Abstract
Background/Aim: We recently reported that miR-126 is down-regulated in malignant pleural mesothelioma (MPM) and can be combined into a 4-microRNA-classifier that can accurately diagnose MPM with high sensitivity and specificity. Herein we analyzed the epigenetic regulation of miR-126 and its host gene EGF-like domain, multiple 7 (EGFL7). Materials and Methods: Resected formalin-fixed paraffin-embedded MPM tissues from 29 patients, 14 patient-matched non-neoplastic pleura (NNP) specimens, 5 MPM diagnostic biopsies (DB), and 5 samples of pneumothorax-induced benign reactive mesothelial proliferation (PTHX) were analyzed. miR-126 and EGFL7 mRNA were quantified by RT-qPCR. CpG-islands' methylation in the EGFL7 promoter was analyzed using methylation-specific PCR and in the MIR126-containing intron 7 was quantified by pyrosequencing. Results: Relative to NNP, EGFL7 was under-expressed more than 4-fold in MPM (p<0.001). EGFL7 mRNA and miR-126 levels correlated in MPM (p<0.01) and NNP (p<0.001). The EGFL7 promoter region was hypermethylated in 69% of MPM and 80% of DB samples, but not in NNP and PTHX samples. EGFL7 promoter hypermethylation was associated with epithelioid histology (p<0.05) and reduced patient-survival (p<0.05). Conclusion: In MPM, DNA-hypermethylation down-regulates miR-126 and its host gene EGFL7, therefore is a poor prognostic factor, and may represent a future therapeutic target for de-methylating strategies re-establishing EGFL7 and miR-126 expression.
Malignant pleural mesothelioma (MPM) is a very aggressive, usually asbestos-induced cancer with dismal prognosis, originating from the mesothelial cells lining the pleura (1). Histologically MPM is classified as epithelioid, sarcomatoid or biphasic sub-types and the proportions of each are 50-60%, 10% and 20-30%, respectively (1).
Epidermal growth factor-like protein 7 (EGFL7) is secreted by endothelial cells and contributes to development of the embryonal vascular system. While EGFL7 is not expressed in the majority of mature vessels, highly vascularized tissues such as the lung, kidney and heart maintain high levels of EGFL7 into adulthood (2). In circumstances requiring angiogenesis, such as wound healing or tumorigenesis, EGFL7 is up-regulated and regulates the endothelial cells' sprouting that forms new vessels (3).
The EGFL7 gene is located on chromosome 9q34.3 and contains 11 introns and 3 independent transcriptional initiation sites (S1-3) that are related to the expression of full-length EGFL7 and two major transcript variants. While the full-length EGFL7 transcript is non-coding, transcript variant 1 (accession no. NM_016215) and transcript variant 2 (no. NM_201446) encode the same EGFL7 protein precursor (4).
MicroRNAs (miRNAs) are short (18-24 nucleotides) non-coding RNAs that regulate gene expression by binding to complementary sequences within mRNAs' 3’-untranslated region, thereby repressing translation and reducing the level of targeted transcripts (5). miRNAs are key regulators of fundamental cellular processes such as cell proliferation, apoptosis, differentiation, organ development and metabolism (6). Accordingly, de-regulated miRNAs may contribute to development of several cancer types (7-9). miR-126 (MIR126) is located within intron 7 of its host gene EGFL7 (10). Through targeting genes such as vascular endothelial growth factor A (VEGFA), EGFL7 itself, insulin receptor substrate-1 (IRS1), Kirsten rat sarcoma viral oncogene homolog (KRAS) and phosphoinositide-3-kinase, regulatory subunit 2 (PI3KR2), miR-126 has been linked to vascular integrity and tumor suppression (11-14).
Previous investigations have identified CpG-islands in intron 2 and 7 of EGFL7 and have demonstrated that miR-126 is not transcribed from its own promoter but is co-expressed with transcript variant 1 of EGFL7. Furthermore, the transcription initiation site (S2) has been located to the CpG-island in intron 2 of EGFL7 (15). The expression of genes harboring CpG-islands within promoter regions can be silenced by methylation of cytosine residues. In a cancer context, the process of aberrant transcriptional silencing by deregulated DNA-methylation can precede genetic mutations and represents an early event in cancer initiation and progression (16, 17).
We recently reported that miR-126 is down-regulated in MPM and the expression levels of miR-126, -143, -145, and miR-652 can be successfully combined into a 4-miRNA classifier that accurately diagnoses MPM with high sensitivity and specificity and correlates with overall survival (18). Herein, we investigated the methylation-associated regulation of miR-126 host gene EGFL7, in order to elucidate the molecular mechanism(s) of miR-126 down-regulation in MPM. We showed that the expression of miR-126 and EGFL7 is concomitantly down-regulated in MPM through methylation of the S2 CpG-island promoter and that silencing of EGFL7 is associated with a poor clinical outcome.
Materials and Methods
Patients and tissue samples. All patients included in the study were diagnosed and treated at the Rigshospitalet–Copenhagen University Hospital (Copenhagen, Denmark) in the period between 2006-2012. Diagnostic biopsies (DBs) were sampled from 5 MPM patients (4 males, 1 female; median age: 65, age range= 46-68; histotypes: 3 epithelioid MPM (EMM), 2 biphasic MPM (BMM)) prior to any neoadjuvant chemotherapy or cytoreductive surgery. Surgical specimens of MPM were collected from 29 MPM patients (21 males, 8 females, median age: 64, age range: 44-72) with the demographic and clinical features shown in Table I. Surgical treatment was either extrapleural pneumonectomy or pleurectomy/decortication as part of the trimodal protocol, that at our Institution is offered to operable patients (with sufficient cardiopulmonary function and no involvement of N2/N3 lymph nodes) diagnosed with EMM or BMM with <50% sarcomatoid component. Patient-matched histologically verified specimens of non-neoplastic pleura (NNP) were available from 14 of the 29 operated MPM patients. Because of the diffuse growth of MPM, sampling of NNP from the remaining 15 MPM patients was unsuccessful. As part of tri-modal therapy all MPM patients were administered 1-3 series of cisplatin/pemetrexed before surgery. To further verify differential miR-126 and EGFL7 expression profiles in cancerous and non-cancerous tissues, benign pleural specimens of reactive mesothelial proliferations caused by spontaneous pneumothorax (PTHX) were collected from 5 patients (all males; median age: 34 years, age range: 20-37 years) undergoing video-assisted thoracic surgery. All the tissue specimens analyzed in this study were formalin-fixed and paraffin-embedded (FFPE) and characterized by two thoracic pathologists (CBA and ES-R) according to current guidelines for histopathological diagnosis of MPM (1).
Ethics. Approval for the study was obtained from the Scientific Ethical Committees of The Capital Region of Denmark (J.NR. H-3-2011-135) and the national Data Protection Agency of Denmark (J.NR. 2011-41-6825).
Reverse-transcription quantitative PCR (RT-qPCR). Quantification of miR-126 was performed as recently reported (18). For quantification of EGFL7 (transcript variant 1) mRNA (200 ng) was reverse-transcribed using AffinityScript cDNA Synthesis Kit (Agilent Technologies, Santa Clara, CA). RT-qPCR amplification was performed using Brilliant III Ultra-Fast SYBR® Green QPCR Master Mix (Agilent Technologies) using primers and method as described (15) on a Rotor-Gene Q 6000 5plex platform (Qiagen, Hilden, Germany). All RT-qPCR data are presented as log2-transformed ratio according to the 2−ΔCq method (ΔCq=Cq, Target – Cq, Reference).
Detection of DNA-methylation. Genomic DNA was extracted from 4 x 10-μm-thick FFPE tissue-sections utilizing QIAamp DNA Mini Kit (Qiagen). Bisulfite treatment was performed using the EpiTect Bisulfite Kit (Qiagen). Methylation of the CpG-islands in the genomic region flanking the S2 transcriptional start site of EGFL7 was detected using methylation-specific PCR (MSP) as previously described (15) and PCR products were analyzed by high-resolution capillary electrophoresis via a QIAxcel Advanced System (Qiagen). The level of CpG-methylation at MIR126-containing intron 7 of EGFL7 was quantified by pyrosequencing amplifying bisulfite-converted genomic DNA using the PyroMark PCR Kit (Cat. No. 978703, Qiagen) and specific primers for the S2 region (forward: 5’-GGAGTTTTATATTAGTTAAGAAGGTAGAAG-3’ / reverse: 5′-biotin-ACCACCCTAAAAAAAATCAAAACTAAAATC-3’). The PCR amplification was performed on a thermal-cycler with enzyme activation at 95°C for 10 min, followed by 40 cycles of denaturing for 15 sec at 94°C, annealing for 30 sec at 55°C, extension for 30 sec at 72°C and final extension for 10 min at 72°C. Thereafter, single-stranded DNA was isolated by Streptavidin Sepharose High Performance beads (GE Healthcare, Little Chalfont, UK) and pyrosequencing performed using a specific sequencing primer (5’-ATTAGTTAAGAAGGTAGAAGT-3’) and PyroMark Q24 System (Qiagen). The analyzed sequence was: TTTC/TGTTTC/TGGGGTTTGTTTGTATTAGC/TGTAG. Pyrograms were analyzed using system-related PyroMark Q24 Analysis Software (version 2.0).
Statistical analysis. Contingency tables were analyzed using the Fisher's exact test. Gene expression in two groups of independent or patient-matched tissue-specimens was compared using the Student's two-tailed unpaired or paired samples t-test, respectively. Correlation of continuous variables was determined using linear regression. Gene expression in three groups or more were compared using one-way analysis of variation (ANOVA). p-Values for direct comparisons were adjusted for mass significance using Bonferroni post-tests. Analysis of overall survival times was performed according to the Kaplan-Meier method. Survival curves of dichotomized patients were compared using Mantel-Cox log-rank test. All statistical calculations were executed using OriginPro (version 8.6.0, OriginLab Corp., Northampton, MA) or SAS (version 9.4, SAS Institute Inc., Cary, NC). p<0.05 was considered significant.
Results
First, we tested whether miR-126 and EGFL7 (transcript variant 1) were concomitantly expressed in MPM, as previously described for primary bladder and prostate cancer (15). Linear regression analysis showed that the expression of EGFL7 and miR-126 was significantly correlated in both MPM (n=29, r=0.58, p<0.01) and NNP (n=14, r=0.94, p<0.001) (Figure 1A). To further investigate the de-regulation of EGFL7 and miR-126 expression in malignant vs. non-neoplastic pleural tissue, we determined the level of EGFL7 mRNA in 29 surgical samples of MPM, 14 patient-matched NNP samples from these MPM patients, 5 DBs from still chemotherapy-naïve MPM patients, and 5 samples of benign reactive mesothelial proliferations from pneumothorax patients (PTHX). The mean EGFL7 expression in the DB and MPM samples was 2.03 (95% CI=1.11-2.95) and 3.17 (95% CI=2.89-3.46), respectively and the difference between the means was not significant (p=0.12; Bonferroni test), indicating that the preoperative chemotherapy did not significantly influence EGFL7 expression. Likewise, no significant difference was found between the mean EGFL7 expression of 5.27 (95% CI=4.66-5.88) in the NNP samples and 5.89 (95% CI=4.55-7.22) in the PTHX samples, respectively (p=1; Bonferroni test). Compared to the group of NNP samples, the MPM specimens displayed significant EGFL7 underexpression by a factor of −4.27 (95% CI=−7.88 - −2.32; p<0.001 Bonferroni test) (Figure 1B). This observation was further substantiated by comparing the subset of patient-matched MPM and corresponding NNP samples directly, which showed a similar significant down-regulation of EGFL7 in the MPM samples by a factor of −4.17 (95% CI=−6.56 - −2.65; p<0.001, Student's paired-sample t-test) (Figure 1C). The set of tested MPM samples consisted of 14 EMM and 15 BMM specimens. When we compared these 2 histological sub-types of MPM, the BMM samples displayed significantly higher EGFL7 expression relative to the EMM specimens by a factor of 1.71 (95% CI=1.19 - 2.46; p<0.01, Students unpaired t-test; Figure 1D).
Next, we determined the methylation status of the EGFL7 S2 promoter region in the tissue samples using the MSP technique. We found methylation of CpG-islands in the S2 region in 4 out of 5 DB samples (methylated/unmethylated ratio: 0.80, 95% CI 0.45-1.00) and in 20 out of 29 MPM samples (methylated/unmethylated ratio: 0.69, 95% CI: 0.52-0.86) but not in any of the NNP or PTHX specimens. Representative results of MSP analysis by high-resolution capillary electrophoresis of the S2 CpG-island promoter region are shown in Figure 2. Accordingly, the expression of EGFL7 (transcript variant 1) in the subset of MPM samples harboring methylated S2 region was significantly lower by a factor of −2.36 (95% CI=−3.19 - −1.75; p<0.001, Student's unpaired t-test) than in the MPM samples without methylated S2 region (Figure 3A).
Then we correlated the methylation status of the EGFL7 S2 promoter region with standard clinico-pathological parameters of the MPM samples (Table I). While we did not observe any significant link between sex, age, smoking history or tumor stage, we did find that the EGFL7 S2 region's methylation was significantly associated with the EMM histological subtype (odds ratio: 14.9, 95% CI=11.1-19.9; p<0.05, Fisher's exact test). This was consistent with the above-described reduced EGFL7 expression in EMM vs. BMM samples (Figure 1D).
The potential correlation of the MPM patients' overall survival and methylation status of the EGFL7 S2 region was investigated using the Kaplan-Meier method. Overall survival data were available for 28 MPM patients and comprised of 8 events and 20 censored data. Dichotomizing MPM patients according to the EGFL7 S2 region's methylation status showed that the presence of S2-methylation was associated with significant worse survival (p<0.05, log-rank test; Figure 3B).
We used pyrosequencing of bisulfite-converted genomic DNA to explore the potential methylation of the CpG-island in the MIR126-containing intron 7 of EGFL7 in a sub-set of samples consisting of 8 MPM and 5 NNP specimens, and obtained equivalent outcomes. Indeed, quantitative pyrograms detected higher methylation levels (expressed as relative C/T content in percentages) in all 3 methylation-sensitive CpG-sites present in the region spanning the MIR126 locus of the MPM samples as compared to the NNP specimens. Representative pyrograms related to two cases of patient-matched MPM and NNP samples are presented in Figure 4.
Thus, taken together the results indicate that in MPM hypermethylation of the EGFL7 S2 promoter region correlates with parallel down-regulation of EGFL7 transcript variant 1 and miR-126.
Discussion
Recently, increased EGFL7 expression has been reported in certain human epithelial malignancies including hepatocellular, mammary, and gastrointestinal carcinomas (19). In contrast, reduced EGFL7 expression due to DNA-hypermethylation of the promoter region was observed in cancers of the lung, breast, bladder and prostate (15, 20). However, equivalent investigations have not been performed in MPM until now. As reported by others and our group, miR-126 is under-expressed in MPM (18, 21, 22) and this made us hypothesize that the expression of its host gene EGFL7 might be concomitantly de-regulated by DNA-methylation. Our current results demonstrate that miR-126 is expressed along with EGFL7 in MPM and that tumors with higher EGFL7 mRNA expression equivalently have high miR-126 levels as opposed to MPM with reduced EGFL7 expression that display parallel miR-126 down-regulation. This may be somewhat surprising, given that EGFL7 itself has been identified as a target of miR-126 (12) and miRNAs predominantly act to decrease target mRNA (23). However, the biological mechanism of several miRNAs, including miR-126, and their target regulation have been shown to be dependent on the cellular context (12, 20). Since high levels of EGFL7 mRNA and miR-126 are positively correlated, this may indicate that additional miRNAs are involved in regulating EGFL7 expression.
As stated above, increased expression of EGFL7 has previously been associated with certain epithelial cancers (19). In contrast, our results showed that EGFL7 was equally down-regulated in both chemotherapy-naïve biopsies and surgical MPM samples compared to benign NNP and PTHX specimens. This was further substantiated by comparing EGFL7 expression in patient-matched MPM and NNP samples directly. By comparing the expression of EGFL7 in the histological subtypes of MPM, we found that BMMs expressed higher levels of EGFL7 mRNA than EMMs. This is in accordance with our previous report (18) that described an equivalent higher miR-126 expression in BMMs and may reflect the composite nature and differentiation pattern of the biphasic histotype and cell-type-specific differences in miR-126 and EGFL7 expression.
We found that DNA-hypermethylation of the CpG-island in EGFL7 intron 2, which harbors the S2 transcriptional initiation site, was associated with decreased levels of EGFL7 mRNA in MPM, in accordance with findings in other cancer types (15, 20). Consistent with the aforementioned results, we found that DNA-hypermethylation of the EGFL7 S2 flanking region was significantly associated with EMM. Furthermore, DNA methylation of the EGFL7 S2 flanking region was a significant prognostic factor associated with poor survival. Interestingly, this is consistent with the previously reported negative prognostic impact of miR-126 down-regulation in MPM (18). In contrast, patients with hypomethylated DNA at the EGFL7 S2 locus were characterized by favorable prognosis and predominantly comprised biphasic tumor histology (7 out of 8 patients). Although based on few patients, this result is somehow surprising, given that BMM generally is associated with poorer patient survival than EMM (1). Nonetheless, the tissue specimens with hypomethylated DNA at the EGFL7 S2 locus are also characterized by increased levels of miR-126, which has tumor-suppressive functions (11-14, 22) and this may in part explain the favorable survival. In addition, miR-126 downregulates the chemokine stromal cell-derived factor 1 isoform α (Sdf-1α) and the related protein C-X-C chemokine receptor type 4 (CXCR4), which can recruit lymphocytes and mesenchymal stem cells and are overexpressed in MPM (20, 24, 25). These changes in the tumor microenvironment due to miR-126 down-regulation and consequent upregulation of Sdf-1α and its cognate receptor CXCR4 have been related to cancer progression by enhancing invasiveness and metastatic potential of neoplastic cells (20, 25). However, further investigations are required to clarify the mechanistic implications and the carcinogenic role of miR-126 and EGFL7 deregulation in MPM. Understanding the epigenetic co-regulation of EGFL7 and miR-126 in MPM may also provide future therapeutic options for this dismal disease based on new DNA-demethylating agents. Finally, considering that miR-126 is one of the miRNAs that has shown very promising diagnostic potential in MPM (18, 21), the concomitant down-regulation of EGFL7 and DNA-methylation of miR-126 and EGFL7 may represent additional tools that could aid in the difficult differential diagnosis between MPM and reactive mesothelial proliferations.
Acknowledgements
The study received financial support by the Danish Cancer Research Foundation, Sawmill Owner Jeppe Juhl & Wife Ovita Juhl Memorial Scholarship, Copenhagen Municipal's 100th Anniversary Foundation, Rigshospitalet – Copenhagen University Hospital Research Grants, Dagmar Marshall's Foundation, and Erasmus Traineeship Program.
Footnotes
↵* These Authors are equal senior authors.
- Received July 17, 2015.
- Revision received September 3, 2015.
- Accepted September 8, 2015.
- Copyright© 2015 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved