powerful computational method that quantifies the relative abundance of immune cell types and key immune molecules. We focused on a comprehensive profile of 29 immune-related molecules and cell ratios based on previous studies [[73]26]. GSVA and GSEA analysis These analyses were used to identify biological pathways and mechanisms enriched in different risk groups. We utilized gene set variation analysis (GSVA) and GSEA for pathway analysis. GSVA scores were computed for gene sets sourced from the Molecular Signatures Database (v7.0) to assess pathway activity across different samples [[74]27]. GSEA was performed to identify differentially enriched pathways [[75]28, [76]29]. Pathways were ranked by consistency scores and those with adjusted p-values < 0.05 were considered significant. These analyses revealed key molecular mechanisms and biological pathways contributing to differences in survival and TME status between the two risk groups. Tissue acquisition and ethical approval To validate the expression of model gene in patient tissue samples, we obtained pre-treatment biopsy specimens from 10 patients diagnosed with SCLC from the Biobank of Zhejiang Cancer Hospital. The use of these samples was approved by the Institutional Review Board of Zhejiang Cancer Hospital (Approval No. IRB-2024-301), and the study was conducted in accordance with the ethical standards of the Declaration of Helsinki. Written informed consent was waived by the ethics committee due to the retrospective nature of the study and the use of anonymized archival specimens. Hematoxylin–eosin (H&E) staining This staining was used to evaluate tissue morphology and confirm the histological features of tumor specimens. Tissue sections were deparaffinized with xylene and then rehydrated through a series of graded ethanol solutions and washes with water. The cells were stained with hematoxylin for 6 min, differentiated with ethanol hydrochloride, and blue with ammonia water. Eosin staining was done for 10–30 s. The sections were then dehydrated through ethanol and xylene, followed by mounting with neutral balsam further analysis. Immunohistochemistry (IHC) staining IHC staining was used to validate the protein expression level of UBE2E2 in SCLC tissues. Following the deparaffinization and rehydration steps, antigen retrieval was carried out using TRIS–EDTA buffer. Endogenous peroxidase activity was inhibited with a 3% hydrogen peroxide solution. The slides were then blocked with 3% BSA and incubated with the primary antibody (Affinity) overnight at 4 °C. Afterward, they were incubated with the secondary antibody. DAB staining was performed to visualize the antigen, and slides were counterstained with hematoxylin. After dehydration and clearing, the slides were mounted and analyzed under a microscope. Cell culture and transfection This step prepared SCLC cells for functional validation of UBE2E2 through in vitro knockdown experiments. The human SCLC cell line DMS53 was kindly provided by Prof. Chen Ming (Sun Yat-sen University Cancer Center, Guangzhou), identified by short tandem repeat assay. Cells were cultured in RPMI-1640 medium (Gibco) with 10% FBS (Sigma) and 1% Penicillin–Streptomycin (Biosharp), and maintained at 37 °C in a 5% CO[2] atmosphere. The UBE2E2 siRNA expression vector and scrambled siRNA nontarget control were purchased from Genepharma. All sequences were transfected using siRNA-Mate Plus Transfection Reagent (Genepharma). Quantitative real-time PCR (qRT-PCR) qRT-PCR was performed to measure gene expression changes following UBE2E2 silencing, confirming the knockdown efficiency. Total RNA was isolated using Triquick Reagent (Solarbio), followed by cDNA synthesis using the HiScript II 1st Strand cDNA Synthesis Kit (Nanjing Vazyme Biotech). qRT-PCR was performed using the Taq Pro Universal SYBR qPCR Master Mix (Nanjing Vazyme Biotech) on the 7500 Fast Real-time PCR system (Thermo Fisher Scientific). β-actin served as the internal control for normalization. Relative expression levels of target genes were calculated using the Livak method (2^−ΔΔCt). The sequences and primers showed in Supplementary Table 1. Immunoblotting Western blotting was used to confirm the knockdown of UBE2E2 at the protein level. Western blotting was used to validate protein-level changes in UBE2E2 expression, complementing the qRT-PCR results. For protein extraction, samples were mixed with 5 × SDS-PAGE loading buffer and boiled at 100 °C for 10 min. Proteins were separated by SDS-PAGE and transferred onto a PVDF membrane. The membrane was blocked with 5% skimmed milk for 90 min at room temperature. Primary antibodies (Abcam) were incubated overnight at 4 °C, followed by three washes with TBST (0.05% Tween 20). The membrane was then incubated with a secondary antibody for 90 min at room temperature. After further washes, protein bands were detected using an enhanced chemiluminescence reagent (ECL, Abbkine). Protein quantification was performed relative to β-actin using ImageJ software. Cell viability assay The CCK-8 assay was used to assess cell viability after UBE2E2 knockdown, determining its effect on SCLC cell survival. DMS53 cells were transfected with siRNA and plated into 96-well plates at a density of 5 × 103 cells per well. After 24 h of incubation, the culture medium was replaced with fresh medium containing 10% CCK-8 solution (APExBIO). Absorbance at 450 nm was measured after 2 h of incubation at 37 °C in the dark at 0, 24, 48, and 72 h to evaluate cell viability. Colony formation assay Colony formation assays were conducted to evaluate the clonogenic potential of cells following UBE2E2 silencing. After 24 h of siRNA transfection, DMS53 cells were seeded into 6-well plates at a density of 1 × 103 cells per well. Ten days later, colonies were fixed, stained, photographed, and counted. Wound-healing assay A wound-healing assay was performed to investigate the migratory ability of DMS53 cells after UBE2E2 knockdown, reflecting its potential role in metastasis. Transfected cells were plated in 6-well plates and cultured until they reached 95% confluence. A sterile 200 μl pipette tip was used to create a scratch in the monolayer. After removing detached cells with PBS, images were acquired at 0 and 48 h and the wound area was analyzed using ImageJ to assess migration. Annexin V/propidium iodide (PI) staining Flow cytometry analysis of Annexin V/PI-stained cells was used to evaluate the impact of UBE2E2 silencing on apoptosis in SCLC cells. At 48 h post-transfection, DMS53 cells were harvested, digested, and stained with the Annexin V and PI apoptosis detection kit (Beyotime) for 20 min. Then the apoptotic populations were analyzed via flow cytometry, where early apoptosis was indicated by green fluorescence, and late apoptosis was marked by both green and red fluorescence. Statistical analysis Survival analysis was performed using the Kaplan–Meier method, with log-rank testing for comparison. Multivariate analysis was carried out using the Cox proportional hazards model. All statistical analyses were conducted using R (version 4.3.0) and GraphPad Prism. Data are presented as the mean ± standard deviation (SD). Differences between two groups were assessed by Student’s t-test, and comparisons among multiple groups were made using one-way ANOVA. Significance was defined as: ****p < 0.0001; ***p < 0.001; **p < 0.01; *p < 0.05. Results Cell type identification of SCLC using scRNA-seq Our study workflow is illustrated in Fig. [77]1. We analyzed the data and retained 5,025 cells for further examination (Supplementary Fig. 1A). The top 10 genes with highest SD values across all cells were identified, and their expression patterns are displayed in Supplementary Fig. 1B. After standardization, normalization, and dimensionality reduction through PCA and Harmony analysis, we performed UMAP to visualize the relationships between the 18 distinct clusters (Fig. [78]2A). These clusters were categorized into six cell types: Epithelial cells, Immune/Tumor/Natural Killer (I/T/NK) cells, B cells, Myeloid cells, Mast cells, and Fibroblasts cells (Fig. [79]2B). The expression of classical markers for these six cell types was illustrated in a bubble plot (Fig. [80]2C). Fig. 1. [81]Fig. 1 [82]Open in a new tab The flowchart of overall study Fig. 2. [83]Fig. 2 [84]Open in a new tab Annotation of cell types in the SCLC TME. A Using PCA-derived significant components, cells were clustered into 18 distinct groups via the UMAP algorithm. B Classification of the 18 clusters into six main cell types. C Dot plot showing the expression of marker genes across the six identified cell types Cellular communication reveals CAFs as core cells of the SCLC immune microenvironment Through the CellChat analysis, we identified complex ligand-receptor interaction networks within the SCLC tumor microenvironment (Fig. [85]3A). Although fibroblasts comprised a relatively small proportion of the cell types in the dataset, statistical analysis revealed that fibroblasts had the most frequent and extensive interactions with other cell types (Fig. [86]3B, [87]C). These findings implied CAFs orchestrate the immune landscape and contribute to the immunosuppressive TME of SCLC. Given this, we selected CAF functional signature genes for subsequent analysis, aiming to further explore their prognostic value and their impact on the immune microenvironment. Fig. 3. [88]Fig. 3 [89]Open in a new tab Cell–cell communication and contribution analysis. A Cell–cell interaction network. B The comparison of the total interaction network among six cell types. C Distribution of contribution scores for the six cell types Establishment of a risk model for predicting prognosis in SCLC Based on Lasso regression analysis of CAF-related functional genes (Fig. [90]4A–C), the prognostic risk model was established by the formula: [MATH: Risk \,Score=AKR1D1×-0.3708+HGF×-0.1907+UBE2E2×-0.09996+TBX4×-0.0850 :MATH] Fig. 4. [91]Fig. 4 [92]Open in a new tab Construction of the prognostic risk model. A Distribution of Lasso coefficients for prognostic genes. B Selection of the optimal lambda value for the Lasso model using tenfold cross-validation. C Gene coefficients in the model. D, E Kaplan–Meier survival analysis for the high-risk and low-risk groups in the GEO training and testing cohorts. F, G Time-dependent ROC curves for 3-, 4-, and 5-year survival in the training and testing cohorts, demonstrating the model’s predictive performance Kaplan–Meier survival analysis showed that high-risk patients had significantly shorter OS compared to low-risk patients in both the training (Fig. [93]4D) and testing cohorts (Fig. [94]4E). ROC curve analysis further demonstrated the model’s strong predictive performance, with AUC values of 0.974 and 0.791 for 3-year survival in the training and testing cohorts, respectively (Fig. [95]4F, [96]G), confirming its reliability in predicting SCLC prognosis. Recognition of independent risk factors and development of nomogram To validate the efficiency of the CAF functional gene risk model, processed survival data were retrieved from two external cohorts: the Geroge’s cohort and the TCGA-NSCLC cohort. Patients were clarified into clinical subgroups due to model-derived risk scores, Kaplan–Meier analyses were performed to evaluate differences in OS between two groups. The results demonstrated consistent findings across both external validation cohorts, with high-risk patients showing significantly shorter OS compared to low-risk patients, thereby confirming the robustness of the model (Fig. [97]5A, [98]B). Fig. 5. [99]Fig. 5 [100]Open in a new tab Validation of the prognostic model in external cohorts and Cox regression analysis. A Kaplan–Meier survival curve for George’s cohort. B Kaplan–Meier survival curve for the TCGA-NSCLC cohort. C Univariate Cox regression analysis to assess the prognostic significance of the risk score and clinical factors. D Multivariate Cox regression analysis highlighting the risk score as an independent prognostic factor in SCLC Additionally, univariate and multivariate analyses identified the risk score as an independent prognostic factor for SCLC (Fig. [101]5C, [102]D). The risk score was then integrated into a nomogram for individualized prognosis prediction, where logistic regression analysis demonstrated a significant contribution of the risk score to the nomogram’s scoring system (Fig. [103]6A). Fig. 6. [104]Fig. 6 [105]Open in a new tab Construction of the nomogram model. A Nomogram combining the risk score and clinical features to predict OS in SCLC patients. B Calibration curves showing the predicted OS at 1-year and 3-year intervals. C Time-dependent ROC curves illustrating the predictive accuracy. D DCA comparing the clinical utility of the risk score and other clinical features Furthermore, one-year and three-year OS predictions were evaluated using the nomogram, supported by the corresponding ROC and decision curve analysis (DCA) curves (Fig. [106]6B–D). In summary, these findings confirm the strong utility of this risk score to predict OS in SCLC patients. Deciphering immune infiltration patterns and mechanisms in SCLC TME via the CAF-related functional gene risk model To explore the relationship between the CAF-based risk score and the immune microenvironment, we conducted a comprehensive analysis of immune cell infiltration in high- and low-risk groups, aiming to uncover distinct immune dynamics. Our findings showed significant differences in immune cell composition between the two groups (Fig. [107]7A), with low-risk patients exhibiting higher levels of immune cell infiltration, including APC co-stimulation markers, chemokine receptors (CCR), checkpoint molecules, dendritic cells (DCs), immature DCs, mast cells, NK cells, and Th1 cells (Fig. [108]7B). The risk score not only reflects immune microenvironment characteristics but also correlates with several key signaling pathways. GSVA analysis identified DNA repair, glycolysis, and WNT signaling pathways enriched in the high-risk patients, which are associated with tumor aggressiveness as well as treatment resistance (Fig. [109]7C). Moreover, the low-risk group showed activation of immune-related pathways such as IL -17 signaling, cytokine-cytokine receptor interaction, and JAK-STAT signaling (Fig. [110]7D), indicating a more favorable immune landscape and potential for immunotherapy. Fig. 7. [111]Fig. 7 [112]Open in a new tab Immune infiltration and molecular mechanism analysis in of two risk groups. A Relative proportions of immune cell subtypes in two different groups. B Differences in 29 immune cells and molecules highlighted significant differences in the immune TME between groups. C GSVA analysis highlighting hallmark pathways associated with the risk score. D GSEA analysis of KEGG pathways demonstrating significant enrichment of specific signaling pathways linked to the risk score Validate the tumor-related functions of UBE2E2 in SCLC To validate the tumor-related functions of UBE2E2 in SCLC, we used the DMS53 cell line to study its role in a controlled environment. Cell lines like DMS53 are essential for cancer research, as they replicate key tumor characteristics and enable gene manipulation. This approach allows us to investigate cellular behaviors such as proliferation, migration, and apoptosis, shedding light on the molecular mechanisms of SCLC. To confirm its expression in SCLC tumor cells, we first performed H&E and IHC staining on patient tissue samples. The results revealed that UBE2E2 was highly expressed in the cytoplasm of SCLC tumor cells, with distinct staining observed at both 10 × and 40 × magnifications (Fig. [113]8A). The knockdown efficiency of UBE2E2 was validated in the DMS53 cell line using qRT-PCR and Western blot assays at 48 as well as 72 h post-transfection. Results confirmed a robust reduction in UBE2E2 expression transfected by siRNA sequences (Si-1 and Si-2) compared to the negative control, establishing the reliability of the silencing approach (Fig. [114]8B, [115]C). Following this, many in vitro experiments were performed to assess UBE2E2's functions in SCLC proliferation, migration, and apoptosis. The CCK-8 assay revealed a notable decrease in cell growth and viability in UBE2E2-silenced cells, compared to the control (Fig. [116]8D). In colony formation assays, silencing UBE2E2 also significantly reduced the colony formation and tumorigenic capacity of DMS53 cells (Fig. [117]8E), highlighting its essential role in supporting the proliferation of SCLC cells. In addition, wound-healing assays revealed that silencing UBE2E2 markedly impaired the migratory capacity of DMS53 cells, as reflected by a reduced migration area (Fig. [118]8F), emphasizing its role in cellular motility. Lastly, flow cytometry analysis of apoptosis further revealed that UBE2E2 knockdown promoted cell death, providing additional evidence of its influence on cell survival. UBE2E2 knockdown led to a marked increase in both early and late apoptotic cell populations (Fig. [119]8G), suggesting that UBE2E2 suppression triggers apoptotic pathways. Fig. 8. [120]Fig. 8 [121]Open in a new tab Functional validation of UBE2E2 knockdown in SCLC cells. A H&E and IHC staining at 10 × and 40 × magnification, demonstrating high expression of UBE2E2 in SCLC tumor cells. B Validation of UBE2E2 knockdown efficiency using qRT-PCR. C Western blot analysis confirming UBE2E2 protein knockdown efficiency after siRNA treatment. D CCK-8 assay demonstrating a significant decrease in cell viability in UBE2E2- knockdown cells. E Colony formation assay and corresponding quantification, showing reduced clonogenic capacity in UBE2E2-knockdown cells. F Wound healing assay and statistical analysis revealing impaired migratory capacity of DMS53 cells following UBE2E2 knockdown. G Flow cytometry analysis of apoptosis, including quantification of early and late apoptotic cells Collectively, these results establish UBE2E2 as a critical regulator of SCLC progression, driving tumor cell proliferation, migration, and resistance to apoptosis. Its prominent role in promoting the aggressive phenotype of SCLC highlights its potential as a therapeutic target, offering a promising avenue for interventions aimed at mitigating tumor aggressiveness and improving clinical outcomes for SCLC patients. Discussion This study presents the first comprehensive analysis in SCLC, demonstrating that CAF functional-related genes can be effective prognostic indicators. By developing a novel four-gene risk model based on CAF-related functions, we introduce a promising tool for prognostic stratification. This model can not only provide insight into OS but also unveil significant revelations about the immune landscape in SCLC. These findings mark an essential milestone in elucidating the multifaceted role of CAFs in this highly aggressive cancer type, where therapeutic options remain scarce and outcomes poor. The limited efficacy of immunotherapy in SCLC can be largely attributed to the unique characteristics of its immune microenvironment. Despite having a high tumor mutational burden (TMB), SCLC exhibits low immunogenicity, primarily due to impaired antigen presentation mechanisms. Such as selective downregulation of MHC class I and II molecules, which hinders the activation of T-cell-mediated anti-tumor immunity [[122]30, [123]31]. Furthermore, the tumor stroma in SCLC is highly fibrotic, resulting in sparse T-cell infiltration, with these immune cells often localized at the tumor periphery rather than within the core [[124]7]. The tumor is also infiltrated by various immunosuppressive cells, further compromising T-cell activity and enabling SCLC to evade T-cell-mediated anti-tumor effects without relying on immune checkpoints. These features underline why T-cell-based immunotherapies, which have been widely anticipated, struggle in SCLC [[125]4]. Traditional biomarkers like PD-L1 and TMB are insufficient for predicting patient prognosis or immunotherapy response. Consequently, there is an urgent need for novel prognostic models and biomarkers that are not dependent on T-cell activity but instead reflect the distinctive characteristics of the SCLC microenvironment. Although CAFs have long been recognized as central regulators of the TME in various malignancies [[126]32–[127]34], their precise contribution to SCLC progression has largely remained enigmatic. In our study, though CAFs are present in relatively low quantities in the SCLC TME, scRNA-seq and CellChat analyses reveal their profound influence as key mediators of intercellular communication. Interacting extensively with both immune and non-immune cells, CAFs act as pivotal hubs, orchestrating immune suppression, ECM remodeling, and facilitating tumor progression in highly dynamic, interdependent ways. This study introduces a novel perspective on the role of CAFs in SCLC and lays the groundwork for targeting CAF-mediated pathways as a therapeutic strategy. We used ssGSEA to decode the TME status of SCLC patients with different CAF function-related gene risk scores. The low-risk patients exhibited a more active TME, characterized by increased infiltration of DCs, NK, and Th1 cells, alongside robust activation of chemokine and cytokine signaling. This paints a picture of a more "immune-hot" TME [[128]35, [129]36]. In stark contrast, high-risk patients display an immune-suppressive TME with sparse immune cell infiltration. This phenomenon is likely due to ECM remodeling driven by CAFs [[130]8]. The physical and functional barrier directly promotes immune exclusion and SCLC evasion [[131]7, [132]37, [133]38]. Our findings align with current research on immune resistance in SCLC, where antigen presentation deficiencies are a hallmark of the disease. In the low-risk group, we observed higher expression levels of antigen-presenting cell (APC) co-stimulatory molecules and human leukocyte antigen (HLA), suggesting that these patients may be more likely to respond to immunotherapy [[134]38]. This observation underscores the importance of CAF-mediated immune modulation in shaping the immune landscape of SCLC. Pathway enrichment analysis further reinforced these immune landscape findings. In the low-risk group, GSVA and GSEA revealed significant enrichment in immune-related pathways, such as chemokine signaling and JAK-STAT signaling, both of which are critical for immune cell recruitment and activation. In contrast, the high-risk group showed salient enrichment in DNA repair, glycolysis, and WNT-β-catenin signaling pathways, all of which are associated with the aggressive biological traits of SCLC [[135]39–[136]41]. SCLC is one of the tumor types most reliant on DNA repair pathways due to its rapid proliferation and the loss-of-function mutations in TP53 and RB1. DNA repair helps ensure the survival of SCLC cells. The selective upregulation of glycolysis pathway is also associated with the harsh microenvironmental features of SCLC, enabling tumor cells to outcompete immune cells for growth in hypoxic and nutrient-restricted TME conditions. Finally, the classic WNT pathway, with an extremely high nonsynonymous mutation rate (up to 80%) in recurrent SCLC, directly drives chemoresistance in SCLC [[137]42]. It has been confirmed that WNT/β-catenin pathway could suppress the recruitment of DCs by downregulating CCL4, leading to impaired CD8 + T cell infiltration and activation. This process transforms "hot" tumors into "cold" tumors, representing a major mechanism by high tumor mutational burden tumors evade anti-tumor immunity. Our findings complement and extend previous research that links CAFs to tumor phenotype reprogramming and chemotherapy resistance in SCLC [[138]43]. The enrichment of DNA repair and WNT-β-catenin pathways in the high-risk group corroborates earlier studies on CAFs and treatment resistance [[139]44–[140]46]. Our work advances these earlier discoveries by providing a more nuanced analysis of how CAFs regulate immune and metabolic pathways within the TME. In this study, UBE2E2 was identified as a key gene within the CAF function-related risk model and was demonstrated to play a significant oncogenic role in SCLC. As a member of the E2 ubiquitin-conjugating enzyme family, UBE2E2 primarily facilitates the transfer of ubiquitin molecules to target proteins, thereby regulating their stability, localization, or function. Our in vitro experiments in the DMS53 SCLC cell line revealed that UBE2E2 knockdown significantly inhibited cell proliferation, migration, and clonogenic capacity while inducing apoptosis. These findings confirm the oncogenic role of UBE2E2 in SCLC and suggest that targeting UBE2E2 may represent a promising therapeutic strategy, particularly for high-risk patients with more aggressive tumor phenotypes. Beyond its direct role in tumor progression, UBE2E2 may drive SCLC advancement through multiple mechanisms, including cell cycle regulation, promotion of EMT, enhancement of DNA damage repair capacity, and activation of antioxidant pathways [[141]47]. However, further studies are needed to elucidate the precise molecular pathways through which UBE2E2 contributes to SCLC pathogenesis. Future research should also explore its potential interactions with the tumor microenvironment, particularly in the context of CAF-mediated immune modulation, to determine whether UBE2E2 acts as a bridge linking stromal dynamics with tumor aggressiveness. Despite the comprehensive nature of our study, certain limitations should be acknowledged. Although the CAF functional-related risk model was validated across multiple external cohorts, we were unable to directly link CAF-derived factors to immune modulation, limiting the depth of our mechanistic insights. Moreover, our in vitro experiments were conducted using a single cell line, which may not fully capture the heterogeneity present in SCLC tumors. Future studies should incorporate in vivo models and a larger, more diverse cohort of patients to expand and validate our findings. Furthermore, the precise molecular mechanisms through which UBE2E2 influences CAF function and immune suppression remain elusive and warrant further investigation. Conclusion In this study, we demonstrated for the first time that CAF functional-related genes can serve as reliable prognostic biomarkers for SCLC, providing new approaches for risk stratification and therapeutic targeting. Our findings highlight the critical role of CAFs in shaping the immune microenvironment and driving tumor progression. The CAF-based four-gene risk model not only provides a powerful tool for prognostic assessment, but also reveals the mechanisms of therapeutic resistance and immune escape. These findings suggest that CAFs and their associated genes are not only important markers of SCLC prognosis but also potential therapeutic targets. Future studies should further explore the mechanism of interaction between CAFs and the immune microenvironment and develop novel targeted therapies against CAF-driven pathways. By combining single-cell multi-omics technology and functional experiments, we are expected to open up new avenues for the precision treatment of SCLC, and ultimately improve the prognosis of patients. Supplementary Information [142]Supplementary file1 ^(1.8MB, pdf) [143]Supplementary file2 ^(1MB, pdf) [144]Supplementary file3 ^(15.2KB, docx) [145]Supplementary file4 ^(602.3KB, docx) Author contributions Conceptualization, Yun Fan; Data curation, Haicheng Wu, Wanchen Zhai and Yuwei Li; Formal analysis, Yehao Yang; Funding acquisition, Hui Li and Yun Fan; Methodology, Yunfei Chen and Qian Zhang; Resources, Haicheng Wu, Wanchen Zhai and Yuwei Li; Software, Linjing Zhou; Supervision, Hui Li and Yun Fan; Validation, Yunfei Chen and Yunfeng Tong; Visualization, Xinyuan Ye; Writing—original draft, Yunfei Chen and Yunfeng Tong; Writing—review & editing, Hui Li and Jing Sun. Funding This work was supported by the National Natural Science Foundation of China [81972718] and the Natural Science Foundation of Zhejiang Province [LY22H160037] and the Beijing Science and Technology Innovation Medical Development Foundation [KC2023-JX-0186-PZ090]. Data availability All data generated or analyzed during this study are included in this article. Further enquiries can be directed to the corresponding author. Declarations Ethics approval and consent to participate This study was approved by the Institutional Review Board of Zhejiang Cancer Hospital (Approval No. IRB-2024-301) and conducted in accordance with the Declaration of Helsinki. Written informed consent was waived by the ethics committee due to the retrospective nature of the study and the use of anonymized archival specimens. Consent for publication All authors agree to publish. Competing interests The authors declare no competing interests. Footnotes Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Yunfei Chen and Yunfeng Tong contributed equally. References