Investigation of miR-133a, miR-637 and miR-944 genes expression and their relationship with PI3K/AKT signaling in women with breast cancer

MicroRNAs (miRNAs) are regulatory molecules capable of positively or negatively regulating signaling pathways, and are involved in tumorigenesis as well as various aspects of cancer. The purpose of this study was to investigate the expression levels of miR-133a, miR-637, and miR-944 in serum and tumor tissues as well as their relationship with the expression level of phosphatidylinositol-3-kinase (PI3K) and protein kinase-B (AKT) genes and proteins along with their clinical significance in breast cancer. The expressions of miR-133a, miR-637, miR-944, PI3K, and AKT genes were examined in the tumor and tumor margin tissues of 40 patients with breast cancer, as well as the serum levels of miR-133a, miR-637, and miR-944 in these patients and 40 healthy groups by quantitative real-time PCR (qRT-PCR). PI3K and AKT proteins expression in tumor and tumor margin tissues were detected using immunohistochemistry (IHC). The expression levels of miR-133a and miR-637 in the tumor tissue and serum of patients were lower than those in the tumor margin tissue and serum of the healthy group, respectively. In addition, the expression level of miR-944 in the tumor tissue was lower than that in the tumor margin tissue, but its expression increased in the serum of cancer patients compared to that in the healthy group. The expression of miR-637 was correlated with tumor location and Her2 receptors, and the expression of miR-944 was correlated with tumor location and family history. PI3K and AKT mRNA and protein levels were higher in the tumor tissues than in the tumor margin tissues (p < 0.05). The results of our study revealed that miR-637 has a better diagnostic value in breast cancer than miR-133a and miR-944.


Introduction
Various mutations are tracked down in the genome of cancerous tissue cells; as such, cancers are considered hereditary illnesses. On the other hand, in only a small percentage of families, cancer is considered a gene mutation that is passed from one generation to another, and most believe that cancers are affected by the environment (Huang et al. 2009). Breast cancer (BC) is a well-known disease affecting women causing a significant burden for females in both developed and non-industrial nations (Zou et al. 2021). It is the main source of mortality in non-industrial nations and the subsequent driving reason for cancer demise in American women. In 2020, the World Health Organization (WHO) recorded over 2.3 million breast cancer cases among women worldwide and 685,000 deaths (Sideris et al. 2022). Despite the progress and improvement of screening methods, diagnosis (mammography), and successful medicines, for example, medical procedures in the initial phases, numerous patients remain undiscovered until the condition causes distal metastases (Zou et al. 2021). Meanwhile, a large number of women who are diagnosed to have breast cancer in the initial phases experience the development of more aggressive phenotype months or a long time after the initial treatment (Marusyk et al. 2020). Two factors that contribute to cancer resistance are tumor cell heterogeneity and tumor cell metastasis to different organs (Sideris et al. 2022). Traditional symptomatic techniques (mammography) are associated with pain, anxiety, and radiation exposure. Meanwhile, its effectiveness is limited by the presence of dense breasts. Most importantly, mammograms are not suggested for individuals more than 40 years old, often showing a high rate of false positives or overdiagnosis. This might truly hurt the patients and significantly limit their use (Itani et al. 2021). In this respect, understanding molecular mechanisms and elaborate variables can assist in distinguishing novel tools for the prognosis and diagnosis of breast cancer patients (Sideris et al. 2022).
One of these elements is microRNAs (miRNAs), which are a large subgroup of non-coding RNAs of 18-25 nucleotides that are evolutionarily conserved (MacFarlane and R Murphy 2010). These molecules control gene expression after transcription by inhibiting mRNA translation or inducing its degradation; they do this by binding to the untranslated region of mRNA ends (Rahmani et al. 2020). They partake in numerous cancer-related processes, including apoptosis, differentiation, and cell cycle regulation (P Bhattacharyya et al. 2016). Dysregulation of explicit miRNAs has been explained to be significant in tumor development and progression. Likewise, miRNAs might function as a tumor suppressor or oncogenic miRNA in cancer (Saikia et al. 2020). Significantly, the steady presence of miRNAs in blood empowers them to be strong markers for the diagnosis of cancer, including BC. The results of studies show that miR-506 diminishes in hepatocellular carcinoma while miR-197 and miR-552 increase in this cancer (Lv et al. 2019).
The results of several studies suggest that miR-637 is involved in the pathogenesis of various cancers such as liver, stomach, melanoma, and adenocarcinoma, which has been shown to fall in cancerous tissues (Du and Wang 2019). The expression level of miR-133a was diminished in breast cancer patients. miR-133a expression level was linked to lymph node metastasis, high clinical stages, and poor survival of breast cancer patients . The next microRNA, miR-944, has a contradictory function and role in breast cancer. Indeed, in one study, it has been shown that the expression of miR-944 is strongly reduced in clinical samples and breast cancer cell lines (Flores-Pérez et al. 2016). However, in another study, an increase in the expression of miR-944 has been seen in the serum and tumor tissue of breast cancer patients, which stimulates cell proliferation and tumor metastasis in breast cancer cell lines (He et al. 2016).
One of the significant signaling pathways in the cell is the phosphatidylinositol 3-kinase/protein Kinase-B (PI3K/AKT) pathway, with crucial functions in different cellular processes such as metabolism, growth, proliferation, survival, transcription, and protein synthesis (Yu and Cui 2016). This signaling pathway consists of three main components, including PI3K, AKT, and the mammalian target of rapamycin (mTOR) (Rahmani et al. 2020). PI3K proteins are a group of heterodimeric lipid kinases characterized into three isoforms, class I, II, and III. Contingent upon their isoform, these proteins are activated by different cell surface receptors, including receptors for growth factors, antigens, cytokines, chemokines, and toll-like receptors (TLRs) (Jiang et al. 2020). After activation, class I of PI3Ks phosphorylates phosphatidyl-inositol 4-5-bisphosphate (PI(4,5)P2) and thus produces the molecule phosphatidyl-inositol triphosphate (PIP3), which is considered an intracellular second messenger (Jiang et al. 2020). With the production and accumulation of PIP3 in the cell membrane, AKT proteins are called to the site, phosphorylated, and exert their catalytic roles by activating their two regulatory domains. AKT1 is an upstream protein in the P13K/AKT signaling pathway, and its phosphorylation activates this pathway and regulates cancer cell proliferation and apoptosis (Du and Wang 2019;Miricescu et al. 2020). Studies have shown that the uncontrolled activity of the PI3K signaling pathway is involved in most malignancies, including breast cancer, which leads to increased growth and proliferation of cancer cells, inhibition of apoptosis, and tumorigenesis (Owusu-Brackett et al. 2019). It has been seen that miRNAs can affect this signaling pathway. The miR-147 inhibits tumorigenesis by targeting this pathway in breast cancer. Also, miR-214 enhances tumor activity by targeting this pathway (Abolghasemi et al. 2020).
In this study, we aimed to investigate the biological mechanism of miR-133a, miR-637, and miR-944 in breast cancer. Also, in light of bioinformatics discoveries and previous studies, to explore breast cancer metastasis, we investigated the correlation between miR-133a, miR-637, and miR-944 and PI3K/AKT pathway.

Materials and methods
This case-control study was carried out in Khatam Al-Anbia Hospital, Tehran, Iran from June 2021 to January 2022. Patients with breast cancer included 40 women with a histologically confirmed diagnosis of invasive ductal cancer (age 29-77 years). The control group consisted of 40 healthy women (29-68 years old) who were confirmed to have had a mammography screening test in the last year. These women and their families had no history of any type of cancer. Two groups were matched in terms of age and race, and written consent was obtained. Anthropometric data and information of the studied subjects were collected from medical records and questionnaires. Venous blood samples were taken from women with breast cancer and the healthy group. The tube containing the samples was placed at laboratory temperature for 30 min and afterward centrifuged at 2465 g for 5 min to segregate the serum, and further stored in cryotubes at − 80 °C. Tumor tissue and tumor margin tissue from women whose pathology was diagnosed as breast cancer were obtained during breast surgery. Before surgery, these patients should not have received any treatment such as radiation therapy, chemotherapy, targeted therapy, immunotherapy, and hormone therapy. A part of the tissues was immediately stored in liquid nitrogen to check the expression of desired genes and targets via quantitative realtime PCR (qRT-PCR) and then transferred to a freezer at -80 °C. Another part of the tissue was kept in formalin for immunohistochemically test. The research was approved by the Ethics Committee of Hamadan University of Medical Sciences (IR.UMSHA.REC.1400.088).

qRT-PCR
Total RNA was extracted from tissues and sera using Qiazol solution (Kiazist, Iran) and according to the manufacturer's instructions. The obtained RNA precipitate was kept in a freezer at − 80 °C. NanoDrop spectrophotometer (Thermo Fisher Scientific, USA) and 1% agarose gel electrophoresis were used to ensure the integrity of the RNA and the accuracy of the RNA extraction process. cDNA synthesis for PI3K (PIK3C2A and PIK3C2B) and AKT genes was performed according to the kit instructions of Parstous Company (Easy cDNA Synthesis Kit-Parstous-Iran). To synthesize specific cDNA for hsa-miR-133a, hsa-miR-637, hsa-miR-944, and U6, specific stem-loop primers were designed for each miRNA separately, where the cDNA synthesis of miRNAs was carried out using Parstous kit by stem-loop method which was done according to the manufacturer's instructions. (Easy cDNA Synthesis Kit-Parstous-Iran). The specific sequence of primers hsa-miR-133a, hsa-miR-637, hsa-miR-944, U6 (as an internal control) and PI3K (PIK3C2A and PIK3C2B), AKT and 18SrRNA (as reference) genes can be seen in Table 1. The qRT-PCR procedure was done utilizing using Real Q Plus 2 × Master Mix Green (Amplicon, Denmark) kits and LightCycler R 96 System (Roche Life Science, Deutschland GmbH Sandhofer, Mannheim, Germany).
Standard curves were drawn for all genes before data analysis. The specificity of PCR products was confirmed by resolving in 3% agarose gel (for miRNAs) and 1% agarose gel (for PI3K and AKT). Finally, the relative gene expression was assessed by the 2 −∆Ct and 2 −∆∆Ct . To compare the healthy group and breast cancer patients and thus tumor margin tissue and tumor tissue, 2 −∆∆Ct was employed. To compare tumor tissues (not margin) and serum of cancer patients (not healthy group), 2 −∆Ct was applied (Pfaffl 2001).

Immunohistochemistry (IHC)
To mold and cut the texture of the newly prepared samples, first the dehydration process was done with alcohol (ethanol). Xylene was used to remove the effect of alcohol. After xylol, paraffin was used to harden and prepare the tissue for molding and tissue cutting. After completing this stage, the samples of tumor tissue and tumor border tissue were cut using a microtome device to a thickness of 3 µm and placed on a silanized slide. The slides were placed in tris-buffered saline (TBS) 1X solution (T5912-Sigma) inside the microwave and after reaching the boiling point, the microwave was turned off the samples being left in the solution for 20 min. The samples were washed with PBS (P4417-Sigma) in 3 steps and 5 min apart. Then, H 2 O 2 (Sigma-7722-84-1) and methanol were mixed with a ratio of 1 to 9 and placed on the samples for 10 min. The samples were washed with PBS, where diluted PI3K and AKT antibodies (Biorbyt-Cambridge-United Kingdom) 1 to 100 were poured on the samples with phosphate-buffered saline (PBS) and left at room temperature for one hour. Next, the samples were washed 3 times with PBS for 5 min each time and 100 µL of Linker (D1000PVP-Diagnostic BioSystems) was added to the samples for 15 min. Next, they were washed 3 times with PBS, with 100 µL of polymer solution (D1000PVP-Diagnostic BioSystems) added to the samples for 30 min. The samples were washed again with PBS and then 100 μl of 3,3′-Diaminobenzidine (DAB) solution (ScyTek-ACV999) was added to the samples. After 5 min, the samples were washed with water and at the end, they were placed in hematoxylin dye for 10 s. They were then washed again with water and after the dehydrating and clarifying steps, the slide was stuck on the sample and photography was done using an optical microscope (LABOMED).

Statistical analysis
Results were presented as the mean ± standard deviation (SD). The student's t test was used to compare the differences between groups. Statistically, p value < 0.05 was considered significant. Pearson correlation coefficients were calculated between the expression of miR-133a, miR-637, and miR-944 expression in serum and tissue, each miRNA with its target genes, both PI3K (PIK3C2A and PIK3C2B) mRNA and PI3K protein and both AKT mRNA and AKT protein. In addition, the area under the curve (AUC) and receiver operating characteristic (ROC) curves were calculated to evaluate the diagnostic ability of miR-133a, miR-637, and miR-944 for breast cancer. Data were analyzed using GraphPad Prism 9.0 software and SPSS version 26.

Demographic and clinical findings of patients
All 80 people were included in this study, which were divided into two groups: healthy (40 people) and breast cancer patients (40 people). All members were female and their mean age was 54.2 ± 12.04 years. The data and anthropometric information of the patient group were gathered through a medical record and a questionnaire Table 3.

The expression level of miR-133a in tissue and serum and its diagnostic value
We investigated the expression of miR-133a in 40 sera of breast cancer patients and 40 healthy controls to recognize the function of miR-133a in breast cancer. As displayed in Fig. 1a, the expression level of miR-133a in the serum of breast cancer patients was significantly lower than that in the healthy group (p = 0.031). Also, we examined the tissue expression level of miR-133a in 40 tumor tissues and 40 tumor margin tissues. The expression level of miR-133a in tumor tissues was significantly lower than that in tumor margin tissues (p = 0.023, Fig. 1b). In addition, the results revealed that there was no significant relationship between the expression levels of miR-133a in the tissue and serum of breast cancer patients (r = 0.263, p = 0.101, Fig. 1c).

The expression level of miR-637 in tissue and serum and its diagnostic value
The expression level of miR-637 was examined in the serum of women with breast cancer (n = 40) and healthy women (n = 40). The results showed that the serum levels of miR-637 in women with breast cancer were significantly lower than that of healthy women (p = 0.005, Fig. 2a). In addition, the expression of miR-637 in tumor tissues (n = 40) and tumor margin tissues (n = 40) was investigated, where a significant reduction in miR-637 expression was seen in tumor tissues compared to the tumor margin tissues (p = 0.0001, Fig. 2b). Interestingly, we found a positive correlation between expression levels of miR-637 in the tissue and serum of women with breast cancer (r = 0.336, p = 0.034, Fig. 2c).

The expression level of miR-944 in tissue and serum and its diagnostic value
We explored the expression of miR-944 in the serum of 40 breast cancer patients and 40 healthy women. According to the results, the expression of miR-944 in the serum of breast cancer patients was increased compared to healthy individuals (p = 0.038, Fig. 3a). Also, we examined the expression of miR-944 in tumor and tumor margin tissues, where the results were not consistent with the expression of miR-944 in serum. The expression of miR-944 in the tumor tissues had decreased compared to the tumor margin tissues (p = 0.033, Fig. 3b). We also found a positive correlation between miR-944 expression levels in tissue and serum of women with breast cancer (r = 0.465, p = 0.003, Fig. 3c).

Relationship of expression levels of miR-133a, miR-637, and miR-944 in tissue and serum with clinicopathological features of breast cancer patients
As reported in Table 4, decreased expression of miR-133a in the tissue and serum of breast cancer patients was not related to any of the clinicopathological parameters of the patients. Based on the results, the expression level of miR-637 in the tissue of breast cancer patients was related to tumor location and HER2 receptor. The expression level of miR-637 was significantly higher in the tissue of patients with right tumor (p = 0.00). Also, the expression of this miR in the serum was higher in people who lacked the HER2 receptor compared to people who had HER2 receptor (p = 0.03). The expression of miR-944 in serum was significantly higher in tumors located in the left breast position (p = 0.03); also, patients who did not have a family history of cancer had a higher tissue expression of miR-944 (p = 0.00).

The gene and protein expression of PI3K in the tissue of breast cancer patients and its diagnostic value
In this study, we investigated the gene expression of PI3K (PIK3C2A and PIK3C2B) in 40 tumor tissues and 40 tumor margin tissues of breast cancer patients using qRT-PCR and its protein level in these tissues by the IHC method. The results of the study revealed that the expression level of PIK3C2A, PIK3C2B gene and PI3K protein level in tumor tissues was significantly higher than in the tumor margin tissues (p = 0.008, p = 0.006, p < 0.0001, respectively) ( Fig. 4a, b, c). IHC was performed to investigate the localization and expression of PI3K protein in tumor and tumor margin tissues (Fig. 6a, b). In addition, the relationship between PIK3C2A and PIK3C2B mRNA and its protein level in the tumor tissues of these patients was investigated, where this relationship was not significant (r = 0.177, p = 0.275, r = 0.186, p = 0.251, respectively) (Fig. 4d, h). The results of the ROC curve analysis for the gene expression of PIK3C2A indicated that the AUC was 0.724 (0.6127 to 0.8354) with the cut-off point > 0.9850 and 65% sensitivity, 80% specificity, Youden's index 0.450, and p value = 0.0006 (Fig. 4e). Also, PIK3C2B showed that the AUC was 0.728 (0.6175 to 0.8388) with the cutoff point > 1.063 and 65% sensitivity, 77.50% specificity, Youden's index 0.425, and p value = 0.0004 (Fig. 4f). In addition, for the PI3K protein level in tissue, the AUC was 1.000 (1.000 to 1.000) where the cut-off point was > 43.19 with a sensitivity of 100%, a specificity of 100%, a Youden's index of 1.00, and p value < 0.0001 (Fig. 4g).

The gene and protein expression of AKT in tissue of breast cancer patients and its diagnostic value
AKT gene was inspected in tumor tissues (n = 40) and tumor margin tissues (n = 40) of breast cancer patients using qRT-PCR technique whose protein level was also detected in these tissues by IHC method. The level of AKT gene expression in the tumor tissues revealed a significant increase compared to the tumor margin tissues (p = 0.003, Fig. 5a). In addition, the level of AKT protein in the tumor tissues was significantly increased compared to the tumor margin tissues (p < 0.0001, Fig. 5b). To localize and express AKT protein in tumor and tumor margin tissues, IHC was done (Fig. 6c, d). There was no significant relationship between AKT mRNA and its protein level in the tumor tissues of these patients (r = 0.185, p = 0.252, Fig. 5c).

Relationship of gene and protein expression levels of PI3K (PIK3C2A and PIK3C2B) and AKT in tumor tissue with clinicopathological features of breast cancer patients
According to the results, the expression level of PIK3C2A and PIK3C2B in the tissues of breast cancer patients was related to tumor location, histopathology, and family history. The expression level of PIK3C2A and PIK3C2B was significantly higher in the tissues of cancer patients whose tumor was located on the right side of the breast (p = 0.01, p = 0.01, respectively). In addition, the expression of this gene was higher in patients whose lobules were involved (p = 0.002, p = 0.002, respectively). In addition, the expression level of PIK3C2A and PIK3C2B was higher in patients who did not have a family history of cancer than those who had a family history (p = 0.01, p = 0.01, respectively) Table 5.
The results of this study indicated a significant correlation between AKT expression level and tumor location as well as histopathology. In patients whose tumor was located on the right side (p = 0.000) and patients whose lobules were involved (p = 0.004), AKT gene expression was higher Table 5.
According to the relationship between PI3K and AKT protein level and clinical features, no significant relationship was observed between the protein of this gene and clinical pathological features Table 5.

Correlation between expression levels of PI3K (PIK3C2A and PIK3C2B) and AKT and miR-133a, miR-637, and miR-944
To explore the diagnostic value of miR-133a, miR-637, and miR-944 in breast cancer patients, the correlation between miR-133a, miR-637, and miR-944, and PI3K (PIK3C2A and PIK3C2B) as well as AKT was examined. The results of the study did not show any relationship between gene and protein expression of PI3K or AKT and the expression of miR-133a, miR-637, and miR-944 Table 6.

Discussion
There are many signaling pathways in the cell that are responsible for gene expression and regulation of cellular processes. In these signaling pathways, there are molecules that do not directly interfere with transcription but help change gene expression by regulating transcription factors (Abolghasemi et al. 2020). Research has shown that miR-NAs play a role in different pathways, and these biomolecules can be effective in identifying and tracking various diseases such as stroke, types of cancer, depression, diabetes, cardiovascular diseases, and infectious diseases (ZiaSarabi et al. 2019). Due to their multigene regulatory features, miRNAs can positively or negatively regulate key signaling pathways, therefore they can affect in processes such as growth, apoptosis, differentiation, angiogenesis, and even inflammation (Abolghasemi et al. 2020;ZiaSarabi et al. 2019). MicroRNAs such as miR-196 and miR-155 can be used for early detection of pancreatic cancer and miR-378 in colorectal cancer (Xu et al. 2018;Pishbin et al. 2021). In colorectal cancer, cell death is inhibited by miR-219 (Saikia et al. 2020). In breast cancer, miR-21 and miR-155 can be In the present study, we researched the clinical significance of miR-133a, miR-637, and miR-944 in patients with breast cancer. The results of our study revealed that the expression levels of miR-133a and miR-637 in the tumor tissue and serum of patients were lower than in the tumor margin tissue and serum of the healthy group, respectively. Further, in the case of miR-944, its expression level in the tumor tissues was lower than in the tumor margin tissues, but its expression was increased in the serum of cancer patients compared to the healthy group. In addition, there was a significant relationship between the expression of miR-637 and miR-944 in the tissue and serum of breast cancer patients. In addition, PI3K and AKT gene and protein expression were higher in tumor tissues compared to tumor margin tissues.
miR-133a is known as a muscle-specific miRNA involved in the regulation of myoblast differentiation and many muscle-related diseases . Several studies have been conducted to investigate miR-133a in various cancers such as bladder, squamous cell carcinoma of the lung, as well as prostate and stomach cancer, suggesting that miR-133a acts as a tumor suppressor Hu et al. 2020). In breast cancer, Cui et al.'s study showed that miR-133a expression was reduced in breast cancer cell lines and tumor tissues (Cui et al. 2013). The results of our study also showed a decrease in miR-133a expression in the serum and tumor tissue of cancer patients. Studies have shown that the reduction of miR-133a expression in cancer cells causes disruption in cell growth and invasion (Wu et al. 2012). In prostate cancer, the reduction of miR-133a expression in cancer patients has been associated with clinicopathological characteristics and poor survival . However, in our study, miR-133a was not associated with any clinicopathological features. In addition, we used the ROC curve to explore the power of miR-133a as a diagnostic biomarker for breast cancer. The ROC curve revealed that serum and tissue levels of miR-133a did not have AUC, sensitivity, and specificity in breast cancer diagnosis.
miR-637 is located on chromosome 19p13.3 and has independent transcription units (Du and Wang 2019). A study was conducted by DU and colleagues on the miR-637 where it was seen that the miR-637 decreased in patients with liver cancer (Du and Wang 2019). The result of Leivonen et al.'s study indicated that miR-637 can be effective as an inhibitor of HER2 signaling and cell growth (Leivonen et al. 2014). The f ROC curve analysis based on gene expression of PIK3C2B. g ROC curve analysis based on the protein level of PI3K and h correlation between PIK3C2B mRNA and its protein level in tumor tissues. **p < 0.01 and ****p < 0.0001 compared to tumor margin tissue. PI3K Phosphatidylinositol-3-kinase, ROC receiver operating characteristic results of our study showed that miR-637 expression was significantly reduced in tumor tissues and serum of breast cancer patients. Also, a significant positive correlation was observed between the expression levels of miR-637 in the tissue and serum of women with breast cancer. In Xu et al.'s study, a drop in miR-637 expression was seen in pancreatic cancer and its tumor suppressor role was emphasized (Xu et al. 2018). The results of our study indicated a relationship between miR-637 and clinicopathological characteristics, including tumor location and HER2 receptors. In addition, we used ROC curve to explore the power of miR-637 as a diagnostic biomarker in breast cancer. The ROC curve showed that the tissue levels of miR-637 had good AUC, sensitivity, and specificity in breast cancer diagnosis. DU and colleagues also investigated the ROC for miR-637 in tumor tissues of liver cancer (Du and Wang 2019) where they found a similar result to our study.
Sometimes, the amount of miRNA detected in blood and tissue samples is different. For example, miR-145 had increased in some studies of breast cancer and decreased in some other reports (Abolghasemi et al. 2020). In the case of miR-944, the results of studies are contradictory, where in one study, its elevation was seen in breast cancer, and it was introduced as oncogenic (He et al. 2016), while in another study, a reduction of miR-944 is seen, and it was identified as a tumor suppressor (Flores-Pérez et al. 2016). Studies have shown that the miR-944 gene is located in the intron of the P63 gene, which as a transcription factor is often suppressed in breast cancer (Flores-Pérez et al. 2016). Increased miR-944expression has been seen in endometrial cancer and cervical cancer (Lv et al. 2019). The results of our studies have also shown a decline in the expression of miR-944 in the tumor tissues compared to the tumor margin tissues, which has a significant relationship with some histopathological parameters such as the location of the tumor and family history. The results of the studies by Lv et al. and PEI et al. show a reduction in miR-944 expression in hepatocellular carcinoma and colorectal cancer, respectively (Lv et al. 2019;Pei et al. 2019). Examination of the serum expression of miR-944 in our study showed an increase in its expression in the serum of patients compared to the healthy group. In the case of miR-195, something similar has been seen in breast cancer, where its serum expression was increased in patients, but its tissue expression was decreased. The results of studies have shown that the cause of this difference in serum and tissue expression can be due to disease or other malignancy or epigenetic changes. It has been seen that in some microRNAs, DNA methylation (CPG sequence) upstream of the promoter affects its expression and causes a reduction in expression (Godfrey et al. 2013;Li et al. 2011). To investigate the power of miR-944 as a biomarker in breast cancer diagnosis, we used the ROC curve. The ROC curve showed that miR-944 tissue levels have AUC, sensitivity, and relative specificity in breast cancer diagnosis. The PI3K/AKT/mTOR complex is a central intracellular signaling pathway which controls essential cellular activities such as cell metabolism, cell growth, cell proliferation, apoptosis, and angiogenesis (Rahmani et al. 2020). This signaling pathway consists of three main components including PI3K, AKT, and mTOR (Kołodziej et al. 2021). PI3Ks are an important family of plasma membrane-bound lipid enzymes activated by RTKs (receptor tyrosine kinase) and by GPCRs (G protein-coupled receptors) (Miricescu et al. 2020). This enzyme family has four different classes, where class I of this enzyme has been studied in cancer. The structure of class I of PI3Kenzyme consists of regulatory (p85) and catalytic subunits (p110). Class I PI3Ks catalyze the conversion of phosphatidyl 4,5-bisphosphate (PI(4,5) P2) by phosphorylation to phosphatidyl 3,4,5-triphosphate (IP3), which is considered an intracellular second messenger. With the production and accumulation of PIP3 in the cell membrane, AKT proteins are called to the site, phosphorylated, and exert their catalytic roles with the activation of their two regulatory domains (Rascio et al. 2021). AKT is an evolutionarily conserved serine/threonine protein kinase. Through the phosphorylation of many downstream proteins such as forkhead box proteins (FOXO), mTOR, glycogen synthase kinase 3 beta (GSK3b), and many other factors, this protein controls many intracellular processes.
AKT has three isoforms, whose expression levels are different in body tissues, including AKT1 (in most tissues), AKT2 (in tissues with high sensitivity to insulin), and AKT3 (in brain and testis) (Miricescu et al. 2020). Akt1 is involved in many cellular processes such as cell proliferation, cell apoptosis, and glucose metabolism. Studies have shown that AKT1 is highly expressed in liver, pancreas, and breast cancer tissues and plays an important role in the development of cancer (Rascio et al. 2021). AKT activates the mTOR complex, which plays a role in regulating protein synthesis, cell survival, and inhibition of autophagy (Jiang et al. 2020). The presence of multiple mutations in sensitive points of the PIK3CA gene has been observed in 80% of breast cancer cases. Also, in 70% of patients with breast cancer, the activated PI3K pathway has been identified, which ultimately results in the activation of tumorigenic pathways (Jiang et al. 2020). Our results indicated that the relative expression of PI3K and AKT increased in tumor tissues compared to tumor margin tissues. In addition, the upregulation of PI3K and AKT was correlated with tumor location and histopathology. In this study, the protein expression of PI3K and AKT was investigated by immunohistochemical method, where the protein levels of PI3K and AKT were significantly higher in the tumor tissues compared to the tumor margin tissues. Also, we used the ROC curve to explore the power of PI3K and AKT as diagnostic biomarker for breast cancer. The ROC curve revealed that PI3K and AKT protein (due to high specificity and sensitivity) may have a better diagnostic value than PI3K and AKT gene expression in breast cancer patients. Studies show that miRNAs can target the key components of this PI3K/AKT pathway (Tarhriz et al. 2019). Zhang et al. found that miR-147 inhibits tumorigenesis in breast cancer cells by targeting the AKT/mTOR signaling pathway (Zhang et al. 2016). Another miRNA is miR-21, which stimulates the PI3K/AKT signaling pathway (Yan et al. 2016). The study of Cu et al. and Hu et al. showed that the expression of miR-133a caused the inactivation of PI3K/AKT pathways (Hu et al. 2020;Cui et al. 2013).
On the other hand, a study was conducted on prostate cancer and it was seen that the reduction in the expression of miR-133a caused the activation of the PI3K/AKT pathway and the induction of metastasis . The results of our study indicated that there was a relationship between the expression of miR-133a and PI3K/AKT, but this relationship was not significant. Studies conducted in various cancers such as hepatocarcinoma (Du and Wang 2019), glioma (Que et al. 2015), and pancreatic cancer (Xu et al. 2018) have shown that the decrease in the expression of miR-637 induces the PI3K/AKT pathway and enhances the activity of AKT. The expression of miR-637 in breast cancer cells has been investigated, showing a reduction (Xu et al. 2018), which confirms our results. In addition, the results of our study showed that there was a negative relationship, but this relationship was not significant. A study conducted on miR-944 in hepatocarcinoma showed that with the increase of miR-944 expression, the PI3K/AKT pathway decreased and there was an inverse relationship (Lv et al. 2019), which is also seen in our results, but the difference has not been significant.

Conclusion
We investigated the clinical significance of miR-133a, miR-637, and miR-944 in breast cancer patients to evaluate a marker for cancer diagnosis. According to the serum and tissue concentrations of miR-133a, miR-637, and miR-944 as well as the degree of correlation between microRNAs in tumor tissue and patients' serum, as along with the correlation between microRNAs and PI3K, AKT genes, miR-637 is considered as a better marker for breast cancer diagnosis than miR-133a and miR-944.