Abstract

The molecular mechanisms governing the secretion of the non-coding genome are poorly understood. We show herein that cyclin D1, the regulatory subunit of the cyclin-dependent kinase that drives cell-cycle progression, governs the secretion and relative proportion of secreted non-coding RNA subtypes (miRNA, rRNA, tRNA, CDBox, scRNA, HAcaBox. scaRNA, piRNA) in human breast cancer. Cyclin D1 induced the secretion of miRNA governing the tumor immune response and oncogenic miRNAs. miR-21 and miR-93, which bind Toll-Like Receptor 8 to trigger a pro-metastatic inflammatory response, represented >85% of the cyclin D1-induced secreted miRNA transcripts. Furthermore, cyclin D1 regulated secretion of the P-element Induced WImpy testis (PIWI)-interacting RNAs (piRNAs) including piR-016658 and piR-016975 that governed stem cell expansion, and increased the abundance of the PIWI member of the Argonaute family, piwil2 in ERα positive breast cancer. The cyclin D1-mediated secretion of pro-tumorigenic immuno-miRs and piRNAs may contribute to tumor initiation and progression.

Introduction

Less than 3% of the transcribed human genome is translated into protein. There has been a surge of interest in the role of the non-coding RNA transcriptome and its contribution to disease. The small non-coding RNA category (20–35 bp) includes microRNAs (miRNAs), tRNA-derived fragment (tRF) [1], P-element-induced wimpy testis (PIWI)-interacting RNAs (piRNAs), C/D Box [2], snoRNA-derived RNA (sdRNAs), and small interfering RNAs (siRNAs) [3]. The molecular mechanisms governing the abundance and location of small non-coding RNA are important for understanding the function of the non-coding genome. miRNAs associate with the Argonaute (Ago) proteins to regulate mRNA expression through the RNA-induced silencing complex (RISC). piRNAs interact with the PIWI members of the Argonaute family, and thereby function in germ stem cell maintenance, spermatogenesis, epigenetic gene regulation, and the silencing of non-long terminal repeat transposons [4,5].

Altered miRNA expression is thought to contribute to the timing of cell-cycle transition, cell proliferation, stem cell self-renewal, cancer onset and progression [6,7]. Specific miRNAs have been associated with the initiation, progression and therapy resistance in human breast cancer [3,7]. Altered expression of piRNAs occurs in specific cancers [2]. Furthermore, miRNAs regulate immunological processes including lineage commitment, and the activation and ageing of innate and adaptive immune cells through immune-miRs [8,9]. miRNAs can be secreted to thereby influence nearby cells [10]. Immuno-miRs may regulate immune function in the tumor microenvironment by binding to the 3′-UTR region of target mRNAs, or by binding to- and activating Toll-like receptor 8 (TLR8) in surrounding cells of the immune system [11].

The cyclin D1 (CCND1) gene, which encodes the regulatory subunit of a holoenzyme that phosphorylates and inactivates the Rb protein [12], is overexpressed in up to 50% of human breast cancer. Cyclin D1 is required for the growth of oncogene-induced breast tumors in mice [13,14], and is sufficient for the induction of mammary tumors with high levels of chromosomal instability in mice [15]. Cyclin D1 governs certain functions of the non-coding genome, first by inducing Dicer and thereby altering miRNA processing [16], and second by regulating the transcription of specific miRNAs [17]. In breast cancer cells, cyclin D1 governs a signature of miRNA which was predictive of breast cancer outcome associated with the induction of Wnt signaling [18].

Recent studies have shown an important role for cyclin D1 in the anti-tumor immune response. Cyclin D1 expression in the stroma enhanced growth of breast cancer cells and supported a pro-tumorigenic immune environment [19]. Myeloid-derived suppressor cells (MDSCs) constitute one population of inflammatory cells that suppress the anti-tumor immune response and thereby enhance tumor growth [20]. Cyclin D1 induced the expansion of the MDSC population [19]. Further analysis revealed that cyclin D1-conditioned medium induced expansion of CD34 positive hematopoietic stem cells (HSCs) and promoted differentiation of CD34 positive HSCs into MDSC [19].Cyclin D1 expression in the stroma increased the number of F4/80+ and CD11b+ macrophages [19]. However, the mechanisms by which cyclin D1 regulates the tumor immune response and MDSC expansion are poorly understood.

The present studies, conducted in order to determine if cyclin D1 governs the release of non-coding RNA transcriptome from cancer cells, demonstrated that cyclin D1 determines the relative proportion of non-coding transcriptome subtypes, and augments secretion of miRNAs that are predominantly immune-miRs promoting restraint of the anti-tumor response. Furthermore, we show that cyclin D1 induces piwil2 expression and promotes secretion of specific piRNAs that regulate breast cancer stem cells. The expression levels of piwil2 and cyclin D1 in breast cancer are positively associated with ERα status. Collectively, these studies demonstrate a previously unrecognized function that cyclin D1 compartmentalizes components of the non-coding genome to augment secretion of immuno-miRs and a pro-tumorigenic immune phenotype in human breast cancer.

Methods and materials

Cell culture and plasmids

MCF-7 cells, obtained from ATCC, were cultured in DMEM containing penicillin and streptomycin (100 mg of each/l) and supplemented with 10% fetal bovine serum (FBS). Additional breast cancer cell lines, SUM-159, Hs-578t, MDA-MB-231, MDA-MB-453, T-47D, BT474 and SKBR3, were cultured as previously described [16]. Germ cells were isolated from testis tissue of male mice for RNA extraction, and used for positive control of piRNA and PIWI family genes. MSCV-cyclin D1 and control vector were used for cyclin D1 overexpression as referenced [16].

Human breast tumor samples

Human breast cancer specimens and matching normal breast tissue samples were collected from Tongji University Shanghai East Hospital. All the procedures were approved by the Institutional Review Board (IRB) of Tongji University School of Medicine. All patients were provided with written informed consent form.

SiRNA transfection

siRNAs were synthesized from Dharmacon (Lafayette, CO. U.S.A.). Target sequence for cyclin D1 siRNA was: si-D1 AACAAGCTCAAGTGGAACCTG. The target sequence for negative control siRNA was AATTCTCCGAACGTGTCACGT; Target sequence for si-cyclin D1-2 (Ribobio Co. China) was: TCGGTGTCCTACTTCAAAT. Lipofectamine RNAiMax (Invitrogen) was used to transfect siRNAs following the manufacturer’s instructions.

Western blotting analysis

The common protocol for Western blot was followed as referenced [16]. The following antibodies were used for Western blot: anti-Cyclin D1 (Cat. No. sc-450, clone 72-13G, Santa Cruz), anti-β-Tubulin (Cat. No. sc-9104, Santa Cruz), anti-Piwil2 (Cat. No. sc-377258, clone G-1, Santa Cruz), anti-β-Actin (Cat. No. sc-47778, clone C4, Santa Cruz). Secondary HRP-linked antibody was anti-mouse (Cat. No. 7076P2, Cell Signaling Technology).

Cyclin D1-conditioned medium preparation and RNA extraction

A total of 5 × 105 breast cancer cells were treated with cyclin D1 siRNA or control for 48 h, followed by 1× PBS solution washing and overlaid with phenol red-free DMEM without serum and allowed to grow for another 24 h before collecting the medium (supernatant) from the cultures. The collected medium was centrifuged at 2000 × g for 10 min and filtered through 0.22-μm membrane to yield the conditioned medium. TRIzol® LS Reagent (Invitrogen) was used for RNA extraction following the manufacturer’s instruction. MirVanaTM miRNA Isolation Kit from Ambion was used to enrich small RNA population before establishing small RNA library for further analysis. Identical amount of RNAs from cyclin D1+ and cyclin D1- cells were used for comparison.

Immunoprecipitation assay

Cyclin D1-overexpressing MCF-7 cells were lysated with RIPA buffer. About 8 µg of Cyclin D1 antibody (DCS-6, sc-20044, Santa Cruz) and same amount of IgG (sc-2025, Santa Cruz) were used for immunoprecipitation at 4°C overnight. About 30 µl of protein A/G beads (sc-2003, Santa Cruz) were used to bind antibody. The eluted lysates (elution buffer contains 50 mM Tris (pH 6.8) and 2% SDS) and input were applied to detect Cyclin D1 and Piwil2 by Western blot.

Small RNA library preparation and deep sequencing

Small RNA libraries were prepared using the SOLiD™ Small RNA Expression Kit (Applied Biosystems) following the manufacturer’s instructions. cDNA libraries were amplified by standard emulsion PCR using primers complementary to adaptors. After amplification, the library was sequenced by the SOLiD™ 4 System (Applied Biosystems).

Mapping of the reads

Reads were preprocessed using “cutadapt” and mapped on the human genome using BWA [21]. During mapping, base quality values were taken into account and no more than two mismatches (but no insertions or deletions) were allowed. Reads that mapped at multiple locations were discarded and not considered further. All miRNA reads were normalized with a housekeeping small RNA SNORD44. Only those miRNAs with reads pileup value more than 25 and fold-change more than 1.5 were considered as differentially expression. The miRNA reference was based on the miRBase 17 database (http://www.mirbase.org/). The piRNA reference was based on the piRNABank (http://pirnabank.ibab.ac.in/).

Validation of miRNA expression profiling

A home-made quantitative real-time PCR-based human miRNA panel containing 104 differentially expressed miRNAs in the cyclin D1-conditioned medium and the most abundant 28 miRNAs in the secretome of MCF-7 cells derived from deep sequencing analysis was designed in triplicates, and applied to validate the miRNA profiling in the secretomes of human breast cancer cells including MCF-7 and T47D. RNAs from the cyclin D1-conditioned medium were prepared as described above. The common protocol for miRNA QRT-PCR was followed as referenced [16].

piRNA expression analysis

Quantitative analysis of piRNAs was performed following the exactly same protocol with miRNA analysis. The forward primer sequences (5′-3′) are: piR-016658 (DQ592931): CACTGCTAAATTTGACTGGC; piR-016792 (DQ593109): AGTGCTGGGATTACAGGCGTG; piR-011188 (DQ585095): CCTGTAATCCCAGCTACTCA; piR-015150 (DQ590704): TGTAATCCCAGCACTTTGG; piR-016975 (DQ593415): ATGATGTTCCGCAACTACCT. 5s ribosomal RNA was used for normalization. Target genes of piRNA were predicted according to literature [22].

Immunohistochemistry (IHC) staining

Antibodies against Cyclin D1 (sc-8396, clone A-12, Santa Cruz) or with a rabbit polyclonal antibody against Piwil2 (from Dr Gao’s lab [23]) were used for IHC staining. IgG was used as negative control. All slides were photographed using fluorescence microscopy (Leica, Germany). Colocalization of Cyclin D1 and Piwil2 was evaluated quantitatively using ImageJ software. Pearson’s correlation coefficient (Pearson-R) and overlap coefficient (Overlap-R) were employed to evaluate colocalization.

Databases

The TCGA portal web server (http://tumorsurvival.org/index.html) and a KM plotter tool (http://kmplot.com/analysis/index.php?p=service&cancer=breast) were used to analyze the gene expression level and relapse-free survival. Onco-miR analysis in breast cancer was performed using 479 breast cancer-related miRNAs from the OncoLnc database (http://www.oncolnc.org/) [24]. Immuno-miR analysis was performed using 834 immune miRNAsfrom the MNDR v2.0 database (http://www.rna-society.org/mndr/index.html) [25].

Statistical analysis

Paired and unpaired t-test of parametric analysis were used for statistical comparison by using GraphPad Prism 8.0.2. Pearson’s correlation coefficient and overlap coefficient were used for colocalization analysis by using ImageJ. Pearson correlation coefficient was used for correlation analysis by using GraphPad Prism 8.0.2. Gene enrichment of the functions and pathways was analyzed by using Clusterprofiler tool. Data are presented as mean ± SEM. P < 0.05 was considered statistically significant.

Results

Cyclin D1 governs the relative abundance of miRNAs secreted from human breast cancer cells

In order to determine the role of endogenous cyclin D1 in compartmentalizing the non-coding RNA, MCF-7 cells transduced with either cyclin D1 siRNA or negative control were assessed (Figure 1A). Forty-eight hours post transfection, the serum-free conditioned-supernatant was collected. Secreted RNA was isolated and subjected to deep sequencing (Figure 1B–H). Electrophoretic analysis of the small RNAs secreted by MCF-7 cells revealed the presence of miRNAs (∼20 nt) and an additional 30 nt band corresponding to the size of piRNAs (Figure 1C). The majority of secreted non-coding RNA determined by deep sequencing analysis were miRNA (65%), with smaller amounts of rRNA (12%), tRNA (11%), CDBox RNA (6%), scRNA (2%), HAcaBox (2%) and scaRNA (2%) (Figure 1D). Cyclin D1 siRNA transfection reduced Cyclin D1 abundance (Figure 1A) and reduced the relative proportion of miRNA (from 65 to 48%), while increasing the relative proportion of rRNA (from 12 to 19%), tRNA (from 11 to 15%), CDBox RNA (from 6 to 9%) and scRNA (from 2 to 4%) (Figure 1E,F). Sequencing of secreted small RNAs from MCF-7 cells identified 339 miRNAs, 415 piRNAs, 91 HAcabox and 145 CDBox (Supplementary Tables S1–4). Notably, ∼45% of the miRNA were increased more than 1.5-fold by cyclin D1, whereas only 6% were reduced over 1.5-fold (Figure 1G). About 103 individual miRNA transcripts were increased and 1 miRNA was decreased more than 2-fold by endogenous cyclin D1 (Supplementary Table S5).

Intracellular cyclin D1 augments secretion of miRNA within the non-coding RNA pool

Figure 1
Intracellular cyclin D1 augments secretion of miRNA within the non-coding RNA pool

(A) Western blot of the MCF-7 cells with either control siRNA (si-NC) or two different cyclin D1 siRNAs (si-D1 and si-D1-2) treated, using antibodies as indicated to Cyclin D1 or β-Actin. Quantitative analysis to the intensities of bands was performed. Data are mean ± SEM (n = 3). ** P<0.01. (B) Schematic representation of experimental protocol. Cyclin D1 siRNA or control transduced MCF-7 cells were compared. The medium from the cells was subjected to RNA isolation and deep sequencing analysis. (C) RNAs were isolated from the conditioned medium (serum free) of MCF-7 cells with or without expression of cyclin D1. The RNAs were normalized to the amount of MCF-7 cells to make sure RNAs secreted from same number of cells were used for comparison. γ32P was applied to end-labeling the RNAs, followed by running a denatured PAGE gel to determine the size of RNA subpopulations. Germ cell RNA was used to indicate the size of piRNAs (first lane). Let-7 miRNA was used to indicate the size of miRNAs (second lane). The miRNA population between 25 and 15 nucleotides (nt) and the 30 nt band corresponding to piRNAs are indicated with arrows. (D–E) Pie diagrams illustrating the relative proportion of the types of small non-coding transcripts obtained from the media of the MCF-7 cells with or without expression of cyclin D1. (F) The enrichment values of miRNA transcripts secreted by MCF-7 cells with or without expression of cyclin D1. (G) Pie diagram showing the relative proportion of secreted miRNA transcripts based on their fold change by endogenous cyclin D1. (H) Read number of the top 10 abundant miRNAs in the MCF-7 medium with or without expression of endogenous cyclin D1. (I) QRT-PCR validation of the cyclin D1-regulated miRNAs in the medium of MCF-7 cells derived from deep sequencing analysis. Approximately 70% of the miRNAs were overlapped between QRT-PCR and sequencing analysis. (J) Pie diagram showing the QRT-PCR validation result of the cyclin D1-regulated miRNAs in the medium of MCF-7 cells. Approximately 75.9% of the tested miRNAs significantly upregulated while only 0.9% down-regulated by endogenous cyclin D1, which is consistent with the deep sequencing result. (K) Pie diagram showing the QRT-PCR validation of the cyclin D1-regulated miRNAs in the medium of T47D cells. Approximately 67.2% of the tested miRNAs significantly up-regulated while only 6.0% down-regulated by endogenous cyclin D1. (L) A overlapping comparison of the secretory miRNA signature of cyclin D1 between MCF-7 and T47D cells indicated a major overlap.

Figure 1
Intracellular cyclin D1 augments secretion of miRNA within the non-coding RNA pool

(A) Western blot of the MCF-7 cells with either control siRNA (si-NC) or two different cyclin D1 siRNAs (si-D1 and si-D1-2) treated, using antibodies as indicated to Cyclin D1 or β-Actin. Quantitative analysis to the intensities of bands was performed. Data are mean ± SEM (n = 3). ** P<0.01. (B) Schematic representation of experimental protocol. Cyclin D1 siRNA or control transduced MCF-7 cells were compared. The medium from the cells was subjected to RNA isolation and deep sequencing analysis. (C) RNAs were isolated from the conditioned medium (serum free) of MCF-7 cells with or without expression of cyclin D1. The RNAs were normalized to the amount of MCF-7 cells to make sure RNAs secreted from same number of cells were used for comparison. γ32P was applied to end-labeling the RNAs, followed by running a denatured PAGE gel to determine the size of RNA subpopulations. Germ cell RNA was used to indicate the size of piRNAs (first lane). Let-7 miRNA was used to indicate the size of miRNAs (second lane). The miRNA population between 25 and 15 nucleotides (nt) and the 30 nt band corresponding to piRNAs are indicated with arrows. (D–E) Pie diagrams illustrating the relative proportion of the types of small non-coding transcripts obtained from the media of the MCF-7 cells with or without expression of cyclin D1. (F) The enrichment values of miRNA transcripts secreted by MCF-7 cells with or without expression of cyclin D1. (G) Pie diagram showing the relative proportion of secreted miRNA transcripts based on their fold change by endogenous cyclin D1. (H) Read number of the top 10 abundant miRNAs in the MCF-7 medium with or without expression of endogenous cyclin D1. (I) QRT-PCR validation of the cyclin D1-regulated miRNAs in the medium of MCF-7 cells derived from deep sequencing analysis. Approximately 70% of the miRNAs were overlapped between QRT-PCR and sequencing analysis. (J) Pie diagram showing the QRT-PCR validation result of the cyclin D1-regulated miRNAs in the medium of MCF-7 cells. Approximately 75.9% of the tested miRNAs significantly upregulated while only 0.9% down-regulated by endogenous cyclin D1, which is consistent with the deep sequencing result. (K) Pie diagram showing the QRT-PCR validation of the cyclin D1-regulated miRNAs in the medium of T47D cells. Approximately 67.2% of the tested miRNAs significantly up-regulated while only 6.0% down-regulated by endogenous cyclin D1. (L) A overlapping comparison of the secretory miRNA signature of cyclin D1 between MCF-7 and T47D cells indicated a major overlap.

From this population of small non-coding RNAs, 28 miRNAs (with reads number from 1200 to 450,000) were found to be the most abundant in the medium (Supplementary Table S6) with the top 10 shown in Figure 1H. More than 85% of the cyclin D1-augmented secreted miRNA was attributable to miR-21, with additional contribution from miR-93, miR-24 and miR-29a and let-7d, let7e (Figure 1H), most of which have been shown to function as immuno-miRs.

In order to validate the results from deep sequencing analysis, fresh RNAs from the cyclin D1-conditioned medium of MCF-7 cells and another breast cancer cell line T47D were applied to a home-made quantitative real-time PCR-based human miRNA panel (Supplementary Table S7) covering miRNAs in Supplementary Tables S5 and S6. As shown in Figure 1I, 70 from 103 cyclin D1-upregulated miRNAs in the secretome of MCF-7 were validated by QRT-PCR analysis using independent samples. In consistence with the deep sequencing result, the QRT-PCR validation data indicated up-regulation of 75.9% of the tested miRNAs by cyclin D1, while only 0.9% showed down-regulation (Figure 1J and Supplementary Table S8). Similar results were obtained from the secretome of breast cancer cell T47D, in which 67.2% of the tested miRNAs showed up-regulation by cyclin D1 and only 6.0% showed down-regulation (Figure 1K and Supplementary Table S9). Notably, more than 70% of the cyclin D1-regulated miRNAs in the secretome of MCF-7 overlapped with that in T47D cells (Figure 1L), confirming the secretory miRNA signature of cyclin D1 in human breast cancers.

Cyclin D1 induces immuno-miRs and onco-miRs secretion in MCF-7 breast cancer cells

A comparison was made between the miRNAs regulated by cyclin D1 collected from the MCF-7 cell supernatant, with the miRNAs regulated by cyclin D1 within the MCF-7 cells [18]. Expression of cyclin D1 altered the expression of 121 miRNA in MCF-7 cells [18] and induced the expression of 103 secreted miRNA (Figure 2A and Supplementary Table S5). About 25 miRNAs were common to both data sets, which were expressed and secreted from MCF-7 cells in a cyclin D1-dependent manner (Figure 2B). A functional analysis of the 25 miRNAs regulated by cyclin D1 within MCF-7 cells and in secretome of the medium, demonstrated the majority of miRNAs were known immuno-miRs and/or oncomiRs (Figure 2C,D). Of the secreted miRNAs that were induced in abundance by cyclin D1, several have a well characterized function in the tumor immune response, including members of the miR-17-92 cluster, miR-15/16, miR-200c-141, and the miR-24-23a cluster (Supplementary Figures S1 and S2B) [26,27]. In contrast, the function of the miRNAs regulated by cyclin D1 within MCF-7 cells involves the induction of Wnt signaling [18] and several oncogenic clusters (miR-106a-363, C19MC, and the miR-17∼92 cluster) [18] (Figure 2B; Supplementary Figures S1 and S2).

Cyclin D1 regulated secreted and intracellular miRNAs convey distinct functions

Figure 2
Cyclin D1 regulated secreted and intracellular miRNAs convey distinct functions

(A) Venn diagram showing the number of distinct miRNAs that are altered in abundance by cyclin D1 either within the cell (121), secreted (103) or both within the cell and secreted miRNA (25). (B) The list of the 25 overlapping miRNAs in (A). (C and D) Functional analysis of the 25 miRNAs altered in abundance by cyclin D1 indicated major overlap with immune-miRs, as well as oncomiRs in breast cancer. (E) Western blot showing the increased levels of Cyclin D1 and Piwil2 in the MCF-7 cells transfected with cyclin D1-overexpressing vector. β-Tubulin was used as loading control. (F) Up-regulation of miRNA secretion by overexpression of cyclin D1 in MCF-7 cells. The top 10 miRNAs in the 25 cyclin D1-regulated miRNAs were selected for analysis. Data are mean ± SEM (n=3). * P<0.05, **P<0.01. (G) Signaling pathway analysis on the predicted target genes of the 25 cyclin D1-regulated miRNAs indicated the enrichment of the pathways regulating breast cancer, cell cycle, stem cell, IL-17 signaling, mTOR signaling, JAK-Stat signaling and PI3K-Akt signaling. The Clusterprofiler tool was used for gene enrichment analysis of functions and pathways. (H) The fold changes (cyclin D1+ vs. cyclin D1) in the cells and medium of the 25 regulated miRNAs by cyclin D1.

Figure 2
Cyclin D1 regulated secreted and intracellular miRNAs convey distinct functions

(A) Venn diagram showing the number of distinct miRNAs that are altered in abundance by cyclin D1 either within the cell (121), secreted (103) or both within the cell and secreted miRNA (25). (B) The list of the 25 overlapping miRNAs in (A). (C and D) Functional analysis of the 25 miRNAs altered in abundance by cyclin D1 indicated major overlap with immune-miRs, as well as oncomiRs in breast cancer. (E) Western blot showing the increased levels of Cyclin D1 and Piwil2 in the MCF-7 cells transfected with cyclin D1-overexpressing vector. β-Tubulin was used as loading control. (F) Up-regulation of miRNA secretion by overexpression of cyclin D1 in MCF-7 cells. The top 10 miRNAs in the 25 cyclin D1-regulated miRNAs were selected for analysis. Data are mean ± SEM (n=3). * P<0.05, **P<0.01. (G) Signaling pathway analysis on the predicted target genes of the 25 cyclin D1-regulated miRNAs indicated the enrichment of the pathways regulating breast cancer, cell cycle, stem cell, IL-17 signaling, mTOR signaling, JAK-Stat signaling and PI3K-Akt signaling. The Clusterprofiler tool was used for gene enrichment analysis of functions and pathways. (H) The fold changes (cyclin D1+ vs. cyclin D1) in the cells and medium of the 25 regulated miRNAs by cyclin D1.

In order to exclude the off-target effects, another cyclin D1 siRNA sequence (si-cyclin D1-2 in Figure 1A) was used to treat MCF-7 cells, followed by medium collection and miRNA detection in secretome. The top 10 miRNAs in the 25 cyclin D1-regulated miRNAs (Figure 2B) were selected for analysis, and all showed down-regulation in secretion by si-cyclin D1-2 (Supplementary Figure S3). In addition, cyclin D1 overexpression was applied to MCF-7 cells (Figure 2E), followed by the miRNA analysis in secretome. As shown in Figure 2F, all of the top 10 miRNAs in the 25 cyclin D1-regulated miRNAs were up-regulated in the medium of MCF-7 cells after cyclin D1 overexpression.

About 18 of the 25 miRNAs have been demonstrated to be onco-miRs in breast cancer. About 19 of the 25 miRNAs are immuno-miRNAs regulating the immune response [28–30]. The pathway analysis for target genes of the 25 miRNAs indicated the enrichment of pathways regulating “breast cancer”, “cell cycle”, “stem cell”, “IL-17 signaling”, “mTOR signaling”, “JAK-Stat signaling” and “PI3K-Akt signaling” (Figure 2G). The 25 miRNAs showed substantial differences in the relative changes induced by cyclin D1 in the medium versus the intracellular pools (Figure 2H). The relative changes by cyclin D were greater in the secreted pool for miR-29b-1, miR-16-1, let-7f-1, miR-199b and miR-18b. In contrast, the relative changes of miRNA abundance by cyclin D1 within MCF-7 cells were greater for miR-92a, miR-33b, miR-502, miR-135b, miR-193b and miR-505 (Figure 2H).

Cyclin D1 regulates piRNA in breast cancer associated with cancer stem cells

PIWI- interacting RNAs (piRNAs), a class of small non-coding RNA with 24–32 nt in length, were identified in the medium of MCF-7 cells. The expression and secretion of piRNAs in breast cancer cells were further analyzed (Figure 3). About 415 piRNA sequences were identified from the medium of MCF-7 cells (Supplementary Table S2) by deep sequencing analysis, demonstrating the altered piRNA secretion in MCF-7 cells by endogenous cyclin D1, with a relative increase in abundance (1.5- to 7-fold) of 17 piRNAs including piR-016658, and a reduction in abundance of 10 piRNAs including piR-016975 (Figure 3A). Representative examples of piRNAs identified in the MCF-7 cell supernatant were shown in Supplementary Figure S4, including a chromosome X piRNA cluster (piR-011188, piR-015150 and piR-016792). Quantitative analyses showed decrease of piR-011188 and piR-016792 in the cells and in the medium of cyclin D1 knockdown MCF-7 cells (Supplementary Figure S4A–C), while piR-015150 showed a ∼3-fold decrease in the cyclin D1 knockdown cells, but a ∼14-fold increase in the cyclin D1 knockdown medium (Supplementary Figure S4D–F).

Cyclin D1 augments secretion of PIWI-interacting RNAs (piRNAs) associated with breast cancer stem cells

Figure 3
Cyclin D1 augments secretion of PIWI-interacting RNAs (piRNAs) associated with breast cancer stem cells

(A) Bar graph showing the fold change of the abundant piRNAs in the medium of MCF-7 cells with regulation by cyclin D1 (>1.5 -fold change (FC)). (B) QRT-PCR analysis showing higher expression of piR-016658 in the triple negative breast cancer cells SUM159, MDA-MB-231 and Hs-578t compared with luminal MCF-7 cells. (C) A comparison between 5,272 predicted target genes of piR-016658 and 160 stem cell-regulating genes indicating an overlap of 55 genes, suggesting the possibility of piR-016658 involvement in regulating breast cancer stem cells. (D) QRT-PCR analysis showing higher expression of piR-016658 in aldehyde dehydrogenase 1 (ALDH1)+ breast cancer stem cells than that in (ALDH1) breast cancer cells from same patient (n=21). (E) Quantitative analysis of (D). (F) QRT-PCR analysis showing lower level of piR-016975 in the triple negative breast cancer cells SUM159, MDA-MB-231 and Hs-578t compared with luminal MCF-7 cells. (G) A comparison between 5079 predicted target genes of piR-016975 and 160 stem cell-regulating genes overlapped 51 genes, suggesting the possibility of piR-016975 involvement in regulating breast cancer stem cells. (H) QRT-PCR analysis showing lower expression of piR-016975 in aldehyde dehydrogenase 1 (ALDH1)+ breast cancer stem cells than that in (ALDH1) breast cancer cells from same patient (n=21). (I) Quantitative analysis of (H). Paired t-test of parametric analysis was applied by using GraphPad Prism 8.0.2. QRT-PCR data are mean ± SEM (n = 3); *P<0.05, **P<0.01.

Figure 3
Cyclin D1 augments secretion of PIWI-interacting RNAs (piRNAs) associated with breast cancer stem cells

(A) Bar graph showing the fold change of the abundant piRNAs in the medium of MCF-7 cells with regulation by cyclin D1 (>1.5 -fold change (FC)). (B) QRT-PCR analysis showing higher expression of piR-016658 in the triple negative breast cancer cells SUM159, MDA-MB-231 and Hs-578t compared with luminal MCF-7 cells. (C) A comparison between 5,272 predicted target genes of piR-016658 and 160 stem cell-regulating genes indicating an overlap of 55 genes, suggesting the possibility of piR-016658 involvement in regulating breast cancer stem cells. (D) QRT-PCR analysis showing higher expression of piR-016658 in aldehyde dehydrogenase 1 (ALDH1)+ breast cancer stem cells than that in (ALDH1) breast cancer cells from same patient (n=21). (E) Quantitative analysis of (D). (F) QRT-PCR analysis showing lower level of piR-016975 in the triple negative breast cancer cells SUM159, MDA-MB-231 and Hs-578t compared with luminal MCF-7 cells. (G) A comparison between 5079 predicted target genes of piR-016975 and 160 stem cell-regulating genes overlapped 51 genes, suggesting the possibility of piR-016975 involvement in regulating breast cancer stem cells. (H) QRT-PCR analysis showing lower expression of piR-016975 in aldehyde dehydrogenase 1 (ALDH1)+ breast cancer stem cells than that in (ALDH1) breast cancer cells from same patient (n=21). (I) Quantitative analysis of (H). Paired t-test of parametric analysis was applied by using GraphPad Prism 8.0.2. QRT-PCR data are mean ± SEM (n = 3); *P<0.05, **P<0.01.

Recent studies have identified altered piRNA abundance in cancer and cancer stem cells [2,31,32]. Among the cyclin D1-regulated piRNAs, the abundance of piR-016658 was increased the most by cyclin D1. As increased secretion of piR-016658 is associated with both pancreatic and prostate cancer [32], we further assessed this piRNA in breast cancer cell lines and breast cancer patient samples. The expression of piR-016658 was higher in the basal-like breast cancer cell lines including MDA-MB-231, SUM159 and Hs578t, compared with MCF-7 cells (Figure 3B). In view of the stemness characteristics of basal-like breast cancer cells, a comparison between the 5272 predicted target genes of piR-016658 and 160 genes from the signaling pathways regulating pluripotency of stem cells of Kyoto Encyclopedia of Genes and Genomes (KEGG) database derived 55 overlapped genes including oct4, abcg2 and rex1 (Figure 3C). The expression pattern of piR-016658 with the breast cancer stem cell marker aldehyde dehydrogenase 1 (ALDH1) was assessed in tumor tissues of 21 patients with breast cancer. About 13 of the 21 samples showed higher level of piR-016658 in ALDH1+ cancer stem cells than matched ALDH1 cancer cells (Figure 3D,E). In contrast, piR-016975, which showed repression by cyclin D1, has been shown to be reduced in expression in cancer (patent WO2017147594A1). piR-016975 showed lower expression level in the basal breast cancer cell lines (Figure 3F). About 51 of the 160 stem cell regulatory genes overlapped with the 5079 predicted target genes of piR-016975, including stem cell markers sox2, sca-1 and CD133 (Figure 3G). Analysis of breast cancer samples from patients demonstrated a decrease in piR-016975 abundance in 12 of the 21 ALDH1+ breast cancer stem cell samples, compared to matched ALDH1 cancer cells (Figure 3H,I).

Cyclin D1 induction of piwil2 in ERα+ breast cancer

The abundance of piwil1 is primarily limited to germ cells, whereas piwil2 has been identified in somatic cells [32]. A recent study has demonstrated increased piwil2 abundance in breast cancer compared with non-cancer mastopathy samples [33]. Having found that piRNAs were secreted by MCF-7 cells, we next assessed the relative abundance of piwil1 and piwil2 in breast cancer cells. The correlations between cyclin D1 and piwil1 and piwil2 in expression were also defined (Figure 4A–C and Supplementary Figure S5). These studies confirmed the presence of piwil1 and piwil2 in germ cells, and the presence of piwil2 in mammary gland tumors from MMTV-ErbB2-transgenic female mice (Figure 4A). The high expression of piwil2 was detected in ERα+ breast cancer cell MCF-7 (Figure 4B), while piwil1 did not show expression in the tested breast cancer cells (Supplementary Figure S5). The expression of piwil2 was positively correlated with the abundance of cyclin D1in human breast cancer cell lines (Figure 4B–D).

Cyclin D1 induces the abundance of piwil2

Figure 4
Cyclin D1 induces the abundance of piwil2

(A) QRT-PCR analysis showing the expression of piwil2 in the breast tumor samples from MMTV-ErbB2 transgenic mice, but no expression of piwil1 was detected. Germ cell RNA was used for positive control. (B) Relative piwil2 abundance in eight human breast cancer cell lines. Germ cell RNA was used for positive control. (C) Relative cyclin D1 abundance in eight human breast cancer cell lines. (D) Positive correlation of cyclin D1 and piwil2 in expression in breast cancer cell lines (R2=0.68, P=0.0006). Pearson correlation coefficient was used for correlation analysis. (E) Western blot showing down-regulation of Piwil2 at the protein levels by Cyclin D1 knockdown in MCF-7 cells. (F) Quantitative analysis of the intensities of the bands in (E). (G) Quantitative RT-PCR analysis showing down-regulation of piwil2 mRNA by cyclin D1 knockdown in MCF-7 cells. (H) Immunoprecipitation assay demonstrating the co-pull-down of Piwil2 with Cyclin D1 in MCF-7 cells. IgG was used as negative control. Data are mean ± SEM (n=3); **P<0.01.

Figure 4
Cyclin D1 induces the abundance of piwil2

(A) QRT-PCR analysis showing the expression of piwil2 in the breast tumor samples from MMTV-ErbB2 transgenic mice, but no expression of piwil1 was detected. Germ cell RNA was used for positive control. (B) Relative piwil2 abundance in eight human breast cancer cell lines. Germ cell RNA was used for positive control. (C) Relative cyclin D1 abundance in eight human breast cancer cell lines. (D) Positive correlation of cyclin D1 and piwil2 in expression in breast cancer cell lines (R2=0.68, P=0.0006). Pearson correlation coefficient was used for correlation analysis. (E) Western blot showing down-regulation of Piwil2 at the protein levels by Cyclin D1 knockdown in MCF-7 cells. (F) Quantitative analysis of the intensities of the bands in (E). (G) Quantitative RT-PCR analysis showing down-regulation of piwil2 mRNA by cyclin D1 knockdown in MCF-7 cells. (H) Immunoprecipitation assay demonstrating the co-pull-down of Piwil2 with Cyclin D1 in MCF-7 cells. IgG was used as negative control. Data are mean ± SEM (n=3); **P<0.01.

Given the correlation between cyclin D1 and piwil2 in MCF-7 cells, we next determined the functional relationship between these two genes. Cyclin D1 siRNA knockdown significantly reduced the abundance of piwil2 in MCF-7 cells at both the mRNA and the protein level (Figure 4E–G), suggesting the expression of piwil2 in MCF-7 breast cancer cells is dependent on the expression of cyclin D1. Immunoprecipitation assays further demonstrated the co-pull-down of Piwil2 with Cyclin D1 in MCF-7 cells (Figure 4H). Piwil2 was detectable in the breast cancer epithelial cells but not in the matched normal breast tissues from the same breast cancer patient with luminal subtype (Figure 5A,B). The expression of Piwil2 and Cyclin D1 was further validated by co-staining in ERα+ MCF-7 cells (Supplementary Figure S6A). Quantitative analysis showed the Pearson-R value and Overlap-R value both reached ∼0.8 (Supplementary Figure S6B), indicating the significant overlap and colocalization between Cyclin D1 and Piwil2 in MCF-7 cells. In consistent with these findings, interrogation of The Cancer Genome Atlas (TCGA) database demonstrated higher expression of cyclin D1 and piwil2 in ERα+ breast cancer patients compared with ERα- breast cancer (Figure 5C,D). Relapse-free survival analysis demonstrated a negative correlation between the cyclin D1 abundance and 15-year survival in ERα+ breast cancer (Figure 5E), and a negative correlation between the piwil2 abundance and 15-year survival in ERα+ PR- breast cancer (Figure 5F).

Overexpression of Cyclin D1 and Piwil2 in the ERα+ luminal subtype of breast cancer

Figure 5
Overexpression of Cyclin D1 and Piwil2 in the ERα+ luminal subtype of breast cancer

(A and B) Immunohistochemistry staining of Piwil2 in the tumor samples from ERα+ luminal breast cancer tissues (A), as well as distal normal breast tissues (B) from breast cancer patients. Representative cancer cells with Piwil2 expression were labeled with arrows. (C and D) Gene expression analysis using TCGA database showing higher levels of cyclin D1 (C) and higher levels of piwil2 (D) in the ERα+ breast cancer samples, compared with ERα- samples. **P<0.01. (E) Kaplan–Meier survival curves showing shorter relapse-free survival in the ERα+ breast cancer patients with higher cyclin D1 mRNA level (P<0.01). (F) Kaplan–Meier survival curves showing shorter relapse-free survival in the ERα+ PR- breast cancer patients with higher piwil2 mRNA level (P<0.05).

Figure 5
Overexpression of Cyclin D1 and Piwil2 in the ERα+ luminal subtype of breast cancer

(A and B) Immunohistochemistry staining of Piwil2 in the tumor samples from ERα+ luminal breast cancer tissues (A), as well as distal normal breast tissues (B) from breast cancer patients. Representative cancer cells with Piwil2 expression were labeled with arrows. (C and D) Gene expression analysis using TCGA database showing higher levels of cyclin D1 (C) and higher levels of piwil2 (D) in the ERα+ breast cancer samples, compared with ERα- samples. **P<0.01. (E) Kaplan–Meier survival curves showing shorter relapse-free survival in the ERα+ breast cancer patients with higher cyclin D1 mRNA level (P<0.01). (F) Kaplan–Meier survival curves showing shorter relapse-free survival in the ERα+ PR- breast cancer patients with higher piwil2 mRNA level (P<0.05).

Discussion

Cyclin D1 governs the secretion and subtype proportion of non-coding RNA in human breast cancer cells

Herein, sequencing of small non-coding RNA demonstrated that cyclin D1 increased the relative proportion of secreted miRNAs. Second, more than 95% of the cyclin D1-induced secreted miRNAs, such as miR-21, miR-24-1, miR-200c and let-7, are immuno-miRs. Third, cyclin D1 regulated secretion of piRNAs. Fourth, the abundance of the Argonaute family member, piwil2, was induced by cyclin D1. Although recent studies have evidenced secretion of miRNA from somatic cells [3], the molecular mechanisms coordinating patterns of secreted miRNAs and the role of cell-cycle control proteins in this process were poorly understood. The present studies revealed a new level of organization coordinating the secretion of miRNAs and regulating the production and processing of piRNAs, which was governed by the cell-cycle control protein Cyclin D1.

Cyclin D1 induces secretion of immuno-miRs from breast cancer cells

In the current studies, several miRNAs induced by cyclin D1 were previously shown to participate in the tumor immune response. TLRs can recognize and bind viral single-stranded RNA (ssRNA) sequences in dendritic cells and B lymphocytes. miR-21 and miR-93 can also bind as ligands to the TLR8 (in human) and its functional equivalent TLR7 (in mouse) in immune cells, thereby activating a TLR-mediated NF-κB signaling and increasing secretion of proinflammatory cytokines interleukin IL-6 and TNFα [11]. More than 85% of the cyclin D1-increased miRNAs in the secretome of MCF-7 cells were attributable to miR-21 and miR-93. In addition, let-7d and let-7e secretion were increased. All of the let-7 family members contain conserved GU-rich 3′ sequences, and bind to TLRs (TLR7 in mice and TLR8 in human) that are expressed by the immune cells. As a result, the binding miRNAs function as agonists of these single-stranded RNA-binding TLRs, leading to NF-κB signaling activation and secretion of interleukin IL-6 and TNFα, which promote cancer cell growth and metastasis [11,34].

Our previous work have demonstrated the cyclin D1 induction of miRNA processing and expression in breast cancer [16–18]. Here we found that cyclin D1 induces miRNA secretion from breast cancer cells (Figures 1 and 2). Targeted knockdown of cyclin D1 in MCF-7 and T47D cells significantly decreased the miRNA levels in the cellular medium (Figure 1C,F; Supplementary Figure S3). In contrary, overexpression of cyclin D1 in MCF-7 cells promoted the secretion of the tested miRNAs (Figure 2F). These studies link non-coding genome to the cell cycle regulation through cyclin D1 in human breast cancer.

The miRNA signature of cyclin D1 within the cells differs from the spectrum of the secreted miRNAs

The miRNAs induced by cyclin D1 within breast cancer cells identified in our prior study [18] differ from the secreted miRNAs described herein. The cyclin D1-regulated miRNA signature including the onco-miRs C19MC, the miR-106a-363 cluster and the miR-17-20 cluster, was associated with activation of the Wnt pathway [18]. Consistent with the importance of cyclin D1 in estrogen signaling in vivo, the miR-17-20 cluster is known to be induced by estrogen [35]. The miRNA cluster C19MC member miR-519a is up-regulated in the tamoxifen-resistant MCF-7 cells.

Cyclin D1 induces secretion of specific piRNA

Cyclin D1 induced secretion of piRNAs that are involved in stem cell function and Wnt signaling. Expression of piR-016658, a piRNA previously associated with both pancreatic and prostate cancer [32], was higher in the basal-like breast cancer cell lines. In contrast, the expression of piR-016975, which has previously been shown to be reduced in cancer (patent WO2017147594A1), was repressed by cyclin D1. Increased expression of cyclin D1 or piRNAs has been associated with the expansion of stem cell compartments. Furthermore, both piRNAs and Cyclin D1 kinase activity participate in regulation of endogenous retroviral elements. piRNAs participate in silencing of non-long terminal repeat transposons [4,5]. The functional significance of the cyclin D1-mediated piRNA expression remains to be determined.

Cyclin D1 induces piwil2 expression in ERα+ breast cancer

Herein, cyclin D1 increased the expression of piwil2 in MCF-7 cells. Emerging evidences suggest that piRNAs/ piwil2 are relevant to cancer biology. Piwil2 plays critical roles in the PIWI/piRNA pathway, and is the key player in the process of tumor initiation, progression and invasion [36,37]. Interestingly, piwil2 has been found to enrich in precancerous stem cells, and in cancer stem cells isolated from breast cancer and cervical cancer [38–41]. Cyclin D1 is overexpressed in ∼50% of human breast cancers, including the luminal subtypes (Supplementary Figure S7). Herein, piwil2 was identified in the ERα+ luminal breast cancer. Our analysis of the TCGA database showed higher expression of both piwil2 and cyclin D1 in ERα+ breast cancer tissues compared with ERα breast cancer, which is consistent with previous study showing cyclin D1 abundance correlates with ERα status in breast cancer [42].

The rational approach to targeted cancer therapy requires an improved understanding of the molecular mechanisms governing the compartmentalization of miRNAs. The role of the cell-cycle factors in controlling activity of the non-coding genome is becoming increasingly well understood, and occurring at multiple levels. The initial transcription of miRNAs is regulated by Cyclin D1 through binding in the context of local chromatin pri-miRNA, and then cleaved by the endonuclease Drosha and its partner Dgcr8 to generate hairpin-folded pre-miRNAs of 60–70 nucleotides in length [7,8]. Following transport to the cytoplasm by Exportin-5, pre-miRNAs are processed by Dicer with its partner TAR RNA binding protein (TRBP) to generate the 20–22 nucleotide mature miRNAs [7,8]. The miRNA maturation step is regulated by cyclin D1 through induction of Dicer expression [16]. Mature miRNAs then associate with Argonaute proteins to regulate target mRNA expression through RISC complex. We show herein that cyclin D1 induces the abundance of piwil2, one of four members of the Argonaute family that associate with mature piRNAs.

Conclusion

The present studies demonstrated that cyclin D1 governs the release of non-coding RNA transcriptome from breast cancer cells, and augments secretion of miRNAs that are predominantly immune-miRs promoting restraint of the anti-tumor response. In addition, cyclin D1 was showed to induce piwil2 expression, and control secretion of specific piRNAs that regulate breast cancer stem cells. Collectively, these findings demonstrate a previously unrecognized function of cyclin D1, in which cyclin D1 compartmentalizes components of the non-coding genome, augmenting secretion of immuno-miRs that govern a protumorigenic immune phenotype.

Clinical perspectives

  • The molecular mechanisms governing the secretion of the non-coding genome in human breast cancer are poorly understood.

  • We found that cyclin D1 induces the secretion of non-coding RNAs including miRNA and piRNA in human breast cancer. Those miRNAs have been reported to govern the tumor immune response and oncogenesis. miR-21 and miR-93, which bind toll-like receptor 8 to trigger a pro-metastatic inflammatory response, represented >85% of the cyclin D1-induced secreted miRNA transcripts. Furthermore, cyclin D1 regulated the expression of piwil2, and the secretion of piRNAs including piR-016658 and piR-016975 that govern breast cancer stem cell expansion.

  • The current findings will be of help in the development of novel therapeutics in treatment of ERα positive breast cancer, and shed light on our better understanding of the mechanisms through which pro-oncogene cyclin D1 regulates breast tumor initiation and progression by regulating non-coding genome secretion and cancer stem cells.

Competing Interests

The Authors declare that there are no competing interests associated with the manuscript.

Funding

This work was supported by grant from the National Key Research and Development Program of China Stem Cell and Translational Research [grant number 2016YFA0101202 (to Z.Y.)]; NIH [grant numbers R01CA070896 and R01CA086072 (to R.G.P.)]; the Breast Cancer Research foundation (to R.G.P.); the Dr Ralph and Marian C. Falk Medical Research Trust (R.G.P.); National Natural Science Foundation of China [grant numbers 81772810 and 81972476 (to Z.Y.)]; and the Top-level Clinical Discipline Project of Shanghai Pudong PWYgf 2018-05.

Author Contribution

R.G.P. and Z.Y. designed the research and wrote the paper. J.L., Q.Z., X.D., Y.G., Y.L. and Z.X. performed all experiments. Q.Z. and H.C. did data analysis. S.L., Z.W. and L.S. collected breast tumor samples and did breast cancer stem cell analysis.

Ethics Approval

Human breast cancer specimens and matching normal breast tissue samples were collected from Tongji University Shanghai East Hospital. All the procedures were approved by the Institutional Review Board (IRB) of Tongji University School of Medicine. All patients were provided with written informed consent form.

Data Availability

The materials of supporting data are available and has been included within the article.

Acknowledgements

We thank Dr J. Gao for providing Piwil2 antibody, and thank Professor Isidore Rigoutsos for providing important comments to the manuscript.

Abbreviations

     
  • 3′ UTR

    3′ untranslated region

  •  
  • C19MC

    chromosome 19 miRNA cluster

  •  
  • CCND1

    cyclin D1

  •  
  • ERα

    estrogen receptor alpha

  •  
  • MMTV

    mouse mammary tumour virus

  •  
  • piRNA

    piwi-interacting RNA

  •  
  • Piwi

    P-element induced wimpy testis

  •  
  • RISC

    RNA-Induced Silencing Complex

  •  
  • sdRNA

    snoRNA-derived RNA

  •  
  • ssRNA

    single-stranded RNA

  •  
  • TLR

    Toll-like receptor

  •  
  • TRBP

    TAR RNA-binding protein

References

References
1.
Li
S.
,
Xu
Z.
and
Sheng
J.
(
2018
)
tRNA-Derived Small RNA: A Novel Regulatory Small Non-Coding RNA
.
Genes
9
,
246
2.
Han
Y.N.
,
Li
Y.
,
Xia
S.Q.
,
Zhang
Y.Y.
,
Zheng
J.H.
,
Li
W.
(
2017
)
PIWI Proteins and PIWI-Interacting RNA: Emerging Roles in Cancer
.
Cell. Physiol. Biochem.: Int. J. Exp. Cell. Physiol. Biochem. Pharmacol.
44
,
1
20
[PubMed]
3.
Drusco
A.
and
Croce
C.M.
(
2017
)
MicroRNAs and Cancer: A Long Story for Short RNAs
.
Adv. Cancer. Res.
135
,
1
24
[PubMed]
4.
Iwasaki
Y.W.
,
Murano
K.
,
Ishizu
H.
,
Shibuya
A.
,
Iyoda
Y.
,
Siomi
M.C.
et al.
(
2016
)
Piwi Modulates Chromatin Accessibility by Regulating Multiple Factors Including Histone H1 to Repress Transposons
.
Mol. Cell
63
,
408
419
[PubMed]
5.
Lenart
P.
,
Novak
J.
and
Bienertova-Vasku
J.
(
2018
)
PIWI-piRNA pathway: Setting the pace of aging by reducing DNA damage
.
Mech. Ageing Dev.
173
,
29
38
[PubMed]
6.
Yu
Z.
,
Li
Y.
,
Fan
H.
,
Liu
Z.
and
Pestell
R.G.
(
2012
)
miRNAs regulate stem cell self-renewal and differentiation
.
Front. Genet.
3
,
191
[PubMed]
7.
Di Leva
G.
,
Cheung
D.G.
and
Croce
C.M.
(
2015
)
miRNA clusters as therapeutic targets for hormone-resistant breast cancer
.
Exp. Rev. Endocrinol. Metab.
10
,
607
617
8.
Mehta
A.
and
Baltimore
D.
(
2016
)
MicroRNAs as regulatory elements in immune system logic
.
Nat. Rev. Immunol.
16
,
279
294
[PubMed]
9.
O'Connell
R.M.
,
Rao
D.S.
and
Baltimore
D.
(
2012
)
microRNA regulation of inflammatory responses
.
Annu. Rev. Immunol.
30
,
295
312
[PubMed]
10.
Yu
Z.
,
Willmarth
N.
,
Zhou
J.
,
Katiyar
S.
,
Wang
M.
,
Liu
Y.
et al.
(
2010
)
microRNA 17/20 inhibits cellular invasion and tumor metastasis in breast cancer by heterotypic signaling
.
PNAS
107
,
8231
8236
[PubMed]
11.
Fabbri
M.
,
Paone
A.
,
Calore
F.
,
Galli
R.
,
Gaudio
E.
,
Santhanam
R.
,
Fabbri
M.
et al.
(
2012
)
MicroRNAs bind to Toll-like receptors to induce prometastatic inflammatory response
.
PNAS
109
,
E2110
E2116
[PubMed]
12.
Kato
J.
,
Matsushime
H.
,
Hiebert
S.W.
,
Ewen
M.E.
and
Sherr
C.J.
(
1993
)
Direct binding of cyclin D to the retinoblastoma gene product (pRb) and pRb phosphorylation by the cyclin D-dependent kinase CDK4
.
Genes Dev.
7
,
331
342
[PubMed]
13.
Lee
R.J.
,
Albanese
C.
,
Fu
M.
,
D'Amico
M.
,
Lin
B.
,
Watanabe
G.
et al.
(
2000
)
Cyclin D1 is required for transformation by activated Neu and is induced through an E2F-dependent signaling pathway
.
Mol. Cell. Biol.
20
,
672
683
[PubMed]
14.
Yu
Q.
,
Geng
Y.
and
Sicinski
P.
(
2001
)
Specific protection against breast cancers by cyclin D1 ablation
.
Nature
411
,
1017
1021
[PubMed]
15.
Casimiro
M.C.
,
Crosariol
M.
,
Loro
E.
,
Ertel
A.
,
Yu
Z.
,
Dampier
W.
et al.
(
2012
)
ChIP sequencing of cyclin D1 reveals a transcriptional role in chromosomal instability in mice
.
J. Clin. Invest.
122
,
833
843
[PubMed]
16.
Yu
Z.
,
Wang
L.
,
Wang
C.
,
Ju
X.
,
Wang
M.
,
Chen
K.
et al.
(
2013
)
Cyclin D1 induction of Dicer governs microRNA processing and expression in breast cancer
.
Nat. Commun.
4
,
2812
[PubMed]
17.
Yu
Z.
,
Wang
C.
,
Wang
M.
,
Li
Z.
,
Casimiro
M.
,
Liu
M.
et al.
(
2008
)
A cyclin D1/microRNA 17/20 regulatory feedback loop in control of breast cancer cell proliferation
.
J. Cell Biol.
182
,
509
517
[PubMed]
18.
Wang
G.
,
Gormley
M.
,
Qiao
J.
,
Zhao
Q.
,
Wang
M.
,
Di Sante
G.
et al.
(
2018
)
Cyclin D1-mediated microRNA expression signature predicts breast cancer outcome
.
Theranostics
8
,
2251
2263
[PubMed]
19.
Pestell
T.G.
,
Jiao
X.
,
Kumar
X.
,
Peck
A.R.
,
Prisco
M.
,
Deng
S.
et al.
(
2017
)
Stromal cyclin D1 promotes heterotypic immune signaling and breast cancer growth
.
Oncotarget
8
,
81754
81775
[PubMed]
20.
Condamine
T.
,
Ramachandran
I.
,
Youn
J.I.
and
Gabrilovich
D.I.
(
2015
)
Regulation of tumor metastasis by myeloid-derived suppressor cells
.
Annu. Rev. Med.
66
,
97
110
[PubMed]
21.
Li
H.
and
Durbin
R.
(
2009
)
Fast and accurate short read alignment with Burrows-Wheeler transform
.
Bioinformatics
25
,
1754
1760
[PubMed]
22.
Weng
W.
,
Liu
N.
,
Toiyama
Y.
,
Kusunoki
M.
,
Nagasaka
T.
,
Fujiwara
T.
et al.
(
2018
)
Novel evidence for a PIWI-interacting RNA (piRNA) as an oncogenic mediator of disease progression, and a potential prognostic biomarker in colorectal cancer
.
Mol. Cancer
17
,
16
[PubMed]
23.
Liu
J.J.
,
Shen
R.
,
Chen
L.
,
Ye
Y.
,
He
G.
,
Hua
K.
et al.
(
2010
)
Piwil2 is expressed in various stages of breast cancers and has the potential to be used as a novel biomarker
.
Int. J. Clin. and Exp. Pathol.
3
,
328
337
[PubMed]
24.
Anaya
J.
(
2016
)
OncoLnc: Linking TCGA survival data to mRNAs, miRNAs, and lncRNAs
.
PeerJ Computer Science
2
,
e67
25.
Cui
T.
,
Zhang
L.
,
Huang
Y.
,
Yi
Y.
,
Tan
P.
,
Zhao
Y.
et al.
(
2018
)
MNDR v2.0: an updated resource of ncRNA-disease associations in mammals
.
Nucleic Acids Res.
46
,
D371
D374
[PubMed]
26.
Fanini
F.
and
Fabbri
M.
(
2017
)
Cancer-derived exosomic microRNAs shape the immune system within the tumor microenvironment: State of the art
.
Semin. Cell Dev. Biol.
67
,
23
28
[PubMed]
27.
Lai
M.
and
Xiao
C.
(
2015
)
Functional interactions among members of the miR-17-92 cluster in lymphocyte development, differentiation and malignant transformation
.
Int. Immunopharmacol.
28
,
854
858
[PubMed]
28.
Yoshihara
K.
,
Shahmoradgoli
M.
,
MartÍnez
E.
,
Vegesna
R.
,
Kim
H.
,
Torres-Garcia
W.
et al.
(
2013
)
Inferring tumour purity and stromal and immune cell admixture from expression data
.
Nat. Commun.
4
,
2612
[PubMed]
29.
Wang
Y.
,
Sparwasser
T.
,
Figlin
R.
and
Kim
H.L.
(
2014
)
Foxp3+ T cells inhibit antitumor immune memory modulated by mTOR inhibition
.
Cancer Res.
74
,
2217
2228
[PubMed]
30.
Sathe
A.
,
Patgaonkar
M.S.
,
Bashir
T.
and
Reddy
K.V.
(
2014
)
MicroRNA let-7f: a novel regulator of innate immune response in human endocervical cells
.
Am. J. Reprod. Immunol.
71
,
137
153
[PubMed]
31.
Hashim
A.
,
Rizzo
F.
,
Marchese
G.
,
Ravo
M.
,
Tarallo
R.
,
Nassa
G.
et al.
(
2014
)
RNA sequencing identifies specific PIWI-interacting small non-coding RNA expression patterns in breast cancer
.
Oncotarget
5
,
9901
9910
[PubMed]
32.
Yuan
T.
,
Hu
J.
,
Yang
Y.
,
Hu
H.
,
Zhou
D.
,
Ma
M.
et al.
(
2016
)
Plasma extracellular RNA profiles in healthy and cancer patients
.
Sci. Rep.
6
,
19413
[PubMed]
33.
Litwin
M.
,
Szczepańska-Buda
A.
,
Michałowska
D.
,
Grzegrzółka
J.
,
Piotrowska
A.
,
Gomułkiewicz
A.
et al.
(
2018
)
Aberrant Expression of PIWIL1 and piwil2 and Their Clinical Significance in Ductal Breast Carcinoma
.
Anticancer Res.
38
,
2021
2030
[PubMed]
34.
Casadei
L.
,
Calore
F.
,
Creighton
C.J.
,
Guescini
M.
,
Batte
K.
,
Iwenofu
O.H.
et al.
(
2017
)
Exosome-Derived miR-25-3p and miR-92a-3p Stimulate Liposarcoma Progression
.
Cancer Res.
77
,
3846
3856
[PubMed]
35.
Castellano
L.
,
Giamas
G.
,
Jacob
J.
,
Coombes
R.C.
,
Lucchesi
W.
,
Thiruchelvam
P.
et al.
(
2009
)
The estrogen receptor-alpha-induced microRNA signature regulates itself and its transcriptional response
.
PNAS
106
,
15732
15737
[PubMed]
36.
Shahali
M.
,
Kabir-Salmani
M.
,
Nayernia
K.
,
Soleimanpour-Lichaei
H.R.
,
Vasei
M.
,
Mowla
S.J.
et al.
(
2013
)
A novel in vitro model for cancer stem cell culture using ectopically expressed piwil2 stable cell line
.
Cell J.
15
,
250
257
[PubMed]
37.
Zhang
D.
,
Wu
X.
,
Liu
X.
,
Cai
C.
,
Zeng
G.
,
Rohozinski
J.
et al.
(
2017
)
Piwil2-transfected human fibroblasts are cancer stem cell-like and genetically unstable
.
Oncotarget
8
,
12259
12271
[PubMed]
38.
Chen
L.
,
Ye
Y.
,
Chen
L.
,
Yan
Q.
,
Barsky
S.H.
,
Gao
J.X.
et al.
(
2007
)
Precancerous stem cells have the potential for both benign and malignant differentiation
.
PLoS ONE
2
,
e293
[PubMed]
39.
Feng
D.
,
Peng
C.
,
Li
C.
,
Zhou
Y.
,
Li
M.
,
Ling
B.
et al.
(
2009
)
Identification and characterization of cancer stem-like cells from primary carcinoma of the cervix uteri
.
Oncol. Rep.
22
,
1129
1134
[PubMed]
40.
Lee
J.H.
,
Jung
C.
,
Javadian-Elyaderani
P.
,
Schweyer
S.
,
Schütte
D.
,
Shoukier
M.
et al.
(
2010
)
Pathways of proliferation and antiapoptosis driven in breast cancer stem cells by stem cell protein piwil2
.
Cancer Res.
70
,
4569
4579
[PubMed]
41.
Lee
J.H.
,
Schütte
D.
,
Wulf
G.
,
Füzesi
L.
,
Radzun
H.J.
,
Schweyer
S.
et al.
(
2006
)
Stem-cell protein Piwil2 is widely expressed in tumors and inhibits apoptosis through activation of Stat3/Bcl-XL pathway
.
Hum. Mol. Genet.
15
,
201
211
[PubMed]
42.
Arnold
A.
and
Papanikolaou
A.
(
2005
)
Cyclin D1 in breast cancer pathogenesis
.
J. Clin. Oncol.
23
,
4215
4224
[PubMed]

Author notes

*

These authors have contributed equally to this work.