Identification of MUC1-C as a Target for Suppressing Progression of Head and Neck Squamous Cell Carcinomas

Abstract The MUC1-C protein is aberrantly expressed in adenocarcinomas of epithelial barrier tissues and contributes to their progression. Less is known about involvement of MUC1-C in the pathogenesis of squamous cell carcinomas (SCC). Here, we report that the MUC1 gene is upregulated in advanced head and neck SCCs (HNSCC). Studies of HNSCC cell lines demonstrate that the MUC1-C subunit regulates expression of (i) RIG-I and MDA5 pattern recognition receptors, (ii) STAT1 and IFN regulatory factors, and (iii) downstream IFN-stimulated genes. MUC1-C integrates chronic activation of the STAT1 inflammatory pathway with induction of the ∆Np63 and SOX2 genes that are aberrantly expressed in HNSCCs. In extending those dependencies, we demonstrate that MUC1-C is necessary for NOTCH3 expression, self-renewal capacity, and tumorigenicity. The findings that MUC1 associates with ∆Np63, SOX2 and NOTCH3 expression by single-cell RNA sequencing analysis further indicate that MUC1-C drives the HNSCC stem cell state and is a target for suppressing HNSCC progression. Significance: This work reports a previously unrecognized role for MUC1-C in driving STAT1-mediated chronic inflammation with the progression of HNSCC and identifies MUC1-C as a druggable target for advanced HNSCC treatment.


Introduction
Patients with locally advanced head and neck squamous cell carcinomas (HN-SCC) have historically had a poor prognosis with less than 50% surviving for 5 years despite curative-intent treatment with surgery, radiotherapy, and chemotherapy (1)(2)(3).Recurrent/metastatic HNSCCs were treated with cisplatin, 5-fluorouracil, and cetuximab, which achieved a response rate of 36% and a progression-free survival of 5.6 months (4).These poor outcomes were substantially improved with development of (i) the immune checkpoint inhibitors (ICI) pembrolizumab and nivolumab as single agents for the secondline treatment of platinum-refractory recurrent/metastatic disease and (ii) pembrolizumab alone or in combination with chemotherapy for first-line treatment (5)(6)(7).Those advances for treating HNSCC with immunotherapy fueled trials to evaluate immunochemoradiotherapy and other combined modalities (3).Nonetheless, intrinsic and acquired resistance has remained a challenge for HNSCC immunotherapy as patients invariably succumb to metastatic disease (3).The progression of HNSCC refractory to chemotherapy and ICIs has emphasized a need for the identification of druggable targets to circumvent DNA damage resistance and immune evasion.In this regard, other than cetuximab, there are no approved targeted agents for HNSCC treatment.Potential targets downstream to receptor tyrosine kinases include the transcription factors (TF) (i) Np63 and SOX2 that are dysregulated in HNSCC, and (ii) NOTCH3, which has been linked to stemness and HNSCC progression (8,9).HNSCCs are associated with chronic exposures to tobacco products and alcohol (8).In addition, human papillomavirus (HPV) infections are linked to an increased incidence of oropharyngeal HNSCCs, in further support of the importance of chronic inflammation in driving HNSCC progression (8)(9)(10)(11)(12).Patients with HPV-positive HNSCCs have a more favorable 5-year survival rate than those with HPV-negative HNSCCs and represent a distinct entity in terms of molecular profiles (13)(14)(15).
The mucin  (MUC) gene encodes a noncovalent heterodimeric complex consisting of MUC1-N and MUC1-C subunits that evolved in mammals to protect barrier tissues from biotic and abiotic insults (16,17).As an adverse outcome of this protective function, persistent activation of the transmembrane MUC1-C subunit in settings of chronic inflammation promotes oncogenesis (17)(18)(19).
In this way, MUC1-C drives lineage plasticity, the epithelial-mesenchymal transition (EMT) and epigenetic reprogramming in the progression of adenocarcinomas derived from barrier glandular epithelia (17)(18)(19)(20).Conversely, little is known about the involvement of MUC1-C in SCCs that are derived from stratified squamous epithelial cells (8).Expression of MUC1-N and the MUC1 cytoplasmic tail has been detected in HNSCCs (21)(22)(23).Other work has reported that MUC1-N promotes migration and invasion of oral SCC cells (24); however, the mechanisms underlying potential roles for MUC1-C have not been defined in squamous cell biology and HNSCC.In the progression of adenocarcinomas, MUC1-C regulates effectors of chronic inflammation, including (i) pattern recognition receptors (PRR) that induce IFNa/b production, (ii) STAT1 with activation of the type I/II IFN pathways, and (iii) IFN-stimulated genes (ISG) that contribute to DNA damage resistance and immune evasion (18,20,(25)(26)(27).MUC1-C-induced regulation of these inflammatory genes is associated with changes in (i) chromatin accessibility conferred by the SWI/SNF BAF and PBAF chromatin remodeling complexes, and (ii) H3K27ac and H3K4me1/3 levels (18,20,(25)(26)(27)(28)(29).These effects of MUC1-C, which in principle are reversible with restitution of homeostasis, become established in settings of chronic inflammation with progression to the cancer stem cell (CSC) state (17)(18)(19)(20).
The barrier role of the head and neck squamous epithelium is disrupted by stress, as for example that induced by exposure to cigarette smoke and alcohol (8).How that loss of homeostasis from chronic inflammation promotes progression to HNSCC remains largely unknown.The current studies demonstrate that MUC1-C drives STAT1-dependent chronic inflammation in HPV-negative HNSCC cells and thereby induction of the Np63 (9, 30-32) and SOX2 (10,33) TFs that have been linked to HNSCC pathogenesis.We also demonstrate that (i) MUC1-C signaling regulates the NOTCH3 pathway that contributes to the HNSCC CSC state (34,35), and (ii) HNSCC cells are dependent on MUC1-C for self-renewal and tumorigenicity.In concert with these results, we report that MUC1 associates with Np63, SOX2, and NOTCH3 expression in HNSCC tumor cells as determined by single-cell RNA sequencing (scRNA-seq) analysis.These findings indicate that MUC1-C promotes HNSCC progression and is a potential target for advanced HNSCC treatment.

Cell Culture
HPV-negative CAL27 cells (ATCC, RRID:CVCL_1107) were cultured in DMEM (Thermo Fisher Scientific) supplemented with 10% FBS (GEMINI Bio-Products).HPV-negative HSC3 and FaDu cells (ATCC, RRID:CVCL_1288 and 1218) were cultured in minimum essential medium (Corning) supplemented with 10% FBS.Authentication of the cells was performed every 3-4 months by short tandem repeat analysis.Cells were monitored for Mycoplasma contamination every 3-4 months using the MycoAlert Mycoplasma Detection Kit (Lonza).Cells were maintained for 3 months when performing experiments.

Colony Formation Assays
Cells were seeded in 24-well plates for 24 hours and then treated with (i) 0.1% DMSO or 500 ng/mL DOX, and (ii) PBS or GO-203.After 7-14 days, cells were stained with 0.5% crystal violet (LabChem) in 25% methanol.Growth was quantified at 590 nm using a spectrophotometer and normalized to DMSO treatment.

qRT-PCR
Total cellular RNA was isolated using TRIzol reagent (Thermo Fisher Scientific).cDNAs were synthesized using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems) as described previously (36).The cDNA samples were amplified using the Power SYBR Green PCR Master Mix (Applied Biosystems) and the CFX96 Real-Time PCR System (Bio-Rad) as described previously (36).Primers used for qRT-PCR are listed in Supplementary Table S1.

Chromatin Immunoprecipitation
Chromatin immunoprecipitation (ChIP) was performed on cells cross-linked with 1% formaldehyde for 10 minutes at 37°C, quenched with 2 mol/L glycine, washed with PBS, and sonicated in a Covaris 220 sonicator to generate 300-600 bp DNA fragments.Immunoprecipitation was performed using a control IgG (3900S, CST, RRID:AB_1550038) and antibodies against MUC1-C (16564S, CST, RRID:AB_2798765) and STAT1 (9172S, CST).Quantitation was performed on immunoprecipitated DNA using SYBR-green and the CFX384 real-time PCR machine (Bio-Rad).Precipitated DNAs were detected by PCR using primers listed in Supplementary Table S2.Data are reported as percentage of input DNA for each sample.

Chromatin-bound Protein Extraction
Chromatin Extraction Kit (Abcam) was used to isolate chromatin-bound proteins according to the manufacturer's instructions.

RNA-seq Analysis
Total RNA from cells cultured in triplicates was isolated using TRIzol reagent (Invitrogen) as described previously (36).TruSeq Stranded mRNA (Illumina) was used for library preparation.Raw sequencing reads were aligned to the human genome (GRCh38.74)using STAR.Raw feature counts were normalized and differential expression analysis using DESeq2 as described previously (36).Differential expression rank order for subsequent gene set enrichment analysis (GSEA) was performed using the fgsea (v1.8.0) package in R. Hallmark Gene Sets were queried through the Molecular Signatures Database (MSigDB).

Mouse Tumor Model Studies
Six-week-old nude mice (Jackson Laboratory) were injected subcutaneously into the flank with 1 × 10 7 CAL27 cells in 100 mL of a 1:1 solution of medium and Matrigel (BD Biosciences).When the mean tumor volume reached 100-150 mm 3 , the mice were pair-matched into groups of 5 mice each.Mice were treated intraperitoneally each day with PBS or GO-203 at a dose of 12 μg/g body weight.Tumor measurements and body weights were recorded twice per week.
Tumor lysates were prepared using the T-PER tissue protein extraction reagent (#78510, Thermo Fisher Scientific).The resource equation method was used for determining the minimum number of mice to achieve significance (37).These studies were conducted in accordance with the ethical regulations required for approval by the Dana-Farber Cancer Institute Animal Care and Use Committee under protocol #03-029.

Analysis of Publicly Available Bulk RNA-seq HNSCC Data
HNSCC RNA-seq and clinical annotation dataset file of GSE136037 was downloaded from the Gene Expression Omnibus (GEO) database.GSE136037 included 49 primary lesions and 23 metastatic lesions.The downloaded data GSE136037 was used in read count format.

Analysis of Publicly Available scRNA-seq HNSCC Data
Data for publicly available scRNA-seq dataset of HNSCC samples (GSE181919) were obtained from GEO. GSE181919 included 20 primary tumor samples ( 13HPV-negative and seven HPV-positive).The expression level of Transcript per million (TPM) was obtained with tumor cells identified using previously determined criteria (38).Uniform manifold approximation and projection (UMAP) representations of remaining tumor cells were produced from total expression profiles, with expression (TPM) compared between given factors.Generated census counts were obtained from GEO. Low-quality cells previously determined were eliminated prior to downstream analysis.Remaining cells were analyzed by Seurat.Normalization and variance stabilization were conducted using regularized negative binomial regression(sctransform).

Statistical Analysis
Each experiment was performed at least three times.Unpaired two-tailed Student t tests were used to assess differences between the mean ± SD of two groups.GraphPad Prism9 was used for all statistical analyses.P values were considered significant at *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001; ****, P ≤ 0.0001 with confidence interval = 95%.

Data Availability
Raw and processed RNA-seq data generated in this study are available from the NCBI GEO under accession GSE252287.

Dependence of HNSCC Cells on MUC1-C for Clonogenicity
Analysis of the GSE136037 dataset demonstrated that MUC gene expression is significantly increased in metastatic versus primary HNSCC tumors (Fig. 1A; Supplementary Fig. S1A).To investigate potential involvement of MUC1-C in HNSCC, we established HPV-negative CAL27 HNSCC cells expressing a CshRNA or MUC1shRNA vector.MUC1-C mRNA levels were decreased in CAL27/MUC1shRNA, and not CAL27/CshRNA, cells (Fig. 1B, left).Moreover, the MUC1-C protein, which is expressed as glycosylated 25 kDa and unglycosylated 17 kDa forms, was decreased in CAL27/MUC1shRNA cells (Fig. 1B, right).We found that silencing MUC1-C in CAL27/MUC1shRNA cells suppresses their capacity for clonogenic survival (Fig. 1C), in support of the potential importance of this oncoprotein in HNSCC progression.MUC1-C consists of a 58 aa extracellular, 28 aa transmembrane, and 72 aa cytoplasmic domain (18).In addressing how MUC1-C promotes CAL27 cell clonogenicity, the MUC1-C cytoplasmic domain (MUC1-CD) is an intrinsically disordered scaffold with nodes for integrating diverse signaling pathways (Fig. 1D; ref. 18).MUC1-CD is phosphorylated by EGFR, FGFR3, and MET (18).In addition, the MUC1-CD CQC motif, which is targeted by the GO-203 inhibitor, binds directly to JAK1 and the SAGNGGSSLS region associates with STAT1 in facilitating their interactions (Fig. 1D; refs.18,39).To further assess the effects of MUC1-C and MUC1-CD, we used an inducible tet-MUC1shRNA targeting a different region, which in response to DOX treatment downregulated MUC1-C expression (Fig. 1E).MUC1-C silencing was rescued by expressing a DOX-inducible Flag-MUC1-CD (tet-Flag-MUC1-CD; Fig. 1E).In this way, we found that MUC1-CD reverses the suppressive effects of silencing MUC1-C on colony formation (Fig. 1F).As an additional control, treatment of CAL27/tet-CshRNA cells with DOX had little if any effect on MUC1-C expression and
In addition, ISG15 links the DNA damage response to regulation of innate immunity (50).Along these lines, we found that MUC1-C regulates expression of guanylate-binding protein 1 (GBP1), which is activated by chronic inflammation and confers treatment resistance (refs.51, 52; Fig. 3G and H; Supplementary Fig. S3D).Silencing MUC1-C also decreased expression of (i) IDO1, which suppresses tryptophan (Trp) levels in the tumor microenvironment necessary for T-cell function (53), and (ii) tryptophanyl-tRNA synthetase (WARS, WRS) that protects cancer cells from Trp depletion (refs.54, 55; Fig. 3G and H).Similar results were obtained when targeting MUC1-C in HSC3 cells which, unlike CAL27 cells, have undetectable levels of IDO1 expression (Supplementary Fig. S3D).These findings supported a role for MUC1-C in regulating the type I and II IFN pathways in HNSCC cells and integrating expression of ISGs that promote DNA damage resistance and immune evasion associated with the CSC state (18,27,(56)(57)(58).

MUC1-C and STAT1 Regulate Np63 Expression
The Np63 TF plays a role in (i) homeostasis of the epidermis, (ii) SCC pathogenesis, and (iii) self-renewal of CSCs (9,(30)(31)(32).To our knowledge, there is no known association between MUC1-C and Np63.We found here that targeting MUC1-C decreases Np gene transcription and mRNA levels in CAL27 and HSC3 cells (Fig. 4A; Supplementary Fig. S4A).We also found that silencing STAT1 suppresses Np63 expression (Fig. 4B; Supplementary Fig. S4B), supporting potential involvement of the MUC1-C/STAT1 autoinductive pathway (39).Np is transcribed from a promoter in the third intron of the TP gene (31), which was of interest in that a promoter-like signature (PLS) in this region contains a potential STAT binding motif (Fig. 4C).
In concert with MUC1-C binding to STAT1 in regulating STAT1 target genes (39), ChIP studies of the PLS demonstrated that MUC1-C and STAT1 occupy this region and that silencing MUC1-C decreases their occupancy (Fig. 4D).
Consistent with these results, targeting MUC1-C and STAT1 downregulated expression of the Np63 protein (Fig. 4E and F; Supplementary Fig. S4C and S4D), indicating that MUC1-C/STAT1 signaling is necessary for Np63 expression.

MUC1-C and STAT1 are Necessary for SOX2 Expression
SOX2 is essential for self-renewal of basal cells, which are the proposed origin of SCCs (30,59).In addition, SOX2 contributes to progression of the SCC CSC state (10,33).As found for Np63, targeting MUC1-C in CAL27 and HSC3 cells downregulated SOX gene transcription and mRNA levels (Fig. 5A; Supplementary Fig. S5A).Silencing STAT1 also decreased activation of the SOX gene (Fig. 5B; Supplementary Fig. S5B), indicating that MUC1-C/STAT1 complexes drive SOX2 expression.In addressing this notion, we identified STAT1 binding motifs in a SOX PLS (Fig. 5C) that were occupied by STAT1 in a MUC1-C-dependent manner (Fig. 5D).Targeting MUC1-C and STAT1 also decreased expression of the SOX2 protein (Fig. 5E and F; Supplementary Fig. S5C and S5D), indicating that MUC1-C/STAT1 signaling integrates regulation of the Np and SOX genes in HNSCC cells.
In addition, targeting MUC1-C with GO-203 significantly inhibited SFE of CAL27 (Fig. 6F) and HSC3 (Supplementary Fig. S6F) cells, demonstrating that MUC1-C is necessary for their self-renewal capacity.GO-203 treatment of mice with established CAL27 xenografts further demonstrated dependence on MUC1-C for tumorigenicity (Fig. 6G).Immunoblot analysis of CAL27 tumors from control and GO-203-treated mice confirmed that targeting MUC1-C downregulates Np63 and SOX2 expression (Fig. 6H).

MUC1 Expression in HNSCC Tissues
To extend our findings to HNSCC tumors, we analyzed the GSE181919 scRNAseq dataset derived from 20 primary HNSCCs and four HNSCCs metastatic to lymph nodes (38).HNSCC cells were distinguished from nonmalignant cell types by the presence of DNA copy-number aberrations (CNA; refs.38,61).Analysis of malignant HNSCC cells with recurrent CNAs demonstrated higher levels of MUC1 expression compared with that in other cell types (Fig. 7A).
The GSE181919 dataset includes HPV-negative and HPV-positive HNSCCs (38), which differ in DNA methylation and gene expression profiles (15).CAL27 and HSC3 cells studied here are HPV-negative; accordingly, we analyzed HPV-negative HNSCCs in the GSE181919 dataset and found similar upregulation of MUC1 expression in the malignant cell population (Fig. 7B).Analysis of the HPV-negative/HPV-positive (Supplementary Fig. S7A) and HPV-negative (Fig. 7C) HNSCC data also detected expression of STAT1, Np63, SOX2, and NOTCH3 in HNSCC cells and, to a variable extent, (Continued) replicates relative to that obtained for vehicle-treated cells (assigned a value of 1).D, Lysates from CAL27 cells treated with vehicle or 5 μmol/L GO-203 for 3 days were immunoblotted with antibodies against the indicated proteins.E, CAL27 cells were treated with vehicle or 5 μmol/L GO-203 while they were analyzed for colony formation.Shown are representative photomicrographs of stained colonies (left).The results (mean ± SD of three determinations) are expressed as relative colony formation compared with that for vehicle-treated cells (assigned a value of 1; right).F, CAL27 cells were treated with vehicle or 5 μmol/L GO-203 while they were analyzed for tumorsphere formation.Shown are representative images of the tumorspheres (bar represents 100 μm).SFE is expressed as the mean ± SD of three independent replicates relative to that obtained for untreated cells (assigned a value of 1).G and H, Six-week-old nude mice were injected subcutaneously in the flank with 1 × 10 7 CAL27 cells.Mice pair-matched into groups of 5 mice each when tumors reached 100-150 mm 3 were treated with vehicle control or GO-203 for the indicated days.Tumor volumes are expressed as the mean ± SEM for 5 mice (G).Lysates from control and GO-203-treated tumors were immunoblotted with antibodies against the indicated proteins (H).
in other cell populations.Further analysis of malignant HNSCC cells in the HPV-negative/HPV-positive (Supplementary Fig. S7B) and HPV-negative (Supplementary Fig. S7C) datasets uncovered significant correlations of MUC1 expression with Np63, SOX2, and NOTCH3.By contrast, correlations of MUC1 with STAT1 were not significant in terms of their transcripts, which may be related to the findings that MUC1-C largely regulates STAT1 function by direct interactions at the posttranscriptional level (39).Of further interest, analysis of HPV-positive HNSCC tumors demonstrated MUC1 expression in the malignant cell population (Supplementary Fig. S7D), providing preliminary evidence for subsequent studies to determine if MUC1-C also plays a role in the progression of HPV-positive HNSCCs.

Discussion
The MUC gene evolved in mammals to protect barrier tissues from loss of homeostasis by inflammatory insults (17,18,62).The MUC1-C subunit activates EMT, epigenetic reprogramming, and repair responses to stress that, if prolonged as in settings of chronic inflammation, contribute to cancer progression (17,18,62).MUC1-C has been largely associated with adenocarcinomas that arise from barrier glandular epithelial cells lining internal organs (17,18,62).The current findings demonstrate that MUC is upregulated in metastatic HNSCCs that are derived from barrier stratified squamous epithelial cells (8).
Advanced HNSCCs invariably become refractory to multimodality treatment with chemotherapy, radiotherapy, and immunotherapy in association with development of DNA damage resistance and immune evasion (3,63,64).In regard to novelty, the current studies uncover involvement of MUC1-C in contributing to HNSCC cell intrinsic pathways of chronic inflammation.In this way, MUC1-C regulates expression of the RIG-I and MDA5 PRRs that are activated by cytosolic RNA (41)(42)(43).MUC1-C promotes DNA replicative stress; however, a role in increasing cytosolic RNA is not known (17,18,62).Nonetheless, RIG-I and MDA5 contribute to intrinsic upregulation of IFNb production and the type I IFN pathway (27,65,66).MUC1-C binds to STAT1 and regulates transactivation of STAT1 target genes (39).In concert with this function and the role of STAT1 in activation of the type I/II IFN pathways, we found that MUC1-C regulates intrinsic activation of downstream ISGs, including OAS1, MX1, and ISG15 that are effectors of DNA damage resistance and immune evasion (Fig. 7D; refs.41,47,49).
Barrier tissues are dependent on resident stem cells (SC) that, following injury, are activated to promote wound healing (67).SC lineage infidelity occurs transiently in wound healing, but persists in settings of chronic inflammation that promote cancer progression (68) .By extension and in addition to Np63 and SOX2, we found that MUC1-C regulates expression of NOTCH3, which is necessary for HNSCC cell self-renewal and progression (35).
Np63, SOX2, and NOTCH3 have each been shown to be necessary for conferring the HNSCC CSC state (8,9,(30)(31)(32)70).To our knowledge, there is no evidence for a common pathway that integrates these TFs in driving HNSCC pathogenesis.Our results support a model in which MUC1-C regulates intrinsic chronic inflammation in HNSCC cells through activation of STAT1 and the IFN type I/II pathways (Fig. 7D).In this way, targeting MUC1-C genetically and pharmacologically with the GO-203 inhibitor regulates downstream ISGs that contribute to DNA damage resistance and immune evasion (Fig. 7D).Our results also support a previously unreported role for MUC1-C/STAT1 signaling in integrating chronic inflammation with regulation of (i) lineage-dictating Np63 and SOX2 TFs, and (ii) NOTCH3 and self-renewal capacity (Fig. 7D).
Collectively, these findings and the demonstration that MUC1 associates with expression of Np63, SOX2, and NOTCH3 in individual HNSCC tumor cells indicate that MUC1-C is a common effector of the HNSCC CSC state (Fig. 7D).Chronic inflammation is an established driver of cancer progression and therapeutic resistance, albeit by unifying mechanisms that have remained unclear (71).MUC1-C is activated in barrier tissues in response to inflammation (18).Squamous epithelia of the head and neck are exposed to cigarette smoke and alcohol that contribute to chronic inflammation (8).In settings of chronic inflammation with repetitive cycles of damage and repair, prolonged MUC1-C activation becomes established in promoting cancer progression (18).The present findings shed light on the potential involvement of MUC1-C activation in chronic inflammation of squamous epithelia that extends to driving progression of the HNSCC CSC state (Fig. 7D).In addition, our findings support MUC1-C as a target for the treatment of HNSCCs with anti-MUC1-C CAR T cells and antibody-drug conjugates that are under clinical and preclinical development.

FIGURE 1
FIGURE 1Expression of MUC1 in HNSCC tumors and effects of silencing MUC1-C on HNSCC cell clonogenicity.A, Analysis of primary (P) and metastatic (M) HNSCC tissues for MUC1 expression using the GSE136037 dataset.B, CAL27/CshRNA and MUC1shRNA cells were analyzed for MUC1-C mRNA levels (left).The results (mean ± SD of four determinations) are expressed as relative levels compared with that obtained for CshRNA cells (assigned a value of 1; left).Lysates were immunoblotted with antibodies against the indicated proteins (right).C, The indicated CAL27 cells were analyzed for colony formation.Shown are representative photomicrographs of stained colonies (left).The results (mean ± SD of three determinations) are expressed as relative colony formation compared with that for CshRNA cells (assigned a value of 1; right).D, Amino acid sequence of the MUC1-C cytoplasmic domain (CD) highlighting direct interactions with JAK1 and STAT1.E and F, CAL27 cells expressing the indicated vectors were treated with vehicle or DOX for 7 days.Lysates were immunoblotted with antibodies against the indicated proteins (E).Cells were analyzed for colony formation (F).Shown are representative photomicrographs of stained colonies (top).The results (mean ± SD of three determinations) are expressed as relative colony formation compared with that for control cells (assigned a value of 1; bottom).

FIGURE 2 FIGURE 3
FIGURE 2 MUC1-C regulates a global transcriptional program enriched for STAT/IRF signaling in HNSCC cells.A, RNA-seq was performed on CAL27/tet-MUC1shRNA and HSC3/tet-MUC1shRNA cells treated with vehicle or DOX for 7 days.Volcano plots depicting downregulated (blue) and upregulated (red) DEGs (FDR<0.05;FC>2).Highlighted are the top DEGs by significance and magnitude.B, Common downregulated and upregulated genes in CAL27 and HSC3 cells with MUC1-C silencing.C, Purified chromatin from CAL27/tet-MUC1shRNA cells treated with vehicle or DOX for 7 days was immunoblotted with antibodies against the indicated proteins.D and E, Candidate enrichment plots for the HALLMARK INTERFERON ALPHA RESPONSE (D), HALLMARK INTERFERON GAMMA RESPONSE (E) gene signatures.F and G, Box plots of selected ISG expression in CAL27 (F) and HSC3 (G) cells with MUC1-C silencing.

FIGURE 4
FIGURE 4 MUC1-C/STAT1 signaling regulates Np63 expression.A, CAL27/tet-MUC1shRNA cells treated with vehicle or DOX for 7 days were analyzed for Np63 gene transcription (left) and mRNA (right) levels.The results (mean ± SD of four determinations) are expressed as relative levels compared with that obtained for vehicle-treated cells (assigned a value of 1).B, CAL27/CshRNA and CAL27/STAT1shRNA cells were analyzed for Np63 gene transcription (left) and mRNA (right) levels.The results (mean ± SD of four determinations) are expressed as relative levels compared with that obtained for vehicle-treated cells (assigned a value of 1).C, Schema of the Np63 gene with highlighting localization of the PLS region that contains potential STAT1 binding motifs.D, Soluble chromatin from CAL27/tet-MUC1shRNA cells treated with vehicle or DOX for 7 days was precipitated with anti-MUC1-C and anti-STAT1.The DNA samples were amplified by qPCR with primers for the Np63 PLS region.The results (mean ± SD of three determinations) are expressed as percentage of the input DNA for each sample.E, CAL27 cells expressing the indicated vectors were treated with vehicle or DOX for 7 days.Lysates were immunoblotted with antibodies against the indicated proteins.F, Lysates from CAL27/CshRNA and CAL27/STAT1shRNA cells were immunoblotted with antibodies against the indicated proteins.

FIGURE 5
FIGURE 5 MUC1-C/STAT1 signaling regulates activation of the SOX2 gene.A, CAL27/tet-MUC1shRNA cells treated with vehicle or DOX for 7 days were analyzed for SOX2 gene transcription (left) and mRNA (right) levels.The results (mean ± SD of four determinations) are expressed as relative levels compared with that obtained for vehicle-treated cells (assigned a value of 1).B, CAL27/CshRNA and CAL27/STAT1shRNA cells were analyzed for SOX2 gene transcription (left) and mRNA (right) levels.The results (mean ± SD of four determinations) are expressed as relative levels compared with that obtained for vehicle-treated cells (assigned a value of 1).C, Schema of the SOX2 gene with highlighting localization of the PLS region that contains STAT1 binding motifs.D, Soluble chromatin from CAL27/tet-MUC1shRNA cells treated with vehicle or DOX for 7 days was precipitated with anti-MUC1-C and anti-STAT1.The DNA samples were amplified by qPCR with primers for the SOX2 PLS region.The results (mean ± SD of three determinations) are expressed as percentage of input DNA for each sample.E, Lysates from CAL27/tet-MUC1shRNA cells treated with vehicle or DOX for 7 days were immunoblotted with antibodies against the indicated proteins.F, Lysates from CAL27/CshRNA and CAL27/STAT1shRNA cells were immunoblotted with antibodies against the indicated proteins.

FIGURE 6
FIGURE 6 MUC1-C regulates NOTCH3 expression and HNSCC cell self-renewal capacity.A, CAL27/tet-MUC1shRNA cells treated with vehicle or DOX for 7 days were analyzed for NOTCH3 mRNA levels.The results (mean ± SD of four determinations) are expressed as relative levels compared with that obtained for vehicle-treated cells (assigned a value of 1; left).Lysates were immunoblotted with antibodies against the indicated proteins (right).B, GSEA of RNA-seq data from CAL27/tet-MUC1shRNA cells treated with vehicle or DOX for 7 days using the REACTOME SIGNALING BY NOTCH3 gene signature.C, The indicated CAL27 cells treated with vehicle or DOX for 7 days were analyzed for tumorsphere formation.Shown are representative images of the tumorspheres (bar represents 100 μm).SFE is expressed as the mean ± SD of three independent (Continued on the following page.)

FIGURE 7
FIGURE 7 Single-cell profiling of the expression of MUC1 and related genes in HNSCC.A-C, UMAP representation of total cells analyzed from the GSE118389 dataset of HPV-negative/HPV-positive HNSCC samples (A), GSE118389 dataset of HPV-negative samples (B), and GSE118389 dataset of HPV-positive samples (C) (left).Distributions of MUC1 expression in individual cells (middle).MUC1 expression in individual cells in each cell type (right).D, Model depicting the current findings that MUC1-C integrates activation of STAT1, IFN type I/II signaling and ISG expression with regulation of Np63, SOX2, and NOTCH3 in driving the HNSCC CSC state that promotes DNA damage resistance and immune evasion.
8,9,[30][31][32]70)er the previously un-recognized findings that MUC1-C-induced chronic activation of the STAT1 inflammatory pathway in HNSCC cells regulates the lineage-dictating Np and SOX genes, which are amplified in HNSCCs and contribute to HNSCC pathogenesis (Fig.7D; refs.8,69).Mechanistically, we found that MUC1-C and STAT1 occupy the Np gene and that silencing MUC1-C decreases STAT1 occupancy and Np63 expression.Np is coamplified with SOX in SCCs and the Np63 and SOX2 proteins form a direct complex that regulates SCC gene expression(8,69).Like Np, we found that the SOX gene is occupied by MUC1-C and STAT1.MUC1-C was also necessary for STAT1 occupancy of SOX and targeting MUC1-C and STAT1 downregulated SOX2 expression.These unanticipated results support a model in which MUC1-C/STAT1 complexes regulate Np63 and SOX2, which are necessary for driving the HNSCC CSC state (Fig.7D; refs.8,9,[30][31][32]70)