MYO6+Epithelial Subpopulation as a Prognostic Hub in Lung Adenocarcinoma Identified by Multi-Omics Integration

MYO6+Epithelial Subpopulation as a Prognostic Hub in Lung Adenocarcinoma Identified by Multi-Omics Integration | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article MYO6 + Epithelial Subpopulation as a Prognostic Hub in Lung Adenocarcinoma Identified by Multi-Omics Integration Miaoyan Liu, Houqiang Li, Shenghan Xu, Yike Zhou, Tiegang Cao, and 3 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7742292/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 7 You are reading this latest preprint version Abstract Background Lung adenocarcinoma (LUAD) exhibits profound cellular heterogeneity, with epithelial subpopulations playing pivotal yet incompletely characterized roles in tumor progression. Here, we sought to delineate the key heterogeneous epithelial subpopulations and to uncover the molecular determinants that govern their influence on LUAD prognosis. Methods We integrated two publicly available single-cell RNA-seq datasets derived from normal lung and LUAD tissues. Epithelial cells were extracted and subpopulations annotated via Seurat-based clustering. Transcriptional dynamics during malignant transformation were quantified using Monocle3 trajectory inference and ternary plots. A machine-learning framework comprising 101 algorithmic combinations screened signature genes from poorly differentiated clusters (11/26). Prognostic models were constructed and validated in TCGA cohorts. MYO6⁺ epithelial cells were functionally characterized through differential-expression analysis and GO/KEGG pathway enrichment.Verification of MYO6 mRNA levels and protein expression levels by quantitative real-time PCR (qRT-PCR) and Western blot. Results Single-cell profiling demonstrated that LUAD epithelial cells transition from lineage-specific identities (AT1, AT2, airway) to a dedifferentiated state. Cluster 11/26 displayed low lineage-specific gene scores and selective enrichment in LUAD. Machine learning identified MYO6, ASPH, and KRT8 as core prognostic genes among 140 intersecting markers. The StepCox[forward]+Ridge model reliably stratified patients into high- and low-risk groups (HR = 2.48, p < 0.001; 3-year AUC = 0.82). MYO6⁺ epithelial cells were spatially distinct, exhibited MAPK-pathway activation, and correlated with proliferative signatures and adverse outcomes. Compared with 16HBE, MYO6 mRNA was significantly elevated in A549 and H1299 cells; LUAD tissue MYO6 protein levels were also significantly higher than those in adjacent non-cancerous tissue. Conclusion We identify a low-lineage-score epithelial subpopulation (Cluster 11/26) in LUAD whose attenuated lineage-specific gene expression is linked to malignant progression. A StepCox+Ridge prognostic model, built upon core genes derived from this subpopulation, delivers significant clinical stratification. MYO6⁺ cells, a central constituent of this subpopulation, propel tumor progression via MAPK signaling and represent a promising target for precision therapy. Health sciences/Biomarkers Biological sciences/Cancer Biological sciences/Computational biology and bioinformatics Health sciences/Oncology Single-cell RNA sequencing Lung adenocarcinoma Epithelial cell Ternary plot Machine learning screening MYO6 Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Introduction Lung cancer continues to rank as the foremost contributor to global cancer incidence and mortality, and lung adenocarcinoma (LUAD) represents its most prevalent histologic subtype. LUAD is distinguished by marked genetic heterogeneity and an intricate tumor microenvironment (TME)[1-3]. Therefore, a comprehensive characterization of the cellular landscape and molecular hallmarks that differentiate LUAD from normal lung tissue is essential for unraveling disease pathogenesis, refining precise diagnostics, and advancing targeted therapeutic strategies. Single-cell RNA sequencing (scRNA-seq) has rapidly evolved to offer unprecedented resolution for dissecting the cellular composition and functional states within the TME[4]. Although recent scRNA-seq studies have delineated the heterogeneity of immune and stromal compartments in LUAD, the dynamic evolution of epithelial subpopulations during tumorigenesis and the underlying regulatory mechanisms remain insufficiently characterized [5-8]. In normal lung tissue, epithelial cells are represented primarily by alveolar type 1 (AT1) cells, alveolar type 2 (AT2) cells, and airway epithelial cells, which mediate gas exchange, surfactant secretion, and mucociliary clearance, respectively [9-13]. Emerging evidence implicates AT1, AT2, and Club cells as potential cells of origin for LUAD; however, the transcriptional reprogramming trajectories during tumor progression and their crosstalk with other epithelial subpopulations are poorly defined [14-16]. KRT8⁺ alveolar intermediate cells (KACs), implicated in early LUAD development, accumulate in pre-malignant lesions and LUAD tissues. These cells display diminished differentiation, heightened plasticity, and KRAS-associated signatures, suggesting that they constitute an intermediate transitional state during the transformation of AT2 cells into malignant cells [17]. A comprehensive single-cell interrogation of early LUAD-specific cellular reprogramming will facilitate the discovery of actionable targets for prevention and intervention, ultimately improving clinical outcomes. By integrating two publicly available scRNA-seq datasets, we systematically interrogated cellular heterogeneity and delineated differential gene-expression landscapes in normal lung and LUAD tissues. Given that epithelial cells constitute the core functional units of the lung and are central to LUAD pathogenesis, we specifically isolated these cells from “Normal” and “LUAD” cohorts. Using a multidimensional analytical framework—encompassing dimensionality reduction, differential-gene screening, and cell-type–specific profiling—we comprehensively characterized epithelial subpopulations across pathological states, thereby mapping their dynamic transcriptional trajectories and elucidating transformation patterns during LUAD evolution. Global trajectory reconstruction coupled with pseudotime analysis (Monocle) and ternary-plot quantification delineated the spatial evolutionary trajectory of AT1, AT2, and airway epithelial marker genes, revealing a centralized cluster distribution in LUAD that reflects a shift from lineage-specific signatures toward a dedifferentiated state. To identify novel tumor-associated subpopulations, we devised a scoring system based on cell-type–specific gene signatures; clusters 11 and 26 exhibited the lowest scores. Leveraging 101 algorithmic combinations, we extracted core signature genes from the 140 intersecting markers of clusters 11/26 and constructed a high-precision prognostic model (StepCox[forward] + Ridge). Our data indicate that MYO6⁺ epithelial cells may foster tumor progression through activation of the MAPK signaling axis. In this study, we delineated the transcriptomic trajectory of lung epithelial cells across the entire continuum from normal tissue to invasive LUAD. We established an integrated analytical framework that combines single-cell RNA sequencing, ternary-plot quantification, and machine-learning algorithms to pinpoint tumor-specific epithelial subpopulations. We further identified MYO6 as an independent prognostic biomarker for LUAD and clarified its mechanistic role in driving tumor progression through the MAPK signaling axis. These findings provide novel mechanistic insights into LUAD pathogenesis and nominate MYO6 - expressing cells as actionable targets for subpopulation-restricted precision therapy. Methods Data source RNA-sequencing data were retrieved from the TCGA-LUAD cohort via the Genomic Data Commons (GDC) Data Portal (https://portal.gdc.cancer.gov/) and used as the training set. The GSE31210 and GSE30219 datasets, obtained from the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/), served as independent validation sets. Single-cell RNA-seq processing We analyzed the public single-cell datasets GSE149655 and GSE189357, encompassing 13 experimental conditions (2 control and 11 treatment groups). Raw 10x Genomics data were imported using Read10X, and Seurat objects were generated with CreateSeuratObject after filtering for genes expressed in ≥3 cells and cells expressing ≥200 genes. Cell barcodes were prefixed with sample identifiers to ensure traceability. Quality-control metrics—including the number of detected genes (nFeature_RNA), unique molecular identifiers (nCount_RNA), and the percentages of mitochondrial (percent.mt) and hemoglobin (percent.HB) transcripts—were calculated and visualized with violin plots. Low-quality cells were excluded using the following thresholds: 200–10,000 genes, 100–30,000 UMIs, percent.mt ≤ 20 %, and percent.HB < 3 %. To correct for sequencing-depth differences, counts were log-normalized (LogNormalize) and highly variable genes were identified by variance-stabilizing transformation (VST) using FindVariableFeatures. Data were then scaled with ScaleData, and technical batch effects were mitigated using SCTransform (negative-binomial regression via glmGamPoi) while regressing out percent.mt. We next performed principal-component analysis (RunPCA, 41 PCs) and visualized the latent space with t-SNE (RunTSNE, top 30 PCs). Clustering was conducted with the Louvain algorithm (FindNeighbors/FindClusters) and projected onto the t-SNE embedding. Cluster-defining markers were computed, and the top five markers per cluster were displayed as heatmaps (DoHeatmap). Cell identities were manually curated against established marker gene sets. Differential expression between “Normal” and “LUAD” groups was assessed with FindMarkers (adjusted p 1), and changes in cell-type composition were illustrated with stacked bar charts and distribution plots. Characterization of Epithelial Cell Subtypes in Normal Lung Tissue Epithelial cells within the “Normal” group were subjected to an in-depth characterization. After subsetting from the global Seurat object, the epithelial compartment was log-normalized (NormalizeData), and highly variable genes were identified (FindVariableFeatures, vst method). Data were then scaled (ScaleData) and dimensionality was reduced by principal-component analysis (PCA). Uniform Manifold Approximation and Projection (UMAP) was subsequently applied for two-dimensional visualization. Cluster-specific markers were obtained with FindAllMarkers using only positive markers (only.pos = TRUE), a minimum expression fraction of 0.25 (min.pct = 0.25), and a log-fold-change threshold of 0.25 (logfc.threshold = 0.25). Genes were considered significant if p-value 0.5. Cell-type composition and differential expression patterns were visualized with UMAP embeddings (DimPlot) and feature maps (FeaturePlot). Expression distributions of lineage-defining genes (SFTPA1, SFTPA2, HOPX, and SCGB1A1) were additionally plotted. Guided by clustering outputs and canonical marker expression, clusters were manually annotated as alveolar type II (AT2) epithelial cells, airway epithelial cells, and alveolar type I (AT1) epithelial cells. Comparative Analysis of Epithelial Cell Subtypes in Normal and LUAD Tissues Epithelial cells from the “Normal” and “LUAD” samples were isolated from the global single-cell atlas by subsetting the Seurat object on the basis of prior cell-type annotations. Highly variable genes were identified (FindVariableFeatures) to reduce dimensionality, followed by PCA. Non-linear embedding was then performed with the UMAP algorithm using the top three principal components. UMAP plots were generated to display cellular distributions, and epithelial clusters were manually assigned to AT2 cells, airway epithelial cells, or AT1 cells. Plots were redrawn with discrete color schemes to reflect the refined cell-type annotations. Pseudotime Analysis Pseudotime ordering was performed with Seurat and Monocle 3. The RNA expression matrix and associated metadata were extracted from the Seurat object and imported into Monocle 3, where data were normalized with preprocess_cds. UMAP embedding was computed from the top three principal components. Cell developmental trajectories were inferred via Louvain-based graph partitioning within Monocle 3. Trajectory-associated genes were identified with graph_test (q-value 0.25); the ten most significant genes were selected for visualization. Expression values were log-normalized, spline-smoothed (smooth.spline), and z-score scaled. The top 200 trajectory genes were subjected to k-means clustering and displayed in a heatmap. Temporal gene-expression dynamics were plotted, and cell distribution along pseudotime was overlaid on the UMAP, revealing transcriptional programs underpinning epithelial differentiation. Ternary Plot Visualization and Gene Scoring Analysis Cell-type–specific marker genes were systematically screened with the Seurat package. Markers for AT1, airway epithelial, and AT2 cells were retained when they met the following cut-offs: adjusted p-value 1, and expression in < 40 % of non-target cells. Per-cell lineage scores were then computed as the mean expression of the corresponding markers: score = rowSums(speci_raw) / length(Lineage_marker), where speci_raw denotes the marker-gene expression matrix. To visualize the relative contribution of the three lineage signatures to each cell, ternary plots were generated with ggtern; the three vertices correspond to AT1, airway, and AT2 scores, and the position of each point reflects its lineage affiliation. Distinct ternary plots were produced for normal, LUAD, and Treat.TD1–9 samples (encompassing in-situ, micro-invasive, and invasive stages). Module scores were calculated with AddModuleScore and projected onto UMAP embeddings (FeaturePlot). Box-and-jitter plots (ggplot2) further displayed score distributions across Seurat clusters. Screening of Core Feature Genes Using 101 Machine Learning Algorithms Differentially expressed genes (log₂FC > 1 and adjusted p < 0.05) between cluster 11 and cluster 26 were identified and subjected to intersection analysis. All downstream modelling was performed with the Mime1 package. Prognostic models were constructed with ML.Dev.Prog.Sig using the candidate genes and training data, then validated across independent datasets; model performance was quantified by the concordance index (C-index) and Kaplan–Meier survival analyses. Time-dependent Area Under the Curves (AUCs) for 1- and 3-year survival were calculated and plotted for cross-dataset comparison. Core predictive genes were subsequently extracted via ML.Corefeature.Prog.Screen, and their rankings were visualized with an UpSet plot and a gene-importance bar chart. Transcriptomic Profiling and Functional Pathway Analysis of MYO6+ Epithelial Cell Subpopulations Epithelial subpopulations were interrogated with Seurat. Cells were dichotomized into MYO6⁺ and MYO6⁻ groups using AddMetaData, with the median MYO6 transcript level as the cutoff. Cellular distributions in low-dimensional space were visualized with t-SNE embeddings (RunTSNE) and DimPlot. Differential expression between MYO6⁺ and MYO6⁻ populations was assessed by FindMarkers (min.pct = 0.25, logfc.threshold = 0.25); genes with log₂FC > 1 and P < 0.05 were considered significant. The resulting differentially expressed genes (DEGs) were subjected to Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses via enrichGO and enrichKEGG, respectively[18-20]. Prognostic analysis The prognostic value of MYO6 expression across multiple cancer types was evaluated by overall survival (OS) analysis on the GEPIA2 web platform (http://gepia2.cancer-pku.cn/). Tissue Sample Collection From October 2022 to November 2022, tissue samples were collected from 12 LUAD patients who underwent surgical treatment at Nantong University Affiliated Hospital. All patients' LUAD was confirmed by pathological examination. We removed all tissues and stored them in a -80 ° C freezer for further research. Provide tumor samples and corresponding normal samples to each patient at the edge of cancer. Ethics statement All procedures involving human participants were conducted in compliance with the ethical standards set by the institutional and national research committees, in alignment with the 1964 Helsinki Declaration and its subsequent amendments. The project was approved by Ethics Committee of the affiliated hospital of Tong University (license number: 2021-L142). Informed consent was obtained from all participants involved in the research. All participants gave informed consent to be included in the study. We fully adhere to SAGER guidelines and our study design was not related to gender. Cell Culture H1299, BEAS-2B and A549 cells were obtained from the American Type Culture Collection (ATCC). H1299 cells were maintained in RPMI-1640 (Cytiva) supplemented with 10 % fetal bovine serum (FBS). BEAS-2B and A549 cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM; Fuhong Biology) containing 10 % FBS (Gibco). All lines were incubated at 37 °C with 5 % CO₂ in a humidified atmosphere (Thermo Scientific). Quantitative Real-Time PCR (qRT-PCR) Total RNA was extracted from cultured cells using TRIzol reagent (Invitrogen, Waltham, MA, USA) and reverse-transcribed to cDNA with a commercial kit (Vazyme, Nanjing, China). qRT-PCR was performed with SYBR Green PCR Master Mix (Takara) in 10 µL reactions run in triplicate. Relative mRNA levels were normalized to β-actin and calculated by the 2^–ΔΔCt method. Primer sequences: MYO6 forward 5′-TGCCGACCAGTACAAAGACC-3′, reverse 5′-ATGGGTGGCTTGTCAAGGATG-3′; β-actin forward 5′-CATGTACGTTGCTATCCAGGC-3′, reverse 5′-CTCCTTAATGTCACGCACGAT-3′. Protein Extraction and Western Blotting Cells or tissues were lysed in RIPA buffer containing 1 % protease-inhibitor cocktail, incubated on ice for 30 min, and centrifuged at 12 000 × g, 4 °C, 15 min. Protein concentrations were determined with a BCA assay kit. Equal amounts (30 µg per lane) were denatured in 5× loading buffer at 95 °C for 10 min, resolved by 10 % SDS-PAGE, and transferred to PVDF membranes. After blocking with 5 % non-fat milk, membranes were incubated overnight at 4 °C with primary antibodies: MYO6 (1:5000; 26778-1-AP, Proteintech, China) and β-actin (1:20 000; 66009-1-Ig, Proteintech, China). Following extensive washing, membranes were probed with HRP-conjugated secondary antibodies (1:5000) for 2 h at room temperature. Protein bands were visualized with ECL substrate and quantified with ImageLab software using GAPDH as the loading control. Statistical analysis Statistical analysis was performed using GraphPad Prism (version 10) and R packages. The Student’s t-test and paired t-test were used for independent and paired groups, respectively. The results for continuous variables were presented as the mean ± standard deviation. A P value ≤0.05 was considered statistically significant. Results Single-cell RNA Sequencing Reveals Cell-type Heterogeneity and Differential Gene Expression in Normal and LUAD Lung Tissues As detailed in the Methods section and Supplementary Figure 1, single-cell analysis of 13 samples from GSE149655 and GSE189357 was performed following stringent quality control. All cells were classified into 41 clusters (Figure 1A). Guided by canonical markers, these clusters were assigned to eight major lineages: B cells, endothelial cells, epithelial cells, fibroblasts, macrophages, mast cells, neutrophils, and T cells (Figure 1B). t-SNE visualization and a corresponding heatmap illustrate the spatial distribution and marker-gene profiles of each population (Figure 1C). Comparative analysis revealed significant shifts in the relative abundance of all eight cell types between LUAD and normal lung tissue (Figure 1D). DEGs identified distinct gene signatures for each lineage across the “Normal” and “LUAD” groups (Figure 1E), and further confirmed proportional alterations in these populations between control and treatment conditions (Figure 1F). Single-cell Transcriptomic Analysis Reveals Heterogeneity and Molecular Signatures of Epithelial Cell Subpopulations in Normal and LUAD Lung Tissues UMAP embedding revealed the spatial distribution of epithelial cells within the “Normal” cohort (Figure 2A). Feature plots for canonical lineage markers (SFTPA1, SFTPA2, HOPX, and SCGB1A1) highlighted their cell-type–specific expression patterns (Figure 2B). Guided by these signatures, clusters 8, 11 and 18 were annotated as AT2 cells, clusters 33 and 35 as airway epithelial cells, and cluster 29 as AT1 cells (Figure 2C). When epithelia from both “Normal” and “LUAD” samples were projected into the same latent space, their distinct spatial organization became evident (Figure 2D). Accordingly, epithelial subpopulations were extracted according to these annotations (Figure 2E). Refinement of the classification assigned clusters 8, 11, 18 and 26 to the AT2 lineage, clusters 33 and 35 to airway epithelial cells, and cluster 29 to AT1 cells (Figure 2F). Ternary Plot Analysis Combined with Single-cell Transcriptomics Resolves the Dynamic Evolution of Epithelial Marker Genes and Identifies Heterogeneous Cell Subpopulations Ternary plots showed that normal epithelial cells clustered near the AT2 vertex (Figure 3A), whereas LUAD cells accumulated in the central region (Figure 3B). Adenocarcinoma in situ (AIS) cells began to shift toward the centre or an edge of the triangle (Figure 3C), and micro-invasive carcinoma cells displayed an even wider dispersion (Figure 3D). Invasive carcinoma cells converged at the centroid with the broadest distribution (Figure 3E). This reflects the dynamic changes in cell type marker gene expression during tumor progression from normal to invasive stages. In our study, the distribution of gene expression scores for AT1, AT2, and Airway Epithelial Cells was analyzed using UMAP plots. The results demonstrated significant heterogeneity in gene scores among these cells. However, we identified a subpopulation of cells that did not exhibit significant gene expression in any of the three scoring sets (Figure 3F-H). Three boxplots combined with jitter plots illustrate the distribution characteristics of gene scores across different cell clusters within the AT1, AT2, and Airway epithelial cell types. Notably, cell clusters 11 and 26 consistently exhibited low gene scores across all categories (Figure 3I-K). Furthermore, the spatial location of clusters 11 and 26 on the UMAP plot coincided with the location of cells lacking significant gene expression. Pseudotime Analysis To investigate dynamic changes during cellular development, a cell developmental trajectory plot illustrates the differentiation paths connecting AT2 Cells, AT1 Cells, and Airway Epithelial Cells (Figure 4A). The UMAP pseudotime trajectory plot presents the UMAP dimensionality reduction visualization of the single-cell transcriptomic data, depicting cell distribution along the developmental trajectory (Figure 4B). We visualized the expression levels of the top 200 feature genes in each epithelial cell (Figure 4C). This visualization clearly demonstrates the dynamic changes in gene expression, further validating the accuracy of the epithelial cell pseudotime analysis. To study the differential expression of key genes across cell populations, the top 10 genes significantly associated with the developmental trajectory were selected for visualization (Figure 4D). Pseudotime analysis of these significant genes revealed distinct dynamic expression patterns across pseudotime. These genes exhibited varying expression trends in different cell clusters (Figure 4E). Screening of Core Feature Genes Using a Combination of 101 Machine Learning Algorithms A total of 140 intersecting feature genes were identified from cell cluster 11 and cell cluster 26 (Figure 5A). These 140 genes were selected for model construction. Using the R package `Mime1`, prognostic risk models were built, and the C-index for each model is displayed (Figure 6A). The StepCox[forward] + Ridge model was selected for further analysis. The C-index of the StepCox[forward] + Ridge model across different datasets is shown (Figure 6B). Survival curves for high-risk and low-risk groups were plotted in both the training and test datasets. The results indicate significantly worse survival prognosis for the high-risk group in both datasets (Figure 6C). One-year and three-year survival prediction Receiver Operating Characteristic (ROC) curves were generated for the training and test sets (Figure 6D). The AUC values for one-year and three-year survival prediction were calculated in both datasets (Figure 6E). The results demonstrate that this risk model can effectively predict the survival of LUAD patients. Core feature gene selection was performed using an Upset plot showing genes filtered by different methods. A ranking plot displays the genes filtered by various methods, revealing that MYO6, ASPH, and KRT8 were the most frequently selected genes (Figure 6F-G). Spatial Distribution of MYO6 Gene Expression in Epithelial Cells and GO/KEGG Enrichment Analysis Spatial mapping of MYO6 expression distinguished MYO6⁻ and MYO6⁺ epithelial cells in two-dimensional space (Figure 7A). DEGs between these populations were subjected to GO and KEGG enrichment analyses. GO revealed marked enrichment in biological processes (e.g., phosphatidylcholine metabolism regulation, Hippo signalling), cellular components (multivesicular bodies, tight junctions) and molecular functions (protein-tyrosine-kinase activator activity, actin binding). KEGG analysis further identified significant involvement of fatty-acid metabolism, signalling cascades and disease-associated pathways, with the MAPK signalling pathway being most prominent (Figure 7B-C). MYO6 is highly expressed in both LUAD cells and tissues GEPIA2 analysis revealed that MYO6 mRNA levels were significantly elevated in 483 LUAD tissues relative to 347 normal lung controls (Figure 8A). Kaplan–Meier survival curves further demonstrated that high MYO6 expression conferred a markedly poorer prognosis (Figure 8B). Compared to the normal bronchial epithelial cell line 16HBE, significantly higher levels of MYO6 mRNA were observed in both LUAD cell lines A549 and H1299 (Figure 8C). Similarly, compared to adjacent non-tumorous tissues, MYO6 protein levels were significantly upregulated in LUAD tissues (Figure 8D-E). Discussion Classical models of tumorigenesis posit that malignant transformation reflects an aberrant recapitulation of normal developmental programmes, typically initiated in stem or progenitor cells whose differentiation is arrested [21-24]. Accumulating evidence underscores the marked plasticity of tumour cell phenotypes [25]; for example, loss of LKB1 can drive lineage switching from lung adenocarcinoma to squamous cell carcinoma [26]. Nevertheless, the precise cellular origins of most human malignancies—particularly the epithelial cancers that account for over 90 % of all tumours—remain poorly defined [27-29]. Consequently, the cellular identity of adenocarcinoma, a major epithelial subtype, is still incompletely understood. Single-cell transcriptomic profiling from pre-invasive to invasive LUAD has recently begun to illuminate the molecular underpinnings of disease progression [30]. Here, we leveraged single-cell RNA sequencing to interrogate epithelial cells from normal and LUAD lung tissues, systematically dissecting cellular heterogeneity and lineage-specific transcriptional programs. Consistent with prior work implicating AT2 cells as the potential cell-of-origin for KRAS-driven LUAD [31-33], we resolved distinct epithelial subpopulations—including AT1, AT2 and airway epithelial cells—whose relative abundances shifted markedly across pathological states. Ternary-plot and pseudotime analyses revealed a progressive loss of lineage fidelity, manifested as a transition from discrete lineage-specific signatures to a centrally aggregated, dedifferentiated state. Notably, clusters 11 and 26 exhibited uniformly low lineage-specific gene scores, suggesting that they represent transitional intermediates or specialized cellular states within the evolving LUAD microenvironment that may exert unique tumorigenic functions [34]. The progression from normal to malignant lung epithelium proceeds through intermediate states. KACs—KRT8⁺ alveolar intermediate cells—represent a transitional population between AT2 cells and tumour cells, exhibiting diminished differentiation, heightened plasticity, and increased susceptibility to KRAS-driven transformation [17]. Similarly, Kim et al. identified transitional alveolar cells during LUAD development that harbour unique molecular signatures implicated in tumour initiation [35]. Analogous transitional epithelial populations have been observed in murine lung injury models and may contribute to early LUAD evolution [36]. The concept of “confused cell identity” has further been proposed in oesophageal squamous cell carcinoma (ESCC), where tumour cells simultaneously display—yet diverge from—features of multiple normal oesophageal epithelial lineages; this identity confusion independently predicts poor prognosis [37]. Nevertheless, the precise roles of epithelial subpopulations within injury niches, and the mechanisms by which they drive progression from normal lung to pre-invasive lesions and invasive LUAD, remain undefined. From the 140 intersecting genes of clusters 11/26, we deployed 101 algorithmic combinations to identify core signature genes and constructed a high-precision prognostic model (StepCox[forward] + Ridge). MYO6, ASPH and KRT8 emerged as central survival-associated genes. We further defined a MYO6-high epithelial subpopulation (MYO6⁺) whose functional enrichment implicates MYO6-mediated activation of the MAPK axis in tumour progression. Thus, MYO6 constitutes a potential prognostic biomarker for LUAD. Differential expression of MYO6 within epithelial cells critically influences tumour behaviour; genes up-regulated in MYO6⁺ versus MYO6⁻ cells are enriched in multiple biological processes and, most prominently, the MAPK pathway. MYO6 (Myosin VI), a unique minus-end-directed motor protein, has been implicated as an oncogene in various cancers (e.g., prostate, colorectal, breast). Elevated MYO6 expression is consistently linked to tumor invasion, metastasis, and poor prognosis [38-42]. Mechanistically, the AR/MYO6/FAK axis drives malignant progression in castration-resistant prostate cancer [39],whereas MYO6 deletion suppresses proliferation and attenuates ERK1/2 and PRAS40 phosphorylation [43]. In breast cancer, MYO6 up-regulation activates the MAPK/ERK pathway and accelerates tumourigenesis [41]. These pan-cancer data corroborate our findings implicating MYO6 in LUAD pathogenesis. Although this study provides valuable insights, several limitations merit consideration. First, the exclusive use of publicly available datasets restricts sample size; validation in larger, independent cohorts is imperative. Second, while we integrated multiple analytical strategies for cell-type annotation and expression profiling, residual subjectivity in lineage assignment may affect interpretation under specific contexts. Future work should address these constraints by expanding sample numbers, incorporating orthogonal experimental platforms, and conducting rigorous clinical validation. Conclusion This study integrates scRNA-seq analysis with ternary plots and machine learning to precisely identify tumor-specific epithelial subpopulations and nominate MYO6 as a novel prognostic biomarker for LUAD. Moreover, we provide evidence suggesting MYO6 may drive tumor progression via the MAPK signaling pathway, offering a direction for future mechanistic investigations. This comprehensive analytical approach not only deepens our understanding of epithelial cell heterogeneity in LUAD but also unveils promising novel therapeutic targets for future diagnostic and therapeutic strategies. Abbreviations LUAD: Lung adenocarcinoma scRNA-seq: Single-cell RNA sequencing TME: Tumor microenvironment AT1: Alveolar type 1 AT2: Alveolar type 2 KACs: KRT8+ alveolar intermediate cells KEGG: Kyoto Encyclopedia of Genes and Genomes GO: Gene Ontology TCGA: The Cancer Genome Atlas UMAP: Uniform Manifold Approximation and Projection PCA: Principal Component Analysis t-SNE: t-distributed Stochastic Neighbor Embedding DEGs: Differentially expressed genes AUC: Area Under the Curve HR: Hazard ratio MAPK: Mitogen-activated protein kinase qRT-PCR: Quantitative real-time PCR cDNA: Complementary DNA DMEM: Dulbecco’s Modified Eagle Medium FBS: Fetal bovine serum RPMI-1640: Roswell Park Memorial Institute 1640 medium PVDF: Polyvinylidene difluoride HRP: Horseradish peroxidase ECL: Enhanced chemiluminescence GAPDH: Glyceraldehyde-3-phosphate dehydrogenase AIS: Adenocarcinoma in situ C-index: Concordance Index ROC: Receiver Operating Characteristic ESCC: Esophageal squamous cell carcinoma AR: Androgen receptor FAK: Focal adhesion kinase ERK: Extracellular signal-regulated kinase PRAS40: Proline-rich Akt substrate of 40 kDa Declarations Acknowledgements The authors acknowledge the efforts of all of the researchers who have contributed the data to the public databases of TCGA and GEO. The interpretation and reporting of these data are the sole responsibility of the authors. Funding This work was supported by the Natural Science Research Project of Nantong Science and Technology Bureau (No. MS12020029). Author information Author notes Miaoyan Liu and Houqiang Li contributed equally to this work. Authors and Affiliations Department of Thoracic Surgery, Affiliated Hospital of Nantong University, Nantong, 226001,China Miaoyan Liu, Houqiang Li, Shenghan Xu, Yike Zhou, Tiegang Cao , Jiahai Shi, Lou Zhong Department of Immunology, Medical School of Nantong University & Research Center of Clinical Medicine, Affiliated Hospital of Nantong University,Nantong, 226019,China Min Yao Contributions LZ and JHS provided guidance throughout the preparation of this manuscript. MYL and HQL conceptualised and wrote the manuscript. HQL,SHX ,YKZ , TGC, and MY analysed and interpreted the data. MYL conducted the experiments and collected the data. MYL, JHS and LZ designed the experiments and analysed the results. All authors approved the final version. Corresponding authors Correspondence to Lou Zhong and Jiahai Shi. Ethics declarations Ethics approval and consent to participate All human samples procedures were approved by the Ethics Committee of the Affiliated Hospital of Nantong University (license number: 2021-L142). Consent for publication Not applicable. Competing interests The authors declare no competing interests. Data availability All datasets used in this study are publicly available from the following repositories: Single-cell RNA-seq data: GSE149655 and GSE189357 from the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/). Bulk RNA-seq data: TCGA-LUAD cohort from the Genomic Data Commons (GDC) Data Portal (https://portal.gdc.cancer.gov/). Validation datasets (GSE31210 and GSE30219) from GEO (https://www.ncbi.nlm.nih.gov/geo/). References Bray F, Ferlay J, Soerjomataram I, Siegel RL, Torre LA, Jemal A. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2018;68:394-424. Devarakonda S, Morgensztern D, Govindan R. Genomic alterations in lung adenocarcinoma. Lancet Oncol. 2015;16:e342-51. Song C, Guo Z, Yu D, Wang Y, Wang Q, Dong Z, Hu W. A Prognostic Nomogram Combining Immune-Related Gene Signature and Clinical Factors Predicts Survival in Patients With Lung Adenocarcinoma. Front Oncol. 2020;10:1300. Sha Y, Wang S, Zhou P, Nie Q. Inference and multiscale model of epithelial-to-mesenchymal transition via single-cell transcriptomic data. Nucleic Acids Res. 2020;48:9505-20. He D, Wang D, Lu P, Yang N, Xue Z, Zhu X, Zhang P, Fan G. Single-cell RNA sequencing reveals heterogeneous tumor and immune cell populations in early-stage lung adenocarcinomas harboring EGFR mutations. Oncogene. 2021;40:355-68. Li D, Yu H, Hu J, Li S, Yan Y, Li S, Sun L, Jiang G, Hou L, Zhang L, et al. Comparative profiling of single-cell transcriptome reveals heterogeneity of tumor microenvironment between solid and acinar lung adenocarcinoma. J Transl Med. 2022;20:423. Travaglini KJ, Nabhan AN, Penland L, Sinha R, Gillich A, Sit RV, Chang S, Conley SD, Mori Y, Seita J, et al. A molecular cell atlas of the human lung from single-cell RNA sequencing. Nature. 2020;587:619-25. Zhu J, Fan Y, Xiong Y, Wang W, Chen J, Xia Y, Lei J, Gong L, Sun S, Jiang T. Delineating the dynamic evolution from preneoplasia to invasive lung adenocarcinoma by integrating single-cell RNA sequencing and spatial transcriptomics. Exp Mol Med. 2022;54:2060-76. Zepp JA, Morrisey EE. Cellular crosstalk in the development and regeneration of the respiratory system. Nat Rev Mol Cell Biol. 2019;20:551-66. Iwasaki A, Foxman EF, Molony RD. Early local immune defences in the respiratory tract. Nat Rev Immunol. 2017;17:7-20. Park KS, Liang MC, Raiser DM, Zamponi R, Roach RR, Curtis SJ, Walton Z, Schaffer BE, Roake CM, Zmoos AF, et al. Characterization of the cell of origin for small cell lung cancer. Cell Cycle. 2011;10:2806-15. Jia Y, You X, Ma N, Li H, Liu G, Wang Y, Xue J, Shi J, Wei J, Li Y, et al. Phenotypic Analysis of BrdU Label-Retaining Cells during the Maturation of Conducting Airway Epithelium in a Porcine Lung. Stem Cells Int. 2019;2019:7043890. Weitnauer M, Mijošek V, Dalpke AH. Control of local immunity by airway epithelial cells. Mucosal Immunol. 2016;9:287-98. Sainz de Aja J, Dost A, Kim CF. Alveolar progenitor cells and the origin of lung cancer. J Intern Med. 2021;289:629-35. Yang M, Shen H, Flodby P, Koss MD, Bassiouni R, Liu Y, Jashashvili T, Neely A, Ogbolu E, Castillo J, et al. Alveolar type I cells can give rise to KRAS-induced lung adenocarcinoma. Cell Rep. 2023;42:113286. Rosigkeit S, Kruchem M, Thies D, Kreft A, Eichler E, Boegel S, Jansky S, Siegl D, Kaps L, Pickert G, et al. Definitive evidence for Club cells as progenitors for mutant Kras/Trp53-deficient lung cancer. Int J Cancer. 2021;149:1670-82. Han G, Sinjab A, Rahal Z, Lynch AM, Treekitkarnmongkol W, Liu Y, Serrano AG, Feng J, Liang K, Khan K, et al. An atlas of epithelial cell states and plasticity in lung adenocarcinoma. Nature. 2024;627:656-63. Kanehisa M, Furumichi M, Sato Y, Matsuura Y, Ishiguro-Watanabe M. KEGG: biological systems database as a model of the real world. Nucleic Acids Res. 2025;53:D672-672D677. Kanehisa M. Toward understanding the origin and evolution of cellular organisms. Protein Sci. 2019;28:1947-51. Kanehisa M, Goto S. KEGG: kyoto encyclopedia of genes and genomes. Nucleic Acids Res. 2000;28:27-30. Batlle E, Clevers H. Cancer stem cells revisited. Nat Med. 2017;23:1124-34. Magee JA, Piskounova E, Morrison SJ. Cancer stem cells: impact, heterogeneity, and uncertainty. Cancer Cell. 2012;21:283-96. Kelly LM, Gilliland DG. Genetics of myeloid leukemias. Annu Rev Genomics Hum Genet. 2002;3:179-98. Blanpain C, Horsley V, Fuchs E. Epithelial stem cells: turning over new leaves. Cell. 2007;128:445-58. Niederst MJ, Sequist LV, Poirier JT, Mermel CH, Lockerman EL, Garcia AR, Katayama R, Costa C, Ross KN, Moran T, et al. RB loss in resistant EGFR mutant lung adenocarcinomas that transform to small-cell lung cancer. Nat Commun. 2015;6:6377. Han X, Li F, Fang Z, Gao Y, Li F, Fang R, Yao S, Sun Y, Li L, Zhang W, et al. Transdifferentiation of lung adenocarcinoma in mice with Lkb1 deficiency to squamous cell carcinoma. Nat Commun. 2014;5:3261. Network CGA. Comprehensive molecular portraits of human breast tumours. Nature. 2012;490:61-70. Perou CM, Sørlie T, Eisen MB, van de Rijn M, Jeffrey SS, Rees CA, Pollack JR, Ross DT, Johnsen H, Akslen LA, et al. Molecular portraits of human breast tumours. Nature. 2000;406:747-52. Dotto GP, Rustgi AK. Squamous Cell Cancers: A Unified Perspective on Biology and Genetics. Cancer Cell. 2016;29:622-37. Xiang C, Zhang M, Shang Z, Chen S, Zhao J, Ding B, Jiang D, Zhu Q, Teng H, Zhu L, et al. Single-cell profiling reveals the trajectory of FOLR2-expressing tumor-associated macrophages to regulatory T cells in the progression of lung adenocarcinoma. Cell Death Dis. 2023;14:493. Lin C, Song H, Huang C, Yao E, Gacayan R, Xu SM, Chuang PT. Alveolar type II cells possess the capability of initiating lung tumor development. PLoS One. 2012;7:e53817. Liu B, Li C, Li Z, Wang D, Ren X, Zhang Z. An entropy-based metric for assessing the purity of single cell populations. Nat Commun. 2020;11:3155. Xu X, Rock JR, Lu Y, Futtner C, Schwab B, Guinney J, Hogan BL, Onaitis MW. Evidence for type II cells as cells of origin of K-Ras-induced distal lung adenocarcinoma. Proc Natl Acad Sci U S A. 2012;109:4910-5. Liebler JM, Marconett CN, Juul N, Wang H, Liu Y, Flodby P, Laird-Offringa IA, Minoo P, Zhou B. Combinations of differentiation markers distinguish subpopulations of alveolar epithelial cells in adult lung. Am J Physiol Lung Cell Mol Physiol. 2016;310:L114-20. Xu JY, Zhang C, Wang X, Zhai L, Ma Y, Mao Y, Qian K, Sun C, Liu Z, Jiang S, et al. Integrative Proteomic Characterization of Human Lung Adenocarcinoma. Cell. 2020;182:245-61.e17. Tata PR, Chow RD, Saladi SV, Tata A, Konkimalla A, Bara A, Montoro D, Hariri LP, Shih AR, Mino-Kenudson M, et al. Developmental History Provides a Roadmap for the Emergence of Tumor Plasticity. Dev Cell. 2018;44:679-93.e5. Pan X, Wang J, Guo L, Na F, Du J, Chen X, Zhong A, Zhao L, Zhang L, Zhang M, et al. Identifying a confused cell identity for esophageal squamous cell carcinoma. Signal Transduct Target Ther. 2022;7:122. Kruppa AJ, Kishi-Itakura C, Masters TA, Rorbach JE, Grice GL, Kendrick-Jones J, Nathan JA, Minczuk M, Buss F. Myosin VI-Dependent Actin Cages Encapsulate Parkin-Positive Damaged Mitochondria. Dev Cell. 2018;44:484-99.e6. Zheng S, Hong Z, Tan Y, Wang Y, Li J, Zhang Z, Feng T, Hong Z, Lin G, Ye D. MYO6 contributes to tumor progression and enzalutamide resistance in castration-resistant prostate cancer by activating the focal adhesion signaling pathway. Cell Commun Signal. 2024;22:517. Jiang M, Xiong C. USP7 accelerates colorectal cancer progression by up-regulating MYO6 through deubiquitination. Mutat Res. 2025;831:111908. Zhan XJ, Wang R, Kuang XR, Zhou JY, Hu XL. Elevated expression of myosin VI contributes to breast cancer progression via MAPK/ERK signaling pathway. Cell Signal. 2023;106:110633. Yu H, Zhu Z, Chang J, Wang J, Shen X. Lentivirus-Mediated Silencing of Myosin VI Inhibits Proliferation and Cell Cycle Progression in Human Lung Cancer Cells. Chem Biol Drug Des. 2015;86:606-13. Wang D, Zhu L, Liao M, Zeng T, Zhuo W, Yang S, Wu W. MYO6 knockdown inhibits the growth and induces the apoptosis of prostate cancer cells by decreasing the phosphorylation of ERK1/2 and PRAS40. Oncol Rep. 2016;36:1285-92. Additional Declarations No competing interests reported. Supplementary Files SupplementFigure1.jpg Supplement Figure 1: Quality control metrics for single-cell RNA sequencing data. (A) Distribution of RNA counts, mitochondrial content (percent), and number of detected features (genes) across different experimental conditions. (B) Distribution analysis of the number of detected features (genes), total RNA counts, mitochondrial RNA percentage, and hemoglobin gene transcript percentage across experimental conditions. (C) Identification of highly variable genes and visualization of their mean-variance relationship. (D) Principal Component Analysis illustrating the distribution of samples under different experimental conditions. fulllengthgelsandblots.pdf Cite Share Download PDF Status: Under Review Version 1 posted Reviews received at journal 01 Mar, 2026 Reviewers agreed at journal 01 Mar, 2026 Reviewers invited by journal 26 Feb, 2026 Editor assigned by journal 28 Jan, 2026 Editor invited by journal 08 Oct, 2025 Submission checks completed at journal 07 Oct, 2025 First submitted to journal 07 Oct, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7742292","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":598852776,"identity":"72e7d685-2516-4dc1-a112-7b265fb3f229","order_by":0,"name":"Miaoyan Liu","email":"","orcid":"","institution":"Affiliated Hospital of Nantong University","correspondingAuthor":false,"prefix":"","firstName":"Miaoyan","middleName":"","lastName":"Liu","suffix":""},{"id":598852778,"identity":"f458b94c-419b-4664-849b-755866ec1390","order_by":1,"name":"Houqiang Li","email":"","orcid":"","institution":"Affiliated Hospital of Nantong University","correspondingAuthor":false,"prefix":"","firstName":"Houqiang","middleName":"","lastName":"Li","suffix":""},{"id":598852785,"identity":"98ee4d83-81c6-445d-a4c3-736504413463","order_by":2,"name":"Shenghan Xu","email":"","orcid":"","institution":"Affiliated Hospital of Nantong University","correspondingAuthor":false,"prefix":"","firstName":"Shenghan","middleName":"","lastName":"Xu","suffix":""},{"id":598852792,"identity":"dcdacc51-2401-4d4f-acf3-b00653b3e9e8","order_by":3,"name":"Yike Zhou","email":"","orcid":"","institution":"Affiliated Hospital of Nantong University","correspondingAuthor":false,"prefix":"","firstName":"Yike","middleName":"","lastName":"Zhou","suffix":""},{"id":598852793,"identity":"45fa6c14-0c16-462a-89c2-a5bb53541ea3","order_by":4,"name":"Tiegang Cao","email":"","orcid":"","institution":"Affiliated Hospital of Nantong University","correspondingAuthor":false,"prefix":"","firstName":"Tiegang","middleName":"","lastName":"Cao","suffix":""},{"id":598852795,"identity":"dabc2d22-d417-4cd4-ab2c-fa2bd837b5f8","order_by":5,"name":"Min Yao","email":"","orcid":"","institution":"Affiliated Hospital of Nantong University","correspondingAuthor":false,"prefix":"","firstName":"Min","middleName":"","lastName":"Yao","suffix":""},{"id":598852796,"identity":"046379a2-d396-4b92-9781-b0ecd928d2f4","order_by":6,"name":"Jiahai Shi","email":"","orcid":"","institution":"Affiliated Hospital of Nantong University","correspondingAuthor":false,"prefix":"","firstName":"Jiahai","middleName":"","lastName":"Shi","suffix":""},{"id":598852798,"identity":"d525a634-e3a0-408a-a344-4e01817c1799","order_by":7,"name":"Lou Zhong","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAxklEQVRIiWNgGAWjYFCCxAcMDAU2EDYPMRp4GJINGBgM0iCqSdFymAQt9uzJbJI/DM7n2UskMD5428Ygb07QFp7HbBISBreLeSQSmA3ntjEY7mwgpEUi/5iEgcHtxB6JBDZp3jaGBIMDBLUks0kkGJwDaWH/TbyWAwYHwLYwE6flzGNmywaD5MSeMw+bJeeckzDcQEgLe3sy480fFXaJ7e3JBz+8KbORJ2gLELBIQGjGBiAhQVg9EDB/IErZKBgFo2AUjFwAAN6yOC+6mDlXAAAAAElFTkSuQmCC","orcid":"","institution":"Affiliated Hospital of Nantong University","correspondingAuthor":true,"prefix":"","firstName":"Lou","middleName":"","lastName":"Zhong","suffix":""}],"badges":[],"createdAt":"2025-09-29 12:53:32","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7742292/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7742292/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":104403393,"identity":"63a66d96-99ea-43b8-8c25-3a3963790677","added_by":"auto","created_at":"2026-03-11 12:18:14","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":2273838,"visible":true,"origin":"","legend":"\u003cp\u003eVisualization analysis of single-cell data. (A) t-SNE algorithm identified 41 cell clusters. (B) Annotation of 8 cell types using marker gene-based t-SNE. (C) Cell type annotation and analysis of marker gene expression patterns. (D) Comparative analysis of cell type proportions in LUAD tissue versus normal lung tissue. (E) Volcano plot of DEGs for each cell cluster, highlighting the top 3 upregulated genes and top 5 downregulated genes per cell type.(F) Comparison of the relative frequencies of various cell types across different experimental groups (Control and Treat).\u003c/p\u003e","description":"","filename":"Figure1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7742292/v1/cd4c0bd3b1192e4ba94056b4.jpg"},{"id":104172052,"identity":"9fac5846-e216-4666-9e60-a9b838404957","added_by":"auto","created_at":"2026-03-08 14:58:44","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":2988570,"visible":true,"origin":"","legend":"\u003cp\u003eSingle-cell transcriptomic analysis reveals differential distribution of epithelial cell subpopulations in normal and LUAD lung tissues. (A) UMAP dimensionality reduction analysis of epithelial cell subpopulations in normal tissue. (B) UMAP expression distribution plots visualizing specific genes. (C) UMAP visualization revealing the distribution of alveolar and airway epithelial cell types in normal tissue. (D) UMAP dimensionality reduction analysis of epithelial cells in normal and LUAD tissues. (E) Epithelial cell subpopulations in normal and LUAD tissues. (F) UMAP visualization revealing the distribution of alveolar and airway epithelial cell types in normal and LUAD tissues.\u003c/p\u003e","description":"","filename":"Figure2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7742292/v1/03e35dc36c26c2fdda293715.jpg"},{"id":104172050,"identity":"c02ddbb8-44f3-4ad4-aaf3-9bf22221b077","added_by":"auto","created_at":"2026-03-08 14:58:44","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1804329,"visible":true,"origin":"","legend":"\u003cp\u003eCombined ternary plot and single-cell transcriptomic analysis resolves the dynamic evolution of epithelial marker genes. (A-E) Ternary plot-based trajectory of lung epithelial cells from normal to LUAD transformation. (F-H) UMAP analysis reveals heterogeneity in AT1, AT2, and Airway epithelial cell gene scores and identifies atypical cell subpopulations. (I-K) Boxplot-jitter plot visualization of gene score heterogeneity within AT1, AT2, and Airway epithelial cell subpopulations.\u003c/p\u003e","description":"","filename":"Figure3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7742292/v1/98fee7c81ef74416e55aa108.jpg"},{"id":104779402,"identity":"216e223d-1cc5-4a43-8b28-8902627f45d2","added_by":"auto","created_at":"2026-03-17 07:39:49","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":3410973,"visible":true,"origin":"","legend":"\u003cp\u003ePseudotime analysis. (A) Cell developmental trajectory plot. (B) UMAP pseudotime trajectory plot. (C) Visualization of top 200 feature gene expression in epithelial cells. Color represents z-score (standardized expression level). Each row represents a gene, each column represents a cell. Color gradient (blue to red) indicates low to high expression levels. (D) Expression differences of the top 10 significant genes across different cell populations. (E) Expression dynamics of the 10 significant genes along the pseudotime trajectory.\u003c/p\u003e","description":"","filename":"Figure4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7742292/v1/e162a2cd2f5a577d30bbfa60.jpg"},{"id":104172049,"identity":"259fdb15-8f55-4925-a7ae-63fb11e1f11e","added_by":"auto","created_at":"2026-03-08 14:58:44","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":106499,"visible":true,"origin":"","legend":"\u003cp\u003eVenn diagram of intersecting feature genes. (A) Venn diagram of intersecting feature genes from cell cluster 11 and cell cluster 26.\u003c/p\u003e","description":"","filename":"Figure5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7742292/v1/5f098d20edf5e1071f57e92e.jpg"},{"id":104172057,"identity":"34173324-8c0e-4920-94fa-7b145dda8d32","added_by":"auto","created_at":"2026-03-08 14:58:45","extension":"jpg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":1664175,"visible":true,"origin":"","legend":"\u003cp\u003eScreening of core feature genes using a combination of 101 machine learning algorithms. (A) C-index of each model. (B) C-index of the StepCox[forward] + Ridge model across different datasets. (C) Survival curve analysis of the StepCox + Ridge model in different datasets. (D) ROC analysis for 1-year and 3-year survival prediction by the StepCox + Ridge model in different datasets. (E) Comparison of AUC values for 1-year and 3-year survival prediction by the StepCox + Ridge model across different datasets. (F) Performance comparison of gene feature selection methods using ROC analysis for the specific model across different datasets. (G) Top 20 selected genes and their selection frequency.\u003c/p\u003e","description":"","filename":"Figure6.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7742292/v1/b22df7939a51c484242690e1.jpg"},{"id":104172054,"identity":"68079b90-e516-4dc6-ac97-1992d963ce2f","added_by":"auto","created_at":"2026-03-08 14:58:44","extension":"jpg","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":1019354,"visible":true,"origin":"","legend":"\u003cp\u003eSpatial distribution of MYO6 gene expression in epithelial cells and GO/KEGG enrichment analysis. (A) t-SNE visualization analysis of MYO6 gene expression levels in epithelial cells. (B-C) Bubble plots of GO and KEGG enrichment analyses for the specific MYO6-associated DEG sets.\u003c/p\u003e","description":"","filename":"Figure7.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7742292/v1/ad900f888158cd7ea49c4a4d.jpg"},{"id":104172058,"identity":"33f702a0-13ae-457a-8089-89bf65fdc517","added_by":"auto","created_at":"2026-03-08 14:58:45","extension":"jpg","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":613187,"visible":true,"origin":"","legend":"\u003cp\u003eAnalysis of MYO6 Expression in LUAD. (A) GEPIA2 analysis revealed MYO6 expression in LUAD vs. normal tissues. (B) KM curves stratified by high/low MYO6 expression. (C) qRT-PCR showed elevated MYO6 mRNA in A549/H1299 LUAD cells vs. HBE normal cells. (D-E). WB confirmed upregulated MYO6 protein in 12 paired tumor vs. adjacent tissues. “ns”: not significant; “*”: P \u0026lt; 0.05; “**”: P \u0026lt; 0.01; “***”: P \u0026lt; 0.001; “****”: P \u0026lt; 0.0001.\u003c/p\u003e","description":"","filename":"Figure8.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7742292/v1/1c87af43d2f2f28d6de3a73d.jpg"},{"id":104786318,"identity":"c9277b8c-9472-4ba7-b07d-253886978dba","added_by":"auto","created_at":"2026-03-17 08:16:17","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":14677640,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7742292/v1/d6540258-5537-4c50-ad8b-ebfe4ab2bc5c.pdf"},{"id":104172056,"identity":"796ae51e-74ea-4bf2-868b-4f3ddccf8ee1","added_by":"auto","created_at":"2026-03-08 14:58:45","extension":"jpg","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":536913,"visible":true,"origin":"","legend":"\u003cp\u003eSupplement Figure 1: Quality control metrics for single-cell RNA sequencing data. (A) Distribution of RNA counts, mitochondrial content (percent), and number of detected features (genes) across different experimental conditions. (B) Distribution analysis of the number of detected features (genes), total RNA counts, mitochondrial RNA percentage, and hemoglobin gene transcript percentage across experimental conditions. (C) Identification of highly variable genes and visualization of their mean-variance relationship. (D) Principal Component Analysis illustrating the distribution of samples under different experimental conditions.\u003c/p\u003e","description":"","filename":"SupplementFigure1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7742292/v1/af7fd871a6e0915eb6c332e0.jpg"},{"id":104404769,"identity":"e3678eb9-c528-45e3-b87e-4fd807b3ca84","added_by":"auto","created_at":"2026-03-11 12:21:03","extension":"pdf","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":3913445,"visible":true,"origin":"","legend":"","description":"","filename":"fulllengthgelsandblots.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7742292/v1/d859b4d69442a151a1e00dff.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"\u003cp\u003e\u003cstrong\u003eMYO6\u003c/strong\u003e\u003csup\u003e\u003cstrong\u003e+\u003c/strong\u003e\u003c/sup\u003e\u003cstrong\u003eEpithelial Subpopulation as a Prognostic Hub in Lung Adenocarcinoma Identified by Multi-Omics Integration\u003c/strong\u003e\u003c/p\u003e","fulltext":[{"header":"Introduction","content":"\u003cp\u003eLung cancer continues to rank as the foremost contributor to global cancer incidence and mortality, and lung adenocarcinoma (LUAD) represents its most prevalent histologic subtype. LUAD is distinguished by marked genetic heterogeneity and an intricate tumor microenvironment (TME)[1-3]. Therefore, a comprehensive characterization of the cellular landscape and molecular hallmarks that differentiate LUAD from normal lung tissue is essential for unraveling disease pathogenesis, refining precise diagnostics, and advancing targeted therapeutic strategies.\u003c/p\u003e\n\u003cp\u003eSingle-cell RNA sequencing (scRNA-seq) has rapidly evolved to offer unprecedented resolution for dissecting the cellular composition and functional states within the TME[4]. Although recent scRNA-seq studies have delineated the heterogeneity of immune and stromal compartments in LUAD, the dynamic evolution of epithelial subpopulations during tumorigenesis and the underlying regulatory mechanisms remain insufficiently characterized\u0026nbsp;[5-8]. In normal lung tissue, epithelial cells are represented primarily by alveolar type 1 (AT1) cells, alveolar type 2 (AT2) cells, and airway epithelial cells, which mediate gas exchange, surfactant secretion, and mucociliary clearance, respectively\u0026nbsp;[9-13]. Emerging evidence implicates AT1, AT2, and Club cells as potential cells of origin for LUAD; however, the transcriptional reprogramming trajectories during tumor progression and their crosstalk with other epithelial subpopulations are poorly defined\u0026nbsp;[14-16]. KRT8⁺ alveolar intermediate cells (KACs), implicated in early LUAD development, accumulate in pre-malignant lesions and LUAD tissues. These cells display diminished differentiation, heightened plasticity, and KRAS-associated signatures, suggesting that they constitute an intermediate transitional state during the transformation of AT2 cells into malignant cells\u0026nbsp;[17]. A comprehensive single-cell interrogation of early LUAD-specific cellular reprogramming will facilitate the discovery of actionable targets for prevention and intervention, ultimately improving clinical outcomes.\u003c/p\u003e\n\u003cp\u003eBy integrating two publicly available scRNA-seq datasets, we systematically interrogated cellular heterogeneity and delineated differential gene-expression landscapes in normal lung and LUAD tissues. Given that epithelial cells constitute the core functional units of the lung and are central to LUAD pathogenesis, we specifically isolated these cells from \u0026ldquo;Normal\u0026rdquo; and \u0026ldquo;LUAD\u0026rdquo; cohorts. Using a multidimensional analytical framework\u0026mdash;encompassing dimensionality reduction, differential-gene screening, and cell-type\u0026ndash;specific profiling\u0026mdash;we comprehensively characterized epithelial subpopulations across pathological states, thereby mapping their dynamic transcriptional trajectories and elucidating transformation patterns during LUAD evolution. Global trajectory reconstruction coupled with pseudotime analysis (Monocle) and ternary-plot quantification delineated the spatial evolutionary trajectory of AT1, AT2, and airway epithelial marker genes, revealing a centralized cluster distribution in LUAD that reflects a shift from lineage-specific signatures toward a dedifferentiated state. To identify novel tumor-associated subpopulations, we devised a scoring system based on cell-type\u0026ndash;specific gene signatures; clusters 11 and 26 exhibited the lowest scores. Leveraging 101 algorithmic combinations, we extracted core signature genes from the 140 intersecting markers of clusters 11/26 and constructed a high-precision prognostic model (StepCox[forward] + Ridge). Our data indicate that MYO6⁺ epithelial cells may foster tumor progression through activation of the MAPK signaling axis.\u003c/p\u003e\n\u003cp\u003eIn this study, we delineated the transcriptomic trajectory of lung epithelial cells across the entire continuum from normal tissue to invasive LUAD. We established an integrated analytical framework that combines single-cell RNA sequencing, ternary-plot quantification, and machine-learning algorithms to pinpoint tumor-specific epithelial subpopulations. We further identified MYO6 as an independent prognostic biomarker for LUAD and clarified its mechanistic role in driving tumor progression through the MAPK signaling axis. These findings provide novel mechanistic insights into LUAD pathogenesis and nominate MYO6\u003csup\u003e-\u0026nbsp;\u003c/sup\u003eexpressing cells as actionable targets for subpopulation-restricted precision therapy.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003eData source\u003c/p\u003e\n\u003cp\u003eRNA-sequencing data were retrieved from the TCGA-LUAD cohort via the Genomic Data Commons (GDC) Data Portal (https://portal.gdc.cancer.gov/) and used as the training set. The GSE31210 and GSE30219 datasets, obtained from the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/), served as independent validation sets.\u003c/p\u003e\n\u003cp\u003eSingle-cell RNA-seq processing\u003c/p\u003e\n\u003cp\u003eWe analyzed the public single-cell datasets GSE149655 and GSE189357, encompassing 13 experimental conditions (2 control and 11 treatment groups). Raw 10x Genomics data were imported using Read10X, and Seurat objects were generated with CreateSeuratObject after filtering for genes expressed in \u0026ge;3 cells and cells expressing \u0026ge;200 genes. Cell barcodes were prefixed with sample identifiers to ensure traceability. Quality-control metrics\u0026mdash;including the number of detected genes (nFeature_RNA), unique molecular identifiers (nCount_RNA), and the percentages of mitochondrial (percent.mt) and hemoglobin (percent.HB) transcripts\u0026mdash;were calculated and visualized with violin plots. Low-quality cells were excluded using the following thresholds: 200\u0026ndash;10,000 genes, 100\u0026ndash;30,000 UMIs, percent.mt \u0026le; 20 %, and percent.HB \u0026lt; 3 %. To correct for sequencing-depth differences, counts were log-normalized (LogNormalize) and highly variable genes were identified by variance-stabilizing transformation (VST) using FindVariableFeatures. Data were then scaled with ScaleData, and technical batch effects were mitigated using SCTransform (negative-binomial regression via glmGamPoi) while regressing out percent.mt. We next performed principal-component analysis (RunPCA, 41 PCs) and visualized the latent space with t-SNE (RunTSNE, top 30 PCs). Clustering was conducted with the Louvain algorithm (FindNeighbors/FindClusters) and projected onto the t-SNE embedding. Cluster-defining markers were computed, and the top five markers per cluster were displayed as heatmaps (DoHeatmap). Cell identities were manually curated against established marker gene sets. Differential expression between \u0026ldquo;Normal\u0026rdquo; and \u0026ldquo;LUAD\u0026rdquo; groups was assessed with FindMarkers (adjusted p \u0026lt; 0.05, |log2FC| \u0026gt; 1), and changes in cell-type composition were illustrated with stacked bar charts and distribution plots.\u003c/p\u003e\n\u003cp\u003eCharacterization of Epithelial Cell Subtypes in Normal Lung Tissue\u003cbr\u003e Epithelial cells within the \u0026ldquo;Normal\u0026rdquo; group were subjected to an in-depth characterization. After subsetting from the global Seurat object, the epithelial compartment was log-normalized (NormalizeData), and highly variable genes were identified (FindVariableFeatures, vst method). Data were then scaled (ScaleData) and dimensionality was reduced by principal-component analysis (PCA). Uniform Manifold Approximation and Projection (UMAP) was subsequently applied for two-dimensional visualization. Cluster-specific markers were obtained with FindAllMarkers using only positive markers (only.pos = TRUE), a minimum expression fraction of 0.25 (min.pct = 0.25), and a log-fold-change threshold of 0.25 (logfc.threshold = 0.25). Genes were considered significant if p-value \u0026lt; 0.05 and |log2FC| \u0026gt; 0.5. Cell-type composition and differential expression patterns were visualized with UMAP embeddings (DimPlot) and feature maps (FeaturePlot). Expression distributions of lineage-defining genes (SFTPA1, SFTPA2, HOPX, and SCGB1A1) were additionally plotted. Guided by clustering outputs and canonical marker expression, clusters were manually annotated as alveolar type II (AT2) epithelial cells, airway epithelial cells, and alveolar type I (AT1) epithelial cells.\u003c/p\u003e\n\u003cp\u003eComparative Analysis of Epithelial Cell Subtypes in Normal and LUAD Tissues\u003cbr\u003eEpithelial cells from the \u0026ldquo;Normal\u0026rdquo; and \u0026ldquo;LUAD\u0026rdquo; samples were isolated from the global single-cell atlas by subsetting the Seurat object on the basis of prior cell-type annotations. Highly variable genes were identified (FindVariableFeatures) to reduce dimensionality, followed by PCA. Non-linear embedding was then performed with the UMAP algorithm using the top three principal components. UMAP plots were generated to display cellular distributions, and epithelial clusters were manually assigned to AT2 cells, airway epithelial cells, or AT1 cells. Plots were redrawn with discrete color schemes to reflect the refined cell-type annotations.\u003c/p\u003e\n\u003cp\u003ePseudotime Analysis\u003cbr\u003e Pseudotime ordering was performed with Seurat and Monocle 3. The RNA expression matrix and associated metadata were extracted from the Seurat object and imported into Monocle 3, where data were normalized with preprocess_cds. UMAP embedding was computed from the top three principal components. Cell developmental trajectories were inferred via Louvain-based graph partitioning within Monocle 3. Trajectory-associated genes were identified with graph_test (q-value \u0026lt; 1 \u0026times; 10⁻\u0026sup3; and Moran\u0026rsquo;s I \u0026gt; 0.25); the ten most significant genes were selected for visualization. Expression values were log-normalized, spline-smoothed (smooth.spline), and z-score scaled. The top 200 trajectory genes were subjected to k-means clustering and displayed in a heatmap. Temporal gene-expression dynamics were plotted, and cell distribution along pseudotime was overlaid on the UMAP, revealing transcriptional programs underpinning epithelial differentiation.\u003c/p\u003e\n\u003cp\u003eTernary Plot Visualization and Gene Scoring Analysis\u003cbr\u003e Cell-type\u0026ndash;specific marker genes were systematically screened with the Seurat package. Markers for AT1, airway epithelial, and AT2 cells were retained when they met the following cut-offs: adjusted p-value \u0026lt; 0.05, |log₂FC| \u0026gt; 1, and expression in \u0026lt; 40 % of non-target cells. Per-cell lineage scores were then computed as the mean expression of the corresponding markers: score = rowSums(speci_raw) / length(Lineage_marker), where speci_raw denotes the marker-gene expression matrix. To visualize the relative contribution of the three lineage signatures to each cell, ternary plots were generated with ggtern; the three vertices correspond to AT1, airway, and AT2 scores, and the position of each point reflects its lineage affiliation. Distinct ternary plots were produced for normal, LUAD, and Treat.TD1\u0026ndash;9 samples (encompassing in-situ, micro-invasive, and invasive stages). Module scores were calculated with AddModuleScore and projected onto UMAP embeddings (FeaturePlot). Box-and-jitter plots (ggplot2) further displayed score distributions across Seurat clusters.\u003c/p\u003e\n\u003cp\u003eScreening of Core Feature Genes Using 101 Machine Learning Algorithms\u003cbr\u003e Differentially expressed genes (log₂FC \u0026gt; 1 and adjusted p \u0026lt; 0.05) between cluster 11 and cluster 26 were identified and subjected to intersection analysis. All downstream modelling was performed with the Mime1 package. Prognostic models were constructed with ML.Dev.Prog.Sig using the candidate genes and training data, then validated across independent datasets; model performance was quantified by the concordance index (C-index) and Kaplan\u0026ndash;Meier survival analyses. Time-dependent Area Under the Curves (AUCs) for 1- and 3-year survival were calculated and plotted for cross-dataset comparison. Core predictive genes were subsequently extracted via ML.Corefeature.Prog.Screen, and their rankings were visualized with an UpSet plot and a gene-importance bar chart.\u003c/p\u003e\n\u003cp\u003eTranscriptomic Profiling and Functional Pathway Analysis of MYO6+ Epithelial Cell Subpopulations\u003cbr\u003e Epithelial subpopulations were interrogated with Seurat. Cells were dichotomized into MYO6⁺ and MYO6⁻ groups using AddMetaData, with the median MYO6 transcript level as the cutoff. Cellular distributions in low-dimensional space were visualized with t-SNE embeddings (RunTSNE) and DimPlot. Differential expression between MYO6⁺ and MYO6⁻ populations was assessed by FindMarkers (min.pct = 0.25, logfc.threshold = 0.25); genes with log₂FC \u0026gt; 1 and P \u0026lt; 0.05 were considered significant. The resulting differentially expressed genes (DEGs) were subjected to Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses via enrichGO and enrichKEGG, respectively[18-20].\u003c/p\u003e\n\u003cp\u003ePrognostic analysis\u003c/p\u003e\n\u003cp\u003eThe prognostic value of MYO6 expression across multiple cancer types was evaluated by overall survival (OS) analysis on the GEPIA2 web platform (http://gepia2.cancer-pku.cn/).\u003c/p\u003e\n\u003cp\u003eTissue Sample Collection\u003c/p\u003e\n\u003cp\u003eFrom October 2022 to November 2022, tissue samples were collected from 12 LUAD patients who underwent surgical treatment at Nantong University Affiliated Hospital. All patients\u0026apos; LUAD was confirmed by pathological examination. We removed all tissues and stored them in a -80 \u0026deg; C freezer for further research. Provide tumor samples and corresponding normal samples to each patient at the edge of cancer.\u003c/p\u003e\n\u003cp\u003eEthics statement\u003c/p\u003e\n\u003cp\u003eAll procedures involving human participants were conducted in compliance with the ethical standards set by the institutional and national research committees, in alignment with the 1964 Helsinki Declaration and its subsequent amendments. The project was approved by Ethics Committee of the affiliated hospital of Tong University (license number: 2021-L142). Informed consent was obtained from all participants involved in the research. All participants gave informed consent to be included in the study. We fully adhere to SAGER guidelines and our study design was not related to gender.\u003c/p\u003e\n\u003cp\u003eCell Culture\u003c/p\u003e\n\u003cp\u003eH1299, BEAS-2B and A549 cells were obtained from the American Type Culture Collection (ATCC). H1299 cells were maintained in RPMI-1640 (Cytiva) supplemented with 10 % fetal bovine serum (FBS). BEAS-2B and A549 cells were cultured in Dulbecco\u0026rsquo;s Modified Eagle Medium (DMEM; Fuhong Biology) containing 10 % FBS (Gibco). All lines were incubated at 37 \u0026deg;C with 5 % CO₂ in a humidified atmosphere (Thermo Scientific).\u003c/p\u003e\n\u003cp\u003eQuantitative Real-Time PCR (qRT-PCR)\u003c/p\u003e\n\u003cp\u003eTotal RNA was extracted from cultured cells using TRIzol reagent (Invitrogen, Waltham, MA, USA) and reverse-transcribed to cDNA with a commercial kit (Vazyme, Nanjing, China). qRT-PCR was performed with SYBR Green PCR Master Mix (Takara) in 10 \u0026micro;L reactions run in triplicate. Relative mRNA levels were normalized to \u0026beta;-actin and calculated by the 2^\u0026ndash;\u0026Delta;\u0026Delta;Ct method. Primer sequences: MYO6 forward 5\u0026prime;-TGCCGACCAGTACAAAGACC-3\u0026prime;, reverse 5\u0026prime;-ATGGGTGGCTTGTCAAGGATG-3\u0026prime;; \u0026beta;-actin forward 5\u0026prime;-CATGTACGTTGCTATCCAGGC-3\u0026prime;, reverse 5\u0026prime;-CTCCTTAATGTCACGCACGAT-3\u0026prime;.\u003c/p\u003e\n\u003cp\u003eProtein Extraction and Western Blotting\u003c/p\u003e\n\u003cp\u003eCells or tissues were lysed in RIPA buffer containing 1 % protease-inhibitor cocktail, incubated on ice for 30 min, and centrifuged at 12 000 \u0026times; g, 4 \u0026deg;C, 15 min. Protein concentrations were determined with a BCA assay kit. Equal amounts (30 \u0026micro;g per lane) were denatured in 5\u0026times; loading buffer at 95 \u0026deg;C for 10 min, resolved by 10 % SDS-PAGE, and transferred to PVDF membranes. After blocking with 5 % non-fat milk, membranes were incubated overnight at 4 \u0026deg;C with primary antibodies: MYO6 (1:5000; 26778-1-AP, Proteintech, China) and \u0026beta;-actin (1:20 000; 66009-1-Ig, Proteintech, China). Following extensive washing, membranes were probed with HRP-conjugated secondary antibodies (1:5000) for 2 h at room temperature. Protein bands were visualized with ECL substrate and quantified with ImageLab software using GAPDH as the loading control.\u003c/p\u003e\n\u003cp\u003eStatistical analysis\u003c/p\u003e\n\u003cp\u003eStatistical analysis was performed using GraphPad Prism (version 10) and R packages. The Student\u0026rsquo;s t-test and paired t-test were used for independent and paired groups, respectively. The results for continuous variables were presented as the mean \u0026plusmn; standard deviation. A P value \u0026le;0.05 was considered statistically significant.\u003c/p\u003e"},{"header":"Results","content":"\u003ch2\u003eSingle-cell RNA Sequencing Reveals Cell-type Heterogeneity and Differential Gene Expression in Normal and LUAD Lung Tissues\u003c/h2\u003e\n\u003cp\u003eAs detailed in the Methods section and Supplementary Figure 1, single-cell analysis of 13 samples from GSE149655 and GSE189357 was performed following stringent quality control. All cells were classified into 41 clusters (Figure 1A). Guided by canonical markers, these clusters were assigned to eight major lineages: B cells, endothelial cells, epithelial cells, fibroblasts, macrophages, mast cells, neutrophils, and T cells (Figure 1B). t-SNE visualization and a corresponding heatmap illustrate the spatial distribution and marker-gene profiles of each population (Figure 1C). Comparative analysis revealed significant shifts in the relative abundance of all eight cell types between LUAD and normal lung tissue (Figure 1D). DEGs identified distinct gene signatures for each lineage across the \u0026ldquo;Normal\u0026rdquo; and \u0026ldquo;LUAD\u0026rdquo; groups (Figure 1E), and further confirmed proportional alterations in these populations between control and treatment conditions (Figure 1F).\u003c/p\u003e\n\u003ch2\u003eSingle-cell Transcriptomic Analysis Reveals Heterogeneity and Molecular Signatures of Epithelial Cell Subpopulations in Normal and LUAD Lung Tissues\u003c/h2\u003e\n\u003cp\u003eUMAP embedding revealed the spatial distribution of epithelial cells within the \u0026ldquo;Normal\u0026rdquo; cohort (Figure 2A). Feature plots for canonical lineage markers (SFTPA1, SFTPA2, HOPX, and SCGB1A1) highlighted their cell-type\u0026ndash;specific expression patterns (Figure 2B). Guided by these signatures, clusters 8, 11 and 18 were annotated as AT2 cells, clusters 33 and 35 as airway epithelial cells, and cluster 29 as AT1 cells (Figure 2C). When epithelia from both \u0026ldquo;Normal\u0026rdquo; and \u0026ldquo;LUAD\u0026rdquo; samples were projected into the same latent space, their distinct spatial organization became evident (Figure 2D). Accordingly, epithelial subpopulations were extracted according to these annotations (Figure 2E). Refinement of the classification assigned clusters 8, 11, 18 and 26 to the AT2 lineage, clusters 33 and 35 to airway epithelial cells, and cluster 29 to AT1 cells (Figure 2F).\u003c/p\u003e\n\u003ch2\u003eTernary Plot Analysis Combined with Single-cell Transcriptomics Resolves the Dynamic Evolution of Epithelial Marker Genes and Identifies Heterogeneous Cell Subpopulations\u003c/h2\u003e\n\u003cp\u003eTernary plots showed that normal epithelial cells clustered near the AT2 vertex (Figure 3A), whereas LUAD cells accumulated in the central region (Figure 3B). Adenocarcinoma in situ (AIS) cells began to shift toward the centre or an edge of the triangle (Figure 3C), and micro-invasive carcinoma cells displayed an even wider dispersion (Figure 3D). Invasive carcinoma cells converged at the centroid with the broadest distribution (Figure 3E). This reflects the dynamic changes in cell type marker gene expression during tumor progression from normal to invasive stages. In our study, the distribution of gene expression scores for AT1, AT2, and Airway Epithelial Cells was analyzed using UMAP plots. The results demonstrated significant heterogeneity in gene scores among these cells. However, we identified a subpopulation of cells that did not exhibit significant gene expression in any of the three scoring sets (Figure 3F-H). Three boxplots combined with jitter plots illustrate the distribution characteristics of gene scores across different cell clusters within the AT1, AT2, and Airway epithelial cell types. Notably, cell clusters 11 and 26 consistently exhibited low gene scores across all categories (Figure 3I-K). Furthermore, the spatial location of clusters 11 and 26 on the UMAP plot coincided with the location of cells lacking significant gene expression.\u003c/p\u003e\n\u003ch2\u003ePseudotime Analysis\u003c/h2\u003e\n\u003cp\u003eTo investigate dynamic changes during cellular development, a cell developmental trajectory plot illustrates the differentiation paths connecting AT2 Cells, AT1 Cells, and Airway Epithelial Cells (Figure 4A). The UMAP pseudotime trajectory plot presents the UMAP dimensionality reduction visualization of the single-cell transcriptomic data, depicting cell distribution along the developmental trajectory (Figure 4B). We visualized the expression levels of the top 200 feature genes in each epithelial cell (Figure 4C). This visualization clearly demonstrates the dynamic changes in gene expression, further validating the accuracy of the epithelial cell pseudotime analysis. To study the differential expression of key genes across cell populations, the top 10 genes significantly associated with the developmental trajectory were selected for visualization (Figure 4D). Pseudotime analysis of these significant genes revealed distinct dynamic expression patterns across pseudotime. These genes exhibited varying expression trends in different cell clusters (Figure 4E).\u003c/p\u003e\n\u003ch2\u003eScreening of Core Feature Genes Using a Combination of 101 Machine Learning Algorithms\u003c/h2\u003e\n\u003cp\u003eA total of 140 intersecting feature genes were identified from cell cluster 11 and cell cluster 26 (Figure 5A). These 140 genes were selected for model construction. Using the R package `Mime1`, prognostic risk models were built, and the C-index for each model is displayed (Figure 6A). The StepCox[forward] + Ridge model was selected for further analysis. The C-index of the StepCox[forward] + Ridge model across different datasets is shown (Figure 6B). Survival curves for high-risk and low-risk groups were plotted in both the training and test datasets. The results indicate significantly worse survival prognosis for the high-risk group in both datasets (Figure 6C). One-year and three-year survival prediction Receiver Operating Characteristic (ROC) curves were generated for the training and test sets (Figure 6D). The AUC values for one-year and three-year survival prediction were calculated in both datasets (Figure 6E). The results demonstrate that this risk model can effectively predict the survival of LUAD patients. Core feature gene selection was performed using an Upset plot showing genes filtered by different methods. A ranking plot displays the genes filtered by various methods, revealing that MYO6, ASPH, and KRT8 were the most frequently selected genes (Figure 6F-G).\u003c/p\u003e\n\u003ch2\u003eSpatial Distribution of MYO6 Gene Expression in Epithelial Cells and GO/KEGG Enrichment Analysis\u003c/h2\u003e\n\u003cp\u003eSpatial mapping of MYO6 expression distinguished MYO6⁻ and MYO6⁺ epithelial cells in two-dimensional space (Figure 7A). DEGs between these populations were subjected to GO and KEGG enrichment analyses. GO revealed marked enrichment in biological processes (e.g., phosphatidylcholine metabolism regulation, Hippo signalling), cellular components (multivesicular bodies, tight junctions) and molecular functions (protein-tyrosine-kinase activator activity, actin binding). KEGG analysis further identified significant involvement of fatty-acid metabolism, signalling cascades and disease-associated pathways, with the MAPK signalling pathway being most prominent (Figure 7B-C).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMYO6 is highly expressed in both LUAD cells and tissues\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eGEPIA2 analysis revealed that MYO6 mRNA levels were significantly elevated in 483 LUAD tissues relative to 347 normal lung controls (Figure 8A). Kaplan\u0026ndash;Meier survival curves further demonstrated that high MYO6 expression conferred a markedly poorer prognosis (Figure 8B). Compared to the normal bronchial epithelial cell line 16HBE, significantly higher levels of MYO6 mRNA were observed in both LUAD cell lines A549 and H1299 (Figure 8C). Similarly, compared to adjacent non-tumorous tissues, MYO6 protein levels were significantly upregulated in LUAD tissues (Figure 8D-E).\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eClassical models of tumorigenesis posit that malignant transformation reflects an aberrant recapitulation of normal developmental programmes, typically initiated in stem or progenitor cells whose differentiation is arrested [21-24]. Accumulating evidence underscores the marked plasticity of tumour cell phenotypes [25]; for example, loss of LKB1 can drive lineage switching from lung adenocarcinoma to squamous cell carcinoma [26]. Nevertheless, the precise cellular origins of most human malignancies\u0026mdash;particularly the epithelial cancers that account for over 90 % of all tumours\u0026mdash;remain poorly defined [27-29]. Consequently, the cellular identity of adenocarcinoma, a major epithelial subtype, is still incompletely understood.\u003c/p\u003e\n\u003cp\u003eSingle-cell transcriptomic profiling from pre-invasive to invasive LUAD has recently begun to illuminate the molecular underpinnings of disease progression\u0026nbsp;[30]. Here, we leveraged single-cell RNA sequencing to interrogate epithelial cells from normal and LUAD lung tissues, systematically dissecting cellular heterogeneity and lineage-specific transcriptional programs. Consistent with prior work implicating AT2 cells as the potential cell-of-origin for KRAS-driven LUAD\u0026nbsp;[31-33], we resolved distinct epithelial subpopulations\u0026mdash;including AT1, AT2 and airway epithelial cells\u0026mdash;whose relative abundances shifted markedly across pathological states. Ternary-plot and pseudotime analyses revealed a progressive loss of lineage fidelity, manifested as a transition from discrete lineage-specific signatures to a centrally aggregated, dedifferentiated state. Notably, clusters 11 and 26 exhibited uniformly low lineage-specific gene scores, suggesting that they represent transitional intermediates or specialized cellular states within the evolving LUAD microenvironment that may exert unique tumorigenic functions\u0026nbsp;[34].\u003c/p\u003e\n\u003cp\u003eThe progression from normal to malignant lung epithelium proceeds through intermediate states. KACs\u0026mdash;KRT8⁺ alveolar intermediate cells\u0026mdash;represent a transitional population between AT2 cells and tumour cells, exhibiting diminished differentiation, heightened plasticity, and increased susceptibility to KRAS-driven transformation\u0026nbsp;[17]. Similarly, Kim et al. identified transitional alveolar cells during LUAD development that harbour unique molecular signatures implicated in tumour initiation\u0026nbsp;[35]. Analogous transitional epithelial populations have been observed in murine lung injury models and may contribute to early LUAD evolution\u0026nbsp;[36]. The concept of \u0026ldquo;confused cell identity\u0026rdquo; has further been proposed in oesophageal squamous cell carcinoma (ESCC), where tumour cells simultaneously display\u0026mdash;yet diverge from\u0026mdash;features of multiple normal oesophageal epithelial lineages; this identity confusion independently predicts poor prognosis\u0026nbsp;[37]. Nevertheless, the precise roles of epithelial subpopulations within injury niches, and the mechanisms by which they drive progression from normal lung to pre-invasive lesions and invasive LUAD, remain undefined.\u003c/p\u003e\n\u003cp\u003eFrom the 140 intersecting genes of clusters 11/26, we deployed 101 algorithmic combinations to identify core signature genes and constructed a high-precision prognostic model (StepCox[forward] + Ridge). MYO6, ASPH and KRT8 emerged as central survival-associated genes. We further defined a MYO6-high epithelial subpopulation (MYO6⁺) whose functional enrichment implicates MYO6-mediated activation of the MAPK axis in tumour progression. Thus, MYO6 constitutes a potential prognostic biomarker for LUAD. Differential expression of MYO6 within epithelial cells critically influences tumour behaviour; genes up-regulated in MYO6⁺ versus MYO6⁻ cells are enriched in multiple biological processes and, most prominently, the MAPK pathway.\u003c/p\u003e\n\u003cp\u003eMYO6 (Myosin VI), a unique minus-end-directed motor protein, has been implicated as an oncogene in various cancers (e.g., prostate, colorectal, breast). Elevated MYO6 expression is consistently linked to tumor invasion, metastasis, and poor prognosis\u0026nbsp;[38-42]. Mechanistically, the AR/MYO6/FAK axis drives malignant progression in castration-resistant prostate cancer\u0026nbsp;[39],whereas MYO6 deletion suppresses proliferation and attenuates ERK1/2 and PRAS40 phosphorylation\u0026nbsp;[43]. In breast cancer, MYO6 up-regulation activates the MAPK/ERK pathway and accelerates tumourigenesis\u0026nbsp;[41]. These pan-cancer data corroborate our findings implicating MYO6 in LUAD pathogenesis.\u003c/p\u003e\n\u003cp\u003eAlthough this study provides valuable insights, several limitations merit consideration. First, the exclusive use of publicly available datasets restricts sample size; validation in larger, independent cohorts is imperative. Second, while we integrated multiple analytical strategies for cell-type annotation and expression profiling, residual subjectivity in lineage assignment may affect interpretation under specific contexts. Future work should address these constraints by expanding sample numbers, incorporating orthogonal experimental platforms, and conducting rigorous clinical validation.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThis study integrates scRNA-seq analysis with ternary plots and machine learning to precisely identify tumor-specific epithelial subpopulations and nominate MYO6 as a novel prognostic biomarker for LUAD. Moreover, we provide evidence suggesting MYO6 may drive tumor progression via the MAPK signaling pathway, offering a direction for future mechanistic investigations. This comprehensive analytical approach not only deepens our understanding of epithelial cell heterogeneity in LUAD but also unveils promising novel therapeutic targets for future diagnostic and therapeutic strategies.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003eLUAD: Lung adenocarcinoma\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;scRNA-seq: Single-cell RNA sequencing\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;TME: Tumor microenvironment\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;AT1: Alveolar type 1\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;AT2: Alveolar type 2\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;KACs: KRT8+ alveolar intermediate cells\u003c/p\u003e\n\u003cp\u003eKEGG: Kyoto Encyclopedia of Genes and Genomes\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;GO: Gene Ontology\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;TCGA: The Cancer Genome Atlas\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;UMAP: Uniform Manifold Approximation and Projection\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;PCA: Principal Component Analysis\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;t-SNE: t-distributed Stochastic Neighbor Embedding\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;DEGs: Differentially expressed genes\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;AUC: Area Under the Curve\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;HR: Hazard ratio\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;MAPK: Mitogen-activated protein kinase\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;qRT-PCR: Quantitative real-time PCR\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;cDNA: Complementary DNA\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;DMEM: Dulbecco\u0026rsquo;s Modified Eagle Medium\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;FBS: Fetal bovine serum\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;RPMI-1640: Roswell Park Memorial Institute 1640 medium\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;PVDF: Polyvinylidene difluoride\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;HRP: Horseradish peroxidase\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;ECL: Enhanced chemiluminescence\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;GAPDH: Glyceraldehyde-3-phosphate dehydrogenase\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;AIS: Adenocarcinoma in situ\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;C-index: Concordance Index\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;ROC: Receiver Operating Characteristic\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;ESCC: Esophageal squamous cell carcinoma\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;AR: Androgen receptor\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;FAK: Focal adhesion kinase\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;ERK: Extracellular signal-regulated kinase\u003c/p\u003e\n\u003cp\u003ePRAS40: Proline-rich Akt substrate of 40 kDa\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors acknowledge the efforts of all of the researchers who have contributed the data to the public databases of TCGA and GEO. The interpretation and reporting of these data are the sole responsibility of the authors.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was supported by the Natural Science Research Project of Nantong Science and Technology Bureau (No. MS12020029).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor information\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAuthor notes\u003c/p\u003e\n\u003cp\u003eMiaoyan Liu and Houqiang Li contributed equally to this work.\u003c/p\u003e\n\u003cp\u003eAuthors and Affiliations\u003c/p\u003e\n\u003cp\u003eDepartment of Thoracic Surgery, Affiliated Hospital of Nantong University, Nantong, 226001,China\u003c/p\u003e\n\u003cp\u003eMiaoyan Liu, Houqiang Li, Shenghan Xu, Yike Zhou, Tiegang Cao , Jiahai Shi, Lou Zhong\u003c/p\u003e\n\u003cp\u003eDepartment of Immunology, Medical School of Nantong University \u0026amp; Research Center of Clinical Medicine, Affiliated Hospital of Nantong University,Nantong, 226019,China\u003c/p\u003e\n\u003cp\u003eMin Yao\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eContributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eLZ and JHS provided guidance throughout the preparation of this manuscript. MYL and HQL conceptualised and wrote the manuscript. HQL,SHX ,YKZ , TGC, and MY analysed and interpreted the data. MYL conducted the experiments and collected the data. MYL, JHS and LZ designed the experiments and analysed the results. All authors approved the final version.\u003c/p\u003e\n\u003cp\u003eCorresponding authors\u003c/p\u003e\n\u003cp\u003eCorrespondence to Lou Zhong and Jiahai Shi.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics declarations\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll human samples procedures were approved by the Ethics Committee of the Affiliated Hospital of Nantong University (license number: 2021-L142).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability \u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll datasets used in this study are publicly available from the following repositories: Single-cell RNA-seq data: GSE149655 and GSE189357 from the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/). \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eBulk RNA-seq data: TCGA-LUAD cohort from the Genomic Data Commons (GDC) Data Portal (https://portal.gdc.cancer.gov/). Validation datasets (GSE31210 and GSE30219) from GEO (https://www.ncbi.nlm.nih.gov/geo/). \u0026nbsp;\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eBray F, Ferlay J, Soerjomataram I, Siegel RL, Torre LA, Jemal A. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2018;68:394-424.\u003c/li\u003e\n\u003cli\u003eDevarakonda S, Morgensztern D, Govindan R. Genomic alterations in lung adenocarcinoma. Lancet Oncol. 2015;16:e342-51.\u003c/li\u003e\n\u003cli\u003eSong C, Guo Z, Yu D, Wang Y, Wang Q, Dong Z, Hu W. A Prognostic Nomogram Combining Immune-Related Gene Signature and Clinical Factors Predicts Survival in Patients With Lung Adenocarcinoma. Front Oncol. 2020;10:1300.\u003c/li\u003e\n\u003cli\u003eSha Y, Wang S, Zhou P, Nie Q. Inference and multiscale model of epithelial-to-mesenchymal transition via single-cell transcriptomic data. Nucleic Acids Res. 2020;48:9505-20.\u003c/li\u003e\n\u003cli\u003eHe D, Wang D, Lu P, Yang N, Xue Z, Zhu X, Zhang P, Fan G. Single-cell RNA sequencing reveals heterogeneous tumor and immune cell populations in early-stage lung adenocarcinomas harboring EGFR mutations. Oncogene. 2021;40:355-68.\u003c/li\u003e\n\u003cli\u003eLi D, Yu H, Hu J, Li S, Yan Y, Li S, Sun L, Jiang G, Hou L, Zhang L, et al. Comparative profiling of single-cell transcriptome reveals heterogeneity of tumor microenvironment between solid and acinar lung adenocarcinoma. J Transl Med. 2022;20:423.\u003c/li\u003e\n\u003cli\u003eTravaglini KJ, Nabhan AN, Penland L, Sinha R, Gillich A, Sit RV, Chang S, Conley SD, Mori Y, Seita J, et al. A molecular cell atlas of the human lung from single-cell RNA sequencing. Nature. 2020;587:619-25.\u003c/li\u003e\n\u003cli\u003eZhu J, Fan Y, Xiong Y, Wang W, Chen J, Xia Y, Lei J, Gong L, Sun S, Jiang T. Delineating the dynamic evolution from preneoplasia to invasive lung adenocarcinoma by integrating single-cell RNA sequencing and spatial transcriptomics. Exp Mol Med. 2022;54:2060-76.\u003c/li\u003e\n\u003cli\u003eZepp JA, Morrisey EE. Cellular crosstalk in the development and regeneration of the respiratory system. Nat Rev Mol Cell Biol. 2019;20:551-66.\u003c/li\u003e\n\u003cli\u003eIwasaki A, Foxman EF, Molony RD. Early local immune defences in the respiratory tract. Nat Rev Immunol. 2017;17:7-20.\u003c/li\u003e\n\u003cli\u003ePark KS, Liang MC, Raiser DM, Zamponi R, Roach RR, Curtis SJ, Walton Z, Schaffer BE, Roake CM, Zmoos AF, et al. Characterization of the cell of origin for small cell lung cancer. Cell Cycle. 2011;10:2806-15.\u003c/li\u003e\n\u003cli\u003eJia Y, You X, Ma N, Li H, Liu G, Wang Y, Xue J, Shi J, Wei J, Li Y, et al. Phenotypic Analysis of BrdU Label-Retaining Cells during the Maturation of Conducting Airway Epithelium in a Porcine Lung. Stem Cells Int. 2019;2019:7043890.\u003c/li\u003e\n\u003cli\u003eWeitnauer M, Mijo\u0026scaron;ek V, Dalpke AH. Control of local immunity by airway epithelial cells. Mucosal Immunol. 2016;9:287-98.\u003c/li\u003e\n\u003cli\u003eSainz de Aja J, Dost A, Kim CF. Alveolar progenitor cells and the origin of lung cancer. J Intern Med. 2021;289:629-35.\u003c/li\u003e\n\u003cli\u003eYang M, Shen H, Flodby P, Koss MD, Bassiouni R, Liu Y, Jashashvili T, Neely A, Ogbolu E, Castillo J, et al. Alveolar type I cells can give rise to KRAS-induced lung adenocarcinoma. Cell Rep. 2023;42:113286.\u003c/li\u003e\n\u003cli\u003eRosigkeit S, Kruchem M, Thies D, Kreft A, Eichler E, Boegel S, Jansky S, Siegl D, Kaps L, Pickert G, et al. Definitive evidence for Club cells as progenitors for mutant Kras/Trp53-deficient lung cancer. Int J Cancer. 2021;149:1670-82.\u003c/li\u003e\n\u003cli\u003eHan G, Sinjab A, Rahal Z, Lynch AM, Treekitkarnmongkol W, Liu Y, Serrano AG, Feng J, Liang K, Khan K, et al. An atlas of epithelial cell states and plasticity in lung adenocarcinoma. Nature. 2024;627:656-63.\u003c/li\u003e\n\u003cli\u003eKanehisa M, Furumichi M, Sato Y, Matsuura Y, Ishiguro-Watanabe M. KEGG: biological systems database as a model of the real world. Nucleic Acids Res. 2025;53:D672-672D677.\u003c/li\u003e\n\u003cli\u003eKanehisa M. Toward understanding the origin and evolution of cellular organisms. Protein Sci. 2019;28:1947-51.\u003c/li\u003e\n\u003cli\u003eKanehisa M, Goto S. KEGG: kyoto encyclopedia of genes and genomes. Nucleic Acids Res. 2000;28:27-30.\u003c/li\u003e\n\u003cli\u003eBatlle E, Clevers H. Cancer stem cells revisited. Nat Med. 2017;23:1124-34.\u003c/li\u003e\n\u003cli\u003eMagee JA, Piskounova E, Morrison SJ. Cancer stem cells: impact, heterogeneity, and uncertainty. Cancer Cell. 2012;21:283-96.\u003c/li\u003e\n\u003cli\u003eKelly LM, Gilliland DG. Genetics of myeloid leukemias. Annu Rev Genomics Hum Genet. 2002;3:179-98.\u003c/li\u003e\n\u003cli\u003eBlanpain C, Horsley V, Fuchs E. Epithelial stem cells: turning over new leaves. Cell. 2007;128:445-58.\u003c/li\u003e\n\u003cli\u003eNiederst MJ, Sequist LV, Poirier JT, Mermel CH, Lockerman EL, Garcia AR, Katayama R, Costa C, Ross KN, Moran T, et al. RB loss in resistant EGFR mutant lung adenocarcinomas that transform to small-cell lung cancer. Nat Commun. 2015;6:6377.\u003c/li\u003e\n\u003cli\u003eHan X, Li F, Fang Z, Gao Y, Li F, Fang R, Yao S, Sun Y, Li L, Zhang W, et al. Transdifferentiation of lung adenocarcinoma in mice with Lkb1 deficiency to squamous cell carcinoma. Nat Commun. 2014;5:3261.\u003c/li\u003e\n\u003cli\u003eNetwork CGA. Comprehensive molecular portraits of human breast tumours. Nature. 2012;490:61-70.\u003c/li\u003e\n\u003cli\u003ePerou CM, S\u0026oslash;rlie T, Eisen MB, van de Rijn M, Jeffrey SS, Rees CA, Pollack JR, Ross DT, Johnsen H, Akslen LA, et al. Molecular portraits of human breast tumours. Nature. 2000;406:747-52.\u003c/li\u003e\n\u003cli\u003eDotto GP, Rustgi AK. Squamous Cell Cancers: A Unified Perspective on Biology and Genetics. Cancer Cell. 2016;29:622-37.\u003c/li\u003e\n\u003cli\u003eXiang C, Zhang M, Shang Z, Chen S, Zhao J, Ding B, Jiang D, Zhu Q, Teng H, Zhu L, et al. Single-cell profiling reveals the trajectory of FOLR2-expressing tumor-associated macrophages to regulatory T cells in the progression of lung adenocarcinoma. Cell Death Dis. 2023;14:493.\u003c/li\u003e\n\u003cli\u003eLin C, Song H, Huang C, Yao E, Gacayan R, Xu SM, Chuang PT. Alveolar type II cells possess the capability of initiating lung tumor development. PLoS One. 2012;7:e53817.\u003c/li\u003e\n\u003cli\u003eLiu B, Li C, Li Z, Wang D, Ren X, Zhang Z. An entropy-based metric for assessing the purity of single cell populations. Nat Commun. 2020;11:3155.\u003c/li\u003e\n\u003cli\u003eXu X, Rock JR, Lu Y, Futtner C, Schwab B, Guinney J, Hogan BL, Onaitis MW. Evidence for type II cells as cells of origin of K-Ras-induced distal lung adenocarcinoma. Proc Natl Acad Sci U S A. 2012;109:4910-5.\u003c/li\u003e\n\u003cli\u003eLiebler JM, Marconett CN, Juul N, Wang H, Liu Y, Flodby P, Laird-Offringa IA, Minoo P, Zhou B. Combinations of differentiation markers distinguish subpopulations of alveolar epithelial cells in adult lung. Am J Physiol Lung Cell Mol Physiol. 2016;310:L114-20.\u003c/li\u003e\n\u003cli\u003eXu JY, Zhang C, Wang X, Zhai L, Ma Y, Mao Y, Qian K, Sun C, Liu Z, Jiang S, et al. Integrative Proteomic Characterization of Human Lung Adenocarcinoma. Cell. 2020;182:245-61.e17.\u003c/li\u003e\n\u003cli\u003eTata PR, Chow RD, Saladi SV, Tata A, Konkimalla A, Bara A, Montoro D, Hariri LP, Shih AR, Mino-Kenudson M, et al. Developmental History Provides a Roadmap for the Emergence of Tumor Plasticity. Dev Cell. 2018;44:679-93.e5.\u003c/li\u003e\n\u003cli\u003ePan X, Wang J, Guo L, Na F, Du J, Chen X, Zhong A, Zhao L, Zhang L, Zhang M, et al. Identifying a confused cell identity for esophageal squamous cell carcinoma. Signal Transduct Target Ther. 2022;7:122.\u003c/li\u003e\n\u003cli\u003eKruppa AJ, Kishi-Itakura C, Masters TA, Rorbach JE, Grice GL, Kendrick-Jones J, Nathan JA, Minczuk M, Buss F. Myosin VI-Dependent Actin Cages Encapsulate Parkin-Positive Damaged Mitochondria. Dev Cell. 2018;44:484-99.e6.\u003c/li\u003e\n\u003cli\u003eZheng S, Hong Z, Tan Y, Wang Y, Li J, Zhang Z, Feng T, Hong Z, Lin G, Ye D. MYO6 contributes to tumor progression and enzalutamide resistance in castration-resistant prostate cancer by activating the focal adhesion signaling pathway. Cell Commun Signal. 2024;22:517.\u003c/li\u003e\n\u003cli\u003eJiang M, Xiong C. USP7 accelerates colorectal cancer progression by up-regulating MYO6 through deubiquitination. Mutat Res. 2025;831:111908.\u003c/li\u003e\n\u003cli\u003eZhan XJ, Wang R, Kuang XR, Zhou JY, Hu XL. Elevated expression of myosin VI contributes to breast cancer progression via MAPK/ERK signaling pathway. Cell Signal. 2023;106:110633.\u003c/li\u003e\n\u003cli\u003eYu H, Zhu Z, Chang J, Wang J, Shen X. Lentivirus-Mediated Silencing of Myosin VI Inhibits Proliferation and Cell Cycle Progression in Human Lung Cancer Cells. Chem Biol Drug Des. 2015;86:606-13.\u003c/li\u003e\n\u003cli\u003eWang D, Zhu L, Liao M, Zeng T, Zhuo W, Yang S, Wu W. MYO6 knockdown inhibits the growth and induces the apoptosis of prostate cancer cells by decreasing the phosphorylation of ERK1/2 and PRAS40. Oncol Rep. 2016;36:1285-92. \u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Single-cell RNA sequencing, Lung adenocarcinoma,Epithelial cell, Ternary plot, Machine learning screening, MYO6","lastPublishedDoi":"10.21203/rs.3.rs-7742292/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7742292/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eLung adenocarcinoma (LUAD) exhibits profound cellular heterogeneity, with epithelial subpopulations playing pivotal yet incompletely characterized roles in tumor progression. Here, we sought to delineate the key heterogeneous epithelial subpopulations and to uncover the molecular determinants that govern their influence on LUAD prognosis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe integrated two publicly available single-cell RNA-seq datasets derived from normal lung and LUAD tissues. Epithelial cells were extracted and subpopulations annotated via Seurat-based clustering. Transcriptional dynamics during malignant transformation were quantified using Monocle3 trajectory inference and ternary plots. A machine-learning framework comprising 101 algorithmic combinations screened signature genes from poorly differentiated clusters (11/26). Prognostic models were constructed and validated in TCGA cohorts. MYO6⁺ epithelial cells were functionally characterized through differential-expression analysis and GO/KEGG pathway enrichment.Verification of MYO6 mRNA levels and protein expression levels by quantitative real-time PCR (qRT-PCR) and Western blot.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSingle-cell profiling demonstrated that LUAD epithelial cells transition from lineage-specific identities (AT1, AT2, airway) to a dedifferentiated state. Cluster 11/26 displayed low lineage-specific gene scores and selective enrichment in LUAD. Machine learning identified MYO6, ASPH, and KRT8 as core prognostic genes among 140 intersecting markers. The StepCox[forward]+Ridge model reliably stratified patients into high- and low-risk groups (HR = 2.48, p \u0026lt; 0.001; 3-year AUC = 0.82). MYO6⁺ epithelial cells were spatially distinct, exhibited MAPK-pathway activation, and correlated with proliferative signatures and adverse outcomes. Compared with 16HBE, MYO6 mRNA was significantly elevated in A549 and H1299 cells; LUAD tissue MYO6 protein levels were also significantly higher than those in adjacent non-cancerous tissue.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusion\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe identify a low-lineage-score epithelial subpopulation (Cluster 11/26) in LUAD whose attenuated lineage-specific gene expression is linked to malignant progression. A StepCox+Ridge prognostic model, built upon core genes derived from this subpopulation, delivers significant clinical stratification. MYO6⁺ cells, a central constituent of this subpopulation, propel tumor progression via MAPK signaling and represent a promising target for precision therapy.\u003c/p\u003e","manuscriptTitle":"MYO6+Epithelial Subpopulation as a Prognostic Hub in Lung Adenocarcinoma Identified by Multi-Omics Integration","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-03-08 14:58:40","doi":"10.21203/rs.3.rs-7742292/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorInvitedReview","content":"","date":"2026-03-01T17:21:08+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"103137259349064744403264483811396763021","date":"2026-03-01T17:13:36+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-02-26T18:15:04+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-01-28T05:59:39+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-10-08T18:40:23+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-10-07T09:18:45+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2025-10-07T08:42:01+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"79e5bc87-3697-4222-8ea3-b30b36d00708","owner":[],"postedDate":"March 8th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":63900101,"name":"Health sciences/Biomarkers"},{"id":63900102,"name":"Biological sciences/Cancer"},{"id":63900103,"name":"Biological sciences/Computational biology and bioinformatics"},{"id":63900104,"name":"Health sciences/Oncology"}],"tags":[],"updatedAt":"2026-03-08T14:58:40+00:00","versionOfRecord":[],"versionCreatedAt":"2026-03-08 14:58:40","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7742292","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7742292","identity":"rs-7742292","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}