Oncogene https://doi.org/10.1038/s41388-018-0280-6
ARTICLE
NUDT21 negatively regulates PSMB2 and CXXC5 by alternative polyadenylation and contributes to hepatocellular carcinoma suppression Sheng Tan1 Hua Li2 Weijie Zhang1 Yunying Shao1 Yuan Liu1 Haiyang Guan1 Jun Wu2 Yani Kang2 Junsong Zhao1 Qing Yu1 Yunzhao Gu2 Keshuo Ding3 Min Zhang1 Wenchang Qian1 Yong Zhu1 Huayong Cai1 Changyu Chen4 Peter E. Lobie5 Xiaodong Zhao6 Jielin Sun6 Tao Zhu1 ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
1234567890();,:
1234567890();,:
Received: 8 September 2017 / Revised: 29 March 2018 / Accepted: 30 March 2018 © Macmillan Publishers Limited, part of Springer Nature 2018
Abstract Alternative polyadenylation (APA) is an important post-transcriptional regulatory mechanism and involved in many diseases, including cancer. CFIm25, a subunit of the cleavage factor I encoded by NUDT21, is required for 3′RNA cleavage and polyadenylation. Although it has been recently reported to be involved in glioblastoma tumor suppression, its roles and the underlying functional mechanism remain unclear in other types of cancer. In this study, we characterized NUDT21 in hepatocellular carcinoma (HCC). Reduced expression of NUDT21 was observed in HCC tissue compared to adjacent nontumorous compartment. HCC patients with lower NUDT21 expression have shorter overall and disease-free survival times than those with higher NUDT21 expression after surgery. Knockdown of NUDT21 promotes HCC cell proliferation, metastasis, and tumorigenesis, whereas forced expression of NUDT21 exhibits the opposite effects. We then performed global APA site profiling analysis in HCC cells and identified considerable number of genes with shortened 3′UTRs upon the modulation of NUDT21 expression. In particular, we further characterized the NUDT21-regulated genes PSMB2 and CXXC5. We found NUDT21 knockdown increases usage of the proximal polyadenylation site in the PSMB2 and CXXC5 3′ UTRs, resulting in marked increase in the expression of PSMB2 and CXXC5. Moreover, knockdown of PSMB2 or CXXC5 suppresses HCC cell proliferation and invasion. Taken together, our study demonstrated that NUDT21 inhibits HCC proliferation, metastasis and tumorigenesis, at least in part, by suppressing PSMB2 and CXXC5, and thereby provided a new insight into understanding the connection of HCC suppression and APA machinery.
Introduction
Electronic supplementary material The online version of this article (https://doi.org/10.1038/s41388-018-0280-6) contains supplementary material, which is available to authorized users.
Nearly all mammalian messenger RNAs (mRNAs) are featured by cleavage and polyadenylation at their 3′ ends [1]. The polyadenylation site (PAS) defines the length of 3′ untranslated region (UTRs), in which the regulatory
* Xiaodong Zhao
[email protected]
3
* Jielin Sun
[email protected]
Department of Pathology, Anhui Medical University, Meishan Road, Hefei, Anhui 230031, China
4
First Affiliated Hospital of Anhui Medical University, Anhui Medical University, Meishan Road, Hefei, Anhui 230031, China
5
Tsinghua-Berkeley Shenzhen Institute, Precision Medicine & Healthcare Research Center, Tsinghua University, Shenzhen 518055, China
6
Key Laboratory of Systems Biomedicine (Ministry of Education), Shanghai Center for Systems Biomedicine, Shanghai Jiao Tong University, 800 Dongchuan Road, Shanghai 200240, China
These authors contributed equally: Sheng Tan, Hua Li, Weijie Zhang.
* Tao Zhu
[email protected] 1
2
Hefei National Laboratory for Physical Sciences at Microscale, the CAS Key Laboratory of Innate Immunity and Chronic Disease, School of Life Sciences, University of Science and Technology of China, Hefei, Anhui 230027, China School of Biomedical Engineering, Bio-ID Center, Shanghai Jiao
Tong University, Shanghai 200240, China
S. Tan et al.
elements of microRNAs (miRNA) or RNA-binding proteins are located. Approximately 70% of mammalian genes contain alternative polyadenylation (APA) sites, and thus potentially encode mRNA isoforms with various 3′UTRs [1]. The 3′UTR APA influence the mRNA cellular localization, stability and translation. Recent transcriptomic analyses have indicated that APA is modulated under various physiological or pathological circumstances [2, 3]. In particular, some fast-growing cell populations (including cancer cells) prefer to utilize proximal PASs and thereby generate mRNA isoforms with shorter 3′UTRs, while the quiescent or differentiated cell populations favor distal PAS to generate mRNA isoforms with longer 3′UTRs [4–7]. Moreover, some studies demonstrated that APA can be tumor specific and the specific APA patterns can be used as biomarkers for cancer diagnosis and prognosis [8–12]. These observations implicate APA is functionally involved in development and disease. The length of 3′UTR is primarily determined by the 3′ processing complex that contains ~85 proteins [13]. Although highly diverse, at the core machinery are four multisubunit protein complexes responsible for cleavage and polyadenylation: CPSF, CstF, CFI-II and poly(A) polymerase [1]. These core complexes make different contribution to the APA. Previous studies has demonstrated that CFIm25 and CFIm68, two members of CFI complex, play an important role in the extensive distal-proximal poly (A) site switching [14–19]. Nevertheless, how these key 3′ processing factors regulate APA and their biological consequence in development or disease remain elusive. The initiation and progression of cancer have long been considered as a result of the accumulation of genetic mutation, by which oncogenes are activated and/or tumor suppressor genes are inactivated. Recent studies demonstrated that APA activates oncogenes without genetic mutations and facilitates cancer transformation. Indeed, there is widespread 3′UTR shortening caused by APA in a wide variety of cancers. One of key regulators, NUDT21, is involved in glioblastoma tumor suppression and neuropsychiatric disease [14, 20]. It remains unclear that whether it exerts the anti-tumor effect on other types of cancer. Therefore, additional investigations are required to understand whether this critical 3′ processing factor possesses a universal role in tumorigenesis. To further understand how NUDT21 affect APA and subsequent cancer progression, in this study we performed a comprehensive characterization of NUDT21 in hepatocellular carcinoma (HCC) cells. Similar to the phenomena in glioblastoma, we found NUDT21 exhibited tumor suppressive activities in HCC. Moreover, we observed that NUDT21 modulated widespread 3′UTR alterations. In particular, we demonstrated that PSMB2 and CXXC5 contribute to NUDT21-mediated tumor suppression in HCC cells.
Together, our study sheds new insights into the link between the 3′ processing factor and hepatoma tumor suppression.
Results NUDT21 is downregulated in HCC and low NUDT21 expression correlates with poor prognosis in HCC patients Firstly, we examined the mRNA levels of many factors driving APA in HCC (n = 59) and adjacent non-tumor tissues (n = 59) and found NUDT21 (CFIm25) mRNA levels were decreased in HCC versus paired adjacent nontumorous tissues using quantitative real-time PCR (Fig. 1a). In addition, we compared the expression of NUDT21 in 135 paired HCC and surrounding non-tumorous tissue samples using immunohistochemical (IHC) assays (Supplementary Table 2). NUDT21 staining was scored as a weak or moderate expression in 69 and 20% of tumor tissues, as compared with 19 and 38% in corresponding adjacent non-tumorous liver tissues (Fig. 1b, c). The patients with low NUDT21 expression (n = 64) had shorter overall survival (OS) and disease-free survival (DFS) than the patients who has high NUDT21 expression (n = 53) (Fig. 1d, e). To further explore the clinical implications of the decreased expression of NUDT21, we examined its correlation with the clinicopathologic features of HCC. The reduced expression of NUDT21 was observed to be associated with the tumor size (p = 0.002), advanced tumornode metastasis (TNM) stage (p = 0.031), and distant metastasis (p = 0.004) in HCC (Supplementary Table 2). We also analyzed a publicly available human clinic data set from The Cancer Genome Atlas (TCGA), and about 40% (146/364) of HCC samples were found to have copy number loss in NUDT21 locus. As a consequence, NUDT21 mRNA levels were significantly decreased in the HCC clinical samples (Fig. 1f). Kaplan–Meier survival curve revealed that patients with NUDT21 deletion (n = 58) had significantly worse OS compared with that of patients with NUDT21 non-deletion (n = 107) (Fig. 1g). However, the somatic loss of NUDT21 did not seem to significantly correlate with the DFS of HCC patients (Fig. 1h). Together, our data suggested that the expression of NUDT21 is reduced due to the somatic loss of gene in HCC. Low NUDT21 expression correlates poor clinical outcomes in the HCC patients.
NUDT21 suppresses cell growth and local invasion in HCC cells We then screened the protein expression of NUDT21 in non-tumorous liver cell (HL-7702 an LO2) and human
NUDT21 negatively regulates PSMB2 and CXXC5 by alternative polyadenylation and contributes to. . .
S. Tan et al. Fig. 1 Somatic loss of NUDT21 predicts poor prognosis in HCC. a qRT-PCR assessing mRNA levels of many factors driving APA in HCC (n = 59) and adjacent non-tumor tissues (n = 59). b Expression of NUDT21 in HCC and surrounding non-tumorous tissues were detected by IHC. c Bar graph shows statistics for IHC staining intensity in training. d Kaplan–Meier overall survival curve of HCC patients. e Kaplan–Meier disease-free survival curve of HCC patients. f Analysis of NUDT21 mRNA in TCGA liver cancer samples with deleted (n = 146) or non-deleted (n = 218) NUDT21 locus. g Kaplan–Meier analysis of overall survival in liver cancer patients stratified according to their NUDT21 locus status in the TCGA liver cancer cohort. h Kaplan–Meier analysis of disease free survival in liver cancer patients stratified according to their NUDT21 locus status in the TCGA liver cancer cohort. *p < 0.05. **p < 0.001
HCC cell lines (HepG2, PLC/PRF/5, SMMC-7721, Hep3B, BEL-7404 and QGY-7703). We found that the protein level of NUDT21 decreased in several HCC cell lines compared to that in non-tumorous liver cell lines (HL-7702 and LO2) (Fig. 2a). To explore the function of NUDT21 in HCC, we depleted NUDT21 with shRNAs in HL-7702 and PLC/PRF/ 5 cells, and ectopically expressed NUDT21 in Hep3B and SMMC-7721 cells. The levels of NUDT21 in these cell lines were verified by western blotting (Fig. 2b, c). Compared to control cells, both HL-7702-shNUDT21 and PLC/ PRF/5-shNUDT21 cells showed significant increases in cell proliferation and viability, and generated more and larger colonies in colony formation assay (Fig. 2d–f). In contrast, forced expression of NUDT21 in Hep3B and SMMC-7721 cells significantly reduced cell proliferation, viability and clonogenicity in vitro (Fig. 2g–i). To further assess the role of NUDT21 in vivo, we injected the stably transfected cells into the flanks of nude mice. The HL-7702-shNUDT21 cells-derived tumors grew at a significantly faster rate, resulting in more than two folds increase in tumor volume at day 32 as compared to tumors from HL-7702-shCtrl cells (Fig. 2j). In contrast, the Hep3BNUDT21 cells derived tumors grew at a significantly slower rate, resulting in a 60% reduction in tumor volume as compared to Hep3B-vec cells at day 32 (Fig. 2l). The difference with the tumor grow in these cells was also confirmed by the weight of harvested tumor tissues at the end point (Fig. 2k, m). In addition, Ki-67 staining revealed that tumors derived from HL-7702-shNUDT21 cells and Hep3B-vec cells exhibited significantly higher percentage of proliferative cells compared to tumors from HL-7702shCtrl cells and Hep3B-NUDT21 cells, respectively (Fig. 2n, o).
Overexpression of NUDT21 suppresses migration and metastasis of HCC cells Given the connection between lower levels of NUDT21 with advanced TNM stages in HCC patients, we further assessed the possible roles of NUDT21 in modulating HCC
cell migration and invasion. As shown in fig. 3a–e, knockdown of NUDT21 in HL-7702 cells significantly increased, while the forced expression of NUDT21 in Hep3B cells decreased the cell migration and invasion. By contrast, forced expression of NUDT21 resulted in the retarded wound closing in Hep3B cells (Fig. 3f). The results indicated that NUDT21 suppresses the migration and invasion of HCC cells in vitro. Furthermore, hematoxylin–eosin (HE) staining of the tumor sections revealed that the tumors derived from HL7702-shNUDT21 cells and Hep3B-vec cells were poorly encapsulated and exhibited significant local invasion into adjacent tissues, whereas those from HL-7702-shCtrl cells and Hep3B-NUDT21 cells were well encapsulated and noninvasive (Fig. 3g,h). To extend our observations in vitro, we explored the effects of NUDT21 on the metastatic capacity of HCC cell in vivo. The cells (HL-7702-shCtrl, HL-7702-shNUDT21, Hep3B-pEZ-vec and Hep3B-pEZ-NUDT21) were injected into the tail vein of nude mice and the metastatic nodules in lungs were detected. The mice injected with HL-7702 shNUDT21 cells showed a significantly increased the frequency of lung metastasis after 2 months inoculation (Fig. 3i). On the other hand, the frequency of lung metastasis was decreased in the mice injected with Hep3B-NUDT21 cells as compared to those injected with Hep3B-vec cells (Fig. 3j). Therefore, the in vivo results further demonstrate the critical role of NUDT21 in HCC metastasis.
NUDT21 modulates the widespread 3′UTR alteration in HCC cells Give the functional involvement of NUDT21 in APA formation and its tumor suppressive activities mentioned above, we hypothesized that NUDT21 acts as a tumor suppressor in HCC, at least in part, by influencing APA and 3′UTR. To test this hypothesis, we examined all instances of NDUT21-related 3′UTR alteration in HCC cells. Technically, we identified poly(A) sites in NUDT21 overexpressed Hep3B cells, NUDT21 knockout HL-7702 cells and the respective control cells with 3T-seq we reported previously [21]. A total of 46.8 million reads were generated, of which 68.9% were uniquely mapped to the reference genome. After filtering internal priming events, we found nearly 80% of qualified reads were mapped to the annotated transcription terminal sites (TTSs) or 3′UTRs (Fig. 4a and Supplementary Fig. 1). These reads were used to delineate poly(A) sites with the strategy described previously [21], yielding 87,894 poly(A) sites. We observed that the majority of identified poly(A) sites located at the 3′ terminus (Fig. 4b). In particular, 30.8% of these poly(A) sites were mapped to UCSC TTS and 42.6%
NUDT21 negatively regulates PSMB2 and CXXC5 by alternative polyadenylation and contributes to. . .
to the 3′UTR regions. 30% of these poly(A) sites have been annotated in the UCSC poly(A) database, and thus the remaining sites are putatively novel. Moreover, 30% of genes in 3T-seq harbor three or more poly(A) sites (Fig. 4c and Supplementary Fig. 2). Based on the poly(A) data, we identified genes that exhibited significantly different APA profiles upon NUDT21 alteration in HCC cells. 135 genes with altered 3′UTR were identified in NUDT21 knockout
HL-7702 cells, and 59.3% showed a shift of distal-toproximal APA usage and thus possessed 3′UTRs shortening (Fig. 4d and Supplementary table 3). Meanwhile, we applied the same approach to characterize the APA profiles in Hep3B cells overexpressing NUDT21, and identified 1328 genes with 3′UTRs shortening in the context of low expression of NUDT21 (Fig. 4d and Supplementary table 4).
S. Tan et al. Fig. 2 NUDT21 suppresses proliferation and local invasion in HCC cells. a Western blotting analysis of NUDT21 expression in normal liver and HCC cell lines. b The protein levels of NUDT21 in HL-7702 and PLC/PRF/5 cells stably transfected with shRNA plasmids were assessed by western blotting. c The protein levels of NUDT21 in Hep3B and SMMC-7721 cells stably transfected with NUDT21 overexpressing plasmids were assessed by western blotting. d Proliferation curves of HL-7702 and PLC/PRF/5 cells stably transfected with shRNAs targeting NUDT21 or a scrambled shRNA are shown. e The total cell viability of HL-7702 and PLC/PRF/5 cells stably transfected with shRNAs targeting NUDT21 or a scrambled shRNA are determined by MTT assay. f Colony formation assays of HL-7702 and PLC/PRF/5 cells stably transfected with shRNAs targeting NUDT21 or a scrambled shRNA are shown. g Proliferation curves of Hep3B and SMMC-7721 cells stably transfected with NUDT21 overexpressing plasmid or an empty control vector are shown. h The total cell viability of Hep3B and SMMC-7721 cells stably transfected with NUDT21 over-expressing plasmid or an empty control vector are determined by MTT assay. i Colony formation assays of Hep3B and SMMC-7721 cells stably transfected with NUDT21 over-expressing plasmid or an empty control vector are shown. j, l Male immunedeficient BALB/c nu/nu (nude) mice were inoculated subcutaneously with HCC cells. The tumor growth curves were determined by measuring the tumor volume (HL-7702, n = 6/group; Hep3B, n = 8/ group). k, m The weight of tumors at the end point is shown. n, o Ki67 staining were performed on the respective HCC cells-derived tumor sections. *p < 0.05. **p < 0.001
To determine whether and how NUDT21-modulated APA contributes to tumor suppression, we went on to examine the functional consequence of NUDT21 responsive genes. We focused on the genes with 3′UTR alterations simultaneously influenced in both NUDT21 knockdown and overexpression cells, which yielded 22 candidates (Supplementary table 5). Then we chose the top five based on their p values and performed initial functional screening. We found modulation of the expression of PSMB2 and CXXC5 obviously affects cellular activities in HCC cells (Supplementary Fig. 3). The 3′UTRs of the PSMB2 and CXXC5 genes contains several alternative PASs (Fig. 5a). To validate that PSMB2 and CXXC5 are subject to NUDT21-mediated APA, 3′RACE was performed (Figs. 5b) and 3′-UTR isoforms with different length were identified in HL-7702 and Hep3B cells. These DNA products were subjected to Sanger sequencing and both the short isoforms of PSMB2 and CXXC5 3′UTR were found containing only one proximal PAS (Supplementary table 6). Of note, the proximal site PAS1 seems to be the most frequently used PAS and NUDT21 drives the utilization shift from PAS1 to other distal PASs (Fig. 5c, d). To further support our observation in 3′RACE, qRTPCR was performed. Six different primer pairs were showed in Fig. 5e, in which primer pairs 1–3 cover both long and short isoforms and primer pairs 4–6 amplify long isoform. The results indicated that the total PSMB2 or CXXC5 mRNA expression significantly increased in the NUDT21 knockdown HL-7702 cells as compared to HL7702-shCtrl cells and the expression of the long isoform of
PSMB2 or CXXC5 mRNA showed no significant difference upon NUDT21 knockdown (Fig. 5f, h). Consistently, forced expression of NUDT21 in Hep3B cells led to a significant decrease of the expression of total PSMB2 or CXXC5 mRNA, as well as a mild increase of the expression of long isoforms (Fig. 5g, i). Collectively, these data indicate that NUDT21 modulates the switch of short and long isoforms and the total abundance of PSMB2 and CXXC5 mRNAs. Recent study had shown that the shortening of 3′-UTRs enables genes to escape the repression of miRNA, which leads to higher protein expression. Interestingly, most of the predicted miRNA binding sites by TargetScan were localized to the region between the PAS1 and the distal poly(A) site on PSMB2 and CXXC5 3′UTR (Supplementary Fig.4). To investigate if the isoforms of 3′UTR differentially affect the production of proteins, the different 3′-UTR fragments were cloned into the luciferase reporter vector. As expected, the reporter plasmids containing short 3’-UTR of PSMB2 or CXXC5 showed significantly higher luciferase activities as compared to those with the long 3′-UTR of PSMB2 or CXXC5 (Fig. 5j, k). Furthermore, knockdown of NUDT21 increased the protein levels of PSMB2 and CXXC5 in HL7702 cells, while forced expression of NUDT21 reduced the protein levels of PSMB2 and CXXC5 in Hep3B cells (Fig. 5l, m). Together, these data suggested that the expression of PSMB2 and CXXC5 are modulated by NUDT21 through APA in HCC.
PSMB2 and CXXC5 promotes cell growth, migration and invasion in HCC cells Since the functional of PSMB2 and CXXC5 in HCC development hasn’t been reported to date, we transfected shRNAs targeting PSMB2, CXXC5, or control shRNA into Hep3B cells. The efficacy of these shRNAs was verified by western blotting (Fig. 6a). The silencing of either PSMB2 or CXXC5 led to a significant decrease in the proliferation and clonogenicity of Hep3B cells (Fig. 6b, d, f). Transwell assay and would healing assay showed knockdown of PSMB2 or CXXC5 significantly suppressed the migration and invasion of Hep3B cells in vitro (Fig. 6g–i). Cell line specificity was excluded by demonstration of similar effects in SMMC 7721 cells (Fig. 6c, e–i), suggesting that PSMB2 and CXXC5 are essential for the growth and motility of HCC cells. We went on to investigate whether NUDT21 acts as tumor suppressor genes through PSMB2 and CXXC5 in HCC. To this end, PSMB2 or CXXC5 knockdown was performed in HL-7702-shNUDT21 cells, in which their protein levels are elevated. As we mentioned above, knockdown of NUDT21 increased the proliferation, total cell viability, colony formation, and in vitro migration and invasion capacity of HL-7702 cells, whereas knockdown of
NUDT21 negatively regulates PSMB2 and CXXC5 by alternative polyadenylation and contributes to. . .
Fig. 3 NUDT21 suppresses metastasis in HCC cell lines. a, b Transwell migration and c, d invasion assays. Representative images were shown. e, f In vitro wound healing assays of HCC cells. g, h Representative H&E staining of respective HCC cells-derived primary tumors. Red arrows indicate invasive tumor front. Results are shown
as mean ± SD. i, j Representative pictures of H&E staining of lungs and incidence of lung metastasis from mice inoculated with HCC cells. Black arrows indicate the lung metastases. Results are shown as mean ± SD. *p < 0.05. **p < 0.001
PSMB2 or CXXC5 genes fully abrogated the phenotypes induced by silencing of NUDT21 (Fig. 7a–d, f–h). The same effects were observed in PLC/PRF/5 cells, indicating the central role for PSMB2 and CXXC5 in NUDT21-mediated cell growth, migration and invasion inhibition of HCC cells.
of HCC progression are poorly characterized. APA has been increasingly recognized to be involved in tumorigenisis. The APA-mediated 3′UTR alterations are regulated by various members of cleavage and polyadenylation machinery (1). In this study we examined several members and found NUDT21 (CFIm25) is most down-regulated in HCC when compared with the adjacent non-tumorous samples (Fig. 1a). The causal relationship between NUDT21-modulated APA and cancer cell proliferation has been demonstrated in glioblastoma tumor [14]. Here, we reported the somatic loss of NUDT21 locus and expression
Discussion HCC is the third leading cause of cancer-related deaths worldwide [22–24]. However, the underlying mechanisms
S. Tan et al. Fig. 4 Characteristics of 3T-seq data. a Genomic locations of qualified 3T-seq reads mapped to the reference genome. b Genomic distribution of the poly (A) sites. c The statistics of genes with various number of detected poly(A) sites. d Scatterplot of CULI in NUDT21 KD and control cells in which genes with shortened (n = 2209) or lengthened (n = 1344) 3′ UTRs upon NUDT21 knockdown (false discovery rate (FDR) = 0.05)
in HCC and their connection to poorer clinical outcomes of HCC patients. Systemic analysis, including both in vitro and in vivo study, demonstrated NUDT21 is a tumor suppressor gene in HCC. The recent study of NUDT21 in glioblastoma tumor also identified a set of NUDT21-regulated genes with shortened 3′UTRs [14]. Because of the heterogeneity among different tissues, we could envision that the functional genes modulated by NDUT21 in HCC may distinguish from those in glioblastoma tumor. To test our hypothesis, we firstly examined the transcriptome-wide APA events in HCC cells with an approach we developed previously [21]. We identified a considerable number of genes with altered 3′UTRs upon the perturbation of NUDT21 expression (Supplementary table 3-5). The data suggested NUDT21 modulates the extensive APA in HCC cells.
A direct consequence of APA is the generation of mRNA isoforms with different 3′UTR, and the isoforms with shortened 3′UTR could be activated probably due to the relief of microRNA-mediated repression. To understand how NUDT21 exert its tumor suppressive function by APA modulation in HCC cells, we further characterized NUDT21regulated genes with shortened 3′UTRs. PSMB2 and CXXC5 were demonstrated to be involved in NUDT21-mediated tumor suppression. PSMB2 is a subunit of the multicatalytic proteinase complex which contain a highly ordered ringshaped 20S core structure [25]. Although proteasomes have been suggested as the therapeutic targets in multiple cancers, the role of PSMB2 is largely unappreciated. Meanwhile, CXXC5 is reported to be overexpressed in solid tumors and an unfavorable prognostic factor in breast cancer [26], but few work has been done illustrating its role in HCC. In this
NUDT21 negatively regulates PSMB2 and CXXC5 by alternative polyadenylation and contributes to. . .
Fig. 5 NUDT21 modulates 3′-UTR shortening of PSMB2 and CXXC5 genes. a Schematic illustration of PSMB2 and CXXC5 mRNAs. Red boxes show protein coding region; black line represents untranslated regions. Positions of the PASs are indicated by vertical lines. b Schematic representation of the primer pairs used in 3′RACE analyses. c, d Validation of the long and short 3′-UTR transcripts by 3′RACE. e Schematic representation of the primer pairs used in mRNA expression analyses. f, h qRT–PCR analyses of the expression of PSMB2 and CXXC5 mRNA upon NUDT21 knock-down in HL-7702 cell was
performed using the primers levels and a slight increase in the long isoform designed in E. g, i qRT–PCR analyses of the expression of PSMB2 and CXXC5 mRNA upon NUDT21 overexpression in Hep3B cell was performed using the primers designed in E. j, k The activity of the isoform of PSMB2 and CXXC5 3′UTR was examined using luciferase reporter assay. l,m The protein level of NUDT21, PSMB2 and CXXC5 was assessed by western blotting. Results are shown as mean ± SD. *p < 0.05. **p < 0.001
study we demonstrated that NUDT21 knockdown results in the utilization of the proximal poly(A) site and ends up with the shortening of 3′UTR in PSMB2 and CXXC5 genes, which consequently leads to the elevated expression of PSMB2 and CXXC5. Moreover, we explore the role of
PSMB2 and CXXC5 in modulating the tumor suppressive function of NUDT21 in HCC. This novel axis, NUDT21APA-PSMB2/CXXC5, may shed new insights into the understanding the link between cancer progression and APA machinery and provide novel therapeutic targets for HCC.
S. Tan et al.
Fig. 6 Silencing of PSMB2 or CXXC5 inhibits the HCC cell growth, migration and invasion properties in vitro. a The efficacy of shRNA targeting PSMB2 or CXXC5 was assessed by western blotting. b, c Proliferation curves of cells transfected with control shRNA, PSMB2 or CXXC5 shRNA. d, e Colony formation assays of cells transfected with control shRNA, PSMB2 or CXXC5 shRNA. f MTT assays of cells
transfected with control shRNA, PSMB2 or CXXC5 shRNA. g Migration assays, h Invasion assays and i In vitro wound healing assays of Hep3B and SMMC-7721 cells transfected with control shRNA, PSMB2 or CXXC5 shRNA. Results are shown as mean ± SD. *p < 0.05. **p < 0.001
Materials and methods
streptomycin (50 mg/ml), penicillin (50 U/ml), and glutamine (2 mM).
Cells culture Patients and HCC specimens All cells were purchased from the Shanghai Cell Bank, (Shanghai, China). All cells were cultured in RPMI-1640 or DMEM medium with 10% FBS (Gibco),
Human liver tissue samples were collected from the First Affiliated Hospital of Anhui Medical University (Hefei,
NUDT21 negatively regulates PSMB2 and CXXC5 by alternative polyadenylation and contributes to. . .
Anhui, People’s Republic of China). The protocol to use these tissues was approved by the Biomedical Ethics Committee of Anhui Medical University with informed
consent of all the involved patients. The tissues for RNA extraction were frozen immediately after resection and kept in RNAlater (Invitrogen), and total RNA were extracted
S. Tan et al. Fig. 7 Co-knockdown of NUDT21 plus PSMB2 or NUDT21 plus CXXC5 rescues the phenotype induced by NUDT21 knockdown alone. a, b Proliferation curves of a HL-7702 and b PLC/PRF/5 cells transfected with the control shRNA, NUDT21 shRNA, NUDT21 plus PSMB2 or NUDT21 plus CXXC5 shRNA. c MTT assays of HL-7702 and PLC/PRF/5 cells transfected with the control shRNA, NUDT21 shRNA, NUDT21 plus PSMB2 or NUDT21 plus CXXC5 shRNA. d, e Colony formation assays of HL-7702 and PLC/PRF/5 cells transfected with the control shRNA, NUDT21 shRNA, NUDT21 plus PSMB2 or NUDT21 plus CXXC5 shRNA. f Migration assays, g Invasion assays and h in vitro wound healing assays of HL-7702 and PLC/PRF/5 cells transfected with the control shRNA, NUDT21 shRNA, NUDT21 plus PSMB2 or NUDT21 plus CXXC5 shRNA. Results are shown as mean ± SD. *p < 0.05. **p < 0.001
using Trizol ((Invitrogen) following the manufacturer’s instruction. The immunohistochemistry assay was performed following the manufacturer’s instruction, and the stained sections were reviewed and scored by two independent pathologists.
qRT-PCR Total RNA was isolated using Trizol (Life Technologies) and reverse transcribed with RevertAid™ First Strand cDNA Synthesis Kit (Life Technologies). The qPCR was performed using gene-specific primer pairs and SYBR Green.
3′RACE assay cDNA was synthesized using Superscript III reverse transcriptase (Invitrogen). The first PCR was done using PSMB2-Fa or CXXC5-Fa forward primer and Outer as reverse primer. The PCR product was cleaned using PCR clean-up Kit (Sangon, Shanghai), which was used as template DNA in nested PCR. Nested PCR was done with a nested PSMB2-Fb or CXXC5-Fb forward primer and inner as reverse primer. The DNA from the PCR was cloned in pUCm-T Vector (Sangon, Shanghai) for Sanger sequencing. Primer pairs were shown in Supplementary table 1.
Western blotting Western blot analysis was carried out as described previously [27]. The antibodies against NUDT21 (1:1000, Abcam), PSMB2 (1:1000, Proteintech), CXXC5 (1:500, Proteintech) or β-ACTIN (1:5000, BD Bioscience) were used.
MTT assay The cellular viability was examined using the MTT assay (Sigma). Cells (1 × 103 cells per well) were plated into 96well plates. After 5 days, 20 µL of MTT solution (5 mg/ml) was added into each well. After 1–2 h incubation, the media was removed and cells was rinsed by PBS. 100 µL DMSO was added and the absorbance at 570 nm wavelength was examined by a 96-well plate reader.
Migration and invasion assays For transwell assays, cells were trypsinized and resuspended in culture medium without FBS. For migration assay, 5–10 × 104 normal liver cells or HCC cells were added to the top 8-µm chamber without Matrigel. For invasion assays, 10–20 × 104 cells were added to the Matrigel-coated upper chamber. The lower chambers were filled with medium containing 10–20% serum. After 24–48 h incubation, the inserts were rinsed by PBS and stained with 0.1% crystal violet solution for 5 min. The noninvading cells were then wiped from the inner surface of the inserts with cotton swabs. Images were taken by the Olympus IX-70 microscope.
Histopathological analysis Briefly, mouse lung tissues were perfused with 4% paraformaldehyde, fixed overnight, dehydrated in ethanol series and stained with HE (Sigma). Ki-67 staining on tumor sections was carried out by using Ki67 antibody (MA514520, Thermo Scientific).
Generation of stable cell lines Luciferase assays The 3′-UTRs of the human PSMB2 and CXXC5 were cloned into the dual-luciferase expression vector, psiCHECK2 vector (Promega). HCC cells were plated in 24well plates. Cells of 40–50% confluence were transfected using Lipofectamine 3000 (Invitrogen). Cell extracts were prepared after 48hs posttransfection with 100 ng plasmid and the luciferase activity was examined by the Dual Luciferase Assay System (Promega). Three independent experiments were performed.
Specific gene knockdowns were achieved by transduction with shRNAs into HCC cells. The sequence of shRNAs and corresponding scramble control are listed: shNUDT21-1,CCGGCAGTGTAGAATAAATGTGGTACTCGAGTACCACATTTATTCTACACTGTTTTT; shNUDT21-2,CCGGCAGCGCATGAGG GAAGAATTTCTCGAGAAATTCTTCCCTCATGCGCTGTTTTT; scramble control, CCGGCAACAAGATGAAGAGCA CCAACTCGAGTTGGTGCTCTTCATCTTGTTGTTTTT.
NUDT21 negatively regulates PSMB2 and CXXC5 by alternative polyadenylation and contributes to. . .
For NUDT21 over-expression, HCC cells were transfected with either empty control or NUDT21-expressing pEZMO2 plasmids. 2 mg/ml G418 selection was performed over a period of 3 weeks and NUDT21 protein was examined by western blot
3T-seq analysis 3T-seq analysis was performed as we described previously [21]. Fifty micrograms total RNA was used for cDNA synthesis and the resulting dsDNA was fragmentated with fragmentase (NEB). The fragments in 3′ terminal were released by Gsu I digestion. Then the products were subjected to deep sequencing. The qualified raw sequencing reads were aligned to the human reference genome (hg19) with bowite using the default parameters. The poly(A) sites were constructed by iteratively clustering the mapping results. The linear trend alternative test was employed to identify the significant 3′UTR switching for each gene (the FDR-adjusted p-value < 0.05). The accession number of raw data is E-MTAB-6368, which are available in the ArrayExpress database (http://www.ebi.ac.uk/arrayexpress).
Xenograft model The design and protocol of in vivo experiments were approved by the Institutional Animal Care and Use Committee, University of Science and Technology of China (USTCACUC1601008). Five weeks old male nude mice were used in all xenograft assays. A total of 5 × 106 tumor cells were mixed with same volumes of matrigel, and injected subcutaneously into the flanks or tail vein of the animals. The volumes of tumors were continuously measured after 2 weeks. Acknowledgements This work was supported by The National Key Scientific Programme of China (2016YFC1302305), The National Natural Science Foundation of China (81672609, 31671299, 81502282, 81472494,) and the Shenzhen Development and Reform Commission Subject Construction Project [2017] 1434. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Compliance with ethical standards Conflict of interest The authors declare that they have no conflict of interest.
References 1. Di Giammartino DC, Nishida K, Manley JL. Mechanisms and consequences of alternative polyadenylation. Mol Cell. 2011;43:853–66. 2. Tian B, Manley JL. Alternative polyadenylation of mRNA precursors. Nat Rev Mol Cell Biol. 2017;18:18–30.
3. Elkon R, Ugalde AP, Agami R. Alternative cleavage and polyadenylation: extent, regulation and function. Nat Rev Genet. 2013;14:496–506. 4. Mayr C, Bartel DP. Widespread shortening of 3’UTRs by alternative cleavage and polyadenylation activates oncogenes in cancer cells. Cell. 2009;138:673–84. 5. Sandberg R, Neilson JR, Sarma A, Sharp PA, Burge CB. Proliferating cells express mRNAs with shortened 3’ untranslated regions and fewer microRNA target sites. Science. 2008;320:1643–7. 6. Morris AR, Bos A, Diosdado B, Rooijers K, Elkon R, Bolijn AS, et al. Alternative cleavage and polyadenylation during colorectal cancer development. Clin Cancer Res. 2012;18:5256–66. 7. Erson-Bensan AE, Can T. Alternative polyadenylation: another foe in cancer. Mol Cancer Res. 2016;14:507–17. 8. Singh P, Alley TL, Wright SM, Kamdar S, Schott W, Wilpan RY, et al. Global changes in processing of mRNA 3’ untranslated regions characterize clinically distinct cancer subtypes. Cancer Res. 2009;69:9422–30. 9. Fu Y, Sun Y, Li Y, Li J, Rao X, Chen C, et al. Differential genome-wide profiling of tandem 3’ UTRs among human breast cancer and normal cells by high-throughput sequencing. Genome Res. 2011;21:741–7. 10. Xia Z, Donehower LA, Cooper TA, Neilson JR, Wheeler DA, Wagner EJ, et al. Dynamic analyses of alternative polyadenylation from RNA-seq reveal a 3’-UTR landscape across seven tumour types. Nat Commun. 2014;5:5274. 11. Lembo A, Di Cunto F, Provero P. Shortening of 3’UTRs correlates with poor prognosis in breast and lung cancer. PLoS One. 2012;7:e31129. 12. Miles WO, Lembo A, Volorio A, Brachtel E, Tian B, Sgroi D, et al. Alternative polyadenylation in triple-negative breast tumors allows NRAS and c-JUN to bypass PUMILIO posttranscriptional regulation. Cancer Res. 2016;76:7231–41. 13. Shi Y, Di Giammartino DC, Taylor D, Sarkeshik A, Rice WJ, Yates JR 3rd, et al. Molecular architecture of the human premRNA 3’ processing complex. Mol Cell. 2009;33:365–76. 14. Masamha CP, Xia Z, Yang J, Albrecht TR, Li M, Shyu AB, et al. CFIm25 links alternative polyadenylation to glioblastoma tumour suppression. Nature. 2014;510:412–6. 15. Martin G, Gruber AR, Keller W, Zavolan M. Genome-wide analysis of pre-mRNA 3’ end processing reveals a decisive role of human cleavage factor I in the regulation of 3’ UTR length. Cell Rep. 2012;1:753–63. 16. Brown KM, Gilmartin GM. A mechanism for the regulation of pre-mRNA 3’ processing by human cleavage factor Im. Mol Cell. 2003;12:1467–76. 17. Venkataraman K, Brown KM, Gilmartin GM. Analysis of a noncanonical poly(A) site reveals a tripartite mechanism for vertebrate poly(A) site recognition. Genes Dev. 2005; 19:1315–27. 18. Coseno M, Martin G, Berger C, Gilmartin G, Keller W, Doublie S. Crystal structure of the 25 kDa subunit of human cleavage factor Im. Nucleic Acids Res. 2008;36:3474–83. 19. Yang Q, Gilmartin GM, Doublie S. Structural basis of UGUA recognition by the Nudix protein CFI(m)25 and implications for a regulatory role in mRNA 3’ processing. Proc Natl Acad Sci USA. 2010;107:10062–7. 20. Gennarino VA, Alcott CE, Chen CA, Chaudhury A, Gillentine MA, Rosenfeld JA, et al. NUDT21-spanning CNVs lead to neuropsychiatric disease and altered MeCP2 abundance via alternative polyadenylation. Elife 2015;4;e10782. 21. Lai DP, Tan S, Kang YN, Wu J, Ooi HS, Chen J, et al. Genomewide profiling of polyadenylation sites reveals a link between selective polyadenylation and cancer metastasis. Hum Mol Genet. 2015;24:3410–7.
S. Tan et al. 22. Forner A, Llovet JM, Bruix J. Hepatocellular carcinoma. Lancet. 2012;379:1245–55. 23. Llovet JM, Zucman-Rossi J, Pikarsky E, Sangro B, Schwartz M, Sherman M, et al. Hepatocellular carcinoma. Nat Rev Dis Prim. 2016;2:16018. 24. Llovet JM. Liver cancer: time to evolve trial design after everolimus failure. Nat Rev Clin Oncol. 2014;11:506–7. 25. Collavoli A, Comelli L, Cervelli T, Galli A. The over-expression of the beta2 catalytic subunit of the proteasome decreases homologous recombination and impairs DNA double-strand
break repair in human cells. J Biomed Biotechnol. 2011; 2011:757960. 26. Knappskog S, Myklebust LM, Busch C, Aloysius T, Varhaug JE, Lonning PE, et al. RINF (CXXC5) is overexpressed in solid tumors and is an unfavorable prognostic factor in breast cancer. Ann Oncol. 2011;22:2208–15. 27. Tan S, Ding K, Li R, Zhang W, Li G, Kong X, et al. Identification of miR-26 as a key mediator of estrogen stimulated cell proliferation by targeting CHD1, GREB1 and KPNA2. Breast Cancer Res. 2014;16:R40.
