High expression of PSMD7 is associated with poor PC prognosis

To investigate the implications of PSMD7 in PC, we first assayed PSMD7 expression. Data obtained from Gene Expression Profiling Interactive Analysis (GEPIA) (http://gepia.cancer-pku.cn/) revealed that the expression of PSMD7 was higher in PC tissues (T) than in non-tumour tissues (NT) (Fig. 1A and Supplementary Fig. 1A). Remarkably, PSMD7 expression was significantly associated with both overall survival (p = 0.0095, Fig. 1B) and disease-free rates (p = 0.0027, Fig. 1C). GSEA also indicated that the dysregulation of gene expression in PC was linked to PSMD7 (Fig. 1D). Furthermore, ROC curve analysis showed that PSMD7 had high sensitivity and specificity in the diagnosis of PC (AUC = 0.9742, 95% CI: 0.956–0.991, p < 0.01) (Supplementary Fig. 1B). Subsequently, we analyzed 70 PC specimens newly collected in recent years by qRT-PCR and Western blotting (Fig. 1E, F). The results showed that the expression of PSMD7 protein and mRNA in PC tissues was higher than that in adjacent tissues. For further evaluation of the association of PSMD7 expression with the clinicopathologic characteristics of PC, IHC assay was conducted on 104 human PC specimens to examine PSMD7 expression (Fig. 1G). The IHC results showed that the expression of PSMD7 in PC tissues was significantly higher than that in adjacent tissues.

Fig. 1

High expression of PSMD7 is associated with poor prognosis in PC. (A) Expression of PSMD7 in pancreatic non-tumour tissues (n = 171) and PC patient specimens (n = 179) evaluated via the GEPIA webtool. (B) and (C) Overall survival ( p = 0.0095) and disease-free survival ( p = 0.0027) times of patients with low versus high expression of PSMD7 in PC. (D) GSEA outcomes were graphed to show the association between poorly regulated genes and PSMD7 expression in PC. (E) and (F) Western blotting and qRT-PCR assay of PSMD7 expression in human PC tissues (n = 70), together with neighbouring normal tissues (n = 70). G, Quantification and representative immunohistochemistry (IHC) images of PSMD7 expression in human PC tissues, together with neighbouring normal tissues. (H) and (I) Kaplan–Meier analysis of the overall survival ( p = 0.0117) and disease-free survival ( p = 0.0206) time of PC patients with various PSMD7 expression. (J) and (K) The PSMD7 expression in BxPC-3, AsPC-1, HPDEC, CFPAC-1, PANC-1, and SW 1990 cells analysed via qRT-PCR and western blotting. *p < 0.05, **p < 0.01

Clinicopathological correlation analysis revealed a correlation between increased PSMD7 expression, TNM stage, and tumour size (Table 1, Supplyment Table 2). Kaplan-Meier analysis revealed a significant correlation between elevated PSMD7 levels and reduced overall survival as well as disease-free survival (Fig. 1H, I). Furthermore, PSMD7 was highly expressed in five PC cell lines (SW 1990, PANC-1, CFPAC-1, AsPC-1, and BxPC-3) compared to in normal pancreatic cells (HPDE6-C7) (Fig. 1J, K). These findings demonstrate that PSMD7 is commonly elevated in PC and is significantly associated with an unfavorable prognosis in individuals afflicted with this disease.

PSMD7 facilitates PC cell growth in vitro and in vivo

To understand the impact of increased PSMD7 expression on PC carcinogenesis, we transfected shPSMD7#1, shPSMD7#2,shPSMD7#3 and shNC interference plasmids into PANC-1 cells, and the qRT-PCR and Western blotting results showed that shPSMD7#2 and shPSMD7#3 interference plasmids had obvious effect on knockout (Supplymentment Fig. 1C, D). Then, a in-of-function test was performed by decreasing PSMD7 expression in the AsPC-1 and PANC-1 cell lines and upregulating PSMD7 expression in the SW 1990 cell line (Figs. 2A and B and 3A and B ). The role of PSMD7 in PC cell proliferation was investigated using EdU and CCK-8 assays. As presented in Figs. 2C and D and 3C-E, in PC cells with PSMD7 knockdown, the cell proliferation capacity was obviously lowered, whereas in PC cells with high PSMD7 expression, the cell proliferation capacity was obviously enhanced. These findings suggest that PSMD7 plays a pivotal role in facilitating PC cell proliferation in vitro.

Fig. 2

Kowndown of PSMD7 expression restrains the PC cell proliferative capacity in vitro and in vivo. (A) The mRNA of PSMD7 in AsPC-1 and PANC-1 cells transfected with shPSMD7 or shNC assayed via qRT-PCR. (B) PSMD7 knockdown in AsPC-1 and PANC-1 cells was validated via western blotting. (C) The proliferative abilities of AsPC-1 and PANC-1 PC cells with shNC or shPSMD7 treatment was detected separately via CCK-8. (D) Proliferative capacities of AsPC-1 and PANC-1 cells that were transfected with shPSMD7 or shNC were assayed with EdU, and the scale bar represents 50 μm, **p < 0.01. (E) Evaluation of mice that were injected with luciferase-expressing PANC-1/shPSMD7 or PANC-1/shNC cells via the IVIS imaging system. (F) Tumour volumes of the PANC-1/shPSMD7 and PANC-1/shNC groups. Volumes of tumours are expressed as the mean ± SD, n = 6. (G) Tumour weight of the shPSMD7 and shNC groups. n = 6. (H) The images of tumors the PANC-1/shPSMD7 and PANC-1/shNC groups. (I) and (J) Immunohistochemistry (IHC) analysis of PSMD7 and Ki-67 expression in tumours of the PANC-1/shPSMD7 and PANC-1/shNC groups, scale bar denotes 100 μm. n = 3, *p < 0.05, **p < 0.01

Fig. 3

Overexpression of PSMD7 promotes PC proliferation. (A) and (B) PSMD7 mRNA along with protein levels in SW 1990 cells transfected with Vector or Flag-PSMD7 plasmids were detected by qRT-PCR as well as western blotting. Tubulin was employed as the loading control. (C) CCK-8 assay was employed to examine the difference in cell viability between both groups. (D) and (E) EdU staining was utilized to establish the impact of PSMD7 overexpression on the PC cell proliferation rate. Scale bar denotes 50 μm, **p < 0.01. (F) Mice injected with luciferase-expressing SW 1990/Flag-PSMD7 or SW 1990/Vector cells were investigated through IVIS imaging system. (G) Tumour volumes for SW 1990/Flag-PSMD7 or SW 1990/Vector groups. Volumes of tumours are expressed as mean ± SD, n = 6, **p < 0.01. (H) Tumour weights in the Flag-PSMD7 and Vector groups, n = 6, **p < 0.01. (I) Tumour volumes for SW 1990/Flag-PSMD7 or SW 1990/Vector groups. (J) and (K), immunohistochemistry (IHC) analysis of Ki-67 and PSMD7 expression in tumours of the SW 1990/Flag-PSMD7 or SW 1990/Vector group, scale bar is 20 μm. n = 3, *p < 0.05, **p < 0.01

To investigate the impact of PSMD7 on in vivo tumor growth, we injected PANC-1 and SW 1990 cells stably transfected with PSMD7-interfering lentivirus or PSMD7-overexpressing lentivirus into the subcutis of both legs of nude mice, respectively. The results of in vivo experiments demonstrated that the suppression of PSMD7 led to a significant inhibition of tumour growth (Fig. 2E). Moreover, the PSMD7-knockout group exhibited significantly reduced tumour weight and volume compared to the control group (Fig. 2F, G and H). IHC analysis demonstrated that the knockdown of PSMD7 resulted in a reduction in the number of Ki-67-positive cells within the tumor (Fig. 2I, J). However, upregulation of PSMD7 increased tumour growth (Fig. 3F-K). The data presented herein provide compelling evidence for the pivotal oncogenic role of PSMD7 in promoting the growth of PC cells.

Notch1 pathway as a PSMD7 downstream component facilitates the role of PSMD7 in PC cells

To detect the latent molecular mechanisms of PSMD7 action, we first carried out GSEA in the TCGA database to investigate potential associations between PSMD7 and a variety of signalling pathways. As shown in Fig. 4A, in PC samples with high levels of PSMD7, the gene set Hallmark_Notch1_Targets was markedly enriched, indicating a strong association between the Notch1 pathway and elevated levels of PSMD7 in PC. We assumed that the Notch1 signalling pathway was activated by PSMD7 because the former had the highest enrichment score. To verify this hypothesis, western blotting and qRT-PCR analyses were conducted, and silencing PSMD7 was found to repress the protein and mRNA levels of the Notch1 target gene HES1 (Fig. 4B, D). Conversely, overexpression of PSMD7 resulted in enhanced protein and mRNA levels of Notch1, together with its target gene, HES1 (Fig. 4C, E). To further validate the link between NOTCH1 and PSMD7 signaling, dual luciferase assays in SW 1990 and PANC-1 cells showed that the transcriptional activity of NOTCH1 was repressed by PSMD7 silencing, whereas that of NOTCH1 was activated by PSMD7 up-regulation (Fig. 4F, G). Collectively, these data demonstrate that PSMD7 may function by augmenting the Notch1 signalling pathway.

Fig. 4

Notch1 pathway, downstream component of PSMD7, facilitates the role of PSMD7 in PC cells. (A) GSEA outcomes were graphed to show the association between genes linked to the Notch1 signalling pathway and PSMD7 expression. (B) and (C) The HES1 and Notch1 mRNA expression levels were assayed by qRT-PCR with PSMD7 overexpression in SW 1990 cells or knockdown in PANC-1 cells. (D) and (E) Protein levels of Notch1 along with its target genes (including HES1) were measured via western blotting after PSMD7 overexpression in SW 1990 cells or silencing in PANC-1 cells. (F) and (G) Luciferase activities were examined after transfection of sh-PSMD7 and the corresponding luciferase reporter plasmids in PSMD7-overexpressing SW 1990 cells or PANC-1 cells. (H) Expression of HES1, Notch1, and PSMD7 in PC cells that were co-transfected with Notch1 and sh-PSMD7. (I) and (J) EdU and CCK-8 analysis of cell proliferation in PANC-1 cells that expressed shPSMD7 or shNC, with or without overexpression of Notch1. (K) Expression of HES1, Notch1, and PSMD7 in PC cells that were co-transfected with sh-Notch1 and Flag-PSMD7. (L) and (M) EdU and CCK-8 analysis of cell proliferation in SW 1990 cells that expressed exogenous PSMD7 or control vector, with or without shNotch1 transfection. (N) Mice that were injected with luciferase-expressing PANC-1/Notch1, PANC-1/shPSMD7, PANC-1/shNC, or PANC-1/shPSMD7 + Notch1 cells were evaluated via IVIS imaging system. (O) Tumour volumes in the PANC-1/Notch1, PANC-1/shPSMD7, PANC-1/shNC, or PANC-1/shPSMD7 + Notch1 groups. Volumes of tumours are represented as mean ± SD, n = 6; (P) Tumour weight in the PANC-1/Notch1, PANC-1/shPSMD7, PANC-1/shNC, or PANC-1/shPSMD7 + Notch1 group. n = 6. *p < 0.05, **p < 0.01

To investigate the involvement of the Notch1 signaling pathway in the oncogenic role of PSMD7 in PC, the plasmid for Notch1 overexpression was first transfected into PSMD7 knockdown cells, and subsequently, HES1, Notch1, and PSMD7 protein expression was analysed by western blotting. The proliferative capacity of the cells was evaluated using EdU and CCK-8 assays. Our data demonstrated that PSMD7 downregulation suppressed HES1 and Notch1 expression, whereas Notch1 up-regulation attenuated the decrease in HES1 and Notch1 expression resulting from the knockdown of PSMD7 (Fig. 4H). CCK-8 and EdU assays revealed that knockdown of PSMD7 repressed proliferation, whereas overexpression of the Notch1 gene alleviated the reduction in cell proliferation resulting from PSMD7 knockdown (Fig. 4I, J). Conversely, PSMD7 overexpression markedly enhanced Notch1 expression, whereas knockdown of Notch1 remarkably inhibited this increase in PSMD7-induced Notch1 expression (Fig. 4K). Additionally, knockdown of Notch1 attenuated cell proliferation resulting from overexpression of PSMD7 (Fig. 4L, M). These results revealed that Notch1 is a pivotal signalling pathway for the proliferation of PC cells induced by PSMD7. Furthermore, in vivo investigations demonstrated that rescuing the expression of Notch1 attenuated the reduction in tumour growth induced by PSMD7 knockdown (Fig. 4N-P). In conclusion, our findings revealed that Notch1 is regulated by PSMD7 and is a crucial mediator of PC cell growth evoked by PSMD7.

PSMD7 activates Notch1 pathway by modulating SOX2 expression

It has been reported that PSMD7 acts by interacting with various substrates. To further elucidate the mechanism of PSMD7 modulation of Notch1 in PC cells, we initially investigated the potential direct interaction between Notch1 and PSMD7. However, co-immunoprecipitation (co-IP) analysis did not reveal any direct interaction (Fig. 5A, B). Subsequently, to identify the intrinsic mechanism of PSMD7 activation of the Notch1 signalling pathway in PC, large-scale proteomic assays were carried out in PSMD7-silenced PC cells using TMT-based LC-MS/MS analysis. SOX2 protein expression was reduced in PSMD7-silenced PC cells (Fig. 5C). Notably, SOX2 has been reported to be an essential transcription factor for the positive transcriptional regulation of Notch1 [24]. To investigated SOX2 regulates the NOTCH1 signal in PC cells, we used SOX2 knockdown PANC-1 cell lines and found that SOX2 downregulation decreased the mRNA levels of Notch1 and downstream targets Hes1. On the contrary, SOX2 upregulation enhanced the mRNA levels of Notch1 and downstream targets Hes1 (Supplementary Fig. 2A and B). Furthermore, dual-luciferase detection in SW 1990 and PANC-1 cells demonstrated that NOTCH1 transcriptional activity was inhibited by SOX2 silencing, whereas NOTCH1 transcriptional activity was repressed by SOX2 upregulation (Supplementary Fig. 2C and D). These findings led us to speculate that SOX2 could activate Notch1 signalling. As expected, SOX2 protein expression was markedly reduced in PSMD7-silenced PC cells (Fig. 5D). In comparison, the expression of the SOX2 protein increased in PSMD7-overexpressing PC cells, implying that SOX2 activated the Notch1 signalling pathway mediated by PSMD7 (Fig. 5E). To validate the modulation of SOX2 mediated by PSMD7 in clinical PC specimens, SOX2 expression in PC tissues was examined by IHC. IHC staining showed that SOX2 accumulated in PC tissues with high PSMD7 expression (Fig. 5F). To further corroborate the connection between SOX2 and PSMD7, we tested 52 PC specimens newly collected in recent years using western blotting (Fig. 5G-I). Importantly, statistical analyses illustrated that SOX2 protein levels were positively associated with PSMD7 protein levels in PC tissues (Fig. 5J). Collectively, the findings collectively demonstrate that PSMD7 activates the Notch1 pathway by modulating the expression of SOX2.

Fig. 5

PSMD7 activates Notch1 pathway by modulating SOX2 expression

(A) and (B) Co-IP test was implemented to identify the protein binding of Notch1 and PSMD7 in SW 1990 and PANC-1 cells, separately. (C) Heatmap displaying the first 15 most differentially expressed genes in the PANC-1 cells that were transfected with shPSMD7 or shNC. (D) and (E) Protein expression of PSMD7 and SOX2 in shNC or shPSMD7 in PANC-1 cells and exogenous PSMD7 or vector in SW 1990 cells examined via western blotting. (F) Representative plots of immunohistochemistry (IHC) staining for SOX2 and PSMD7 in clinical PC samples. 100× plots with a scale bar of 200 μm and 400× plots with a scale bar of 50 μm. (G-I) western blotting assay of PSMD7 expression in human PC tissues (n = 52), together with neighbouring normal tissues (n = 52). **p < 0.01. (J) Correlations of data from western blotting of PSMD7 expression with respect to the level of SOX2. r = 0.3649, p = 0.0067. *p < 0.05, **p < 0.01.

PSMD7 interacts with SOX2

To identify how PSMD7 modulates SOX2 expression, qRT-PCR was performed to investigate SOX2 mRNA expression in PSMD7-overexpressing and PSMD7-knockdown PC cells. The results revealed that PSMD7 did not affect SOX2 mRNA levels, indicating that PSMD7 may control SOX2 expression at the post-transcriptional level rather than at the transcriptional level (Fig. 6A, B). The protein interactome was further explored via Co-IP mass spectrometry, which revealed that PSMD7 binds to SOX2 (Fig. 6C, D). Western blotting and endogenous IP analyses were subsequently performed to verify that PSMD7 interacted with SOX2 (Fig. 6E, F). Additionally, the co-localization of SOX2 and PSMD7 in pancreatic cancer cells was confirmed by confocal microscopy, providing further substantiation for the protein-protein interaction (Fig. 6G). Docking analysis revealed binding as well as interactions between SOX2 and PSMD7 (Fig. 6H). An array of SOX2 truncated plasmids tagged with haemagglutinin (HA) markers was constructed to characterise the specific structural domains that interact directly with PSMD7. SOX2 mainly contains two domains: the HMG (high mobility group) domain at the N-terminal, which is a DNA binding domain, and the TAD (trans-activation domain) domain at the C-terminal. Based on these structural information, the truncated form of the SOX2 mutant is constructed to determine the domain of its interaction with PSMD7. The subsequent localization experiments showed that the HMG domain of SOX2, the nterminal region of SOX2 (amino acid 41–109), was related to PSMD7 (Fig. 6I-K). Hence, our data indicate that there is an interaction between PSMD7 and SOX2.

Fig. 6

PSMD7 interacts with SOX2. (A) and (B) qRT-PCR analysis of PSMD7 and SOX2 mRNA expression in PANC-1 cells transfected with either shNC or shPSMD7, and SW 1990 cells transfected with exogenous PSMD7 or a control vector. n = 3; ns, not significant. **p < 0.01. (C) The first 5 proteins that co-precipitated with PSMD7 were identified via LC-MS/MS. #PSMs with matching peptide profiles. (D) Mass spectra displaying distinct peptides of PSMD7 characterized by 2D-LC-MS/MS after immunoprecipitation of PANC-1 cell lysates with anti-PSMD7. (E) and (F) Immunoprecipitation of SW 1990 and PANC-1 cell lysates with control IgG, anti-SOX2, or anti-PSMD7 antibodies. (G) SW 1990 and PANC-1 cells were fixed and stained using SOX2 (green) and PSMD7 (red) antibodies. Cell nuclei were stained using DAPI (blue). Scale bar: 20 μm. (H) Top ranked docking conformations and 3D structures of SOX2 and PSMD7. SOX2 and PSMD7 are displayed in cyan and green, respectively. (I) Schematic illustration of SOX2 structure. (J) Diagram of SOX2 truncated mutant constructs. (K) HMG region of SOX2 was required for interaction with PSMD7. *p < 0.05, **p < 0.01

PSMD7 stabilizes SOX2 protein expression via suppressing SOX2 degradation mediated by proteasome

Since PSMD7 is a deubiquitylase, we speculated that PSMD7 might modulate the degradation and ubiquitylation of SOX2 in PC. We found that MG132, a proteasome inhibitor, repressed the degradation of SOX2 (Fig. 7A, B). These findings imply that SOX2 undergoes proteasomal degradation in PC cells. Consequently, we examined whether PSMD7 modulates SOX2 protein stability by influencing proteasomal degradation. As presented in Fig. 7C, PSMD7 knockdown caused a pronounced decrease in SOX2, whereas treatment with MG132 completely blocked the decrease in SOX2 resulting from the knockdown of PSMD7. In contrast, treatment with MG132 inhibited the increase in SOX2 expression induced by the overexpression of PSMD7 (Fig. 7D). Additionally, the influence of PSMD7 on the half-life of SOX2 in CHX-treated PC cells was characterised. The findings suggest that the knockdown of PSMD7 shortened the half-life of the SOX2 protein, whereas the overexpression of PSMD7 prolonged its half-life (Fig. 7E, F). Furthermore, we conducted an in vitro proteasome activity assay and found that the knockdown of PSMD7 decreased the expression activity of proteasome in PANC-1 cells (Fig. 7G). In contrast, the overexpression of PSMD7 increased the expression activity of proteasome in SW 1990 cells (Fig. 7H). Co-IP analysis demonstrated that PSMD7 knockdown enhanced the endogenous SOX2 ubiquitination level, whereas overexpression of PSMD7 lowered it (Fig. 7I, J). Taken together, these findings indicate that PSMD7 represses SOX2 protein degradation, thereby contributing to SOX2 stability.

Fig. 7

PSMD7 stabilizes SOX2 protein expression via suppressing SOX2 degradation mediated by proteasome. (A) and (B) SOX2 protein levels at various times were measured by western blotting after MG132 addition (10 µM) to SW 1990 and PANC-1 cells. (C) and (D) Western blot analysis of SOX2 and PSMD7 protein expression in PANC-1 cells transfected with shPSMD7 or shNC and SW 1990 cells transfected with exogenous PSMD7 or a control vector, with or without treatment with 10 µM MG132. (E) and (F) PANC-1 cells transfected with shPSMD7 or shNC and SW 1990 cells transfected with exogenous PSMD7 or a control vector were subjected to treatment with 20 µg/mL CHX, followed by assessment of SOX2 protein levels using western blotting. n = 3, **p < 0.01. (G) and (H) PANC-1 and SW 1990 cells were treated with shPSMD7 or PSMD7 for 72 h and the intracellular proteasome activity in the treated cells were assessed using proteasome activity fluorometric assay kit. Experiment was repeated three times. A statistically significant difference in the proteasome activity in cells treated with SAHA vs. without SAHA (control) is denoted by *p < 0.05, **p < 0.01. (I) and (J) Lysates of SW 1990 along with PANC-1 cells that were transfected with Flag-PSMD7, sh-PSMD7, and HA- ubiquitin (HA-Ub) were analysed via immunoblotting and then immunoprecipitated using anti-SOX2 and probed with anti-HA. *p < 0.05, **p < 0.01

PSMD7 oncogenesis relies on the SOX2-Notch1 pathway

Finally, rescue experiments were conducted to investigate the dependence of PSMD7 oncogenic effect on SOX2 stabilization in PC. As shown in Fig. 8A-C, the knockdown of PSMD7 dramatically diminished the proliferative capacity of PC cells, while simultaneous SOX2 overexpression diminished this capacity. Likewise, augmentation with SOX2 rescued PSMD7 knockdown-induced PC cell proliferation (Fig. 8D-F). Simultaneously, we injected PANC-1 cells stably expressing sh-NC, sh-PSMD7, SOX2 or sh-PSMD7 + SOX2 into the subcutaneous part of the leg of nude mice, and tumour growth was monitored. As shown in Fig. 8G-I, the tumor volume and weight of mice harboring PSMD7-silenced cells exhibited a remarkable reduction, whereas simultaneous SOX2 overexpression completely abrogated the antitumour effects of PSMD7 knockdown. Consequently, the deubiquitylase PSMD7 activated the Notch1 pathway via altered degradation of SOX2, thereby promoting PC progression (Fig. 8H).

Fig. 8

PSMD7 oncogenesis depends on the SOX2-Notch1 pathway. (A) Expression of HES1, Notch1, and PSMD7 in PC cells that were co-transfected with SOX2 and sh-PSMD7. (B) and (C) EdU and CCK-8 analysis of cell proliferation in the PANC-1 cells that express shPSMD7 or shNC, with or without overexpression of SOX2. (D) Expression of HES1, Notch1, and PSMD7 in PC cells that were co-transfected with sh-SOX2 and Flag-PSMD7. (E) and (F) EdU and CCK-8 analysis of cell proliferation in SW 1990 cells that express exogenous PSMD7 or control vector, with or without the transfection of sh-SOX2. (G) Mice injected with luciferase-expressing PANC-1/SOX2, PANC-1/shPSMD7, PANC-1/shNC, or PANC-1/shPSMD7 + SOX2 cells were examined through IVIS imaging system. (H) Tumour volumes in the PANC-1/SOX2, PANC-1/shPSMD7, PANC-1/shNC, or PANC-1/shPSMD7 + SOX2 group. Volumes of tumours are expressed as the mean ± SD, n = 6, **p < 0.01. (I) Tumour weight in the PANC-1/SOX2, PANC-1/shPSMD7, PANC-1/shNC, or PANC-1/shPSMD7 + SOX2 group. n = 6. **p < 0.01. (J) Suggested mechanistic scheme for PSMD7 to facilitate the Notch1 signalling pathway mediated by SOX2 in PC, r = 0.3649, p = 0.0067. *p < 0.05, **p < 0.01