Introduction

Renal cell carcinoma (RCC), a leading cause of global cancer-related deaths, is widely recognized as a significant threat1. Clear cell RCC (ccRCC) constitutes 80% of renal malignancies, making it the predominant type of kidney cancer2,3. ccRCC is a highly vascularized subtype of renal cell carcinoma, a characteristic that drives its high invasiveness4. This vascularization facilitates metastasis to organs such as the lungs, liver, and bones through hematogenous spread5. At the time of initial diagnosis, metastasis is observed in 25–30% of patients6. Metastatic RCC has limited treatment options due to its resistance to conventional radiation and chemotherapy. As a result, molecular targeted therapy and immunotherapy have become the primary treatment approaches for advanced kidney cancer in recent years4. However, due to the heterogeneity of ccRCC, some patients still show resistance to these regimens7,8. The overall prognosis for patients with ccRCC remains poor, particularly for those with advanced clinical stages, which imposes a substantial societal burden. Therefore, it is imperative to effectively impede the metastatic progression of ccRCC and to develop targeted interventional therapies during the pre-metastatic stage to enhance patient outcomes. Achieving this objective necessitates comprehensive exploratory studies to elucidate the metastatic process and its molecular mechanisms, with the aim of identifying effective therapeutic targets.

It is well known that metastasis is a complex process and circulating tumor cells (CTCs) undoubtedly play a crucial role in this process9. However, the regulatory factors leading to the production of CTCs by primary tumors are not clear10,11. In recent years, work on cell lineage reprogramming has been heavily investigated, leading to the discovery of transcription factors capable of inducing pluripotency in somatic cells12,13, as well as the important mechanism related to metastasis10,14. It has been shown that cell reprogramming leading to adherent-to-suspension transition (AST), between adherent (A) and suspended (S) cell types is one of the key mechanisms during the dissemination of CTCs15. At present, there is a lack of studies addressing the significant role of AST, an emerging tumor reprogramming process, in the progression and metastasis of ccRCC. A thorough characterization of AST factors within ccRCC is essential to enhance our understanding of the critical importance of AST in this context. Concurrently, the development of prognostic models and the investigation of molecular mechanisms may offer novel insights for clinicians, facilitating improved prognostic evaluations and the development of innovative therapeutic strategies for patients.

In this study, we performed a comprehensive analysis of the expression levels and genomic variations of 20 AST factors in human tumors. Through detailed analysis, we categorized ccRCC patients into four subgroups based on AST factor scores and expression levels, investigating variations in prognosis, immune microenvironment, and therapeutic responsiveness among these groups. Furthermore, we formulated an innovative prognostic model specifically tailored for ccRCC patients to better assess the patients’ prognosis. Finally, we identified the SPIB-SAA1-AKT pathway as one of the potential mechanisms by which AST factors promote ccRCC metastasis in combination with differential and enrichment analysis, and verified it by single-cell sequencing, spatial transcriptome, and cellular experiments. In conclusion, this study thoroughly assessed the role of AST factors in ccRCC and proposed potential molecular mechanisms underlying its metastasis. Our study aims to provide a foundational understanding that facilitates clinicians to gain insight into ccRCC metastatic mechanisms.

Results

Pan-cancer characteristics of AST factors

To enhance understanding of the genetic traits, expression patterns, and prognostic importance of AST factors, an extensive analysis was carried out across multiple human tumors. Initially, the occurrence of copy number variation (CNV) mutations in 20 AST factors was recorded (Fig. 1A, B). A notable prevalence of amplification CNV mutations was identified in ACC, KIRC, and UCS, involving a substantial proportion of the AST factors. It is important to highlight that there were notable increases in the frequencies of amplification CNV mutations for prominent genes such as BTG2, IKZF1, and IKZF5 across different types of malignancies. Meanwhile, the frequencies of deletion CNV mutations were notably higher in HNSC and MESO. Additionally, most AST factors exhibited a tendency toward low-frequency single nucleotide variant (SNV) mutations (Fig. 1C), with elevated SNV occurrences observed only in SKCM and UCS. Considering that abnormal gene methylation is a recurrent epigenetic feature in tumorigenesis16, we conducted an investigation into the methylation patterns of AST factors across various cancer types (Fig. 1D). Significant differences in the methylation status of AST factors were observed between tumor and normal tissues. Among them, AKAN and SPI1 showed consistent hypermethylation across most cancer types.

Fig. 1
figure 1

Pan-cancer analysis of 20 AST factors. (A) Depiction of amplification CNV mutations of 20 AST factors in human tumors. (B) Depiction of deletion CNV mutations of 20 AST factors in human tumors. (C) The profile of SNV data of 20 AST factors in human tumors. (D) The methylation status of 20 AST factors in human tumors.

Full size image

Subsequently, a comparative analysis was conducted to evaluate the expression patterns of 20 AST factors in tumor and normal samples, revealing significant differences in the expression of most AST factors between the two tissue types (Fig. 2A). Notably, a substantial increase in expression levels was observed in 16 out of 20 AST factors in KIRC samples. To identify genes with potential prognostic implications, univariate Cox regression analysis was employed (Fig. 2B). It is worth noting that the prognostic relevance of AST factors differs among various cancer types. In ACC, LGG, and LIHC, the majority of AST factors were found to be associated with increased risk for patient prognosis. Conversely, in the instances of CESC, HNSC, LUAD, and SARC, the majority of AST factors were observed to be protective in nature.

Fig. 2
figure 2

Expression pattern and survival landscape of 20 AST factors. (A) Expression pattern among 20 AST factors in human tumors. (B) Survival landscape among 20 AST factors in human tumors.

Full size image

Consensus clustering analysis based on AST factors

Given the potential involvement of AST factors in ccRCC development and their possible underlying mechanisms, we conducted an exploratory study utilizing molecular typing to classify ccRCC patients based on 20 AST factors. Using the consensus clustering algorithm, we classified ccRCC patients from the TCGA-KIRC cohort into four distinct subgroups: cluster 1, cluster 2, cluster 3, and cluster 4 (Fig. 3A–C). Notably, we observed a higher proportion of patients with low stage and grade in cluster 2 (Fig. 3D), and this subgroup exhibited a significantly prolonged overall survival (OS) (Fig. 3E). The analysis focused on identifying DEGs between cluster 2 and the other three clusters to investigate potential mechanisms associated with poor prognosis (Fig. 3F). The DEGs primarily consisted of highly expressed genes in cluster 2, such as APOLD1, CA4, and FOSB. Additionally, SAA1 was identified as one of the few genes significantly under-expressed in cluster 2, indicating its potential role as a downstream target of AST in facilitating cancer-related effects. Furthermore, enrichment analysis of the DEGs highly expressed in cluster 2 revealed a significant enrichment of cell adhesion-related pathways (Fig. 3G, H).

Fig. 3
figure 3

Identification of AST-related subtype. (A) Consensus clustering cumulative distribution function (CDF) curve. (B) Relative change in the area under the CDF curve when cluster number varying from k − 1 to k. (C) Consistency of clustering results heatmap (k = 4). (D) The distribution of clinical characteristics among different groups. (E) Kaplan–Meier survival analysis of the different groups. (F) Volcano plot presenting the DEGs between cluster 2 and the other three clusters. (G,H) KEGG and GO Enrichment analysis of upregulated DEGs in the cluster 2.

Full size image

Correlations between the immune microenvironment and subclusters

Initially, the study aimed to investigate the associations between AST factors GSVA scores and immune cell infiltrations. The results revealed significant associations between GSVA scores and the infiltration levels of various immune cells, indicating that AST factors might influence immune-related pathways in KIRC patients and potentially affect their clinical prognosis (Fig. 4A). Given the complexity of this mechanism, further research is necessary to fully understand its specific mechanisms. Consequently, a comprehensive analysis was performed using multiple computational tools to examine variations in immune cell infiltration levels (Fig. 4B). Significantly, cluster 1 exhibited the most pronounced infiltration of immune cells, whereas cluster 2 displayed the highest infiltration of endothelial cells. We subsequently analyzed the correlation between endothelial cell infiltration and prognosis in patients with ccRCC, and the findings aligned with the previously observed trend (Fig. 4C, D). Given the crucial function of immune checkpoints in facilitating tumor cells’ evasion of immune surveillance, our study also involved the analysis of immune checkpoint genes (ICGs) across different clusters (Fig. 4E). Our results showed that almost all 8 ICGs were significantly higher expressed in cluster 1 and significantly lower expressed in cluster 3. Based on the synthesis of the observations, it is reasonable to deduce that AST factors exert a substantial regulatory impact on the immune microenvironment in ccRCC, thus presenting the possibility of improving prognostic outcomes for affected individuals.

Fig. 4
figure 4

Correlations between the AST factors and immune microenvironment. (A) The bubble plot underscored the correlation between immune cell infiltrates and AST factors GSVA enrichment score. (B) Heatmap showed the immune cell infiltration levels by TIMER, CIBERSORT, EPIC, and the XCELL algorithm among different clusters. (C) Kaplan–Meier survival curves were generated between the high and low Endothelial cell (EPIC) infiltration groups. (D) Kaplan–Meier survival curves were generated between the high and low Endothelial cell (Xcell) infiltration groups. (E) The box plots encompassed the expression distribution of immune checkpoints genes among different clusters.

Full size image

Drug sensitivity analysis across the subclusters

Considering the significant role of molecular targeted therapy in the treatment of ccRCC, using highly precise pharmaceutical agents customized to the distinct characteristics of therapeutic targets exhibits substantial potential. Stemness, being a crucial mechanism of drug resistance in tumors17, was initially evaluated in subclusters to ascertain any variations. The findings revealed notable disparities in stemness levels across four subclusters with the mRNA stemness index18 (Fig. 5A). Consequently, it can be inferred that there may also exist dissimilarities in drug resistance among these subgroups. We estimated the IC50 concentration values for individual KIRC patients, enabling the formulation of personalized treatment strategies (Fig. 5B–L). Of the eleven drugs analyzed, some were commonly used in renal cancer targeted therapy, while others were chemotherapeutic agents. Our results reveal that cluster 1 is more sensitive to chemotherapeutic agents, including cisplatin, gemcitabine, 5-fluorouracil, vorinostat, and vinorelbine. In terms of targeted agents, cluster 1 shows higher sensitivity to ruxolitinib, nilotinib, and gefitinib, while cluster 2 demonstrates greater sensitivity to axitinib and savolitinib. Additionally, cluster 3 exhibits increased sensitivity to lapatinib. The results provide insight into the potential of AST factors as valuable indicators for guiding the design of targeted therapies against ccRCC.

Fig. 5
figure 5

Drug sensitivity analysis. (A) The box plot illustrates the mRNA stemness index (mRNAsi) pattern among different clusters. (B–L) The box plot illustrates the IC50 values of 11 commonly used drugs (cisplatin, gemcitabine, 5-Fluorouracil, vorinostat, vinorelbine, ruxolitinib, nilotinib, axitinib, savolitinib, lapatinib and gefitinib) among different clusters.

Full size image

AST-related prognostic risk model development and validation

Based on the discovery that AST factors hold significant value in ccRCC, we established an AST-related prognostic risk model to enhance the prediction of patient prognosis. We performed dimensionality reduction and constructed prognostic models based on the LASSO regression algorithm in the TCGA-KIRC cohort. 10-fold cross-validation results show that the model is optimal when the optimal adjustment parameter is 0.0411 (Fig. 6A, B). The final generated prognostic model equation is as follows: Risk score  = (− 0.0428) * TAL1 + (− 0.1521)  * KLF2 + (− 0.1521) * BTG2 + (0.2068) * POU2F2 + (0.1304) * SPIB + (− 0.0365) * EAF2  + (− 0.0203) * IRF8. After determining the optimal risk score threshold, patients were classified into high-risk and low-risk groups (Fig. 6C). OS was significantly shorter in the high-risk group compared to the low-risk group (Fig. 6D). The AUC values for 1-, 3-, and 5-year survival were 0.71, 0.70, and 0.70, respectively (Fig. 6G). The predictive validity of the risk model was additionally confirmed through validation with the independent E-MTAB-1980 cohort and NIHMS1611472 dataset (Fig. 6E and Supplementary Fig. 1A–C). Survival analysis revealed a consistent trend, with the high-risk group exhibiting significantly shorter OS (Fig. 6F). In the validation cohort, the AUC values for the 1-, 3-, and 5-year periods were all 0.78 (Fig. 6H). Collectively, these findings underscore the diagnostic efficacy of the risk model established in our study.

Fig. 6
figure 6

Establishment and validation of the AST-related prognostic risk model. (A) LASSO coefficient profiles of the AST factors. (B) The optimal tuning parameter in the LASSO model. (C) Scatterplot and heatmap presenting the risk score, survival time, survival status and 7 AST factors expression of the training set (TCGA-KIRC cohort). (D) KM curves between the high-risk group and low-risk group in the training set. (E) Scatterplot and heatmap presenting the risk score, survival time, survival status and 7 AST factors expression of the test set (E-MATB-1980 cohort). (F) KM curves between the high-risk group and low-risk group in the test set (E-MATB-1980 cohort). (G) The 1-year, 3-year and 5-year ROC curves in the training set. (H) The 1-year, 3-year and 5-year ROC curves in the test set.

Full size image

The potential AST factors-driven mechanism to promote the progression of CcRCC

Next, we analyzed the DEGs between the high-risk and low-risk groups to explore potential mechanisms underlying poor prognosis. Using a fold change cutoff of 2 and a P-value threshold of 0.05, 1607 genes were upregulated and 344 genes were downregulated in the high-risk group (Fig. 7A). SAA1 was identified as a significantly upregulated gene in the high-risk group, consistent with our previous findings (Fig. 3F). These results substantiate our earlier conjecture. Enrichment analysis of upregulated DEGs revealed significant pathways related to cell migration and immune response, such as acute inflammatory response and cell migration (Fig. 7B). Conversely, enrichment analysis of downregulated DEGs revealed significant pathways involving ion transport, ECM-receptor interaction, and calcium signaling (Fig. 7B). These results also demonstrate that AST factors may promote tumor progression by affecting cell adhesion status.

Fig. 7
figure 7

The potential mechanisms behind AST-driven ccRCC progression. (A) Volcano plot presenting the DEGs between the high-risk and low-risk groups. (B,C) KEGG and GO Enrichment analysis of upregulated and downregulated DEGs in the high-risk group. (C–I) The expression correlation of SAA1 and 7 AST factors was analyzed with Spearman test. (J) Transcription factor binding sites predicted by JASPAR. (K) CUT&RUN-qPCR showed that SPIB could bind to the SAA1 promoter.

Full size image

SPIB contributes to CcRCC migration by AST via promoting SAA1 transcription

Drawing on previous findings, we have identified SAA1 as a prospective downstream target. In order to enhance our understanding of the regulatory mechanisms governing SAA1 and AST factors, we conducted a correlation analysis between the seven AST factors encompassed in the prognostic model and SAA1 (Fig. 7C–I). Our findings indicate a significant positive correlation solely between SAA1 and SPIB as well as POU2F2. Subsequently, we investigated the potential binding sites for SPIB and POU2F2 within the SAA1 promoter region, using the JASPAR database (Fig. 7J). Notably, the relative score of SPIB binding to the SAA1 promoter region reached the maximum value of 1.0, indicating that SPIB might contribute to facilitating SAA1 transcription by modulating its promoter activity. Consistent with this hypothesis, CUT&RUN-qPCR analysis confirmed that SPIB can bind to the SAA1 promoter (Fig. 7K). To substantiate our findings, we further corroborated the aforementioned results using single-cell and spatial transcriptome sequencing data (Fig. 8A–F). Notably, a substantial overlap was observed between the regions exhibiting high expression of SPIB in ccRCC and those displaying elevated expression of SAA1, thereby strengthening our initial hypothesis. Subsequently, we validated the significantly elevated expression of SAA1 in ccRCC tumor samples through Western blotting, immunohistochemistry, and qPCR (Fig. 8G–I).

Fig. 8
figure 8

SPIB contributes to AST by promoting SAA1 transcription. (A) The distribution and annotations of single cells labelled by cluster identity. (B,C) Expression profiles of SPIB and SAA1 for each cell based the single-cell transcriptome sequencing data. (D) HE staining of ccRCC tissues corresponding to the spatial transcriptomics results. (E,F) Expression profiles of SPIB and SAA1 for each cell based the spatial transcriptome sequencing data. (G) The expression of SAA1 between ccRCC tissues and normal tissues using western blotting (5 pairs). (H) Representative images of SAA1 immunohistochemistry. (I) The expression of SAA1 between ccRCC tissues and normal tissues using RT-qPCR (5 pairs). (J) Expression of SAA1, phosphorylated AKT, AKT, phosphorylated mTOR, mTOR, β-actin detected after transfection with siRNA in Caki-1 cell lines using western blotting. (K) Representative cell images of si-NC and si-SAA1 cells in culture plate and media. Original blots are presented in Supplementary Fig. 2.

Full size image

In light of previous enrichment analysis findings and the impact of AST factors on cellular status, the effect of SAA1 on AST was further substantiated at the cellular level. To achieve this, we utilized siRNAs specifically designed to target SAA1 and suppress its expression in Caki-1 cells. The effectiveness of SAA1 knockdown was validated using western blotting (Fig. 8J). Subsequently, a significant reduction in the number of suspended cells in the media and a decline in AST efficiency were observed following SAA1 knockdown (Fig. 8K). Furthermore, the AKT-mTOR pathway, a well-established signaling cascade critical for cell migration, exhibited significantly reduced activity upon SAA1 knockdown (Fig. 8J). Collectively, our study demonstrates that SAA1 knockdown inhibited the AKT signaling pathway, which subsequently enhanced ccRCC cell adhesion.

Discussion

Metastasis, a formidable characteristic of cancer, continues to pose a significant challenge in enhancing clinical outcomes for cancer patients, despite notable advancements in cancer diagnosis and treatment facilitated by the continuous progress of medical technology. This predicament is particularly pronounced in the context of RCC19. The surgical intervention for early stage RCC demonstrates significant efficacy, leading to improved prognosis for patients. Conversely, the therapeutic options available for metastatic RCC are constrained20, with metastasis emerging as a primary determinant of unfavorable clinical prognoses in individuals diagnosed with RCC2. Extensive research has been conducted on the intricate biological mechanisms underlying tumor cell metastasis, including epithelial-mesenchymal transition (EMT)21, the formation of circulating tumor cells (CTCs)22, and the immune microenvironment23. Cancer metastasis is a multifaceted and sequential process encompassing invasion, intravasation, extravasation, and eventual attachment and colonization24,25. CTC formation undeniably assumes a pivotal function in the metastatic process26. CTCs are defined as malignant cells that detach from a solid tumor mass and enter the blood circulation27. Despite their origin from the primary tumor, CTCs exhibit distinct EMT attributes and augmented with stemness traits, enabling them to evade immune-mediated elimination and establish distant colonies more effectively28,29,30. Recent research has revealed that an emerging biological phenomenon, referred to as AST, significantly contributes to the development of CTCs15. The AST factors trigger a spontaneous cellular rounding and detachment from the extracellular matrix, thereby transforming adherent cells into suspension cells. Furthermore, investigations have demonstrated the pivotal involvement of AST in the metastasis of breast cancer and melanoma15.

The role of AST in ccRCC progression, as well as potential therapeutic strategies targeting AST-related factors, remains underexplored. In this study, we performed a novel and comprehensive investigation into the significance of AST factors in human tumors across the pan-cancer spectrum. Our analysis encompassed various factors such as CNV, SNV, expression profiles, prognostic significance, and methylation patterns. Through this systematic investigation, we uncovered the critical role of AST factors in various tumor types, where they either promote or inhibit tumorigenesis depending on the specific cancer context. Considering the scarcity of research on AST factors in human tumors, our study offers valuable insights and guidance for future investigations in this expansive pan-cancer domain. Simultaneously, our study has yielded intriguing results. Notably, a substantial upregulation of AST factors was observed in tumor tissues of ccRCC patients, with the highest prevalence observed across various cancer types. This observation suggests a potential distinctive involvement of AST in ccRCC patients.

Therefore, we prioritized ccRCC as the central focus of our further investigation. Employing cluster analysis, we successfully classified the ccRCC samples into four distinct subgroups, wherein patients belonging to cluster 2 exhibited a noteworthy extension in survival duration. We conducted a differential and enrichment analysis of DEGs between cluster 2 and the remaining three clusters to delve deeper into potential mechanisms. Our findings revealed SAA1 was found to be one of the few genes which were significantly reduced in cluster 2, thereby highlighting SAA1 as a potential downstream target of interest. Given that immunotherapy and targeted therapies are presently the primary therapeutic modalities for advanced ccRCC31,32, our study aimed to examine the immune microenvironment, the expression of immune checkpoint molecules33,34, and the sensitivity to classical targeted and chemotherapeutic agents35 across various subgroups. Notably, cluster 1 demonstrated the most substantial infiltration of immune cells, while cluster 2 exhibited the highest infiltration of endothelial cells. Furthermore, all eight ICGs were significantly upregulated in cluster 1. This may directly influence the responsiveness of various clusters to immunotherapy. The analysis of sensitivity predictions for targeted and chemotherapeutic agents revealed distinct sensitivity patterns across different clusters. Notably, patients in cluster 1 exhibited heightened sensitivity to common chemotherapeutic agents, whereas the sensitivity to common targeted therapeutic agents varied within each cluster. These observations underscore the significance of personalized treatment approaches and patient stratification based on AST factors.

Based on the above findings, it can be concluded that AST factors significantly impact ccRCC progression. However, due to the inherent molecular heterogeneity of AST factors, the precise categorization of each individual factor does not consistently enable accurate prognostication of patient outcomes. As a result, we established a risk prediction model based on AST factors to accurately predict the survival duration of ccRCC patients. The validation of the risk model’s robustness and accuracy was performed in the TCGA-KIRC cohort and the E-MATB-1980 cohort. The results of our survival analysis indicate a significant disparity between high-risk and low-risk groups. In the context of clinical decision-making, the prognosis of patients with ccRCC is primarily influenced by pathological stage and grade. Nonetheless, tumors with identical clinical stages and grades do not consistently exhibit the same prognosis. Our risk model seeks to provide an innovative approach for predicting the prognosis of patients with ccRCC. Furthermore, differential analysis reveals that patients with varying prognostic scores display significantly different gene expression profiles, suggesting a pathway for personalized treatment strategies. Additionally, we compared our prognostic model, which is based on AST factors, with five other recently published prognostic models. The AUC-ROC values of the studies assessed ranged from 0.70 to 0.7736,37,38,39,40. In contrast, our model achieved a maximum ROC value of 0.78, underscoring its superior predictive performance in clinical decision-making for patient prognosis.

Further investigations were conducted to explore the potential mechanism facilitated by AST in the progression of ccRCC. It was observed that SAA1 exhibited significant up-regulation in the high-risk group, which coincided with the downstream targets identified during the risk stratification process. Given that all AST factors are transcription factors, our inquiry revolves around whether AST factors induce AST by regulating the transcriptional activity of SAA1. By examining the JASPAR database41 and conducting CUT&RUN-qPCR analysis, we have discovered that SPIB can bind to the SAA1 promoter. Additionally, the outcomes of the differential gene enrichment analysis, conducted between high and low risk groups, have revealed significant enrichment of various extracellular matrix-related pathways. Hence, it was postulated that AST factors might impede the adhesion of tumor cells through the facilitation of SAA1 transcription, consequently heightening the likelihood of metastasis. Finally, additional experiments were conducted to partially validate the findings in ccRCC. Key gene expressions were first confirmed through single-cell sequencing and spatial transcriptome data42. Following this, we conducted an evaluation of the modifications in cellular adhesion capabilities, as well as the impacts on the AKT signaling pathway43, through the knockdown of SAA1 in the Caki-1 cell line. Our results suggest that SAA1 induces AST and regulates the invasive potential of tumor cells by means of the AKT pathway.

It is important to note that this study has certain limitations. First, the data used was sourced from publicly available databases, without including data generated by our own research facility. The absence of in vivo experiments to validate these findings should be acknowledged. In future investigations, we aim to reinforce the current findings and explore potential mechanisms.

In conclusion, this study provides a detailed analysis of the critical role and prognostic significance of AST factors in ccRCC progression. Additionally, it elucidates the underlying pathways and mechanisms through which AST factors can impact the clinical outcomes of ccRCC. Notably, this study proposes innovative risk stratification approaches and a prognostic model for ccRCC, offering valuable insights to advance clinical strategies in its management.

Methods

Data collection and processing

All 20 AST factors were identified based on the references15. Transcriptional and clinical prognostic data for the TCGA-KIRC cohort were retrieved from the TCGA official website, supplemented with additional samples from the E-MTAB-1980 cohort in the ArrayExpress database. The NIHMS1611472 dataset was sourced from Toni K Choueiri44. Pan-cancer genomic and transcriptomic data were also obtained from the TCGA database. The genome sequence data used in this study is accessible via the European Genome-Phenome Archive: EGAD00001008030 for single-cell RNA sequencing data and EGAD00001008781 for spatial transcriptomic data42. Analysis of single-cell RNA sequencing and spatial transcriptomic data was conducted using the online portal provided by the Wellcome Sanger Institute (https://www.sanger.ac.uk/).

Pan-cancer analyses of AST factors

The fold change (FC) value was first calculated to assess differences in gene expression between tumor and normal tissues. Next, univariate Cox regression analysis was conducted to evaluate the prognostic significance of each gene. CNVs, including both amplifications and deletions, were then quantified. For SNVs, mutation frequency was determined by dividing the number of samples with SNVs by the total sample count. The results were visualized as a heatmap using the ggplot2 package. Survival analysis was carried out with the “forestplot,” “survival,” and “survminer” packages.

Molecular subtyping of CcRCC based on AST factors

To characterize the expression patterns of AST factors in ccRCC, a methodology aligned with prior research was employed to identify distinct molecular subtypes. The “ConsensusClusterPlus” package was used for consensus clustering analysis, “survival” for survival analysis, “Limma” for differential expression analysis, and “ClusterProfiler” for enrichment analysis.

Immune related analysis and therapeutic sensitivity analysis

GSVA was utilized to estimate immune cell infiltration for each ccRCC patient. The immune microenvironment across different molecular subtypes was further examined by analyzing immune cell infiltration and immune checkpoint expression levels. Multiple algorithms, including XCELL, CIBERSORT, EPIC, and TIMER, were employed to evaluate immune cell infiltration abundance. Drug sensitivity among various KIRC subtypes was analyzed using the “oncoPredict” package. Correlation heatmaps, bubble plots, and scatter plots were generated with the “ggplot2” package, and statistical differences were determined using the Kruskal–Wallis test.

Development and validation of AST factors related prognostic model

In the TCGA-KIRC cohort, the development of the AST factors related prognostic risk model was carried out using the LASSO regression algorithm and Cox regression analysis. LASSO regression is recognized for its capability to eliminate non-essential variables by penalizing the magnitude of the regression coefficients. This method is particularly effective in addressing multicollinearity among variables. LASSO achieves this by shrinking the coefficient estimates towards zero, with the extent of shrinkage governed by a tuning parameter, λ. To ascertain the optimal value of λ, a ten-fold cross-validation procedure was employed, selecting λ based on the minimum criterion. The model was validated using the E-MTAB-1980 cohort and the NIHMS1611472 dataset. Patients were divided into high-risk and low-risk groups based on the median risk score in the training set, with similar methods applied for internal validation. The LASSO regression algorithm was implemented using the R package “glmnet,” while Cox regression analysis was conducted with the “survival” package. Time-dependent receiver operating characteristic (ROC) curves were employed to evaluate and compare the predictive performance of the risk score.

Transcription factor binding site analysis

The 2000 bp regions upstream of the genes were selected as candidate promoter sequences. These sequences, presumed to contain promoters, were obtained from the National Center for Biotechnology Information (NCBI). Transcription factor binding sites were predicted using the JASPAR tool, with the binding probability (relative score) set at 0.9.

Cell culture and antibody

The Caki-1 cell lines were supplied by the Tianjin Institute of Urology (Tianjin, China). Cell culture was performed in accordance with ENCODE standards. Primary antibodies, including anti-SAA1 (A1655), anti-Phospho-AKT1-S473 (AP0637), anti-AKT (A17909), anti- p-mTOR-S2448 (AP1413), anti-mTOR (A25581) and anti-β-actin (AC026), were prepared by diluting them as per the manufacturer’s instructions (ABclonal).

SiRNA and transfection

The siRNA targeting SAA1, along with its corresponding control, was supplied by JiKai Gene (Shanghai, China) and transfected into cells using the Lipofectamine 6000 reagent (C0526, Beyotime).

Western blotting

Total proteins were extracted using RIPA buffer. Proteins were resolved on 10% polyacrylamide gels and subsequently transferred to PVDF membranes via SDS-PAGE. The membranes were incubated overnight with diluted primary antibodies, followed by a 30-min incubation with secondary antibodies.

Induction of AST

Caki-1 cells were seeded in 24-well culture plates. Representative cell images of adherent and suspended cells were taken using the microscope separately.

Real-time quantitative PCR (RT-qPCR)

Total RNA was isolated using the SPARKeasy RNA kit, and first-strand cDNA was synthesized using a Multi-Block thermal cycler. RT-qPCR was performed with FastStart Universal SYBR Green Master kits, using GAPDH as the reference for normalization. The primers used for RT-qPCR were as follows:

  • SAA1-F: ATCACTTCCGACCTGCTG; SAA1-R: CCTCTCCCCGCTTTGTA;

  • GAPDH-F: ACAACTTTGGTATCGTGGAAGG; GAPDH-R: GCCATCACGCCACAGTTTC.

Immunohistochemical analysis

Tumor tissues embedded in paraffin were sectioned and immunostained using the streptavidin-peroxidase method following antigen retrieval through heat treatment. Detection was performed with 3,3′-diaminobenzidine and hematoxylin, and the slides were analyzed under a microscope.

CUT&RUN-qPCR

CUT&RUN-qPCR was conducted utilizing the Hyperactive pG-MNase CUT&RUN Assay Kit for PCR/qPCR (HD101). All experimental procedures adhered strictly to the manufacturer’s instructions. The antibody targeting SPIB was procured from Proteintech (Catalog No. 15768-1-AP). The Fold Enrichment method was employed for calculations, with Spike-in DNA serving as an internal reference. The primers used for CUT&RUN-qPCR were as follows: SAA1-Pro-F: CCAGGCACATCTTGTTCCCT and SAA1-Pro-R: AGTCCCTGCAGGTCATTTCC.

Statistical analysis

Statistical analyses were conducted using SPSS 23 and R software v4.03. Differences in gene expression levels, risk scores, and immune cell infiltration abundance between groups were evaluated using the Wilcoxon test or Kruskal–Wallis test. Correlations were assessed with Spearman’s correlation test. A p-value below 0.05 was considered statistically significant.