Proliferation-associated 2G4 P48 is stabilized by malignant T-cell amplified sequence 1 and promotes the proliferation of head and neck squamous cell carcinoma

Abstract

Background/purpose

Proliferation-associated protein 2G4 (PA2G4) has alternative transcriptional and translational initiation. One dominant transcript ENST00000303305 could be translated into two protein isoforms (PA2G4-P42 and PA2G4-P48). In this study, we aimed to explore the effects of PA2G4-P42 and PA2G4-P48 on the proliferation of head and neck squamous cell carcinoma (HNSCC) and the mechanisms regulating PA2G4-P48 stability.

Materials and methods

HNSCC cell lines HSC2 and SCC25 with relatively low PA2G4 expression were used for in-vitro cell studies. PA2G4-P42 and PA2G4-P48 overexpression lentiviruses were generated. In vitro cell proliferation was assessed by CCK-8 and colony formation. In vivo tumor cell proliferation was assessed by HSC2 cell-derived xenograft tumors. Liquid chromatography-mass spectrometry (LC-MS)/MS and co-immunoprecipitation (co-IP) assays were applied to check PA2G4-P48 interacting partners. Cycloheximide (CHX) chase and ubiquitin-based co-IP assays were also performed.

Results

PA2G4-P48 was the dominant isoform, with substantially higher expression than PA2G4-P42 in HNSCC. PA2G4-P48 overexpression enhanced HNSCC cell proliferation, but PA2G4-P42 overexpression slowed the proliferation. MCTS1 interacted with PA2G4-P48, but not PA2G4-P42. PA2G4 protein but not its mRNA expression was decreased in cells with MCTS1 knockdown. MG132 treatment abrogated this alteration. MCTS1 overexpression significantly elevated the half-life of PA2G4-P48, while its knockdown drastically reduced the half-life compared with the control cells. In addition, MCTS1 overexpression significantly decreased the polyubiquitination of exogenous flag-tagged PA2G4-P48. MCTS1 overexpression-induced cell proliferation was hampered by knocking down of PA2G4-P48.

Conclusion

PA2G4-P42 and PA2G4-P48 exert growth-suppressive and growth-promoting effects in HNSCC, respectively. MCTS1 can interact with PA2G4-P48 and prolong its half-life by reducing its poly-ubiquitination.

Introduction

Proliferation-associated protein 2G4 (PA2G4), also known as Erb-B2 Receptor Tyrosine Kinase 3 (ERBB3)-binding protein 1 (EBP1), can regulate cell growth and differentiation in both normal and tumor cells.^1,2 Its upregulation is observed in multiple types of cancers, including head and neck squamous cell carcinoma (HNSCC).3, 4, 5, 6 In addition, its upregulation was associated with poor prognosis in patients with HNSCC.³

PA2G4 gene has alternative splicing during the transcriptional process.² One dominant transcript, ENST00000303305 (NM_006191) could be translated into two different protein isoforms by alternative translational initiation, including a long PA2G4-P48 (394 aa) isoform and a short PA2G4-P42 (340 aa).² These two isoforms have distinct regulations in cancer biology.⁷ PA2G4-P48 shows tumor-promoting effects via multiple mechanisms, such as inhibiting the tumor-suppressing activity of p53,⁸ enhancing the expression of antiapoptotic proteins,⁹ and binding to oncogenic proteins and reduce their proteasomal degradation, such as MYCN,¹⁰ cyclin E, c-Myc and Aurora A.⁵ In comparison, PA2G4-P42 showed potent tumor suppressive effects.^2,11 In oral squamous cell carcinoma, PA2G4 upregulation can enhance colony formation, invasion and migration of the tumor cells.⁶ However, the functional differences between the two isoforms in HNSCC are not well understood.

PA2G4-P42 expression is significantly lower than PA2G4-P48 in tumors due to complex underlying mechanisms.² PA2G4-P42, but PA2G4-P48 could be rapidly polyubiquitination via binding of the E3 ubiquitin-protein ligase, Bre1. Therefore, PA2G4-P42 was subjected to rapid proteasomal degradation after translation.¹² One recent study found that FBXW7 (F-box and WD40 domain protein 7), a substrate recognition subunit of the SCF (SKP1/CUL1/F-box protein) E3 ubiquitin ligase complex could bind to PA2G4-P48 and facilitate EBP1 PA2G4-P48 turnover through the ubiquitin-dependent proteasome pathway.⁵ However, why PA2G4-P48 is stabilized and kept away from proteasomal degradation is not clear.

In this study, we aimed to compare the effects of PA2G4-P42 and PA2G4-P48 on the proliferation of HNSCC and the mechanisms regulating PA2G4-P48 stability.

Materials and methods

Bioinformatic analysis

Gene expression and survival data in The Cancer Genome Atlas-HNSCC subset were extracted as we previously described. Kaplan-Meier (K-M) survival analysis was conducted to assess the difference in overall survival (OS) between patients with high and low ENST00000303305 expression.

Cell culture and treatment

Human oral squamous carcinoma cells (HSC2 and SCC25) were obtained from Otwo Biotech (Shenzhen, China) and were maintained in Dulbecco's Modified Eagle's medium (DMEM) supplemented with streptomycin (100 μg/mL), penicillin (100 units/mL), 2 mM glutamine, and 10% fetal bovine serum (FBS).

PA2G4 P48 and P42 human cDNA were amplified from human PA2G4 cDNA clone (NM_006191.2) and were inserted into the pLVX-IRES-Puro-3xFlag vector, for generation of flag-tagged PA2G4 P48 (flag-P48) and flag-tagged PA2G4 P42 (flag-P42). Myc-tagged MCTS1 (NM_014060) vector was generated using pLVX-IRES-Puro plasmid. pLenti-puro-HA-Ubiquitin (#74218) was obtained from Addgene (Watertown, MA, USA). Lentivirus for gene knockdown was generated based on pLKO.1-puro vector. The validated shRNA sequences by one previous study were used: shMCTS1#1, 5′-CCCTAAGATTACTTCACAAAT-3′ and shMCTS1#2, 5′-TGTACTCAGTGGAGCAAATAT-3′.¹³ The validated shRNA sequence targeting PA2G4-P48 was used: shPA2G4, 5′-GAGCAACAGGAGCAAACTA-3′.⁵ Lentivirus for infection was produced by co-transfecting the recombinant vector plasmid, packaging plasmids, psPAX2 and envelope plasmid, pMD2.G into packaging cell line (293T), following the protocol introduced previously.¹⁴ Cycloheximide (CHX) (a protein synthesis inhibitor) and MG132 (a proteasome inhibitor) were purchased from Selleck (Houston, TX, USA).

Western blotting assays

Western blotting assays were performed using the method described previously.¹⁵ The protein band intensities were quantified by the Image J software. The following primary antibodies and dilutions were used: anti-PA2G4 (1:1000, 11147-2-AP, Proteintech, Wuhan, China); anti-cyclin E (1:1000, 11554-1-AP, Proteintech), anti-c-Myc (1:2000, 10828-1-AP, Proteintech), anti-Aurora A (1:2000, 66757-1-Ig, Proteintech), anti-β-actin (1:2000, 20536-1-AP, Proteintech).

Cell proliferation assays

CCK-8 and colony formation assays were conducted to assess cell proliferation. For CCK-8 assays, HSC2 and SCC25 cells with flag-P42 or flag-P48 overexpression separately, or with flag-P48 overexpression and MCTS1 knockdown in combination were seeded into 96-well plate at 5 × 10³ cells/well. Then, the cells were cultured for 72 h after. The absorbance at 450 nm was detected at 0 h and 72 h using the Cell Counting Kit-8 (CCK-8) assay (Dojindo, Tokyo, Japan), with a microplate reader (Bio-Rad, Irvine, CA, USA). For colony formation, HSC2 and SCC25 cells with flag-P42 or flag-P48 overexpression separately, or with flag-P48 overexpression and MCTS1 knockdown in combination were plated into 24-well plates (500 cells/well) and were cultured for 10–14 days. Then, the colonies were fixed with methanol and stained with crystal violet (0.2%). Images of the plates were taken with a scanner. Cell colonies were counted if they had at least 50 cells. Results were obtained based on three technical repeats of three independent assays.

Animal studies

Nude mice (6 weeks old) were purchased from Vital River Laboratory Animal Technology (Beijing, China). All animals were used following institutional guidelines. Animal studies were approved by the Institutional Animal Care and Use Committee of Binzhou Medical University Hospital (approval no. 20221014-099). HSC2 cells (2 × 10⁶) with lentivirus-mediated PA2G4-P42 or P48 overexpression were prepared in single cell suspension in 100 μl volume (PBS: matrigel = 1:1) and were injected into the subcutaneous tissue of mice using a 27-gauge needle. After 40 days, the mice were sacrificed by CO₂ inhalation. The tumors were dissected for weighing and immunohistochemistry (IHC) staining.

Immunoprecipitation (IP) and Co-IP

IP was conducted using rabbit anti-PA2G4 (15348-1-AP, Proteintech). Rabbit IgG was used as the negative control. Lysate from HSC2 cells was prepared using RIPA lysis buffer (Beyotime, Shanghai, China) following the standard protocol. Protein G sepharose bead slurry was added to the mixture to capture the immunocomplex, by gently rocking overnight at 4 °C. Then, the immunocomplex was collected by centrifugation, washed and subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). The following Liquid chromatography-mass spectrometry (LC-MS)/MS assay was conducted to analyze the protein samples immunoprecipitated by anti-PA2G4 or the IgG control, using the methods introduced previously.¹⁶ For co-IP assays, the presence of MCTS1 in the anti-PA2G4 precipitated samples were detected by anti-MCTS1 (1:1000, GTX50967, GeneTex, Irvine, CA, USA).

To explore the interaction between myc-MCTS1 and flag-P42 or flag-P48, lysate from HSC2 or SCC25 cells with selective overexpression of myc-MCTS1, flag-P42 and flag-P48 were prepared using RIPA lysis buffer (Beyotime), after 10 μM MG132 treatment for 6 h. IP was conducted using mouse anti-Myc tag (60003-2-Ig, Proteintech) or rabbit anti-Flag tag (20543-1-AP, Proteintech). Flag-tagged protein in the anti-Myc precipitated samples and Myc-tagged protein in the anti-Flag precipitated samples were then detected by western blotting.

Molecular docking

Molecular docking between PA2G4 and MCTS1 was performed using HDOCK SERVER (http://hdock.phys.hust.edu.cn/),¹⁷ using protein structural files obtained from RCSB Protein Data Bank (RCSB PDB). The PDB IDs were 6WM8 for PA2G4 and 5ONS for MCTS1.¹⁸

Immunofluorescent staining

Immunofluorescent staining was conducted following the protocol introduced previously.¹⁹ Briefly, HSC2 and SCC25 cells were grown on coverslips, fixed, permeabilized and blocked. Then, the cells were incubated with primary rabbit anti-PA2G4 (1:200, 15348-1-AP, Proteintech) and mouse anti-MCTS1 (1:200, GTX50967, GeneTex) overnight at 4 °C. Then, the cells were washed and incubated with goat anti-rabbit IgG H&L (Alexa Fluor 488) and goat anti-mouse IgG H&L (Alexa Fluor 647) for 1 h at room temperature. Nuclei were stained with DAPI in SlowFade Gold Antifade Mountant (S36938, ThermoFisher Scientific, Carlsbad, CA, USA). Immunofluorescent images were captured using a fluorescence microscope (IX83, Olympus, Tokyo, Japan).

Quantitative reverse transcription PCR (qRT-PCR) assays

RNA extraction and the subsequent qRT-PCR analysis were conducted as described previously.¹⁵ Relative gene expression was calculated using the 2^−ΔΔCt method by normalizing to the expression of ACTB. All the primers used in this study are shown in Table 1.

Table 1.

The primers used in this study.

Gene name	Forward primer sequence (5′–3′)	Reversed primer sequence (5′–3′)
ACTB	TCGTACCACTGGCATTGTGAT	CGAAGTCTAGGGCAACATAG
MCTS1	ATGCTCTATGTGTTGGAGTCATGA	CAAGTGCATTCCTTCTGAGGCTC
PA2G4	GCTCACCTTTGTGCTGAAGCTG	GCTGCTTCAACTGGTGTGACAG

ACTB: Actin Beta; MCTS1: malignant T-cell amplified sequence 1; PA2G4: proliferation-associated 2G4.

Statistical analysis

Data integration and statistical analysis were conducted using GraphPad Prism 8.10 (GraphPad Software, San Diego, CA, USA). Data were reported as mean ± SD based on at least three technical repeats of three independent studies. Unpaired Welch's t-test and analysis of Variance (ANOVA) with Tukey correction for multiple comparisons was performed for two-group and multiple-group comparison respectively. P < 0.05 was considered statistically significant.

Results

PA2G4-P48 but not PA2G4-P42 promotes head and neck squamous cell carcinoma proliferation

PA2G4 has been characterized as an oncogene in oral tumorigenesis.⁶ However, it has complex alternative splicing profile. Based on RNA-seq data from TCGA-HNSCC, we checked the expression of two dominant protein-coding transcripts of PA2G4 (Fig. 1A, top) between the tumor (n = 518) and tumor-adjacent normal (n = 44) groups. Results showed that ENST00000303305 was significantly upregulated in the tumor group (p < 0.001) (Fig. 1A, bottom). However, no significant difference was observed in ENST00000552766 expression (Fig. 1A). High ENST00000303305 expression is associated with significantly shorter OS compared to the low expression group (Fig. 1B). ENST00000303305 could be alternatively translated into two isoforms: PA2G4-P48 and PA2G4-P42 (Fig. 1C). PA2G4-P48 and PA2G4-P42 elicits different functions in multiple cancers.² Therefore, we decided to explore whether these two protein isoforms exert different regulatory effects in HNSCC. HSC2 and SCC25 with relatively low expression of PA2G4 in all HNSCC cell lines (Supplementary Table 1, data from Cancer Cell Line Encyclopedia) were subjected to PA2G4-P48 or PA2G4-P42 overexpression (Fig. 1D). Western blotting data showed that PA2G4-P48 overexpression elevated the expression of multiple tumor-growth related proteins, including cyclin E, c-Myc and Aurora-A (Fig. 1D). However, PA2G4-P42 overexpression suppressed their expression (Fig. 1D). CCK-8 and colony formation assays confirmed that PA2G4-P48 overexpression promoted cell proliferation. In contrast, PA2G4-P42 overexpression significantly suppressed cell proliferation (Fig. 1E and F). By performing in vivo tumor growth analysis, we observed that PA2G4-P48 overexpression enhanced tumor growth in vivo, but PA2G4-P42 overexpression slowed the growth (Fig. 1G–I).

Figure 1.

PA2G4-P48 but not PA2G4-P42 promotes HNSCC proliferation.

A. Exonic structure of the two dominant protein-coding transcripts (ENST00000303305 and ENST00000552766) of PA2G4 (top panel) and their expression between the tumor (n = 518) and tumor-adjacent normal (n = 44) groups (bottom panel), using RNA-seq data from TCGA-HNSCC. B. In patients with primary HNSCC in TCGA, Kaplan-Meier plots were generated to compare the differences in OS between high and low PA2G4 expression groups. Log-rank test was conducted to compare the difference. C. The alternative translational initiation of ENST00000303305. D. Western blotting assays were performed to detect the expression of cyclin E, c-Myc and Aurora-A in HSC2 and SCC25 cells, 48 h after lentivirus-mediated PA2G4-P42 (P42) or PA2G4-P48 (P48) overexpression. E-F. CCK-8 (E) and colony formation (F) assays were performed to compare the relative cell proliferation in HSC2 and SCC25 cells with PA2G4-P42 (P42) or PA2G4-P48 (P48) overexpression. G-I. HSC2-derived xenograft tumors (G) and weight (H), with PA2G4-P42 (P42) or PA2G4-P48 (P48) overexpression. I. Representative images of Ki-67 expression (by IHC staining) in HSC2-derived xenograft tumors panel G. Scale bar: 100 μm. ∗∗, P < 0.01; ∗∗∗, P < 0.001.

Malignant T-cell amplified sequence 1 interacts with PA2G4-P48 but not PA2G4-P42

To identify the binding proteins of PA2G4, we performed IP-SDS-PAGE-LS-MS/MS, using cellular lysate from HSC2 cells. Then, the potential candidates were compared with the evidence-based interactors identified in BioGRID (https://thebiogrid.org/) (Fig. 2A). Malignant T-cell amplified sequence 1 (MCTS1) was a candidate identified by both methods (Fig. 2A). Immunofluorescent staining confirmed the co-localization of MCTS1 and PA2G4 in HSC2 and SCC25 cells (Fig. 2B). To validate the binding, we performed co-IP assays using anti-PA2G4 (Commercial MCTS1 antibody suitable for IP assay is not available). Results confirmed that MCTS1 could be precipitated by anti-PA2G4 in both HSC2 and SCC25 cells (Fig. 2D).

Figure 2.

MCTS1 interacts with PA2G4-P48 but not PA2G4-P42.

A. A work chart showing the strategy to identify proteins interacting with PA2G4. B. Immunofluorescent staining was performed to visualize the distribution and expression of PA2G4 (green) and MCTS1 (red) in HSC2 and SCC25 cells. Scale bar: 10 μm. C. Molecular docking model (left) between MCTS1 (rainbow) and PA2G4 (yellow), using HDOCK SERVER (http://hdock.phys.hust.edu.cn/). The amino acid residues of PA2G4 in the interacting interfaces (right). D. Co-IP assays were conducted using anti-PA2G4 in the lysates from HSC2 and SCC25 cells. E. Co-IP assays were conducted using lysates from HSC2 cells with lentivirus-mediated flag-PA2G4-P42 and myc-MCTS1, or flag-PA2G4-P48 and myc-MCTS1 overexpression.

Molecular docking analysis revealed that some amino acid residues within the PA2G4-P48 specific N terminal region (ASN21, THR22, TYR33, ALA45, ILE46, PRO47) were critical for the binding between PA2G4 and MCTS1 (Fig. 2C). Therefore, we hypothesized that there might be different binding capabilities of MCTS1 with PA2G4-P48 and PA2G4-P42. To validate this hypothesis, we generated flag-tagged PA2G4-P48 (flag-P48), PA2G4-P42 (flag-P42) and myc-tagged MCTS1 (myc-MCTS1), respectively. HSC2 cells were subjected to lentivirus-mediated flag-PA2G4-P42 and myc-MCTS1 or flag-PA2G4-P48 and myc-MCTS1 overexpression (Fig. 2E). Then, co-IP assays were performed using the lysates from these cells. IP assays were performed using anti-flag or anti-myc antibodies, respectively. Co-IP assays indicated that the myc-tagged MCTS1 could only be detected in the immunoprecipitants from the flag-PA2G4-P48 group (Fig. 2E). In addition, only flag-PA2G4-P48 but not flag-PA2G4-P42 was detected in the immunoprecipitants prepared by anti-myc (Fig. 2E).

Malignant T-cell amplified sequence 1 interacts with PA2G4-P48 and reduces its proteasomal degradation

MCTS1 has been characterized as an oncoprotein in laryngeal squamous cell carcinoma, with stabilizing effect on La Ribonucleoprotein 7 (LARP7).¹³ Therefore, we decided to investigate whether the interaction between MCTS1 and PA2G4-P48 contributes to PA2G4-P48 stabilization. MCTS1 is usually upregulated in HNSCC,²⁰ thus we generated MCTS1 shRNAs. PA2G4 transcription was unchanged in HSC2 and SCC25 cells with MCTS1 knockdown (Fig. 3A and B). However, PA2G4-P48 expression at the protein level was decreased after MCTS1 knockdown (Fig. 3C–F). When MG132 (a proteasome inhibitor) was added, MCTS1 knockdown could not induce PA2G4-P48 downregulation (Fig. 3C–F). These findings indicated that MCTS1 might mediate PA2G4-P48 stability by reducing its proteasomal degradation. To validate this hypothesis, we performed cycloheximide (CHX) chase assays. Results showed that MCTS1 overexpression significantly elevated the half-life of PA2G4-P48, while its knockdown drastically reduced the protein half-life compared with the control cells (Fig. 3G-L). In addition, through in vitro ubiquitination assays using HA-tagged wild-type ubiquitin (HA-Ub), we showed that MCTS1 overexpression significantly decreased the poly-ubiquitination of exogenous flag-P48. In comparison, MCTS1 knockdown drastically reduced the poly-ubiquitination level (Fig. 3M − N).

Figure 3.

MCTS1 interacts with PA2G4-P48 and reduces its proteasomal degradation.

A-B. QRT-PCR analysis was performed to detect the expression of MCTS1 (A) and PA2G4 (B) mRNA in HSC2 and SCC25 cells 48 h after lentivirus-mediated MCTS1 knockdown. C–F. Western blotting assays were conducted to detect the expression of MCTS1 and PA2G4 at the protein level in HSC2 and SCC25 cells, 48 h after lentivirus-mediated MCTS1 knockdown. For MG132 treatment group, MG132 (10 μM) was added 10 h before cell collection and lysis. The protein levels were quantified by densitometry using ImageJ (C-D, n = 3). G-L. HSC2 (G, I and J) and SCC25 (H, K and L) cells were co-transfected with flag-tagged P48 in combination with myc-MCTS1 or shMCTS1, followed by treatment with 50 μg/mL CHX for 0, 4, 8 and 12 h (h). Western blotting shows the degradation of flag-P48. The protein levels were quantified by densitometry using ImageJ (n = 3) (I–L). M-N. HSC2 (M) and SCC25 (N) cells were infected with indicated selective combination of flag-P48, HA-tagged Ub (HA-Ub), MCTS1 shRNA or overexpression lentiviruses for 48 h. MG132 (10 μM) was added 10 h before cell collection and lysis. Then, immunoprecipitation was conducted with flag antibody followed by immunoblotting with HA antibody. All results are repeated at least three times. ∗P < 0.05; ∗∗P < 0.01; ∗∗∗P < 0.001; ns: not significant.

Malignant T-cell amplified sequence 1 promotes head and neck squamous cell carcinoma proliferation via stabilizing PA2G4-P48

Since MCTS1 showed oncogenic properties in HNSCC,¹³ and we confirmed that MCTS1 could stabilize PA2G4-P48, we further explored whether PA2G4-P48 is a downstream effector of MCTS1. HSC2 and SCC25 cells were subjected to lentivirus-mediated knocking down of PA2G4-P48 (Fig. 4A). CCK-8 and colony formation assays showed that MCTS1 overexpression promoted cell proliferation and colony formation. However, these alterations were hampered by PA2G4-P48 knocking down (Fig. 4B–C and E). Western blotting results showed that MCTS1 overexpression also elevated the expression of cyclin E, c-Myc and Aurora-A, but these trends were reversed by knocking down of PA2G4-P48 (Fig. 4D).

Figure 4.

MCTS1 promotes head and neck squamous cell carcinoma proliferation via stabilizing PA2G4-P48.

A. QRT-PCR analysis was performed to detect the expression of PA2G4 mRNA in HSC2 cells, 48 h after lentivirus-mediated PA2G4-P48 knockdown. B–C and E. CCK-8 (B) and colony formation (C and E) assays were performed to compare the relative cell proliferation in HSC2 and SCC25 cells with MCTS1 overexpression alone or in combination with PA2G4-P48 knockdown. D. Western blotting assays were performed to detect the expression of cyclin E, c-Myc and Aurora-A in HSC2 and SCC25 cells, 48 h after lentivirus-mediated MCTS1 overexpression alone or in combination with PA2G4-P48 knockdown. ∗P < 0.05; ∗∗P < 0.01; ∗∗∗P < 0.001; ns: not significant.

Discussion

The tumor-promoting role of PA2G4-P48 in oral squamous cell carcinoma was observed in one previous study. Mei et al. showed that PA2G4-P48 could bind to the promoter region of podoplanin (encoded by the PDPN gene) and enhance its transcription.⁶ Activation of the PA2G4-P48-podoplanin axis can promote anchorage-dependent growth, invasion, and wound healing of oral squamous cell carcinoma.⁶ In consistent with these findings, we showed that PA2G4-P48 exerts growth-promoting effects in representative HNSCC. In addition, we observed that PA2G4-P42 presents growth-inhibiting effects in HNSCC. However, these two isoforms have distinct expression profiles. PA2G4-P48 is the dominant isoform, with substantially higher expression than PA2G4-P42 in tumors.²

The differences in the N terminal region between PA2G4-P48 and PA2G4-P42 might be responsible for different interactions with different binding partners.² For instance, PA2G4-P42, but not PA2G4-P48 can bind to ERBB3.²¹ The N-terminal 54-residues of PA2G4-P48 can mask the ERBB3-binding motif on the PA2G4-P42 isoform, thereby preventing the binding.²¹ Both PA2G4-P48 and PA2G4-P42 can be degraded through the ubiquitin-mediate proteasomal pathway.^5,12 The WD40 domain of FBXW7 can bind to the N-terminal tail of PA2G4-P48 (specifically Ser40 and Ser44), leading to its ubiquitination and proteasomal degradation.⁵ Therefore, we hypothesized that the expressional difference might be related to the different stabilities of the isoforms. In addition, there might be some interacting proteins modulating their stability. To validate this hypothesis, we conducted LC-MS/MS and co-IP assays. Our data indicated that MCTS1 interacts with PA2G4-P48, but not PA2G4-P42.

MCTS1 has been identified as an oncogenic protein in multiple cancers. In lung adenocarcinoma, it promotes cell proliferation and migration, and inhibits apoptosis of the cancer cells via elevating E2F transcription factor 1 (E2F1) expression and activating the c-Myc signaling pathway.²² It binds to twinfilin actin binding protein 1 (TWF1) and cooperatively enhances cyclin D1 and c-Myc translation in luminal A/B breast cancer cells.²³ In HNSCC, MCTS1 upregulation might serve as a potential prognostic biomarker for unfavorable overall survival and recurrence-free survival.²⁰ In laryngeal squamous cell carcinoma (a subtype of HNSCC), MCTS1 interacts with LARP7, increases LARP7 protein half-life and reduces its poly-ubiquitination.¹³ Via interacting with density-regulated reinitiation and release factor (DENR), MCTS1-DENR complex can promote translation reinitiation to drive the expression of multiple oncogenes, such as activating transcription factor 4 (ATFF), a-Raf, c-Raf, and CDK4.²⁴ These findings indicated the oncogenic properties of MCTS1 might be related to its interaction with other proteins. In this study, we observed that MCTS1 can enhance the stability of PA2G4-P48 and reduce its poly-ubiquitination.

Since the oncogenic properties of MCTS1 have been characterized in multiple cancers, we checked whether PA2G4-P48 is a downstream functional effector of MCTS1 in HNSCC. PA2G4-P48 specific shRNA was applied in HNSCC cells with MCTS1 overexpression. Our data suggested that PA2G4-P48 specific shRNA significantly impaired MCTS1 overexpression-induced cell proliferation. These findings confirmed that MCTS1 promotes HNSCC proliferation via stabilizing PA2G4-P48.

Based on the findings of this study, we speculate that the minor structural difference between PA2G4-P42 and PA2G4-P48 might be an intriguing site for developing therapeutic reagents (such as small molecules) in the future. The truncation of different domains in AA1-54 may render the functional dissection of oncogenic elements in PA2G4-P48, including the MCTS1 binding. Mutation of specific key predicted resides in PA2G4-P48 might even alter the confirmation of MCTS1 binding and reduce PA2G4-P48 stability. Future studies are supposed to be performed to identify key resides required for the binding between MCTS1 and PA2G4-P48. In addition, selective inhibition of PA2G4-P42 by repressing its specific interacting protein partner, such as MCTS1, might be a potential drug discovery approach. Drugs selectively targeting PA2G4-P48 or MCTS1 would have the potential therapeutic value in HNSCC, without interrupting the tumor-suppressive role of PA2G4-P42.

In summary, this study revealed that PA2G4-P42 and PA2G4-P48 exert growth-suppressive and growth-promoting effects in HNSCC, respectively. MCTS1 can interact with PA2G4-P48 and prolong its half-life by reducing its poly-ubiquitination.

Conflicts of interest

The authors have no conflicts of interest relevant to this article.

Appendix A. Supplementary data

The following is the Supplementary data to this article: