Screening of the entire let-7 family of microRNAs (miRNA) by in situ hybridization identified let-7g as the only member, the diminished expression of which was significantly associated with lymph node metastasis and poor survival in breast cancer patients. Abrogation of let-7g expression in otherwise nonmetastatic mammary carcinoma cells elicited rapid metastasis from the orthotopic location, through preferential targets, Grb2-associated binding protein 2 (GAB2) and fibronectin 1 (FN1), and consequent activation of p44/42 mitogen-activated protein kinase (MAPK) and specific matrix metalloproteinases. Treatment with estrogen or epidermal growth factor specifically reduced the expression of mature let-7g through activation of p44/42 MAPK and subsequently stimulated expression of GAB2 and FN1, which, in turn, promoted tumor invasion. We thus identify let-7g as a unique member of the let-7 miRNA family that can serve as a prognostic biomarker in breast cancer and also propose a paradigm used by specific signaling molecules via let-7g to cooperatively promote breast cancer invasion and metastasis. Thus, let-7 family members neither possess equivalent clinicopathologic correlation nor function in breast cancer. Cancer Res; 71(20); 6463–74. ©2011 AACR.
microRNAs (miRNA) are small noncoding RNAs ranging in size by 20 to 25 nucleotides. miRNAs posttranscriptionally repress gene expression mostly by recognizing complementary target sites in the 3′-untranslated region (3′-UTR) of target mRNAs (1–3). miRNAs are involved in the regulation of a continuous biological processes leading to the acquisition of metastatic potential, such as aberrant adhesion, migration, and invasion, and neoangiogenesis (2–4). A limited number of miRNAs, either upregulated (such as miR-10b, miR-21, miR-373, and miR-520c) or downregulated (such as miR-98, let-7a, miR-31, and miR-146), have thus far been identified to play a role in cancer metastasis (5).
The human let-7 miRNA family consists of 13 members located in 8 genomic locations frequently deleted in human cancers (6). Nine distinct mature let-7 miRNAs with identical seed sequences are produced from 12 precursor sequences from miRBase (7). Aberrant expression of let-7 miRNA has been associated with poor prognosis of several types of cancer (8–11). For example, expression of let-7 miRNAs is reduced in non–small-cell lung cancer patients and associated with poor prognosis (12), and increased expression of let-7a substantially reduces tumor burden in a K-Ras murine lung cancer model (13). While let-7 is widely viewed as a tumor suppressor miRNA, upregulation of certain let-7 family members has also been observed (6), although less frequently. Although let-7 family members were generally believed to have overlapping targets and therefore redundant roles, whether individual members of the let-7 family possess specific roles in cancer development and metastasis is largely undefined (6, 14).
In this report, amongst all mature let-7 miRNAs in the let-7 family, let-7g was identified as the only one, reduced expression of which correlates with worse survival outcome in breast cancer patients. Grb2-associated binding protein 2 (GAB2) and fibronectin 1 (FN1) were identified as the novel targets of let-7g that increased activity of p44/42 mitogen-activated protein kinase (MAPK) and matrix metalloproteinase (MMP)-2/MMP-9. Estrogen and epidermal growth factor (EGF) specifically reduced let-7g expression and increased expression of GAB2 and FN1. We therefore propose a model whereby estrogen and EGF promote breast cancer metastasis through several feedback loops via the specific interplays between let-7g and its downstream cascades.
Materials and Methods
Cell lines and cell culture
All human breast cancer cell lines used in this study were obtained from the American Type Culture Collection and cultured in conditions as recommended. All cells were maintained in a humidified incubator at 37°C and 5% CO2.
Detection of miRNA expression by in situ hybridization in formalin-fixed paraffin-embedded tissues
For detecting expression of let-7g family miRNAs in invasive mammary carcinoma and normal mammary tissues, in situ hybridization was carried out as previously described (15) with modifications (16). Briefly, 3 μm thick tissue microarray sections were deparaffinized, rehydrated, and then digested and refixed in 4% paraformaldehyde. Sections were then replaced with hybridization solution and incubated with locked nucleic acid (LNA)-modified probes (Exiqon) for let-7 miRNAs, U6 (positive control), and scrambled RNA (negative control) at 60°C for 20 hours. The slides were then incubated with mouse anti-digoxin antibody followed by binding to streptavidin-biotin-peroxidase complex solution, and the sections were stained with 3,3′-diaminobenzidine solution and counterstained with hematoxylin solution. The stained sections were reviewed and scored for expression of let-7 miRNAs under the microscopy (Olympus). The sections were scored on the basis of the intensity and the percentage of stained cells.
Cultured cell lines were used for miRNA extraction using the mirVana miRNA Isolation Kit (Ambion). miRNA fraction was then converted to cDNA using SuperScript III Reverse Transcriptase (Invitrogen). All available TaqMan miRNA assays were used for expression analysis as indicated. For mRNA analysis, cultured cells were used for extraction of mRNA using TRIzol Plus RNA Purification system. SYBR Premix Ex Taq Kit (Takara) was used to determine the expression levels of GAB2, FN1, and other analyzed genes (Supplementary Table S5). The relative amount of gene transcripts was normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH).
Activity of MMP-2 and MMP-9 was measured by the method of gelatin zymography as previously described (17) with several modifications. Conditioned media were obtained by incubation of MCF cells with serum-free medium for 24 hours and were then concentrated 80-fold using Amicon Ultra Centrifugal Filter Units (Millipore) and normalized by protein concentrations. Samples were loaded on 10% SDS-PAGE gels containing 0.1% gelatin. Electrophoresis was carried out under nonreducing conditions at 100 V and 4°C. Gels were washed in 2.5% Triton X-100, incubated in substrate buffer (50 mmol/L Tris-HCl, pH 8.0, 50 mmol/L NaCl, 10 mmol/L CaCl2, and 0.05% Brij 35) for 40 hours at 37°C, stained with Coomassie stain solution (Bio-Rad), and destained in 20% methanol and 10% acetic acid. Gelatinolytic activity was identified as a clear band on a blue background.
Reduced let-7g expression is prognostic for poor survival of breast cancer patients
To specifically define let-7 miRNA species involved in mammary carcinoma progression, we used digoxigenin-labeled LNA–miRNA probes to detect abundance of 9 mature let-7 miRNAs in archived breast cancer specimens (n = 86) and breast tissue specimens from patients with benign breast diseases (n = 21) using in situ hybridization (15, 16). The expression of let-7 miRNAs were detected predominantly within the cytoplasm of the benign luminal epithelial cells or carcinoma cells (Fig. 1A, Supplementary Fig. S1A). Of all the let-7 miRNAs examined and graded double blindly, we observed significantly diminished expression of only let-7a, let-7c, and let-7g (Fig. 1A) in breast cancer compared with normal specimens (P = 0.031, Supplementary Table S1). For direct comparison, detection of the different let-7 miRNAs on consecutive sections of one sample of invasive ductal carcinoma is provided in Supplementary Fig. S1A. A heat map displaying the expression of all let-7 miRNAs in normal and cancer tissues is provided in Supplementary Fig. S1G. Heat maps displaying expression of let-7 miRNAs according to estrogen receptor, progesterone receptor, or HER2 status are also provided in Supplementary Fig. S1G. The association between the expression of all let-7 family members and clinicopathologic characteristics of breast cancer and these data are summarized in Supplementary Table S2. A statistically significant association was observed between lower expression of let-7a and tumor size (P = 0.041) and Ki-67 labeling index (P = 0.036), between lower expression of let-7b and Ki-67 labeling index (P = 0.036), between lower expression of let-7c and progesterone receptor-negative status (P = 0.011) and HER2-negative status (P = 0.040), between lower expression of let-7g and clinical stage (P = 0.017), lymph node metastasis (P < 0.0001), and Ki-67 labeling index (P = 0.001), and between lower expression of let-7i with lymph node metastasis (P = 0.040), progesterone receptor-negative status (P = 0.021), and HER2-negative status (P = 0.030).
We next examined the correlation between the expression of let-7 miRNAs and survival in the breast cancer patient cohort with 5-year follow-up (n = 80), using Kaplan–Meier survival analyses. Patients with lower expression levels of let-7g exhibited a significantly worse survival outcome than those with higher expression level of let-7g (Fig. 1B, Supplementary Table S5; Fig. 1C, Supplementary Table S5). Recurrence-free survival and overall survival outcomes according to estrogen receptor, progesterone receptor, or HER2 status are provided in Supplementary Tables S3 and S4. Again, let-7g was the only let-7 family member whose reduced expression was associated with decreased survival in the patient subgroups. In contrast, the expression levels of none of the remaining 8 miRNAs could predict survival outcome in breast cancer patients (Supplementary Table S5). We further showed that the reduced expression level of let-7g was a significant predicator for patient survival independent of clinical stage and lymph node metastasis in breast cancer by multivariate analyses. We therefore further concentrated on let-7g to gain insight on its functional role and mechanism. For comparison of speci-ficity, let-7d and let-7f were included in parallel in further studies.
We next examined let-7g expression in fresh mammary carcinoma specimens by quantitative real-time PCR analysis using stem-looped miRNA-specific reverse transcription primers and TaqMan probes. The expression level of let-7g was significantly lower in breast cancer patients with lymph node metastases than those from patients with no detectable lymph node metastases (Fig. 1D). However, the expression level of neither let-7d nor let-7f miRNA exhibited significant differences between the breast cancer patient cohorts with or without lymph node metastases (Supplementary Fig. S1B and C).
We also examined the relative expression of let-7g in benign breast tissue, premalignant proliferative lesions, and mammary carcinoma specimens by in situ hybridization (Fig. 1E). Interestingly, progressively reduced expression of let-7g was observed from normal breast epithelium to ductal carcinoma in situ to invasive ductal carcinoma, supporting a role of reduced expression of let-7g in both the initiation and progression of breast cancer. Finally, we examined the relative expression of let-7g in an array of mammary epithelial and mammary carcinoma cell lines using quantitative real-time PCR (Supplementary Fig. S1D). The highest expression levels of let-7g were observed in 2 immortalized but otherwise normal human mammary epithelial cells, whereas low or undetectable levels of let-7g expression were observed in highly invasive mammary carcinoma cell lines (such as MDA-MB-231 and MDA-MB-468). In comparison, expression levels of let-7d (Supplementary Fig. S1E) and let-7f (Supplementary Fig. S1F) were not significantly different between the low-invasive and high-invasive mammary carcinoma cell lines.
let-7g depletion promotes mammary carcinoma cell migration and invasion in vitro
Given the inverse correlation between let-7g levels and invasiveness in mammary carcinoma cells, we next assessed the potential role of let-7g in mammary carcinoma cell motility through manipulation of the expression level of let-7g by transfection of either antisense oligonucleotide (ASO) or synthetic miRNA mimic (Supplementary Fig. S2A and B). let-7g ASO exhibited a high specificity to deplete its own target sequence but did not appreciably deplete other let-7 miRNAs by using quantitative real-time PCR analysis (Fig. 2A).
let-7g depletion in low-invasive MCF-7 cells led to a more rapid closing of the wound (Supplementary Fig. S2C) and a potent increase in cell migration and invasion than the control cells (Fig. 2B). Conversely, forced expression of let-7g in the highly invasive MDA-MB-231 cells resulted in retarded wound closing (Supplementary Fig. S2D) and significant reduction in cell migration and invasion (Fig. 2C). The same experimental approach was adopted in 2 other mammary carcinoma cell lines (low-invasive T47D and highly invasive MDA-MB-468; Supplementary Fig. S2E–H).
let-7g depletion initiates distant metastasis in vivo
We further determined whether let-7g expression depletion would initiate invasion and metastasis of otherwise nonmetastatic MCF-7 cells in vivo. To this end, MCF-7-luc cells were infected with either lentivirus-expressing let-7g ASOs or control oligonucleotides (Supplementary Fig. S2I) and injected orthotopically into the mammary fat pad of female BALB/c nude mice. Exposed bioluminescent imaging of lung and liver was used to prevent anatomical superimposition of the primary tumor sites. Primary tumors derived from let-7g–depleted cells were poorly encapsulated and highly invasive (Fig. 2D). Tumor emboli (red arrow) were observed in lymphatic vessels. In contrast, tumors derived from control cells remained well-confined and noninvasive. Interestingly, in the primary tumors, cells of tumor emboli within tumor vasculature exhibited no expression of let-7g (Fig. 2E, red arrow), further indicative that let-7g depletion potently promotes breast cancer invasion and metastasis.
Orthotopic injection of let-7g–depleted MCF-7-luc cells resulted in the formation of pulmonary (8 of 8, P = 0.014) and hepatic (7 of 8, P = 0.021) metastases in host mice observed by bioluminescent imaging 5 weeks after injection (Fig. 2F and G, Supplementary Fig. S2K). No metastases were detected in the mice with orthotopic transplanted with control cells. In addition, metastatic tumors were readily detectable in the lungs of all mice (7 of 7) injected with let-7g–depleted MCF-7-luc cells by bioluminescent imaging 4 weeks post–tail vein injection, whereas no metastases were detectable in the lungs of host mice (0 of 7) injected with the control cells (Fig. 2H and I, Supplementary Fig. S2L). It was interesting to note that bone metastases were only observed in the mice injected with let-7g–depleted cells (Supplementary Fig. S2L).
Concordantly, forced expression of let-7g by infection with lentivirus-expressing let-7g was sufficient to prevent the pulmonary macrometastases of highly invasive MDA-MB-231 cells (0 of 8, P = 0.015) in a xenograft model by tail vein injection (Supplementary Fig. S2M–O).
GAB2 and FN1 are targeted by let-7g to regulate mammary carcinoma cell migration and invasion
To identify downstream targets of let-7g, 2 sets of paired cells were used to examine the mRNA expression profile of an array of genes that have been reported to be involved in cancer invasion and metastasis (18, 19). MDA-MB-231 cells transfected with either let-7g mimic or control and MCF-7 cells transfected with either let-7g ASO or scrambled sequence control were used. The data derived from the 2 complementary pairs of screening systems were largely concordant (Supplementary Table S6).
To further identify the putative direct targets of let-7g, we used 3 algorithms for mRNA target prediction—MiRanda, RNAhybrid, and TargetScan (20–22). GAB2 and FN1 were of particular interest among the candidate genes, as both have been implicated in breast cancer metastasis in animal models and the clinical setting (23–29). Both the GAB2-encoded mRNA and FN1-encoded mRNA contain a 3′-UTR element that is partially complementary to let-7g and carries the identical sequence in the multiple mammalian mRNA orthologues (Supplementary Fig. S3A–C).
Forced expression of let-7g reduced the activity of a luciferase reporter gene containing the 3′-UTR of GAB2 (Fig. 3A), indicating that let-7g directly targets GAB2. We further identified that a single let-7g cognate binding site, LCS2, within the GAB2 3′-UTR, was the major target site for let-7g and thereby sufficient to regulate GAB2 expression (Fig. 3A, Supplementary Fig. S3B). Consistently, forced expression of let-7g diminished GAB2 expression at both mRNA and protein levels (Fig. 3B), indicating GAB2 is a bona fide target of let-7g.
We further determined whether GAB2 was required for let-7g–mediated mammary carcinoma cell motility. Specific depletion of GAB2 by siRNAs significantly abrogated let-7g depletion–enhanced MCF-7 cell motility (Fig. 3C and D, Supplementary Fig. S3E and F). The functional importance of the interaction between let-7g and GAB2 3′-UTR was also determined by using either full-length GAB2 (open reading frame + 3′-UTR) with or without mutation at LCS2. Thus, GAB2 is a critical mediator for the let-7g depletion–enhanced cell migration and invasion (Fig. 3E).
Similarly, we identified FN1 as a direct target of let-7g and mapped the major let-7g binding site within the FN1 3′-UTR (Fig. 3F and G, Supplementary Fig. S3C and D). We showed that siRNA-mediated depletion of FN1 expression significantly abrogated the enhanced migration and invasion consequent to let-7g depletion (Fig. 3H and I). Thus, both GAB2 and FN1 play important roles downstream of let-7g.
p44/42 MAPK and MMPs are essential for let-7g–mediated mammary carcinoma cell migration and invasion via GAB2 and FN1
We further investigated the downstream signaling pathway used by GAB2 to mediate let-7g effects on cell motility. Increased expression of GAB2 has been shown to play an important role in promoting mammary carcinoma metastasis through activation of the p44/42 MAPK pathway (23, 25). Elevated GAB2 expression and p44/42 MAPK activity were observed in let-7g–depleted MCF-7 cells compared with the control oligonucleotide–transfected cells. Moreover, this effect was substantially abrogated by GAB2-specific siRNA (Fig. 4A), indicating that let-7g depletion directly increased GAB2 expression to promote p44/42 MAPK activation.
To define the p44/42 MAPK downstream molecules that mediate the effect of let-7g, we focused on the gelatinases MMP-2 and MMP-9, which are regulated by the p44/42 MAPK pathway (30–33), as well as possessing the capacity to degrade type IV collagen and mediate tumor cell invasion (34, 35). Indeed, let-7g depletion potently increased the expression and activity of MMP-2/MMP-9 (Fig. 4B and C). Furthermore, abrogation of GAB2 expression by specific siRNAs diminished the expression and enzymatic activity of MMP-2/MMP-9 consequent to let-7g depletion (Fig. 4B and C). Similarly, we showed that FN1 expression was required for the increased MMP-2/MMP-9 expression and invasion consequent to let-7g depletion (Supplementary Fig. S3G). To further show that MAPK activity was required for the increased MMP-2/MMP-9 activation, let-7g–depleted MCF-7 cells were treated with either the MAP/ERK kinase 1/2 (MEK1/2)-specific inhibitor U0126 or MMP-2/MMP-9–specific inhibitor SB-3CT. Inhibition of MEK1 activity was shown to be sufficient to diminish MMP-2/MMP-9 activity and let-7g depletion–promoted cell motility (Fig. 4D and E). It is apparent that the p44/42 MAPK–MMP-2/MMP-9 pathway was used to enhance cell migration and invasion consequent to let-7g depletion.
GAB2 and FN1 are preferentially targeted by let-7g for p44/42 MAPK activation
We next sought to determine why let-7g depletion alone should possess such striking biological effects and its cellular context in breast cancer metastasis. To answer both questions, we sought to determine possible differential effects of let-7g depletion on cell behavior compared with the depletion of other let-7 family members. Both let-7d and let-7f ASOs were included as controls, and the specificity of each ASO for the 3 let-7 miRNAs was first determined by quantitative real-time PCR (Fig. 5A). Both let-7d and let-7f were chosen as controls for full functional analyses, as let-7d with a 4-nucleotide difference to let-7g would represent a distant structural control and let-7f with only a 2-nucleotide difference would represent a closer structural control. Functionally, let-7g depletion was more potent to promote the migration and invasion of MCF-7 cells than the depletion of either let-7d or let-7f (Fig. 5B). Thus, let-7g depletion exhibited significantly higher efficacy to promote cancer cell migration and invasion.
Mechanistically, let-7g more potently inhibited the activity of 3′-UTR–luciferase constructs for GAB2 or FN1 genes than let-7d or let-7f in 2 types of cells (Fig. 5C, Supplementary Fig. S4H). Consistently, let-7g ASO also more potently increased the activity of 3′-UTR–luciferase construct for GAB2 or FN1 genes than let-7d or let-7f (Fig. 5D). The effect of forced expression of all of the different let-7 family members on the 3′-UTR of GAB2 and FN1 is presented in Supplementary Fig. S3H. Concordantly, we observed that the expression level of let-7g but not let-7d or let-7f negatively correlated with GAB2 or FN1 mRNA level in an array of breast epithelial cell lines (Supplementary Fig. S4B–G). GAB2 and FN1 were therefore preferentially targeted by let-7g in mammary carcinoma cells. Consistently, we showed that let-7g depletion, but not let-7d or let-7f depletion, potently increased the expression of GAB2 or FN1 and significantly activated p44/42 MAPK (Fig. 5E). We therefore propose that increased GAB2 and FN1 consequent to let-7 depletion represents an integrated modulation of the p44/42 MAPK pathway specific for let-7g but not for the other let-7 miRNAs examined. The functional specificity of let-7g shown herein was highly consistent with our clinical studies that only decreased expression of let-7g significantly correlated with proliferative status, lymph node metastasis, and prognosis of breast cancer patients (Supplementary Tables S2 and S5).
Estrogen and EGF modulate the expression of let-7g and GAB2
We further sought to determine whether etiologic factors stimulatory of breast cancer metastasis, such as estrogen and epidermal growth factor receptor (EGFR)/HER2 ligands, might contribute to promotion of mammary carcinoma cell invasion through regulation of let-7g expression.
Treatment of MCF-7 cells with 17β-estradiol (E2) resulted in rapid and specific reduced let-7g expression (Fig. 6A and B, Supplementary Fig. S5A and B). E2 increased GAB2 expression at both the mRNA and protein levels (Fig. 6B and C). The specificity of E2 treatment via ERα was verified by using fulvestrant (Fig. 6D). Transfection of the let-7g mimic substantially abrogated the effect of E2 on GAB2 expression, indicating that E2 stimulated GAB2 expression via the interaction between let-7g and GAB2 3′-UTR (Fig. 6E). We further showed that estrogen promoted migration and invasion of MCF-7 (Fig. 6F) and T47D cells (Supplementary Fig. S5C) via let-7g depletion and elevated GAB2 expression.
Similarly, the effects of EGF on let-7g and GAB2 expression and cell motility were also detected. EGF promoted mammary carcinoma cell migration and invasion (Fig. 6I, Supplementary Fig. S5D) through let-7g depletion and consequently increased GAB2 expression (Fig. 6G–L) via EGFR/HER2 (Fig. 6J). Thus, both E2 and EGF reduced let-7g expression and consequently increased GAB2 expression, leading to enhancement of cell migration and invasion.
Similarly, we showed that FN1 expression was increased by E2 and EGF via their repression of let-7g expression (Supplementary Fig. S5E–H). Increased FN1 expression was also required for E2 and EGF stimulation promoted mammary carcinoma cell migration and invasion via let-7g repression (Supplementary Fig. S5I and J).
We also determined what pathways might mediate E2- and EGF-promoted let-7g depletion. It has been recently reported that proliferative signals can promote let-7 family downregulation via the p44/42 MAPK pathway (36). Serum-deprived MCF-7 cells were therefore pretreated with U0126 and followed by E2 treatment. Interestingly, neither estrogen nor EGF stimulation with or without U0126 affected the expression of primary let-7g (Supplementary Fig. S5K), despite that both ERα and c-Myc binding sites at the let-7g promoter have been either identified or predicted (Supplementary Fig. S5L), indicative that reduction of mature let-7g expression by estrogen or EGF stimulation may be achieved through a posttranscriptional mechanism. In contrast, U0126 significantly abrogated E2- or EGF-induced repression of mature let-7g (Fig. 6M). Thus, E2 or EGF stimulation specifically decreased functional let-7g expression via p44/42 MAPK pathway.
The aberrant expression of miRNAs in cancer progression has previously been determined mostly by microarray profiling, bead-based technologies, and quantitative real-time PCR in either fresh or archived formalin-fixed paraffin-embedded tumor specimens or blood samples. However, these approaches are inevitably confounded by the heterogeneity of the biospecimens, especially the cell population of the tumor microenvironment including stromal and inflammatory cells. This is probably one explanation for the substantial conflicting data in the field (37). For example, miR-10b has been reported prometastatic (38, 39) or antimetastatic (40, 41) in breast cancer independently by multiple groups. In this study, the expression of the entire let-7 family miRNAs was examined in clinical specimens by in situ hybridization with clinicopathologic analysis. Strikingly, reduced expression of let-7g in mammary carcinoma was uniquely linked with tumor metastasis and poor patient survival. The functional significance of this negative correlation was further supported by in vitro and in vivo studies (Fig. 2). We have therefore identified a unique member of the let-7 family as an independent prognostic marker for breast cancer.
let7g seems to be specifically involved in breast cancer metastasis for 2 reasons. First, pathologic and prognostic analyses of survival outcome indicated that let-7g was the only let-7 miRNA family member to exhibit a significant correlation of diminished expression with poor survival in breast cancer patients. Second, of the all the mature let-7 miRNAs examined, let-7g showed the highest efficacy to diminish the reporter activities of the 3′-UTR of GAB2 and FN1. Although our data indicated that let-7e, let-7i, or miRNA-98 also significantly reduced the reporter activities of 3′-UTR of GAB2 and/or FN1, however, none of these miRNAs exhibited reduced expression in clinical specimens, indicating that these in vitro molecular events do not possess any pathophysiologic relevance. Moreover, even though reduced expression of let-7a and let-7c was observed in breast cancer tissue, neither of these miRNAs could efficiently target the 3′-UTR of either GAB2 or FN1. Therefore, our results show that amongst all let-7 family members, only let-7g fulfilled 2 critical criteria at the same time: (i) reduced expression in breast cancer and an expression level that inversely correlated with metastasis and survival and (ii) efficacious targeting of the 3′-UTR of both GAB2 and FN1 to promote tumor invasion and metastasis. Amongst let-7 family members, diminished expression of let-7g is therefore specifically involved in breast cancer metastasis. Our results do not exclude possible roles of other let-7 family members in promoting tumor growth or proliferation. For example, reduced expression of let-7a in breast cancer was associated with larger tumor size and higher proliferative status, indicative that reduced let-7a expression may contribute to tumor growth. That let-7a does not influence survival outcome of breast cancer patients can easily be understood, in that death from breast cancer is primarily due to complications of metastasis (42), a process that let-7a is not correlated with clinically nor mechanistically. We have therefore garnered both clinical and mechanistic evidence to propose a paradigm, used by specific growth factors upstream and mediated by specific signaling molecules downstream, that let-7g is pivotal to promote breast cancer invasion and metastasis.
It is of interest to note that let-7g has also been reported to play a unique role among the let-7 family members in the irradiation response of lung cancer (43). Contrary to other let-7 miRNAs, only let-7g exhibited an increased expression in response to γ-irradiation in 3 lung cancer cell lines. However, the mechanism whereby let-7g played the unique role in lung cancer was not determined.
In this study, we showed that E2 and EGF, major prometastatic growth factors in breast cancer (44–46), specifically repress let-7g expression and subsequently increase GAB2 and FN1 expression, which, in turn, cooperatively increased the activity of p44/42 MAPK and MMP-2/MMP-9. It has been reported recently that the p44/42 MAPK cascade may repress let-7 expression either via activating LIN28 transcription to affect primary let-7g processing (47) or via phosphorylating TAR RNA–binding protein, a critical component of miRNA-generating complex, to suppress the expression of let-7 miRNAs (36). Our study thus adds another dimension to the signaling network, that reduced expression of let-7g specifically increased p44/42 MAPK activity, which, in turn, further repressed let-7 expression, generating a positive feedback loop to potentially enhance the biological processes such as metastasis and proliferation (Figs. 4–6, Supplementary Tables S2 and S4). Our results have thus identified a feedback loop (E2/EGF-p44/42 MAPK-let-7g-GAB2/FN1-p44/42 MAPK) used by let-7g to integrate multiple signaling molecules for the facilitation of mammary carcinoma metastasis (Supplementary Fig. S6). We propose that the critical growth factors for breast cancer (estrogen and EGF) might specifically deplete let-7g expression and further trigger synergistic interaction between let-7g and p44/42 MAPK to promote mammary carcinoma cell invasion and metastasis via positive feedback loops.
It should be noted that let-7g should be expected to possess a wide range of functionalities due to its pleiotrophic regulation of genes. Further studies will be required to fully understand its functional roles and interactions.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
This work was supported by the National Key Scientific Program of China (2010CB912804, 2007CB914801, 2012CB934002, and 2011CBA01103), the National NSF of China (30971492, 30725015, and 30873047), the Anhui NSF (090413090 and KJ2009A162), the Fundamental Research Funds for the Central Universities (WK2070000008), the CAS Visiting Professorship for Senior International Scientists (2010T2S03), and the Cancer Science Institute of Singapore.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).
- Received April 19, 2011.
- Revision received July 19, 2011.
- Accepted August 4, 2011.
- ©2011 American Association for Cancer Research.