The roles of miRNAs in human breast cancer and canine mammary tumor

MicroRNAs have become a hot topic in cancer research nowadays due to their important role not only on cancer development, progression, invasion but also on repression of cancer related genes. With advanced technologies, these microRNAs can easily be detected from biopsy samples and blood for early diagnosis, prognosis and treatment. Due to increasing demand of research in exploring expression profile of microRNAs with respect to different subtypes of breast cancer, this review aimed to provide an update on microRNA database available resources, canine breast cancer models, the role of microRNA as oncomir or oncosupressor, detection of microRNAs and potential of miRNAs for breast cancer treatment.


Background
MicroRNA (miRNA) was discovered in 1993 in Caenorhabditis elegans [1]. These non-coding short RNAs known as miRNAs from eukaryotes such as in plants and animals are highly important for regulation of gene expression at post-transcriptional level. Further, miRNAs are also responsible for cellular growth, differentiation, proliferation and apoptosis [2]. Predominant expression of a particular miRNA varies according to the type of animal tissue. Changes of normal miRNA expression levels indicate abnormal or disease conditions of tissues or body systems. Researchers were convinced to focus more about miRNAs after they found the first miRNA lin-4 of C. elegans that inhibited the expression of lin-14 gene [1]. Since then researchers introduced more and more worthy new findings related to several miRNAs and their roles in different diseases and cancers including breast cancers that potentially led clinicians to be able use circulating miRNAs as affordable, non-invasive detectable biomarkers for earlier diagnosis, prognosis and treatments of breast cancers.

Processing of microRNAs in the body
MicroRNAs (miRNAs, miRs) are 20 to 23 nucleotide RNAs and modulate gene expressions [3]. MicroRNAs processing in our body can be observed two parts as they started working in the nucleus of cells and continue its further processing in the cytoplasm [4]. The process in the nucleus initiates with the transcription of noncoding miRNAs genes with their own promoters. These primary transcriptions are carried out mainly by RNA polymerase II and the products are known as pri-miRNAs [5]. The stem-loop structure of pri-miRNAs are further processed by enzyme Drosha and dsRNAbinding protein DGCR8 [6][7][8]. The cleavage of pri-miRNAs produces 70 nucleotide long pre-miRNAs [9]. These pre-miRNAs are then exported into the cytoplasm by exportin 5 for further processing [10]. In the cytoplasm, pre-miRNAs are cleaved by RNase III enzyme, Dicer 1 that works together with transactivation responsive RNA binding protein 2 (TRBP) and AGO2 to generate double stranded miRNA-miRNA* duplex [11][12][13]. From the two miRNA strands, the guide strands or mature miRNA strands are incorporated into RISC (RNAinduced silencing complex) in order to bind its complementary target mRNA sequences of the target genes to block translation process. The passenger strands miRNA* are degraded [14,15]. However, it was also reported that some miRNAs; e.g. miRNA 145 are processed without Dicer, alternative precursors of miRNAs called mirtrons are processed without Drosha. Some miRNAs are processed along the pathway of tRNA [16][17][18][19]. Up to now, scientists have found miRNAs 2619 in human [20].
The process of miRNA begins in the nucleus of cells and continue in the cytoplasm of cells. Primary miRNA (pri-miRNA) transcription is done mainly by RNA polymerase II. Drosha and DGCR8 trim 100 nt to 120 nt long pri-miRNA to become a shorter 70 nt pre-miRNA and it is exported into cytoplasm by exportin 5 for further processes. Pre-miRNA in the cytoplasm is cleaved by RNase III, Dicer I and Argonaute 2 (AGO2) to become a 20 nt to 30 nt long duplex miRNA consists of a guide strand and a passenger strand. From the duplex miRNA, the strand, 3′ to 5′ miRNA known as passenger miRNA is degraded due to PIWI domain in RISC. The guide strand 5′ to 3′ miRNA which is engaged with PAZ domain in RISC binds to its target mRNAs and causes translational repression and cleavage of target mRNAs (Fig. 1).

Mechanisms of miRNA action
Single stranded mature microRNAs bind 3′ UTR (untranslated region) of target mRNAs as a post transcriptional regulation of gene expressions that inhibits translation of target mRNAs for proteins or cleavage of the target mRNAs by Argonaut ribonuclease of RISC (RNA-induced silencing complex) [21]. MicroRNA does not need to bind its whole length of complete nucleotide sequences to target mRNA but miRNA can perform gene expression regulation as long as minimum 2 to 8 base pairs are perfectly complementary between 3′ region of target mRNA and 5′ region of miRNA (miRNA seed sequence) in human and animals [22]. In human and animals, seed region of miRNA is usually 100% complementary to target mRNA but the rest sequences of miRNA may include mismatches to target mRNA that causes distended appearance due to limited base paring [23]. Two silencing mechanisms; slicer dependent that shows targeted mRNA cleavage irreversibly by Ago2 and slicer independent way that causes reversible translation repressions [24].
One miRNA can regulate concurrently the expression of many genes and thus it is involved in multiple cellular signaling pathway [3]. Furthermore, in recent studies, miRNAs was found to be involved in increasing the translation of a target mRNA directly and they were also be able to indirectly increase target mRNA levels by interacting repressor proteins which deter the translation of target mRNA [25]. MicroRNAs (miRNAs), either be as oncogenes or tumor suppressor genes play crucial role in cell differentiation, development, apoptosis and cell cycle. MicroRNAs are known to be expressed differentially in cancers but they show unique expression signature for a specific tumor type. It was reported that miRNAs influence on metastasis and resistance to different types of therapies [26]. Furthermore, researchers have found pro and anti-metastatic miRNA [27].

MicroRNAs for diagnosis, prognosis and therapeutics
Detecting circulating miRNA has also become a promising diagnostic tool. MicroRNA profiles are useful not only for diagnosis and prognosis but also for treatments in terms of choosing effective chemotherapeutic drug in cancers. This is why nowadays detecting miRNA profiles from breast cancer become more and more common in research field and also this happens to clinicians. In a study with zebra fish, researcher found that certain miR-NAs dominantly expressed in specific tissues and they became noticeably reduced expression levels when the tissues are in tumors state [28]. Commonly used breast cancers diagnostic methods such as mammography, ultrasound, X-rays and MRI still exist limitations of its use in patients and also required skills to read the images for proper interpretations. Although detecting circulating miRNA is not a substitute for other conventional diagnostic methods, it simply becomes essential for most cancer diseases due to its less invasive than tissue biopsy which has to be done for genetic tests such as specific mRNA expression. It has been reported several miRNAs have been proved to be deregulated. Up-regulated miRNAs that may be used for diagnosis, prognosis and therapeutic are miR21, miR155, miR221/ 222, miR9, miR10b, miR29a, miR96, miR146a, miR181, miR373, miR375, miR520c and miR589 [29,30]. Some of these up-regulated miRNAs in tumors play major role in controlling and working together with multiple targets genes and led them to be invasive types. Downregulated miRNAs are those with tumor suppressor properties. Recent study reported that let7c is possible biomarker which can discriminate breast cancer samples grade 1 to grade 3 [3]. Other consistently downregulated miRNAs are miR30a, miR31, miR34, miR92a, miR93, miR125, miR126, miR146a, miR195, miR200, miR205, miR206, miR503 and let7 family.
In breast cancer stem cells, miR15/16, miR103/107, miR128b, miR145, miR200 and miR335 are found to be down-regulated. With great efforts, now total 130 over miRNAs signatures were identified to distinguish between normal and tumor breast tissue accurately [31]. MicroRNAs in breast cancers can be classified into Oncomirs, Oncosupressors and MetastamiRs [32]. Oncomirs disturb expression of oncosupressors and cells apoptosis. In human breast cancer, oncomirs miRNA21 and miRNA155 are outstandingly upregulated and oncosupressors miRNA10b, miRNA125b and miRNA145 are noticeably downregulated. In triple negative breast cancers, miRNA21, miRNA 210 and miRNA 221 were found to be upregulated. Anti-miRNAs to those upregulated oncomirs not only inhibited cancer cells growth but also increased apoptosis and decreased cell proliferation. Oncosupressors such as miR204, miR34a,b,c and let-7 family of microRNAs are found to be downregulated in breast cancers. In one study, let-7 mimics showed effectiveness in the treatment of lungs cancer in mouse model when it was administered I.V. in the form of neutral lipid emulsion. MetastamiRs are recently recognized and introduced to those small non-coding miR-NAs that regulates migration and invasion of cancer cells [33,34]. These metastamiRs plays as major role in epithelial-mesenchymal transition (EMT), apoptosis and angiogenesis. It was reported that miRNAs 10b expression is 50% higher in invasive type of cancer cell line MDA-MB-231 compare to non-invasive type MCF-7 cancer cell line. Although miRNA10b was found to be down-regulated in non-invasive breast cancers, its change to up-regulation indicates the initiation of breast cancer metastasis. A study shown that applying antisense miR10b inhibitor oligonucleotides reduces invasiveness of transfected cells. It is required to restore the level of miRNAs to normal state and this can be achieved by dosing miRNA mimics or inhibitors, introducing miRNA genes with a DNA vector, application of small molecules to reverse epigenetic silencing of miRNAs [28]. Micro-RNAs mimics are short double stranded RNA guide strand and passenger stand oligonucleotides. They may be fully or partially complementary each other. The list of down-regulated miRNAs has been published and each of them are indeed potential cancer therapeutic miR-NAs. It was also known that one therapeutic miRNA is Fig. 1 MicroRNA Processing Steps Pathways. The process begins in the nucleus of cells and continue in the cytoplasm of cells. Primary miRNA (pri-miRNA) transcription is done mainly by RNA polymerase II. Drosha and DGCR8 trim 100 nt to 120 nt long pri-miRNA to become a shorter 70 nt pre-miRNA and it is exported into cytoplasm by exportin 5 for further processes. Pre-miRNA in the cytoplasm is cleaved by RNase III, Dicer I and Argonaute 2 (AGO2) to become a 20 nt to 30 nt long duplex miRNA consists of a guide strand and a passenger strand. From the duplex miRNA, the strand, 3′ to 5′ miRNA known as passenger miRNA is degraded due to PIWI domain in RISC. The guide strand 5′ to 3′ miRNA which is engaged with PAZ domain in RISC binds to its target mRNAs and causes translational repression and cleavage of target mRNAs able to target multiple mRNAs. In vivo study further provided an evidence that silencing of overexpression miR10b with systemic antagomirs suppressed breast cancer metastasis in mice [35]. Another study reported that miR373 was up-regulated in patients with breast cancer metastasized to lymph nodes. Like other circulating miRNAs, metastamiRs also become a popular molecular marker for breast cancers. Circulating miRNAs biomarkers offered affordable and non-invasive method of early detection, differentiating subtypes of breast cancers which allows clinicians to predict prognosis and further adjust their treatment regime as required. Table 1 miRNAs expression level, their targets, functional roles and potentials as biomarker in human breast cancer (female).

Methods to detect miRNAs in breast cancer
Micro-RNAs are quite a smaller size and impossible to detect them with the methods used for mRNA. However, since miRNAs roles become important and also found to be significant biomarkers of diseases including cancers, scientists invested enormous efforts in order to detect them. Nowadays, about 30 methods in total are available to detect miRNA expression profiles. Micro-RNA expression profiles in certain diseases, give critical useful information about diagnosis and prognosis. Among these miRNA detection methods, commonly applied to detect miRNA are Northern blotting, Microarray, Bead based flow cytometry, qRT-PCR, In sithu hybridization and Next-Generation Sequencing [36]. Among them, the most standardized method for miRNA analysis is Northern blotting although it is lowthroughput [37].
Being a high throughput, miRNA microarrays can be applied several miRNAs in a large number of samples. However, miRNA microarrays is low specificity and low sensitivity due to their extremely short sequence which allows only very little room to fine-tune the hybridization conditions of miRNAs [38]. Another high throughput, high specificity and high sensitivity method for assessment of miRNA is qRT-PCR and this method can also be used for validation of data obtained from other detection platforms [39]. Bead-based flow cytometry and in situ hybridization detection methods are limited to known miRNA structures. However, the latest detection method which gives high throughput, high specificity and high sensitivity is Next-generation sequencing and this method allows to discovery of new miRNA as well as to confirm known miRNA [40].
In breast cancers, miR10b, miR125b and miR145 are consistently found to be down-regulated and miR21, miR155 are consistently up-regulated [41,42]. It was also reported that miRNAs can still be detected from formalin fixed and paraffin embedded tissue samples, collected blood and serum too. MicroRNAs are stable due to their smaller size that showed tolerance to ribonuclease degradation.

Breast cancer model
All previous studies have identified that canine mammary tumor models are great to study the molecular pathogenesis of human breast cancer due to several similarities such as epidemiological factors and histopathological aspects [43][44][45][46]. In addition, it is very common that dogs and humans are exposed to equivalent carcinogens as frequently they are living in the same environment. Risk factors and molecular changes in affected mammary tissues for both species are comparable. Most of canine mammary tumors and human breast cancers are also originated from epithelial tissues [47]. In fact, different types of human cancers can be found in dogs too. For example, lymphoma, breast cancer, bone cancer and etc. Furthermore, studies have pointed out specifically that the age of onset and peak incidence among humans and dogs are about the same if their age is calculated proportionately [43]. Spontaneous mammary tumors incidence rate in female dogs is about 25% (1 in 4) while breast cancer incidence rate in women is about 12% (1 in 8). Readily availability of spontaneous canine mammary tumors is one of the factors that makes them a good model to study breast cancer as an alternative to experimental animal models. In particular, similarities including expression of hormone receptors, tumor growth markers, epidermal growth factor receptor, and P53 mutations also strengthen the canine mammary tumor model. Moreover, overexpression of cyclooxygenase-2 (COX-2), a molecule directly involved with mammary carcinogenesis, and matrix metalloproteinase-2 (MMP-2) which plays a role in invasion and metastasis process was detected in both species too [48,49].
It is noted that miRNA genes between human and canine are highly conserved. A study had been reported that miR15a, miR16, miR17-5p, miR21, miR29b, miR125b, miR155, miR181b and let7 family expression patterns were the same between human breast cancer and canine mammary tumors except miR145 which does not showed changes in canine [50,51]. As in human breast cancer, miR15a and miR16 expressions were also reduced in canine mammary carcinomas. Furthermore, miR15a and miR16 were shown to be involved in differentiation of malignancy [51]. Computationally analysis results showed that 300 miRNAs amongst canine miR-NAs were found to have the same sequences as of those found in human [52]. Also, it has been described that miRNA210 which is regulated by hypoxia, was significantly upregulated in continuous manner throughout canine mammary tumor progression [53]. In human breast   cancer, upregulated miRNA210 was reported as a diagnostic and prognostic marker [3,36].
A study reported that protein expression profiles were comparable between canine mammary tumors and human breast cancers. However, 12 other differentially expressed proteins in metastatic canine mammary tumors have not been studied yet and it remains to be established whether they are associated or not with metastasis of human breast cancers [54]. It is noted that researchers are able to study myoepithelial cell proliferation in canine breast cancer model rather than in human breast cancer due to their incidence rate > 20% and <0.1% respectively [55].
Besides the aforementioned similarities between canine and human breast cancers, other aspects that make canine tumors a good model for human breast cancer is faster aging (7 times more) compared to human due to their shorter life span and additionally the average shorter overall survival rate of dogs diagnosed with cancers [56].

Resources of miRNAs database
MicroRNA database including targets of miRNAs can be retrieved from the following web-based online programs.   [96] miRGator is based in Seoul, Korea. Current version 3 provides deep sequencing data of miRNAs that facilitate users to comprehend precursor of miRNAs, its sequences, their final products. Furthermore, miRGator has another two features such as miRNA catalogues and Expression profiles, and miRNA-mRNA target relations and expression correlations.
6. miRDB: http://mirdb.org/miRDB/ This data based was created by the department of radiation oncology, Washington university school of medicine in St. Louis. This web page provides predicted microRNA targets in human, mouse, rat, dog and chicken. This miRDB search engine allows users to search targets of miRNAs by miRNA name directly or gene accession number from GenBank, Gene Symbol and NCBI Gene ID.  Other useful online miRNA related sites were microrna.org, targetscan.org, ChIPBase, TarBase, starBase, ebi.ac.uk, PmmR (Putative microRNA-microRNA Regulations) and pictar.mdc-berlin.de.

Conclusion
MicroRNAs can be easily detected with advanced technologies and the continuous investigations supply miRNA as a novel clue for diagnosis and prognosis marker to therapeutic targets. Furthermore, research on delivering miRNAs mimics and/or inhibitors directly into breast cancer tissue to regulate the balance of miR-NAs is gaining attention in the field. The miRNA resources mentioned in this paper are useful to generate predicted targeted mRNAs for interested miRNAs candidate. Aberrant miRNA expression in our body such as oncomirs (upregulated miRNAs) and oncosupressors (downregulated miRNAs) could be the sign of diseases and they can be clinically detected by RT-qPCR, digital PCR, microarrays, and next generation sequencing technologies. Signature miRNA profile can be used to distinguish the stages of breast cancer progression and help in early diagnosis, prognosis, and effective treatment for human breast cancer and canine mammary tumor.