Regular Contribution
Tumor Promoting Role of NRIP1 in Oral Squamous Cell Carcinoma: The Involvement of NSD2-Mediated Histone Methylation of DGCR8
2023 Volume 260 Issue 3 Pages 193-204
Details
2023 Volume 260 Issue 3 Pages 193-204
Oral squamous cell carcinoma (OSCC) remains the most prevalent malignance in the head and neck with highly aggressive attributes. This study investigates the functions of nuclear receptor interacting protein 1 (NRIP1) and its target transcripts in the progression of OSCC. By analyzing four OSCC-related Gene Expression Omnibus (GEO) datasets (GSE9844, GSE23558, GSE25104 and GSE74530) and querying bioinformatics systems, we obtained NRIP1 as an aberrantly highly expressed transcription factor in OSCC. Increased NRIP1 was detected in OSCC cell lines. Artificial downregulation of NRIP1 significantly suppressed proliferation, migration and invasion, resistance to apoptosis, tumorigenicity, and in vivo metastatic potential of OSCC cells. Moreover, the bioinformatics analyses suggested nuclear receptor binding SET domain protein 2 (NSD2) as a target of NRIP1 and DGCR8 microprocessor complex subunit (DGCR8) as a target of NSD2. Indeed, we validated by chromatin immunoprecipitation and luciferase assays that NRIP1 activated the transcription of NSD2, and NSD2 increased DGCR8 transcription by modulating histone methylation near the DGCR8 promoter. Either NSD2 or DGCR8 upregulation in OSCC cells rescued their malignant properties. Collectively, this study demonstrates that NRIP1 augments malignant properties of OSCC cells by activating NSD2-mediated histone methylation of DGCR8.
Head and neck squamous cell carcinoma (HNSCC) are the most prevailing malignancies that arise in the head and neck, and oral squamous cell carcinoma (OSCC) stands for the most common subtype of HNSCC and is featured with high metastasis and recurrence rates as well as treatment resistance (Yang et al. 2021). According to the global cancer statistics in 2020, the estimated annual cases for OSCC were 377,713 and the estimated deaths were 177,757 (Sung et al. 2021). Pathogenic factors of OSCC oftentimes involve abuse of alcohol, tobacco consumption, or both (Johnson et al. 2020). Distant metastasis is a major characteristic of OSCC and a decisive factor affecting the prognosis of patients (Irani 2016). The 5-year survival rate of patients can reach 90% if the malignance is diagnosed at the initial stage, while it would fall to less than 50% once recurrent and metastatic diseases occur (Bloebaum et al. 2014; Panzarella et al. 2014). It remains an urgent task to identify more predictive markers for an early diagnosis and more key molecules implicated in the malignant progression of tumor cells.
Transcriptional regulators are well-recognized key mediators in cell differentiation and transition, and their regulatory networks have causative correlation with tumor progression (Marengo et al. 2016; Xiong et al. 2022). In the present work, we obtained nuclear receptor interacting protein 1 (NRIP1) as a candidate transcription regulator with an aberrant expression profile in OSCC by analyzing four Gene Expression Omnibus (GEO) datasets. NRIP1, also known as RIP140, is a transcriptional coregulator that was initially identified in breast cancer cells via its interaction with estrogen receptor (Cavailles et al. 1995). It exerts crucial physiological roles by orchestrating the activity of a variety of transcription regulators (Lapierre et al. 2015). NRIP1 has been identified to be upregulated with oncogenic roles in several human malignancies, such as esophageal squamous cell carcinoma (Chen et al. 2020), breast cancer (Aziz et al. 2015), and nasopharyngeal carcinoma (Chao et al. 2019). On the other hand, studies have also demonstrated the tumor suppressive effect of NRIP1 in several cases (Zhang et al. 2015; Jacquier et al. 2022a, b), indicating that it might affect tumor progression in a cellular-dependent manner. However, its expression profile and function in OSCC remain unstudied yet.
When it comes to the downstream molecules affected by NRIP1, we obtained nuclear receptor binding SET domain protein 2 (NSD2) as a target of NRIP1 and DGCR8 microprocessor complex subunit (DGCR8) as a target of NSD2 via integrated bioinformatics predictions. The reason why we focused on NSD2 is that we previously found that it promoted the malignant progression of OSCC by activating the E2F transcription factor 1 (E2F1)/Y-box binding protein 2 (YBX2) axis (Zhang and Hu 2022). NSD2, also known as MMSET or WHSC1, is a SET domain-containing protein lysine methyltransferase principally catalyzing dimethylation of histone H3 at lysine 36 (H3K36me2) that favors transcriptional activation of certain oncogenes (Kuo et al. 2011). Moreover, we further obtained via bioinformatics tools that DGCR8 shows close correlation with NSD2. DGCR8 is an RNA-binding protein that plays key roles in the biogenesis of canonical microRNAs by coordinating with the ribonuclease III enzyme Drosha, which is also closely linked to the pathogenesis of cancers (Lu et al. 2020; Mirahmadi et al. 2021). However, the exact role of DGCR8 in OSCC remains untouched. Taken together, we hypothesized that there might be an NSD/DGCR8 interaction implicated in OSCC by which NRIP1 mediates the progression of the malignancy.
Four OSCC-related datasets (GSE9844, GSE23558, GSE25104 and GSE74530) were downloaded from the NCBI GEO system (https://www.ncbi.nlm.nih.gov/gds/). GSE9844 includes 26 OSCC samples and 12 normal control samples; GSE23558 contains 27 OSCC samples and 4 independent normal control samples; GSE25104 (GPL5175 platform) contains 57 OSCC samples and 22 control samples; GSE74530 includes six pairs of OSCC samples and adjacent tissue samples.
Differentially expressed genes (DEGs) between OSCC and control samples were screened using a p < 0.05 threshold, and volcano plots for DEGs in each dataset were generated using Sangerbox (http://www.sangerbox.com/tool.html). A Venn diagram of the intersection of DEGs in the four datasets was drawn using jvenn (http://jvenn.toulouse.inra.fr/app/example.html). The expression patterns of NRIP1, NSD2, and DGCR8 and genes correlated with NSD2 in HNSCC were predicted using TCGA-related datasets in UALCAN (http://ualcan.path.uab.edu/index.html), a comprehensive analysis tool for cancer omics databases.
The downstream targets of NRIP1 were predicted in hTFtarget (http://bioinfo.life.hust.edu.cn/hTFtarget/#!/), a comprehensive database available for the prediction of human transcription factors and their target genes. The predicted targets were cross-screened with the four sets of DEGs to obtain candidate downstream targets of NRIP1 in OSCC. The core proteins among the downstream candidates were obtained by establishing a protein-protein interaction (PPI) network using STRING system (https://cn.string-db.org/).
Cell incubationNormal human oral keratinocytes HOK (YS1199C), and human OSCC cell lines SCC-9 (YS336C), CAL-27 (YS498C) and HSC-3 (YS1645C) were procured from Yaji Biological (Shanghai, China). All cells were authenticated by short tandem repeat analysis and incubated in Dulbecco’s modified Eagle’s medium (D0697, Sigma-Aldrich, Merck KGaA, Darmstadt, Germany) along with 10% fetal bovine serum (FBS) and 1% antibiotics. The culture condition was maintained at 37°C with 5% CO2.
Lentivirus-based cell infectionLentivirus vectors carrying overexpression plasmid of NSD2 (oe-NSD2), oe-DGCR8, short hairpin (sh) RNA of NRIP1 (sh-NRIP1), sh-NSD2, and the corresponding negative control (NC; oe-NC and sh-NC) were constructed by Sangon Biotech Co., Ltd. (Shanghai, China). Exponentially growing SCC-9 and CAL-27 cells were seeded in six-well plates and infected with the lentivirus solution when the cell confluence reached approximately 60-70%. After 48 h of incubation at 37°C with 5% CO2, stably infected cells were screened by antibiotics, and the gene expression in cells was analyzed by reverse transcription quantitative polymerase chain reaction (RT-qPCR) to evaluate the infection efficacy.
RT-qPCRCellular RNA was extracted using the TRIzol RNA extraction reagent (N065, Nanjing Jiancheng Bioengineering Institute, Nanjing, China) and reverse-transcribed to cDNA using the Transcriptor High Fidelity cDNA Synthesis Kit (5081955001, Sigma-Aldrich). The cDNA was amplified for real-time qPCR analysis using the KiCqStart® SYBR® Green qPCR ReadyMix™ (KCQS03, Sigma-Aldrich). The PCR primers (Table 1) were designed and synthesized using the primer designing tool (http://www.ncbi.nlm.nih.gov/tools/primer-blast/). Relative gene expression was evaluated by the 2−ΔΔCt method.
Primer sequences for reverse transcription quantitative polymerase chain reaction (RT-qPCR).
NRIP1, nuclear receptor interacting protein 1; NSD2, nuclear receptor binding SET domain protein 2; DGCR8, DGCR8 microprocessor complex subunit.
Total protein in cells or xenograft tumor tissues (see details below) was extracted using the mixture of cell lysis buffer and protease inhibitor. The protein sample was separated by SDS-PAGE and transferred onto nitrocellulose membranes. The membranes were blocked with non-fat milk and then probed with the primary antibodies including NRIP1 (1:500, A9955, ABclonal Technology Co., Ltd., Wuhan, China), B-cell lymphoma-2 (Bcl-2; 1:500, A0208, ABclonal), Bcl-2-associated X (Bax; 1:500, A0207, ABclonal), NSD2 (1:500, GTX106306, GeneTex Inc., San Antonio, TX, USA), DGCR8 (1:1,000, ab191875, Abcam Inc., Cambridge, MA, USA), H3K36me2 (1:2,000, ab176921, Abcam), and β-actin (1:500, GTX109639, GeneTex) overnight at 4°C. Thereafter, the membranes were further incubated with HRP-conjugated mouse anti-rabbit IgG (1:2,000, D110065, Sangon Biotech) at room temperature for 1 h. The blot bands were developed by enhanced chemiluminescence reagent. The gray value of the protein bands was evaluated by Image J to analyze expression of target proteins.
Cell counting kit-8 (CCK-8) methodTreated SCC-9 and CAL-27 cells were digested and seeded in 96-well plates. At 0 h, 6 h, 12 h, 18 h, 24 h, 36 h, and 48 h, respectively, each well was added with 10 μL CCK-8 reagent (E606335, Sangon Biotech), followed by another 2 h of incubation at 37℃. The optical density at 450 nm was read by microplate reader.
Colony formation assayTreated SCC-9 and CAL-27 cells were seeded in six-well plates at 103 cells per well and cultured for 10 d. The colonies formed by cells were fixed by 4% paraformaldehyde (E672002, Sangon Biotech) for 15 min, stained with 0.1% crystal violet (A100528, Sangon Biotech), and counted under microscopy.
Flow cytometryAccording to the instructions of the Annexin V-fluorescein isothiocyanate (FITC)/propidium iodide (PI) apoptosis detection kit (G003-1-2, Nanjing Jiancheng), the SCC-9 and CAL-27 cells were digested and resuspended. The cell suspension (1 × 106 cells/mL) was added to the tubes and incubated with 5 μL Annexin V FITC and 5 μL PI reagent at room temperature in the dark for 10 min. The apoptosis rate of cells was then analyzed by flow cytometer.
Transwell migration and invasion assaysTreated SCC-9 and CAL-27 cells resuspended in serum-free medium was seeded into the Transwell apical chambers (Matrigel pre-coating was applied for invasion test), and the basolateral chambers were added with 500 µL fresh medium containing 10% FBS. After 24 h of incubation, the migratory or invasive cells were fixed and stained with crystal violet, and the cell number was calculated under microscopy.
Luciferase reporter assayThe promoter sequences of NSD2 and DGCR8 were inserted in the pGL4.10 luciferase vector (E6651, Promega Corporation, Madison, WI, USA) to construct corresponding luciferase reporter vectors. The constructed vectors were transfected into SCC-9 and CAL27 cells along with sh-NRIP1, sh-NSD2, or sh-NC. After 48 h, the luciferase activity in cells was analyzed according to the protocol of the dual luciferase reporter system (E1910, Promega).
Chromatin-immunoprecipitation (ChIP)-qPCRAccording to the instruction manual of the ChIP kit (Cat. No. 492024, Thermo Fisher Scientific, Rockford, IL, USA), the SCC-9 or CAL27 cells were cross-linked in 1% formaldehyde and ultrasonicated to truncate DNA to 200-500 bp fragments. Thereafter, the lysates were reacted with the antibodies of NRIP1 (customized by Sangon Biotech), NSD2 (1:50, ab75359, Abcam), and H3K36me2 (1:50, ab176921, Abcam) overnight for immunoprecipitation, with normal rabbit IgG (A7016, Beyotime Biotechnology Co., Ltd., Shanghai, China) or normal mouse IgG (A7028, Beyotime) used as the NC antibodies. The DNA in the immunoprecipitated complexes was eluted and purified for qPCR analysis.
Xenograft tumor modelsSixty male BALB/c mice (6 weeks old) were procured from Cavens Laboratory Animal Co., Ltd. (Jiangsu, China) and used in protocols approved by the Animal Ethics Committee of Affiliated Hospital of Beihua University. The animal experiment procedures also adhered to the Guide for the Care and Use of Laboratory Animals (NIH, Bethesda, MD, USA). Thirty mice were used to construct subcutaneous xenograft tumor models, and the rest 30 mice were used to analyze tumor metastasis in vivo.
For xenograft tumor induction, the 30 mice were allocated into sh-NC, sh-NRIP1, sh-NRIP1 + oe-NC, sh-NRIP1 + oe-NSD2, and sh-NRIP1 + oe-DGCR8 groups, n = 6 in each. In short, 100 μL cell suspension (containing 1 × 106 lentivirus-infected SCC-9 cells) was injected subcutaneously at the right abdomen. The length (L) and width (W) of the tumors were examined using the caliper once a week, and the tumor volume (V) was calculated as follows: V = L × W2/2. After four weeks, the mice were euthanized by excessive anesthesia, and the xenograft tumor tissues were collected.
As for tumor metastasis model, the mice were also divided into the above five groups, n = 6 in each. The same volume of SCC-9 cell suspension was injected into mice via tail vein. After four weeks, the mice were euthanized, and the lung tissues were collected for hematoxylin and eosin (HE) staining.
Immunohistochemistry (IHC)Tumor tissue sections (4 μm) were deparaffined, rehydrated, and blocked with 3% H2O2 in methanol for 30 min. After that, the sections were incubated with anti-Ki67 (1:100, SAB5700770, Sigma-Aldrich) at 4°C overnight and then with HRP-conjugated mouse anti-rabbit IgG (1:2,000, D110065, Sangon Biotech) for 1 h. After color development by DAB kit (GTX73338, GeneTex) and counter-staining by hematoxylin, followed by microscopy observation.
HE stainingThe mouse lung tissue sections (5 μm) were deparaffined, rehydrated, and then stained with 10% hematoxylin reagent and with 0.2% glacial acetic acid containing 1% eosin. After that, the tissue sections were rehydrated again, cleared, and then sealed by neutral resin for microscopy observation.
Statistical analysisExperimental data analysis and the histogram plotting were performed by GraphPad Prism 8 (GraphPad, La Jolla, CA, USA). All data were expressed as the mean ± standard deviation (SD). Differences were compared by the unpaired t test, or by the one- or two-way analysis of variance (ANOVA) plus Tukey’s post-hoc test. A p-value less than 0.05 was statistically significant.
Differentially expressed genes (DEGs) (p < 0.05) between OSCC and control samples were identified by analyzing four OSCC-related GEO datasets (GSE9844, GSE23558, GSE25104 and GSE74530) (Fig. 1A). The DEGs screened from the four datasets were cross-analyzed via the jvenn system, with 397 common outcomes obtained (Fig. 1B). We further queried these candidate genes in the hTFtarget system and identified 17 transcription factors, amongst which NRIP1 attracted our interests as its oncogenic role has been demonstrated in several cancer types, but not in OSCC. According to data in the UALCAN system, NRIP1 is highly expression in HNSC (Fig. 1C). Indeed, we confirmed that the mRNA and protein levels of NRIP1 were conspicuously elevated in OSCC cell lines compared to the normal HOK cells (Fig. 1D, E). The SCC-9 and CAL-27 cell lines with the highest levels of NRIP1 were selected for subsequent use.
NRIP1 shows a high-expression profile in oral squamous cell carcinoma (OSCC).
(A) Differentially expressed genes (DEGs) between OSCC and control samples screened from four Gene Expression Omnibus (GEO) datasets (GSE9844, GSE23558, GSE25104 and GSE74530). (B) Intersecting genes among the four sets of DEGs. (C) Transcription level of NRIP1 in head and neck squamous cell carcinoma (HNSC) (n = 520) and normal tissues (n = 44) in the UALCAN system (Transcripts per million). (D) mRNA and (E) protein levels of NRIP1 in OSCC cell lines and in normal HOK cells determined by RT-qPCR and western blot assays (n = 3). Data are expressed as the mean ± SD. Differences were compared by the t test (C) or the one-way ANOVA (D-E). *p < 0.05.
Three shRNAs of NRIP1 (sh-NRIP1 1#, 2#, and 3#) were introduced in SCC-9 and CAL-27 cells for gene knockdown, and the sh-NRIP1 1# with the best suppressive effect (Fig. 2A) was used in the subsequent assays. Under NRIP1 knockdown, both SCC-9 and CAL-27 cell lines showed significantly reduced viability and colony formation ability (Fig. 2B, C). Transwell assays also showed that the number of migratory or invasive cells was decreased (Fig. 2D, E). Moreover, flow cytometry showed that the apoptosis of the SCC-9 and CAL-27 cells was promoted by the NRIP1 knockdown (Fig. 2F), which could also be shown by the western blot assays that a decline in Bcl-2 protein and an increase in Bax protein were detected (Fig. 2G).
Knockdown of NRIP1 suppresses the malignant phenotype of oral squamous cell carcinoma (OSCC) cells.
(A) NRIP1 mRNA level in SCC-9 and CAL-27 cells after shRNA administration determined by RT-qPCR (n = 3). (B) Viability of SCC-9 and CAL-27 cells analyzed by CCK-8 assay (n = 3). (C) Colony formation ability of SCC-9 and CAL-27 cells (n = 3). (D) Migration and (E) invasion of SCC-9 and CAL-27 cells analyzed by Transwell assays (n = 3). (F) Apoptosis of SCC-9 and CAL-27 cells analyzed by flow cytometry (n = 3). (G) Protein levels of Bcl-2 and Bax in SCC-9 and CAL-27 cells determined by western blot analysis (n = 3). Differences were compared by two-way ANOVA (A-G). *p < 0.05.
We next predicted the downstream targets of NRIP1 in the hTFtarget system and had the candidate targets compared with the four sets of DEGs above, with 72 genes found intersected (Fig. 3A). A PPI network based on the 72 factors was established using the STRING system (the interaction-free factors were concealed), and the transcriptional regulation of NRIP1 on NSD2 was confirmed in the hTFtarget system (Fig. 3B). We noticed NSD2 (WHSC1) as we previously found that it played an oncogenic role in OSCC by activating the E2F1/YBX 2 axis (Zhang and Hu 2022). Similarly, NSD2 presents a high-expression profile in the UALCAN system (Fig. 3C). We validated increased NSD2 expression in the OSCC cell lines (Fig. 3D). Interestingly, in SCC-9 and CAL-27 cells induced with sh-NRIP1, reduced NSD2 mRNA was detected (Fig. 3E). The subsequent ChIP-qPCR assay revealed that the NRIP1 downregulation in cells reduced the binding of NSD2 promoter fragments with NRIP1 (Fig. 3F). Alike, the NRIP1 downregulation reduced the luciferase activity of the NSD2 promoter reporter vector in the SCC-9 and CAL-27 cells (Fig. 3G).
NRIP1 silencing suppresses nuclear receptor binding SET domain protein 2 (NSD2) transcription.
(A) Common outcomes of the downstream targets of NRIP1 predicted in the hTFtarget system with the four sets of DEGs. (B) A protein-protein interaction (PPI) network based on the 72 factors. (C) Transcription level of NSD2 in head and neck squamous cell carcinoma (HNSC) (n = 520) and normal samples (n = 44) in the UALCAN system (Transcripts per million). (D) NSD2 mRNA expression in OSCC cell lines examined by RT-qPCR (n = 3). (E) NSD2 mRNA expression in SCC-9 and CAL-27 cells induced with sh-NRIP1 determined by RT-qPCR (n = 3). (F) Abundance of NSD2 promoter fragments bound with NRIP1 examined by ChIP-qPCR (n = 3). (G) Binding between NRIP1 and NSD2 promoter examined by luciferase reporter gene assay (n = 3). Data are expressed as the mean ± SD. Differences were compared by the t test (C), or by the one-way (D) or two-way (E-G) ANOVA. *p < 0.05.
To examine more molecules implicated in NSD2-mediated oncogenic events, we queried genes showing positive correlation with NSD2 in HNSC in the UALCAN system. Meanwhile, the enrichments of NSD2 and H3K36me2 in the promoters of genes of interests were queried via the Cistrome database (http://cistrome.org/db/#/) (Fig. 4A). DGCR8, with the closest correlation with NSD2 and with the enrichments of NSD2 and H3K36me2 in its promoter, was selected for subsequent research. DGCR8 also presents a high-expression profile in HNSC in the UALCAN system (Fig. 4B). Indeed, elevated DGCR8 mRNA was detected in OSCC cell lines as well (Fig. 4C). We conjectured that NSD2 near the DGCR8 promoter might modulate the DGCR8 expression through histone methylation. Importantly, we introduced sh-NSD2 in the SCC-9 and CAL-27 cells, which not only suppressed the protein level of NSD2, but also led to a decline in the protein level of H3K36me2 (Fig. 4D). The RT-qPCR results also showed that the sh-NSD2 in cells resulted in decreased mRNA expression of DGCR8 (Fig. 4E). The ChIP-qPCR assay showed that the abundant DGCR8 promoter fragments bound by NSD2 or H3K36me2 were significantly reduced by NSD2 knockdown (Fig. 4F). The luciferase assay also revealed that sh-NSD2 reduced the transcriptional activity of DGCR8 (Fig. 4G). These results indicate that NSD2 can modulate H3K36me2 modification near the DGCR8 promoter to activate the DGCR8 transcription.
NSD2 upregulates DGCR8 through histone methylation.
(A) Enrichments of NSD2 and H3K36me2 in the promoter of DGCR8. (B) Transcription level of DGCR8 in head and neck squamous cell carcinoma (HNSC) (n = 520) and normal samples (n = 44) in the UALCAN system (Transcripts per million). (C) Expression of DGCR8 in OSCC cell lines and in normal HOK cells determined by RT-qPCR (n = 3). (D) Protein levels of NSD2 and H3K36me2 in SCC-9 and CAL-27 cells administrated with sh-NSD2 determined by western blot analysis (n = 3). (E) DGCR8 mRNA expression in SCC-9 and CAL-27 cells administrated with sh-NSD2 determined by RT-qPCR (n = 3). (F) Abundance of DGCR8 promoter fragments bound with NSD2 or H3K36me2 examined by ChIP-qPCR (n = 3). (G) Binding between NSD2 and DGCR8 promoter examined by luciferase reporter gene assay (n = 3). Data are expressed as the mean ± SD. Differences were compared by the t test (B), or by the one-way (C) or two-way ANOVA (D-G). *p < 0.05.
In SCC-9 and CAL-27 cells, silencing of NRIP1 led to reduced levels of NSD2 and DGCR8. Further administration of oe-NSD2 restored the NSD2 and DGCR8 levels in cells, and oe-DGCR8 rescued the DGCR8 level as well, though it did not affect the expression of NSD2 (Fig. 5A). Of note, either oe-NSD2 or oe-DGCR8 restored the viability and colony formation ability of the SCC-9 and CAL-27 cells (Fig. 5B, C) and increased cell migration and invasion (Fig. 5D, E). Flow cytometry showed that the cell apoptosis induced by NRIP1 silencing was largely blocked upon NSD2 or DGCR8 overexpression (Fig. 5F). In line, the western blot assay showed that NSD2 or DGCR8 upregulation also restored the protein level of Bcl-2 and reduced the protein level of Bax in cells (Fig. 5G). These results indicate that NRIP1 possibly mediates the NSD2/DGCR8 axis to induce OSCC progression.
Upregulation of NSD2 or DGCR8 rescues the malignant properties of oral squamous cell carcinoma (OSCC) cells.
(A) NSD2 and DGCR8 expression in SCC-9 and CAL-27 cells after sh-NRIP1, oe-NSD2 or oe-DGCR8 administration determined by RT-qPCR (n = 3). (B) Viability of SCC-9 and CAL-27 cells analyzed by CCK-8 assay (n = 3). (C) Colony formation ability of SCC-9 and CAL-27 cells (n = 3). (D) Migration and (E) invasion of SCC-9 and CAL-27 cells analyzed by Transwell assays (n = 3). (F) Apoptosis of SCC-9 and CAL-27 cells analyzed by flow cytometry (n = 3). (G) Protein levels of Bcl-2 and Bax in SCC-9 and CAL-27 cells determined by western blot analysis (n = 3). Data are expressed as the mean ± SD. Differences were compared by two-way ANOVA (A-G). *p < 0.05.
The SCC-9 cells infected with sh-NRIP1 alone or with the additional oe-NSD2 or oe-DGCR8 were injected into mice to induce subcutaneous tumors. It was observed that the growth rate of tumors was slowed down by NRIP1 silencing but enhanced by NSD2 or DGCR8 overexpression (Fig. 6A). Moreover, the NRIP1 silencing led to reduced immunohistochemical staining of Ki67, decreased Bcl-2 but increased Bax in the xenograft tumors, but further overexpression of NSD2 or DGCR8 led to inverse trends (Fig. 6B, C). In the tail vein injection model, the HE staining showed that NRIP1 silencing decreased the number of metastatic nodules in mouse lung tissues, and the tumor metastasis was promoted by NSD2 or DGCR8 overexpression again (Fig. 6D).
The NRIP1/NSD2/DGCR8 axis regulates tumor growth and metastasis in vivo.
(A) Representative images of the xenograft tumors and the tumor growth rate (n = 6). (B) Expression of Ki67 in mouse xenograft tumors examined by immunohistochemistry (n = 6). (C) Protein levels of Bcl-2 and Bax in mouse xenograft tumors determined by western blot analysis (n = 6). (D) Metastatic nodules in mouse lung tissues after tail vein injection of SCC-9 cells (n = 6). Data are expressed as the mean ± SD. Differences were compared by one-way (B, D) or two-way (A-C) ANOVA. *p < 0.05.
The incidence of OSCC is particularly high in Asia-Pacific countries including China due to the betel nut chewing (Ma et al. 2022). Despite advances in the historical approaches such as surgery and chemoradiotherapy, the patient’s prognosis remains unsatisfactory, especially for those with locally advanced, therapy-resistant, recurrent, and metastatic diseases (Liu et al. 2021; Wu et al. 2021). This necessitates the identification of more biomolecular mechanisms underpinning the disease progression. In the present work, the authors report a novel pathogenic network comprising the transcription regulator NRIP1, the lysine methyltransferase NSD2, and the oncogenic RNA-binding protein DGCR8 that favors the malignant properties of OSCC cells.
First, we identified that NRIP1 as a key transcriptional regulator consistently shows aberrant expression in four OSCC-related GEO datasets (GSE9844, GSE23558, GSE25104 and GSE74530). Indeed, by finely tuning the activity of transcription factors, NRIP1 has been increasingly recognized to play substantial functions in the development of solid tumors (Lapierre et al. 2015). It can suppress the transactivation ability of transcription factors through carboxyl-terminal binding proteins or by interacting with histone deacetylases (Augereau et al. 2006), or activate gene transcription through the Sp1 transcription factor (Castet et al. 2006) or nuclear factor kappa B (Zschiedrich et al. 2008). Its exact correlation with tumorigenesis has been well demonstrated. For instance, NRIP1 depletion in breast cancer cell lines has been witnessed to affect the expression of many critical genes implicated in the transformation of luminal A ductal cells (Binato et al. 2021). Likewise, siRNA silencing of NRIP1 in ESCC cells suppressed viability, colony formation ability, invasiveness, and apoptosis resistance of the cells (Chen et al. 2020). In acute myeloid leukemia, NRIP1 upregulation has been linked to poor prognosis of patients and its downregulation suppressed proliferation and survival of cells (Grasedieck et al. 2022). Interestingly, the NRIP1 has also been reported to be linked to increased infiltration of macrophages in the tumor microenvironment of stomach adenocarcinoma, therefore leading to poor prognosis (Fang and Lu 2020). Due to its differential influence on transcriptional regulation, NRIP1 also presented tumor suppressive roles in several cases such as breast cancer (Jacquier et al. 2022a, b) and hepatocellular carcinoma (Zhang et al. 2015). Overall, the expression profile and function of NRIP1 in OSCC were less investigated. Here, we validated increased NRIP1 expression in OSCC cell lines and then confirmed that the artificial downregulation of NRIP1 significantly suppressed proliferation, migration and invasion, resistance to apoptosis, tumorigenicity, and in vivo metastatic potential of the OSCC cells. The evidence suggests that NRIP1 plays a significant oncogenic role in OSCC.
We thereafter predicted the downstream targets of NRIP1 and established a PPI network, with NSD2 identified as one of the core factors. We validated downregulated NSD2 in the OSCC cell lines and confirmed that the NSD2 upregulation rescued the malignant properties of the cells. These trends were in concert with our previous findings that NSD2 silencing blocked proliferation, invasion, and tumorigenicity of OSCC cells (Zhang and Hu 2022). By catalyzing H3K36me2, NSD2 could lead to transcriptional activation of many oncogenes and augment progression of several human malignancies (Song et al. 2021), such as colorectal cancer (Zhao et al. 2021), renal cancer (Han et al. 2020), and lung adenocarcinoma (Sengupta et al. 2021). We previously identified that the oncogenic role of NSD2 in OSCC is attributive to the activation of the E2F1/YBX2 axis (Zhang and Hu 2022). In the present work, we obtained from the UALCAN system that DGCR8 showed a strong positive correlation with NSD2. Intriguingly, high DGCR8 expression has reportedly been associated with radiosensitivity of HNSCC by upregulating microRNA-106 (Zhang et al. 2020). However, by inducing maturation of several pri-microRNAs, the microprocessor protein has been found to enhance proliferation of tumor cells (Belair et al. 2015; Zhang et al. 2019). However, the exact function of DGCR8 in the malignant phenotype of OSCC cells remains unknown. Here, we confirmed that the NRIP1 downregulation led to a decline in DGCR8 in OSCC cells as well, and the overexpression of DGCR8 similarly restored proliferation, migration and invasion, resistance to apoptosis, tumorigenicity, and in vivo metastatic potential of OSCC cells as NSD2 did. The ample evidence demonstrates that the NSD2/DGCR8 axis is implicated in the OSCC progression mediated by NRIP1.
By the way, we previously found that the NSD2 could activate E2F1 to promote OSCC progression (Zhang and Hu 2022). Intriguingly, NRIP1 has also been reported to have close crosstalk between the E2F family members including E2F1 (Lapierre et al. 2015). Docquier et al. (2010) found that NRIP1 could suppress E2F1 transactivation on the promoters of several E2F target and inhibit the expression levels of the E2F1 targets. They further reported that the E2F1 could induce the transcription and expression of the NRIP1 gene (Docquier et al. 2012). Tsai et al. (2022) found a similar result that the E2F1 activation increased intracellular NRIP1 levels in glioma cell lines. Therefore, it would be interesting to investigate the potential interactions between NRIP1, NSD2, and E2F1 in OSCC and perhaps in other malignancies. We would like to focus on this issue in our later research.
In summary, the present work demonstrates that NRIP1 augments malignant properties of OSCC cells by activating NSD2-mediated histone methylation of DGCR8. The findings may shed new light in the management of OSCC that suppressing any member of the NRIP1/NSD2/DGCR8 may suppress the aggressiveness of OSCC cell lines. However, there remains a major limitation of the present work. The downstream molecules, most likely microRNAs or any signaling pathways that regulated by DGCR8, remain undefined. We would like to study on this issue in our future investigations.
This work was supported by Science and Technology Research Project of Jilin Provincial Department of Education (JJKH20200073KJ) and 2023 Jilin Provincial Department of Education Scientific Research Project (JJKH20230084KJ).
The authors declare no conflict of interest.
