MicroRNA expression profiles and clinicopathological implications in lung adenocarcinoma according to EGFR, KRAS, and ALK status

Lung adenocarcinoma has distinctive clinicopathological features that are related to specific genetic alterations, including EGFR and KRAS mutations and ALK rearrangement. MicroRNAs are small non-coding RNAs that post-transcriptionally regulate many important biological processes and influence cancer phenotypes. This study retrospectively investigated microRNA expression profiles, and their clinicopathological implications, in lung adenocarcinoma according to genetic status (EGFR, KRAS, ALK, and triple negative). A total of 72 surgically resected lung adenocarcinoma specimens (19 EGFR-mutated, 17 KRAS-mutated, 16 ALK-rearranged, and 20 triple negative cancers) were screened for 23 microRNAs using quantitative real-time reverse transcriptase polymerase chain reaction. We then evaluated the associations between the microRNA expressions and the cancers’ genetic and clinicopathological features. Eight microRNAs were associated with clinicopathological features, such as male sex and ever-smoker status (high miR-373-3p, miR-1343-3p, miR-138-1-3p, and miR-764; low miR-27b-3p) and vascular invasion (high miR-27b-3p; low miR-1343-3p and miR-764). Clustering and discriminant analyses revealed that the microRNA expression patterns in the ALK group were different from those in the EGFR and KRAS groups. Five microRNAs (high miR-1343-3p; low miR-671-3p, miR-103a-3p, let-7e, and miR-342-3p) were especially distinctive in the ALK group, compared to the EGFR and KRAS groups. Moreover, a significant association was observed between ALK-rearrangement, decreased miR-342-3p expression, and immunohistochemical loss of E-cadherin. Therefore, microRNA expression profiles appear to have distinctive clinicopathological implications in ALK-rearranged lung adenocarcinoma. Furthermore, the association of ALK rearrangement, decreased miR-342-3p expression, and E-cadherin loss might indicate that miR-342-3p is involved in the ALK-associated phenotypes and epithelial-mesenchymal transition.


INTRODUCTION
Lung adenocarcinoma is the most common histological type of primary lung cancer [1], and the identification of key oncogenic drivers has created a paradigm shift in our understanding of lung adenocarcinoma biology.Furthermore, lung adenocarcinoma is a heterogeneous tumor with diverse molecular, clinical, and pathological characteristics [2][3][4], and the mutational status of driver oncogenes can be used to create personalized strategies for treating the advanced disease.Lung adenocarcinomas with activating mutations of the epidermal growth factor receptor gene (EGFR) have been reported to respond to EGFR tyrosine kinase inhibitors (TKIs) [5][6][7], and rearrangement of the anaplastic lymphoma kinase gene (ALK) is the best predictor of lung adenocarcinoma response to crizotinib (an ALK TKI) [8][9][10].
MicroRNAs are small endogenous non-coding RNAs that can modulate protein expression by regulating translational efficiency or cleavage of the target mRNA [11].Deregulated microRNA expression has been identified in a variety of human malignancies, which suggests potential oncogenic or tumor-suppressive roles, depending on the cancer cell, tissue, and target gene [12][13][14][15][16].In this context, microRNA expression profiles are emerging as potentially useful diagnostic and prognostic biomarkers, which can facilitate personalized therapy and disease management [17,18].Oncogenic and tumor-suppressive microRNAs are deregulated in the various tumor entities of lung cancer [19][20][21][22][23], and accumulating evidence indicates that dysregulation of specific microRNAs contributes to the development and progression of lung cancer.For example, overexpression of miR-17-92, miR-21, and miR-128 has been observed in lung cancer and has implications in its carcinogenesis [19][20][21]24].Compared to the normal lung, lung adenocarcinoma exhibits dysregulation of several microRNAs, such as let-7a-2, let-7a-3, miR-15b, miR-21, miR-155, and miR-200b, which are associated with the development and progression of cancer (altered tumor suppression) and shorter patient survival (reduced prognosis) [25].Among these microRNAs, let-7 was the first to be confirmed as having aberrant expression in lung carcinoma [26,27], and it is now known that the let-7 microRNA family controls the activity of the RAS oncogene [28].Other microRNAs, such as miR-200b, regulate the chemosensitivity of human lung adenocarcinoma cell lines, mainly through their effects on cell proliferation, cell cycle distribution, and apoptosis [29].
Few studies have investigated microRNA expressions and genetic profiles in lung cancer, although recent studies have demonstrated that EGFR may be a functional target of microRNAs [24,30].In this context, miR-128b is frequently deleted in lung cancers and directly downregulates EGFR, but is not associated with EGFR mutational status and prognosis [24].Another study demonstrated that miR-96 decreased the phosphorylation of ALK target proteins, and reduced the proliferation, colony formation, and migration of ALK-expressing cancer cells, which suggests that miR-96 may be involved in the aberrant expression of ALK in cancer cells [31].However, little is known regarding any differences in microRNA expressions in lung adenocarcinomas with different oncogene mutational statuses.Therefore, the present study aimed to 1) investigate the microRNA expression profiles in surgically resected lung adenocarcinomas with different oncogenic mutations, 2) analyze their clinicopathological implications, and 3) evaluate any associations between the mutation-related phenotypes and microRNAs.This information may help improve our understanding of tumor behavior according to mutational status, and might help identify therapeutic strategies for lung adenocarcinoma that are based on the use of specific microRNAs.

Clustering and discriminant analyses according to the four genetic groups
Hierarchical clustering analyses of the 23 microRNA expression levels revealed two large clusters: an EGFR/KRASrich group and an ALK/TN-rich group (Figure 1A).Discriminant analyses of the 23 microRNAs revealed that the EGFR and KRAS group were centered close to each other, and were relatively distant from the ALK and TN groups (Figure 1B).These findings suggest that the microRNA expression profiles were similar in the EGFR and KRAS groups, and that these groups had relatively distinct microRNA expression profiles, compared to the ALK and TN groups.

Distinctive expressions of microRNAs in the ALK group and the EGFR/KRAS groups
Based on the assumption that the TN group was composed of genetically heterogeneous cases, and the observation that the microRNA expression patterns were similar in the EGFR and KRAS groups, we focused on comparing the ALK group to the EGFR and KRAS groups.

Association between microRNA and E-cadherin protein expressions in the ALK group and the EGFR/KRAS groups
To investigate how ALK-distinctive microRNAs could contribute to the clinicopathological features of the ALK group, we focused on the frequent epithelialmesenchymal transition (EMT) phenomenon in the ALK group and its relationship with the ALKdistinctive microRNAs [32].For these analyses, we immunohistochemically evaluated the loss of E-cadherin expression (a well-known EMT marker) and two ALK-distinctive microRNAs (let-7e and miR-342-3p), which are associated to EMT and cancer cell migration and invasion [33,34].As shown in Table 3 and Supplementary Table 1, the ALK group exhibited low levels of let-7e and miR-342-3p expression, and loss of E-cadherin expression.The expressions of miR-342-3p, let-7e and E-cadherin tended to be closely related (Figure 3). Figure 4 shows the immunohistological staining results for E-cadherin.These findings indicate that the ALK group could be characterized by decreased miR-342-3p and let-7e expression and frequent loss of E-cadherin, which suggests increased EMT.

DISCUSSION
Lung adenocarcinoma represents a heterogeneous group of tumors with broad spectrums of morphology, clinical behavior, genetic alteration, and therapeutic response.Furthermore, the discovery of driver oncogenes (e.g., EGFR, KRAS and ALK) has changed our understanding of and approach to treating lung carcinogenesis, which has highlighted the importance of the genotype in lung cancer research [35,36].Thus, we aimed to provide insight into lung adenocarcinoma biology according to various genotypes, and investigated whether microRNA expression profiles could be used to characterize the genotypic groups of lung adenocarcinomas.Several reports demonstrated that unique miRNA expressions were related with genotypes in lung adenocarcinoma [37][38][39].Although a few miRNAs such as miR-184 was also identified in our study, it was difficult to compare or integrate the results due to the differences in study design and methods between these studies and ours.In addition, we compared with four different genotypes in lung adenocarcinoma, which has not yet been performed in other studies.The present study revealed that 1) several microRNAs' expression levels were associated with clinicopathological features of lung cancers (e.g., miR-27b-3p expression was closely associated with vascular invasion), 2) the microRNA expression profiles in ALK-rearranged cancers were different from those in EGFR/KRAS-mutated cancers, 3) ALK-rearranged cancers exhibited characteristic expression changes for five microRNAs (including let-7e and miR-342-3p), compared to EGFR/KRAS-mutated cancers, and 4) the loss of E-cadherin expression was associated with ALK-rearrangement and decreased miR-342-3p expression, which suggests that miR-342-3p may play a role in the EMT of ALK-rearranged cancers.
MicroRNAs are precise regulators of various physiological processes and disease models, such as cancers and vascular diseases [11,12,14,15,40,41].In addition, vascular invasion is an important step in metastasis to distant sites, and angiogenesis and its regulators are associated with tumor growth and prognosis in various solid tumors, including lung cancer [42,43].In this context, miR-27b-3p is a pro-angiogenic microRNA [40,44] that is up-regulated in vascular endothelial cells [45][46][47].In a recent mouse study, increased miR-27b-3p expression resulted in enhanced neovascularization and contributed to tumor growth [44].These findings may help explain the association between increased miR-27b-3p and vascular invasion in our study.However, given our limited ability to measure microRNA expressions in the complete specimens (intermixed tumor, vascular, stromal, www.impactjournals.com/oncotargetand other microenvironmental cells), the precise source of the increased miR-27b-3p expression in lung cancer with vascular invasion remains unclear.For example, proangiogenic miR-27b-3p may be actively produced and secreted by tumor cells to permeabilize the vessels and facilitate vascular supply and tumor growth, or there may be secondary production as part of a vascular regenerative process after invasion by the tumor cells.Further studies are needed to evaluate this issue.Other reports have also revealed that miR-1343-3p reduces the expression of transforming growth factor-β (TGF-β) receptor-1, which induces angiogenesis through vascular endothelial growth factor (VEGF)-mediated apoptosis [48,49].Therefore, miR-1343-3p may also play an anti-angiogenic role that could oppose the pro-angiogenic role of miR-27b-3p.Moreover, little is known regarding about the angiogenic effects of miR-764.Databases of microRNAs (e.g., microrna.org)suggest that RAF-1 may be an important target of miR-764, and may also promote angiogenesis [50,51].Thus, miR-764 may mediate angiogenesis through RAF-1, although further studies are also needed to evaluate this possibility.Moreover, our data indicate that miR-1343-3p and miR-764 expressions were inversely associated with vascular invasion and may play antiangiogenic roles, and while miR-27b-3p expression was positively associated with vascular invasion and may play a pro-angiogenic role.
ALK-rearranged cancers have distinct and aggressive clinicopathological characteristics, such as a young age at presentation, frequent nodal metastasis, high disease stage at the diagnosis, cribriform formation, presence of mucin-containing cells, presence of psammoma bodies, predominantly solid patterns, and frequent EMT [32].In our previous study, ALKrearranged tumors were closely related to the adjacent bronchial epithelium, especially the proximal bronchus, and tended to be centrally located [32].These distinctive features of ALK-rearranged cancers may be the result of pathologically activated ALK signaling, although the detailed underlying molecular mechanisms remain unclear.In addition, activated ALK signaling can influence the expressions of several important proteincoding genes that are involved in the JAK/STAT, PI3K/ AKT and RAS/ERK pathways [59], as well as non-coding RNAs, such as miR-150 and miR-155 [60,61].Given the complex roles of microRNAs as "fine tuners" of gene expression [62], it is conceivable that the microRNA expression profile of ALK-rearranged cancers might be at least partially different from EGFRand KRAS-driven cancers, which partly share signaling pathways.Consistent with this hypothesis, our findings revealed that ALK-rearranged cancer had distinctive microRNA expression profiles, while the EGFR-and KRAS-mutated cancers had similar microRNA expression profiles, despite having several differentially expressed microRNAs.It is possible that commonly activated RAS/ MAPK signaling pathways in these groups could largely influence their microRNA expression profiles, although further clarification is needed.
Another remarkable finding was the distinctive low expressions of miR-342-3p and let-7e in ALK-rearranged cancer, compared to EGFR/KRAS-mutated cancers.A previous study of a pancreatic cancer cell line revealed that the expression of the let-7 family (including let-7e) was decreased in gemcitabine-resistant cells with EMT features, such as fibroblastoid morphology and low E-cadherin expression [33].Another study of the A549 lung cancer cell line revealed reduced let-7 levels in side population cells (an enriched source of cancer stem cells), compared to non-side population cells [63].These findings suggest that let-7e down-regulation might be associated with stemness, chemoresistance, and EMT, which might contribute to the potentially aggressive tumor biology of ALK-rearranged lung cancer.Interestingly, a recent study suggested that the RAS-mitogen-activated protein kinase (MAPK) pathway was an important mechanism of resistance to ALK inhibitors in ALK-rearranged lung cancer [64].As let-7e and the let-7 family are regulators of RAS [28,65], it is possible that RAS activation is mediated by low let-7e levels.This provides an hypothetical mechanism whereby the RAS-MAPK axis is activated in ALK-rearranged tumor cells and allows them to resist suppression by ALK inhibitors.
ALK-rearranged lung cancer was also associated with low miR-342-3p expression and E-cadherin loss.In this context, E-cadherin is a cell-cell adhesion molecule and its loss is an important marker for the EMT process that allows cancer cells of epithelial origin to gain migratory and invasive properties [66].Furthermore, the significant association between ALK rearrangement, low miR-342-3p expression, and E-cadherin loss might contribute to the unique clinicopathological features of ALK-rearranged lung cancer [32].Several reports have suggested that miR-342-3p expression might be related to E-cadherin through DNMT1, which is a target mRNA for miR-342-3p, based on miRNA databases (e.g., microrna.org).Although DNMT1 is an essential DNA methyltransferase, it might affect E-cadherin expression through its direct and DNA methylation-independent interaction with SNAIL (an E-cadherin transcriptional repressor) [67].In non-small cell lung cancer, miR-342-3p has been suggested to play a tumor-suppressive role by negatively regulating RAP2B (a member of the RAS oncogene family) [68] and MYC transcriptional activity by targeting E2F1 (a MYC-cooperating molecule) [69].In this context, the simultaneous down-regulation of miR-342-3p and let-7e in ALK-rearranged lung cancer might help amplify the RAS family effector signals and/or MYC activity, which could result in resistance to anti-cancer drugs [64,70].Therefore, it might be useful to investigate miR-342-3p and let-7e mimics in combination with ALK inhibitors for ALKrearranged lung cancers, in order to prevent drug resistance and EMT-mediated metastasis.However, this approach requires further support from functional validation studies, pre-clinical studies, and clinical trials.
In conclusion, our results revealed that microRNA expression profiles had clinicopathological implications that were related to EGFR and KRAS mutations, as well as ALK-rearrangement.ALK-rearranged lung cancer had a unique microRNA expression profile that was distinct from that of EGFR-and KRAS-mutated lung cancers.In addition, ALK-rearranged lung cancer was associated with decreased miR-342-3p and let-7e expression, and the loss of E-cadherin.These characteristics may provide insight regarding the mechanisms for the relatively aggressive biology of ALK-rearranged lung cancer, as well as potential therapeutic strategies to overcome resistance to ALK inhibitors and prevent EMT-mediated metastasis.

Patients and samples
We retrieved 72 lung adenocarcinoma paraffinembedded specimens that had been surgically resected at Seoul National University Bundang Hospital between January 2004 and June 2011.Among these specimens, 19 were EGFR-mutated, 17 were KRAS-mutated, 16 were ALK-rearranged, and 20 were TN (negative for EGFR mutations, KRAS mutations, and ALK rearrangement).Hematoxylin and eosin-stained slides of all specimens were reviewed by two pulmonary pathologists (HK and JHC) and confirmed to be adenocarcinoma.All cases were classified according to the seventh edition of the Union for International Cancer Control/American Joint Committee on Cancer TNM classification [71].Clinicopathological data were obtained from the patients' medical records and pathology reports.This study's retrospective design was approved by the institutional review board at Seoul National University Bundang Hospital.

Detection of EGFR and KRAS mutations
Genomic DNA was extracted from paraffinembedded tissues.After deparaffinization using xylene, tissue sections were stained using hematoxylin and eosin, and the target lesions were selectively dissected to minimize contamination with normal tissue.Genomic DNA was isolated using the QIAamp DNA Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions.EGFR mutations at exons 18-21 and KRAS mutations at codons 12, 13, and 61 were analyzed using nested polymerase chain reaction (PCR) and direct DNA sequencing, as previously described [72].PCR products were processed using the Big Dye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, Foster, CA, USA), and the sequence data were generated using the ABI PRISM 3100 DNA Analyzer (Applied Biosystems).

Detection of ALK gene rearrangement
ALK rearrangement in formalin-fixed paraffinembedded (FFPE) tumor tissues was detected using fluorescence in situ hybridization and a break-apart probe that is specific for the ALK locus (Vysis LSI ALK dualcolor, break-apart rearrangement probe; Abbott Molecular, Abbott Park, IL, USA).Positive cases were defined as cases with >15% split signals or an isolated red signal in the tumor cells, as previously described [73][74][75].

RNA extraction from FFPE tissue samples
Total RNA was extracted as previously described [76].The RecoverAll Total Nucleic Acid Isolation Kit (Applied Biosystems) was used to isolate RNA from 10 mm-thick FFPE tissue sections.The concentrations were measured using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific Inc., Waltham, MA, USA), and then the RNAs were stored at 80°C until the time of use.

Construction of tissue microarrays
The tissue microarray blocks were constructed by Superbiochips Laboratories (Seoul, Korea) using the most representative areas of the paraffin blocks, as previously described [77].

Immunohistochemical analysis of E-cadherin expression
Immunohistochemistry was performed using tissue microarray sections.Four-micrometer-thick sections were transferred to poly-l-lysine-coated glass slides and incubated in a dry oven at 60°C for 1 h.The sections were then dewaxed using xylene (3 changes), rehydrated in a descending series of graded ethanol concentrations, and rinsed using Tris-buffered saline (pH 7.4).The endogenous peroxidase activity was blocked using 5% hydrogen peroxide in methanol for 15 min at 37ºC.For antigen retrieval, the slides were placed in citrate buffer (10% citrate buffer stock in distilled water, pH 6.0) and microwaved for 10 min.Non-specific staining was blocked using 1% horse serum in Tris-buffered saline (pH 7.4) for 3 min.The primary antibody for E-cadherin was diluted 1:100 (BD Biosciences, San Jose, CA), and the immunostaining was developed using an avidin-biotinperoxidase complex (Universal Elite ABC Kit; PK-6200; Vectastain, Burlingame, CA, USA) and diaminobenzidine tetrahydrochloride solution (HK153-5K; Biogenex, San Ramon, CA, USA).Positive controls (samples with known reactivity for the antibody) and negative controls (omission of the primary antibody) were included in each assay.Positive E-cadherin staining was defined as a cytoplasmic membranous pattern, and was classified as high (>90%), low (0-90%), or absent (0%) [78].

Statistical analysis
All data analyses were performed using SPSS software (version 21.0; SPSS Inc., Chicago, IL, USA).The associations between the clinicopathological parameters and the microRNA expression patters were evaluated using Pearson's R value, the chi-square test, and Fisher's exact test.The Mann-Whitney U test and Kruskal-Wallis test were used to compare the microRNA expression levels between the genetic groups.Hierarchical clustering analyses were performed using Ward's method, with intervals measured using the square Euclidian distance and standardized values (0 to 1).All tests were two-tailed, and differences with p-values of <0.05 were considered statistically significant.

Figure 1 :
Figure 1: Clustering and discriminant analysis using the microRNAs expression profiles in lung adenocarcinomas.A. A dendrogram using Ward's linkage and the expression profiles of 23 microRNAs revealed a trend towards two large groups: lung adenocarcinomas with EGFR or KRAS mutations (EGFR/KRAS-rich group) and lung adenocarcinomas with ALK rearrangement or triplenegative (TN) for these three genes (ALK/TN-rich group).B. Discriminant analysis of lung adenocarcinomas with EGFR and KRAS mutations and ALK rearrangement reveals relatively distinct grouping of the EGFR-and KRAS-mutated cancers, compared to the ALKrearranged and triple negative cancers (wild type for EGFR, KRAS and ALK).

Figure 2 :
Figure 2: Dot plots of microRNA expressions in lung adenocarcinomas according to genetic alterations.Dot plots A. and schematic illustration B. reveal differentially expressed microRNAs in ALK-rearranged cancers, compared to EGFR-or KRAS-mutated cancers.(n=52).

3p mean± SD p value mean± SD p value mean± SD p value mean± SD p value mean± SD p value mean± SD p value
(Continued) www.impactjournals.com/oncotarget