Introduction

Lung cancer is the leading cause of cancer-related deaths around the world. In 2015, approximately 2.2 million individuals were diagnosed with lung cancer, resulting in the death of 1.7 million patients [1]. Non-small cell lung cancer (NSCLC) is the most common type of lung cancer (approximately 80ā€“85% of this disease). It is divided into three subtypes: squamous cell carcinoma (LUSQ), adenocarcinoma (LUAD), and large cell carcinoma [2]. About 40% of lung cancers are LUAD, and this type tends to grow more slowly than other types of lung cancer [3]. There are currently numerous treatment options available for patients with LUAD depending on the stage of the cancer, the patientā€™s overall health and their cancer genomic status [4]. Due to recently developed molecularly targeted drugs and immune checkpoint inhibitors, the treatment outcome of patients with LUAD has been markedly improved. However, the therapeutic effectiveness for patients with distant metastasis is limited and the prognosis is extremely poor. To develop effective therapies, genomic studies of LUAD are being conducted worldwide using the latest genome analysis technology.

microRNAs (miRNAs) classified as small non-coding RNAs exist in the human genome and have been actually transcribed. They regulate the expression of protein coding/non-coding RNAs by repressing their translation or by cleaving transcripts. This is achieved in a sequence-dependent manner [5]. A single miRNA species can regulate a large number of RNA transcripts in physiological and pathological conditions [5]. Therefore, aberrantly expressed miRNAs can disrupt RNA and protein populations critical for normal cell maintenance and growth [6]. In fact, many studies have shown that dysregulated expression of miRNAs enhances cancer cell development, metastasis, and drug resistance [7]. Thus, aberrantly expressed miRNA can be closely involved in human cancer pathogenesis [8].

We have used RNA sequencing techniques to obtain miRNA expression signatures in cancer tissues. The data revealed that the expression of some miRNA passenger strands was upregulated or downregulated, suggesting that certain passenger strands might be involved in cancer pathogenesis [9, 10]. Our recent studies revealed that some passenger strands (miR-145-3p, miR-150-3p, miR-199a-3p, miR-216b-3p, and miR-99a-3p) targeted oncogenes in several cancers [11,12,13,14,15,16]. In fact, our results showed that both strands of the miR-145 duplex (miR-145-5p, the guide strand and miR-145-3p, the passenger strand) possess anti-tumor functions in LUSQ cells and bladder cancer [17, 18]. However, the functional significance of miR-145-3p and its targets in LUAD cells is still obscure.

In our continuing analysis of LUAD pathogenesis, we aimed here to identify important oncogenes that are controlled by miR-145-5p and miR-145-3p. We identified four possible target genes (LMNB2, NLN, SIX4, and DDC) that might be regulated by both strands of the miR-145 duplex. Among the targets, high expression of LMNB2 was significantly associated with poor prognosis of the patients with LUAD (disease-free survival, pā€‰=ā€‰0.0353 and overall survival, pā€‰=ā€‰0.0017). Overexpression of LMNB2 was observed in LUAD clinical specimens and its expression enhanced cancer cell aggressiveness. Defining the targets of anti-tumor miRNAs will enhance our understanding of the molecular pathology of LUAD. Moreover, these insights could lead to the development of early detection markers of LUAD and new therapeutic drugs.

Materials and methods

Human LUAD specimens

Formalin-fixed, paraffin-embedded specimens (FFPE) (19 LUAD tissues and 24 normal lung tissues) were collected at Kagoshima University Hospital (from 2010 to 2013) where patients underwent lung surgery. The clinical features of these specimens are summarized in Supplemental TableĀ 1. The pathological stages of these LUAD patients were determined according to the Association for the Study of Lung Cancer TNM classification, 7th edition.

Tissue samples for this study were obtained with informed consent and approval from each patient. Our study was approved by the Institutional Review Board for Clinical Research of Kagoshima University Hospital (approval number 26-164).

RNA extraction, reverse transcription, and quantitative real-time PCR

RNA extraction from FFPE specimens and cell lines and quantitative real-time reverse transcription-PCR (qRT-PCR) were described in our previous studies [10, 19, 20].

Expression levels of miR-145-5p and miR-145-3p were analyzed by stem-loop RT-PCR (assay ID: 002278 and 002149; Applied Biosystems). They were normalized to RNU48 (assay ID: 001006; Applied Biosystems). TaqMan probes and primers for LMNB2 (P/N: Hs00383326_m1) and GUSB (P/N: Hs00939627_m1) as an internal control were obtained from Applied Biosystems.

Cell culture and transfection of miRNAs and siRNAs

Two human LUAD cell lines (A549 and H1299) were used in this study. Both were purchased from the ATCC (Manassas, VA, USA). The procedures for transfection of miRNA or siRNA species into cancer cells were described in our recent reports [20,21,22].

The following miRNA species were used in this study: Pre-miRTM miRNA precursors (hsa-miR-145-5p, assay ID: PM 11480; hsa-miR-145-3p, assay ID: PM 13036; negative control miRNA, catalog no: 4464059, AM17111; Applied Biosystems). Also, we obtained the following siRNAs from Invitrogen (Carlsbad, CA, USA): Stealth Select RNAi siRNA, si-LMNB2 (P/N: HSS189213 and HSS189214; Invitrogen), and negative control miRNA/siRNA, Anti-miR Negative Control #1 (catalog no: AM 17010; Applied Biosystems).

Incorporation of miR-145-5p or miR-145-3p into the RISC by Ago2 immunoprecipitation

To determine whether exogenous miR-145-5p and miR-145-3p were incorporated into the RISC, immunoprecipitation assays were performed as described previously [14, 15, 23].

Cell proliferation, migration and invasion assays

The procedures for assessing cell proliferation, migration and invasion were described previously [19,20,21].

Identification of putative target genes regulated by miR-145-5p and miR-145-3p in LUAD cells

Microarray gene expression analysis and in silico analysis were used to identify putative target genes regulated by miR-145-5p and miR-145-3p in LUAD cells. The microarray data were deposited in the Gene Expression Omnibus (GEO) repository under accession number GSE77790. Our strategy for identification of miR-145-5p and miR-145-3p target genes is outlined in Supplemental Fig.Ā 1.

Plasmid construction and dual-luciferase reporter assay

The wild-type or deletion-type sequences targeted by miR-145-5p or miR-145-3p were inserted in the psiCHECK-2 vector (C8021; Promega, Madison, WI, USA).

After co-transfecting miRNA and the constructed vector in A549 and H1299Ā cells, firefly and Renilla luciferase activities were measured using a dual-luciferase assay kit (Promega Madison, WI, USA). The procedure was described previously [17, 20, 24].

Clinical database analysis

The clinical significance of miRNAs and their target genes in LUAD was investigated with TCGA database (The Cancer Genome Atlas: https://tcga-data.nci.nih.gov/tcga/). The gene expression and clinical data were retrieved from cBioPortal (http://www.cbioportal.org/) and OncoLnc (data downloaded on April 24, 2018) [25]. The normalized mRNA expression values in the RNA sequencing data were processed and shown as Z-scores. In the present study, patients were divided into two groups: those with Z-scoreā€‰ā‰„ā€‰0 and those with Z-scoreā€‰<ā€‰0. The TNM stage of LUAD patients was determined by the American Joint Committee on Cancer staging system in TCGA database.

Western blotting and immunohistochemistry

The procedures for Western blotting and immunohistochemistry were described previously [19, 20, 24]. Membranes were immunoblotted with monoclonal anti-LMNB2 antibody (1:1000 dilution) (ab8983; Abcam, Cambridge, UK) and monoclonal anti-GAPDH antibody (1:10,000 dilution) (MAB374; EMD Millipore, Billerica, MA, USA).

Immunohistochemistry was performed with a VECTASTAINĀ® Universal Elite ABC Kit (catalog no: PK-6200; Vector Laboratories, Burlingame, CA, USA) according to the manufacturerā€™s protocol. Primary mouse monoclonal antibody against LMNB2 (ab8983; Abcam) was used at a 1:100 dilution at 4ā€‰Ā°C overnight.

Statistical analysis

Mannā€“Whitney U tests were used to analyze the differences between two groups. Bonferroni-adjusted Mannā€“Whitney U tests were used to analyze the differences among more than three groups. Statistical tests were performed using Expert StatView software (version 5.0, SAS Institute Inc., Cary, NC, USA). Survival analysis was performed using the Kaplanā€“Meier method, and log-rank tests and multivariable Cox hazard regression analyses with JMP software (version 14; SAS Institute Inc., Cary, NC, USA).

Results

Downregulation of miR-145-5p and miR-145-3p in LUAD clinical specimens and cell lines

The expression levels of miR-145-5p and miR-145-3p were remarkably decreased in LUAD tissues in comparison with non-cancerous tissues (pā€‰<ā€‰0.0001 and pā€‰=ā€‰0.0039, respectively, Fig.Ā 1a, b). Positive correlations between the expression of miR-145-5p and miR-145-3p were confirmed by Spearmanā€™s rank test (Rā€‰=ā€‰0.862 and pā€‰<ā€‰0.0001, Fig.Ā 1c). In cancer cell lines, A549 and H1299, the expression levels of miR-145-5p and miR-145-3p were extremely low (Fig.Ā 1a, b). There was no correlation between expression levels of miR-145-5p and miR-145-3p, and clinical features of LUAD (Supplemental Fig.Ā 2).

Fig. 1
figure 1

The expression levels of miR-145-5p and miR-145-3p in lung adenocarcinoma (LUAD) clinical specimens and functional assays of miR-145-5p-transfected and miR-145-3p-transfected LUAD cell lines. a, b Expression levels of miR-145-5p and miR-145-3p in clinical specimens (LUAD and non-cancerous tissue) and LUAD cell lines (A549 and H1299) were determined by quantitative real-time reverse transcription-PCR (qRT-PCR). Data were normalized to RNU48 expression. The expression levels of miR-145-5p and miR-145-3p in cancer tissues were significantly downregulated in comparison with non-cancerous tissues (pā€‰<ā€‰0.0001, pā€‰=ā€‰0.0039, respectively). c Spearmanā€™s rank test indicated a positive correlation between the expression of miR-145-5p and miR-145-3p (Rā€‰=ā€‰0.862 and pā€‰<ā€‰0.0001). d Cell proliferation was evaluated by XTT assays. e Cell migration was determined by wound healing assay. f Cell invasion was determined by Matrigel invasion assay. Inhibition of proliferation, migration, and invasion was found in miR-145-5p-transfected and miR-145-3p-transfected cell lines (pā€‰<ā€‰0.0001). *pā€‰<ā€‰0.0001, **pā€‰<ā€‰0.05

Full size image

Ectopic expression of miR-145-5p and miR-145-3p suppressed cancer cell proliferation, migration, and invasive abilities

To verify the anti-tumor roles of miR-145-5p and miR-145-3p, we conducted gain-of-function studies by miRNA transfection into A549 and H1299 cells. Cell proliferation assays indicated the inhibition of cancer cell growth in miR-145-5p and miR-145-3p-transfected cells compared with mock-transfected or miR-control-transfected cells (both pā€‰<ā€‰0.0001, Fig.Ā 1d). Also, cell migration activities were reduced in the cells transfected with miR-145-5p or miR-145-3p (both pā€‰<ā€‰0.0001, Fig.Ā 1e). Moreover, Matrigel invasion assays revealed that transfection with miR-145-5p and miR-145-3p remarkedly decreased cell invasive capacity (both pā€‰<ā€‰0.0001, Fig.Ā 1f).

Identification of putative target genes regulated by miR-145-5p and miR-145-3p in LUAD cells: clinical significance of LMNB2, NLN, SIX4, and DDC

Using TargetScan database analysis, there were a total of 3773 and 2727 putative target genes that had binding sites for miR-145-5p and miR-145-3p, respectively. Among them, we selected genes that were upregulated in NSCLC clinical expression profiles from the GEO database (accession no: GSE19188). Then, we merged gene expression analysis data using miR-145-5p-transfected or miR-145-3p-transfected A549 cells (GEO accession number: GSE77790) and chose the downregulated genes. Finally, a total of four putative target genes (LMNB2, NLN, SIX4, and DDC) regulated by both miR-145-5p and miR-145-3p were identified (TablesĀ 1A,Ā 1B,Ā 1C).

Table 1A Candidate target genes regulated by miR-145-50070
Full size table
Table 1B Candidate target genes regulated by miR-145-3p
Full size table
Table 1C Candidate target genes regulated by miR-145-5p and miR-145-3p
Full size table

To investigate the relationship between these four genes and the course of LUAD, we examined these genes with the TCGA database (Fig.Ā 2aā€“d). Among these candidates, high expression of LMNB2 was associated with poor prognosis (disease-free survival, pā€‰=ā€‰0.0353, and overall survival, pā€‰=ā€‰0.0017). In this study, we focused on LMNB2 and validated the effect on LUAD cells.

Fig. 2
figure 2

Clinical significance of LMNB2, NLN, SIX4, and DDC expression based on The Cancer Genome Atlas (TCGA) database. aā€“d The relationship between the expression of four genes (LMNB2, NLN, SIX4, and DDC) and disease-free survival analyses, overall survival analyses, T stage, N stage, M stage, and disease stage. **pā€‰<ā€‰0.05

Full size image

Incorporation of miR-145-5p and miR-145-3p into the RISC in LUAD cells

We hypothesized that the miR-145-3p passenger strand might be incorporated into the RISC in LUAD cells. If so, it could serve as a tumor suppressor. To verify that concept, we immunoprecipitated Ago2 in cells that had been transfected with either miR-145-5p or with miR-145-3p. Note that Ago2 is an essential component of the RISC. Isolated Ago2-bound miRNAs were analyzed by qRT-PCR to determine whether miR-145-5p or miR-145-3p or both bound to Ago2 (Supplemental Fig.Ā 3A). In A549 cells, miR-145-3p transfectants showed higher expression levels of miR-145-3p than did mock transfectants or miR-control (pā€‰<ā€‰0.0001). Similarly, after miR-145-5p transfection, miR-145-5p was detected by Ago2 immunoprecipitation (pā€‰<ā€‰0.0001) (Supplemental Fig.Ā 3B).

Direct regulation of LMNB2 by miR-145-5p or miR-145-3p in LUAD cells

We examined whether the expression of LMNB2 mRNA and LMNB2 protein was suppressed by the restoration of miR-145-5p and miR-145-3p in LUAD cells. The expression levels of LMNB2 mRNA and LMNB2 protein were significantly decreased by miR-145-5p or miR-145-3p transfection compared with mock- or miR-control-transfected cells (Fig.Ā 3a).

Fig. 3
figure 3

Both strands of miR-145 (miR-145-5p and miR-145-3p) directly regulated LMNB2 in LUAD cell lines. a Expression levels of LMNB2 mRNA and LMNB2 protein 48ā€‰h after transfection of 10ā€‰nM miR-145-5p or miR-145-3p into LUAD cell lines. GUSB and GAPDH was used as an internal control. *pā€‰<ā€‰0.0001. b, c miR-145-5p or miR-145-3p binding sites in the 3ā€²-UTR of LMNB2. Luciferase assay using the vectors encoding putative miR-145-5p or miR-145-3p target sites of the LMNB2 3ā€²-UTR for both wild-type and deletion-type constructs. Renilla luciferase values were normalized to firefly luciferase values. *pā€‰<ā€‰0.0001

Full size image

Next, we sought to confirm that miR-145-5p and miR-145-3p directly regulated the LMNB2 gene in a sequence-dependent manner. TargetScan database projected the existence of binding sites in the 3ā€²-UTR of LMNB2 (miR-145-5p, position 2700ā€“2706; miR-145-3p, position 1689ā€“1696, Fig.Ā 3b, c). Accordingly, we carried out luciferase reporter assays with vectors that included either the wild-type or deletion-type 3ā€²-UTR of LMNB2. The luminescence intensities were markedly reduced by transfection with miR-145-5p or miR-145-3p and the vector carrying the wild-type 3ā€²-UTR of LMNB2. In contrast, transfection with the deletion-type vector did not reduce the luminescence intensities in LUAD cells (A549 and H1299) (Fig.Ā 3b, c). These findings showed that both strands of the miR-145 duplex directly bound the 3ā€²-UTR of LMNB2.

In human genome, miR-143 and miR-145 are known to be located close together on human chromosome 5q32 region, it called as the clustered miRNA. The TargetScan database search revealed that a putative binding site of miR-143-3p (the guide strand) has annotated in the 3ā€²-UTR of LMNB2. We also investigated the regulation of LMNB2 by miR-143-3p in LUAD cells. The mRNA and protein expression levels of LMNB2 were markedly downregulated in miR-143-3p expression (Supplemental Fig.Ā 4A, B). The luciferase reporter assay showed that miR-143-3p was directly bound to the 3ā€²-UTR of LMNB2 (Supplemental Fig.Ā 4C). These findings indicated that LMNB2 was directly regulated by miR-143-3p in LUAD cells.

LMNB2 knockdown inhibited cell proliferation, migration, and invasion in LUAD cells

To analyze the oncogenic role of LMNB2 in LUAD cells, loss-of-function assays were performed using si-LMNB2. We confirmed that both siRNAs (si-LMNB2-1 and si-LMNB2-2) significantly suppressed the expression levels of LMNB2 mRNA and LMNB2 protein by RT-PCR and western blotting (Fig.Ā 4a, b).

Fig. 4
figure 4

Loss-of-function studies by siRNAs. a, b Expression levels of LMNB2 mRNA and LMNB2 protein 48ā€‰h after transfection with si-LMNB2 into LUAD cell lines. c Cell proliferation was evaluated by XTT assays. d Cell migration was determined by wound healing assay. e Cell invasion was determined by Matrigel invasion assay. These assays showed that inhibition of proliferation, migration, and invasion were observed in si-LMNB2-transfected cell lines (pā€‰<ā€‰0.0001). *pā€‰<ā€‰0.0001

Full size image

Functional assays showed that cell proliferation, migration and invasive abilities were strongly inhibited by si-LMNB2 transfection in LUAD cells in comparison with mock- or miR-control-transfectants (Fig.Ā 4cā€“e).

Expression of LMNB2 in LUAD clinical specimens

LMNB2 was mainly detected around the nuclear membrane. Our results showed that the LMNB2 expression levels of cancer tissues were higher than those of non-cancerous tissues (Fig.Ā 5aā€“c).

Fig. 5
figure 5

Immunohistochemical staining of LMNB2 protein in LUAD specimens. Immunostaining of the LUAD specimens respectively collected from three patients who underwent lung surgery at Kagoshima University Hospital. a Lung adenocarcinoma with acinar growth pattern. b Lung adenocarcinoma with papillary and micropapillary growth pattern. c Lung adenocarcinoma with lepidic growth pattern. In all cases, high expression of LMNB2 was confirmed in the nuclear membrane of cancer cells

Full size image

Discussion

A large number of studies have shown that expression of miR-145-5p is deeply involved in human physiological and pathological conditions, e.g., the cell cycle, differentiation, migration, invasion and apoptosis [26, 27]. In human embryonic stem cells (ESCs), expression of miR-145-5p promotes the differentiation of ESCs by suppressing OCT4, SOX2, and KLF4, thereby silencing the self-renewal program [28]. In the field of cancer, many studies have demonstrated that miR-145-5p is frequently downregulated in a wide range of diseases and acts as an anti-tumor regulator through targeting several oncogenes [29,30,31]. Importantly, the promoter region of pre-miR-145 has a p53 binding region. Indeed, its transcriptional activity is controlled by the pivotal tumor suppressor p53 [32].

It has been generally believed that the passenger strands of miRNAs are degraded and therefore have no function [33,34,35]. Thus, it was rather surprising that our analyses of miRNA expression signatures by RNA sequencing showed that some passenger strands were dysregulated in cancer tissues [9, 10]. Our functional assays showed that both strands of the miR-145 duplex and the miR-150 duplex act as anti-tumor miRNAs in several cancers [9, 14, 17, 18, 23]. We hypothesized that genes in which both strands of miRNAs were coordinately regulated were important in cancer cells. Based on this hypothesis, we showed that MTDH was regulated by both miR-145-5p and miR-145-3p in lung squamous cell carcinoma. Moreover, UHRF1 was modulated by both strands in bladder cancer [17, 18]. In addition, SPOCK1 was coordinately regulated by miR-150-5p and miR-150-3p in esophageal cancer and in head and neck cancer [9, 23]. These oncogenes (MTDH, UHRF1, and SPOCK1) were overexpressed in cancer tissues and their expression was significantly associated with poor prognosis of the patients [17, 18, 23].

In this study, we identified 4 genes (LMNB2, NLN, SIX4, and DDC) that were coordinately regulated by miR-145-5p and miR-145-3p in LUAD cells. Analysis of these genes should provide important information regarding the molecular pathogenesis of LUAD. In human genome, the miR-143 and miR-145 formed clustered miRNA in chromosome 5q32 region. Several studies showed that the miR-143/miR-145 clustered miRNAs were frequently downregulated in several types of cancers and acted as anti-tumor miRNAs [36, 37]. Our previous studies showed that miR-143-3p and miR-145-5p were coordinately regulated oncogenic Golgi membrane protein 1 (GOLM1) and hexokinase-2 (HK2) in prostate cancer and renal cell carcinoma, respectively [36, 37]. Interestingly, our present data showed that LMNB2 was directly regulated by miR-143-3p, miR-145-5p, and miR-145-3p in LUAD cells. Moreover, overexpression of LMNB2 was significantly associated with the poor prognosis of LUAD patients. For these reasons, we focused on the LMNB2 and further analyses were performed in this study.

Lamins are pivotal components of the nuclear lamina, and their interactions with chromatin and its binding partner act to regulate cell proliferation and differentiation [38]. Lamins are subdivided into A-types and B-types based on their biochemical properties [38]. In the human genome, lamin proteins are coded by three genes (LMNA, LMNB1, and LMNB2) [38]. Previous studies showed that aberrantly expressed lamin family members are observed in a wide range of cancers [39,40,41,42,43]. Data indicated that ectopic expression of anti-tumor miR-193a-3p suppressed expression levels of HNRNRU, G3BP1, UHRF1, and LMNB2 in the highly metastatic human lung cancer cell line SPC-A-1sci [44]. More recently, knockdown of LMNB2 in A549 and H1299 cells inhibited cell proliferation and increased apoptosis, indicating that aberrantly expressed LMNB2 enhanced cancer cell aggressiveness [45]. This study also showed that minichromosome maintenance complex component 7 (MCM7) was a binding partner of LMNB2 [45]. Interestingly, overexpression of the MCM family was reported in several cancers and their expression was involved in cancer pathogenesis [46,47,48]. TCGA database showed that high expression of MCM7 was significantly associated with poor prognosis of LUAD patients (pā€‰=ā€‰0.0083). Identification of the proteins with which LMNB2 interacts will lead to the elucidation of the molecular pathogenesis of LUAD.

In conclusion, downregulation of both strands of the miR-145 duplex (miR-145-5p and miR-145-3p) was detected in LUAD clinical specimens. Data showed that both strands possessed tumor suppressive abilities in LUAD cells. To the best of our knowledge, this is the first report showing that anti-tumor miR-145-5p and miR-145-3p coordinately regulated LMNB2 in LUAD cells. Overexpression of LMNB2 enhanced cancer cell aggressiveness and its expression was involved in LUAD pathogenesis. Further studies of anti-tumor miRNA-modulated pathways in LUAD may provide novel insights into the molecular pathogenesis of this disease.