Introduction

Head and neck squamous cell carcinoma (HNSCC) is a biologically heterogeneous malignancy arising from the mucosal epithelium of the oral cavity, pharynx, and larynx. Despite advances in surgery, radiotherapy, and systemic treatment, survival remains unsatisfactory in a substantial proportion of patients, particularly in those with locally advanced, recurrent, or treatment-refractory disease. Among HNSCC subsets, oral squamous cell carcinoma (OSCC) represents a major clinical burden and frequently requires multimodal treatment strategies [1,2]. Cisplatin remains one of the most widely used cytotoxic agents in the treatment of OSCC, either as part of definitive chemoradiotherapy or in systemic treatment settings. However, the clinical effectiveness of cisplatin is limited by intrinsic and acquired resistance, which contributes to treatment failure, recurrence, and poor prognosis [3,4]. The molecular basis of cisplatin resistance in OSCC is multifactorial and includes dysregulated DNA damage response, apoptosis escape, oxidative stress adaptation, and changes in oncogenic signaling networks [3,5]. These observations underscore the need for molecular markers that can help identify tumors with reduced cisplatin responsiveness.

A number of candidate predictive biomarkers for cisplatin efficacy have been proposed in HNSCC, but relatively few have been consistently validated across cohorts or translated into clinically useful tools. The current trend in biomarker discovery increasingly relies on integrated analyses of large-scale transcriptomic and clinical datasets, including The Cancer Genome Atlas (TCGA), as well as accessible analytical platforms such as UALCAN. These resources provide an opportunity to identify candidate genes associated with tumor biology, survival, and treatment-related outcomes in a reproducible manner [6-8].

LIM homeobox 1 (LHX1) encodes a LIM-domain-containing transcription factor classically involved in embryonic patterning, tissue differentiation, and developmental regulation. Although LHX1 has historically been studied in developmental biology, growing evidence suggests that members of the LIM homeobox family may also contribute to malignant progression through transcriptional control of cell identity, phenotypic plasticity, and invasive behavior [9]. Nevertheless, the role of LHX1 in HNSCC remains insufficiently characterized. In the present study, we investigated the clinical and biological relevance of LHX1 in HNSCC using a combined strategy incorporating TCGA-derived transcriptomic analyses, survival comparisons, cisplatin-treated subgroup evaluation, and experimental validation in cisplatin-resistant OSCC cell lines and primary organoid models. By integrating public cancer genomics data with functional observations, we aimed to assess whether LHX1 may serve as a candidate biomarker associated with cisplatin response in HNSCC.

This study was approved by the Institutional Review Board (IRB) of Kyungpook National University Hospital (IRB no. 2021-03-002-001), and written informed consent was obtained from all patients prior to the collection and use of human tissue specimens, in accordance with the principles of the Declaration of Helsinki.

Materials and Methods

1. Analysis of LHX1 mRNA expression in HNSCC and non-tumor tissues

LHX1 mRNA expression was analyzed using publicly available transcriptomic data from TCGA-HNSC through the UALCAN web portal (The University of Alabama at Birmingham). Expression levels were compared between primary tumor tissues and non-tumor tissues. The dataset included 520 primary tumor samples and 44 non-tumor samples. Gene expression was presented as transcripts per million (TPM). According to the UALCAN platform, differences in expression between the two groups were evaluated using an unpaired two-sample t-test implemented by the portal. The boxplot was reconstructed based on the reported five-number summary values (minimum, first quartile, median, third quartile, and maximum) for each group.

2. Survival and differential expression analyses of LHX1 in the TCGA-HNSC cohort

Publicly available transcriptomic and clinical data from the TCGA-HNSC cohort were used to evaluate the prognostic relevance of LHX1 in HNSCC. mRNA expression data were derived from the Illumina HiSeq RNASeqV2 platform and processed using the RSEM pipeline. Expression values were log₂-transformed as log₂(RSEM + 1) for downstream analyses. The overall cohort included 514 patients with available expression and survival data. A subgroup of 92 cisplatin-treated patients was analyzed separately for survival analysis. Patients were stratified into low- and high-LHX1 expression groups using the median expression value as the cutoff, and Kaplan–Meier survival curves were generated for the overall cohort (n = 257 per group) and the cisplatin-treated subgroup (n = 46 per group). Survival differences were assessed using the log-rank test.

To examine the association between LHX1 expression and clinical outcome, differential expression analysis was performed between disease-specific survival (DSS)-living and DSS-deceased patients in both the overall cohort and the cisplatin-treated subgroup. Differentially expressed genes were visualized using volcano plots, with the x-axis representing log₂(fold change) and the y-axis representing −log₁₀(q-value). Welch t-test was applied to log₂(RSEM + 1) expression values, followed by multiple testing correction using the Benjamini–Hochberg false discovery rate (FDR) method; adjusted p-values were reported as q-values. Positive log₂ fold-change values indicate higher expression in DSS-living patients relative to DSS-deceased patients. LHX1 was highlighted in the volcano plots, and its expression distributions in DSS-living and DSS-deceased patients were further visualized using boxplots of log₂(RSEM + 1).

3. Reagents

For cell cultures, Dulbecco’s Modified Eagle’s Medium (DMEM), fetal bovine serum (FBS), and penicillin–streptomycin used for cell culture were obtained from Invitrogen. MTT (3-[4, 5-dimethyl-2-thiazolyl]-2, 5-diphenyl-2H-tetrazolium bromide) and cisplatin were purchased from Sigma-Aldrich. QIAzol™ reagent was obtained from Qiagen, and PCR Master Mix was obtained from Takara Bio. Mouse anti-LHX1 antibody was purchased from R&D systems (MAB2725), and secondary antibody was purchased from Cell Signaling Technology (7076). Mouse anti-β-actin (HRP-conjugated) antibody was obtained from Santa Cruz Biotechnology. A pooled set of LHX1-targeting siRNAs composed of 2–3 target oligonucleotides was purchased from Santa Cruz Biotechnology.

4. Cell lines and patient-derived organoids

UMSCC1 cells, established from a floor-of-mouth squamous cell carcinoma, were purchased from Merck KGaA. The cisplatin-resistant subline, UM-CIS was generated from UMSCC1 by chronic cisplatin exposure as previously described [2]. YD-9 cells, derived from buccal mucosal squamous cell carcinoma, were obtained from the Korean Cell Line Bank, and the cisplatin-resistant subline YD-9/CIS was kindly provided by Prof. Jong In Yook (Department of Oral Pathology, Yonsei University College of Dentistry, Seoul, South Korea). Cells were maintained in DMEM supplemented with 10% FBS and 1% penicillin–streptomycin at 37℃ in a humidified incubator containing 5% CO₂. Mycoplasma contamination was checked every 2 months using the CellSafe® Mycoplasma PCR Detection Kit (CellSafe Co.).

Patient-derived OSCC organoids were established from surgical tumor specimens with minor modifications of previously described protocols [2,10]. Briefly, tumor tissues were washed with cold Advanced DMEM/F12 containing GlutaMAX, penicillin–streptomycin, HEPES, and Primocin, mechanically minced into small fragments, and digested with TrypLE. After filtration and centrifugation, the cell pellet was resuspended in cold basement membrane extract and plated as droplets onto culture dishes. Following matrix solidification, organoid culture medium was added, and the medium was replaced every 2–3 days. Organoids were passaged every 1–2 weeks and used for cisplatin response assays after reaching an approximate diameter of 300 μm.

5. Measurement of LHX1 mRNA and protein expression in OSCC cells

To compare LHX1 expression between cisplatin-sensitive and cisplatin-resistant OSCC cells, total RNA was isolated using QIAzol™ reagent and reverse-transcribed into cDNA according to standard procedures. Quantitative polymerase chain reaction (PCR) was carried out in triplicate using an ABI 7600 Real-Time PCR System (Applied Biosystems). LHX1 transcript levels were normalized to GAPDH, and relative expression was calculated by the 2^−ΔΔCt method using the corresponding parental cell line as the calibrator. Primer sequences were as follows: LHX1 forward, 5′-GACTTCTTCCGGTGTTTCGG-3′; LHX1 reverse, 5′-CACATCATGCAGGTGAAGCA-3′; GAPDH forward, 5′-AGATCATCAGCAATGCCTCCTG-3′; and GAPDH reverse, 5′-CTGGGCAGGGCTTATTCCTTTTCT-3′.

For LHX1 protein analysis, whole-cell lysates were prepared and protein concentrations were determined before electrophoresis. Equal amounts of protein (30 μg) were separated on 8% SDS–polyacrylamide gels and transferred to nitrocellulose membranes. After blocking with 5% skim milk, membranes were incubated overnight at 4℃ with anti-LHX1 antibody, followed by HRP-conjugated secondary antibody. β-Actin served as the loading control. Signals were detected using an enhanced chemiluminescence system, and band intensities were quantified by densitometry after normalization to β-actin.

6. Assessment of cisplatin sensitivity after LHX1 knockdown in cisplatin-resistant OSCC cells

To determine whether LHX1 contributes to cisplatin resistance, UM-CIS and YD-9/CIS cells were transfected with either control siRNA or LHX1-targeting siRNA. Cells were seeded in 96-well plates at 1 × 10⁴ cells per well. On the following day, transfection was performed in serum-free medium using Lipofectamine® 3000 (Thermo Fisher Scientific) with a final siRNA concentration of 10 nM. After 24 hours, cells were exposed to cisplatin (5 µg/mL) or vehicle control (0.1% DMSO in PBS) for 48 hours, and cell viability was determined by MTT assay. Knockdown efficiency was confirmed by reverse transcription quantitative PCR (RT-qPCR) in parallel samples.

7. Assessment of cisplatin efficacy after LHX1 knockdown in OSCC organoids

To evaluate the effect of LHX1 depletion in a cisplatin-sensitive three-dimensional (3D) model, two independent patient-derived OSCC organoids were transferred to 24-well plates. Organoids were transfected with control siRNA or LHX1-targeting siRNA using Lipofectamine® 3000 at a final siRNA concentration of 15 nM per well. After 24 hours, organoids were treated with cisplatin or vehicle and monitored for 5 additional days by phase-contrast microscopy at 5× magnification. Organoid growth was quantified as the Day 5/Day 0 area ratio based on projected organoid area measured from captured images. For each organoid model, the experiment was independently repeated twice, and at least five organoids per condition were included in each experiment for quantitative analysis.

8. Statistical analysis

All analyses were performed using two-sided tests, and p < 0.05 was considered statistically significant unless otherwise stated. Two-way analysis of variance was used for experiments with two independent variables, including siRNA treatment and cisplatin exposure, whereas unpaired Student’s t-tests were used for pairwise comparisons. For differential expression analyses, q < 0.05 after Benjamini–Hochberg FDR correction was considered statistically significant. Data processing and visualization were performed using Python (version 3.11) with pandas, numpy, scipy, statsmodels, lifelines, and Matplotlib. Exact statistical details, including sample size, number of biological replicates, and significance thresholds, are provided in the corresponding figure legends. All in vitro experiments were performed in biological duplicate or triplicate.

Results

1. LHX1 mRNA expression is increased in HNSCC tissues

LHX1 mRNA expression was significantly higher in primary HNSCC tissues than in normal tissues (p < 0.001). Normal tissues showed minimal expression, whereas primary tumor tissues exhibited a broader distribution of LHX1 transcript levels, with a median TPM of 0.227, an upper quartile of 0.784, and a maximum value of 2.566 (Fig. 1).

2. Association of LHX1 expression with survival and clinical outcome in HNSCC

Kaplan–Meier survival analysis showed that high LHX1 expression was associated with poorer survival in both the overall TCGA-HNSC cohort and the cisplatin-treated subgroup (log-rank p < 0.05 for both) (Fig. 2A and 2B). The cisplatin-treated group included 92 patients, and the adverse prognostic effect of high LHX1 expression appeared more pronounced in this subgroup. To further assess the relationship between LHX1 expression and clinical outcome, differential expression analysis was performed according to DSS status. In the cisplatin-treated subgroup, LHX1 was significantly different between DSS-living and DSS-deceased patients, with a log₂ ratio [Living/Deceased] of −1.9953 (p < 0.05, q < 0.05), indicating that LHX1 expression was higher in DSS-deceased patients (Fig. 2D and 2F). In contrast, although LHX1 tended to be higher in DSS-deceased patients in the overall cohort, the difference did not remain significant after multiple-testing correction (q > 0.05) (Fig. 2C and 2E). These findings suggest that elevated LHX1 expression is associated with adverse clinical outcome in HNSCC, with a more evident subgroup-specific association in cisplatin-treated patients.

3. mRNA and protein expression of LHX1 in cisplatin-sensitive and paired cisplatin-resistant OSCC cells

To determine whether LHX1 expression is associated with cisplatin resistance in OSCC cells, we compared its mRNA and protein levels between parental cells and their cisplatin-resistant counterparts. RT-qPCR analysis showed that LHX1 mRNA expression was significantly increased in UM-CIS cells compared with UMSCC1 cells and was also elevated in YD-9/CIS cells compared with YD-9 cells (Fig. 3A). Consistent with the RT-qPCR data, western blot analysis showed higher LHX1 protein expression in cisplatin-resistant cells than in the corresponding parental cells. Densitometric analysis normalized to β-actin demonstrated an approximately 2.2-fold increase in UM-CIS cells and a 1.5-fold increase in YD-9/CIS cells, relative to their parental controls (Fig. 3B). These findings suggest that LHX1 is upregulated in cisplatin-resistant OSCC cells and may be associated with the cisplatin-resistant phenotype.

4. Evaluation of cisplatin sensitivity in cisplatin-resistant OSCC cells after LHX1 knockdown

To assess whether LHX1 is functionally associated with cisplatin resistance in OSCC cells, cisplatin-resistant UM-CIS and YD-9/CIS cells were transfected with LHX1-targeting siRNA and then exposed to cisplatin. RT-qPCR confirmed effective suppression of LHX1 mRNA expression in both resistant cell lines compared with the siControl groups (Fig. 4A). Under vehicle-treated conditions, LHX1 knockdown had little effect on relative cell viability. In contrast, following treatment with cisplatin, the viability of siLHX1-transfected cells was significantly lower than that of siControl-transfected cells in both UM-CIS and YD-9/CIS cells (Fig. 4B). These findings suggest that LHX1 knockdown may enhance cisplatin sensitivity in cisplatin-resistant OSCC cells and that LHX1 may contribute to maintenance of the resistant phenotype.

5. Evaluation of cisplatin responsiveness in cisplatin-sensitive OSCC organoids after LHX1 knockdown

The effect of LHX1 knockdown on cisplatin responsiveness was further examined in cisplatin-sensitive patient-derived OSCC organoids. Representative phase-contrast images showed continued organoid growth under vehicle treatment in both the siControl and siLHX1 groups. Upon cisplatin treatment, however, growth suppression was more pronounced in organoids transfected with LHX1 siRNA than in those transfected with control siRNA (Fig. 5A). Quantitative analysis demonstrated that the relative Day 5/Day 0 organoid area ratio was significantly lower in the siLHX1 plus cisplatin group than in the siControl plus cisplatin group, whereas no substantial difference was observed under vehicle-treated conditions (Fig. 5B).

Discussion

In this study, we identified LHX1 as a gene associated with adverse clinical outcome in HNSCC and showed that LHX1 expression was elevated in cisplatin-resistant OSCC models. Functional experiments further demonstrated that LHX1 knockdown enhanced cisplatin sensitivity in resistant OSCC cells and increased the growth-inhibitory effect of cisplatin in patient-derived organoids. In addition, LHX1 expression was increased in tumor tissues compared with normal tissues and was associated with poor clinical outcome. Although high LHX1 expression was associated with poorer survival in the overall cohort by Kaplan–Meier analysis, LHX1 did not remain significantly different between DSS-living and DSS-deceased patients after multiple-testing correction. Therefore, the current public-dataset analysis does not support LHX1 as a definitive cisplatin-specific or independent prognostic marker. Rather, when interpreted together with the experimental findings, LHX1 may have value as an adjunctive biomarker for estimating treatment responsiveness. Although the association appeared more evident in the cisplatin-treated subgroup, this is more appropriately interpreted as a subgroup-specific association under treatment pressure that requires further validation. This view is consistent with previous transcriptomic studies in HNSCC indicating that LHX1 behaves as a risk-associated transcription factor and that its higher expression is linked to unfavorable clinical outcomes [11].

Our findings are also in line with emerging OSCC-focused multi-omics studies. Recent work has placed LHX1 within transcriptional programs associated with malignant epithelial subpopulations, including features related to cell motility, migration, and invasiveness [12]. These characteristics are relevant to treatment resistance, as invasive and plastic cellular states are often accompanied by enhanced survival under cytotoxic stress. Accordingly, the tumor-associated upregulation of LHX1 observed in our study may reflect not only differential expression between tumor and normal tissues, but also a broader transcriptional state linked to tumor aggressiveness. The experimental results further support this interpretation. In UM-CIS and YD-9/CIS cells, LHX1 expression was elevated at both the mRNA and protein levels relative to the corresponding parental cells, and siRNA-mediated LHX1 suppression significantly increased cisplatin sensitivity. A similar trend was observed in patient-derived OSCC organoids, in which LHX1 knockdown enhanced the growth-inhibitory effect of cisplatin. Taken together, these findings suggest that LHX1 is associated with the resistant phenotype and may functionally contribute to reduced cisplatin responsiveness. This interpretation is also supported by previous studies showing that experimentally established cisplatin-resistant OSCC cell lines are informative models for identifying resistance-associated molecular changes [13].

Several biological mechanisms may underlie this association. First, LHX1 has been implicated in epithelial–mesenchymal transition (EMT)-related transcriptional regulation in other tumor types, including changes linked to migration, invasion, and mesenchymal-like phenotypes [14]. Because EMT is closely associated with drug resistance and phenotypic plasticity, such mechanisms may partly explain the relationship between LHX1 expression and cisplatin response. Second, treatment resistance in HNSCC has been strongly linked to cancer stemness, which supports survival, self-renewal, and therapeutic escape [15-17]. In this context, recent mechanistic work in HNSCC is noteworthy, showing that the LHX1-LDB1 complex can repress STING-dependent senescence surveillance, promote cancer stem cell maintenance, and facilitate tumor progression [18]. Although these pathways were not directly tested in the present study, they provide a biologically plausible framework through which LHX1 may be associated with cisplatin resistance.

This study has several limitations. First, the clinical analyses were primarily based on HNSCC datasets, rather than exclusively on OSCC-specific patient cohorts; therefore, anatomical site heterogeneity should be considered when interpreting the results. Second, the cisplatin-treated subgroup was relatively small, limiting the precision of survival and biomarker analyses. In particular, the small subgroup size and limited number of outcome events reduced the statistical robustness of additional stratified or multivariable survival analyses. Human papilloma virus (HPV) status is a major determinant of prognosis in HNSCC, but robust HPV-stratified survival analysis was not feasible in the cisplatin-treated subgroup because only 15 HPV-positive, 70 HPV-negative, and 7 HPV-unannotated cases were available. Although HPV distribution did not differ significantly between the cisplatin-treated and non-cisplatin groups (chi-square p = 0.685; Fisher’s exact p = 0.404), the present findings should not be interpreted as evidence of HPV-independent prognostic value. Third, although our knockdown experiments support a functional association between LHX1 and cisplatin response, the downstream signaling pathways were not directly examined. In addition, the cell line and organoid experiments addressed related but distinct biological questions: the cell-line experiments were intended to test whether LHX1 contributes to maintenance of an established resistant phenotype, whereas the organoid experiments were used as a complementary 3D patient-derived model to assess cisplatin responsiveness in a more physiologically relevant setting. Because cisplatin-resistant organoid models were technically difficult to establish and were not available in the present study, validation in such models will be necessary to confirm these findings in a resistant 3D patient-derived context.

Despite these limitations, the overall pattern of evidence was generally consistent across public datasets and experimental models. LHX1 was upregulated in tumor tissue, associated with poor clinical outcome, elevated in cisplatin-resistant OSCC cells, and functionally linked to reduced cisplatin sensitivity. This interpretation is also compatible with recent efforts to construct prognostic models based on cisplatin-resistance-related genes in OSCC [19]. Accordingly, our findings support LHX1 as a candidate biomarker associated with adverse outcome and cisplatin responsiveness. Future studies should focus on independent validation in well-annotated OSCC cohorts, confirmation at the protein level in patient tissues, and mechanistic dissection of the LHX1-centered transcriptional network under cisplatin exposure.