Genomic DNA Copy-Number Alterations of the let-7 Family in Human Cancers

In human cancer, expression of the let-7 family is significantly reduced, and this is associated with shorter survival times in patients. However, the mechanisms leading to let-7 downregulation in cancer are still largely unclear. Since an alteration in copy-number is one of the causes of gene deregulation in cancer, we examined copy number alterations of the let-7 family in 2,969 cancer specimens from a high-resolution SNP array dataset. We found that there was a reduction in the copy number of let-7 genes in a cancer-type specific manner. Importantly, focal deletion of four let-7 family members was found in three cancer types: medulloblastoma (let-7a-2 and let-7e), breast cancer (let-7a-2), and ovarian cancer (let-7a-3/let-7b). For example, the genomic locus harboring let-7a-3/let-7b was deleted in 44% of the specimens from ovarian cancer patients. We also found a positive correlation between the copy number of let-7b and mature let-7b expression in ovarian cancer. Finally, we showed that restoration of let-7b expression dramatically reduced ovarian tumor growth in vitro and in vivo. Our results indicate that copy number deletion is an important mechanism leading to the downregulation of expression of specific let-7 family members in medulloblastoma, breast, and ovarian cancers. Restoration of let-7 expression in tumor cells could provide a novel therapeutic strategy for the treatment of cancer.


Introduction
The characterization of heterochronic mutations in Caenorhabditis elegans revealed an evolutionarily conserved genetic pathway that orchestrates both the timing of cell divisions and cellular fates, as appropriate for the developmental stage of the organism [1][2][3][4][5]. In this pathway, the microRNA (miRNA) let-7 controls the progression of timing events, ensuring that cell cycle exit and terminal differentiation occur at the correct time [6,7]. The let-7 target genes lin-28 (an RNA-binding protein) and lin-41 (a putative ubiquitin ligase) block let-7 maturation and interact with argonaute proteins, respectively. These heterochronic genes appear to form bistable switches via double-negative regulatory feedback loops [2][3][4][5]. Expression of LIN28 and LIN41 is highly restricted at undifferentiated stages, such as in embryonic stem cells, early embryos, and certain somatic stem/progenitor cells. Conversely, their regulatory miRNA, let-7, shows a reciprocal temporal expression pattern that is dramatically increased during differentiation and development, and it is extensively expressed in adult tissues [2][3][4][5]. Thirteen members of the let-7 family have been identified in the human genome [7,8] which display both distinct and overlapping functions [8].
The role of let-7 in cancer was first discovered by Johnson et al. when they found that the let-7 family negatively regulates let-60/ RAS in C. elegans by binding to multiple let-7 complementary sites in its 39 untranslated region (39UTR) [9]. Moreover, having found that let-7 expression is lower in lung tumors than in normal lung tissue, while RAS protein is significantly higher in lung tumors, they proposed that let-7 is a tumor suppressor gene [9], which is consistent with previous clinical observations in lung cancer [10]. Reduced expression of let-7 has been associated with shortened postoperative survival in patients with cancer [7,11,12], and forced expression of let-7 family members can suppress cancer cell growth both in vitro and in vivo [13][14][15][16]. The inhibitory function of the let-7 family in cancer has been corroborated by a number of groups and in various types of tumors [7,11,12]. It is likely that let-7 performs these functions by targeting various genes. For example, let-7 inhibits many well-characterized oncogenic proteins, including KRAS [9,17,18], HRAS [9,17,18], HMGA2 [18][19][20][21], c-Myc [22], and NF2 [23]. In addition, let-7 targets multiple cell cycle associated genes, including CDC25A [24], CDK6 [24], and CDK4 [25] as well as Cyclin A [25], D1 [25], D2 [24], and D3 [25]. Finally, let-7 represses expression of the reprogramming factor LIN28 that functions to block differentiation and maintain cancer stem cell populations [26].
Since it has been found that human cancers show a significantly reduced expression of the let-7 family, and that this is associated with shorter survival times in these patients [7,11,12], the characterization of the mechanisms leading to let-7 downregulation in cancer has important clinical significance. One cause of genetic deregulation in cancer is an alteration in the somatic copynumber of a given gene [27]. It has also been reported that miRNA copy numbers may be altered during tumorigenesis [28,29]. This led us to examine whether the copy-numbers of the let-7 family were altered in cancer.

Results
Certain Members of the let-7 Family have Deletions in Copy Number in Medulloblastoma, Breast Cancer, and Ovarian Cancer To determine the copy number of the let-7 family members, we analyzed a high-resolution SNP array (Affymetrix 250 K Sty array) dataset, Tumorscape, created by the Broad Institute of MIT and Harvard [27]. Fourteen types of human cancers (a total of 2,969 tumor specimens) were included in our study ( Table 1). The segmented raw data of the above SNP arrays were retrieved and analyzed following a standard protocol provided by the Tumorscape database [27] and visualized by the Integrative Genomics Viewer (IGV) (Figure 1) [30]. The genomic regions where the copy number alterations occurred significantly more frequently than the background rate were identified using GISTIC algorithm [31]. The following terms, which were defined by the Tumorscape database [27], were used to describe our results. Frequency indicates the fraction of cancers which exhibit amplification/ deletion at the genomic locus harboring a given let-7 gene. Low qvalues (upper threshold = 0.25) suggest that amplifications/deletions at this locus are significant and enriched by selective pressures.
As shown in Figure 2, the let-7 family is present as thirteen members located at eight loci in the human genome [7,11,12]. All eight genomic loci were analyzed in our study. Most of the family members, such as let-7a-1, let-7f-1, and let-7d, cluster together. (The let-7 genes are named by including a letter to denote a distinct mature sequence, and a number to indicate that the same mature sequence is present at distinct genomic loci.) Briefly, we found that four let-7 loci harboring five let-7 members showed significant deletions in copy number in a cancer-type specific manner ( Figure 2). Focal deletions of these let-7 family members were found in three cancer types: medulloblastoma (let-7a-2, frequency 25%; let-7e, frequency 9%), breast cancer (let-7a-2, frequency 47%), and ovarian cancer (let-7a-3/let-7b, frequency 44%). This suggests that genomic focal copy number deletions of let-7 may play an important role during tumorigenesis in the above cancer types. In addition, two non-focal deletions (let-7a-3/let-7b, frequency 40%; let-7g, frequency 36%) were also found in breast cancer, and single non-focal deletions were found in melanoma (let-7a-2, frequency 43%) and non-small cell lung carcinoma (let-7e, frequency 31%). Finally, only one let-7 family member, let-7i, was found to be amplified (in non-small cell lung carcinoma). However, this was appeared a chromosome arm-level, non-focal amplification, indicating that it could be a ''passenger'' genetic alteration. Taken together, our data indicate that a reduction in copy number of specific let-7 family member genes were frequent in medulloblastoma, breast, and ovarian cancers.

Copy Number Alteration of let-7b is Positively Correlated with Mature let-7b Expression in Ovarian Cancer
To determine whether copy number alterations of let-7b affect mature let-7 expression in cancer, we examined an ovarian cancer dataset from The Cancer Genome Atlas (TCGA) [32], because this independent genomic dataset contains matched data on both genome-wide copy number (SNP array) and mature miRNA expression (miRNA array) from a large collection of human ovarian tumor specimens. Level 3 data (segmented SNP array data and normalized miRNA array data) was retrieved from TCGA [32]. A total of 537 tumors with data from both the SNP and miRNA arrays were identified. Consistent with the data from our Tumorscape analysis, there was a deletion in copy number at the let-7a-3/let-7b locus in the specimens from ovarian cancer patients ( Figure 3A). Since the mature let-7a sequence is encoded by three let-7a genes, which are located at three different chromosomal loci, we could not examine the correlation between copy number and expression for let-7a-3. Thus, we analyzed the correlation between deletions at the let-7a-3/let-7b locus and expression of mature let-7b in this data set. As shown in Figure 3 B and C, the expression of mature let-7b was significantly and positively correlated with let-7b copy number in ovarian cancer specimens (p,0.0001, R = 0.46, n = 537). We did not find a correlation between any other let-7 family members with the let-7b copy number ( Figure 3D). This demonstrates that a reduction in the copy number of let-7b leads to a reduction in the expression of mature let-7b in ovarian cancer.

Restoration of let-7b Expression Significantly Reduces Ovarian Tumor Growth in vitro
Focal loss in copy number of thelet-7 family members in medulloblastoma, breast, and ovarian cancers strongly suggests that let-7 may have an important role in tumorigenesis. Therefore, restoring its expression could suppress tumor growth in these cancers. To test this hypothesis in vitro, we transfected two ovarian cancer cell lines, A2780 and 2008, with a let-7b mimic (a doublestrand RNA oligonucleotide). A cultured human ovarian surface epithelial (HOSE) cell line was used as a control, and cell growth was monitored via an MTT assay. At 72 hours post-transfection, we found that the let-7b mimic significantly reduced cell growth in the cancer cells ( Figure 3 A and B). There was no reduction in cell growth seen in the control cells ( Figure 4C). This suggests that the restoration of let-7b expression in tumor cells may have great therapeutic potential for the treatment of ovarian cancer.

Restoration of let-7b Expression Significantly Reduces Ovarian Tumor Growth in vivo
To evaluate the in vivo therapeutic potential of the let-7b mimic, an orthotopic late stage ovarian cancer mouse model was generated by intraperitoneal injection of 10610 6 ovarian cancer (A2780) cells into female nude mice ( Figure 5A). After two weeks, the mice were randomly assigned to two groups, to be treated with either the let-7 mimic or the control oligonucleotide by i.p. injection. The mice were treated every three days for a total of four treatments ( Figure 5A), and the tumor nodes were collected three days after the last treatment ( Figure 5B). We found that the weight of the tumor nodes was significantly reduced in the mice treated with the let-7b mimic compared to controls (Figure 5 B and C). We also monitored endogenous let-7b activity using a constitutively expressed let-7b luciferase reporter that contained sequences complementary to let-7 in the 3'UTR [33,34]. We consistently found that treatment with the let-7b mimic significantly decreased luciferase activity in vivo compared to the control group ( Figure 5 D and E). Finally, we found that treatment with the let-7b mimic significantly increased the let-7b expression levels in the treatment compared to the control group (7.0365.73 fold, p = 0.047). Meanwhile, the mRNA expression levels of well-known let-7 target genes such as CCND1, CDC25A, HMGA2, IL6 and LIN28B were significantly decreased by let-7b mimic treatment. Taken together, this demonstrates that the in vivo delivery of a let-7b mimic can functionally restore let-7 expression and remarkably reduce tumor growth in a pre-clinical animal model of ovarian cancer.

Discussion
The let-7 family is one of the first miRNA tumor suppressor families shown to be involved in human cancer. Expression of members of the let-7 family has been reported to be significantly downregulated in multiple cancer types, and this decreased let-7 expression has been correlated with poorer clinical outcomes. Two molecular mechanisms have been proposed that may lead to global downregulation of let-7 expression in cancer. First, key proteins in the miRNA biogenesis pathway, such as Dicer and Drosha, can be remarkably deregulated in cancer [35]. This may result in an unselective, global downregulation of miRNAs, including the let-7 family. In addition, it has been shown that the RNA-binding protein, LIN28, which selectively inhibits some miRNA families, including the let-7 family [36][37][38][39][40][41], is activated in a large percentage of cancer patients [42][43][44][45][46][47][48][49][50][51]. However, the above two mechanisms cannot explain the finding that, in most cancer types, only some let-7 family members are downregulated. For example, the expression of let-7a, let-7c, and let-7g have been found to be selectively downregulated in breast cancer [52], suggesting that there are other independent mechanisms affecting the expression of each individual let-7 family member. Recent advances in high-throughput genome characterization technologies have revealed that cancer genomes are highly disordered, with extensive changes in chromosome structure and gene copy numbers. It has also been reported that copy numbers of miRNAs are commonly altered in human cancer [28,29]. For example, the mir-16-1/mir-15a cluster on chromosome 13q14 was deleted in more than 50% of the chronic lymphocytic leukemia patients [53]. In addition, amplification of the C13orf25/mir-17,92 cluster on chromosome 13q31-32 has been reported in lymphoma patients [54,55]. In the present study, we have shown that three let-7 loci, which harbor four let-7 members (let-7a-2, let-7a-3, let-7b, and let-7e), have deletions in copy number in a cancertype specific manner in medulloblastoma, breast cancer, and ovarian cancer. Most importantly, we confirmed the correlation between let-7 copy number alterations and mature let-7 expression in ovarian cancer. These results indicate that deletion in copy number is an important mechanism leading to the downregulation of specific let-7 family members in at least these three types of human cancers. However, this mechanism is likely cancer-type specific, since we did not find significant copy number alterations of the let-7 family in other cancer types, such as colon and prostate cancers.
Importantly, focal reductions in the copy numbers of the let-7 family suggest that deletion of let-7 may play an important role    during tumorigenesis, and suggests that restoring expression of these let-7 family members may be a novel strategy to treat medulloblastoma, breast cancer, and ovarian cancer. Current rapid advances in oligonucleotide/nanoparticle therapy create realistic optimism for the establishment of miRNAs as a new and potent therapeutic target and/or chemoresistant modulator in cancer. To test this hypothesis, we delivered a small RNA mimic for let-7b, the most frequently deleted let-7 family member in ovarian cancer patients, to ovarian cancer cells in vitro and in vivo. We found that restoration of let-7b expression dramatically inhibited tumor growth. In agreement with our observations, restoration of let-7 expression has also been shown to reduce tumor growth in pre-clinical models of other cancer types, such as lung cancer [13][14][15][16], in which the let-7 family is globally decreased [9,10]. Given that let-7 simultaneously inhibits multiple oncogenic pathways that are involved in most steps of tumorigenesis (such as RAS, MYC, and HMGA2), restoration of let-7 expression in tumor cells provides a novel therapeutic strategy to treat cancer. Interestingly, we found that our let-7b treatment did not significantly affect normal ovarian surface epithelial cell growth, suggesting that treatment to restore let-7 expression may be less toxic than traditional chemotherapy. This finding may be due to the fact that normal cells already express higher levels of endogenous let-7 and therefore the delivery of additional let-7 does not significantly increase its gene silencing activity in normal cells. Another explanation could be that the downstream cellular context is different in the tumor cells vs. the normal cells. For example, some direct targets of let-7, such as LIN28, RAS, MYC and HMGA2, are not expressed or activated in normal cells, but are the 'driver' genes promoting cell growth in tumors.
It has been demonstrated by independent laboratories that the repression on their target gene expression levels by miRNAs is relatively mild [56][57][58][59], although each miRNA can specifically and simultaneously target hundreds of mRNA transcripts [56,59]. In consistent with this common behavior of miRNAs, we showed that the mRNA expression levels of multiple well-known let-7 target genes such as CCND1, CDC25A, HMGA2, IL6 and LIN28B were significantly decreased by let-7b mimic treatment (all p,0.05). However, the repression effects on the mRNA levels by let-7b are typically modest (from 58.0% 60.03% to 78.9% 60.08%). It could bring important therapeutic benefits for the let-7b replacement therapy compared to the siRNA based strategies: First, the side effects of restoration of let-7b in patients with cancer may be mild and limited. Second, multiple oncogenic genes/pathways such as cell cycle regulation, DNA damage response, cancer stem cell differentiation and inflammation could be targeted simultaneously. In agree with this hypothesis, indeed we did not find that the let-7b replacement therapy significantly affected on normal epithelial cell growth at a therapeutic dosage ( Figure 4C). In addition, miRNAs including let-7 negatively regulate target gene expression by two major mechanisms, i.e. mRNA cleavage (transcriptional level) and/or translational repression (translational level), in a sequence-specific manner [7,8,11,12]. This could be another explanation why we only found modest repression of the target gene expression in mRNA levels that were detected by realtime RT-PCR. Finally, the dosage for let-7b delivery in the present study was determined by a dose-dependent experiment. We used the minimal functional response dosage (30 nM) for our real-time RT-PCR experiments. When we increased the dose of let-7b delivery, we found the effects of the let-7b replacement therapy on its target genes were remarkably increased.
It has been well demonstrated that the mature let-7 expression is a robust biomarker to predict clinical outcome in patients with cancer. For example, a systematic review of 43 published studies shows that let-7 is the miRNA most frequently and significantly associated with clinical outcomes in patients with cancer [60]. To examine the correlation between the let-7b/let-7a3 cluster copy number alterations (deletion v.s. non-deletion) and clinical outcome of patients, a total of 491 late-stage (stage-III and -IV) tumors with well-annotated survival information was examined by Kaplan-Meier survival analysis. However, we did not find significant correlation between the let-7b/let-7a3 cluster copy number alterations and overall survival in this sample set (p = 0.343). This result indicated that the DNA copy number alteration is not the only reason by which the mature let-7 expression is reduced in cancer. In agree with this observation, we and other independent groups have reported that miRNA deregulation in human ovarian cancer was led by multiple mechanisms [29,61,62]. First, it has been shown that most primary miRNAs are transcripted from Pol II promoter and regulated by transcriptional factors. Several examples showing miRNA deregulation in ovarian cancer by transcriptional deregulation have been reported. Second, recent studies suggest that epigenetic alterations play a critical role in deregulating miRNA expression in human ovarian cancers. Third, mutation might also contribute to down-regulation of mature miRNAs. Finally, the key proteins in the miRNA biogenesis pathway may be dysfunctional or deregulated in cancer, and may enhance tumorigenesis. Therefore, transcriptional deregulations, epigenetic alterations, mutations, DNA copy number abnormalities and defects in the miRNA biogenesis machinery might each contribute, either alone but more likely together, to the let-7 family deregulation in human cancer [29,61,62]. In the present study, we reported that let-7b/let-7a3 cluster deleted in more than 40% percentage of ovarian and breast cancers. We believe it is one of important mechanisms to decrease let-7 expression in these diseases, although other mechanisms such as transcriptional deregulations, epigenetic alterations, mutations, and defects in the miRNA biogenesis machinery are needed to be further characterized in ovarian and breast cancers.

SNP and miRNA Microarray Data Retrieval and Analysis
The segmental raw data from the SNP microarrays (Affymetrix 250 K Sty array) was downloaded from the Tumorscape database (created by the Broad Institute of MIT and Harvard) [27] and analyzed following a standard protocol provided by Tumorscape, then visualized using the Integrative Genomics Viewer (IGV) (Figure 1) [30]. Fourteen types of human cancers (a total of 2,969 tumor specimens) were included in our study ( Table 1). The level 3 data (segmented SNP array data and normalized miRNA array data) was download from The Cancer Genome Atlas (TCGA) dataset [32], and a total of 537 tumors with both SNP and miRNA array data were identified. The genomic regions where copy number alterations occurred significantly more frequently than the background rate were identified using the Genomic Identification of Significant Targets in Cancer (GISTIC) algorithm [31]. The information for the tumor types of the specimens from the Tumorscape database was available at the portal of the Tumorscape (http://www.broadinstitute.org/tumorscape) [27].
Breast and ovarian tumors clinical information was available from the published profiles (GSE7545 and GSE19399) [27,63,64]. All TCGA ovarian cancer specimens are high-grade serous ovarian tumors, which were collected from newly diagnosed patients with ovarian serous adenocarcinoma who were undergoing surgical resection and had received no prior treatment for their disease, including chemotherapy or radiotherapy. All cases had to be of serous histology but were collected regardless of surgical stage or histologic grade. Cases were staged according to the 1988 FIGO staging system. The detail clinical information of the specimens from the TCGA database was available at the TCGA data portal (https://tcga-data.nci.nih.gov/tcga/) [32].

Cell Lines and Cell Culture
Ovarian cancer cell lines were purchased from the American Type Culture Collection (ATCC) and the Division of Cancer Treatment and Diagnosis (DCTD) Tumor/Cell Line Repository. All cancer cell lines were cultured in RPMI 1640 medium (Invitrogen) supplemented with 10% fetal bovine serum (Invitrogen). Human ovarian surface epithelial (HOSE) cells were generously provided by Dr. Nelly Auersperg (University of British Columbia) [65]. HOSE cells were cultured 1:1 in Media 199: MCDB 105 (Sigma), supplemented with 15% FBS.

let-7b Mimic in vitro Transfection
The let-7b mimic and control oligonucleotides were purchased from Invitrogen. Cells were plated 24 hours before transfection. Transfection of the miRNA oligonucleotides (30 nM) in vitro was performed using Lipofectamine RNAiMAX (Invitrogen).

MTT Assay
MTT assays were performed in 96-well plates using a Cell Proliferation Kit (Roche) following the manufacturer's instructions. Four to six wells were analyzed for each sample. The resulting colored solution was quantified using an ELx800 Absorbance Microplate Reader (BioTek) at 570 nm with a reference wavelength of 630 nm.

Generation of Stable Cell Lines
The let-7 reporter vector was transfected into A2780 cells using the FuGene6 Transfection Reagent (Roche). The medium was changed 48 hrs post-transfection, and the stable clones were selected by neomycin for 14 days. The stable clones expressing the let-7 reporter were further confirmed by a luciferase assay.

Generation of Orthotopic Ovarian Cancer Xenograft Model
Six to eight week old female nude mice were obtained from The Jackson Laboratory. Subconfluent ovarian cancer cells (A2780) were trypsinized and suspended in phosphate buffered saline (PBS), then 10610 6 cells in a total volume of 0.1 ml were injected into the mouse peritoneal cavity. The animal study protocol was reviewed and approved by the institutional animal care and use committee (IACUC) of the University of Pennsylvania.

let-7b Mimic Delivery in vivo
The let-7b mimic and control oligonucleotides were chemically synthesized by Sigma, and the jetPEI reagent (Polyplus Transfection) was used to deliver the miRNA in vivo. Two weeks after the tumor cell injection (Figure 5A), the mice were randomly assigned to two groups, to be treated with either the let-7 mimic or the control oligonucleotide (40 ug per animal). The let-7b mimic (or the control oligonucleotide) was combined with the jetPEI reagent at an N/P ratio of 8, following the manufacturer's instructions. The let7-mimic/jetPEI or control-mimic/jetPEI was administrated to the mice by intraperitoneal injection every three days for a total of four treatments ( Figure 5A). Three days after the last treatment, the animals were euthanized and the tumor nodes were collected and weighed.

Bioluminescence Optical Imaging
The IVIS Lumina II Bioluminescence and Fluorescence Imaging System (Caliper Life Sciences) was used for in vivo bioluminescent imaging of the xenograft tumors. Animals were anesthetized with isoflurane and injected intraperitoneally with Dluciferin substrate (150 mg/kg body weight) (Caliper Life Sciences). Pilot studies revealed that the peak bioluminescent intensity of the tumors was reached 10 minutes after D-luciferin injection; therefore, this time point was chosen for imaging. Images of the tumor were taken under the following settings: exposure time = 0.5 seconds, f/stop = 16, medium binning. Living Image software was used to quantify the bioluminescent signal, reported as units of tissue radiance (photons/s/cm2/sr).

RNA Isolation
Total RNA was isolated from 100-500 mg of frozen tissue using TRIzol reagent (Invitrogen). The quality and quantity of the isolated RNA was analyzed using a Bioanalyzer 2100 system (Agilent).

TaqMan miRNA Assays
Expression of mature let-7b was analyzed using TaqMan miRNA assays (Applied Biosystems, Foster City, CA, USA) under conditions recommended by the manufacture. Briefly, singlestranded cDNA was synthesized from 5 ng of total RNA in a 15 ml reaction volume using a TaqMan microRNA reverse transcription kit (Applied Biosystems, Foster City, CA, USA). The reactions were incubated first at 16uC for 30 min, at 42uC for 30 min, and then inactivated by incubation at 85uC for 5 minutes. Each cDNA generated was amplified by quantitative PCR using sequencespecific primers from the TaqMan microRNA assay on an 7900HT sequence detection system (Applied Biosystems, Foster City, CA, USA). Each 20 ml PCR included 10 ml of 26 Universal PCR Master Mix (without AmpErase UNG), 1 ml of 206TaqMan microRNA assay mix, and 2 ml of reverse transcription product. The reactions were incubated in a 384-well plate at 95uC for 10 min, followed by 40 cycles of 95uC for 15 seconds and 60uC for 1 minute.

Quantitative QRT-PCR
Total RNA was reverse-transcribed using a High Capacity RNA-to-cDNA Kit (Applied Biosystems) according to the manufacturer's instructions. cDNA was quantified by real-time PCR on an ABI Prism 7900 Sequence Detection System (Applied Biosystems). PCR was performed using SYBR Green PCR Core reagents (Applied Biosystems) according to manufacturer's instructions. PCR amplification of the housekeeping gene GAPDH was performed for each sample as a control for sample loading and to allow for normalization across samples.

Statistical Analysis
Statistical analysis was performed using the SPSS and SAS statistical software packages.