Introduction
Lung cancer is the second most common diagnosed malignancy and the leading cause of cancer death worldwide [
1]. Lung adenocarcinoma (LUAD) is the most common histological subtype of lung cancer [
2], and the 5-year survival rate is less than 30% [
3]. Metastatic disease is a poor prognostic feature and a leading cause of death in LUAD patients [
4,
5]. Rapid disease progression and organ failure due to metastasis account for the majority of patient mortality [
6,
7]. Metastatic lung adenocarcinoma cells can spread through the blood and lymphatic circulation, eventually colonizing the liver, bone, brain, and other organs. Unfortunately, despite significant advances that have been achieved in understanding the precise molecular mechanisms involved in LUAD, its mechanisms of development, progression, and metastasis remain elusive.
PCBPs are a subgroup belonging to the heterogeneous nuclear ribonuclear protein (hnRNP) family and contain three conserved RNA-binding domains, termed the heterogeneous nuclear protein k-homology (KH) domain, which has specific roles in interactions with RNAs and proteins, thereby regulating gene expression at various level [
8]. PolyC-RNA-binding protein 1 (PCBP1) is also known as hnRNP E1 [
9], which was identified as a component of heterogeneous nuclear ribonucleoprotein complexes and regulates alternative splicing, translation, and RNA stability of many cancer-related genes [
10]. PCBP1 can reduce the stability of LC3B and p62/SQSTM1 mRNA [
11,
12]. PCBP1 also plays a role in CD44 alternative splicing [
13]. PCBP1 regulates gene expression by binding to specific elements of its target mRNAs (e.g., p21, p63, and c-myc) [
14‐
16].
PCBP1 has been shown to be a tumour suppressor and is closely related to the occurrence and development of various tumours. In haematological tumours, PCBP1 was found to be a more common mutation type after the common mutations in Burkitt lymphoma with respect to ID3, TCF3, CCND3, and TP53 [
17]. Large-scale sequencing results revealed that mutations in PCBP1 also frequently occur in colon cancer [
18]. In prostate cancer, TGF-β1 enhances tumour stemness by downregulating expression of PCBP1 [
19]. In ovarian cancer, PCBP1 inhibits tumour progression by regulating the mRNA stability of p27 [
20]. In breast cancer, knockdown of PCBP1 enhances the stemness of breast cancer cells and promotes breast cancer invasion and metastasis by regulating the ILEI/LIFR signalling pathway [
21]. In gastric and pancreatic cancer, downregulation of PCBP1 expression promotes peritoneal metastasis [
22‐
24]. In recent years, researchers have attempted to elucidate the multiple regulatory roles of PCBP1 in tumorigenesis, development and metastasis; however, the function of PCBP1 in LUAD tumorigenesis and metastasis remains unknown.
Cancer metastasis, a process involving the spread of cancer cells from a primary lesion to distant organs, is one of the greatest contributors to cancer-related death [
25]. The Wnt signalling pathway and its involvement in cancers have been extensively investigated and play an important role in the development and metastasis of cancer [
26,
27]. This signalling pathway affects the maintenance of metastasis in lung cancer and provides targets for developing cancer therapy agents [
28]. Since the reduction of PCBP1 is a key alteration contributing to the acquisition of metastatic characteristics in tumour cells, understanding the function of PCBP1 is critical for its development as a therapeutic target to slow cancer progression. Utilizing functional studies, mechanistic investigations and mouse models, we demonstrate that PCBP1 inhibits LUAD development by upregulating DKK1 to inactivate the Wnt/β-catenin pathway. Our study highlights the important function of PCBP1 in LUAD, indicating its promising clinical potential as a prognostic marker and novel therapeutic target.
Materials and methods
Preprocessing of clinical and sequencing data and the heterogeneous nuclear ribonucleoprotein (hnRNP) family
This study incorporated data from two publicly available databases, the Cancer Genome Atlas (TCGA) and the Gene Expression Omnibus (GEO). LUAD samples from TCGA and normal tissue from Genotype-Tissue Expression (GTEx) were acquired from the UCSC Xena (
https://tcga.xenahubs.net). Additional LUAD cases (n = 82) were selected from GSE19188. Patients with an expression value of zero detected by more than 50% probes or missing prognostic data were excluded, while others were not included due to incomplete clinicopathological data. More gene expression and prognosis data of breast cancer, ovarian cancer, and gastric cancer patients were downloaded from the Kaplan–Meier (K-M) plotter database [
29,
30]. A list of the hnRNP family was extracted from previously published studies [
31].
Differentially expressed, survival, and functional analysis
LUAD patients from TCGA were divided into an early death group (death within one year) and a long-term survival group (more than 5-year overall survival). Propensity score (PS) matching analysis was performed between the two groups to adjust for clinical features. The linear model for microarray data (LIMMA) method was used to evaluate differentially expressed genes (DEGs) between cancer and adjacent tissues in LUAD patients and those who had different prognoses. Kaplan–Meier survival curve analyses were performed on TCGA and GSE19188 datasets using the R package “survcomp” to determine the prognostic value of genetic markers [
32]. Based on the "ggpubr" R package, the research team constructed box plots to show the specific gene expression levels of different samples. The correlations between gene expression and molecular or immune subtypes across cancers were explored using the TISIDB database [
33]. Meanwhile, the research group performed Gene Set Enrichment Analysis (GSEA) using “javaGSEA” [
34].
Human samples
We obtained fresh tumour tissues and paired normal tissues from 31 patients and paraffin-embedded tissue samples from 72 patients with pathologically confirmed LUAD (Additional file
1: Table S1) who underwent surgery at the National Cancer Centre/National Clinical Research Centre for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College (Beijing, China).
Clinical data were collected by reviewing the patients’ medical histories. Pathological staging was performed according to the 8th edition of the American Joint Committee on Cancer/Union for International Cancer Control TNM classification system.
Cell culture
The human lung adenocarcinoma A549 and H358 cell lines were cultured in RPMI-1640 (Corning) supplemented with 10% foetal bovine serum (FBS, Gibco) and 100 U/mL penicillin–streptomycin (Gibco). Cells were cultured at 37 °C in a humidified atmosphere containing 5% CO2.
Small interfering RNA (siRNA) and plasmid transfection
To establish stably interfered or target gene-expressing cells, we transfected lentivirus-mediated shRNA specifically against PCBP1 and lentiviral PCBP1 overexpression constructs into A549 and H358 cells according to the manufacturer’s protocols. Positively infected cells were selected using puromycin treatment for 3 weeks. Transfection of cells with siDKK1 was performed using Lipofectamine™ 3000 (Invitrogen, Carlsbad, USA) according to the manufacturer’s instructions.
RNA isolation, reverse transcription, and qPCR
Total RNA was isolated from tissues or cells using RNAiso Plus (9108, TakaRa), and then total RNA was used for complementary DNA (cDNA) synthesis using the Prime Script RT reagent kit according to the manufacturer’s instructions (RR036A, TaKaRa). Relative mRNA expression levels were determined by qPCR using SYBR Green Master Mix (RR82LR, TaKaRa). β-actin and GAPDH served as internal controls, and the 2
−ΔΔCT method was utilized to calculate relative expression levels. Detailed primer sequences are listed in Additional file
1: Table S2.
Western blot
Proteins were extracted from cells and animal tissues. Next, the proteins were transferred onto polyvinylidene fluoride membranes (Millipore, Bedford, MA) using electrophoresis and transferred to a 0.2-µm nitrocellulose membrane (GE Healthcare, Chicago, IL USA). Membranes were blocked with 5% nonfat milk and incubated with primary antibodies overnight at 4 ℃. The primary antibodies used for western blotting were as follows: PCBP1 (1:2000, ab74793, Abcam), DKK1 (1:1000, ab109416, Abcam), phosphor-β-catenin (1:1000, 9561S, CST), Claudin-1 (1:1000, 13255 s, CST), β-catenin (1:1000, 8480 s, CST), and Vimentin (1:1000, 5741 s, CST). The signalling detection was performed using an ECL detection kit (WBKLS0500, Millipore).
mRNA stability assay
A549 cells stably expressing PCBP1 and control cells were treated with 2 mg/ml actinomycin D (SBR00013, Sigma) to inhibit mRNA transcription. Total RNA was extracted for cDNA synthesis and detected by semiquantitative and real-time RT–PCR as indicated. Relative mRNA levels were normalized to the starting point of treatment.
IncuCyte™ cell proliferation and wound healing assays
IncuCyte live-cell imaging enables noninvasive, fully kinetic measurements of cell growth based on area (confluence) metrics. Cell cultures were imaged every twelve hours using IncuCyte (Essen BioScience, USA). A total of 3 × 103 cells were seeded into 96-well plates, and cell proliferation was evaluated by the degree of cell confluence. A total of 8 × 104 cells were plated in duplicate in 96-well plates and grown to 90% confluence. Wounds were then created using IncuCyte ZOOM™. After washes to remove cellular debris with PBS, serum-free medium was added. Plates were then cultured in an IncuCyte ZOOM™ incubator. The results of the cell proliferation and migration assays were analysed 72 h later.
RNA sequencing (RNA-seq) and bioinformatics analyses
RNA-seq assays were performed at Novogene (China) to measure the mRNA expression profiles of A549 cells and shPCBP1 A549 cells using Illumina PE150. Differentially expressed genes (DEGs, |logFC|≥ 0.5) were clustered and visualized using the pheatmap R package. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses of DEGs were performed using the clusterProfiler R package.
RNA pulled down and silver staining
The RNA–Protein Pull-Down Kit (21,115, Thermo Fisher Scientific) was used according to the manufacturer’s instructions for RNA pulldown. Biotin-labelled DKK1 RNA and control RNA were obtained in vitro using the MEGAscript™ T7 Transcription Kit (AM1334, Invitrogen) and bio16-UTP (AM8452, Invitrogen). Biotinylated DKK1 RNA was captured using streptavidin magnetic beads and subsequently mixed with the A549 cell protein lysates. The compounds were separated by SDS–PAGE and subjected to western blotting and silver staining.
RNA immunoprecipitation (RIP) assay
The EZ-Magna RIP Kit (17–701, Millipore) was used according to the manufacturer’s instructions. Immunoprecipitated RNA was extracted from the eluent and reverse transcribed. Protein–RNA complexes immunoprecipitated by anti-PCBP1 or IgG were determined through qPCR.
Transwell assay
A transwell plate with an 8 μm porous polycarbonate membrane was used to assess tumour cell migration (3422, Corning). The tumour cells were suspended in serum-free RPMI 1640 and placed on the upper portion of a transwell chamber. The 10% FBS RPMI1640 placed in the lower portion of the chamber was used as a chemoattractant. After 24 h of incubation, the cells on the upper side were removed with cotton swabs, whereas the migrating cells on the lower side were fixed in 4% paraformaldehyde at room temperature for 30 min and then stained with 0.1% crystal violet for 30 min. Finally, the migrating cells were imaged and counted using ImageJ software.
Immunohistochemistry (IHC)
The tissue slices were harvested and then processed with formalin, dehydrated, paraffinized and sectioned. Immunohistochemical staining was performed using antibodies against PCBP1 (1:200, ab109577, Abcam), DKK1 (1:200, ab109416, Abcam), β-catenin (1:200, 8480, CST) and Ki67 (1:200, ab 15,580, Abcam). The immunopathology score was assessed by two pathologists, and the total result was given according to the positive intensity and area extent. Each specimen received a score according to staining intensity (negative = 0, weak = 1, moderate = 2, and strong = 3) and the percentage of positive cells (0% = 0, 0–1% = 1, 2–10% = 2, 11–30% = 3, 31–70% = 4, and 71–100% = 5).
CCK-8 assay
Transfected A549 and H358 cells were seeded into 96-well plates (2000 cells/well). Cell viability was assessed using the Cell Counting Kit-8 (CCK-8) assay (CK04, Dojindo) according to the manufacturer’s instructions. The absorbance was measured at 450 nm using a microplate reader (Thermo Fisher Scientific, Singapore).
A549 and H358 cells were seeded into 6-well plates with 200 cells per well supplemented with 2 mL of 10% FBS cell culture medium, and the medium was changed every 3 days. After 10–14 days, harvested cells were fixed in 4% PFA for 30 min and stained with crystal violet. Three images of each well were acquired.
Mouse models
Five-week-old female BALB/c nu mice were used to establish a subcutaneous xenograft model. A549 and H358 cells transfected with sh-PCBP1 or control and PCBP1 or NC were resuspended in PBS. Transfected A549 and H358 cells were inoculated subcutaneously in 100 μL PBS at a density of 1 × 106 cells per mouse, with 6 mice in each group. Starting 12 days after inoculation, the tumours were measured every 3 days. The tumour volume (mm3) was assessed as follows: volume = (length × width2)/2.
Five-week-old female NSG mice were used to establish a metastatic lung model. In the metastatic lung model, 1 × 106 cells were injected into the tail vein. After inoculation, mouse weight was measured every week. Eight weeks after the H358 injection or sixteen weeks after the A549 injection, mice were sacrificed to collect the lungs. Metastatic nodules were assessed using HE staining. This study was approved by the Institutional Animal Care and Use Committee of Peking Union Medical College, Chinese Academy of Medical Sciences.
Statistical analysis
Statistical analyses were performed using Prism 7.0 or R. The paired Student's t test, chi-square test, Kaplan–Meier method and Spearman's correlation analyses were used (*P < 0.05; **P < 0.01). The data are reported as the mean ± S.D.
Discussion
The hnRNP family contains many members involved in alternative splicing, mRNA stability, and translation regulation, and has been shown to play a regulatory role in a variety of diseases [
36‐
38]. PCBPs is a subgroup belonging to the hnRNP family. PCBPs can be divided into two subgroups: hnRNP K/J and hnRNPE. HnRNP E protein consists of four members, namely PCBP1, PCBP2, PCBP3 and PCBP4. PCBPs has the common characteristics of three hnRNP K homology (KH) domains [
39]. The amino acid sequences of PCBP 1 and PCBP 2 were highly similar (89%), but about 50% similar to those of the other three family members [
40,
41]. However, according to current reports, the role of PCBP1 and PCBP2 in tumor is completely opposite, PCBP1 can be considered as a tumor suppressor, and PCBP2 plays a carcinogenic role. PCBP3 and PCBP4 have rarely been studied in cancer [
38]. Through data analysis, we found that the expression of PCBP1 in patients with long survival time was higher than that in patients with short survival time, and the expression of PCBP1 in tumor tissues was significantly different from that in normal tissues. Therefore, we chose to functionally elucidate the role of PCBP1.
Loss of PCBP1 is implicated in various types of carcinogenesis [
10,
13,
42]. Recently, PCBP1 has been shown to be a negative regulator of ovarian carcinoma [
20], breast cancer [
43], gastric cancer [
22], and other cancers. EMT is a process in which epithelial cells acquire mesenchymal characteristics. In cancer, EMT is related to tumorigenesis, invasion, metastasis, and resistance to treatment [
44]. Several recent studies have reported that PCBP1 is implicated in EMT markers [
45,
46]. In addition, PCBP1 confers drug sensitivity in colorectal cancer, prostate cancer, and breast cancer [
47]. All these findings highlight the clinical significance of PCBP1 in cancers. However, the exact role of PCBP1 in LUAD remains largely undefined.
In our study, we found that PCBP1 was decreased in LUAD tumour tissues compared to matched normal tissues. LUAD patients with high PCBP1 expression exhibited a longer survival time, and the data presented suggest that PCBP1 may act as a novel marker associated with good prognosis. Here, LUAD cells with silenced or overexpressed PCBP1 were established. Using colony formation assays, CCK8 assays, IncuCyte™ cell proliferation and wound healing and Transwell assays, we confirmed that PCBP1 was closely related to the proliferation and migration of LUAD cells. Mechanistically, the RNA-binding protein PCBP1 represses LUAD by directly binding to DKK1 mRNA. Downregulating expression of PCBP1 shortened the half-life of DKK1 mRNA and reduced its expression. Subsequently, the decreased expression of DKK1 protein downregulated beta-catenin phosphorylation, which relieves the inhibition of the Wnt/β-catenin signalling pathway. Disrupted β-catenin and other EMT markers (claudin-1 and vimentin) contribute to the malignant phenotype of LUAD cells. To fully elucidate the role of PCBP1 in LUAD, metastatic lung and subcutaneous xenograft models were successfully constructed. The data indicated that PCBP1 knockdown promoted lung metastasis and proliferation in vivo. Taken together, these results reveal that PCBP1 acts as a tumour suppressor gene, inhibiting the tumorigenesis of LUAD. Our findings highlight the potential of PCBP1 as a promising therapeutic target.
In addition to participating in tumour metastasis, PCBP1 has other important functions. PCBP1 is involved in regulating the stability of intracellular iron [
48], and our study found that PCBP1 regulates ferroptosis in tumour cells (data not shown). PCBP1 also plays an important role in immune regulation. PCBP1 can promote the production of GM-CSF in T cells [
49], PCBP1 is an intracellular immune checkpoint for shaping T-cell responses in cancer immunity [
50], and PCBP1 modulates the innate immune response [
51]. These studies suggest that PCBP1 may play a more complex role in tumours, which needs to be further investigated.
The tumorigenesis mechanisms of PCBP1 include execution of alternative splicing of oncogenes, inhibition of oncogene translation and regulation of mRNA stability [
10]. PCBP1 is known to recognize poly(rC)- or CU-rich elements in its targets. It has been reported that PCBP1 can bind to the GCCCAG motif in the 5′-UTR of PRL-3 [
42]. PCBP1 binds to the CU-rich motif in the 3′-UTR of P63 mRNA to stabilize p63 mRNA [
16]. PCBP1 stabilizes sortilin mRNA by binding to atypical C-rich elements in the 3′-UTR [
52]. PCBP1 also recognizes the atypical AU-rich element in the POLH mRNA 3′-UTR to stabilize POLH mRNA [
53]. In this study, we found that PCBP1 directly binds to DKK1 to regulate the stability of DKK1 mRNA, but the specific binding region was not identified. Future studies are needed to clarify the specific binding elements or regions of PCBP1 and DKK1.
DKK1 is a secretory Wnt antagonist, and abnormal expression of DKK1 has become recognized as an important regulator of many human cancers. Some reports indicate that DKK1 is a tumour promoter in pancreatic cancer, oesophageal cancer, and hepatocellular carcinoma [
54,
55]. In contrast, DKK1 expression is downregulated in thyroid cancer [
56], cutaneous squamous cell carcinoma [
57], and breast cancer [
58], suggesting that DKK1 may have a tumour suppressive effect. The role of DKK1 in tumours is controversial and may play different roles in different organs and tumours [
59]. The function of DKK1 expression in lung cancer progression and prognosis remains unclear. Knockdown of DKK1 sensitizes NSCLC cells to cisplatin [
60]. However, it has also been reported that a decrease in DKK-1 in metastatic lung cancer cells eliminates microglial inhibition and increases the risk of metastasis [
61]. We found that DKK1 inhibits LUAD development via degradation of β-catenin. These results may indicate that DKK1 may play different roles in different organs, cell populations, or tumour states. Extensive efforts have been made to block Wnt signalling using small molecules, and the most effective target in cancer is the complex between TCF and β-catenin because it modulates signalling at downstream nodes of the pathway, a goal that has proven difficult to achieve despite extensive efforts [
62]. DKK1 significantly reduces the expression of β-catenin, and targeting PCBP1 reduces catenin to achieve the therapeutic effect of inhibiting the WNT signalling pathway. Therefore, by studying the regulatory mechanism of PCBP1, we provide a new approach for blocking the WNT signalling pathway through β-catenin, providing a new therapeutic perspective for LUAD patients.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.