Abstract
Head and neck squamous cell carcinoma (HNSCC) is one of the most prevalent and lethal cancers worldwide. Despite multimodal therapeutic advances, long-term survival remains poor, underscoring the need to identify novel molecular drivers of disease aggressiveness. Hypertranscription is a genome-wide increase in total RNA output that has emerged as a hallmark of oncogenic transformation. However, the role of mRNA-specific hypertranscription in HNSCC and its underlying molecular drivers remain undefined. In the present study, we investigated the association between mRNA hypertranscription and malignant phenotypes in HNSCC. Single-cell transcriptomics data revealed that elevated mRNA hypertranscription was significantly associated with the activation of oncogenic pathways and poor clinical outcomes. Through transcription factor activity analysis, we identified the transcription factor Spi-1 Proto-Oncogene (SPI1) as a potential regulator of mRNA hypertranscription in HNSCC malignant cells. Loss- and gain-of-function experiments in HNSCC cell lines and xenograft models established that SPI1 drives cell proliferation, invasion, migration, and tumor growth in vitro and in vivo. Mechanistically, inducible SPI1 overexpression elevated nascent RNA synthesis as measured by EU incorporation, and integrative ChIP-seq/RNA-seq profiling identified direct genomic targets of SPI1 enriched in oncogenic transcriptional programs. Collectively, our findings show that SPI1-driven mRNA hypertranscription is important in HNSCC progression and provide novel insights into the transcriptional dysregulation underlying aggressive malignancies.
Similar content being viewed by others
Introduction
Head and neck squamous cell carcinoma (HNSCC) is the seventh most common cancer worldwide, with more than 1.1 million new cases and accounting for approximately 0.5 million deaths annually.1 More than 60% of HNSCC patients are in a locally advanced stage at their initial diagnosis. Even when they receive appropriate treatment, 40–60% of patients eventually develop recurrence and metastasis.2 Although immunotherapy has been introduced as a first-line treatment for these patients, the overall survival of late-stage HNSCC patients has not significantly improved.3 This high failure rate underscores the need for deeper insights into HNSCC pathogenesis and novel therapeutic strategies.4
While classical studies have focused predominantly on individual oncogenes (e.g., EGFR) or tumor-suppressor genes (e.g., TP53),5,6 such gene-centric approaches often fail to capture broader transcriptional states that may better reflect tumor heterogeneity and malignant potential.7 Recent studies have defined hypertranscription as a transcriptome-wide increase in total RNA output, which is characteristic of highly proliferative cells, including embryonic stem cells and certain cancers.8,9 However, hypertranscription has remained underrecognized in cancer biology,10 largely because technical normalization procedures obscure the variation in total transcript levels across cells or samples.11,12,13 Advances in single-cell RNA sequencing and the application of unique molecular identifiers (UMIs) now enable more accurate quantification of RNA output at the individual cell level.14,15 Although hypertranscription broadly encompasses all RNA species, the specific upregulation of mRNA transcripts represents a biologically distinct dimension of this phenomenon, directly linked to protein-coding capacity and cellular biosynthetic output.16 Elevated levels of such protein-coding transcripts fuel ribosome loading and protein synthesis, supporting the biosynthetic demands of rapid tumor cell proliferation, invasion, and survival.17 Emerging analytical strategies now enable the estimation of global mRNA abundance on the basis of RNA-seq data, revealing that elevated mRNA output is a common and clinically relevant feature across multiple tumor types.16 Building on these observations, we define “mRNA hypertranscription” as a state characterized by widespread upregulation of mRNA transcripts and hypothesize that this transcriptional phenotype may contribute to the malignant progression of HNSCC. This unresolved gap highlights the need to define actionable regulators of transcriptional amplification in HNSCC.
TFs orchestrate complex regulatory networks and may drive hypertranscription in a tumor-specific manner.18 Although c-Myc is a known hypertranscription regulator,19 MYC-centered mechanisms alone may be insufficient to fully explain the extent and heterogeneity of the transcriptional amplification observed in HNSCC. Through integrated single-cell and bulk analyses, we identified Spi-1 Proto-Oncogene (SPI1, also known as PU.1) as a strong candidate orchestrator of mRNA hypertranscription in HNSCC. Previous studies have shown that SPI1 is an ETS family transcription factor located on the chromosome 11p11.2 region.20 It is considered an essential gene involved in the development of myeloid and lymphoid cells,21 and its gene deletion or expression dysregulation is associated with the occurrence of acute myeloid leukemia.22,23 Although the role of SPI1 in hematologic malignancies is context dependent, as it functions as either a tumor suppressor or an oncogene depending on the cell type and differentiation state, its involvement in solid tumors is more consistently tumor promoting.24,25 In these contexts, SPI1 has been shown to enhance malignant phenotypes by regulating glycolysis, angiogenesis, and immune remodeling.26,27,28 Recent methodological advances now allow in vitro quantification of mRNA hypertranscription and SPI1-associated transcriptional effects. At the molecular level, quantifying hypertranscription most reliably involves metabolic labeling with EU. This approach has proven valuable in diverse oncologic contexts, revealing transcriptional dynamics otherwise masked in bulk analysis.29,30,31 For example, in Ewing sarcoma, EU incorporation resulted in constitutively high transcriptional activity and delayed repression upon DNA damage, indicating that EWS-FLI1 is involved in overriding genotoxic checkpoint control.32 However, EU labeling indiscriminately captures all RNA species, including noncoding and precursor transcripts, limiting its specificity for mature mRNA output. To overcome this limitation, poly(A)-selected mRNA purification provides a practical in vitro surrogate for quantifying total mRNA abundance.33,34
In this study, we aimed to systematically characterize the prevalence, transcriptional landscape, and clinical significance of mRNA hypertranscription in HNSCC, to identify its upstream transcriptional regulators through unbiased transcription factor activity inference applied to single-cell data, and to determine whether SPI1 is both necessary and sufficient to drive this transcriptional phenotype and its associated malignant behaviors. To this end, we integrated single-cell and bulk transcriptomic profiling of HNSCC patient tumors with systematic gain- and loss-of-function experiments, EU incorporation-based nascent RNA synthesis assays, integrative ChIP-seq profiling, and independent clinical cohort validation. This multi-modal strategy further enabled us to map the direct genomic targets of SPI1 across the HNSCC transcriptome, to delineate the transcriptional programs underlying its oncogenic activity, and to validate its independent prognostic significance in patient cohorts. Together, these findings provide a mechanistic link between SPI1-associated transcriptional amplification and HNSCC progression, and highlight SPI1 as a candidate therapeutic vulnerability.
Results
mRNA hypertranscription is associated with poor prognosis and tumor progression in HNSCC
To investigate whether mRNA hypertranscription contributes to malignant phenotypes in HNSCC, we performed scRNA-seq on 12 surgical samples, resulting in 116,855 cells that passed stringent quality controls (Supplementary Table 1 and Supplementary Fig. 1a, b). Epithelial cell populations were identified on the basis of the expression of EpCAM, KRT8, and KRT18 (Supplementary Table 2, Supplementary Fig. 1c–f and Fig. 1a, b). From these epithelial populations, we identified and reclustered 18,798 aneuploid malignant cells by inferring genomic copy number variations via copyKAT analysis (Fig. 1c, d). To assess mRNA hypertranscription, we quantified UMI counts per cell, which reflect the number of captured mRNA molecules. After normalization to account for sequencing depth and technical variation using the method introduced by Cao et al.16, normalized total UMI counts serve as a reliable proxy for global transcriptional output at the single-cell level. Normalized total UMI counts were strongly positively correlated with total RNA abundance, raw UMI counts, and the number of genes detected per cell, supporting their use as robust proxies for global transcriptional output at single-cell resolution (Supplementary Fig. 2). Consistent with its global nature, hypertranscription was detectable across multiple cell types in the tumor microenvironment, while malignant epithelial cells exhibited the highest overall levels of transcriptional output (Supplementary Fig. 3). Accordingly, cells were stratified into low, medium, and high hypertranscription groups on the basis of tertile distribution (n = 6266 per group), revealing marked transcriptional heterogeneity across distinct clusters (Fig. 1e). Previous studies have suggested that the total number of genes expressed at the single-cell level may better predict cellular phenotype and developmental potential than individual gene markers or pathways.35 Building on these findings, we analyzed gene expression features across 18,798 tumor cells and reported a strong positive correlation between the number of expressed genes and normalized total UMI counts (R = 0.85, p < 0.0001; Fig. 1f), indicating that mRNA hypertranscription may be linked to cellular differentiation status. CytoTRACE analysis further demonstrated that compared with cells in the medium and low groups, cells in the high hypertranscription group had significantly higher scores, indicating a less differentiated state (p < 0.001; Fig. 1g). The differentially expressed genes in the high- hypertranscription group were significantly enriched in pathways related to protein synthesis (RPS23 and RPL35A), cytoskeletal organization (MYL12A/B and TMSB10), stress response (HSPE1, S100A6, and S100A11), and inflammation/immunity (LYZ and TGFBI; Fig. 1h). Consistently, GO and Reactome enrichment analyses revealed coordinated activation of developmental and SUMOylation-related pathways, reflecting transcriptionally and posttranscriptionally regulated programs likely orchestrated by context-specific transcription factors (Supplementary Fig. 4). Furthermore, we observed significant upregulation of the vast majority (48/50) of tumor-associated hallmark pathways, which correlated positively with increased mRNA hypertranscription levels in cancer cells (Supplementary Table 3 and Fig. 1i). To further validate the clinical relevance of tumor hypertranscription, we assessed its status in 210 HNSCC patients from the TCGA-HNSC cohort using the TmS metric (Supplementary Fig. 5).16 After additionally adjusting for HPV status and other established clinicopathological covariates, tumor hypertranscription remained an independent adverse prognostic factor for overall survival in patients with HNSCC (Supplementary Fig. 6). These findings reinforce the prognostic value of hypertranscription while extending prior work by incorporating HPV-related disease heterogeneity.16
mRNA hypertranscription is associated with poor prognosis and tumor progression in HNSCC. a UMAP projection of all single cells from 12 HNSCC samples, annotated by major cell types. b Dot plot showing the average expression and detection frequency of canonical marker genes used for cell type identification. c Aneuploid and diploid cell prediction using copyKAT analysis. d UMAP of reclustered epithelial tumor cells (n = 18,798), defining 11 subpopulations (epi_0-epi_10). e mRNA hypertranscription status (low, medium, and high) overlaid on the tumor UMAP. f Scatter plot showing the correlation between the normalized total UMI count and the number of expressed genes. g Box plot comparing the CytoTRACE scores among the hypertranscription groups; higher scores reflect lower differentiation. Statistical significance was assessed using the Wilcoxon rank-sum test (all p < 0.001). h Heatmap of differentially expressed genes across mRNA hypertranscription groups. i Heatmap showing enrichment scores of the top 20 hallmark pathways across transcriptional states
SPI1 is a key transcription factor associated with mRNA hypertranscription in HNSCC
Our subsequent efforts focused on identifying TFs associated with HNSCC mRNA hypertranscription and demonstrating their correlation with this enhanced mRNA hypertranscriptional state. TF activity inference analysis identified SPI1 as the most highly activated TF in hypertranscriptional tumors (Fig. 2a) on the basis of a weighted-sum scoring approach leveraging TF-target annotations from the CollecTRI database, a comprehensive resource integrating curated regulatory interactions from 12 distinct sources.36 Notably, downstream target genes of SPI1 were predominantly upregulated in cells with high global mRNA transcriptional activity (Supplementary Fig. 7), suggesting a functional role for SPI1 in driving the hypertranscriptional phenotype. Consistent with its activation status, SPI1 expression was significantly elevated in tumor tissues compared with adjacent normal tissues in bulk RNA-seq data from the TCGA HNSCC cohort (Supplementary Fig. 8). Clinically, patients with high SPI1 activity had markedly poorer overall survival (Fig. 2b). Furthermore, after adjustment for clinicopathological variables, HPV status, and tumor microenvironment-related scores (stromal, immune, and ESTIMATE), high SPI1 activity remained independently associated with poor overall survival in patients with HNSCC (Fig. 2c).
SPI1 is associated with the hypertranscriptional phenotype in HNSCC. a TF activity inference identifying SPI1 as the top candidate. Scatter plot of TF activity scores (weighted t statistic) versus significance. b Kaplan‒Meier analysis showing that patients with high SPI1 activity exhibit significantly worse overall survival. P values were calculated using the log-rank test. c Forest plot derived from an expanded multivariate Cox regression model. High SPI1 activity remained an independent predictor of poor prognosis after we adjusted for clinicopathological factors (age, sex, stage, smoking/alcohol history, tumor subsite, and HPV status) and tumor microenvironment characteristics (stromal, immune, and ESTIMATE scores). d UMAP plots of single cells illustrating graded SPI1 activity (left) and mRNA hypertranscription status (right). e Heatmap of representative differential genes across low-, medium-, and high-SPI1-activity groups, showing distinct gene expression modules associated with SPI1 activation. f Scatter plot showing a positive correlation between SPI1 activity and normalized total UMI counts (R = 0.42, p < 0.0001). g Violin plots of the SPI1 activity across tertile-defined low, medium, and high mRNA hypertranscription groups (n = 6266 cells per group; p < 0.0001). h Heatmap of the top 20 hallmark pathways enriched in cells with high SPI1 activity. i Rank-ordered distribution of TF activity t statistics from an independent single-cell HNSCC dataset (GSE234933)
To investigate the cellular underpinnings of this association, we evaluated SPI1 activity in our scRNA-seq dataset. SPI1 activity was positively correlated with mRNA expression across all major cell types (Supplementary Table 4). Given our primary interest in delineating the relationship between SPI1 activity and hypertranscription within malignant cells, all subsequent analyses were restricted to the aneuploid tumor cell population. UMAP projections revealed marked spatial colocalization between high SPI1 activity and mRNA hypertranscription (Fig. 2d). Differential expression analysis further revealed representative marker genes across low, medium, and high SPI1 activity states (Fig. 2e). Quantitatively, SPI1 activity displayed a strong positive correlation with normalized total UMI counts (R = 0.42; p < 0.0001; Fig. 2f) and increased progressively across hypertranscription tertiles (Fig. 2g), reinforcing its link to elevated transcriptional output at the single-cell level. Consistently, analysis of HNSCC cell lines from the Cancer Cell Line Encyclopedia (CCLE) revealed a strong positive correlation between SPI1 regulon-based activity and global transcriptional output (Pearson r = 0.706, P < 0.001; Supplementary Fig. 9). Transcriptomic profiling of cells stratified by SPI1 activity revealed that high-SPI1 cells exhibited activation of many of the same pathways observed in hypertranscriptional cells, especially those linked to RNA processing and nuclear receptor signaling (Supplementary Fig. 10). Notably, the AUCell scores of cancer hallmark pathways revealed that most hallmark pathways were significantly enriched in the high-SPI1-activity group (41/50, all adjusted p < 0.05; Fig. 2h; Supplementary Table 5), indicating that SPI1 is involved in widespread oncogenic reprogramming. To strengthen these findings, we analyzed an independent single-cell dataset comprising 187,399 cells from 51 HNSCC patients (Supplementary Fig. 11a, b). This external validation confirmed the correlation between SPI1 activity and mRNA hypertranscription (Supplementary Fig. 11c–e, Fig. 2i) and recapitulated the transcriptional enrichment of oncogenic pathways in cells with high SPI1 activity (Supplementary Fig. 11f, g). Together, these multilevel analyses establish SPI1 as a key transcriptional regulator associated with mRNA hypertranscription in HNSCC.
SPI1 is associated with the malignant phenotypes of HNSCC cells in vitro
To define the role of SPI1 in HNSCC progression, we systematically evaluated its expression, subcellular localization, and functional effects in a panel of representative HNSCC cell lines and the immortalized human keratinocyte line HaCaT. RT‒qPCR and western blot analyses demonstrated that SPI1 was abundantly expressed in CAL27 and CAL33 cells but remained low in SAS and FaDu cells (Fig. 3a, b and Supplementary Fig. 12). Immunofluorescence staining and nuclear–cytoplasmic fractionation followed by Western blot analysis demonstrated that SPI1 was predominantly localized to the nucleus, with a minor fraction detectable in the cytoplasm (Fig. 3c and Supplementary Fig. 13). We next performed loss-of-function experiments by generating stable SPI1 knockdown lines in CAL27 and CAL33 cells using shRNA. The knockdown efficiency was confirmed at both the transcript and protein levels (Fig. 3d and Supplementary Fig. 14). Functionally, SPI1 silencing significantly reduced cell proliferation (Fig. 3e) and impaired colony-forming ability (Fig. 3f, g). Consistent with the impaired proliferative capacity, the results of the EdU incorporation assays revealed a significant decrease in DNA synthesis following SPI1 knockdown, indicating reduced S-phase entry (Fig. 3h, i). Additionally, SPI1 knockdown markedly suppressed cell migration and invasion, as determined by Transwell assays (Fig. 3j–l). In parallel, western blot analysis revealed a reduction in RB phosphorylation without appreciable changes in total RB levels in SPI1 knockdown cells (Supplementary Fig. 15), which is consistent with attenuated cell cycle progression. Moreover, SPI1 depletion increased apoptosis, as assessed by Annexin V/7-AAD staining and TUNEL assays (Supplementary Figs. 16 and 17). To complement these findings, we overexpressed SPI1 in SAS and FaDu cells using lentiviral vectors (Fig. 3m and Supplementary Fig. 14), resulting in the opposite phenotype: increased proliferation (Fig. 3n), enhanced clonogenicity (Fig. 3o, p), and elevated motility and invasion (Fig. 3q, s). These comprehensive gain- and loss-of-function assays demonstrate that SPI1 plays a critical role in maintaining the malignant phenotype of HNSCC cells in vitro.
SPI1 regulates proliferation, survival, migration, and invasion in HNSCC cells in vitro. a Representative Western blot analysis showing the SPI1 protein levels in the normal human keratinocyte cell line HaCaT and the head and neck squamous cell carcinoma cell lines SAS, FaDu, CAL27, and CAL33. β-actin was used as a loading control. b Quantification of SPI1 protein expression normalized to that of β-actin and expressed relative to that in HaCaT cells. Data are presented as the mean ± SD from independent experiments. Statistical significance was assessed using one-way ANOVA; P < 0.05 is indicated by an asterisk. c Immunofluorescence staining showing the nuclear localization of SPI1 in CAL27 and CAL33 cells; the right panels show line-scan intensity profiles. Scale bar, 20 μm. d Western blot confirmation of the SPI1 knockdown efficiency in CAL27 and CAL33 cells using two independent shRNAs. e Cell viability was measured by a CCK-8 assay at 72 h posttransduction in control and SPI1-knockdown cells (n = 3 independent replicates). Quantification (f) and representative images (g) of colony formation assays showing the clonogenic potential of CAL27 and CAL33 cells after SPI1 knockdown (n = 3 independent replicates). h, i EdU incorporation assay for cell proliferation. h Representative immunofluorescence images showing EdU incorporation in CAL27 and CAL33 cells transduced with shNC or SPI1-targeting shRNAs (scale bar = 50 μm); i quantification of EdU-positive cells (n = 3 independent replicates). j Representative images from Transwell migration and invasion assays in CAL27 and CAL33 cells with or without SPI1 knockdown (scale bar = 200 μm). Quantification of migrated and invaded CAL27 (k) and CAL33 (l) cells (n = 3 independent replicates). m Western blot showing SPI1 overexpression in SAS and FaDu cells. n CCK-8 assay showing enhanced proliferation in SPI1-overexpressing SAS and FaDu cells (n = 3 independent replicates). Colony formation assay results showing increased clonogenic capacity following SPI1 overexpression (o), with corresponding representative colony images shown in (p) (n = 3 independent replicates). q Representative transwell migration and invasion assay images of SPI1-overexpressing SAS and FaDu cells (scale bar = 200 μm). Quantification of migrated and invaded SAS (r) and FaDu (s) cells following SPI1 overexpression (n = 3 independent replicates). In (d–l), the shNC group indicates cells transduced with a lentiviral vector carrying nontargeting shRNA. In (m–s), the NC group indicates cells transduced with an empty lentiviral vector as the control for SPI1 overexpression. The data are presented as the mean ± SD. Statistical significance was assessed using one-way ANOVA followed by Tukey’s post hoc test for multiple group comparisons or two-tailed unpaired Student’s t test for two-group comparisons. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001
SPI1 modulates HNSCC tumor growth and malignant phenotypes in vivo
To investigate the functional significance of SPI1 in vivo, we established subcutaneous xenograft models using CAL33 cells stably transduced with two independent SPI1-targeting shRNAs. Compared with shNC, SPI1 knockdown markedly suppressed tumor growth, as evidenced by reduced tumor volume over time and significantly decreased tumor weight at the experimental endpoint (Fig. 4a–c). Histological examination by H&E staining revealed a more dispersed and less compact tumor architecture in the SPI1-deficient xenografts (Fig. 4g). To specifically assess SPI1 expression in tumor epithelial cells, we performed multiplex immunofluorescence costaining for SPI1, EpCAM, and Ki-67. Quantitative analysis demonstrated a reduction in the proportion of SPI1⁺ cells among EpCAM⁺ tumor cells following SPI1 knockdown (Fig. 4d, g). Consistent with impaired tumor growth, multiplex immunofluorescence further revealed a significant decrease in the fraction of Ki-67⁺ cells within the EpCAM⁺ tumor cell compartment in the SPI1-knockdown xenografts (Fig. 4e, g), indicating attenuated proliferative activity. In parallel, TUNEL staining revealed a marked increase in the number of apoptotic cells in tumors lacking SPI1 expression (Fig. 4f, g), suggesting that tumor cell apoptosis increased upon depletion of SPI1. To validate these findings in an independent gain-of-function setting, we generated xenograft models using SAS and FaDu cells with stable SPI1 overexpression. In both models, compared with control tumors, tumors with increased SPI1 expression had significantly increased tumor volumes, elevated endpoint tumor weights, and accelerated tumor growth kinetics (Fig. 4h–m and Supplementary Fig. 18). Multiplex immunofluorescence staining revealed a robust increase in the proportion of SPI1⁺ cells within the EpCAM⁺ tumor cell compartment in SPI1-overexpressing SAS and FaDu xenografts (Fig. 4n–p). Concomitantly, SPI1 overexpression was associated with a significant increase in the fraction of Ki-67⁺ cells among EpCAM⁺ tumor cells (Fig. 4q–s), indicating enhanced proliferative activity in vivo. Taken together, these data demonstrate that SPI1 promotes HNSCC tumor growth and sustains malignant phenotypes in vivo by increasing tumor cell proliferation and survival.
a Representative images of xenograft tumors derived from CAL33 cells transduced with nontargeting shRNA (shNC) or two independent SPI1-targeting shRNAs (n = 6 biologically independent mice per group). b Growth curves of CAL33 xenografts with or without SPI1 knockdown (n = 6 biologically independent mice per group). c Tumor weights at the endpoint in CAL33 xenografts (n = 6 biologically independent mice per group). d Quantification of the percentage of SPI1⁺ cells among EpCAM⁺ tumor cells in CAL33 xenografts on the basis of immunofluorescence analysis (n = 6 independent tumors). e Quantification of the percentage of Ki-67⁺ cells among EpCAM⁺ tumor cells in CAL33 xenografts on the basis of immunofluorescence analysis (n = 6 independent tumors). f Quantification of apoptotic cell percentages in CAL33 xenografts by TUNEL staining (n = 6 independent tumors). g Representative histological and immunofluorescence images of CAL33 xenograft tumors. Hematoxylin and eosin (H&E) staining is shown in the top row. Immunofluorescence images show costaining of SPI1 (magenta) and EpCAM (yellow), Ki-67 (green) and EpCAM (yellow) with DAPI (blue), and TUNEL-positive apoptotic cells (green) with DAPI (blue). SPI1/EPCAM and Ki-67/EPCAM images were derived from the same multiplex-stained section and identical field, with channels displayed separately for visualization. Scale bar, 100 μm. h Representative images of xenograft tumors derived from SAS cells with SPI1 overexpression or control cells (n = 6 biologically independent mice per group). i Tumor weights at the endpoint in SAS xenografts (n = 6 biologically independent mice per group). j Growth curves of SAS xenografts with or without SPI1 overexpression (n = 6 biologically independent mice per group). k Representative images of xenograft tumors derived from FaDu cells with SPI1 overexpression or control cells (n = 6 biologically independent mice per group). l Tumor weights at the endpoint in the FaDu xenografts (n = 6 biologically independent mice per group). m Growth curves of FaDu xenografts with or without SPI1 overexpression (n = 6 biologically independent mice per group). Quantification of the percentage of SPI1⁺ cells among EpCAM⁺ tumor cells in SAS (n) and FaDu (o) xenografts on the basis of immunofluorescence analysis (n = 6 independent tumors). Representative immunofluorescence images showing the costaining of SPI1 (magenta) with EpCAM (yellow) and Ki-67 (green) with EpCAM (yellow) in SAS (p) and FaDu (q) xenografts; DAPI (blue). SPI1/EPCAM and Ki-67/EPCAM images were derived from the same multiplex-stained section and identical field, with channels displayed separately for visualization. Scale bar, 100 μm. Quantification of the percentage of Ki-67⁺ cells among EpCAM⁺ tumor cells in SAS (r) and FaDu (s) xenografts on the basis of immunofluorescence analysis (n = 6 independent tumors). In (a–g), the shNC group indicates cells transduced with a lentiviral vector carrying nontargeting shRNA. In (h–s), the NC group indicates cells transduced with an empty lentiviral vector as the control for SPI1 overexpression. In (b, j, m), the data are presented as the mean ± SEM. Statistical significance was determined using one-way ANOVA followed by Tukey’s post hoc test for multiple group comparisons or two-tailed unpaired Student’s t test for two-group comparisons. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; ns, not significant
RNA sequencing reveals the molecular pathways regulated by SPI1 in HNSCC
To elucidate the transcriptional programs regulated by SPI1 in HNSCC cells, we performed bulk RNA sequencing followed by TF activity inference and pathway enrichment analyses in both the SPI1 knockdown and overexpression models. In the loss-of-function setting, SPI1 knockdown in CAL33 cells led to a significant reduction in inferred SPI1 TF activity (Fig. 5a) and coordinated remodeling of the TF activity landscape (Fig. 5b). At the pathway level, gene set enrichment analysis (GSEA) revealed a broad attenuation of oncogenic and inflammation-associated transcriptional programs in SPI1-deficient cells, indicating global suppression of protumorigenic signaling upon loss of SPI1 expression (Supplementary Fig. 19). In the gain-of-function setting, enforced SPI1 expression in SAS cells resulted in a marked increase in SPI1 TF activity (Fig. 5c). Consistent with this regulatory activation, SPI1 overexpression was accompanied by coordinated upregulation of epithelial–mesenchymal transition (EMT) and invasion-associated genes, as reflected by increased expression of representative mesenchymal markers (Fig. 5d). Hallmark pathway analysis further demonstrated that SPI1 overexpression activated multiple oncogenic signaling programs related to inflammation, growth signaling, and cellular stress responses (Fig. 5e). Complementary functional enrichment analyses of SPI1-induced genes revealed pathways involved in cytoskeletal remodeling and cell motility, which is consistent with the enhanced invasive potential (Supplementary Fig. 20). Together, these results demonstrate that SPI1 acts as a central transcriptional regulator in HNSCC cells, suppressing oncogenic programs upon depletion while promoting EMT-associated transcriptional reprogramming and malignant signaling upon overexpression.
Transcriptional reprogramming and phenotypic alterations driven by SPI1 perturbation in HNSCC cells. a Violin plot showing inferred SPI1 TF activity in CAL33 cells transduced with shNC or shSPI1 #1. b Heatmap displaying the top differentially active TFs between shNC and shSPI1 #1 inferred by decoupleR. The red arrow highlights the specific downregulation of SPI1 activity. c Violin plot confirming significantly increased SPI1 TF activity in SAS cells upon SPI1 overexpression (OE) compared with that in the negative control (NC). d Heatmap showing the expression profiles of key EMT- and invasion-related markers (e.g., MMP9, VIM, and CDH2) in SAS-NC versus SAS-OE cells. The side bar indicates the log2-fold change (log2FC). e Gene set enrichment analysis (GSEA) of cancer hallmark pathways in SPI1-overexpressing cells. Red circles indicate upregulated pathways, while blue circles indicate downregulated pathways. Dot size represents statistical significance (−log10 FDR). *, P < 0.05
SPI1 promotes mRNA hypertranscription in HNSCC
To directly assess whether SPI1 modulates global transcriptional activity in HNSCC, we employed two complementary strategies: 5-ethynyl uridine (EU) incorporation to measure nascent RNA synthesis and quantification of total mRNA content normalized to cell number. The EU is a uridine analog that labels all newly transcribed RNA, providing a sensitive readout of global transcription.37 Furthermore, to specifically evaluate mRNA output independent of noncoding RNA contributions, we further performed poly(A)-selected mRNA purification followed by quantification adjusted for cell count. In CAL27 and CAL33 cells, stable knockdown of SPI1 led to a marked reduction in EU fluorescence intensity (Fig. 6a, b) and a significant decrease in per-cell mRNA abundance (Fig. 6c), indicating impaired transcriptional output. Conversely, SPI1 overexpression in SAS and FaDu cells significantly increased EU incorporation and mRNA levels per cell (Fig. 6d–f), establishing that SPI1 is necessary to drive bulk mRNA synthesis.
SPI1 promotes mRNA hypertranscription in HNSCC through a time-dependent mechanism. Representative images (a) and quantification (b) of EU incorporation in CAL27 and CAL33 cells after stable knockdown of SPI1 using two independent shRNAs (n = 3 independent replicates). Red: EU-labeled nascent RNA; blue: DAPI-stained nuclei. Scale bar, 100 μm. c Relative per-cell mRNA yield in CAL27 and CAL33 cells, determined after mRNA purification and normalized by cell number (n = 3 independent replicates). Representative EU-stained images (d) and quantification (e) of SAS and FaDu cells with or without SPI1 overexpression (n = 3 independent replicates). f Relative per-cell mRNA yield in SPI1-overexpressing SAS and FaDu cells, based on poly(A)-selected RNA quantification (n = 3 independent replicates). g Time-course RT‒qPCR analysis of SPI1 expression in Dox-inducible Tet-ON SAS and FaDu cells following doxycycline treatment (n = 3 independent replicates). The data were normalized to that of GAPDH and expressed relative to 0 h. h Immunoblot analysis of the SPI1 and c-MYC protein levels. Representative images (i) and quantification of the fluorescence intensity (j) of cells with incorporated EU at the indicated time points after Dox induction (n = 3 independent replicates). The data are presented as the mean ± SD. Scale bars, 50 μm. In (a–c), the shNC group indicates cells transduced with a lentiviral vector carrying nontargeting shRNA. In (d–f), the NC group indicates cells transduced with an empty lentiviral vector as the control for SPI1 overexpression. Statistical significance was determined using one-way ANOVA followed by Tukey’s post hoc test or an unpaired two-tailed Student’s t test. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001
To define the kinetics of this transcriptional amplification, we developed a doxycycline-inducible Tet-ON system in SAS and FaDu cells. Upon Dox exposure, the SPI1 mRNA and protein levels increased in a time-dependent manner, peaking at 24 h (Fig. 6g, h). EU fluorescence likewise increased progressively, with a detectable upregulation beginning at 3–6 h and peaking at 24 h (Fig. 6i, j), confirming the time-resolved activation of transcription. To rule out potential leakiness inherent to doxycycline-inducible systems, we further implemented a Shield-1-regulated destabilization domain (DD)-SPI1 system to achieve tighter posttranslational control of SPI1 protein abundance. Shield-1 treatment resulted in time-dependent stabilization of the DD-SPI1 protein and a concordant increase in EU incorporation in SAS and FaDu cells (Supplementary Fig. 21), independently corroborating that acute SPI1 induction is sufficient to increase globax bh transcriptional output. Notably, although MYC target gene sets were enriched upon SPI1 overexpression, MYC protein levels were not elevated under these conditions (Fig. 6h), suggesting the activation of MYC-associated transcriptional programs without direct MYC upregulation.
ChIP and RNA-seq analyses revealed that SPI1 widely activates the cell transcriptional program
Given that SPI1 is an important transcription factor in cellular development, we hypothesized that SPI1 may promote global mRNA output in HNSCC through its intrinsic transcriptional regulatory activity. To directly map SPI1 chromatin engagement, we performed ChIP-seq in SPI1-overexpressing SAS cells using two independent ChIP-seq replicates with matched input controls. Across two independent SPI1-OE ChIP-seq replicates, compared with their corresponding input controls, immunoprecipitated (IP) samples presented substantially higher FRiP scores, indicating robust and reproducible SPI1 chromatin enrichment and overall high ChIP-seq data quality (Supplementary Table 6). ChIP analysis of SAS-overexpressing cells revealed that SPI1 binding sites were highly enriched around transcription start sites (TSSs), with symmetric peak signals within ±1.5 kb flanking regions, indicating preferential promoter occupancy (Fig. 7a, b and Supplementary Fig. 22). SPI1 binding events were evenly distributed across all chromosomes without overt regional bias (Fig. 7c). Among these genes, SPI1 directly binds to the transcription sites of 7250 protein-coding genes, indicating its important role in inducing the transcription program in HNSCC cells (Supplementary Table 7). To define SPI1-dependent transcriptional programs, we conducted RNA sequencing of SAS cells with enforced SPI1 expression versus vector controls. Differential expression analysis revealed that SPI1 overexpression upregulated a large set of genes (Fig. 7d). GSEA demonstrated significant enrichment of hallmark oncogenic pathways, prominently hypoxia and glycolysis, in SPI1-overexpressing cells (Fig. 7e and Supplementary Table 8). Integration of the RNA-seq and ChIP-seq data revealed 2154 direct target genes, highlighting its role as a direct transcriptional activator (Fig. 7f–h). To summarize network-level organization, we constructed a protein‒protein interaction (PPI) network from the SPI1-regulated gene set and performed module detection with functional enrichment; these module-level results are provided in the Supplementary Information (Supplementary Fig. 23 and Supplementary Table 9). To validate the transcriptional activity of SPI1, we also performed a dual-luciferase reporter assay, which revealed time-dependent increases in the firefly/Renilla ratio upon doxycycline-induced SPI1 expression in SAS cells, confirming its activation via its cognate motif (Fig. 7i). Together, these results identify SPI1 as a broad transcriptional activator that drives pro-oncogenic programs in HNSCC.
Integrative transcriptomic and epigenomic analyses of SPI1 function in HNSCC. a Heatmaps and averaged ChIP-seq signal profiles showing SPI1 occupancy around transcription start sites (TSSs) in SPI1-overexpressing SAS cells, with SPI1 ChIP-seq Rep1 and its corresponding input control shown separately. b Aggregated metagene profiles illustrating normalized SPI1 ChIP-seq read density relative to the TSS across two independent SPI1-OE ChIP replicates and matched input samples. c Circos plot showing the genome-wide chromosomal distribution of SPI1 ChIP-seq binding peaks. d Volcano plot of DEGs identified by RNA-seq following SPI1 overexpression. DEGs were defined using an FDR < 0.05 and a |log2FC| > 1. e Representative gene set enrichment analysis (GSEA) plots showing hallmark pathways enriched in SPI1-overexpressing cells. f Venn diagram depicting the overlap between genes upregulated upon SPI1 overexpression and genes annotated with SPI1 ChIP-seq binding peaks. g Bubble plot showing Gene Ontology (GO) biological process enrichment of genes coregulated by SPI1 transcriptional activation and direct chromatin binding; the dot size indicates the gene count, and the color represents the adjusted P value. h De novo motif analysis of SPI1 ChIP-seq binding peaks. i Dual-luciferase reporter assay measuring SPI1-dependent transcriptional activity in a doxycycline-inducible system over time (n = 3 independent experiments; mean ± SD; one-way ANOVA with Tukey’s post hoc test)
Elevated SPI1 expression is correlated with poor prognosis in clinical HNSCC cohorts
To assess the clinical significance of SPI1, we performed IHC staining in two independent HNSCC patient cohorts from West China Hospital. In the laryngeal cancer cohort (n = 90; Supplementary Table 10), SPI1 protein levels were significantly higher in tumor tissues than in adjacent nontumor epithelium, as measured by the IRS (Fig. 8a). Kaplan‒Meier survival analysis revealed that high SPI1 expression was associated with poor overall survival (Fig. 8b). Notably, multivariate Cox regression analysis demonstrated that SPI1 remained a significant independent prognostic factor after adjusting for clinical stage, indicating that its prognostic impact was not confounded by tumor burden (Fig. 8c). Similar patterns were observed in the hypopharyngeal cancer cohort (n = 81; Supplementary Table 11 and Supplementary Fig. 24). SPI1 protein levels were again significantly elevated in tumor tissues (Fig. 8d), and higher expression correlated with poorer survival outcomes (Fig. 8e). Multivariate Cox regression further confirmed the independent prognostic value of the SPI1 in this cohort (Fig. 8f). To explore whether this expression pattern extends beyond laryngeal and hypopharyngeal cancers, we additionally assessed SPI1 expression in a small cohort of oropharyngeal cancer specimens, including HPV-positive cases. SPI1 immunoreactivity was significantly greater in tumor tissues than in peritumoral tissues and tended to be consistent with that observed in laryngeal and hypopharyngeal cancers (Supplementary Fig. 25). Together, these findings establish SPI1 as a clinically relevant biomarker of poor prognosis in HNSCC patients, with potential utility for risk stratification across subtypes.
Elevated SPI1 expression is associated with poor prognosis in laryngeal and hypopharyngeal cancer patients. a Validation of SPI1 expression in a local cohort of laryngeal cancer patients. b Kaplan‒Meier survival curve and c multivariate Cox forest plot for laryngeal cancer patients. d Validation of SPI1 expression in a local cohort of hypopharyngeal cancer patients. e Kaplan‒Meier survival curve and f multivariate Cox forest plot for hypopharyngeal cancer patients. For Kaplan‒Meier analyses (b, e), follow-up was truncated at 1000 days to ensure stable risk set sizes and avoid overinterpretation of sparse late follow-up
Discussion
In the present study, we sought to elucidate the functional role and regulatory mechanisms of mRNA hypertranscription in HNSCC. Our integrative analysis revealed that mRNA hypertranscription was strongly associated with aggressive tumor phenotypes and adverse clinical outcomes in HNSCC patients. We identified SPI1 as a key upstream regulator of this transcriptional state. Functional experiments confirmed that SPI1 not only drives malignant behaviors in vitro and in vivo but also promotes global mRNA output in a temporally dynamic manner that is not accompanied by detectable MYC upregulation under our experimental conditions. Moreover, integrative transcriptomic and ChIP-seq analyses demonstrated that SPI1 broadly engages transcriptional regulatory elements across the genome, orchestrating a transcriptional network that reinforces mRNA hypertranscription and tumor progression.
The first goal of our study was to determine whether mRNA hypertranscription could promote the malignant progression and poor prognosis of HNSCC. Zatzman et al. demonstrated that hypertranscription in tumor cells was associated with malignant phenotypes in nearly 31 types of tumors.15 This suggests that the overall upregulation of mRNA expression may be a breakthrough point for studying the malignant phenotypes of tumor cells and a determinant of malignant progression in tumor hypertranscription. However, in the past, the mRNA expression level of individual cells could be assessed only through single-cell transcriptome sequencing methods, limiting large-scale studies.15 Cao et al. developed a method to predict total mRNA expression within tumors on the basis of bulk sequencing results, enabling us to evaluate and analyze mRNA hypertranscription in HNSCC. On the basis of Cao’s method, we found that HNSCC mRNA hypertranscription is associated with the activation of tumor-related malignant pathways, both at the individual patient level and at our own single-cell level. Consistent with this state-based view, differential expression analysis revealed lincRNA enrichment in low-hypertranscription malignant epithelial cells, whereas in hypertranscriptional cells, lincRNAs preferentially upregulated ribosomal, cytoskeletal, and stress-response genes, which is consistent with a poised versus globally amplified transcriptional state.38 Building upon the findings of previous literature, our study provides evidence supporting the hypothesis that tumor hypertranscription can promote the malignant progression and poor prognosis of HNSCC.
We identified SPI1 as a potentially important transcription factor in the hypertranscription process of HNSCC through TF activity analysis. In recent years, an increasing number of studies have reported that SPI1 plays a key role in the development and progression of solid tumors. Gao et al. reported that SPI1 can promote the growth, migration, and invasion of non-small cell squamous carcinoma by transcriptionally activating the expression of SNHG6.27 Moreover, Yang et al. reported that downregulating SPI1 expression in lung cancer enhances the radiosensitivity of LUSC cells by inhibiting the transcription of TPX2 and RNF2, providing potential targets for radiotherapy in LUSC.39 In this study, we provide the first evidence suggesting that SPI1 may be a crucial transcription factor involved in the process of mRNA hypertranscription. Furthermore, using both in vivo and in vitro experiments, we demonstrate for the first time that SPI1 plays a significant role in the malignant progression of HNSCC.
From a translational perspective, although SPI1 has emerged as a functionally important regulator of hypertranscription and malignant progression in HNSCC, direct pharmacological targeting of SPI1 remains challenging. At present, no SPI1-directed agents have entered clinical application. While functional assays revealed that SPI1 is predominantly localized in the nucleus, its subcellular distribution may represent an additional regulatory layer influencing transcriptional output rather than a distinct cytoplasmic function. Nevertheless, accumulating evidence suggests that transcription factor-driven oncogenic programs are not inherently undruggable.40 Notably, proof-of-concept studies have demonstrated that small molecules can interfere with the DNA-binding activity or transcriptional function of SPI1 in preclinical settings, supporting the chemical tractability of this transcription factor in specific contexts.41,42 In parallel, indirect strategies that modulate transcriptional dependency, including the use of epigenetic modulators that disrupt transcription factor-associated chromatin states and transcriptional CDK inhibitors that attenuate global RNA output, have been shown to be effective.43 Together, these studies suggest that although direct SPI1 inhibition remains an unmet challenge, SPI1-driven transcriptional programs may be therapeutically targeted through convergent regulatory nodes, providing a rational framework for future translational exploration in HNSCC.
Although the responses to PD-1/PD-L1 or CTLA-4 blockade were not directly evaluated in this study, SPI1 overexpression was associated with the activation of immune-related transcriptional programs.44 Consistent with this notion, recent studies have reported an association between SPI1 expression and immunotherapy tolerance in gastric cancer, suggesting that SPI1-mediated immune regulation may extend across tumor types.28 Together, these observations suggest that SPI1 may influence tumor sensitivity to ICI therapy, suggesting that SPI1 could serve as a biomarker in immunotherapy-based strategies for HNSCC. Notably, the efficacy of immune checkpoint blockade in HNSCC is strongly influenced by HPV-defined tumor subtypes, underscoring the importance of considering how SPI1-driven immune programs may operate within distinct HPV-associated regulatory contexts.45 Given that SPI1 can engage cytokine-NF-κB/STAT3-associated networks and myeloid-linked transcriptional modules, its downstream consequences may be shaped by the regulatory context of HPV, including interferon signaling, antigen presentation programs, and macrophage polarization states.46 Therefore, SPI1-driven hypertranscription may intersect with HPV-related immune surveillance in a context-dependent manner. Dedicated HPV-stratified mechanistic studies will be needed to determine whether SPI1 contributes to immune evasion or therapeutic resistance differently across HPV-defined subtypes.
Despite these strengths, several limitations of the present study should be acknowledged. First, although we incorporated major clinicopathological factors, residual confounding factors (e.g., resection margin and perineural invasion) cannot be entirely excluded. The local clinical validation cohort was retrospectively assembled and limited to patients with laryngeal and hypopharyngeal cancers because of tissue availability, with some variation in clinical characteristics across subsites. While the generalizability of our findings was further supported by complementary analyses in the TCGA-HNSC cohort, additional validation in broader, multisubsite cohorts will be valuable. In addition, as SPI1 is also expressed in immune cells, SPI1-associated signals from bulk TCGA transcriptomic data may partially reflect contributions from the tumor immune microenvironment; in contrast, xenograft experiments selectively profiling human tumor–derived transcripts minimized such confounding. Finally, while our data support a broad role for SPI1 in regulating transcriptional output, the precise molecular mechanisms underlying its interaction with the core transcriptional machinery have not been directly investigated and warrant further investigation.
In summary, we identified SPI1 as an important regulator of mRNA hypertranscription in HNSCC. SPI1 exerts its effects through both widespread promoter engagement and the selective activation of oncogenic transcriptional programs, collectively contributing to increased global mRNA output. While the precise underlying mechanisms, including potential roles in RNA polymerase II recruitment or chromatin remodeling, remain to be elucidated, our findings support SPI1 as a broad transcriptional activator and highlight its potential relevance as a therapeutic target in hypertranscriptional HNSCC subtypes.
Materials and methods
Cell lines and cell culture
The human HNSCC cell lines FaDu, SAS, CAL33, and CAL27 were acquired from GuangZhou Jennio Biotech Co., Ltd. CAL27, CAL33, and SAS are derived from the tongue, whereas FaDu originates from the hypopharynx. Short tandem repeat (STR) analyses confirmed the identities of the cell lines used. All cells were cultured in high-glucose Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum (Thermo Fisher Scientific, Inc.). The cells were incubated in a humidity-controlled incubator with an atmosphere consisting of 95% air and 5% CO2 at a constant temperature of 37 °C. All of these cells were regularly tested for Mycoplasma every 12 months.
Animals
Five- to six-week-old female BALB/c nude mice were purchased from Gempharmatech Co., Ltd. The mice were housed in a specific pathogen-free (SPF) environment maintained at 22 °C with a 12-h light/dark cycle at West China Hospital. All regulatory and ethical requirements were followed. Specific ethical approval was obtained for the study from the Animal Ethics Committee, West China Hospital of Sichuan University (No. 20240923008).
In vivo experiments
Five- to six-week-old female BALB/c nude mice were randomly assigned to the following groups: CAL33/shSPI1#1, CAL33/shSPI1#2, CAL33/shNC, SAS/SPI1 OE, SAS/SPI1 NC, FaDu/SPI1 OE, and FaDu/SPI1 NC. For each group, 3 × 106 corresponding cells were subcutaneously injected into the right flank of the mice (n = 6 per group). The weight and tumor volume were measured every 4 days. The tumor volumes were estimated using the following formula: V = (length × width × width)/2. After the mice were sacrificed, the tumors were carefully excised, weighed, and photographed for subsequent analysis. The mice were sacrificed when the tumor volume of any individual mouse reached the predefined limit (800 mm³), and the experiment was terminated accordingly to ensure animal welfare. The tumor tissues were harvested, paraffin-embedded, sectioned for histological analyses, and used for RNA sequencing.
Real-time quantitative PCR (RT‒qPCR)
An Orizol Chloroform-Free RNA Extraction Kit (Oriscience) was used to extract total RNA from the cell lines following the manufacturer’s protocol. A NanoDrop 2000 spectrophotometer was used to assess RNA purity and concentration. Reverse transcription of 1 μg of total RNA into cDNA was performed using the PrimeScript RT Reagent Kit (TaKaRa) according to the manufacturer’s instructions. Quantitative real-time PCR (qPCR) was then performed using SYBR Green Master Mix, with initial denaturation at 95 °C for 600 s, followed by 42 cycles of 95 °C for 5 s and 58 °C for 30 s. GAPDH served as the reference gene for normalization in in vitro experiments. The primers were produced by Beijing Tsingke Biotech Co., Ltd. The primer sequences are available in Supplementary Table 12.
Western blot
Total cell lysates were extracted from cells using RIPA lysis buffer (Beyotime) supplemented with a phosphatase inhibitor cocktail. Protein concentrations were determined using a BCA protein quantitation kit (Beyotime). Protein lysates were separated by 10% SDS‒PAGE (Shanghai Epizyme Biomedical Technology Co., Ltd) and transferred to PVDF membranes (Millipore, Massachusetts, USA). The membranes were then blocked with 5% nonfat milk powder and incubated overnight at 4 °C with primary antibodies against SPI1 (1:500; Cell Signaling Technology, cat. 2258), RB1 (1:1000; Proteintech, 10048-2-Ig), phospho-RB1 (Ser780) (1:1,000; Proteintech, 84692-1-RR), Akt (pan) (1:1000; Cell Signaling Technology, #4691), phospho-Akt (Ser473) (1:1000; Cell Signaling Technology, #4060), and β-actin (1:10,000; Proteintech, 60008-1-Ig). The membranes were subsequently incubated with either anti-mouse (1:10,000; ZSGB-BIO, ZB-2305) or anti-rabbit (1:10,000; ZSGB-BIO, ZB-2301) horseradish peroxidase-labeled secondary antibodies. The protein bands were detected using an ECL system (Thermo Fisher Scientific) and visualized with an iBright CL1000 instrument.
Cell transfection
For SPI1 knockdown, small-hairpin RNAs targeting human SPI1 mRNA were cloned and inserted into the LV2N-U6/Puro vector (GenePharma, Shanghai, China). The shRNA sequences used were as follows: shSPI1#1 (5′-ACGCCAAACGCACGAGTATTA-3′), shSPI1#2 (5′-AGCAGATGCACGTCCTCGATA-3′), and a nontargeting control, shNC (5′-TTCTCCGAACGTGTCACGT-3′). To establish SPI1-overexpressing cell lines, the full-length coding sequence of human SPI1 (NM_003120) was cloned and inserted into the pcDNA3.1-puromycin lentiviral vector (GenePharma). Cells transduced with empty vector served as controls. Additionally, we obtained the TetOn inducible expression system for SPI1 from GeneCopoeia, Inc., which utilizes the PLVX-TetOne-Puro vector. This system allows for the controlled expression of SPI1 in response to doxycycline treatment. Lentiviral particles containing the SPI1 shRNA, overexpression, or inducible constructs were generated and transduced into HNSCC cell lines following the manufacturer’s protocol. Stable cell lines were established through puromycin selection using optimized concentrations for each cell line (Solarbio, Beijing, China). The efficiency of lentiviral transduction and the resulting changes in SPI1 expression were validated by Western blot analysis and quantitative reverse transcription–PCR (qRT‒PCR). To construct an orthogonal inducible system with tighter posttranslational control, an FKBP12 destabilization domain (DD, L106P)-SPI1 expression plasmid (pLVX-IRES-Puro; EcoRI/XhoI) was constructed by Tsingke Biotechnology, and the insert/flanking sequences were sequence verified. This DD-SPI1 plasmid was introduced into SAS and FaDu cells by transient transfection, followed by Shield-1 treatment for the indicated durations to stabilize DD-SPI1.
Estimating nascent RNA generation and mRNA output
Single-cell nascent RNA synthesis was quantified using the BeyoClick™ EU Cell Proliferation Kit with Alexa Fluor 555 (Beyotime) and visualized by fluorescence microscopy according to the manufacturer’s protocol. To assess mRNA output, the cells were harvested, and the number of viable cells was determined using an automated cell counter in combination with trypan blue exclusion, thereby excluding nonviable cells from downstream normalization. Total RNA was extracted, and mRNA was purified using the BeyoMag™ mRNA Purification Kit with Magnetic Beads (Beyotime). The mRNA output was quantified using a Qubit® RNA BR Assay Kit (Thermo Fisher Scientific) and normalized to the number of viable cells.
Clinical sample collection and preparation
Fresh surgical samples from 12 HNSCC patients were collected. The tissues were rinsed twice with 1× PBS, after which the necrotic part was removed. Afterward, the samples were divided into small pieces of approximately 2 mm. The gentleMACS C Tube containing a Human Tumor Dissociation Kit (Miltenyi Biotec, #130-095-929) was used to dissociate the cells according to the manufacturer’s protocols, after which the cells were filtered through 40 µm cell strainers. A cell viability greater than 80% and a clump rate less than 20% were considered acceptable. Then, the scRNA-seq libraries were prepared and sequenced using the MobiNova-100 platform (MobiDrop, Zhejiang, China) following the manufacturer’s instructions. A total of 12 samples were collected, and their primary tumor locations are shown in Supplementary Table 1. Specific ethical approval was obtained for the study from the Ethics Committee on Biomedical Research, West China Hospital of Sichuan University (Nos. 20211673 and 2025341).
Statistical analysis
Data are presented as the mean ± standard deviation (SD) unless otherwise specified. Statistical significance was determined using one-way ANOVA followed by Tukey’s post hoc test or an unpaired two-tailed Student’s t test. All experiments with statistical analysis were performed independently at least three times. A two-sided p value < 0.05 was considered to indicate statistical significance.
Data availability
Available raw sequencing data generated in this study have been deposited in the Genome Sequence Archive at the BIG Data Center, Beijing Institute of Genomics (BIG), Chinese Academy of Sciences, under accession number CRA018102 (https://ngdc.cncb.ac.cn/gsa/search?searchTerm=CRA018102). Transcriptomic and clinical data from The Cancer Genome Atlas (TCGA) Head and Neck Squamous Cell Carcinoma (HNSC) cohort were obtained from the Genomic Data Commons (GDC) Data Portal (https://portal.gdc.cancer.gov/). The single-cell RNA-sequencing dataset GSE234933 can be acquired from the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/) under accession number GSE234933. Gene expression data for head and neck squamous cell carcinoma cell lines were obtained from the Cancer Cell Line Encyclopedia (CCLE), which is available through the DepMap portal (https://depmap.org/portal/). data involving human genetic resources from Chinese participants are available from the corresponding author upon reasonable request, subject to the approval of a material transfer agreement (MTA) and in compliance with the Regulations of the People’s Republic of China on the Management of Human Genetic Resources. All other data supporting the findings of this study are available within the article and its Supplementary Information files.
References
Chow, L. Q. M. Head and neck cancer. N. Engl. J. Med. 382, 60–72 (2020).
Mody, M. D., Rocco, J. W., Yom, S. S., Haddad, R. I. & Saba, N. F. Head and neck cancer. Lancet 398, 2289–2299 (2021).
Kaidar-Person, O., Gil, Z. & Billan, S. Precision medicine in head and neck cancer. Drug. Resist. Update 40, 13–16 (2018).
Meng, Y., Deng, J., Deng, W. & Sun, Z. Intra-tumoral bacteria in head and neck cancer: Holistic integrative insight. Cancer. Biol. Med. 22, 113–136 (2025).
Dong, Y., Tu, R., Liu, H. & Qing, G. Regulation of cancer cell metabolism: oncogenic MYC in the driver’s seat. Sig. Transduct. Target Ther. 5, 124 (2020).
Wilkie, M. D. et al. TP53 mutations in head and neck cancer cells determine the Warburg phenotypic switch creating metabolic vulnerabilities and therapeutic opportunities for stratified therapies. Cancer Lett. 478, 107 (2020).
Swanton, C. et al. Embracing cancer complexity: hallmarks of systemic disease. Cell 187, 1589–1616 (2024).
Lin, C. Y. et al. Transcriptional amplification in tumor cells with elevated c-Myc. Cell 151, 56–67 (2012).
Percharde, M., Bulut-Karslioglu, A. & Ramalho-Santos, M. Hypertranscription in development, stem cells, and regeneration. Dev. Cell 40, 9–21 (2017).
Bowry, A., Kelly, R. D. W. & Petermann, E. Hypertranscription and replication stress in cancer. Trends Cancer 7, 863–877 (2021).
Udvardi, M. K., Czechowski, T. & Scheible, W.-R. Eleven golden rules of quantitative RT-PCR. Plant Cell 20, 1736–1737 (2008).
Chung, M. et al. Best practices on the differential expression analysis of multi-species RNA-seq. Genome Biol. 22, 121 (2021).
Heumos, L. et al. Best practices for single-cell analysis across modalities. Nat. Rev. Genet. 24, 550–572 (2023).
Kim, Y.-K. et al. Absolute scaling of single-cell transcriptomes identifies pervasive hypertranscription in adult stem and progenitor cells. Cell Rep. 42, 111978 (2023).
Zatzman, M. et al. Widespread hypertranscription in aggressive human cancers. Sci. Adv. 8, eabn0238 (2022).
Cao, S. et al. Estimation of tumor cell total mRNA expression in 15 cancer types predicts disease progression. Nat. Biotechnol. 40, 1624–1633 (2022).
Kovalski, J. R., Kuzuoglu-Ozturk, D. & Ruggero, D. Protein synthesis control in cancer: Selectivity and therapeutic targeting. EMBO J. 41, e109823 (2022).
Shiah, J. V., Johnson, D. E. & Grandis, J. R. Transcription factors and cancer: approaches to targeting. Cancer J. 29, 38–46 (2023).
Baluapuri, A. et al. MYC recruits SPT5 to RNA polymerase II to promote processive transcription elongation. Mol. Cell 74, 674–687.e11 (2019).
Oikawa, T. et al. The role of Ets family transcription factor PU.1 in hematopoietic cell differentiation, proliferation and apoptosis. Cell Death Differ. 6, 599–608 (1999).
Le Coz, C. et al. Constrained chromatin accessibility in PU.1-mutated agammaglobulinemia patients. J. Exp. Med. 218, e20201750 (2021).
He, Y. et al. Dasatinib-therapy induced sustained remission in a child with refractory TCF7-SPI1 T-cell acute lymphoblastic leukemia. Pediatr. Blood Cancer 69, e29724 (2022).
Verbiest, T., Bouffler, S., Nutt, S. L. & Badie, C. PU.1 downregulation in murine radiation-induced acute myeloid leukaemia (AML): from molecular mechanism to human AML. Carcinogenesis 36, 413–419 (2015).
Korczmar, E. A. et al. Transcriptional regulation of the lineage-determining gene PU.1 in normal and malignant hematopoiesis: current understanding and therapeutic perspective. Front. Biosci. 16, 10 (2024).
Curtiss, B. M. et al. PU.1 and MYC transcriptional network defines synergistic drug responses to KIT and LSD1 inhibition in acute myeloid leukemia. Leukemia 36, 1781–1793 (2022).
Wang, J. et al. Therapeutic targeting of SPIB/SPI1-facilitated interplay of cancer cells and neutrophils inhibits aerobic glycolysis and cancer progression. Clin. Transl. Med. 11, e588 (2021).
Gao, N. & Ye, B. SPI1-induced upregulation of lncRNA SNHG6 promotes non-small cell lung cancer via miR-485-3p/VPS45 axis. Biomed. Pharmacother. 129, 110239 (2020).
Huang, J., Chen, W., Jie, Z. & Jiang, M. Comprehensive analysis of immune implications and prognostic value of SPI1 in gastric cancer. Front. Oncol. 12, 820568 (2022).
Filippakopoulos, P. et al. Selective inhibition of BET bromodomains. Nature 468, 1067–1073 (2010).
Shorstova, T., Foulkes, W. D. & Witcher, M. Achieving clinical success with BET inhibitors as anti-cancer agents. Br. J. Cancer 124, 1478–1490 (2021).
Zhu, X. et al. Bromodomain and extraterminal protein inhibitor JQ1 suppresses thyroid tumor growth in a mouse model. Clin. Cancer Res. 23, 430–440 (2017).
Gorthi, A. et al. EWS-FLI1 increases transcription to cause R-loops and block BRCA1 repair in ewing sarcoma. Nature 555, 387–391 (2018).
Bulut-Karslioglu, A. et al. Chd1 protects genome integrity at promoters to sustain hypertranscription in embryonic stem cells. Nat. Commun. 12, 4859 (2021).
Guzman-Ayala, M. et al. Chd1 is essential for the high transcriptional output and rapid growth of the mouse epiblast. Development 142, 118–127 (2015).
Gulati, G. S. et al. Single-cell transcriptional diversity is a hallmark of developmental potential. Science 367, 405–411 (2020).
Müller-Dott, S. et al. Expanding the coverage of regulons from high-confidence prior knowledge for accurate estimation of transcription factor activities. Nucleic Acids Res. 51, 10934–10949 (2023).
Kotsantis, P. et al. Increased global transcription activity as a mechanism of replication stress in cancer. Nat. Commun. 7, 13087 (2016).
Statello, L., Guo, C.-J., Chen, L.-L. & Huarte, M. Gene regulation by long non-coding RNAs and its biological functions. Nat. Rev. Mol. Cell Biol. 22, 96–118 (2021).
Yang, J. et al. SPI1 mediates transcriptional activation of TPX2 and RNF2 to regulate the radiosensitivity of lung squamous cell carcinoma. Arch. Biochem. Biophys. 730, 109425 (2022).
Tao, Z. & Wu, X. Targeting transcription factors in cancer: from ‘undruggable’ to ‘druggable’. Methods Mol. Biol. 2594, 107–131 (2023).
Antony-Debré, I. et al. Pharmacological inhibition of the transcription factor PU.1 in leukemia. J. Clin. Investig. 127, 4297–4313 (2017).
Wang, X. et al. Small-molecule inhibitors of transcription factor PU.1 for the treatment of acute T cell lymphoblastic leukemia and organ fibrosis. CCS Chem. 8, 319–340 (2025).
Marei, H. E. et al. Targeting CDKs in cancer therapy: advances in PROTACs and molecular glues. npj Precis. Oncol. 9, 204 (2025).
Tufail, M., Jiang, C.-H. & Li, N. Immune evasion in cancer: mechanisms and cutting-edge therapeutic approaches. Signal Transduct. Target. Ther. 10, 227 (2025).
Sabatini, M. E. & Chiocca, S. Human papillomavirus as a driver of head and neck cancers. Br. J. Cancer 122, 306–314 (2020).
Gorvel, L. & Olive, D. Tumor associated macrophage in HPV+ tumors: between immunosuppression and inflammation. Semin. Immunol. 65, 101671 (2023).
Acknowledgements
This work was supported by the Noncommunicable Chronic Diseases-National Science and Technology Major Project (2023ZD0503000, 2023ZD0503004), the Regional Innovation and Development Joint Fund Key Project of the National Natural Science Foundation of China (U24A20735), the National Natural Sciences Foundation of China (82473434), the Fundamental and Interdisciplinary Disciplines Breakthrough Plan of the Ministry of Education of China, the Institutional Joint Innovation Fund from Sichuan University and Nuclear Power Institute of China (1-KJ-FWHT-WU-20240150), the Sichuan Science and Technology Program (2025YFHZ0087), the Science and Technology Project of Sichuan Provincial Health Commission (Clinical Research Special Project JH2023082), the International Science and Technology Cooperation Program of Chengdu Science and Technology Bureau (2024-YF06-00011-HZ and 2022-GH03-00004-HZ), the Health Research Project of Chengdu Eastern New Area Management Committee (202304), 1.3.5 project for disciplines of excellence from West China Hospital of Sichuan University (ZYYC23006, ZYGD23023), and the Clinical Research Incubation Project of West China Hospital (23HX001).The funders played no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Author information
Authors and Affiliations
Contributions
Conceptualization: P.X.C. and D.L.; Methodology: L.Z.R., L.H.L., and Q.Z.J.; Investigation: L.Z.R., L.H.L., and Q.Z.J.; Visualization: H.L., C.N., L.Q.H., and Z.D.; Supervision: P.X.C., D.L.; Writing—original draft: L.Z.R., L.H.L., Z.L.L, and Q.Z.J. All the authors have read and approved the article.
Corresponding authors
Ethics declarations
Competing interests
The authors declare no competing interests.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
About this article
Cite this article
Liu, Z., Qin, Z., Li, H. et al. Spi-1 proto-oncogene regulates mRNA hypertranscription and malignant progression in head and neck cancer. Sig Transduct Target Ther 11, 102 (2026). https://doi.org/10.1038/s41392-026-02669-6
Received:
Revised:
Accepted:
Published:
Version of record:
DOI: https://doi.org/10.1038/s41392-026-02669-6
