Background
Small cell lung carcinoma (SCLC) is the most aggressive subtype of all lung tumors [
1]. The poor survival rate of patients with SCLC is largely due to late detection and the lack of therapeutic regimens specifically targeted to SCLC [
2,
3]; thus, therapeutic improvement depends on a better understanding of the mechanisms underlying SCLC tumorigenesis and developing targeted therapy for this class of lung cancers. Although decades of work have led to better understanding of the genetic abnormalities in SCLC [
1,
4], these still cannot completely explain the aggressive phenotype that distinguishes it from other lung cancer subtypes. There is clearly an urgent need for continued efforts to understand SCLC tumorigenesis and to identify early diagnostic markers and therapeutic targets for SCLC.
A recently discovered class of small noncoding RNAs, microRNAs (miRNAs), regulates gene expression primarily by binding to sequences in the 3' untranslated region (3'UTR) of expressed mRNAs, resulting in decreased protein expression either by repression of translation or by enhancement of mRNA degradation. miRNAs have been shown to have a variety of regulatory functions and to play roles in controlling cancer initiation and progression [
5]. Many studies have demonstrated dysregulation of particular miRNAs in various cancer types and investigated the mechanisms of specific miRNAs in tumorigenesis [
5‐
7]. In the context of lung cancer, several studies have attempted to distinguish the miRNA profiles of histological subtypes showing the potential of miRNA profiles as diagnostic markers for distinguishing specific subtypes, such as squamous cell carcinoma and adenocarcinoma [
8,
9]. Moreover, tumor suppressor genes and oncogenes that play crucial roles in lung tumorigenesis have been demonstrated to be targets of miRNAs [
10‐
12], and manipulation of miRNA levels has been used to control lung cancer cell survival and proliferation
in vitro and
in vivo[
13‐
16]. Few studies, however, have focused on the role of miRNAs in the pathogenesis of SCLC [
17]. Primary tissue specimens are difficult to obtain as most SCLC tumors are not surgically resected [
4,
18], underscoring the importance of cell lines for studying this disease [
19,
20]. In order to characterize the expression of miRNAs in SCLC and explore the potential role of miRNAs in SCLC tumorigenesis, we profiled and compared the expression levels of a group of miRNAs in a set of lung cancer cell lines, including SCLC and non-small cell lung cancer (NSCLC) cell lines and immortalized human bronchial epithelial cells (HBECs).
Discussion
miRNAs have been intensively investigated as diagnostic markers in various cancers and cancer subtypes [
5,
6,
30]. However, few studies have specifically investigated the diagnostic value of miRNA profiles in SCLCs. In this study, we show that more miRNAs are differentially expressed between SCLC cell lines and HBECs than between NSCLC cell lines and HBECs; only two miRNAs were significantly differentially expressed between the NSCLCs and HBEC cell lines. The similarity between the HBEC and NSCLC miRNA profiles reflects the close histological relationship between HBECs and NSCLCs [
31‐
33]. On the other hand, the distinctive miRNA expression signature of SCLCs as compared with NSCLCs and HBECs suggests the possibility of developing miRNA profiling as a diagnostic tool for distinguishing SCLCs from NSCLCs and normal lung tissues. The development of miRNA profiling as a diagnostic tool could potentially benefit SCLC diagnosis from two perspectives. First, miRNA profiling would add a more quantitative aspect to the diagnosis of SCLC. Although the SCLCs share common genetic abnormalities and histological features and represent the most aggressive subtype of lung cancer in general, the survival and prognosis of SCLC patients diagnosed at the same stage vary [
34,
35], suggesting that quantitative molecular traits are related to the degree of malignancy. However, current diagnosis of SCLC is primarily determined histologically [
36], which is not sufficient to quantitatively evaluate malignancy and prognosis. Several studies have shown that miRNA expression levels are related to cancer prognosis [
37‐
40]. Similarly, the quantification of aberrant expression levels of miRNAs in SCLCs may serve as a reliable tool for the prediction of SCLC prognosis. Second, the miRNAs identified as over-expressed in SCLCs may serve as early and non-invasive detection markers. Recent findings have shown that miRNAs are secreted into blood and are detectable in serum, showing potential as non-invasive markers for diseases [
41,
42]. Inexpensive, non-invasive detection methods are suitable for the development of large-scale screening of high-risk populations and may therefore significantly advance the early diagnosis of cancers. Given the aggressive nature of most SCLCs, the development of highly sensitive and specific non-invasive molecular diagnostics based on miRNA profiling could be of great clinical benefit. Overall, the miRNAs identified as differentially expressed in SCLC compared to NSCLC and normal cells hold promise as early, noninvasive and quantitative markers of SCLCs and warrant further investigation.
Our results suggest that miRNAs may play an important role in the pathogenesis of SCLCs. Although there is evidence to support NSCLCs as originating from HBECs [
31‐
33], the findings on the histological origin of SCLC remain somewhat controversial [
43‐
45]. Previous studies suggest that a transition between NSCLC and SCLC can occur during lung tumor progression and that neuroendocrine differentiation of NSCLCs, which has been postulated to be an intermediate step between NSCLC and SCLC, is related to poor prognosis and early metastasis [
46‐
48]. However, the mechanisms involved in this transition between the two subtypes are not completely understood. Our results show that of the 41 miRNAs that are differentially expressed between the three groups of cell lines, 34 (83%) show a trend of progressive differential expression from HBECs to NSCLCs to SCLCs (Table
2). These results support the hypothesis that differential expression of miRNAs could contribute to the differentiation of lung cancer cells from one subtype to another, in which SCLC could result from NSCLC cells by gradually acquiring SCLC properties through the cumulative dysregulation of miRNAs, and that manipulating the levels of specific miRNAs levels might prevent the differentiation of lung cancer cells toward a more malignant phenotype.
Changes in miRNA expression can lead to tumorigenesis, but the many complex interactions between miRNAs and their targets that occur during these processes are not fully understood. A variety of studies have linked miRNA dysregulation with malignant transformation [
49]. However, the role of miRNAs in SCLC pathogenesis has not been extensively studied. Our investigation identified a group of miRNAs that show a progressive differential expression from HBECs to NSCLC and SCLC cells. Several of the miRNAs identified in this study have been shown to be associated with various cancer types in previous studies. For example, we found significant overexpression of miR-103, miR-107, miR-301 and miR-338 in lung cancer cells as compared to HBECs. These miRNAs have been shown to be over-expressed in several types of cancers including lung cancers [
17,
50,
51], and high expression of miR-103 and miR-107 were correlated with poor survival in cancer patients (esophageal squamous and pancreatic tumors) [
51,
52]. These miRNAs might contribute to common pathways during the transformation of normal cells to tumor cells during lung cancer pathogenesis, and the greater extent of aberrant expression of these miRNAs in SCLCs relative to NSCLCs might contribute to the more aggressive phenotype of the former.
Our study also identified a group of miRNAs that might contribute to the establishment of SCLC features and the specific phenotypes that differentiate SCLC from NSCLC. For example, we found over-expression of miR-17-5p in SCLCs compared to NSCLCs. This miRNA was recently shown to target Rbl2, a member of the Rb family [
53]. Rb is a tumor suppressor that induces arrest of the cell cycle at G1 [
54]. SCLCs have been shown to exhibit loss of Rb expression in 87-100% of tumors compared to less than 15% in NSCLC [
55‐
57]. SCLC cells were also previously shown to be addicted to continued over-expression of miR-17-5p [
58], and forced over-expression of the miRNA cluster that includes miR-17-5p (miR-17-92) was shown to induce embryonic lung epithelial cell proliferation [
59]. Coupled with these data, our results suggest that dysregulation of this miRNA could be an important distinction that defines the pathogenesis and phenotypic characteristics of SCLC compared to NSCLC. We also observed a significant increase in miR-135 expression in SCLC cells compared to NSCLC cells. miR-135 has recently been shown to inhibit expression of the tumor suppressor gene Adenomatous Polyposis Coli (APC) in colorectal cancer [
60]. Loss of heterozygosity of APC has been shown in both small cell and non-small cell lung cancers, but appears to be more frequent in SCLC [
61]. Silencing of this gene by CpG hypermethylation, however, is more frequent in NSCLC compared to SCLC [
62], suggesting that various lung tumor subtypes could use different means to down-regulate this tumor suppressor. These findings suggest that SCLC preferentially utilizes microRNA-based regulatory mechanisms to reduce APC expression. miR-29a, -29b and -29c expression was shown be significantly down-regulated in SCLC cells compared to HBECs, whereas these reductions were not seen in NSCLC cells. Expression levels of miR-29a, miR-29b, and miR-29c were previously shown to be inversely correlated with levels of DNA methyltransferase (DNMT) -3A and -3B [
14], two key enzymes involved in DNA methylation that have been shown to promote tumorigenesis [
63]. Forced expression of these miRNAs also inhibited tumorigenicity
in vitro and
in vivo[
14]. In SCLC cells, but not NSCLC cells, we also observed significant reductions in miR-24, inhibition of which was previously shown to enhance cell proliferation [
64]. These miRNAs might contribute to the specific pathogenesis pathways during the transformation of SCLCs but not NSCLCs.
Several miRNAs identified in our study exhibited expression levels not consistent with previous observations in other cancer types, suggesting contextual dependence of miRNA function in the regulation of tumorigenesis pathways. For example, we observed significantly increased levels of miR-148b in SCLC compared to HBECs; miR-148b has been shown to target DNMT3B [
65], with down-regulation of miR-148b observed in metastatic cancers [
66]. miR-21, miR-221 and miR-222, which have been shown to be oncogenic miRNAs and up-regulated in certain lung cancer subtypes [
67,
68], are significantly down-regulated in SCLC. We speculate that these miRNAs may not be the primary driving force for controlling SCLC cell proliferation and survival. Given the large number of miRNAs that are found aberrantly expressed in SCLCs, it is possible that some of these miRNAs play crucial roles in pathogenesis of SCLC. The oncogenic pathways up-regulated by these miRNAs might lead to feedback up-regulation of certain tumor suppressor miRNAs and down-regulation of certain oncogenic miRNAs. Further studies are certainly needed to address this question. We also observed up-regulation of miR-142-3p in SCLC compared to HBECs, although a previous report showed significant repression of this miRNA in lung adenocarcinomas versus normal tissue [
69]. Another study showed down-regulation of this miRNA early in tumor development followed by increased expression at the later stages of lung tumorigenesis [
70]. Expression levels of this miRNA could therefore vary both with lung tumor subtype and stage of tumor development. miR-1 has also been shown to be expressed at lower levels in lung cancer cell lines, including both NSCLC and SCLC, than in bronchial epithelial cells [
71], whereas our results show significant over-expression of miR-1 in lung cancer cell lines compared to HBECs. However, given the extremely low expression levels observed in both the normal bronchial epithelial cells and lung cancer cells in our study, and in normal lung tissues in other studies [
71,
72], the aberrant expression of miR-1 in lung cancers relative to normal lung cells needs to be evaluated further.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
JDM and AFG derived the cell lines, LG isolated the RNA, SMH ran the arrays, and JJS and I performed data analysis. LD and AP designed the study, analyzed the data and wrote the manuscript. All authors read and approved the final manuscript.