Background
Pancreatic cancer (PC) is one of the most malignant cancers, and ranks as the fourth leading cause of cancer-related deaths, amounting to 4.5% of all cancer-related deaths globally [
1]. Pancreatic ductal adenocarcinoma (PDAC) is responsible for over 80% of PC cases, with a 5-year survival of approximately 10% [
2]. The poor prognosis of PDAC patients is associated with highly aggressive phenotype and early cancer recurrence and metastasis following surgical treatment. Surgical resection represents the only possibility of a cure for resectable and borderline resectable cases, while advancements in chemotherapy including the FOLFIRINOX regimen and gemcitabine plus nab-paclitaxel has improved the long-term outcomes of patients with metastasis [
3]. Despite advances achieved in PDAC management in recent years, limited breakthroughs in the identification of effective biomarkers or treatment strategies have emerged. A deeper understanding of the molecular mechanism regulating cancer initiation and progression remains urgent to identify early diagnostic and prognostic biomarkers, as well as to discover novel therapeutic targets for PDAC.
Almost all PDAC patients carry at least one of four frequently-mutated driver genes including the oncogene KRAS and the tumor suppressors TP53, SMAD4, and CDKN2A [
4,
5]. As a transcription factor, TP53 encodes the p53 protein that regulates ~ 300 target genes that coordinate diverse cellular effector functions including cell cycle arrest, apoptotic cell death, damaged DNA repair, free radical scavenging, and immune response regulation [
6]. Importantly, p53 suppresses tumor initiation by transcriptionally inducing the critical cyclin-dependent kinase inhibitor p21 and essential pro-apoptotic modulators (PUMA, BAX, NOXA), leading to cell-cycle arrest and apoptosis of abnormal or damaged cells [
7‐
9]. Mutation of TP53 or loss of p53 may induce tumor initiation and tumor growth [
10]. Zinc finger matrin-type 1 (ZMAT1), maps to Xq22.1 and encodes a protein that contains three U1-like zinc fingers of Cys
2His
2(C2H2)-type zinc fingers family similar to those found in the nuclear matrix protein matrin 3. ZMAT1 belongs to a 5-member family (ZMAT1-5) in humans, in which all encoded proteins contain zinc finger domains, but are otherwise dissimilar [
11]. Such associations have been reported between the zinc-finger proteins and cellular stress response pathways including those involved in DNA damage, cell cycle arrest, and apoptosis [
12]. ZMAT3, a transcriptional target of p53, acts as a tumor suppressor by triggering cell cycle arrest and apoptosis [
13]. However, the biological function of ZMAT1 in the context of tumorigenicity and tumor progression is unknown, as is its association with p53.
Herein, we found that ZMAT1 was down-regulated in PDAC and the reduced expression of ZMAT1 was associated with unfavorable clinicopathological characteristics and poor survival of PDAC. The loss of ZMAT1 promoted PDAC tumorigenicity and progression in vitro and in vivo. Moreover, we found ZMAT1 functioned in a p53-dependent manner and identified SIRT3 as an enhanced target and effector of ZMAT1 regulation of p53. The findings indicated a role for ZMAT1-SIRT3-p53 signaling during tumor growth, highlighting that ZMAT1 is a novel prognostic and therapeutic biomarker of PDAC.
Methods
Data acquisition
Gene expression data for 171 PDAC samples were obtained from The Cancer Genome Atlas (TCGA) database up to June 2021 (
https://portal.gdc.cancer.gov/repository). The GSE62165 dataset based on the GPL13667 platform (contains 13 pancreas and 118 PDAC samples), the GSE62452 dataset based on the GPL6244 platform (contains 61 pancreas and 69 PDAC samples), the GSE15471 dataset based on the GPL570 platform (contains 36 pancreas and 36 PDAC samples), and the GSE16515 dataset based on the GPL570 platform (includes 16 pancreas and 36 PDAC samples) were downloaded from the Gene Expression Omnibus (GEO) database for expression validation (
https://www.ncbi.nlm.nih.gov/geo/). Gene Expression Profiling Interactive Analysis (GEPIA,
http://gepia.cancer-pku.cn/index.html) and Oncomine (
https://www.oncomine.org) databases were also used to validate the transcriptional levels of ZMAT1 in PDAC and other cancer types [
14,
15]. Correlation analysis of gene expression was performed using the GEPIA database using TCGA and Genotype-Tissue Expression Project (GTEx) data.
Tissue samples and patient follow-up
For validation studies, tumor and adjacent normal tissue samples were obtained from 122 PDAC patients from the Guangdong Provincial People’s Hospital and the First Affiliated Hospital of Sun Yat-sen University (Validation cohort). All patients were followed to June 2021 and the median follow-up time was 12.3 months (range 3–42.3). Overall survival (OS) was defined as the date from surgical resection to death, while disease-free survival (DFS) was defined as the date from surgical resection to tumor metastasis or recurrence. The clinicopathological data of the enrolled patients, including sex, age, carbohydrate antigen 19–9 (CA19-9), lymph node metastasis, perineural invasion, TNM stage, and differentiation were manually collected. The study was approved by the Ethics Association of Guangdong Provincial People’s Hospital and the First Affiliated Hospital of Sun Yat-sen University. All enrolled patients provided written informed consent before participation. Each tissue sample was evaluated and diagnosed as PDAC by two different pathologists.
Cox regression and Kaplan–Meier analysis
Univariate and multivariate Cox regression analyses were used to identify independent prognostic factors for PDAC prognosis in the validation cohort. Kaplan–Meier analyses and log-rank tests were conducted for the high- and low-ZMAT1 expression groups in both TCGA and validation cohorts to assess the ability to predict patient survival.
Functional and pathway enrichment analysis
Genes co-expressed with ZMAT1 (|Spearman’s correlation coefficient|> 0.5 and
P < 0.05) were screened from the cBioPortal database (
https://www.cbioportal.org/). Co-expressed genes and ZMAT1-binding genes identified by Chromatin immunoprecipitation (ChIP) sequencing were integrated into DAVID 6.8 separately for Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses (
https://david-d.ncifcrf.gov/). GO and KEGG analyses results were visualized using the R “ggplot2” package. Gene set enrichment analysis (GSEA) determined the gene sets and functional pathways that differed significantly between the high- and low-ZMAT1 expression groups. ZMAT1 expression was used as a phenotype label and 1000 gene set permutations were performed per analysis. The STRING database was used to perform functional enrichment analysis and to construct a protein–protein interaction (PPI) network, to subsequently identify P53-associated genes (
https://string-db.org/cgi/) [
16].
Materials
Antibodies against p53 (ab26, 1/1000), p21 (ab109520, 1/1000), CDK2 (ab32147, 1/2000), CCNA2 (ab185619, 1/1000), BAD (ab32445, 1/1000), BAX (ab32503, 1/5000), Bcl-2 (ab32124, 1/1000), SIRT3 (ab217319, 1/1000), Ki-67 (ab16667, 1/200), Flag (ab205606, 1/30), Caspase-3 (ab32351, 1/5000), GAPDH (ab8245, 1/2000) and β-actin (ab8226, 1/5000) were from Abcam (Cambridge, UK). Anti-ZMAT1 (NBP1-81375, 1/100) was from Novus Biologicals (Colorado, USA) and anti-ZMAT1 (bs-4387R, 1/200) was from Bioss (Beijing, China). Tunel apoptosis assay kit (#8109) was from Cell Signaling Technology (Danfoss, USA). Pifithrin-α (PFT-α, S2929) was from Selleck (Texas, USA).
Cell culture and transfection
HPDE6, PANC-1, BxPC-3, Capan-2, SW1990, and AsPC-1 cells lines were obtained from Procell (Wuhan, China). Cells were cultured in RPMI 1640 medium (Gibco, New York, USA), supplemented with 10% fetal bovine serum (FBS) at 37℃ with 5% CO2. PFT-α treatment was administered at a final concentration of 100 ng/ml for 24 h.
For the generation of cell lines with ZMAT1 overexpression or knockdown, the EGFP-tagged/3 × Flag hZMAT1-CMV Puro vector and three siRNAs targeting ZMAT1 were transfected into cells to overexpress and silence ZMAT1, respectively. Cell transfection was performed as previously described [
17]. After antibiotic selection, the over-expression and depletion efficiency were assessed by Real-time quantitative polymerase chain reaction (RT-qPCR) and western blot. Reconstituted cells with stable overexpression of ZMAT1 and cells with depleted ZMAT1 were utilized for cell proliferation assays, cell migration assays, RT-qPCR, immunoblotting, and animal experiments as indicated below. The siRNA and PCR primer oligonucleotide sequences used in our study are shown in Tables S1-S2.
Cell proliferation assays
Cell counting Kit-8 (CCK-8) and colony formation assays were used to determine cell viability. For CCK-8 assays, 1500 cells were seeded per well of a 96-well plate. At a specific timepoint, the CCK-8 solution was added to each well. The cells were then cultured at 37 °C under 5% CO2 and the absorbance (OD450) was assessed in a microplate reader after 0, 24, 48, and 72 h. For colony formation assays, cells of each cell type (2000 cells/ well) were seeded into 6-well plates, gently shaken, and cultured at 37 °C with 5% CO2 for 7–14 days. The medium was subsequently removed and the cells were stained with 0.1% crystal violet to quantify positive colonies (diameter > 30 µm).
Cell migration assays
Transwell plates (Corning Costar, USA) were used for cell migration analysis. A total of 5 × 104 cells were seeded in the upper chambers of Transwell plates in 200 µL of serum-free DMEM while DMEM containing 10% FBS was added to the lower chambers. After incubating for 24 h, migrated cells in the lower chambers were fixed in methanol and then stained with crystal violet. Migrated cells were imaged using an inverted microscope and quantified from three different fields.
Cell migration was also evaluated by the wound-healing assay. Cells (1 × 106) were seeded into each well of a 6-well plate until 80–90% confluence. A sterile 200-µL pipette tip was then used to draw a wound in the cell monolayer, following which, the cells were washed twice with phosphate buffer saline (PBS). Images of the wounds were obtained at 0 and 48 h using a photomicroscope and wound closure was evaluated in at least three different fields using Image J 1.52 (National Institute of Health, USA).
Flow cytometry of cell cycle and apoptosis
Flow cytometry was performed as previously described [
17].
Immunohistochemistry and immunofluorescence analysis
Paraffin-embedded sample tissues were consecutively cut into 4-μm slices and then mounted on glass slides for immunohistochemistry (IHC) staining. The sections were processed and stained using ZMAT1, p53, SIRT3, Ki67, and Tunel antibodies, respectively. After drying, the sections were examined and photographed under an Olympus BX63 microscope. ZMAT1 immunoreactivity was determined by staining intensity and distribution to obtain an H-score: < 50% staining was considered low expression, while ≧50% staining was considered high expression. The specimens were assessed independently by two pathologists. For immunofluorescence (IF) analysis, SW1990 cells were incubated with the primary antibody against ZMAT1 overnight at 4℃, followed by incubation with the secondary antibody for 2 h at 37℃ in the dark. DAPI was added for 10 min and images were taken with confocal microscopy (Olympus, FV3000) in a dark room.
Real-time quantitative polymerase chain reaction and western blotting
Real-time quantitative polymerase chain reaction (RT-qPCR) was used to assess mRNA expression. Total RNA was extracted with TRIzol reagent (Life Technologies, USA) and then reversed transcribed to cDNA with a reverse transcription kit (Toyobo, Japan). RT-qPCR was performed with SYBR green mix (Toyobo, Japan) on the Light Cycler 480II (Roche, USA). The relative mRNA expression level was determined using the 2−∆∆Ct method.
For western blot, cells were collected and lysed in an ice bath by RIPA buffer containing 1% phenylmethanesulfonylfluoride fluoride for 30 min to extract total protein. Protein concentrations were determined using the BCA protein assay. Equal amounts of protein were separated on an 8% SDS-PAGE gel and transferred onto a polyvinylidene fluoride membrane. After blocking for 1 h with a 3% BSA in TBST buffer, the membranes were probed with specific antibodies at 4℃ overnight. The membranes were washed by TBST buffer 5 times and incubated with secondary antibody marked with HRP-conjugated goat anti-rabbit IgG for 1 h at 25℃. Finally, the membranes were subjected to an enhanced chemiluminescence system (Pierce, USA) and the targeted protein bands were visualized.
RNA sequencing
RNA was collected from SW1990/Vector and SW1990/ZMAT1-OV cells. After RNA reverse transcription, amplification, and quality control, the clustering of the index-coded samples was performed on a cBot Cluster Generation System using the TruSeq PE Cluster Kit v3-cBot-HS (Illumina, California, USA) according to the manufacturer’s instructions. After cluster generation, the library preparations were sequenced on an Illumina Novaseq platform (Sinotech Genomics, Shanghai, China) and analyzed with Counts v1.5.0-p3 and Stringtie v1.3.3b software. Differential mRNA expression analysis was carried out using the R package “DESeq2.” Differentially expressed genes (DEGs) were obtained as a Log2|Fold Change|> 1.5 and a false discovery rate < 5%. The RNA sequencing data were deposited to Sequence Read Archive (SRA) of National Center for Biotechnology Information (NCBI) and the accession numbers are SRR17968055, SRR17968056, SRR17968057, SRR17968058, SRR17968059 and SRR17968060.
Luciferase reporter assay
The plasmid encoding the luciferase reporter gene for p53 (pGL3-p53) was purchased from Novel Biotechnology (Guangzhou, China). The reporter plasmid was transfected into HER-293 T cells together with blank plasmid or 0.01 μg, 0.05 μg, 0.1 μg, 0.5 μg, and 1 μg of ZMAT1. After 48 h of incubation, the luciferase activity was analyzed by Dual-Luciferase® Reporter (DLRTM) Assay System (Promega, Cat. No. E1910) and the relative luciferase activity was calculated by normalizing the firefly luciferase activity against the renilla luciferase activity. pGL3 vectors containing SIRT3 wild type and mutant promoters were chemically synthesized by Hedgehog BioScience and Technology Ltd (Shanghai, China). DNA segments were respectively cloned into the pGL3-basic vector purchased from Promega (Wisconsin, USA). Subsequently the transfection and luciferase reporter assay were performed as indicated as above.
Chromatin immunoprecipitation (ChIP) assay, ChIP-based quantitative polymerase chain reaction, and ChIP-based sequencing
Chromatin immunoprecipitation (ChIP) assays were conducted using the Simple CHIP Plus Enzymatic Chromatin IP kit (Cell Signaling Technology, USA). SW1990/Vector and SW1990/ZMAT1-OV cells were cross-linked in 1% formaldehyde for 10 min and then glycine solution and ice-cold PBS were added to inactivate the formaldehyde and wash the cells, respectively. After centrifugation, the cells were lysed in 100 μg Lysis Buffer 1 to remove the cell membranes and then 0.25 μL Micrococcal Nuclease was added for 15 min. Finally, 10 µL of MNase Stop Solution was added to stop the reaction and the samples were centrifuged to recover the nuclei. Next, the samples were lysed in 50 µL of Lysis Buffer 2 on ice for 15 min and centrifuged to obtain the supernatant containing the digested chromatin. Aliquots of 5 µL of the supernatant were stored at -20 °C and used as the input sample. The remaining 45 µL of supernatant was supplemented with 450 µL of 1 × IP Dilution Buffer, and subsequently incubated with 15 µL of anti-Flag antibody or 1 µL of Normal Rabbit IgG at 4 °C, overnight, respectively. After incubation with 20 µL of ChIP Grade Protein A/G Plus Agarose for 1 h on a rocking platform, each sample was washed with IP Wash buffer 1, IP Wash buffer 2, and IP Wash buffer 3. The samples were incubated with 150 µL of 1 × IP Elution Buffer at 65 °C for 30 min, and then placed in a collection tube with 2 µL of 20 mg/ml Proteinase K and 6 µL of 5 M NaCl. Meanwhile the input sample was thawed and added to 150 µL of the 1 × IP Elution buffer, 2 µL of 20 mg/ml Proteinase K, and 6 µL of 5 M NaCl. Input and IP samples were incubated in a 65 °C heat block for 1.5 h to reverse cross-links. The ChIP DNA and input DNA were purified using a DNA Clean-Up Column, rinsed with the DNA Column Wash Buffer, and eluted using the DNA Column Elution Solution.
The purified DNA was used for subsequent RT-qPCR and sequencing. RT-qPCR was performed with the SYBR Green Mix (Toyobo, Japan) on a Light Cycler 480II (Roche, USA). Sequences were generated using the Illumina NovaSeq 6000 genome analyzer and aligned to the Human genome (HG19) using BOWTIE software (V2.2.7). The mapped reads were used for peak detection by Model-based Analysis of ChIP-Seq (MACS) v1.4.2 software. Statistically significant ChIP-enriched regions (peaks) were identified by comparison of IP vs Input or comparison to a Poisson background model (cut-off P = 1 × 10–3).
Animal experiments
A total 1 × 106 SW1990/ZMAT1-OV cells and control cells were subcutaneously injected to BALB/c nude mice to establish the xenograft mouse models (n = 6 per group). The tumor sizes and volumes of mice were recorded by a digital caliper every 10 days. On Day 60, the mice were sacrificed and the tumors were dissected, photographed, weighed, and subjected to IHC analysis with the indicated antibodies. All 6-week old male BALB/c nude mice were purchased from GemPharmatech (Jiangsu, China). The assignment of mice to different groups was done randomly. All animal experiments were approved by the Ethics Association of Guangdong Provincial People’s Hospital.
Statistical analysis
The statistical analysis of continuous parameters between the two groups was determined by Student’s t-test, while one-way and repeated-measures analysis of variance (ANOVA) was used to compare multiple groups. χ
2 test or Fisher’s exact test was used to explore qualitative variables as appropriate. Spearman’s correlation was performed to analyze the correlation between variables. All statistical analyses were performed using R software version 4.0.1 (
https://www.r-project.org/) and SPSS version 24.0 (SPSS, Inc., Chicago, IL, USA). A
P < 0.05 was considered statistically significant unless otherwise specified.
Discussion
Finding novel driver genes regulating PDAC initiation and progression is of great value for identifying potential therapeutic targets. In our study, we demonstrated that ZMAT1 was down-regulated in PDAC and the expression of ZMAT1 was correlated with tumor differentiation, tumor stage, and patient survival, indicating it served as a predictive biomarker for PDAC. Furthermore, in vitro and in vivo experiments showed that the over-expression of ZMAT1 suppressed PDAC cells proliferation, migration, and tumor growth. Taking these together, our study identified ZMAT1 as a tumor suppressor in PDAC.
Mutation or loss of P53 is a critical molecular event leading to PC initiation and progression [
10,
22]. Studies have shown p53 regulates cancer cell cycle progression by inducing p21 expression [
23]. Similarly, p53 also impaired cancer cell apoptosis by affecting BAX expression [
24]. In our study, we determined that P53 was a key downstream factor of ZMAT1 and the effects of ZMAT1 on the cell cycle phase S/G2 arrest and cell apoptosis were dependent on p53. A p53 inhibitor rescued the inhibitory effect of ZMAT1 overexpression on cancer cells viability and proliferation. Of note, ZMAT1 influenced P53 expression in BXPC-3 cell line harboring the P53 mutation and SW1990 and Capan-2 cell lines not presenting any P53 mutations, which indicated that ZMAT1 exerted a general regulatory mechanism on P53. In future studies, it will be interesting to investigate association between ZMAT1 and P53 expression in PDAC patients with or without P53 mutation. In addition, by analyzing the survival data of PDAC cohorts, we found patients with low expression of both ZMAT1 and p53 exhibited the worst survival compared to those with ZMAT1 or high expression of p53. Thus, co-analyzing expression of ZMAT1 and p53 may be beneficial for identifying high-risk and aggressive PC.
ZMAT1 is a member of the C2H2-type zinc finger proteins, which are believed to bind to DNA and act as transcription factors. Indeed, the C2H2-type zinc finger proteins are the largest group among all zinc finger classes. Some binding motifs of C2H2-type zinc finger proteins like ZFP335 were identified [
25]. However, direct experimental data are missing supporting whether other particular C2H2-type zinc finger proteins bind to DNA, and their associated biological functions. In our study, we showed ZMAT1 directly bound to DNAs using the ChIP assay. Indeed, ChIP sequencing identified 2567 significant ZMAT1-binding peaks and 1079 putative ZMAT1-binding genes. We identified three ZMAT1 binding sites in the promoter region of SIRT3 and this interaction induced the transcription of SIRT3. These results suggested that ZMAT1 may serve as a transcriptional activator in human cells. It will be of great interest to further investigate the structure of ZMAT1 and explore physical interaction between the corresponding binding domains of ZMAT1 and the corresponding DNA fragments.
Our results showed that ZMAT1-induced p53 expression was dependent on SIRT3. Consistently, other studies also found that SIRT3 up-regulated p53. SIRT3 is a member of the Sirtuin family of class III histone deacetylases and is also considered a tumor suppressor in some solid tumors like hepatocellular carcinoma [
26]. In HCC cells, Zhang et al. reported that SIRT3 over-expression up-regulated p53 protein levels by reducing Mdm2-mediated p53 degradation [
27]. Xiao et al. revealed that SIRT3 activated p53/p21 signaling and caused apoptosis in the A549 cell line [
28]. Taken together with these results and our data, SIRT3 is sufficient to activate P53 in pancreatic cancer cells. Specifically, we illustrate a model that ZMAT1 functions in a p53-dependent manner and SIRT3 is an enhanced target and effector of ZMAT1. ZMAT1 binds to the promoter of SIRT3 and promotes the SIRT3 transcription, which activates p53 signaling pathway and affects pancreatic cancer cell proliferation and apoptosis (Fig.
7I). Furthermore, we found overexpression of ZAMT1 inhibited tumor growth in a xenograft tumor model, along with higher expression of SIRT3 and p53. Future studies may involve developing a treatment strategy of forced expression of ZMAT1 via an adenovirus or mRNA-mediated delivery system in a preclinical PDAC model.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.