Abstract

   MicroRNAs (miRNAs) are short non-coding RNA molecules involved in
   post-transcriptional control. They are closely related to
   cancer-associated processes, making them promising candidates for
   diagnostic and prognostic biomarkers, as well as targets for cancer
   therapy. This study investigated the expression levels of let-7b,
   miR-29b, miR-125b, and miR-146 in normal and neoplastic canine mammary
   gland samples by quantitative reverse transcription polymerase chain
   reaction (RT-qPCR). Furthermore, the miRNA profile of the tumors was
   compared to adjacent mammary glands and to normal control mammary
   tissues. MiR-29b and miR-125b were downregulated in both benign and
   malignant lesions compared to the control, while let-7b was
   significantly downregulated only in malignant lesions. Additionally,
   miR-29b was upregulated in the malignant group compared to the benign
   group. Regarding histological type analysis, miR-125b was downregulated
   in simple carcinomas compared to complex carcinomas, whereas let-7b and
   miR-29b were upregulated in complex and simple carcinomas compared to
   the other subtypes, respectively. MiR-29b has been shown to be able to
   differentiate malignant from benign, while miR-125b can discriminate
   simple carcinomas from complex carcinomas. In addition, let-7b,
   miR-29b, and miR-125b, together, can discriminate simple and complex
   histological types from other histological types. In conclusion, the
   present study demonstrated that the expression of these miRNA changes
   during the tumorigeneses process, representing an epigenetic mechanism
   to sustain canine mammary tumor development, and identified these
   miRNAs as promising diagnostic biomarkers.

1. Introduction

   Epigenetics is the study of changes in organisms that result from
   modifications in gene expression rather than alterations in the genetic
   code itself. These changes affect chromatic dynamics and, consequently,
   gene regulation and expression [[36]1,[37]2,[38]3]. The main mechanisms
   of epigenetics can be divided into four types: DNA methylation, histone
   variants, histone post-translational modifications, and non-coding RNAs
   (ncRNAs) [[39]4]. This last type, in particular, encompasses a class of
   RNA sequences that are transcribed but do not encode proteins [[40]5].
   NcRNAs are often grouped according to their size: small ncRNAs (<200
   nucleotides) and long ncRNAs (>200 nucleotides) [[41]6]. MicroRNAs
   (miRNAs) are small ncRNA molecules with an average length of 22
   nucleotides and are involved in post-transcriptional gene regulation in
   diverse biological contexts [[42]7]. In cancer, they can act as
   oncogenes or tumor suppressor genes [[43]8]. The deregulation of miRNA
   expression in both tumor development and progression has made these
   molecules an important biomarker in cancer research, namely in breast
   cancer, where they were described as potential diagnostic and/or
   prognostic biomarkers [[44]9,[45]10,[46]11]. In veterinary medicine,
   miRNA research is still an emerging field. However, it has been
   receiving an increasing level of attention, as ongoing research
   suggests that microRNAs can serve as valuable tools for diagnostic and
   prognostic biomarkers in this field as well [[47]12]. Canine mammary
   tumors (CMTs) are the most prevalent neoplasms in intact female dogs,
   and approximately 50% of these tumors are malignant [[48]13]. CMTs
   constitute a heterogeneous group of tumors characterized by variations
   in morphology and biological behavior. CMTs exhibit several
   similarities to human breast cancer (HBC), establishing them as a
   valuable translational model for breast cancer research [[49]13]. One
   of these characteristics is their genetic similarity, namely miRNA
   deregulation, which can be very useful when studying potential
   biomarkers [[50]14]. In CMTs, research into miRNA expression has been
   relatively limited; nevertheless, existing studies suggest that miRNA
   expression patterns in CMTs resemble those of HBC
   [[51]15,[52]16,[53]17]. In the present study, we analyzed a panel of
   four miRNAs (let-7b, miR-29b, miR-125b, and miR-146) in canine mammary
   tumors (CMTs) and in normal/adjacent mammary gland tissues that were
   previously associated with diagnostic and/or prognostic importance in
   HBC [[54]18,[55]19,[56]20,[57]21]. Our goal was to evaluate the changes
   in expression levels of these miRNAs in CMT samples and assess their
   potential to discriminate between the defined groups and their
   consequent applicability in the veterinary field.

2. Materials and Methods

2.1. Ethical Statement

   This study was conducted in accordance with the European Directive
   (2010/63/EU). All procedures were performed after ethical approval by
   DGAV (Direção Geral de Agricultura e Veterinária, Lisbon, Portugal)
   with reference number 004582.

2.2. Animal Population, Sample Collection and Histopathology

   Forty-nine mammary tumor samples from 31 female dogs were surgically
   excised in veterinary hospitals or private clinics in Portugal. A
   portion of each tissue sample was frozen for molecular analyses, while
   the remaining samples were fixed in 10% neutral buffered formalin for
   histopathological diagnosis. Sixteen adjacent mammary tissues were
   collected from animals harboring tumor samples. Two normal mammary
   gland tissue samples were obtained from two dogs undergoing postmortem
   examination. Age and breed were registered.

   Histological sections (two µm-thick) were stained with hematoxylin and
   eosin (H&E) and analyzed under a light microscope with a digital camera
   by a trained veterinary pathologist (A.G.) (Nikon Eclipse E6000^®,
   Nikon Instruments Inc., Melville, NY, USA). Tumor samples were
   classified histologically using the classification system proposed by
   the Davis-Thompson DVM Foundation [[58]22]. The samples were divided
   into three groups: 18 control samples (16 tumor-adjacent and 4 normal
   mammary tissues), 6 benign samples (4 benign mammary tumors and 4
   mammary hyperplasias), and 27 malignant samples (all malignant mammary
   carcinomas).

2.3. Total RNA Extraction

   Tissue samples were processed for total RNA extraction using the
   semi-automated high-throughput magnetic-bead extraction kit (MagMAX™
   mirVana™ RNA Isolation Kit; product no: A27828, Thermo Fisher
   Scientific^®, Waltham, MA, USA) on the KingFisher Duo Prime magnetic
   particle processor (Thermo Fisher Scientific^®, Waltham, MA, USA)
   following the manufacturer’s instructions.

   RNA concentration and purity were analyzed using the NanoDrop Lite
   spectrophotometer (Thermo Fisher Scientific^®, Waltham, MA, USA) and
   further stored at −80 °C until use.

2.4. cDNA Synthesis

   RNA samples served as a template for cDNA synthesis using a TaqMan™
   MicroRNA Reverse Transcription kit (Applied Biosystems^®, Foster City,
   CA, USA) according to the manufacturer’s guidelines. The reverse
   transcription reactions were performed in MyCycler™ Thermal cycler
   (Bio-Rad Laboratories, Hercules, CA, USA) for 30 min at 16 °C, 60 min
   at 42 °C, 10 min at 85 °C, and then held at 4 °C, after which the cDNA
   was stored at −20 °C.

2.5. Relative Quantification of microRNAs

   The miRNA expression levels were measured by quantitative real-time
   polymerase chain reaction (RT-qPCR) using the StepOnePlus^TM qPCR
   system (Applied Biosystems^®, Foster City, CA, USA). Data analysis was
   performed using the StepOne software (version 2.3, Applied
   Biosystems^®, Foster City, CA, USA).

   The RT-qPCR reaction was performed for each sample in a final reaction
   volume of 10 µL, using 3.5 µL of nuclease-free water, 5 µL of 2×
   TaqMan^® Fast Advanced Master Mix (Applied Biosystems^®, Foster City,
   CA, USA), 0.5 µL of 20× TaqMan^TM Gene Expression Assays (miR-29b:
   000413, miR-125b: 000449, miR-146a: 000468, let-7b: 002619, mir-16:
   000391, Applied Biosystems^®, Foster City, CA, USA), and 0.5 µL of cDNA
   sample. miR-16 was selected as the endogenous control for
   normalization, as described previously [[59]23,[60]24,[61]25,[62]26].
   The thermal cycling conditions were as follows: 10 min at 95 °C
   followed by 40 cycles of 15 s at 95 °C and 1 min at 60 °C. The
   quantification of miRNAs was performed in duplicate, and CT standard
   deviation values superior to 0.5 were excluded from the analysis.
   Negative controls (without cDNA) were also included in all reactions.

2.6. MicroRNAs Predicted Target and Enrichment Analysis

   The potential biological implications of the studied miRNAs were
   explored through in silico analyses. Target genes prediction for
   let-7b, miR-29b, and miR-125b were performed using the online database
   miRWalk 3.0 [[63]http://mirwalk.umm.uni-heidelberg.de (accessed on 24
   May 2024)], which includes 3 miRNA-target prediction databases (miRDB,
   TargetScan and miRTarBase) [[64]27]. The analysis targeted miRNA
   binding sites in the 5′ untranslated region (5′UTR), in the coding
   sequence (CDS), and the 3′UTR, selecting all potential targets with a
   binding probability of 1. To avoid redundancy, gene targets linked to
   more than one miRNA were removed from the final set.

   These miRNA targets were then visualized in a protein–protein
   interaction (PPI) network using STRINGapp Protein Query from Cytoscape
   software (v3.10.2; [65]https://cytoscape.org, accessed on 7 June 2024)
   with a confident score cut-off of 0.4. A Markov Cluster Algorithm (MCL)
   was further applied to cluster the proteins based on their STRING
   interaction score, with an inflation value of 3.0. Lastly, functional
   enrichment of formed clusters was performed using the canine genome as
   background. This analysis was performed using the enrichment analysis
   tool of the STRINGapp Protein Query, considering a False Discovery Rate
   (FDR) of p < 0.05. The enrichment results were filtered to remove
   redundant terms using a cut-off of 0.5. Gene Ontology (GO), Kyoto
   Encyclopedia of Genes and Genomes (KEGG), and Reactome terms were
   selected.

2.7. Statistical Analysis

   Data analysis was performed using GraphPad Prism^® 9.5.0 software for
   Windows (GraphPad Software Inc., La Jolla, CA, USA). The Shapiro–Wilk
   or Kolmogorov–Smirnov test was applied to determine the data for normal
   distribution. Depending on the distribution, the Student’s t-test (for
   normal distribution) or Mann–Whitney U test (for non-normal
   distribution) was used to assess statistical differences between the
   expression levels of the normalized miRNAs. All values were considered
   statistically significant at p < 0.05. Pearson’s correlation
   coefficients were calculated for the gene expression data (-∆Ct values)
   of several of the microRNAs to identify any potential relationships
   among their expression levels.

3. Results

3.1. Clinicopathological Features

   In this study, fifty-three mammary samples from 34 female dogs were
   included with a mean age of 8.25 ± 2.86 years (mean ± standard
   deviation), ranging from 4 to 15 years. The clinicopathological
   findings are summarized in [66]Table 1. Most of the animals were mixed
   breeds (n = 12; 38.8%), followed by German Shepherds (n = 5; 16.1%),
   and the majority were less than 10 years old (n = 19; 67.9%). Malignant
   tumors were removed from female dogs aged 7.67 ± 1.37 years, while the
   benign samples were harvested from animals aged 9.05 ± 3.11 years. No
   significant statistical differences were observed in age between female
   dogs with malignant and benign mammary tumors.

Table 1.

   Clinicopathological findings of female dogs included in this study.
        Features      Frequency (%)
     Age (n = 28)
       >10 years        9 (32.1%)
       ≤10 years       19 (67.9%)
    Breed (n = 31)
         Mixed         12 (38.8%)
    German shepherd     5 (16.1%)
         Boxer          4 (12.9%)
       Pekingese        2 (6.5%)
   Portuguese Pointer   2 (6.5%)
     Cocker Spaniel     1 (3.2%)
       Dachshund        1 (3.2%)
     French bulldog     1 (3.2%)
   Labrador Retriever   1 (3.2%)
        Pinscher        1 (3.2%)
         Poodle         1 (3.2%)
   [67]Open in a new tab

   The histological diagnosis of the canine mammary samples is detailed in
   [68]Table 2. Benign tumors were diagnosed as complex adenomas (n = 2;
   3.8%) and benign mixed tumors (n = 2; 3.8%) ([69]Supplementary Figure
   S1). Malignant tumors were mostly classified as complex carcinomas (n =
   12; 22.6%) and tubulopapillary carcinomas (n = 7; 13.2%).

Table 2.

   Distribution of canine mammary samples according to histological
   classification.
 Histopathological Group    Histological Classification            Frequency (%)
 Controls (n = 18)          Normal mammary gland                       2 (11.1%)
                            Adjacent mammary gland                    16 (88.9%)
 Benign lesions (n = 6)     Lobular hyperplasia                        2 (33.3%)
                            Complex adenoma                            2 (33.3%)
                            Benign mixed tumor                         2 (33.3%)
 Malignant lesions (n = 27) Tubulopapillary carcinoma                  7 (25.9%)
                            Solid carcinoma                             2 (7.4%)
                            Comedocarcinoma                             1 (3.7%)
                            Complex carcinoma                         12 (44.5%)
                            Carcinoma-and-malignant myoepithelioma      2 (7.4%)
                            Carcinosarcoma                             3 (11.1%)
   [70]Open in a new tab

3.2. miRNAs Expression Analysis

   [71]Figure 1 displays the expression of miRNAs across the three groups:
   control, benign, and malignant lesions. Let-7b (p = 0.0004) was seen as
   downregulated in the malignant group when compared to the control
   group. miR-29b (p = 0.0001; p = 0.0003) and miR-125b (p = 0.0070; p <
   0.0001) exhibited downregulation in both benign and malignant groups
   when compared to the control group. In addition, only miR-29b (p <
   0.0001) showed significant differences between the benign and malignant
   groups, being upregulated in the malignant group. No significant
   differences were observed among the groups for miR-146a. Based on the
   two major histological types (simple and complex carcinoma), the
   malignant group was divided into simple carcinomas (including
   tubulopapillary, solid, and comedocarcinoma), complex carcinomas, and
   other carcinoma subtypes (carcinoma-and-malignant myoepithelioma and
   carcinosarcoma) for further comparison. Let-7b was significantly
   downregulated in simple carcinomas (p = 0.0295), complex carcinomas (p
   = 0.0189), and other carcinomas (p = 0.0085) compared to the control
   group ([72]Figure 2). In addition, let-7b was significantly upregulated
   in complex carcinomas (p = 0.0313) in relation to other subtypes.
   MiR-29b exhibited a significant downregulation in all groups in
   comparison with the control group. Furthermore, miR-29b was upregulated
   in the simple carcinoma group (p = 0.0002) and complex carcinoma group
   (p = 0.0001) compared to benign cases. Moreover, miR-29b was
   significantly upregulated in simple carcinomas in relation to other
   subtypes (p = 0.0050). Regarding miR-125b, all groups were
   downregulated compared to the control group. In addition, in simple
   carcinomas, miR-125b was downregulated in comparison with the complex
   carcinoma group (p = 0.0437). No significant differences were observed
   among the groups for miR-146a.

Figure 1.

   [73]Figure 1
   [74]Open in a new tab

   Relative expression levels of the miRNAs let-7b (A), miR-29b (B),
   miR-125b (C), and miR-146a (D) among benign lesions, malignant lesions,
   and the control group. Data are presented as −∆Ct mean values ±
   standard error of the mean (SEM). Statistical analysis was performed
   using Student’s t-test for normally distributed variables and
   Mann–Whitney U test for non-normally distributed variables.
   Significance was declared at p < 0.01 (**), p < 0.001 (***), and p <
   0.0001 (****).

Figure 2.

   [75]Figure 2
   [76]Open in a new tab

   Relative expression levels of the miRNAs let-7b (A), miR-29b (B),
   miR-125b (C), and miR-146a (D) among benign lesions, simple, complex,
   and other subtypes carcinomas, and the control group. Data are
   presented as −∆Ct mean values ± standard error of the mean (SEM).
   Statistical analysis was performed using Student’s t-test for normally
   distributed variables and Mann–Whitney U test for non-normally
   distributed variables. Significance was declared at p < 0.05 (*), p <
   0.01 (**), p < 0.001 (***).

   Different degrees of positive correlation were found between all
   miRNAs, with significant correlation values observed between miR-29b
   and let-7b and (r = 0.5776, p < 0.001), miR-125b and let-7b (r =
   0.8095, p < 0.001), and miR-125b and miR-29b (r = 0.5581, p < 0.001)
   ([77]Supplementary Figure S2).

3.3. miRNA Target Prediction and Enrichment Analysis

   According to our results, let-7b, miR-29b, and miR-125b were the miRNAs
   found to be downregulated in the neoplastic lesions, suggesting that
   they may have interfered in tumor formation/progression. Thus, our in
   silico analysis focused on the mRNA targets of these 3 miRNAs. The
   miRNA–mRNA interaction resulted in a total of 7729 target genes (5441
   for let-7b, 500 for miR-29b, and 2566 for miR-125b), of which 1227 were
   considered for STRING analysis due to their interaction with two or
   three miRNAs ([78]Figure 3, [79]Table S1). Fifty-one genes were found
   in common in the three miRNAs.

Figure 3.

   [80]Figure 3
   [81]Open in a new tab

   Venn diagram showing the number of unique and common target genes of
   let-7b, miR-29b, and miR-125b obtained from the miRWalk database, with
   a binding probability of 1. Graph obtained using Venny 2.1.

   To analyze the biological impact of the three miRNAs, we analyzed the
   1227 target genes using the STRINGapp Protein Query from Cytoscape
   (v3.10.2.) and by applying MCL. PPI networks were generated with a
   total of 1170 nodes and 629 edges, all with a significant enrichment (p
   = 1.0 × 10^−16) ([82]Figure 4).

Figure 4.

   [83]Figure 4
   [84]Open in a new tab

   Protein–protein interaction (PPI) network of let-7b, miR-29b, and
   miR-125b. PPI was generated using the STRINGapp Protein Query from
   Cytoscape (v3.10.2), and clustering was performed by applying the
   Markov Cluster Algorithm (MCL), using an inflation value of 3.0.
   Proteins with fewer than five interaction partners in the network have
   been omitted from the visualization.

   For the functional enrichment analysis, we focused on clusters
   comprising five or more interactions, obtaining a total of 149 nodes
   with 243 edges across 21 clusters ([85]Table S2). The analysis was
   conducted with an FDR threshold of 5%, and redundant terms were removed
   using a redundancy cut-off of 0.5, which resulted in a total of 28
   functional enriched terms among the Reactome, KEGG, and GO categories
   ([86]Figure 5, [87]Tables S3–S6). Reactome pathways revealed that the
   targets of the deregulated miRNAs are involved in signaling by GPCR,
   whereas KEEG pathways demonstrated that miRNAs are involved in cAMP
   signaling and fatty acid degradation.

Figure 5.

   [88]Figure 5
   [89]Open in a new tab

   Gene ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), and
   Reactome pathways of the functional enrichment analysis. The analysis
   was performed with STRINGapp Protein Query from Cytoscape (v3.10.2).
   Redundant terms were eliminated with a redundancy cut-off of 0.5, and
   an FDR threshold of 5% was applied.

4. Discussion

   This study aimed to understand the differences in the expression of
   four miRNAs (let-7b, miR-29b, miR-125b, and miR-146a) in dogs diagnosed
   with mammary tumors and to determine the diagnostic capabilities of
   single and combined miRNAs. The mean age of the female dogs included
   was similar to previous studies, which reported ages between 7 and 13
   years [[90]28,[91]29,[92]30,[93]31]. Although not significant, more
   malignant tumors were found at younger ages than benign tumors (7.67 ±
   1.37 vs. 9.05 ± 3.11 years), which is contradictory with other studies
   indicating that malignant neoplasm occurs in older dogs [[94]32]. Our
   small cohort may be a reason for this result since it may not be large
   enough for epidemiological studies. The predisposition of mixed-breed
   dogs to develop CMTs was aligned with other studies
   [[95]30,[96]33,[97]34]. However, this issue is controversial in
   literature and has been attributed to variations in the canine
   population included in the research studies [[98]35]. Regarding
   histological classification, tubulopapillary and complex carcinomas
   were the most common subtypes, as previously described by several
   authors [[99]29,[100]30,[101]36].

   We selected benign and malignant lesions to assess whether there was a
   differential expression of miRNAs at different stages of canine mammary
   tumor progression. To analyze differential expression in specific
   cancer phenotypes, we selected homogeneous groups of tumors that were
   classified as simple (tubulopapillary, solid, and comedocarcinoma) and
   complex carcinomas. The remaining samples were grouped into a single
   group that included carcinoma-and-malignant myoepithelioma and
   carcinosarcoma.

   A significant downregulation of let-7b was observed in malignant
   lesions compared to control cases. These results are in concordance
   with HBC, where it was inversely associated with lymph node metastasis,
   overall survival, and disease-free survival [[102]18]. A recent study
   reported that the use of let-7b mimics in the MCF-7 cell line increased
   the expression of the pro-apoptotic BAX gene and decreased the
   expression level of anti-apoptotic BCL2 gene after let-7b
   overexpression, suggesting that this miRNA was associated with
   apoptosis-regulating genes in HBC [[103]37]. Regarding the three
   histological subtype groups, upregulation was also found in complex
   carcinomas in relation to other subtypes. In our study, the
   downregulation of let-7b was not observed in benign lesions, which
   suggests that let-7b could be involved in the oncogenic process and
   differentiation of lesions. Contrary to our findings, let-7b and
   let-7b-5p were recently found upregulated in canine mammary cell lines
   and mammary tissues, respectively [[104]38,[105]39]. One possible
   explanation for this difference is that these studies used different
   endogenous microRNAs. Additionally, the cell lines were derived from a
   mixed tumor and a comedocarcinoma, which are not representative of our
   study. Furthermore, the study on mammary tissue samples did not
   stratify the data by malignancy status (benign vs. malignant lesions)
   or by histological type.

   MiR-29b was shown to be downregulated in both benign and malignant
   lesions compared to control; however, malignant lesions were
   upregulated in relation to benign lesions. Our results are in
   concordance with studies in which miR-29b was downregulated in benign
   and malignant canine lesions compared to control [[106]40,[107]41]. In
   these studies, miR-29b showed no significant differences between benign
   and malignant lesions [[108]40,[109]41]. In our study, this difference
   was observed, demonstrating its ability to differentiate benign from
   malignant lesions. Curiously, after the analysis of miR-29b expression
   between the histological subtypes, our data showed an upregulation of
   two groups (simple and complex carcinomas) in relation to benign
   lesions. These results indeed demonstrate the complex role of this
   miRNA, which might have both tumor suppressor and oncogenic roles.
   Additionally, the expression of miR-29b was significantly increased in
   simple carcinomas compared to other subtypes. A study observed a
   downregulation of miR-29b in metastatic canine samples compared to the
   primary tumor [[110]16]. In contrast, another study described the
   upregulation of miR-29b in canine mammary cancer samples compared to
   the control, but the authors used RNU6 as the reference gene for
   normalization, which could explain these differences [[111]15].
   Interestingly, in CMT cell lines, miR-29b was also found to be
   upregulated [[112]42,[113]43], as well as in exosomal RNA [[114]26].

   According to our results, miR-125b was shown to be downregulated in
   both benign and malignant lesions, as well as in the three groups
   (simple, complex, and other subtypes), compared to the control group.
   These findings suggest that miR-125b can be a diagnostic biomarker,
   confirming the presence of a neoplastic lesion. In addition, miR-125b
   seems to have a malignant differentiation capacity, since it can
   distinguish simple carcinomas from complex carcinomas. In another
   study, miR-125b was found to be downregulated in canine mammary
   neoplastic lesions [[115]16]. Moreover, Von Deetzen and co-workers
   evaluated the same microRNA and found no significant difference between
   canine mammary tissues (n = 30) at different stages of tumor
   progression (adenoma, non-metastasizing carcinoma, metastasizing
   carcinoma, and metastasis) and normal gland that used miR-181b,
   miR-155, and U44 for normalization [[116]41]. Accordingly, in HBC,
   several studies reported that miR-125b was found downregulated in
   tissue samples [[117]19,[118]44]. Contradictory, a recent study
   observed a higher expression level of this miRNA in the serum of 56
   patients with non-responsive breast cancer, indicating that it was
   associated with chemoresistance [[119]45]. Another study using
   paraffin-embedded breast cancer samples of 221 patients described a
   positive association between high expression of miR-125b and tumor size
   and stage in HER2-positive breast cancer patients [[120]46].

   Since let-7b can distinguish between complex and other subtypes,
   miR-29b between simple carcinomas and other subtypes, and miR-125b
   between simple and complex carcinomas, using these three miRNAs
   together could allow a precise diagnosis in distinguishing among these
   three histological types. In sum, the differential expression of
   let-7b, miR-29b, and miR-125b together may suggest the ability to
   discriminate these two histological groups (simple and complex
   carcinoma) from the other subtypes. However, further studies with other
   specific histological types will be necessary to consolidate these
   results.

   Regarding the bioinformatic analysis, our key findings from Reactome
   pathways suggest that these miRNAs are involved in cancer-related key
   pathways, namely G Protein Coupled Receptor (GPCR) that are
   transmembrane signal transducers that regulate physiological and
   pathological processes represent the largest family of therapeutics
   targets [[121]47]. As a member of the GPCRs family, protein-coupled
   estrogen receptor (GPER), is a seven-transmembrane receptor that
   mediates estrogen signals in both normal and malignant cells and has
   been involved in diverse signaling pathways that characterize the
   progression of breast cancer [[122]48]. GPER expression was associated
   with tumor size, metastasis, and tamoxifen resistance [[123]49]. It has
   also been observed that it is more prevalent in triple-negative breast
   cancer and is associated with younger age and more aggressive disease
   [[124]50]. The other GPCR, GPR110, influences the growth of
   HER2-positive breast cancer cells and potentiates drug resistance in
   these cells [[125]51]. In the KEGG pathway, other cancer-related
   pathways were observed, including the cAMP pathway and fatty acid
   degradation, which also plays an important role in the development of
   HBC [[126]52,[127]53]. Previous studies have demonstrated that
   activation of the cAMP pathway can inhibit cell proliferation and
   migration in breast cancer [[128]54,[129]55]. Deregulation of fatty
   acid metabolism is recognized as a critical factor for malignant
   transformation in HBC and has been implicated in the promotion of cell
   proliferation, migration, and invasion, as well as associated with poor
   prognosis in breast cancer patients [[130]53,[131]56,[132]57,[133]58].

   This study could be a breakthrough in the research of complex carcinoma
   (malignant adenomyoepithelioma) in HBC. Complex carcinoma is
   characterized by the proliferation of both luminal and myoepithelial
   cells, a rare condition in HBC. However, this type of tumor shows
   molecular similarities to those found in humans, making canine complex
   carcinoma a unique model for investigating the roles of myoepithelial
   cells, the second major cell lineage in the mammary gland involved in
   normal developmental and pathogenic processes [[134]59].

   Our study’s limitations include the use of tumor-adjacent mammary
   tissues as controls and only two normal mammary glands. In HBC, healthy
   control tissue is generally readily available through breast reduction
   surgeries [[135]60]. However, this is not the case for canine patients,
   as mammary surgeries are not performed in veterinary clinics unless
   there is an existing pathology. Another limitation is the in silico
   analysis, which, although helpful in identifying biological mechanisms
   associated with these miRNAs, requires validation to provide an
   integrated view of adjacent mechanisms, as other genes and pathways
   with significant roles may have been overlooked. However, this issue
   was mitigated by the fact that the analysis was conducted only on genes
   whose mRNAs were regulated by at least two of the miRNAs studied.

   Further studies are required to understand the mechanisms of these
   miRNAs and to consider them as potential biomarkers in dogs. Studies
   with larger sample sizes, using exclusively normal mammary gland tissue
   as controls, and investigating miRNAs not only in different
   histological types but also with diverse immunophenotype subtypes,
   could provide valuable insights. In HBC, miRNAs have been shown to
   exhibit distinct regulation across immunophenotype subtypes [[136]61].
   It is also necessary to associate these miRNAs with clinicopathological
   features, such as tumor histological grade, tumor size, and lymphatic
   invasion, for these miRNAs to be used as prognostic biomarkers. The use
   of animal samples with follow-up will also facilitate a better
   understanding of the prognostic value of these miRNAs. In addition,
   functional assays will allow us to understand the involvement of these
   miRNAs in tumorigenesis, including their roles in proliferation,
   apoptosis, migration, and invasion in cell culture models, as well as
   in the evaluation of tumorigenicity and metastasis in xenograft animal
   models. Additionally, the use of antimiRNAs or miRNAs mimics in both in
   vivo and in vitro models would further help to consolidate the role of
   these miRNAs in tumorigenesis. Finally, incorporating these miRNAs into
   non-invasive techniques, such as serum or exosomal miRNA profiling,
   could improve the early detection and prognosis of CMTs, thereby
   increasing their translational relevance and potential clinical
   applications.

5. Conclusions

   Histological classification guides diagnostic and prognostic
   evaluation, suggesting that the differential expression of miRNAs
   across different histological types is a promising line of research.
   The expression of miR-29b differed significantly between benign and
   malignant groups, indicating its potential to differentiate malignancy,
   while let-7b, miR-29b, and miR-125b appear capable of discriminating
   between histological phenotypes. Further studies with these miRNAs and
   their combined use could improve diagnostic accuracy.

   This work establishes a miRNA signature for CMT to guide future miRNA
   selection for early validation and reinforces the evidence supporting
   microRNAs as important biomarker candidates for this tumor type in
   dogs.

Supplementary Materials

   The following supporting information can be downloaded at:
   [137]https://www.mdpi.com/article/10.3390/ani15010020/s1, Figure S1.
   Histopathology microphotographs representative of the canine mammary
   gland tissues analyzed in the present study: controls (A. normal
   mammary gland); benign lesions (B. complex adenoma); malignant lesions
   (C.tubulopapillary carcinoma and D. complex carcinoma). Hematoxylin and
   eosin staining. Figure S2. Heatmap of Pearson correlation coefficient
   matrix. The correlation matrix heatmap displays the Pearson correlation
   coefficients between microRNA expression data. In this matrix, 1
   indicates a perfect positive correlation, 0 indicates no correlation
   and −1 indicates a perfect negative correlation between the studied
   microRNAs. Statistical significance: * p < 0.05, ** p < 0.01, *** p <
   0.001, **** p < 0.0001. Table S1. List of 1227 gene targets obtained
   from the mRNA/miRNA interaction, selected for STRINGapp Protein Query.
   Table S2. Protein-protein-interaction (PPI) network clusters according
   to Marckov Clusterring (MCL) analysis. Only clusters with five or more
   interactions were considered. Table S3. Gene Ontology (GO) Cellular
   Component enrichment analysis for terms with FDR p-value < 0.05. Table
   S4. Gene Ontology (GO) Molecular Function enrichment analysis for terms
   with FDR p-value < 0.05. Table S5. KEGG enrichment analysis for terms
   with FDR p-value < 0.05. Table S6. Reactome pathway enrichment analysis
   for terms with FDR p-value < 0.05.
   [138]animals-15-00020-s001.zip^ (463.2KB, zip)

Author Contributions

   Conceptualization, T.F., F.D., and R.M.; methodology, T.F., F.D., Â.A.,
   and A.G.; formal analysis, T.F., F.D., A.G., and R.M.; data curation,
   T.F. and F.D.; writing—original draft preparation, T.F.; writing—review
   and editing, Â.A., F.D., J.F.M., A.G., P.A.O., and R.M.; supervision,
   J.F.M., P.A.O., and R.M.; funding acquisition, P.A.O. and R.M. All
   authors have read and agreed to the published version of the
   manuscript.

Institutional Review Board Statement

   Ethical approval was granted by DGAV (Direção Geral de Agricultura e
   Veterinária, Lisbon, Portugal) with reference number 004582.

Informed Consent Statement

   Informed consent was obtained from the owners of the animals involved
   in the study.

Data Availability Statement

   The data are contained in the manuscript and [139]Supplementary
   Materials.

Conflicts of Interest

   The authors declare no conflicts of interest.

Funding Statement

   This work was supported by National Funds from FCT–Portuguese
   Foundation for Science and Technology, under the projects
   UIDB/04033/2020 ([140]https://doi.org/10.54499/UIDB/04033/2020),
   LA/P/0126/2020, UIDB/00772/2020
   ([141]https://doi.org/10.54499/UIDB/00772/2020), LA/P/0059/2020, a
   doctoral grant with 2020.04789.BD
   ([142]https://doi.org/10.54499/2020.04789.BD) and FD as a junior
   researcher contract (UIDP/00776/2020-4B). This work was also developed
   within the scope of the Project CICECO-Aveiro Institute of Materials,
   UIDB/50011/2020 (DOI 10.54499/UIDB/50011/2020), UIDP/50011/2020 (DOI
   10.54499/UIDP/50011/2020), LA/P/0006/2020 (DOI
   10.54499/LA/P/0006/2020), financed by national funds through the
   FCT/MEC (PIDDAC).

Footnotes

   Disclaimer/Publisher’s Note: The statements, opinions and data
   contained in all publications are solely those of the individual
   author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI
   and/or the editor(s) disclaim responsibility for any injury to people
   or property resulting from any ideas, methods, instructions or products
   referred to in the content.

References