Research ArticleArticles
Open Access

CRISPR/Cas9-mediated Knockout of LYVE1 In Human Tongue Cancer Cells Reveals Transcriptomic Changes in Metastasis-associated Pathways

SINI KARINEN, JOHANNA FORERO-RODRÍGUEZ, ANNIKA JÄRVINEN, KARI K. EKLUND, GONCALO BARRETO and ABDELHAKIM SALEM
Cancer Genomics & Proteomics July 2025, 22 (4) 525-537; DOI: https://doi.org/10.21873/cgp.20519

Abstract

Background/Aim: Tongue squamous cell carcinoma (TSCC), a highly aggressive subtype of head and neck cancers, is characterized by frequent lymphatic metastasis and poor prognosis. Recently, we showed that lymphatic vessel endothelial hyaluronan receptor 1 (LYVE-1) is involved in TSCC progression, yet the underlying molecular mechanisms remain unclear.

Materials and Methods: CRISPR/Cas9 gene editing was employed to generate LYVE1 knockout (KO) TSCC cell lines. Single-cell clones were isolated, screened, and validated through sequencing and Inference of CRISPR Edits (ICE) analysis and qRT-PCR. RNA sequencing was performed on LYVE1 KO and wild-type (WT) cells to identify differentially expressed genes (DEGs). Bioinformatic analyses, including Gene Ontology (GO) enrichment and protein-protein interaction (PPI) network mapping, were conducted to explore affected pathways. Finally, network topology was examined using NetworkAnalyzer and cytoHubba plugins.

Results: Transcriptomic analysis revealed significant down-regulation of pro-metastatic pathways, including epithelial-mesenchymal transition (EMT), extracellular matrix remodeling, and immune modulation. DEG analysis identified 263 genes, with key down-regulated targets such as WNT5A, TGFB2, and MMP2, and up-regulation of tumor-suppressive genes including PTGS2. GO and PPI analyses highlighted LYVE1’s pivotal role in regulating cell adhesion, migration, and immune response.

Conclusion: LYVE1 KO reduces TSCC invasive potential by disrupting EMT and tumor-stroma interactions, aligning with previous experimental findings. These results suggest LYVE1 as a critical driver of metastasis, highlighting its potential as a therapeutic target.

Keywords:

Introduction

Oral squamous cell carcinoma (OSCC) represents the most prevalent malignancy of the head and neck region, accounting for over 90% of cases. Unfortunately, it presents with a dismal 5-year survival rate of approximately 50%. Despite advancements in both diagnostic and therapeutic strategies, the prognosis remains poor, with a five year survival rate stagnant approximately at 50-64%, largely attributable to late-stage detection and a high propensity for lymphatic and distant metastases (1). Among the various subtypes of OSCC, tongue squamous cell carcinoma (TSCC) is particularly aggressive, with frequent lymph node involvement contributing significantly to patient morbidity and mortality.

Lymphatic vessel endothelial hyaluronan receptor-1 (LYVE-1) is primarily expressed in lymphatic endothelial cells, where it functions as a key hyaluronan receptor involved in fluid balance and immune cell trafficking (2, 3). Beyond its canonical role in lymphatic transport, LYVE-1 has been implicated in tumor-associated lymphangiogenesis and modulation of the tumor microenvironment, signifying its role as a potential biomarker and therapeutic target in cancer (4, 5).

Previous clinical studies indicate that elevated LYVE-1 expression correlates with increased invasiveness and nodal metastases in multiple cancers, including OSCC (4). Recently, we showed that LYVE-1-positive TSCC cells can form vessel-like structures, a process termed lymphatic mimicry (LM), which may facilitate tumor cell dissemination (6). Importantly, interference with LYVE-1 not only restricted the migratory capacity of OSCC cells but also reduced vessel-like formation in vitro, and dampened metastasis in vivo (6, 7). However, despite these advances, the molecular processes underlying LYVE-1-mediated cancer progression remain largely unexplored.

CRISPR/Cas9 gene-editing technology offers a powerful platform to dissect the functional contribution of specific genes to cancer progression (8, 9). Building on our previous findings that link LYVE-1 to OSCC metastasis, we employed CRISPR/Cas9 to generate a LYVE1 knockout (KO) in TSCC cells. We then performed a comprehensive transcriptomic analysis to identify differentially expressed genes (DEGs) associated with LYVE1 disruption and to explore cellular networks that underpin cancer cell aggressiveness. By elucidating these molecular pathways, our study aimed to lay a foundation for developing novel therapeutic interventions that target LYVE-1, potentially limiting the lymphatic spread and improving clinical outcomes for patients with TSCC.

Materials and Methods

Cell lines and culture conditions. The TSCC cell line HSC-3, representing a highly invasive OSCC model, was sourced from the Japan Health Sciences Foundation (Tokyo, Japan). Cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM; Gibco, Grand Island, NY, USA), supplemented with 10% fetal bovine serum (FBS; Sigma-Aldrich, Burlington, MA, USA), 1% penicillin-streptomycin (Sigma-Aldrich), and 1% L-glutamine. Cells were maintained at 37°C in a humidified atmosphere with 5% CO2 and were routinely passaged upon reaching 80-90% confluence.

CRISPR/Cas9-mediated knockout of LYVE-1. To generate LYVE1 knockout (KO) HSC-3 cell lines, cells were seeded at a density of 50×103 cells per well in a 24-well plate with 500 μl of growth media. The following day, cells were transfected with a CRISPR/Cas9 ribonucleoprotein (RNP) complex targeting LYVE1.

A multi-guide sgRNA hybrid (Synthego, Redwood City, CA, USA) was designed, consisting of three guide sequences: one in the forward orientation (cttTGCAGAGCTTTCCATCC) and two were in the reverse orientation (cttCGCCTTTTTGC TCACAA, cttTCAAGGCTGTTTCAACT).

The RNP complex was prepared using a 2:1 sgRNA:Cas9 ratio, based on prior optimization. Specifically, 0.8 μl of the sgRNA hybrid was combined with 0.4 μl of TrueCut Cas9 protein V2 (Invitrogen, Waltham, MA, USA), 2.5 μl of CRISPRMAX Cas9 Plus reagent (Invitrogen), and 25 μl of Opti-MEM (Gibco) in an Eppendorf tube. Separately, 25 μl of Opti-MEM was mixed with 1.5 μl of Lipofectamine CRISPRMAX reagent (Invitrogen). The two tubes were gently mixed and incubated at room temperature for 7 min to form the RNP-lipid complex, as per the manufacturer’s protocol. The transfection mixture was then added dropwise to the cells, which were then incubated under standard culture conditions (37°C, 5% CO2) for 24 h. After incubation, the media was refreshed to remove residual transfection reagents. Although not experimentally validated in this study, potential off-target effects were predicted in silico using the Custom Alt-R™ CRISPR-Cas9 guide RNA design tool (IDT, Coralville, IA, USA). The top-ranking predicted off-target sites for each sgRNA including sequences, protospacer adjacent motif (PAM) and number of mismatches are presented in Supplementary Table I.

Screening of LYVE1 knockout using PCR and sanger sequencing. To identify LYVE1 KO cell lines, single-cell clones were isolated using fluorescence-activated cell sorting (FACS) from the CRISPR/Cas9-treated HSC-3 cell population. FACS was conducted at the Biomedicum Flow Cytometry Unit, University of Helsinki. Sorted cells were plated in 96-well plates, expanded in 48-well plates, and cultured until confluency.

For initial screening, genomic DNA was extracted from single-cell clones. Cells were detached using 1× trypsin-EDTA, incubated at 37°C for five min, and neutralized with serum-containing growth media. After centrifugation, the supernatant was discarded, and the cell pellets were resuspended in 100 μl of 1:10 Proteinase K solution (Qiagen, Venlo, the Netherlands). DNA extraction was performed using the BioSite Direct PCR Kit (Qiagen), following the manufacturer’s protocol. The lysate underwent overnight digestion at 55°C, followed by enzyme inactivation at 85°C for 1 h.

PCR was performed using the following primers: Forward: 5′-GGTGCCTTGTGGAAATGCCTGG-3′; Reverse: 5′-TGGTGCCTAGGTTACCTTCC-3′. Each 20 μl PCR reaction contained: 5× HF buffer, 0.4 μl Phusion Hot Start II High-Fidelity DNA Polymerase (Thermo Fisher Scientific, Waltham, MA, USA), 2.5 mM of each dNTP (Invitrogen), and 8 μl of betaine (Thermo Fisher Scientific). Amplification was performed using a Bio-Rad C1000 Touch Thermal Cycler (Bio-Rad, Hercules, CA, USA), and PCR products were visualized on a 2% agarose gel stained with Midori Green (Nippon Genetics, Tokyo, Japan). Samples were run alongside Quick-Load 1kb Plus DNA ladder (New England biolabs, Ipswich, MA, USA). Bands corresponding to the expected amplicon size were excised and purified using the NucleoSpin Gel and PCR Clean-Up Kit (Macherey-Nagel, Düren, Germany). Selected clones were further analyzed based on band intensity and fragment size, and DNA concentrations were determined using a NanoDrop spectrophotometer (Thermo Fisher Scientific). To streamline screening, an initial pooled approach was used, where samples from multiple wells were combined. In subsequent rounds, individual samples from positive pools were screened separately to identify LYVE1 KO clones.

For definitive confirmation, selected KO clones (clones 134 and 56) and an WT control were subjected to Sanger sequencing (Eurofins Genomics, Luxembourg City, Luxembourg) using the same primers as above. Sequencing and synthego ICE CRISPR analysis tool validated the presence of CRISPR/Cas9-induced indels in the LYVE1 gene.

RNA extraction, cDNA synthesis and quantitative RT-PCR. Quantitative real-time PCR (qRT-PCR) was employed to confirm the absence of LYVE1 expression in the selected KO clones (HSC-3-56 and HSC-3-134) and untreated control cells. Cells were cultured under standard conditions in 12-well plates, and total RNA was extracted using the RNeasy Kit (Qiagen) following the manufacturer’s instructions. RNA concentrations were determined using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific). For cDNA synthesis, 400 ng of total RNA was reverse-transcribed using the iScript cDNA Synthesis Kit (Bio-Rad), as per the manufacturer’s protocol. qRT-PCR was conducted using a QuantStudio 5 system (Applied Biosystems, Waltham, MA, USA) with Fast SYBR Green Master Mix (Life Technologies Europe BV, Bleiswijk, the Netherlands, REF 4385612) and the following LYVE1-specific primers: Forward: 5′-GTGAGCAAAAAGGGCAACC-3′; Reverse: 5′-AATCCATCTCC AACCCAGC-3′.

The reaction mixture included 2 μl of cDNA template and 0.4 μM of each primer. PCR conditions were set according to the thermal cycler’s protocol. The expression of LYVE1 was normalized to the housekeeping gene GAPDH, amplified using the following primers: Forward: 5′-AAGGTGAAGGTCGGAGTCAACG-3′ Reverse: 5′-TGGTG AGCCAAAAGTTGTCG-3′.

To confirm successful LYVE1 KO, Ct values (threshold cycles) were compared. Since no LYVE1 expression was detected in the KO clones, the DDCT method was not applicable, as it requires a measurable expression level. Instead, samples where amplification did not exceed the crossing point threshold (40 cycles) were interpreted as no detectable LYVE1 mRNA expression.

RNA quality assessment and RNA sequencing. To assess RNA quality, the Agilent TapeStation 4200 RNA ScreenTape system (Agilent, Santa Clara, CA, USA) was used at the Functional Genomics Unit, University of Helsinki. A minimum of 150 ng of total RNA was extracted from LYVE1 KO clone (HSC-3-134) and an untreated HSC-3 control and sent to Eurofins Genomics for transcriptome sequencing.

The RNA-seq workflow included sequence quality control, read mapping, alignment, and differential gene expression analysis. Fastp software (10) was used for quality control, employing a sliding window approach to retain only high-quality reads. Reads with Phred quality scores below Q20 were discarded, while those with Q30 or higher were retained for downstream analysis. High-quality reads were then aligned to the human reference genome (hg38) using the Spliced Transcripts Alignment to a Reference (STAR) algorithm (11) under the Sention framework (12). Transcript quantification was performed using the featureCounts tool (13).

Differential gene expression analysis was conducted using the edgeR (14). Genes with low expression [counts per million (CPM) <1] were excluded and normalization for RNA composition was performed using the calcNormFactors function, ensuring accurate comparisons between knockout and control samples. Statistical analyses were then performed to compare gene expression between LYVE1 KO and wild-type (WT) cells, with p-values adjusted using the Benjamini-Hochberg method to control for false discovery rates.

Raw sequencing data were also processed using FastQC (v0.20.0) to assess read quality and filter out low-quality bases (Phred score <20). High-quality reads were mapped to hg38 (UCSC) using STAR (v2.7.8a), incorporating known gene models. MultiQC was used to assess overall sequencing quality. To account for sequencing depth biases, CPM values were calculated and reported.

Preprocessing and normalization of RNA-seq data were handled by Eurofins Genomics, and differential expression analysis was conducted using DESeq2 (v1.42.0) in RStudio (v4.3.0) (15). Genes with log2 fold change ≤−1 or ≥1 and adjusted p-value <0.05 were considered significantly DEGs, which were further analyzed for biological relevance.

Pathway analysis. Gene Ontology (GO) enrichment analysis was performed to determine the biological processes and molecular functions associated with the DEGs (16). The analysis was conducted using the clusterProfiler R package (v4.10.0). GO terms with adjusted p-value ≤0.05 were considered significantly enriched. Pathway visualization was conducted using GOplot (v1.0.2) and enrichplot (v1.2.0) packages.

Protein–protein interaction (PPI) network analysis. To explore functional interactions between the identified DEGs, a PPI network was constructed using the STRING database (v12.0) (17). Both up-regulated and down-regulated genes were analyzed for functional associations. The PPI network was visualized using Cytoscape (v3.10.1), and network topology was examined using the NetworkAnalyzer and cytoHubba plugins. The top 10 hub genes were identified based on degree centrality and their interactions with key regulatory partners.

Results

Generation of CRISPR/Cas9-mediated LYVE1 KO cell lines. CRISPR/Cas9-mediated LYVE1 KO cell lines were generated by isolating single cell clones, ensuring that each colony originated from a single edited cell. In the first screening phase, DNA samples from multiple clones were pooled to compress the screening process. The WT LYVE1 fragment size was 700 bp, while successful KO exhibited a 120 bp deletion, indicative of LYVE1 gene disruption. Clones displaying this deletion were advanced to the second screening phase, where individual members of the positive groups were screened separately to confirm KO status. Following this screening strategy, two LYVE1 KO HSC-3 cell lines (134 and 53) were successfully identified and expanded (Figure 1A and B). The KO status was further validated through quantitative RT-PCR (qRT-PCR), where neither of the KO clones exhibited LYVE1 amplification within 40 PCR cycles, confirming the absence of LYVE1 mRNA expression (Figure 1C and D). For further transcriptomic and functional analyses, the KO clone with the best RNA quality (clone 134) was selected as the representative LYVE1 KO model.

Figure 1.
Figure 1.

Screening and confirmation of LYVE-1 knockout (KO) cell lines. A) Agarose gel electrophoresis of PCR-amplified DNA from clones 134 and 56 and WT. Clones 134 and 56 show a band at 580 bp, indicating successful LYVE-1 deletion, while the WT sample exhibits 700 bp fragment. Clone 134 (red) was selected as the representative LYVE-1 KO model. B) KO efficiency obtained by Inference of CRISPR Edits (ICE) analysis. Purple indicates clone 56 and blue 134. C) RT-PCR validation of LYVE-1 KO. The table summarizes amplification results, where KO clones (134 and 56) show no LYVE-1 amplification, confirming the absence of LYVE-1 mRNA. Housekeeping gene GAPDH was amplified as a control for RNA integrity. D) qRT-PCR amplification curves for LYVE-1 expression in WT and 134 KO samples. WT samples (purple curves) exhibit amplification, whereas KO samples (134-1, 134-2, 134-3) show no detectable signal, confirming complete LYVE-1 KO.

Analysis of differentially expressed genes (DEGs). Transcriptomic analysis was performed on LYVE1 KO and WT HSC-3 cells, with an average read depth of 37.12 million reads per sample, covering 60,605 genes. Differential gene expression analysis revealed 263 significantly differentially regulated genes after applying a log2 fold change (log2FC) threshold of ≥±1.0 and an adjusted p-value ≤0.05. Of these, 149 genes were up-regulated, and 114 genes were down-regulated in LYVE1 KO cells compared to WT controls (Figure 2A).

Figure 2.
Figure 2.

Volcano plot and heatmap of differentially expressed genes (DEGs) in LYVE-1 knockout (KO) versus wild-type (WT) HSC-3 cells. A) Volcano plot depicts DEGs between LYVE-1 KO and WT cells. Up-regulated genes (log2 fold change ≥1.0, adjusted p-value ≤0.05) are shown in yellow, while down-regulated genes (log2 fold change ≤−1.0, adjusted p-value ≤0.05) are shown in blue. B) Heatmap of the top 70 DEGs, illustrating expression differences between KO and WT samples. The color scale represents normalized expression counts, with blue indicating lower expression and red indicating higher expression across samples.

A heatmap representing the top 70 DEGs (Figure 2B) illustrates the most significantly altered genes, highlighting distinct expression patterns between knockout and wild-type samples. These findings suggest that LYVE1 deletion modulates key transcriptional programs, potentially influencing tumor-related pathways. Supplementary Table II and Supplementary Table III list the top 20 down-regulated and up-regulated genes, respectively.

Overrepresentation analysis using Gene Ontology (GO). We performed GO analysis on DEGs across three major domains: biological processes (BP), cellular components (CC), and molecular functions (MF) (Figure 3). In total, 85 GO terms were significantly enriched. In the CC domain, differentially expressed mRNAs were primarily associated with extracellular matrix (ECM) components, including collagen-related genes. Within the MF domain, significant enrichment was observed in pathways related to metallopeptidase activity, peptide binding, amide binding, and metalloendopeptidase activity. Furthermore, the BP domain highlighted pathways involved in protein secretion regulation, endocrine pancreas development, smooth muscle cell proliferation, and cell adhesion. Notably, positive regulation of cell adhesion and regulation of smooth muscle cell proliferation emerged as key down-regulated pathways in LYVE1 KO cells compared to wild-type cells. These findings suggest that LYVE-1 may play a role in tumor progression in TSCC by modulating cell adhesion and proliferative signaling, potentially contributing to metastatic behavior.

Figure 3.
Figure 3.

Gene Ontology (GO) enrichment analysis of differentially expressed genes (DEGs) in LYVE-1 knockout (KO) versus wild-type (WT) HSC-3 cells. A) Dot plot representation of the top 20 enriched GO biological processes. The x-axis shows the rich factor (ratio of input genes annotated to a specific GO term relative to the total number of genes in that term). The color gradient represents the negative log10 of the adjusted p-value. The size of the dots corresponds to the number of gene counts in each term. B) CNET plot illustrating the relationships between enriched GO terms and their associated genes. Larger red nodes represent highly enriched biological processes, while smaller blue nodes indicate DEGs associated with each process. The analysis highlights significant pathways related to protein secretion, smooth muscle cell proliferation, and cell adhesion, revealing potential mechanism through which LYVE-1 influences tumor progression.

Protein-protein interaction analysis. To investigate the interactions between DEGs in human TSCC, we constructed a PPI network using the STRING database. The network was visualized and analyzed using Cytoscape, providing insights into key functional associations (Figure 4). The PPI network consisted of 121 nodes and 233 edges, with a combined confidence score ≥0.6, indicating a robust network of protein interactions. A total of 235 interactions were identified, illustrating the complex interconnectivity among DEGs in LYVE1 KO cells compared to WT cells.

Figure 4.
Figure 4.

Protein-protein interaction (PPI) network of differentially expressed genes (DEGs) between LYVE-1 knockout (KO) and wild-type (WT) HSC-3 cells. A) STRING-based PPI network visualizing interactions between DEGs. Nodes represent proteins and edges represent protein-protein interactions. Each protein is color-coded using the STRING glass ball design, with embedded protein structure representations. B) Refined visulaizations of the DEGs’ interaction network, where node borders are colored based on fold change (dark yellow=up-regulation, blue=down-regulation). This network highlights key functional interactions altered in LYVE-1 KO cells, providing insights into the molecular consequences of LYVE-1 loss.

Further hub gene analysis using cytoHubba identified the following 11 key hub genes: TGFB2, IL6, MMP2, CCL5, CDH2, APOE, MMP13, PTGS2 (COX-2), PPARG, along with CXCL11 and WNT5A, which shared the last position (Figure 5). These hub genes are strongly associated with cancer progression, particularly in processes such as ECM remodeling, immune response regulation, and tumor cell migration.

Figure 5.
Figure 5.

Identification of hub genes using the cytoHubba plugin in Cytoscape. A) The top 11 hub genes identified based on degree centrality, visulaized with STRING’s glass ball design. Each node represents a hub gene, displayed with its direct interactions within the network. B) Hierarchical interaction network of the top 11 hub proteins, where node border colors indicate fold change (yellow=up-regulation, blue=down-regulation). The node fill color represents degree centrality (red=highest degree of connectivity, yellow=lower connectivity), highlighting the most functionally relevant hub genes.

Discussion

Metastasis remains the leading cause of most cancer-related mortality, particularly in head and neck cancers, where lymphatic dissemination is a predominant pathway (1). The lymphatic system, essential for tissue fluid balance and immune surveillance, possesses highly permeable vessels due to loose endothelial junctions, a lack of structured basement membranes, and the lack of a central pumping mechanism – factors that render it highly susceptible to tumor cell infiltration and dissemination (18). Tumor cells can further exploit these vulnerabilities through LM, a process whereby tumor cells mimic lymphatic endothelium in vitro and in vivo (6, 19). Similar to how its homologue CD44 drives vasculogenic mimicry, LYVE-1 appears instrumental in LM formation, which has been associated with increased metastasis and poor survival outcomes in OSCC (6, 20). Building on these findings, our current study aimed to elucidate the molecular pathways underpinning the tumorigenic effects of LYVE-1 in TSCC, using CRISPR/Cas9 and transcriptomic profiling.

Multiple lines of evidence suggest a vital role of LYVE-1 in lymphangiogenesis and tumor progression (5, 6). Preclinical studies have highlighted the potential of anti-LYVE-1 therapy to halt both primary tumor formation and metastasis to lymph nodes (7). Consistent with these observations, our transcriptomic profiling of LYVE1 KO TSCC cells revealed marked alterations in key pathways underlying epithelial-mesenchymal transition (EMT), ECM degradation and metastasis. Notably, genes such as WNT5A, TGFB2, MMP2, PPARG, and PTGS2 (COX2) displayed differential expression patterns upon LYVE-1 depletion, signifying its role in regulating the aggressive phenotype of TSCC cells.

Among the most prominent changes observed was the down-regulation of TGF-β (TGFB2), an established facilitator of EMT and lymphatic metastasis (21). TGF-β supports tumor cell migration and invasion (22), and its reduction in LYVE1-depleted cells may explain, at least in part, the diminished capacity for lymphatic metastasis. Intriguingly, we also detected up-regulation of apolipoprotein E (APOE), a factor that can drive OSCC cell invasion through modulation of cholesterol metabolism and matrix metalloproteinases (MMP) expression, notably MMP7 (23). The elevated APOE in LYVE1 KO cells likely reflects a compensatory mechanism aimed at sustaining invasive potential in the face of reduced TGF-β signaling.

MMPs are critical mediators of ECM remodeling and tumor progression, and their dysregulation promotes invasion and metastasis (24). Two MMPs of relevance to OSCC, MMP2 and MMP13, were significantly down-regulated in LYVE1 KO cells. Over-expression of these two MMPs has been proposed as a predictor for lymph node metastasis and poor prognostic outcomes in patients with head and neck cancers including OSCC (25, 26).

This observation parallels the reduced expression of CDH2 (N-cadherin) and CCL5, genes closely associated with enhanced invasive capabilities, tumor cell migration, and immune cell recruitment (27, 28). While high CCL5 expression is linked to poor survival outcomes, CDH2 has been reported to regulate cell proliferation, migration, and invasion in OSCC (27, 29). Such widespread suppression of pro-invasive signals strongly suggests that LYVE1 may govern a network of transcriptional programs essential for OSCC metastasis.

Further reinforcing this concept, IL6, a cytokine implicated in OSCC pathogenesis and a promising salivary biomarker, was also decreased in LYVE1 KO cells (30). IL-6 is pivotal in activating the JAK2/STAT3/Sox4 cascade, fueling tumor growth and metastasis (31). Coupled with the down-regulation of CXCL11, which can orchestrate intravasation, extravasation, and immune evasion (32), these shifts denote a broader impairment of metastatic signaling pathways. Likewise, WNT5A, instrumental in non-canonical Wnt signaling and the facilitation of OSCC cell migration and invasion (33, 34), showed decreased expression, strengthening the hypothesis that LYVE1 is intricately connected to multiple, intersecting oncogenic pathways.

Conversely, PTGS2 (COX2) was up-regulated in LYVE1 KO cells. Although COX-2 often stimulates pro-tumorigenic events, such as angiogenesis via increased prostaglandin levels (35), its role here appears more complex. COX-2 inhibition can up-regulate PPARγ, promoting apoptosis and reducing tumor growth, and PPARγ can further attenuate COX-2 activity (36, 37). Thus, the observed reduction of PPARγ in LYVE1-depleted cells may partially explain COX2 up-regulation. In a broader context, disruption of LYVE1 seems to tip the balance of multiple interacting pathways (e.g., WNT5A, and TGFβ) in a manner largely unfavorable to tumor progression, even if isolated pro-angiogenic signals remain.

Several limitations require consideration when interpreting these data. First, our current work focuses on in vitro transcriptomic analyses without corresponding in vivo or proteomic validation, limiting our ability to definitely link these gene expression changes to functional outcomes in tumor progression. Moreover, although CRISPR/Cas9 system was employed as an RNP and therefore minimizes off-target effects, further genome editing studies will be necessary to corroborate these findings (38). Finally, while our previous siRNA-based experiments in HSC-3 cells demonstrated functional consequences of LYVE-1 silencing (6), the present CRISPR study in TSCC did not include parallel phenotypic assays, leaving important questions about the exact biological impact of these transcriptional shifts unanswered.

Despite these constraints, our findings suggest novel therapeutic strategies for TSCC. Given its key role in lymphangiogenesis, targeting LYVE-1 might selectively block lymphatic dissemination and LM while sparing the blood vasculature. The downstream impact on factors such as TGFB2, MMP2, and COX-2 highlights how targeting LYVE-1 could constrain multiple drivers of cancer, including invasion, metastasis, and angiogenesis. This multi-faceted approach holds promise for improving patient outcomes, particularly in advanced TSCC, where lymph node involvement remains a grim prognostic factor.

Conclusion

In summary, KO of LYVE1 in TSCC significantly perturbs the expression of key genes integral to EMT, ECM degradation, immune modulation, and angiogenesis. These results build on our previous siRNA-based findings, reinforcing the concept that LYVE-1 critically orchestrates pro-metastatic pathways in TSCC. Future studies should incorporate in vivo models and comprehensive proteomics validation to elucidate the full scope of LYVE-1’s influence on tumor biology. Ultimately, targeting LYVE-1 may represent a compelling route to mitigate lymphatic metastasis and improve therapeutic outcomes in head and neck cancers.

Acknowledgements

HSC-3 cell line has been authenticated by Technology Centre, Institute for Molecular Medicine Finland FIMM, University of Helsinki. The HSC-3 cell line was provided by the Oral Cancer Group.

Footnotes

  • Authors’ Contributions

    Study conceptualization, S.K., G.B., and A.S.; Methodology, S.K., A.J., G.B., and A.S.; Validation, S.K., A.J., G.B., and A.S; Formal analyses, S.K., J.F.-R., K.K.E., G.B., and A.S.; Investigation, S.K., and A.J.; Data Curation, S.K., and J.F.-R.; Writing – Original Draft, S.K., and J.F.-R.; Writing – Review & Editing, K.K.E., G.B., and A.S.; Visualization, S.K., and J.F.-R.; Project administration, A.S.; Funding acquisition, A.S. All Authors have read and agreed to the published version of the manuscript.

  • Supplementary Material

    Supplementary material is available at: https://doi.org/10.5281/zenodo.15075901

  • Conflicts of Interest

    The Authors have no conflicts of interest to declare in relation to this study.

  • Funding

    Open Access funding provided by University of Helsinki (including Helsinki University Central Hospital). This study was funded by the Research Council of Finland (Academy of Finland; No. 362035); Sigrid Jusélius Foundation (No. 250210); The PhD Study Track (Tutkijalinja LTT), Faculty of Medicine, University of Helsinki; The Finnish Dental Society, Apollonia; The Emil Aaltonen Foundation; Minerva Foundation. J F-R is supported by the EU Horizon Europe ENDOTARGET project (No. 101095084).

  • Received March 24, 2025.
  • Revision received April 15, 2025.
  • Accepted May 16, 2025.

This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).

References

  1. ↵
    1. Tan Y,
    2. Wang Z,
    3. Xu M,
    4. Li B,
    5. Huang Z,
    6. Qin S,
    7. Nice EC,
    8. Tang J,
    9. Huang C
    : Oral squamous cell carcinomas: state of the field and emerging directions. Int J Oral Sci 15(1): 44, 2023. DOI: 10.1038/s41368-023-00249-w
    OpenUrlCrossRefPubMed
  2. ↵
    1. Wu M,
    2. Du Y,
    3. Liu Y,
    4. He Y,
    5. Yang C,
    6. Wang W,
    7. Gao F
    : Low molecular weight hyaluronan induces lymphangiogenesis through LYVE-1-mediated signaling pathways. PLoS One 9(3): e92857, 2014. DOI: 10.1371/journal.pone.0092857
    OpenUrlCrossRefPubMed
  3. ↵
    1. Banerji S,
    2. Ni J,
    3. Wang SX,
    4. Clasper S,
    5. Su J,
    6. Tammi R,
    7. Jones M,
    8. Jackson DG
    : LYVE-1, a new homologue of the CD44 glycoprotein, is a lymph-specific receptor for hyaluronan. J Cell Biol 144(4): 789-801, 1999. DOI: 10.1083/jcb.144.4.789
    OpenUrlAbstract/FREE Full Text
  4. ↵
    1. Karinen S,
    2. Hujanen R,
    3. Salo T,
    4. Salem A
    : The prognostic influence of lymphatic endothelium-specific hyaluronan receptor 1 in cancer: A systematic review. Cancer Sci 113(1): 17-27, 2022. DOI: 10.1111/cas.15199
    OpenUrlCrossRefPubMed
  5. ↵
    1. Jackson DG,
    2. Prevo R,
    3. Clasper S,
    4. Banerji S
    : LYVE-1, the lymphatic system and tumor lymphangiogenesis. Trends Immunol 22(6): 317-321, 2001. DOI: 10.1016/s1471-4906(01)01936-6
    OpenUrlCrossRefPubMed
  6. ↵
    1. Karinen S,
    2. Juurikka K,
    3. Hujanen R,
    4. Wahbi W,
    5. Hadler-Olsen E,
    6. Svineng G,
    7. Eklund KK,
    8. Salo T,
    9. Åström P,
    10. Salem A
    : Tumour cells express functional lymphatic endothelium-specific hyaluronan receptor in vitro and in vivo: Lymphatic mimicry promotes oral oncogenesis? Oncogenesis 10(3): 23, 2021. DOI: 10.1038/s41389-021-00312-3
    OpenUrlCrossRefPubMed
  7. ↵
    1. Hara Y,
    2. Torii R,
    3. Ueda S,
    4. Kurimoto E,
    5. Ueda E,
    6. Okura H,
    7. Tatano Y,
    8. Yagi H,
    9. Ohno Y,
    10. Tanaka T,
    11. Masuko K,
    12. Masuko T
    : Inhibition of tumor formation and metastasis by a monoclonal antibody against lymphatic vessel endothelial hyaluronan receptor 1. Cancer Sci 109(10): 3171-3182, 2018. DOI: 10.1111/cas.13755
    OpenUrlCrossRefPubMed
  8. ↵
    1. Bhattacharjee G,
    2. Gohil N,
    3. Khambhati K,
    4. Mani I,
    5. Maurya R,
    6. Karapurkar JK,
    7. Gohil J,
    8. Chu DT,
    9. Vu-Thi H,
    10. Alzahrani KJ,
    11. Show PL,
    12. Rawal RM,
    13. Ramakrishna S,
    14. Singh V
    : Current approaches in CRISPR-Cas9 mediated gene editing for biomedical and therapeutic applications. J Control Release 343: 703-723, 2022. DOI: 10.1016/j.jconrel.2022.02.005
    OpenUrlCrossRefPubMed
  9. ↵
    1. Rafii S,
    2. Tashkandi E,
    3. Bukhari N,
    4. Al-Shamsi HO
    : Current status of CRISPR/Cas9 application in clinical cancer research: opportunities and challenges. Cancers (Basel) 14(4): 947, 2022. DOI: 10.3390/cancers14040947
    OpenUrlCrossRefPubMed
  10. ↵
    1. Chen S,
    2. Zhou Y,
    3. Chen Y,
    4. Gu J
    : fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics 34(17): i884-i890, 2018. DOI: 10.1093/bioinformatics/bty560
    OpenUrlCrossRefPubMed
  11. ↵
    1. Dobin A,
    2. Davis CA,
    3. Schlesinger F,
    4. Drenkow J,
    5. Zaleski C,
    6. Jha S,
    7. Batut P,
    8. Chaisson M,
    9. Gingeras TR
    : STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29(1): 15-21, 2013. DOI: 10.1093/bioinformatics/bts635
    OpenUrlCrossRefPubMed
  12. ↵
    1. Sention
    : Sentieon provides complete solutions for secondary dna/rna analysis for a variety of sequencing platforms, including short and long reads, 2021. Available at: https://www.sentieon.com/
  13. ↵
    1. Liao Y,
    2. Smyth GK,
    3. Shi W
    : featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics 30(7): 923-930, 2014. DOI: 10.1093/bioinformatics/btt656
    OpenUrlCrossRefPubMed
  14. ↵
    1. Robinson MD,
    2. McCarthy DJ,
    3. Smyth GK
    : edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26(1): 139-140, 2010. DOI: 10.1093/bioinformatics/btp616
    OpenUrlCrossRefPubMed
  15. ↵
    1. RCoreTeam
    : R: A language and environment for statistical computing. 2013. Available at: http://www.R-project.org/
  16. ↵
    1. Ashburner M,
    2. Ball CA,
    3. Blake JA,
    4. Botstein D,
    5. Butler H,
    6. Cherry JM,
    7. Davis AP,
    8. Dolinski K,
    9. Dwight SS,
    10. Eppig JT,
    11. Harris MA,
    12. Hill DP,
    13. Issel-Tarver L,
    14. Kasarskis A,
    15. Lewis S,
    16. Matese JC,
    17. Richardson JE,
    18. Ringwald M,
    19. Rubin GM,
    20. Sherlock G
    : Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat Genet 25(1): 25-29, 2000. DOI: 10.1038/75556
    OpenUrlCrossRefPubMed
  17. ↵
    1. Szklarczyk D,
    2. Gable AL,
    3. Nastou KC,
    4. Lyon D,
    5. Kirsch R,
    6. Pyysalo S,
    7. Doncheva NT,
    8. Legeay M,
    9. Fang T,
    10. Bork P,
    11. Jensen LJ,
    12. von Mering C
    : The STRING database in 2021: customizable protein-protein networks, and functional characterization of user-uploaded gene/measurement sets. Nucleic Acids Res 49(D1): D605-D612, 2021. DOI: 10.1093/nar/gkaa1074
    OpenUrlCrossRefPubMed
  18. ↵
    1. Hu Z,
    2. Zhao X,
    3. Wu Z,
    4. Qu B,
    5. Yuan M,
    6. Xing Y,
    7. Song Y,
    8. Wang Z
    : Lymphatic vessel: origin, heterogeneity, biological functions, and therapeutic targets. Signal Transduct Target Ther 9(1): 9, 2024. DOI: 10.1038/s41392-023-01723-x
    OpenUrlCrossRef
  19. ↵
    1. Wang QS,
    2. He R,
    3. Yang F,
    4. Kang LJ,
    5. Li XQ,
    6. Fu L,
    7. Sun B,
    8. Feng YM
    : FOXF2 deficiency permits basal-like breast cancer cells to form lymphangiogenic mimicry by enhancing the response of VEGF-C/VEGFR3 signaling pathway. Cancer Lett 420: 116-126, 2018. DOI: 10.1016/j.canlet.2018.01.069
    OpenUrlCrossRefPubMed
  20. ↵
    1. Paulis YW,
    2. Huijbers EJ,
    3. van der Schaft DW,
    4. Soetekouw PM,
    5. Pauwels P,
    6. Tjan-Heijnen VC,
    7. Griffioen AW
    : CD44 enhances tumor aggressiveness by promoting tumor cell plasticity. Oncotarget 6(23): 19634-19646, 2015. DOI: 10.18632/oncotarget.3839
    OpenUrlCrossRefPubMed
  21. ↵
    1. Itoh F,
    2. Watabe T
    : TGF-β signaling in lymphatic vascular vessel formation and maintenance. Front Physiol 13: 1081376, 2022. DOI: 10.3389/fphys.2022.1081376
    OpenUrlCrossRef
  22. ↵
    1. Kim S,
    2. Lee J,
    3. Jeon M,
    4. Nam SJ,
    5. Lee JE
    : Elevated TGF-β1 and -β2 expression accelerates the epithelial to mesenchymal transition in triple-negative breast cancer cells. Cytokine 75(1): 151-158, 2015. DOI: 10.1016/j.cyto.2015.05.020
    OpenUrlCrossRefPubMed
  23. ↵
    1. Jayakar SK,
    2. Loudig O,
    3. Brandwein-Gensler M,
    4. Kim RS,
    5. Ow TJ,
    6. Ustun B,
    7. Harris TM,
    8. Prystowsky MB,
    9. Childs G,
    10. Segall JE,
    11. Belbin TJ
    : Apolipoprotein E promotes invasion in oral squamous cell carcinoma. Am J Pathol 187(10): 2259-2272, 2017. DOI: 10.1016/j.ajpath.2017.06.016
    OpenUrlCrossRefPubMed
  24. ↵
    1. Kessenbrock K,
    2. Plaks V,
    3. Werb Z
    : Matrix metalloproteinases: regulators of the tumor microenvironment. Cell 141(1): 52-67, 2010. DOI: 10.1016/j.cell.2010.03.015
    OpenUrlCrossRefPubMed
  25. ↵
    1. Pramanik KK,
    2. Mishra R
    : ERK-mediated upregulation of matrix metalloproteinase-2 promotes the invasiveness in human oral squamous cell carcinoma (OSCC). Exp Cell Res 411(1): 112984, 2022. DOI: 10.1016/j.yexcr.2021.112984
    OpenUrlCrossRefPubMed
  26. ↵
    1. Li S,
    2. Pritchard DM,
    3. Yu LG
    : Regulation and function of matrix metalloproteinase-13 in cancer progression and metastasis. Cancers (Basel) 14(13): 3263, 2022. DOI: 10.3390/cancers14133263
    OpenUrlCrossRefPubMed
  27. ↵
    1. Qiu J,
    2. Xu L,
    3. Zeng X,
    4. Wu H,
    5. Liang F,
    6. Lv Q,
    7. Du Z
    : CCL5 mediates breast cancer metastasis and prognosis through CCR5/Treg cells. Front Oncol 12: 972383, 2022. DOI: 10.3389/fonc.2022.972383
    OpenUrlCrossRefPubMed
  28. ↵
    1. Luo Y,
    2. Yu T,
    3. Zhang Q,
    4. Fu Q,
    5. Hu Y,
    6. Xiang M,
    7. Peng H,
    8. Zheng T,
    9. Lu L,
    10. Shi H
    : Upregulated N-cadherin expression is associated with poor prognosis in epithelial-derived solid tumours: A meta-analysis. Eur J Clin Invest 48(4): e12903, 2018. DOI: 10.1111/eci.12903
    OpenUrlCrossRefPubMed
  29. ↵
    1. Yao Y,
    2. Chen S,
    3. Lu N,
    4. Yin Y,
    5. Liu Z
    : LncRNA JPX overexpressed in oral squamous cell carcinoma drives malignancy via miR-944/CDH2 axis. Oral Dis 27(4): 924-933, 2021. DOI: 10.1111/odi.13626
    OpenUrlCrossRefPubMed
  30. ↵
    1. Ferrari E,
    2. Pezzi ME,
    3. Cassi D,
    4. Pertinhez TA,
    5. Spisni A,
    6. Meleti M
    : Salivary cytokines as biomarkers for oral squamous cell carcinoma: a systematic review. Int J Mol Sci 22(13): 6795, 2021. DOI: 10.3390/ijms22136795
    OpenUrlCrossRefPubMed
  31. ↵
    1. Xiao L,
    2. Li X,
    3. Cao P,
    4. Fei W,
    5. Zhou H,
    6. Tang N,
    7. Liu Y
    : Interleukin-6 mediated inflammasome activation promotes oral squamous cell carcinoma progression via JAK2/STAT3/Sox4/NLRP3 signaling pathway. J Exp Clin Cancer Res 41(1): 166, 2022. DOI: 10.1186/s13046-022-02376-4
    OpenUrlCrossRef
  32. ↵
    1. Wang J,
    2. Ouyang X,
    3. Zhu W,
    4. Yi Q,
    5. Zhong J
    : The role of CXCL11 and its receptors in cancer: prospective but challenging clinical targets. Cancer Control 31: 10732748241241162, 2024. DOI: 10.1177/10732748241241162
    OpenUrlCrossRefPubMed
  33. ↵
    1. Xie J,
    2. Huang L,
    3. Lu YG,
    4. Zheng DL
    : Roles of the Wnt signaling pathway in head and neck squamous cell carcinoma. Front Mol Biosci 7: 590912, 2021. DOI: 10.3389/fmolb.2020.590912
    OpenUrlCrossRefPubMed
  34. ↵
    1. Asem MS,
    2. Buechler S,
    3. Wates RB,
    4. Miller DL,
    5. Stack MS
    : Wnt5a signaling in cancer. Cancers (Basel) 8(9): 79, 2016. DOI: 10.3390/cancers8090079
    OpenUrlCrossRefPubMed
  35. ↵
    1. Frejborg E,
    2. Salo T,
    3. Salem A
    : Role of cyclooxygenase-2 in head and neck tumorigenesis. Int J Mol Sci 21(23): 9246, 2020. DOI: 10.3390/ijms21239246
    OpenUrlCrossRefPubMed
  36. ↵
    1. Liu B,
    2. Qu L,
    3. Yan S
    : Cyclooxygenase-2 promotes tumor growth and suppresses tumor immunity. Cancer Cell Int 15: 106, 2015. DOI: 10.1186/s12935-015-0260-7
    OpenUrlCrossRefPubMed
  37. ↵
    1. Heudobler D,
    2. Rechenmacher M,
    3. Lüke F,
    4. Vogelhuber M,
    5. Pukrop T,
    6. Herr W,
    7. Ghibelli L,
    8. Gerner C,
    9. Reichle A
    : Peroxisome proliferator-activated receptors (PPAR)γ agonists as master modulators of tumor tissue. Int J Mol Sci 19(11): 3540, 2018. DOI: 10.3390/ijms19113540
    OpenUrlCrossRefPubMed
  38. ↵
    1. Wei T,
    2. Cheng Q,
    3. Min YL,
    4. Olson EN,
    5. Siegwart DJ
    : Systemic nanoparticle delivery of CRISPR-Cas9 ribonucleoproteins for effective tissue specific genome editing. Nat Commun 11(1): 3232, 2020. DOI: 10.1038/s41467-020-17029-3
    OpenUrlCrossRef
Back to top

In this issue

Print
Download PDF
Article Alerts
Email Article
Citation Tools
Reprints and Permissions
CRISPR/Cas9-mediated Knockout of LYVE1 In Human Tongue Cancer Cells Reveals Transcriptomic Changes in Metastasis-associated Pathways
SINI KARINEN, JOHANNA FORERO-RODRÍGUEZ, ANNIKA JÄRVINEN, KARI K. EKLUND, GONCALO BARRETO, ABDELHAKIM SALEM
Cancer Genomics & Proteomics Jul 2025, 22 (4) 525-537; DOI: 10.21873/cgp.20519
Twitter logo Facebook logo Mendeley logo

Jump to section

Keywords