Interplay of lncRNA H19/miR‐675 and lncRNA NEAT1/miR‐204 in breast cancer

Long noncoding RNAs (lncRNAs) are frequently precursor RNAs of microRNAs (miRNAs) or act as competing endogenous RNAs (ceRNAs) to interact with miRNAs. To better understand the shared impact of lncRNAs and miRNAs in the regulatory post‐transcriptional network, we focused here on the relationships between (a) lncRNA H19 and miR‐675, (b) NEAT1 and miR‐204, and (c) HOTAIR and miR‐331 in plasma of early breast cancer (BC) patients. We quantified each RNA in plasma samples of 63 BC patients and 10 healthy women by quantitative real‐time PCR. In cell culture experiments, the influence of these noncoding RNAs (ncRNAs) on proliferation and apoptosis of BC cell line MCF‐7 was examined. Plasma levels of H19 (P = 0.030), NEAT1 (P = 0.030), and miR‐331 (P = 0.012) were deregulated in BC patients compared with healthy women. In both cohorts, the concentrations of H19 correlated with those of miR‐675 (P = 0.0001). Higher H19 (P = 0.001) along with lower miR‐675 (P = 0.007) levels and higher miR‐204 (P = 0.017) along with lower NEAT1 (P = 0.030) levels were detected in plasma of HER2‐positive patients compared with the other BC subgroups. Whereas the expression of HOTAIR was below the detection level, miR‐331 levels correlated with nodal status (P = 0.002) and recurrence (P = 0.012). In cell culture experiments, a competitive impact on cell proliferation and apoptosis by these ncRNAs was also documented. Our findings describe a relationship of the plasma levels of H19/miR‐675 and NEAT1/miR‐204 in the different BC subtypes; in addition, they reveal an interplay between these lncRNAs and miRNAs in the regulatory network in MCF‐7 cells, which should also be considered in the search for new diagnostic and therapeutic markers.

Long noncoding RNAs (lncRNAs) are frequently precursor RNAs of microRNAs (miRNAs) or act as competing endogenous RNAs (ceRNAs) to interact with miRNAs. To better understand the shared impact of lncRNAs and miRNAs in the regulatory post-transcriptional network, we focused here on the relationships between (a) lncRNA H19 and miR-675, (b) NEAT1 and miR-204, and (c) HOTAIR and miR-331 in plasma of early breast cancer (BC) patients. We quantified each RNA in plasma samples of 63 BC patients and 10 healthy women by quantitative real-time PCR. In cell culture experiments, the influence of these noncoding RNAs (ncRNAs) on proliferation and apoptosis of BC cell line MCF-7 was examined. Plasma levels of H19 (P = 0.030), NEAT1 (P = 0.030), and miR-331 (P = 0.012) were deregulated in BC patients compared with healthy women. In both cohorts, the concentrations of H19 correlated with those of miR-675 (P = 0.0001). Higher H19 (P = 0.001) along with lower miR-675 (P = 0.007) levels and higher miR-204 (P = 0.017) along with lower NEAT1 (P = 0.030) levels were detected in plasma of HER2-positive patients compared with the other BC subgroups. Whereas the expression of HOTAIR was below the detection level, miR-331 levels correlated with nodal status (P = 0.002) and recurrence (P = 0.012). In cell culture experiments, a competitive impact on cell proliferation and apoptosis by these ncRNAs was also documented. Our findings describe a relationship of the plasma levels of H19/miR-675 and NEAT1/miR-204 in the different BC subtypes; in addition, they reveal an interplay between these lncRNAs and miRNAs in the regulatory network in MCF-7 cells, which should also be considered in the search for new diagnostic and therapeutic markers.

Introduction
At least 90% of the genome is transcribed into noncoding RNAs (ncRNAs) (Monroig et al., 2015;Qi et al., 2016). Based on their lengths, ncRNAs are classified into two main groups: Small ncRNAs, for example, microRNAs (miRNAs), comprise transcripts of < 200 nucleotides (nt) in length, while long ncRNAs (lncRNAs) include transcripts of 200-100 000 nt in size. These ncRNA molecules are shed into the bloodstream by various cell physiological events, such as apoptosis, necrosis, and active secretion. Apoptotic and necrotic cells release ncRNAs that form complexes with specific RNA-binding proteins, such as AGO2 and HDL proteins, and/or integrate them in apoptotic bodies (Li et al., 2012;Schwarzenbach et al., 2014). ncRNAs are also released by exosomessmall membrane vesicles in size of approximately 100 nm that are actively secreted from multiple cell types, including dendritic cells, lymphocytes, and tumor cells, by exocytosis (Schwarzenbach, 2015(Schwarzenbach, , 2016b. This packaging protects ncRNAs against RNase digestion in the blood circulation and, consequently, provokes their highly stable circulation in blood (Mitchell et al., 2008). The levels of ncRNA can fluctuate in blood during the development and progression of cancer and display cancer stage-specific expression profiles. In this regard, the quantification of aberrantly expressed ncRNAs in blood as well as in other liquid biopsies (e.g., urine, saliva, pleural effusion, and cerebrospinal fluid) could serve as an assay for the development of new bloodbased biomarkers for cancer detection and course of the disease. In addition, their screening and subsequent functional analyses may deliver information on their target molecules and deregulated signaling pathways which they are involved in. Therefore, understanding of their mechanism of action could improve treatment regimens and establish new targeted therapies (Schwarzenbach, 2014).
MicroRNAs are single-stranded and about 25-ntlong ncRNA molecules. Their binding to two proteins (GW182 protein and AGO2) generates a complex called miRNA-induced silencing complex (miRISC) (Fabian and Sonenberg, 2012;H€ ock and Meister, 2008). In this complex, miRNAs bind to complementary sequences in the 3 0 UTR of their target mRNAs and repress post-transcriptionally their translation into protein (Bartel, 2009). In almost the same manner, lncRNAs suppress the translation of specific mRNAs, especially adjacent protein-encoding genes. LncRNAs can also modify gene expression through recruitment of histone modification complexes. Moreover, they are involved in alternative splicing and nuclear import (Schmitt and Chang, 2016;Schwarzenbach, 2016b). A particular feature is their interaction with miRNAs (Liu et al., 2014;Lu et al., 2016;Yoon et al., 2014) and/or to serve as precursors of miRNAs (Schwarzenbach, 2016a). Thus, lncRNAs can act as ceRNAs (competing endogenous RNAs) to de-repress gene expression by competing with miRNAs for interaction with shared target mRNAs, or they produce miRNAs, leading to repression of their target mRNAs. This miRNA/lncRNA crosstalk modulates gene expression patterns that drive physiological and pathological processes (Yoon et al., 2014), implicating that ncRNAs could be of significant relevance in breast cancer (BC).
Of particular interest for BC might be the detection of H19 which was the first lncRNA identified as a riboregulator (Brannan et al., 1990). H19 is involved in embryonic development, paternally imprinted and accordingly maternally expressed. Modifications of imprinting cause its aberrant expression and are responsible for the development of diseases. To date, numerous studies have reported its key regulatory functions in tumor development and progression (Barsyte-Lovejoy et al., 2006;Raveh et al., 2015;Vennin et al., 2015;Zhang et al., 2016). In 2007, Cai and Cullen demonstrated that this oncofetal lncRNA is the precursor RNA of miR-675 which is transcribed from the first exon of H19 (Cai and Cullen, 2007;Tsang et al., 2010). The association of H19 with tumorigenesis and invasion is assumed to be owed to the regulation of its integrated carcinogenic miR-675 (Schwarzenbach, 2016a;Vennin et al., 2015). Also in 2007, HOX antisense intergenic RNA (HOTAIR) was identified. Rinn et al. found that this lncRNA is at the boundary of two diametrical chromatin domains in the HOXC locus and transcribed in an antisense manner. It plays a critical role in chromatin dynamics and regulates gene silencing by its association with histone methylation (Rinn et al., 2007). Moreover, HOTAIR interacts with miRNAs to link miRNAs and the posttranscriptional network in cancer pathogenesis. In gastric cancer, HOTAIR forms complementary base pairing with miR-331 and inversely correlates with the occurrence of miR-331. Acting as a ceRNA of miR-331, it regulates the expression of human epithelial growth factor receptor 2 (HER2) through competition for miR-331 (Liu et al., 2014). A physical interaction was also reported between nuclear paraspeckle assembly transcript 1 (NEAT1) and miR-204, forming a reciprocal repression feedback loop in nasopharyngeal carcinoma (Lu et al., 2016). NEAT1 is localized specifically to nuclear bodies, called paraspeckles, which are irregularly shaped compartments in the interchromatin of the nucleus and involved in the development of the mammary gland. Loss of NEAT1 leads to the disintegration of paraspeckles (Standaert et al., 2014).
In the present study, we analyzed the associations of lncRNAs H19, NEAT1, and HOTAIR with miR-675, miR-204, and miR-331, respectively, in the plasma of BC patients and the BC cell line MCF-7. These investigations shall contribute to a better understanding on the shared impact of these lncRNA/miRNA pairs on regulation of the expression of their target mRNAs whose protein products are involved in BC-associated signaling pathways.

Plasma samples
Blood samples from 63 early BC patients with no signs of distant metastases were collected before surgery from June 2013 to August 2016 at the Department of Gynecology of the University Medical Center Hamburg-Eppendorf. All patients gave written informed consent to examine their blood and review their medical records according to our investigational review board and ethics committee guidelines. Blood collection and experiments were performed in compliance with the Helsinki Declaration and were approved by the ethics committee (Ethik-Kommission der € Arztekammer, Hamburg, Germany, PV5392). Regarding blood processing, uniform management concerning the specific described protocols was performed. In Table 1, patient characteristics at the time of BC diagnosis are summarized. Additionally, blood samples were collected from 10 healthy women and stored under the same conditions.
For the preparation of plasma, blood samples were centrifuged at 2700 g for 10 min. Then, the supernatant was centrifuged at 13 500 g for 10 min, to remove residual cells.

Extraction of RNA
RNA was extracted from plasma using the miRNeasy kit (Qiagen, Hilden, Germany) corresponding to the manufacturer 0 s protocol. Briefly, 1 mL QIAzol Lysis Reagent was mixed with 200 lL plasma. Following incubation at RT for 5 min, 3.5 lL of 100 fmol cel-miR-39 was added as an exogenous spike-in control to the lysate. RNA precipitation was carried out with 200 lL chloroform and 900 lL 100% ethanol in two separate steps. Then, 700 lL of the samples was added to an RNeasy MinElute spin column and centrifuged at 9300 g at RT for 15 s. Following washing the columns with 700 lL RWT and 500 lL RPE buffer, and centrifugation at 9300 g at RT for 15 s, RNA precipitation was carried out with 500 lL of 80% ethanol. RNA was eluted from the column with 14 lL RNasefree water.

cDNA synthesis and preamplification of lncRNAs
Extracted RNA was reverse transcribed into cDNA and preamplified using the RT2 PreAMP cDNA Synthesis Kit (Qiagen). To eliminate genomic DNA, 8 lL RNA was incubated with 2 lL GE buffer at 42°C for 5 min. Reverse transcription was carried out with this 10 lL mix supplemented with 4 lL BC3 buffer, 1 lL control P2, 1 lL cDNA Synthesis Enzyme Mix, 1 lL RNase, and 3 lL RNase-free water, incubated at 42°C for 30 min, and stopped at 95°C for 5 min on a MJ Research PTC-200 Peltier Thermal Cycler (Global Medical Instrumentation, Ramsey, MN, USA). For preamplification, 5 lL cDNA was mixed with 12.5 lL RT2 PreAMP PCR Mastermix and 7.5 lL RT2 PreAMP Pathway Primer Mix specific for H19, HOTAIR, and NEAT1. Cycling conditions comprised 95°C for 10 min to activate HotStart DNA Taq polymerase and 15 cycles of 95°C for 15 s and 60°C for 2 min on a MJ Research PTC-200 Peltier Thermal Cycler (Global Medical Instrumentation).

Quantitative real-time PCR of lncRNAs
For this experiment, lncRNA-specific RT 2 qPCR Primer Assays (Qiagen) for b-actin (reference RNA), H19, NEAT1, and HOTAIR were used. In a 20 lL reaction, 1 lL preamplified product was mixed with 10 lL RT 2 SYBR Green Mastermix and 0.8 lL lncRNA-specific RT 2 lncRNA qPCR assay on a twintec real-time PCR plate (Eppendorf, Hamburg, Germany). The quantitative real-time PCR was carried out at 95°C for 10 min and in 40 cycles at 95°C for 15 s and 60°C for 60 s on an Applied Biosystems 7500 fast real-time PCR device (Applied Biosystems, Darmstadt, Germany).

cDNA synthesis and preamplification of miRNAs
Extracted RNA was reverse transcribed into cDNA and preamplified using the Creating Custom RT and Preamplification Pool and TaqMan MicroRNA assays (Thermo Fisher Scientific, Darmstadt, Germany). Three microlitre RNA was reverse transcribed into cDNA with 6 lL RT primer pool, 0.30 lL 100 mM dNTPs, 3 lL 50 UÁlL À1 Multi-Scribe Reverse Transcriptase, 1.5 lL RT buffer, and 0.19 lL 20 UÁlL À1 RNase inhibitor. The 15 lL reaction was incubated on ice for 5 min and then at 16°C for 30 min, at 42°C for 30 min, and at 85°C for 5 min on a MJ Research PTC-200 Peltier Thermal Cycler (Global Medical Instrumentation). One microlitre cDNA was preamplified in a 10 lL reaction containing 5 lL TaqMan PreAmp Master Mix and 1.5 lL Custom PreAmp primer pool (Thermo Fisher Scientific). PCR was run on a MJ Research PTC-200 Peltier Thermal Cycler (Global Medical Instrumentation): one cycle at 95°C for 10 min, 55°C for 2 min, 72°C for 2 min; 16 cycles at 95°C for 15 s, 60°C for 4 min; and a terminal cycle at 99.9°C for 10 min. To avoid false-positive data (e.g., primer dimer formation or unspecific PCR products), a negative control without any templates was included from the starting point of reverse transcription.

Quantitative TaqMan real-time PCR of miRNAs and data normalization
For quantitative real-time PCR, the miRNA-specific TaqMan miRNA assays (Thermo Fisher Scientific) for miR-484 (reference miRNA), miR-675, miR-204, and miR-331 were used (Table S1). In a 10 lL reaction, 1.5 lL preamplified product was mixed with 10 lL TaqMan Universal PCR Master Mix No AmpErase UNG and 1 lL miRNA-specific TaqMan MicroRNA Assay Mix on a twin-tec real-time PCR plate (Eppendorf). The quantitative real-time PCR was performed at 95°C for 10 min and in 40 cycles at 95°C for 15 s and 60°C for 60 s on an Applied Biosystems 7500 fast real-time PCR device (Applied Biosystems).
As there is no consensus concerning the data normalization, we chose miR-484 and b-actin as references to normalize our miRNA and lncRNA data, respectively, because these RNAs showed the smallest variations. The obtained data of miRNA and lncRNA expression levels were calculated and evaluated by the DC t method as follows: DC t = mean value C t (reference miR-484 + cel-miR-39) À mean value C t (miRNA of interest) and DC t = mean value C t (reference b-actin) À mean value C t (lncRNA of interest), respectively. The relative miRNA and lncRNA levels refer to the value of 2 DCt (Table S2) .

MTT assay
Following 24, 48, and 72 h of transfection with miRNA mimics, inhibitors, or negative controls, cells were incubated with 20 lL 5 mgÁmL À1 MTT (thiazolyl blue tetrazolium bromide; Sigma-Aldrich, Saint Louis, MO, USA) in PBS at 37°C for 3 h. Then, the cells were lysed with lysis buffer (4 mM HCl, 0.1% NP40 in isopropanol), to solubilize the colored crystals. OD (optical density) was measured at 540 and 650 nm (reference) on a microplate reader (Tecan, M€ annerdorf, Switzerland). Each experiment contained three replicate wells and was repeated three times.

Apoptosis assay and flow cytometry
To examine the effect of the ncRNAs on apoptosis, MCF-7 cells were additionally treated with the topoisomerase I inhibitor camptothecin (Biovision, Milpitas, CA, USA). Twenty-four hours after transfection, MCF-7 cells on a six-well plate were treated with 4.8 lM camptothecin for 4 h and incubated for further 24 h to induce apoptosis. Nontransfected and transfected cells untreated or treated with camptothecin were incubated with 3 lL Annexin-V-FITC (BD Biosciences, San Jose, CA, USA) and 5 lL propidium iodide (Sigma-Aldrich) in the dark at 4°C for 15 min. Following incubation, 400 lL Annexin-FITC binding buffer (0.1 M Hepes pH 7.4, 1.4 M NaCl, 25 mM CaCl 2 in PBS) was added to the cells. The cells were then analyzed on a FACSCanto II flow cytometer (BD Biosciences).

Statistical analyses
Statistical analyses were performed using SPSS software package, version 22.0 (SPSS Inc., Chicago, IL, USA). Because of the skewed distribution of miRNA and lncRNA concentrations, differences in group levels for nonparametric comparisons were bivariately assessed by univariate analyses of the Mann-Whitney U-test of two independent variables. Bivariate analyses of the Spearman-Rho test were also used. Changes in proliferation and apoptosis were calculated by using ANOVA Tukey's HSD test. Missing data were handled by pairwise deletion. A P-value < 0.05 was considered as statistically significant. All P-values are two-sided. Due to the explorative nature of the study, no formal adjustment for multiple testing was performed.

Quantification of lncRNAs and miRNAs in blood plasma
The aim of our study was to investigate whether there is a relationship of the expression levels of miRNAs with those of their precursor or interaction partner miRNAs in plasma of BC patients. For quantification of the lncRNA/miRNA pairs, we selected H19/miR-675, NEAT1/miR-204, and HOTAIR/miR-331, and applied the most commonly used kits from Qiagen and Thermo Fisher Scientific. Although we have longstanding experience on liquid biopsy analyses, our attempts of establishing and comparing these kits showed that both kits had different specificities. The miRNA-specific TaqMan miRNA assay from Thermo Fisher Scientific provided most solid data on plasma miRNAs, but not on plasma lncRNAs. Inversely, the Qiagen kit delivered most solid data on plasma lncRNAs, but not on plasma miRNAs. Therefore, to obtain reliable data and avoid false-positive data (e.g., primer dimer products), we were compelled to quantify plasma miRNAs with the Thermo Fisher Scientific and lncRNAs with the Qiagen kit. However, since the measurements provide relative data, the values of lncRNAs and miRNAs can be evaluated together. Unfortunately, our measurements showed that the expression levels of HOTAIR were too low in most of our plasma samples, to be correctly interpreted. Since only a few plasma samples delivered detectable values of plasma HOTAIR, we excluded these analyses from our study. On the other hand, the quantification of miR-675 with its precursor lncRNA H19, miR-204 with its ceRNA NEAT1, and miR-331 provided robust data in all plasma samples (Table S2).

Deregulated plasma levels of H19, NEAT1, and miR-331 in BC patients
At first, we quantified and compared the relative plasma levels of H19, miR-675, NEAT1, miR-204, and miR-331 in 63 BC patients with those in 10 healthy women. The plasma levels of H19 (P = 0.030) and NEAT1 (P = 0.030) were upregulated, and those of miR-331 (P = 0.012) were downregulated in BC patients compared with healthy women. In contrast, the levels of miR-675 (P = 0.228) and miR-204 (P = 0.757) were similar in both cohorts (Fig. 1). When we compared the ncRNA levels in the single patient subgroups with those in healthy women, we found that the plasma levels of H19 were higher in patients with lymph node-negative status (P = 0.033) and HER2-positive patients (P = 0.008) than in healthy women, indicating that the aberrant expression of H19 is rather associated with early and HER2-positive tumors. Compared with healthy women, the plasma levels of NEAT1 were higher in lymph nodenegative patients (borderline P = 0.045) as well as in lymph node-positive patients (P = 0.034) and particularly, in triple-negative (TNBC) patients (P = 0.002), indicating that the deregulation of NEAT1 was rather associated with more aggressive tumors. Finally, the plasma levels of miR-331 were lower in patients with lymph node-negative status (P = 0.031) as well as with lymph node metastases (P = 0.002) and HER2-positive patients (P = 0.043) than in healthy women, indicating that the decrease in plasma miR-331 was particularly associated with advanced tumors. As observed in the entire BC patient cohort, the expression levels of miR-675 and miR-204 were also not deregulated in the single patient subgroups (Table 2).

Differences in plasma levels of H19/miR-675, NEAT1/miR-204, and miR-331 between the BC subgroups
The elevated plasma levels of H19 did not only differ HER2-positive patients from healthy women (P = 0.008) but also differ HER2-positive patients from the other patient subgroups (P = 0.001), indicating an association of plasma H19 with tumors that overexpress HER2. Although plasma miR-675 was not deregulated in the entire BC group as well as in the single BC subgroups, its levels differed between lymph node-negative and lymph node-positive BC patients (P = 0.004) as well as HER2-positive BC and the other subgroups (P = 0.007). Strikingly, H19 and miR-675 displayed opposite levels: Those of H19 were increased (P = 0.001) and those of miR-675 were decreased (P = 0.007) in HER-positive compared with the other subgroups. Alike, NEAT1 and miR-204 also showed opposite levels in HER2-positive patients: Those of NEAT1 were decreased (P = 0.030), and those of miR-204 were increased (P = 0.017). Interestingly, the plasma levels of miR-331 were associated with ER (P = 0.034) and PR (P = 0.022) status. Our BC patient cohort contained a small subgroup of recurrent patients at diagnosis who got a relapse in average of 16 years (range from 12 to 22 years). In this subgroup, the plasma levels of miR-331 were upregulated (P = 0.012, Table 2).

Significant association of plasma levels of H19 with miR-675
As shown by the scatter plot in Fig. 2 and summarized in Table 2,   between the plasma levels of H19 and miR-675 (P = 0.007) in both cohorts of BC patients and healthy women, supporting the fact that miR-675 is a derivative of H19 ( Fig. 2A). No correlation between the plasma levels of NEAT1 and miR-204 (P = 0.070) could be observed, indicating that the link of NEAT1 as ceRNA to miR-204 does not explicitly mean a reciprocal dependency of their concentrations (Fig. 2B).

Impact of H19/miR-675 and NEAT1/miR-204 on cell proliferation and apoptosis
To analyze the functions of lncRNAs (H19 and NEAT1) those relatively basal levels are high in MCF-7 cells, they were inhibited in this cell line by the corresponding siRNAs. We performed our analysis with four siRNAs of H19 (H19.1, H19.2, H19.3, and H19.4 inhibitors) that bind to different sequences of H19 and four siRNAs of NEAT1 (NEAT1.1, NEAT1.2, NEAT1.3, and NEAT1.4 inhibitors) that bind to different sequences of NEAT1. The use of siRNAs of NEAT1.1, NEAT1.2, H19.1, and H19.4 showed the best results, and their impacts on cell proliferation and apoptosis are depicted. We also transfected both miR-675 mimic and miR-675 inhibitor into MCF-7 cells to analyze their influence on the interplay with its derivative of H19, permitting us to better determine the single effects of H19 and miR-675. In addition, we overexpressed miR-204 by miR-204 mimic because its relatively basal levels were about 20 times lower than those of the other ncRNAs and because of the reciprocal repression of NEAT1 and miR-204. This allows us to intensify and localize the effect of miR-204 that could otherwise be influenced by the inhibitory effect of NEAT1.

Discussion
In the current study, we demonstrated the significant, positive relationship of the plasma levels of miR-675 with those of its precursor lncRNA H19 in BC patients. In particular, the plasma levels of H19 seem to be associated with early HER2-positive BC, since they were significantly higher in lymph node-negative and HER2-positive patients than in healthy women. In contrast to its precursor H19, the levels of miR-675 were not deregulated in the entire BC cohort and Fig. 4. Effects on apoptosis by H19/miR-675 and NEAT1/miR-204. MCF cells were transfected with miScript inhibitor negative control, AllStars negative control small interfering RNA, NEAT1.1 inhibitor and/or miR-204 mimic, and inhibitors of H19.1 and H19.4 (A), with miScript inhibitor negative control, AllStars negative control small interfering RNA, NEAT1.2 inhibitor and/or miR-204 mimic, and miR-675 (B), and additionally treated with the topoisomerase I inhibitor camptothecin (C), with miScript inhibitor negative control, AllStars negative control small interfering RNA, inhibitors of H19.1 or H19.4, miR-675 mimic or inhibitor and both, H19.1 inhibitor plus miR-675 mimic or inhibitor and H19.4 inhibitor plus miR-675 mimic or inhibitor (D), and additionally treated with the topoisomerase I inhibitor camptothecin (E). The bar charts summarize the data on the apoptotic effect by these lncRNAs and miRNAs as derived from the FACSCanto II device. The experiments were repeated two times and showed similar results, indicated by error bars representing standard deviation. The P-values refer to the comparison of the indicated ncRNAs with the average of the two neg. controls which displayed similar values. Below the bar charts, exemplary diagrams of the FACS analyses using MCF cells untreated and treated with Camptothecin are shown. Cells were labeled with Annexin-V-FITC and propidium iodide. Cell fragments only positive for propidium iodide can be found in the upper left corner (Q1). Late apoptotic as well as necrotic cells can be found in the upper right corner (Q2), since they are positive for Annexin and propidium iodide. Living cells are negative for Annexin and propidium iodide and, therefore, can be found in the lower left corner (Q3). Early apoptotic cells are only positive for Annexin and located in the lower right corner (Q4). The size for each population (%) is given in the corresponding area.
subgroups. In our experimental model, H19 stimulated cell proliferation and inhibited apoptosis, whereas miR-675 did not affect both cell processes, but affected cotransfected H19. Our findings suggest that in our study, H19 displayed oncogenic features in contrast to miR-675. As expected, there was no correlation of the plasma levels between NEAT1 and miR-204, since their formation of a reciprocal repression feedback loop does not necessarily postulate such a relationship. High plasma levels of NEAT1 were most significantly associated with lymph node-positive and TNBC patients, suggesting rather its role in advanced BC. Inhibition of NEAT1 inhibited cell proliferation and increased apoptosis, indicating that NEAT1 may stimulate cell proliferation and impede apoptosis. However, overexpression of miR-204 could alleviate the stimulatory effect on cell proliferation by NEAT1. Our findings suggest that NEAT1 acts as an oncogene, whereas miR-204 acts as a tumor suppressor. Finally, our data on plasma HOTAIR were not analyzable. As recognizable by real-time amplification curves, TaqMan PCR frequently delivered false-positive PCR products containing primer dimer products. So, the data on plasma miR-331 were analyzed without considering the data on its ceRNA. In particular, we found that the plasma levels of miR-331 decreased in lymph node-positive BC and increased with recurrence.
The significant correlation between the plasma levels of H19 and miR-675 that we detected in our investigations is due to the fact that miR-675 is derived from and processed out of H19 (Cai and Cullen, 2007). In our study, the levels of H19 were increased in the plasma of BC patients, in particular its increase could be observed in the lymph node-negative and aggressive subtype of HER2-positive BC compared with healthy women. In contrast to its precursor lncRNA H19, we detected no deregulated levels of miR-675 in the entire BC cohort as well as in the single BC cohorts. Our findings suggest that the excision of miR-675 out of H19 occurs more infrequently in our BC population. However, we observed that the levels of miR-675 significantly differed between nodal-positive and nodalnegative status, and along with the levels of H19, they differed between HER2-positive subtype and the other subtypes. Interestingly, in HER2-positive patients, H19 and miR-675 concentrations were significantly higher and lower, respectively, indicating, again, a relationship between H19 and miR-675. So far as we know, there are numerous publications on H19 measurements, whereas only one study exists on assessment of the expression status of miR-675 in BC patients. In this study, Zhai et al. showed that the levels of miR-675 were significantly upregulated in formalin-fixed paraffin-embedded (FFPE) tissue from BC patients compared with controls, but they did not differ between lymph node-positive and node-negative BC. These inconsistent data indicate that the miR-675 levels in plasma do not reflect its expression in FFPE tissues. In line with our data, there are some studies showing increased plasma levels of H19 in BC patients (Jiao et al., 2018;Yu et al., 2018;Zhang et al., 2016), and similar to our study, Yu et al. detected significant differences in H19 concentrations between TNBC and non-TNBC in a bivariate meta-analysis model (Yu et al., 2018). In addition, our cell culture experiments show that H19 stimulated proliferation and inhibited apoptosis in MCF-7 cells, whereas miR-675 had no effect on both processes. However, miR-675 mimic could alleviate the inhibitory effect on cell proliferation by siRNA H19. Discrepant data were obtained for the apoptotic effect of miR-675 on H19. Curiously, both overexpression and inhibition of miR-675 alleviated the apoptotic effect by siRNA H19.1although the experiment was repeated twice and an opposite effect was expected. Possibly, an interaction between cotransfected H19 and miR-675 could explain this conflict in data. Analog to our findings Berteaux et al. also demonstrated the oncogenic character of H19. They reported that H19 promoted cell proliferation through the G1-S transition, as the H19 promoter is activated by E2F1 in BC cells (Berteaux et al., 2005), while Li et al. found that H19 promoted proliferation and invasion of BC cells through the miR-152/DNA methyltransferase 1 axis (Li et al., 2017a). Furthermore, Sun et al. (2015) described H19 as an estrogen-inducible gene that plays a key role in survival and estrogeninduced proliferation of MCF-7 cells. Si et al. (2016) detected that H19 attenuated cell apoptosis in response to paclitaxel treatment by inhibiting the transcription of pro-apoptotic genes BIK and NOXA. As reported by Vennin et al. (2015), the aggressive phenotype of BC cells may depend on the increased cell proliferation and migration in vitro, as well as the increased tumor growth and metastasis in vivo by the overexpression of H19/miR-675.
Whereas there is a positive relationship between the levels of H19 and miR-675 triggered by processing of miR-675 out of its precursor H19 (Cai and Cullen, 2007;Tsang et al., 2010), there is a physical interaction and a reciprocal repression feedback loop between NEAT1 and miR-204 (Lu et al., 2016). In our study, we detected that the plasma levels of NEAT1 were upregulated in the entire BC patient cohort. Notably, the levels of NEAT1 were upregulated in lymph nodepositive and TNBC patients. Along with miR-204, the levels of NEAT1 differed between HER2-positive status and the other respective status. A significant upregulation of NEAT1 was observed in tumor tissues of BC patients by several previous studies (Arshi et al., 2018;Choudhry et al., 2015;Li et al., 2017b;Zhang et al., 2017b;Zhao et al., 2017). Similar to our findings, Zhang et al. (2017b) observed that high expression of NEAT1 in BC tissues was closely related to lymph node metastasis and stimulated cell proliferation. Likewise, Zhao et al. detected that NEAT1 expression was significantly upregulated in BC tissues compared to adjacent normal tissues and that higher NEAT1 expression was positively associated with lymph node metastasis. They also reported that NEAT1 promoted cell invasion and proliferation by negatively regulating miR-218 in BC (Zhao et al., 2017). Moreover, Choudhry et al. (2015) demonstrated that induction of NEAT1 in hypoxia led to accelerated cellular proliferation, improved clonogenic survival, and reduced apoptosis. As well, Ke et al. (2016) showed that downregulation of the expression of NEAT1 induced apoptosis in BC cells. These and our findings indicate that NEAT1 seems to be an oncogene. On the other hand, Shen et al. (2017) revealed that miR-204 acts as a tumor suppressor, since miR-204 suppressed the malignant behavior of BC cells. Wang et al. (2015) reported that overexpression of miR-204 inhibited proliferation and promoted apoptosis in BC cells by targeting the Janus kinase 2 (JAK2). In addition,  discovered a decreased miR-204 expression in tissues of BC patients that was significantly associated with tumor stage, metastasis, and survival. Our BC patient cohort displayed no deregulated levels of miR-204 in plasma, but cotransfected miR-204 could alleviate the inhibitory effect on cell proliferation by siRNA NEAT1. The effect on apoptosis by downregulated NEAT1 was weaker than that by overexpressed miR-204. However, the effect on apoptosis by cotransfected siRNA NEAT1 and miR-204 mimic was similar to the effect by miR-204 alone. Supplementing camptothecin to MCF-7 cells supported further our findings, suggesting that NEAT1 affects cell proliferation and apoptosis in context of miR-204, and pointing to a reciprocal repression feedback loop in BC as observed in nasopharyngeal carcinoma (Lu et al., 2016).
Finally, we also planned to investigate the relationship between HOTAIR and miR-331. However, we obtained false-positive data, possibly because of the too low plasma levels of HOTAIR. Our amplification curves showed that in most cases, the PCR products contained primer dimers products. Whereas its ceRNA HOTAIR was not detectable in plasma of our BC cohort, the deregulated amounts of miR-331 that particularly correlated with lymph node status and recurrence could be detected in BC plasma. To date, the inverse correlation of the expression levels between HOTAIR and miR-331, and the involvement of miR-331 in the post-transcriptional regulatory network of HOTAIR have been reported for cervical and gastric cancer (Liu et al., 2014;Zhang et al., 2017a;Zhou et al., 2014). Chang et al. showed that in hepatocellular carcinoma (HCC), a highly invasive tumor with a high recurrence rate, miR-331 is one of the most significantly overexpressed miRNAs and highly associated with metastasis. In xenograft mice, the inhibition of miR-331 resulted in marked inhibition of proliferation and metastasis of HCC (Chang et al., 2014). These and our findings demonstrate the involvement of miR-331 in advanced and recurrent tumors.

Conclusion
In conclusion, our quantifications of plasma H19/miR-675 and NEAT1/miR-204 suggest a crosstalk between these lncRNAs and miRNAs that might drive BC development and progression. This crosstalk between lncRNAs and miRNAs further complicates the mechanism of post-transcriptional regulation of gene expression. Based on these preliminary data, further analyses in a larger BC cohort as well as functional analyses in cell lines from different BC subtypes are needed to reveal the mechanisms of post-transcriptional regulation in more detail. Notably, future targeted therapies should consider that there is an interplay between lncRNAs and miRNAs, and that they do not act alone to affect mRNA expression.

Supporting information
Additional supporting information may be found online in the Supporting Information section at the end of the article. Table S1. Sequences of ncRNA assays used in the study. Table S2. Measured data of miRNA and lncRNA expression levels normalized and evaluated by the DC t method.