19 publications mentioning hsa-mir-661

Open access articles that are associated with the species Homo sapiens and mention the gene name mir-661. Click the [+] symbols to view sentences that include the gene name, or the word cloud on the right for a summary.

1

3.7To demonstrate that miR661‐associated downregulation of H6PDH and PKM2 had a direct effect on the observed changes in cell metabolism, we transiently overexpressed H6PDH (DLD1‐miR661‐H6PD), PKM2 (DLD1‐miR661‐PKM2) and both targets simultaneously (DLD1‐miR661‐H6PD‐PKM2) in DLD1‐miR661 cells.

To demonstrate that miR661‐associated downregulation of H6PDH and PKM2 had a direct effect on the observed changes in cell metabolism, we transiently overexpressed H6PDH (DLD1‐miR661‐H6PD), PKM2 (DLD1‐miR661‐PKM2) and both targets simultaneously (DLD1‐miR661‐H6PD‐PKM2) in DLD1‐miR661 cells.

Moreover, the expression levels of Snail, Slug and Vimentin were not affected by miR661 overexpression in SW620 cells compared with control cells (Nectin expression was used as a validated described target for miR661) (Fig.   S3C).

We hypothesized that miR661 may target, in addition to EMT‐related targets, some other metabolic targets that will affect cell homeostasis.

In ovarian cancer, miR661 promotes cell proliferation by targeting INPP5J (Zhu et al., 2015), but in gliomas it inhibits cell proliferation, migration and invasion by targeting hTER (Li et al., 2015).

In contrast, Reddy et al. (2009) reported that miR661 inhibited the expression of metastatic tumor antigen 1 (MTA1) in invasive breast cancer cells and reintroduction of this miR was able to inhibit motility, invasiveness, and tumorigenesis.

Kaplan–Meier plots for disease‐free survival showed an association between expression levels of miR661 and clinical outcome in stage‐ II CC patients (n = 136; 77.2% low miR661 expression with 11.4% relapse vs.

3.2miR661 induces an increase of SO [·] anions and high Ψm at mitochondriaWe hypothesized that miR661 may target, in addition to EMT‐related targets, some other metabolic targets that will affect cell homeostasis.

Although we have not demonstrated a direct effect of miR661 in PKM and H6PD downregulation by luciferase assays using short regions of their 3′UTRs, our results showed that miR661 downregulated mRNA and protein levels of PKM2 and H6PD in non‐metastatic and metastatic CC cell lines.

Table  1A indicates significantly altered biochemicals with statistical significance (P ≤ 0.05) or with approaching significance (0.05 <  P < 0.10) from a dataset analysis with a total 323 named biochemicals differentially detected in the study (red indicates significantly upregulated metabolites and green indicates downregulated ones) (File S1: total of significant metabolites ratio DLD1‐miR661/DLD1‐control (P < 0.05).

First, we confirmed the effect of miR661 on the downregulation of Nectin1, a described miR661 target (Vetter et al., 2010).

When ATPase was inhibited with oligomycin to block ATP production from mitochondria, DLD1‐miR661 cells were not able to increase glycolysis, indicating an altered mitochondrial respiration as oligomycin injection, does not contribute to upregulate ECAR.

To further segregate the EMT‐associated effects on cell metabolism from other miR661 metabolic targets, we overexpressed miR661 in the invasive SW620 colon cancer cell line (Fig.   S3A).

Kaplan–Meier plots for disease free‐survival showed an association between high expression levels of miR661 and poorer clinical outcome in stage‐II CC patients.

Regardless of PK, we hypothesized that miR661 may also target the M2 isoform, which is frequently overexpressed in cancer cells, similar to embryonic tissues.

Figure 5Pathways and targets modulated by miR661 overexpression in colon cancer DLD1 cell line.

As shown in Fig.   1D, overexpression of miR661 diminished the expression of the epithelial marker E‐cadherin.

In silico prediction of miR661 targets related to cell metabolism pointed to PKLR and H6PD as the two main candidates, which was validated by gene expression analysis (Fig.   5B).

In contrast, miR661 increased the expression levels of the EMT‐associated genes N‐cadherin, Slug and Vimentin, which are normally not expressed in the markedly epithelial DLD‐1 cells.

Overexpression of miR661, both in DLD1 and SW620 cells, diminished the expression levels of PKLR and H6PD (Fig.   5B).

22.8% high expression with 25.8% relapse) and stage‐ III CC patients (n = 81; 75.3% low miR661 expression with 39.3% relapse vs.

In contrast, high expression levels of miR661 in stage‐III CC patients were correlated with a better clinical outcome, in accordance with miR661 overexpression in metastatic CC cell results.

We therefore hypothesized that miR661 may also target the M isoform, which has been described to be highly expressed in cancer cells (Mazurek, 2011).

Nevertheless, although not demonstrated by direct interaction, the final read‐out of miR661 overexpression in both cell lines (DLD1 and SW620) is the depletion of H6PDH and PKM at the transcriptional (qRT‐PCR, Fig.   5B) and post‐transcriptional levels (WB, Fig.   5C).

To summarize, miR661 induces profound changes on cell metabolism: diminished aerobic glycolysis, downregulation of pentose phosphate metabolism, and breakdown of redox homeostasis.

In this sense, depending on p53 status, miR661 may suppress (p53‐wild‐type) or promote (p53‐mutated) cancer aggressiveness by a direct effect on mdm2 and mdm4 (Hoffman et al., 2014).

In addition, it is possible that miR661 interactions may need secondary structures only achieved with the complete 3′UTR sequences, and/or the observed differences in the mRNA levels of H6PD and PKM after miR661 overexpression may occur by indirect mechanisms.

2.3HEK 293T cells were transfected using Lipofectamine 2000 (Life Technologies, Carlsbad, CA, USA) with a lentiviral vector expressing miR661 (DNA 2.0, Menlo Park, CA, USA) and packaging plasmids (Addgene, Cambridge, MA, USA).

Although we were not able to provide a mechanistic demonstration that miR661 directly regulates H6PD and PKM, the complete 3′UTR sequences contain additional predicted binding sites for miR661 that have not been analyzed here.

miR661 overexpression confers EMT and invasive properties to non‐metastatic CC cells.

Data from functional cell bioenergetic analysis indicated that overexpression of miR661 diminished the glycolytic function in the glycolytic cell line DLD1.

PKLR and H6PDH are diminished in miR661 overexpressing CC cells.

As illustrated in Fig.   1E, when miR661 was overexpressed, poorly invasive DLD1 cells gained the ability to invade through Matrigel.

Regarding glycolytic function, DLD1‐miR661‐PKM2 and DLD1‐miR661‐H6PD‐PKM2 upregulated basal ECAR (basal glycolysis after 1 h of glucose starvation) as well as the glycolytic reserve compared with DLD1‐miR661.

AcCoA Aacetyl‐CoA BNGE blue native gel electrophoresis BRR basal respiration rate CC colon cancer DFS disease‐free survival DHAP dihydroxyacetone phosphate ECAR extracellular acidification rate EMT epithelial to mesenchymal transition ETC electron transport chain FAO fatty acid oxidation FFPE formalin‐fixedparaffin‐embedded H6PD hexose‐6‐phosphate dehydrogenase HR hazard ratio IDH1 isocitrate dehydrogenase ME malic enzyme miR661 microRNA‐661 MRR maximal respiration rate OCR oxygen consumption rate OS overall survival PEP phosphoenolpyruvate PKM2 pyruvate kinase M2 PPP pentose phosphate pathway Pyr pyruvate ROS reactive oxygen species SC supercomplexes SO [‐] superoxide anions TCA tricarboxylic acid cycle Ψm mitochondrial membrane potential Tumor cells reprogram their energetic metabolism to support rapid and uncontrolled growth.

As shown in Fig.   S3B, unlike DLD1 cells, both SW620‐control and SW620‐miR661 expressed high levels of Vimentin.

This effect seems to be a general effect of miR661 in colon cancer, as overexpression of miR661 in the metastatic SW620 colon cancer cell line diminishes its tolerance to metabolic stress without altering its invasiveness properties.

To determine the major pathways altered by miR661 overexpression in DLD1 CC cells, we performed a metabolomic analysis.

After the injection of glucose, DLD1‐miR661 cells upregulated glycolysis from the basal situation, to a higher extent compared with control cells (Fig.   3A).

Lentivirus‐mediated stable overexpression of miR661.

For this purpose, we analyzed miR661 expression levels in a panel of CC cell lines.

miR661 overexpression in metastatic CC cells alters functional cell bioenergetics.

However, this controversy disappears when one takes into account that, in the first work, Snail drives the EMT phenotype and miR661 is an associated player, whereas in the second work, miR661, by targeting MTA1, induces the opposite effect.

We generated a stable DLD1 cell line overexpressing miR661 by transduction with a lentiviral vector (DLD1‐miR661).

Finally, high levels of miR661 were correlated with poorer prognosis in stage‐II CC, consistent with EMT and invasiveness promotion upon miR661 overexpression in non‐metastatic CC cells.

3.3To determine the major pathways altered by miR661 overexpression in DLD1 CC cells, we performed a metabolomic analysis.

Among them, we selected the DLD1 non‐metastatic cell line, whose miR661 expression levels were lower than that of primary colon cell lines (Ccd18‐Co, CCD841) (Fig.   S1).

Reexpression of H6PDH and PKM2 in DLD1‐miR661 leads to a partial rescue of cell bioenergetics.

We found a diminished carbon flux through aerobic glycolysis (reduced l‐lactate production and altered glycolytic function) upon miR661 overexpression.

The main metabolic pathways enriched upon miR661 overexpression are also shown (Table  1B).

Finally, the preliminary analysis of miR661 status in CC patients indicated an association between high expression levels of miR661 and poorer clinical outcome in early‐stage CC patients, contrasting with the effects observed for more advanced stage‐III patients, in agreement with in vitro data with non metastatic and metastatic CC cells.

To elucidate the role of miR661 in colon cancer (CC) we present an integrative approach, including metabolomics, analysis of cell bioenergetics, and molecular targets identification.

Values were normalized using U6snRNA (for miR661) and GAPDH expression.

In silico bioinformatic prediction of miR661 targets related to cell metabolism pointed to PKLR and H6PD as two main candidates.

But importantly, high miR661 expression levels in stage‐III CC patients tends to correlate with better, although not significant, prognosis (Fig.   7).

Figure 6Re‐expression of H6 PDH and PKM2 in DLD1‐miR661 leads to a partial rescue of cell bioenergetics.

As shown in Fig.   5C, both H6PDH and PKM2 protein levels were diminished in DLD1‐ and SW620 miR661‐overexpressing cells.

The expression data of the miR661 was categorized in two categories, high and low, selecting one cut‐point based in the minimum P‐value of the log‐rank test.

It seems that miR661 does not alter the expression of EMT markers or the invasiveness capability of SW620 metastatic cancer cell line.

In silico bioinformatic prediction of miR661 targets related to cell metabolism, indicated that PKLR and H6 PD were the two main candidates.

Expression levels of miR661 are normalized to CCD841 primary colon cancer cell line.

Expression levels of miR661 in colon cancer patients.

miR661 and gene‐expression assays were performed in a HT‐7900 Fast Real time PCR.

Based on all these findings, we suggest that miR661 is a potential relevant epigenetic regulator of redox homeostasis and cell metabolism in CC.

The translational repression of PKM or H6PDH upon miR661 binding was determined after transfection using the Dual‐Luciferase Reporter Assay System (Promega Biotech Ibérica S. L., Madrid, Spain).

These data suggest that miR661 plays an important role in regulation of energy and redox homeostasis in DLD1 human colon cancer cells.

miR661‐induced metabolic profile supports its role as a regulator of redox and metabolic homeostasis.

On one hand, the decreased levels of CI‐CIII‐CIV SC in DLD1‐miR661 indicate that these cells switch off respiration through CI to favor respiration through CII and/or directly to CoQ.

Metabolomic data and functional bioenergetics analysis were combined to gain insight into the molecular mechanisms involved in the regulatory role of miR661 on oxidative stress and metabolic reprogramming.

To assess the direct effect of miR661, 100 ng of psiCheck2‐3′‐UTR‐short‐wt‐ PKM, 3′‐UTR‐short‐mut‐ PKM, 3′‐UTR‐short‐wt‐ H6PDH or 3′‐UTR‐short‐mut‐ H6PDH were co‐transfected with 30 nM of LNA‐miR661 in HEK293T cells.

3.6Metabolomic data and functional bioenergetics analysis were combined to gain insight into the molecular mechanisms involved in the regulatory role of miR661 on oxidative stress and metabolic reprogramming.

Moreover, it cannot be discarded that PKM2 protein depletion may also be the result of a combination of the effect of miR661 at the post‐transcriptional level and/or an indirect effect of the oxidative stress on protein stability.

By doing this, DLD1‐miR661 cells may preserve the limited pool of NAD [+] and NADH (File S1).

These effects were less marked in DLD1‐miR661‐H6PD cells.

miR661 induces an increase of SO [·] anions and high Ψm at mitochondria.

3.8To investigate the clinical relevance of miR661 in CC we analyzed the putative association between the expression levels of miR661 and the clinical outcome in tumor samples from stage II (n = 136) and stage III (n = 81) CC patients (Table  S1).

This indicates that DLD1‐miR661 cells do not rely on aerobic glycolysis in a basal situation; however, when challenged to use exogenously added glucose after starvation, they have capacity to undergo glycolysis.

Importantly, NAC diminished the levels of SO [·] in DLD1‐miR661 (galactose condition), although it did not change the mitochondrial membrane potential (Ψm) in DLD1‐miR661 cells.

We conclude that miR661 may play a dual role in CC, being involved in the aggressiveness of the tumor at the onset of the metastasis but offering a potential therapeutic window against invasive tumors.

Immunoblot probed with ScafI (Cox7a2L) shows that III [2] IV [1] SCs were reduced in DLD1‐miR661.

By in silico prediction, the H6PD‐3′UTR region contains at least 24 putative binding sites for miR661, and PKM‐3′UTR at least five.

Based on these findings, we propose miR661 as a potential modulator of redox and metabolic homeostasis in CC.

To gain insights about the overall ROS status in DLD1‐miR661 cells, we analyzed the distribution of ROS species after treatment with low glucose or galactose (5 m m, 24 h) in the presence or absence of the antioxidant scavenger N‐acetylcysteine (NAC; 5 and 10 m m) (Fig.   2A).

Although these data need to be further validated in independent cohorts of stage‐II and stage‐III CC patients, it is tempting to hypothesize that the effect of miR661 in DLD1 non‐metastatic cells might mimick the situation in stage‐II CC, and the effect on SW620 may mimick the more advanced stage‐III CC.

John Wiley & Sons, LtdGlycolysis and pyruvate metabolism were profoundly altered by miR661.

Significant differences were found in the spare respiratory capacity and, more specifically, in the presence of ‘ROSgenic’ galactose as substrate (Fig.   4C, red arrow), again pointing towards a role of miR661 in increasing the overall oxidative stress.

In contrast, a dramatic decrease in the basal OCR for DLD1‐miR661 was found.

However, in a more advanced stage, similar to results in SW620 metastatic cells, increased miR661 levels might raise the oxidative stress to a threshold that is incompatible with cell survival.

Thus, we wanted to address the effect of miR661 in SW620 metastatic CC cell line.

Unlike DLD1‐Control cells, DLD1‐miR661 cells displayed morphological changes resembling an EMT phenotype (Fig.   1B).

SW620‐miR661 stable cell line.

miR661 compromises glycolysis and mitochondrial respiration.

Comparison of the glycolytic function and mitochondrial function of DLD1‐miR661 vs.

Blue‐native gel electrophoresis (BNGE) showed that DLD1‐miR661 had decreased levels of CIII associated in CI‐CIII‐CIV SC and increased levels of free‐CIII complex and CIII‐CIV SC (Fig.   3D).

Hence, miR661 seems to induce an oxidative stress which promotes EMT in the non‐metastatic DLD1 cells, correlating with the poorer prognosis in stage‐II CC patients.

This indicates that miR661 increases SO [·] at mitochondria, which can be partially alleviated with NAC; more importantly, miR661 seems to induce a metabolic reprogramming that favors the maintenance of high levels of mitochondrial membrane potential (Ψm) which is not rescued by NAC treatment.

However, miR661 seems dramatically to compromise both functions in DLD1 cells.

To obtain an insight into how DLD1‐miR661 cells manage mitochondrial respiration in the presence of high ROS, we studied the assembly of mitochondrial complexes into supercomplexes (SCs).

2.11DLD1‐miR661 and DLD1‐Control cells were prepared as indicated by Metabolon Inc.

This indicates that miR661 is sufficient to induce EMT properties in DLD1 non‐metastatic colon cancer cells.

A clear example of this situation is miR661.

As DLD1‐miR661 cells are sensitive to energy and oxidative stress, we also checked AMPK activation as a survival mechanism.

In a mo del of Snail1‐induced EMT in breast cancer cells, Vetter and collaborators identified miR661 as a key Snail1‐induced miR required for efficient invasion (Vetter et al., 2010).

DLD1‐miR661 and DLD1‐Control cells were prepared as indicated by Metabolon Inc.

When glucose was diminished in the growing medium, DLD1‐miR661 cells became hypersensitive to glucose limitations, as seen by the reduction in the total number of cells.

Moreover, DLD1‐miR661 cells displayed similar levels of both MRR and BRR, which is an indication that DLD1‐miR661 cells are at their top limit of respiration and therefore have a reduced mitochondrial function (Fig.   3C).

To summarize, miR661 induces a clear depletion of anabolic pathways from glycolytic intermediates and TCA cycle, and displays a higher dependency on FAO intermediates (Table  1B, File S1).

We hypothesized that miR661 might target metabolic genes, explaining its biological function and characterizing the role of miR661 in cell metabolism, integrating data from metabolomic analysis with functional cell bioenergetics.

In addition, as revealed by the analysis of the SCAFI adaptor protein, it was found that DLD1‐miR661 cells favor free CIII over CIII‐CIV assembly (Fig.   3D, right panel).

Ratios between mean values in DLD1‐ miR661 and DLD1‐Control for each metabolite are indicated.

The effect on these parameters was less prominent in DLD1‐miR661‐PKM2.

In summary, these results reflect a miR661‐dependent metabolic reprogramming affecting both glycolysis and mitochondrial respiration to preserve anabolic processes (mitochondrial anabolic mode).

miR661 diminishes the tolerance to metabolic stress in both non‐metastatic and metastatic CC cells and we propose hexose‐6‐phosphate dehydrogenase (H6PD), and pyruvate kinase M2 (PKM2) as two major players on miR661‐induced metabolic reprogramming.

To investigate the clinical relevance of miR661 in CC we analyzed the putative association between the expression levels of miR661 and the clinical outcome in tumor samples from stage II (n = 136) and stage III (n = 81) CC patients (Table  S1).

We checked this possibility in DLD1‐miR661 cells.

However, when FCCP (0.4 μ m) was used to free the H [+] gradient through the inner mitochondrial membrane and uncouple the electron transporter system (ETC) from the synthesis of ATP, the maximal respiration rate (MRR) was strongly reduced in DLD1‐miR661 cells (Fig.   3B).

John Wiley & Sons, Ltd Glycolysis and pyruvate metabolism were profoundly altered by miR661.

By doing so, DLD1‐miR661 cells maintain high levels of membrane potential, which also contributes to the use of TCA intermediates for biosynthesis instead of oxidative phosphorylation.

This might be a consequence of the extreme membrane potential that DLD1‐miR661 cells face.

2

Strikingly, the ectopic expression of miR-661 strongly modified the spatial phosphorylation of myosin II, while in contrast, overexpression of either miR-612 or miR-940 inhibited myosin II phosphorylation (Figure 3).

Interestingly, miR-940 is moderately to highly expressed (pink to red node) in 7 tissues out of 8 whereas miR-661 and miR-612 both show little or no expression (white node) in the same tissues.

Lastly, 7 hubs from the 11 hubs discovered by DIANA-microT are retrieved within the 40 degree sorted nodes on the TargetScan network (in decreasing order by degree: miR-548c-3p, miR-590-3p, miR-579, miR-186, miR-513a-3p, miR-661, miR-495 and lastly miR-940).

Similarly, our results appear to be in agreement with reports from Vetter et al., who have shown that miR-661 contributes to breast cancer cell invasion through the targeting of Nectin-1 and StarD10 52.

These two results show that miR-661 greatly enhances cell motility and division when overexpressed.

Furthermore, we demonstrated here that miR-612 and miR-661 also regulate cell motility via opposite effects on myosin II phosphorylation.

miR-940 is located on chromosome 16, miR-661 on chromosome 8, and miR-612 on chromosome 11 (Supplementary Figure S5).

Each condition (miR-612, miR-661, miR-940, and siRNA-AllStars) was represented in triplicate, leading to 30 observations per experiment and per condition (Supplementary Figure S14).

The second central hub group, named “assorted club 2”, was composed of miR-612, miR-661 and miR-940.

RPE1 cells were independently transfected with mimics of miR-661, miR-612 and miR-940.

Finally, an increase in actin filament staining was observed in RPE1 cells treated with miR-661 (P-value = 0.012, Figure 3e).

In contrast, the miR-661 mimic revealed a higher number of myosin-decorated stress fibres and highly contracted cells with dense fibres (Figure 3c).

The 5 microRNAs are, in decreasing order of centrality: miR-548c-3p, miR-590-3p, miR-661, miR-186 and miR-940.

RPE1 cells were transfected with AllStars siRNA or mimics of miR-612, miR-661 or miR-940 at 20 nM for 48 h. The positive control (Y27632) was added to the siRNA AllStars -transfected cells at 10 µM for the last 24 h. The transfected cells were then plated on the micropatterns.

Effect of miR-612, miR-661 and miR-940 on RPE1 migration and proliferation.

RPE1 cells were seeded in six-well microtiter plates, cultured for one day, and then transfected with mimics of miR-612, miR-661, or miR-940 or with a negative control siRNA (siRNA AllStars) at a final concentration of 20 nM.

In the near future, we will further investigate the mechanism of action of miR-661 with regard to the p53 status of the cells, as it was recently reported that miR-661 may either suppress or promote cancer aggressiveness, depending on the p53 status 53.

The microRNA mimics (miR-612, miR-940 and miR-661) were purchased from Thermo Scientific (Waltham, Massachusetts) Dharmacon (miRIDIAN).

To pool the four experiments together, each condition of each experiment was normalised to the median number of cells in the siRNA-AllStars condition (), considering all replicates: where N is the number of counted cells and Ñ the normalised number of cells for each condition, replicate and experiment; x is the experiment (1, 2, 3 or 4), r, the replicate for each experiment (1, 2 or 3), and, the different conditions (siRNA-AllStars or mimics of miR-612, miR-661 or miR-940).

is composed of 3 microRNAs (Supplementary Table IV), namely miR-940, miR-661 and miR-612.

Involvement of miR-661, miR-612 and miR-940 in small GTPase signalling.

However, we clearly observed that miR-661 induced a spatial reorganization of MLCII from the border of the cells to the entire cell surface.

Assorted club 2 is composed of 3 microRNAs (Supplementary Table IV), namely miR-940, miR-661 and miR-612.

RPE1 cells were transfected with miR-612, miR-661 or miR-940 mimics and immunolabeled for phosphomyosin II and actin fibres on 1000 µm [2] circular fibronectin patterns.

3

After overexpression, the mimic miRNAs were present at higher expression levels than the endogenous miRNAs, especially miR-641 and miR-661 (Supplementary Figure S3C).

These results indicated that overexpression of the candidate miRNAs (miR-302c, miR-320a, miR542-3p, miR-641, miR-661 and miR-940) could, indeed, reduce the surface expression of MICB protein.

Finally, we concluded that not only MICB expression was regulated by miR-320a and miR-940 but could also be controlled by other candidate miRNAs such as miR-302c, miR-542-3p, miR-641 and miR-661.

Using a computational prediction, we have found that some candidate miRNAs binding sites were on or close to transcription factor binding sites such as miR-641 and miR-661 having binding sites on the activator protein 1 (AP-1) which regulates gene expression in response to cellular stress, bacterial and viral infections [39].

The overexpressing miRNA mimic vectors (miR-302c, miR-320a, miR-542-3p, miR-641, miR-661 and miR-940 mimics) were transiently transfected into 293T cells and were co -transfected with luciferase reporter constructs containing wild-type 3-′UTR of MICB (pMICB_3U).

Thus, the candidate miRNAs (miR-302c, miR-320a, miR-542-3p, miR-641, miR-661 and miR-940) could regulate 5′-UTR of MICB by direct binding.

Consequently, to confirm hypothesis of these studies we constructed plasmids to overexpress candidate miRNAs (miR-302c, miR-320a, miR-542-3p, miR-641, miR-661 and miR-940 mimic) and then these mimic miRNAs or control mimic plasmids were co -transfected with reporter construct containing wild-type 5′-UTR (pMICB_5U) into 293T cells.

Overexpression of candidate miRNAs in HeLa was performed by transfection with miRNA mimic vectors (miR-302c, miR-320a, miR542-3p, miR-641, miR-661 and miR-940) or control mimic vector.

These results indicated that miRNA candidates (miR-320a, miR-542-3p, miR641, miR-940, miR-302c and miR-661) interacted with a predicted binding site on 3′-UTR of MICB.

One type of constructs contained the mutated binding sites of both known miRNAs (miR-20a, miR-93 and miR-106b) and nine novel miRNAs (our candidate miRNAs, miR-320c, miR-320a, miR-320b, miR-320c, miR-320d, miR-542-3p, miR-641, miR-661 and miR-940) and another type contained only the mutated binding sites of known miRNAs as a positive control (Figure 3A).

4

7 miRNAs (hsa-miR-144,hsa-miR-133a,hsa-miR-365, hsa-miR-424,hsa-miR-500, hsa-miR-661,hsa-miR-892b) had different gene expression levels between active TB and healthy controls; 4 of them (hsa-miR-144,hsa-miR-365 and hsa-miR-133a, hsa-miR-424) were up-regulated and 3 of them (hsa-miR-500, hsa-miR-661,hsa-miR-892b) were down-regulated in active TB patients.

Five miRNAs (hsa-miR-130a*, hsa-miR-296-5p, hsa-miR-493*, hsa-miR-520d-3p, hsa-miR-661) had different expression levels between latent TB and healthy controls; all of them except hsa-miR-296-5p were up-regulated in healthy controls.

Hsa-miR-661 was up-regulated in comparisons of healthy controls with active TB, as well as healthy controls with LTBI.

5

In another study, Vetter and coworkers showed that miR-661 expression in MCF7 breast cancer cells conditionally overexpressing the EMT master regulator SNAI1 contributes to breast cancer cell invasion by targeting cell-cell adhesion Nectin-1 and the lipid transferase StarD10 messengers [69].

6

Two other miRNAs have been shown to indirectly modulate iNOS expression: miR-155 and miR-661.

When the miRNA miR-661 was depleted in these HBx -expressing cells, HBx activity was impaired, leading to enhanced iNOS and nitrite production.

7

miR-25, miR-32, miR-661 and miR-339–5p target MDM2 to up-regulate p53 protein levels and function [27– 30].

8

Interestingly, many of the genes containing such highly differentiated SNPs in the Alu-miRNA sites within their 3′UTRs (listed in Table 1) have target sites (encompassing these SNPs) for miRNAs that are primate-specific (miR-661 29, miR-1202 53), human-specific (miR-4739, miR-5095) 54, involved in the regulation of p53 signaling (miR-660 55, miR-661 29, miR-1285 56) or in the apoptosis pathway (miR-17 44 45, miR-30b 57, miR-106a-3p 45, miR-612 58).

We believe that interaction of miR-15a-3p with Alus may have role in skin aging and adaptation of skin to stress in primates, similar to what has been shown before for interaction of exonized Alu with miR-661, in apoptosis 29 48.

9

For example, we found that miR-520d*/miR-661 co-down-regulate EGFR and KRAS.

10

Expression levels of miR-16-1 [*], miR-541, miR-637, miR-661 or miR-608 are not altered in breast cancer tissue (Supplementary Figure 2A).

11

Conversely, miR-335, miR-206, miR-31, miR-145, miR-661 and miR-126 have been identified as metastasis suppressor miRNAs in human breast cancer [21], [22], [23], [24], [25], [26], [27].

12

Individual analysis of four miRNAs (miR-125b-5p, miR-423-5p, miR-661 and miR-3934-5p), chosen alleatorily, performed by qRT-PCR in a subset of CRCs and corresponding PBTs (cases 2, 3, 5 and 6), showed no statistical difference of their expression levels among the pairs (P<0.05) (Fig 6).

13

miR-661 is another C/EBPα target [25].

14

With the same binding energy cutoff, we found two human miRNAs hsa-miR-370 and hsa-miR-661 having MREs within NS1 genes of all types of dengue virus.

The start binding position 771 of hsa-miR-370 was highly conserved among gene population of types 1-3 while hsa-miR-661 had a highly conserved start binding position 789 among gene population of all types.

15

Based on the high predicted frequency, we chose 10 candidates (has-miR-144-3p, hsa-miR-18b-5p, hsa-miR-222-5p, hsa-miR-221-3p, hsa-miR-24-3p, hsa-miR-27a-5p, hsa-miR-27b-3p, hsa-miR-9-5p, hsa-miR-210-3p, hsa-miR-661) that promote BC development (confirmed from previous studies) [21– 30] for verification (Supplementary Figure S1).

16

Tuttelmann et al. also identified an azoospermic man with a deletion and another with a duplication on 8q24.3, encompassing the genes PLEC1 and MIR661 [24].

17

In this study, we used time-lapse video microscopy to observe cell movement in the scratch-wound assay, validated the results using the transwell migration assay, and identified the migration-suppressing miRNA (miR-134) and migration-facilitating miRNAs (miR-1247, miR-1244, miR-146b-3p, miR-1471, miR-188-3p, miR-661, miR-891a, miR-891b and miR-767-5P) in SK-HEP-1 cells.

18

miRNAs with a role in metastasis in BC include miR-7 [138, 139], miR-17/20 [140, 141], miR-22 [142– 144], miR-30 [145, 146], miR-31 [147– 149], miR-126 [150], miR-145 [151], miR-146 [152], miR-193b [153], miR-205 [154], miR-206 [155], miR-335 [156], miR-448 [157], miR-661 [158] and let-7 [159].

19

Thus, in breast cancer, which represents the most common malignancy among women in the world, miRNAs such as miR-9, miR-10b, miR-21, miR-103/107, miR-132, miR-373, and miR-520 stimulate metastasis, while miR-7, miR-30, miR-31, miR-126, miR-145, miR-146, miR-200, miR-205, miR-335, miR-661, and miRNAs of the let-7 families in contrast impair the different steps of metastatic process, from epithelial-to-mesenchymal transition to local invasion to colonisation and angiogenesis [61].