Circulating microRNAs and serum proteins in breast cancer patients: Diagnostic relevance and grade-specific expression patterns

Abstract

BACKGROUND

Breast cancer is a prominent contributor to female cancer-related mortality. Early detection and accurate diagnosis are essential for effective management.

AIM

To evaluate the diagnostic relevance of a panel of circulating microRNAs (miRNAs) independently or in combination with other tumor biomarkers and evaluate their sensitivity and specificity in classifying breast cancer patients by grade.

METHODS

In the present study, we analyzed the aberrant expression of miR-21, miR-221, miR-1246, miR-145, and miR-382, in addition to the tumor biomarkers cancer antigen 15-3 (CA15-3) and 8-hydroxy-2′-deoxyguanosine (8-OHdG) in breast cancer patients with varying grades.

RESULTS

Our results revealed distinct expression patterns of these miRNAs between grade II and III patients. Specifically, miR-21, miR-221, and miR-1246 were significantly elevated, while miR-145 and miR-382 were downregulated. Elevated serum levels of CA15-3 and 8-OHdG were observed in breast cancer patients compared to healthy controls, with CA15-3 showing greater diagnostic efficacy in differentiating between grades. Our study revealed strong correlations among evaluated miRNAs, suggesting their interconnected roles in breast cancer progression. Receiver operating characteristic curve analysis demonstrated high diagnostic accuracy for all investigated miRNAs, with miR-21 and miR-1246 showing the highest diagnostic power for differentiating patients from healthy individuals and distinguishing breast cancer grades. Moreover, the combination of multiple miRNAs and conventional tumor biomarkers revealed enhanced diagnostic accuracy and sensitivity.

CONCLUSION

These findings suggest that circulating miRNAs may play a significant role in distinguishing breast cancer patients based on tumor grade, with superior diagnostic performance over some tumor biomarkers, supporting the development of multi-analyte liquid biopsy approaches in the diagnostic process and personalized management of breast cancer patients.

Key Words: Breast cancer; Liquid biopsy; Tumor grading; MicroRNAs; Cancer antigen 15-3; 8-hydroxy-2′-deoxyguanosine; Sensitivity; Specificity

Core Tip: This study highlights the diagnostic utility of a panel of circulating microRNAs (miRNAs), miR-21, miR-221, miR-1246, miR-145, and miR-382, in stratifying breast cancer patients by tumor grade. Notably, miR-21 and miR-1246 exhibited the highest diagnostic power in distinguishing patients from healthy controls. Importantly, integrating these miRNAs with traditional protein biomarkers cancer antigen 15-3 and 8-hydroxy-2′-deoxyguanosine significantly enhanced diagnostic accuracy and sensitivity, underscoring the value of a multi-marker approach for non-invasive breast cancer detection and grading.

INTRODUCTION

Breast cancer is a widespread health issue of worldwide significance, which also affects Egypt[1]. Breast cancer can be classified into different subtypes according to receptor status as well as distinct grades and stages based on the degree of tumor development. Current clinical diagnosis and classification of breast cancer rely on histological grading of the tumor and imaging techniques that are costly and tedious[2]. However, liquid biopsy is emerging as an increasingly essential approach, offering cost-effective, specific, and sensitive molecular diagnostic markers with the ability to detect the different molecular subtypes, grades, and stages. Thus, it enhances the clinical management of cancer, making it a valuable asset in clinical practice[3].

The discovery of non-coding RNAs and their involvement in gene regulation transformed the area of clinical research. This ground-breaking discovery has provided a new paradigm for the development and improvement of novel disease status predictions[4,5]. MicroRNAs (miRNAs) are short, non-coding RNA molecules, 19-22 nucleotides long, and have emerged as important regulators of gene expression by sequence-specific base pairing of target mRNA. Over the last decade, substantial research has highlighted the crucial role of miRNAs as oncogenic miRNAs (oncomiRs) and tumor suppressors, highlighting their contribution to tumor initiation and progression[6-8].

Growing evidence suggests that miRNAs, whether free-circulating or encapsulated in extracellular vesicles, play crucial roles in the development of breast cancer, with their abnormal expression patterns serving as potential indicators of tumor status[9,10]. Circulating miRNAs are stable and detectable in serum/plasma, and levels of some miRNAs in breast cancer patients could be used as biomarkers for diagnosis[11], prognosis[10], and therapeutic monitoring indicators[12]. Previous studies have shown that several miRNAs are closely associated with the clinicopathological features of the tumor, such as the grade, stage, and hormonal receptor status, suggesting their potential use as grading biomarkers[13]. Numerous studies have focused on using a single miRNA as a marker for diagnosing and/or prognosis. However, it is evident that combining multiple miRNAs or combining miRNAs with clinically applied biomarkers would yield better performance than using a single miRNA[14-16]. The combination of several biomarkers may present a more thorough molecular profile of the disease and overcome the shortcomings of individual indicators.

Cancer antigen 15-3 (CA15-3) is a cost-effective blood tumor marker that is frequently employed in clinical practice to diagnose breast cancer. However, since CA15-3 Levels can be elevated in benign and malignant conditions, it is not sufficiently selective for diagnosing or discriminating between cancer stages. As a result, the combination of CA15-3 and other markers is one of the most widely used cancer screening approaches today[17,18].

8-hydroxy-2′-deoxyguanosine (8-OHdG), a byproduct of oxidative damage, is a valuable marker for measuring oxidative DNA damage and has been linked to malignancies[19]. Previous studies have linked the expression of 8-OHdG to tumor size, clinical stage, and overall survival in various types of cancer[20,21]. For breast cancer, 8-OHdG expression differed significantly between patients and healthy controls. As breast cancer progresses, the oxidative stress burden often elevates, making 8-OHdG a potential indicator of tumor grade and aggressiveness[22]. Therefore, researchers suggest that 8-OHdG, the biomarker of DNA damage caused by oxidative stress, can be utilized for breast cancer diagnosis and prognosis.

In the present study, we evaluated the aberrant expression of a panel of circulating miRNAs, along with the tumor biomarkers CA15-3 and 8-OHdG, in breast cancer patients with various grades. Our aim is to assess the diagnostic relevance of the identified panel of miRNAs independently or in combination with serum CA15-3 and 8-OHdG and evaluate their sensitivity and specificity in classifying breast cancer patients into grades. By integrating these biomarkers, we seek to improve the accuracy of disease grading, which will eventually contribute to advancing the field of breast cancer diagnostics and optimizing treatment regimens for breast cancer patients.

MATERIALS AND METHODS

Ethics statement

Patients were recruited from Baheya Centre of Early Detection and Treatment of Breast Cancer, Giza, Egypt. The study has been approved by the Ethical Committee of the National Research Centre following the ethical standards of the Declaration of Helsinki.

Patient selection

The current study involved 49 adult female patients with primary breast cancer, confirmed through clinical, radiological, and pathological examinations. A control group consisting of 20 healthy female volunteers, matched for age, was also recruited from the local community. The number of participants in this study was determined based on similar previously published research and the availability of eligible patients during the study period.

Control participants were confirmed to have no previous history of malignancy and reported no significant chronic diseases. Exclusion criteria applied to all participants included pregnancy or lactation and the presence of significant concomitant illnesses such as liver disease, renal failure, heart disease, or diabetes mellitus. Clinicopathological data of all patients were recorded from patients' data sheets, including tumor size, grade, estrogen receptor, progesterone receptor, human epidermal growth factor receptor 2 (HER2) status, tumor subtypes, and TNM stage (T: Tumor size, N: Lymph node status, M: Metastasis). TNM stage was calculated according to the classification of the 8^th Edition of the American Joint Committee on Cancer[23].

Sample collection

Blood samples from patients and controls were collected following the same protocol. A 5 mL sample of whole blood was withdrawn from each participant using gel vacutainer tubes. The blood samples were allowed to coagulate at room temperature for 10-20 minutes before being centrifuged at 3000 rpm for 20 minutes, and the serum was separated. A second centrifugation process was then carried out to eliminate any remaining platelets, allowing for more efficient cell-free RNA extraction. Serum samples were stored at -80 °C for further analysis.

Selection of miRNAs for analysis

The rationale for selecting the miRNAs panel (miR-21, miR-145, miR-221, miR-382, and miR-1246) as potential diagnostic biomarkers was based on an extensive literature review. Our strategy involved reviewing relevant studies to identify miRNAs with established roles in breast cancer pathogenesis and potential diagnostic and prognostic significance. We prioritized miRNAs associated with clinical outcomes, such as tumor grade, metastasis, and patient survival. The chosen panel of miRNAs (miR-21, miR-145, miR-221, miR-382, and miR-1246) represents a combination of well-established and new potential biomarkers for breast cancer. This approach allows for the evaluation of both widely studied and potentially novel diagnostic markers in the context of our study.

RNA extraction and quantitative real-time polymerase chain reaction

For the isolation of total RNA from serum samples, the miRNeasy Mini Kit (Qiagen, Hilden, Germany) was used. Briefly, 200 μL of serum was mixed with 1 mL of QIAzol Lysis Reagent. After phase separation using chloroform, the aqueous phase containing RNA was separated and loaded into the miRNeasy column. Following several washing steps, RNA was eluted with RNase-free water. For quality assessment, eluted RNA was evaluated for purity and concentration using a NanoDrop microvolume spectrophotometer (Thermo Scientific^TM). Only high-quality RNA samples were used for subsequent steps. For reverse transcription, the miRCURY LNA Reverse Transcription Kit (Qiagen, No. 339340) was employed to synthesize cDNA from the isolated RNA.

Real-time polymerase chain reaction (PCR) was performed using the miRCURY LNA SYBR Green PCR Kit (Qiagen, No. 339346). Primers of miRNAs were purchased from the Qiagen collection miRCURY LNA miRNA PCR Assays as follows: (1) Hsa-miR-16-5p [GeneGlobe ID: YP00205702 (No. 339306)]; (2) Hsa-miR-21-5p [GeneGlobe ID: YP00204230 (No. 339306)]; (3) Hsa-miR-145-5p [GeneGlobe ID: YP00204483 (No. 339306)]; (4) Hsa-miR-221 [GeneGlobe ID: YP00204032 (No. 339306)]; (5) Hsa-miR-382-5p [GeneGlobe ID: YP00204169 (No. 339306)]; and (6) Hsa-miR-1246 [GeneGlobe ID: YP00205630 (No. 339306)]. Real-time PCR was performed on the Applied Biosystems QuantStudio^TM 5 Real-Time PCR System (Thermo Fisher Scientific). The Ct value of each target miRNA was normalized to the Ct value of the internal reference control, miR-16. The differential expression of each miRNA was represented as fold change and estimated using the 2^-ΔΔCt analysis method.

Estimation of CA15-3 and 8-oHdG by enzyme-linked immunosorbent assay

Sandwich-enzyme-linked immunosorbent assay kits were used to evaluate levels of CA15-3 (SunLong Biotech Co., China, No. SL0383Hu) and 8-OHdG (SunLong Biotech Co., China, No. SL2044Hu) in the sera of the collected samples. Standard curves were plotted to calculate the concentrations of CA15-3 (U/mL) and 8-OHdG (pg/mL) after measuring their absorbance using a microplate reader.

Statistical analysis

The statistical analyses and calculations were carried out with the Statistical Package for the Social Sciences (SPSS, Inc., Chicago, United States) version 21.0 software. A one-way analysis of variance (ANOVA) test was conducted to investigate differences across groups. For parametric data, the least significant difference test was used to make multiple comparisons across all groups. In contrast, the Kruskal-Wallis test was employed for non-parametric data, along with the Mann-Whitney test applied to multiple comparisons. Spearman correlation analysis was utilized to determine the strength and direction of the variables' correlations. A receiver operating characteristic (ROC) analysis was performed to assess the diagnostic performance represented by the area under the curve (AUC), sensitivity, and specificity of the investigated biomarkers. Furthermore, a Combined ROC analysis was used to evaluate the combined performance of multiple biomarkers. Significance was set at P < 0.05.

RESULTS

Clinicopathological and molecular characteristics of participants

A sample of 69 women, 49 breast cancer patients (mean age: 49.2 years ± 11.6 years), and 20 age-matched healthy individuals participated in our study. All data is collected and presented in Table 1.

Table 1 The clinicopathological features of breast cancer patients, n (%).

Parameters		Groups		χ²	P value
		Grade II	Grade III
Estrogen receptor	Positive	20 (80.00)	14 (58.30)	2.706	0.100
	Negative	5 (20.00)	10 (41.70)
Progesterone receptor	Positive	22 (88.00)	13 (54.20)	6.868	0.009
	Negative	3 (12.00)	11 (45.80)
Human epidermal growth factor receptor 2	Positive	19 (76.00)	10 (41.70)	5.975	0.015
	Negative	6 (24.00)	14 (58.30)
Family history	Positive	12 (48.00)	7 (29.20)	1.829	0.176
	Negative	13 (52.00)	17 (70.80)
Tumor size	1	9 (36.00)	2 (8.30)	10.599	0.014
	2	13 (52.00)	10 (41.70)
	3	0 (0.00)	2 (8.30)
	4	3 (12.00)	10 (41.70)
Lymph nodal involvement	0	18 (72.00)	8 (33.30)	7.772	0.050
	1	5 (20.00)	9 (37.50)
	2	1 (4.00)	4 (16.70)
	3	1 (4.00)	3 (12.50)
Metastasis	0	22 (88.00)	18 (75.00)	1.380	0.240
	1	3 (12.00)	6 (25.00)
Metastatic	Yes	3 (12.00)	6 (25.00)	1.380	0.240
	No	22 (88.00)	18 (75.00)
Pathological subtype	Invasive ductal carcinoma	21 (84.00)	24 (100.00)	4.181	0.124
	Invasive lobular carcinoma	2 (8.00)	0 (0.00)
	Other	2 (8.00)	0 (0.00)

Expression profiles of the selected miRNAs and their associations with the patients' grade

The aberrant expression of our selected panel of miRNAs revealed distinct expression patterns when comparing breast cancer patients and controls. Among the five candidate miRNAs, miR-21, miR-221, and miR-1246 were significantly elevated in breast cancer patients compared to healthy controls. Conversely, miR-145 and miR-382 were found to be downregulated in breast cancer patients relative to controls (Figure 1, Table 2). Our study focused on analyzing the expression level of these miRNAs between breast cancer patients, relevant to their pathological grades. We observed significant differences in the expression levels of miR-21, miR-145, miR-221, miR-1246, and miR-382 when comparing grade II patients with those labeled as grade III (grading comparison). These findings indicate that these miRNAs may have a significant impact on differentiating breast cancer patients according to tumor grade. However, no significant difference was detected in the expression of miR-382 between patients with different grades. This suggests that miR-382 may not have a substantial role in differentiating between patients based on grading.

Figure 1 Relative expression values for the selected panel of microRNAs (miR-21, miR-221, miR-1246, miR-145, and miR-382) in breast cancer patients (grade II and grade III) and healthy controls. Significance at ^aP < 0.05 compared to healthy control; significance at ^bP < 0.05 between grade II and grade III. MiRNAs: MicroRNAs.

Table 2 Relative expression values of the selected panel of microRNAs and levels of serum proteins (8-hydroxy-2′-deoxyguanosine and cancer antigen 15-3) in breast cancer patients and healthy control.

Parameters	Groups	Minimum	Maximum	mean ± SD	Median	P value
MiR-21	Control	0.64	1.69	0.94 ± 0.24	0.89	0.000
	Grade II patients	1.28	3.79	2.31 ± 0.73a	2.11
	Grade III patients	3.78	8.54	6.59 ± 1.47a,b	6.84
MiR-221	Control	0.64	1.87	1.07 ± 0.34	0.95	0.000
	Grade II patients	0.89	2.48	1.47 ± 0.37a	1.41
	Grade III patients	1.36	4.56	2.53 ± 0.95a,b	2.40
MiR-1246	Control	0.12	2.89	1.08 ± 0.57	1.04	0.000
	Grade II patients	1.21	8.32	3.40 ± 1.57a	3.31
	Grade III patients	1.67	30.45	12.02 ± 7.85a,b	9.17
MiR-145	Control	0.41	1.09	0.80 ± 0.21	0.88	0.000
	Grade II patients	0.21	0.71	0.38 ± 0.14a	0.34
	Grade III patients	0.08	0.56	0.22 ± 0.13a,b	0.19
MiR-382	Control	0.26	1.18	0.80 ± 0.26	0.86	0.000
	Grade II patients	0.13	0.79	0.39 ± 0.16a	0.35
	Grade III patients	0.04	0.79	0.26 ± 0.24a	0.19
8-hydroxy-2′-deoxyguanosine (Pg/mL)	Control	643.00	2648.00	917.60 ± 515.30	740.00	0.000
	Grade II patients	1849.00	3997.00	3292.92 ± 611.56a	3096.00
	Grade III patients	1849.00	3997.00	3480.00 ± 662.59a	3757.00
Cancer antigen 15-3 (U/mL)	Control	15.87	135.60	38.12 ± 41.91	20.67	0.000
	Grade II patients	119.70	420.00	241.94 ± 86.34a	222.40
	Grade III patients	122.40	435.30	311.00 ± 90.83a,b	319.85

^aP < 0.05, significant difference compared to control at a significant level.

^bP < 0.05, significant difference between grade II and grade III breast cancer patients at a significant level.

MiRNAs: MicroRNAs.

Levels of hydroxy-2′-deoxyguanosine (8-OHdG) and CA15-3

A significant increase in the serum levels of 8-OHdG and CA15-3 was observed in all breast cancer patients compared to control individuals. The ANOVA test analysis demonstrated a statistically significant difference in 8-OHdG levels between breast cancer patients (grade II and grade III) and the control group. Nevertheless, no significant difference was found between grade II and grade III breast cancer patients. In contrast, CA15-3 Levels showed significant differences between controls and breast cancer patients, as well as between patients with different grades (Figure 2, Table 2).

Figure 2 Serum levels of 8-hydroxy-2′-deoxyguanosine (pg/mL) and cancer antigen 15-3 (U/mL). Significance at ^aP < 0.05 compared to healthy control; significance at ^bP < 0.05 between grade II and grade III. CA15-3: Cancer antigen 15-3; 8-OhdG: 8-hydroxy-2′-deoxyguanosine.

Correlation analysis between circulating miRNAs and tumor biomarkers

In our study, we conducted a Spearman correlation coefficient analysis to delineate the potential connection among the detected circulating miRNAs and among the conventional tumor biomarkers CA15-3 and 8-OHdG. Additionally, correlations between circulating miRNAs and tumor biomarkers were assessed. The results revealed that miR-21 showed the highest correlation coefficient values with the other miRNAs. MiR-21 revealed a strong positive correlation with miR-221 (r = 0.678, P = 0.000) and miR-1246 (r = 0.764, P = 0.000), while a strong negative correlation with miR-145 (r = -0.738, P = 0.000) and miR-382 (r = -0.755, P = 0.000). Additionally, strong positive correlations between miR-21 and the tumor biomarkers 8-OHdG (r = 0.644, P = 0.000) and CA15-3 (r = 0.694, P = 0.000) were recorded. Similarly, but to a lesser extent, miR-1246 displayed a strong positive correlation with miR-221 (r = 0.625, P = 0.000) and a strong negative correlation with miR-145 (r = -0.681, P = 0.000) and miR-382 (r =-0.632, P = 0.000). Moreover, miR-221 also revealed positive correlations with 8-OHdG (r = 0.537, P = 0.000) and CA15-3 (r = 0.472, P = 0.000). These correlations suggest that the expression of these miRNAs is interconnected with cooperative function, either in a positive or negative direction, and may play important roles in breast cancer progression and grading. Interestingly, the degree of correlation between the protein biomarkers 8-OHdG and CA15-3 (r = 0.546) was lower than the correlation detected between the miRNAs (Table 3).

Table 3 Spearman’s correlation coefficient between microRNAs and tumor biomarkers.

	MiR-21	MiR-221	MiR-1246	MiR-145	MiR-382	8-OHdG	CA15-3
MiR-21	1.00	0.678a	0.764a	-0.738a	-0.755a	0.644a	0.694a
MiR-221		1.00	0.625a	-0.600a	-0.542a	0.537a	0.472a
MiR-1246			1.00	-0.681a	-0.632a	0.650a	0.684a
MiR-145				1.00	-0.681a	-0.515a	-0.663a
MiR-382					1.00	-0.576a	-0.500a
8-OHdG						1.00	0.546a
CA15-3							1.00

^aP < 0.01, correlation is significant.

CA15-3: Cancer antigen 15-3; MiRNAs: MicroRNAs; 8-OhdG: 8-hydroxy-2′-deoxyguanosine.

Diagnostic accuracy of miRNAs and tumor biomarkers in breast cancer patients

The ROC curve was used to analyze the performance of miRNAs and protein tumor biomarkers, either individually or combined, in distinguishing between subgroups of breast cancer patients and healthy controls. For breast cancer patients (grade II) as compared to control, the results revealed that circulating miRNAs, miR-21, miR-1246, and miR-145 have the highest diagnostic power with an AUC value of 0.988 for miR-21 (95%CI: 0.962-1.000), an AUC value of 0.974 for miR-1246 (95%CI: 0.935-1.000), and an AUC value of 0.950 for miR-145 (95%CI: 0.893-1.000). Interestingly, the 8-OHdG biomarker showed greater diagnostic efficacy than the routinely measured biomarker CA15-3, with an AUC of 0.996 for 8-OHdG (95%CI: 0.985-1.000). ROC curve analyses also showed that combining more than one miRNA could offer a higher sensitivity, specificity, and diagnostic ability. The combination of miR-145 + miR-382 + miR-1246 showed the highest diagnostic accuracy and the highest sensitivity and specificity for grade II patients, with an AUC of 1.000 (95%CI: 1.000-1.000). Similarly, the combination of miR-145 and 8-OHdG offered the same highest diagnostic accuracy, sensitivity, and specificity, emphasizing the value of using multiple markers in cancer diagnostics (Figure 3, Table 4).

Figure 3 Receiver operating characteristic curve analysis demonstrating the diagnostic performance of selected microRNAs and protein tumor biomarkers in differentiating breast cancer patients (grade II) from healthy controls. AUC: Area under the curve; CA15-3: Cancer antigen 15-3; MiRNAs: MicroRNAs; 8-OhdG: 8-hydroxy-2′-deoxyguanosine.

Table 4 Diagnostic performance of microRNAs and other tumor biomarkers in breast cancer patients (grade II) relative to the control group.

Parameters	Area under the curve	Cut-off value	Sensitivity (%)	Specificity (%)	P value	95%CI
MiR-21	0.988	1.25	100.0	95.0	0.000	0.962-1.000
MiR-221	0.804	1.06	88.0	75.0	0.001	0.664-0.944
MiR-1246	0.974	1.99	92.0	95.0	0.000	0.935-1.000
MiR-145	0.950	0.46	84.0	95.0	0.000	0.893-1.000
MiR-382	0.888	0.77	96.0	75.0	0.000	0.780-0.996
8-OHdG	0.996	1822.00	100.0	95.0	0.000	0.985-1.000
Cancer antigen 15-3	0.982	158.15	88.0	100.0	0.000	0.953-1.000
MiR-145 + MiR-382 + MiR-1246	1.000	-	100.0	100.0	0.000	1.000-1.000
MiR-145 + 8-OHdG	1.000	-	100.0	100.0	0.000	1.000-1.000

MiRNAs: MicroRNAs; 8-OhdG: 8-hydroxy-2′-deoxyguanosine.

On the other hand, the ROC curve analysis for discrimination breast cancer patients (grade III) from healthy individuals revealed high diagnostic accuracy for all investigated miRNAs (Figure 4, Table 5), listing miR-21 (AUC = 1.000, 95%CI: 1.000-1.000), miR-1246 (AUC = 0.990, 95%CI: 0.969-1.000), and miR-145 (AUC = 0.978, 95%CI: 0.945-1.000) with the highest diagnostic score. Similar to the previous ROC curve analysis for breast cancer patients (grade II), the biomarker 8-OHdG expressed a more diagnostic efficacy than the routinely measured biomarker CA15-3, with an AUC of 0.994 for 8-OHdG (95%CI: 0.979-1.000). Interestingly, ROC curve analysis for combined biomarkers demonstrated that the combination of miR-382 and miR-221 was the only biomarker combination that achieved high diagnostic accuracy (AUC = 0.992, 95%CI: 0.973-1.000) in comparison to the control.

Figure 4 Receiver operating characteristic curve analysis demonstrating the diagnostic performance of selected microRNAs and protein tumor biomarkers in differentiating breast cancer patients (grade III) from healthy controls. AUC: Area under the curve; CA15-3: Cancer antigen 15-3; MiRNAs: MicroRNAs; 8-OhdG: 8-hydroxy-2′-deoxyguanosine.

Table 5 Diagnostic performance of microRNAs and other tumor biomarkers in breast cancer patients (grade III) relative to the control group.

Parameters	Area under the curve	Cut-off value	Sensitivity (%)	Specificity (%)	P value	95%CI
MiR-21	1.000	2.73	100.0	100.0	0.000	1.000-1.000
MiR-221	0.946	1.20	100.0	75.0	0.000	0.886-1.000
MiR-1246	0.990	2.035	95.8	95.0	0.000	0.969-1.000
MiR-145	0.978	0.35	87.5	100.0	0.000	0.945-1.000
MiR-382	0.950	0.305	79.2	95.0	0.000	0.894-1.000
8-hydroxy-2′-deoxyguanosine	0.994	1822.00	100.0	95.0	0.000	0.979-1.000
Cancer antigen 15-3	0.985	178.00	87.5	100.0	0.000	0.961-1.000
MiR-382 + MiR-221	0.992	-	95.8	100.0	0.000	0.973-1.000

MiRNAs: MicroRNAs.

Interestingly, by evaluating the diagnostic performance of our selected panel of miRNAs and protein tumor biomarkers to differentiate between various grades of breast cancer patients (specifically grade II vs grade III), we found that circulating miRNAs were more diagnostically powerful than 8-OHdG or the conventional tumor biomarker CA15-3. Where miR-1246 and miR-21 showed the highest diagnostic score with an AUC of 1.000 for miR-1246 (95%CI: 0.927-1.000) and an AUC of 0.997 for miR-21 (95%CI: 0.921-1.000) (Figure 5, Table 6).

Figure 5 Receiver operating characteristic curve analysis demonstrating the diagnostic performance of selected microRNAs and protein tumor biomarkers in differentiating breast cancer patients (grade II vs grade III). AUC: Area under the curve; CA15-3: Cancer antigen 15-3; MiRNAs: MicroRNAs; 8-OhdG: 8-hydroxy-2′-deoxyguanosine.

Table 6 Diagnostic performance of microRNAs and other tumor biomarkers for differentiating between breast cancer patients (grade II vs grade III).

Parameters	Area under the curve	Cut-off value	Sensitivity (%)	Specificity (%)	P value	95%CI
MiR-21	0.997	3.79	95.83	100.00	0.001	0.921-1.000
MiR-221	0.843	2.11	75.00	96.00	0.001	0.711-0.931
MiR-1246	1.000	4.26	100.00	100.00	0.001	0.927-1.000
MiR-145	0.941	0.26	91.67	88.00	0.001	0.834-0.988
MiR-382	0.952	0.24	87.50	92.00	0.001	0.849-0.992
8-hydroxy-2′-deoxyguanosine	0.627	3232	79.17	56.00	0.125	0.477-0.761
Cancer antigen 15-3	0.768	235.3	87.50	68.00	0.001	0.626-0.877

MiRNAs: MicroRNAs.

DISCUSSION

The identification of new discriminatory noninvasive biomarkers for use in the diagnosis of different types of cancer is a topic of significant research. Although a plethora of studies on breast cancer diagnosis suggest many biomarkers as diagnostic and/or prognostic biomarkers, there are significant gaps in their sensitivity and specificity during clinical applications[24,25]. For example, the serum tumor marker CA15-3 is the gold-standard biomarker used for breast cancer diagnosis. However, several limitations are hindering the efficacy of this biomarker[26]. Among the many biomarkers suggested as promising liquid biopsies, miRNAs have shown great promise. They exhibit higher expression levels in cancer patients compared to healthy individuals, and they are involved in the development and progression of cancer; thus, their expression reflects the tumor state[16,27,28]. More importantly, miRNAs express a highly stable profile in bodily fluids[29,30].

In the present study, we investigated the expression levels of five miRNAs (miR-21, miR-145, miR-221, miR-382, and miR-1246) in serum samples from breast cancer patients, categorized into grade II and grade III, to assess their diagnostic potential compared with the conventional tumor biomarker CA15-3 and the promising tumor biomarker 8-OHdG[21,31]. Our results revealed distinct miRNA expression patterns between breast cancer patients and healthy controls. Notably, miR-21, miR-221, and miR-1246 Levels were significantly elevated in breast cancer patients. Conversely, miR-145 and miR-382 expression displayed a downregulated trend in patients compared to healthy individuals. Furthermore, significant differences in miRNA expression were observed within the patient groups (grade II vs grade III) for all miRNAs except miR-382. These findings suggest that our selected panel of miRNAs, with the exception of miR-382, might contribute to disease progression within breast cancer patients and act as grading biomarkers.

Our findings underscore the potential application of these miRNAs as reliable biomarkers. These miRNAs have been previously reported to be implicated in cancer pathogenesis. Of particular interest is miR-21, the most highly prevalent “oncomiR” that is up-regulated in most types of cancer, especially breast cancer[32,33]. The miR-21 has been extensively studied as a diagnostic biomarker that can distinguish breast cancer patients at any stage with high sensitivity and specificity[34,35]. The elevated expression level of miR-21 was reported to be associated with the clinicopathological features of breast cancer cases, such as the histological tumor grades, highlighting the value of miR-21 as a marker that can discriminate between grades of breast cancer patients[36,37]. In our study, miR-21 displayed the highest correlation coefficient values with other miRNAs, demonstrating a significant positive correlation with miR-221 or miR-1246, thus strengthening the potential of this panel of miRNAs for diagnosis. Additionally, miR-21 strongly correlated with the protein biomarkers CA15-3 and 8-OHdG, further supporting its potential as a molecular marker to be combined with the conventional protein biomarkers. Interestingly, according to our findings, miR-21 revealed a higher diagnostic performance in discriminating between subgroups of breast cancer patients (AUC = 0.997) than the conventional tumor biomarker CA15-3 (AUC = 0.768) or the novel biomarker 8-OHdG (AUC = 0.627), revealing the promising use of miR-21 as a grade-specific diagnostic and biomarker.

The same diagnostic behavior of miR-21 was also detected for miR-221 and miR-1246. Previous studies have indicated that elevated miR-221 expression is linked to tumor size and histological tumor grades in various cancers, such as colorectal cancer and pancreatic cancer, gastric cancer, and others[38]. The expression of miR-221 was significantly elevated in high-grade gliomas than in low-grade gliomas. Higher expression levels of miR-221 were positively correlated with the degree of glioma infiltration, which is linked to lower overall survival[39]. In breast cancer, circulating miR-221 was associated with pathological grade and the occurrence of distant metastases in breast cancer, suggesting its potential involvement in poor prognosis[40,41]. The diagnostic power of miR-1246 was the focus of several investigations searching for the optimum cancer diagnostic biomarkers[42-44]. A recent meta-analysis of 29 individual studies from 9 countries, covering 12 cancer types and over 5900 samples, concluded that circulating miR-1246 has good sensitivity, specificity, and robust performance in cancer screening[45]. The diagnostic power of miR-1246 for breast cancer patients was exceptional, with an AUC of 0.950 and a diagnostic odds ratio of 98.5, surpassing its performance in other cancer types[45]. These data imply that circulating miR-1246 outperforms some presently utilized breast cancer tumor biomarkers. Moreover, miR-1246 was considered as a staging biomarker, where measuring the level of exosomal miR-1246 in the serum of patients with gastric cancer was able to discriminate patients with stage I gastric cancer from benign patients or healthy controls[46]. According to our results, miR-1246 was reported to be the second most significant miRNA, following miR-21, in terms of correlation significance with other markers and diagnostic power in differentiating between breast cancer patients and healthy individuals. Additionally, miR-1246 showed the highest diagnostic power in the discrimination between the subgroups of breast cancer patients (AUC = 1.00), exceeding the other miRNAs and the tumor biomarkers CA15-3 and 8-OHdG.

In the present study, the reduced expression of miR-145 and miR-382 in breast cancer patients compared to the control has also been reported in previous studies, implying that these two miRNAs function as tumor suppressors and that their decreased expression may contribute to breast cancer pathogenesis. MiR-145 is typically downregulated in breast cancer, with lower expression levels often correlating with higher tumor grades and more aggressive phenotypes. Low expression of miR-145 has been associated with lymph node metastasis and tumor size, which are typically poor prognostic factors for breast cancer[47-49]. Therefore, the diagnostic potential of miR-145 in breast cancer emerges from its capacity to distinguish between different grades of breast cancer patients, making it a viable marker for early detection and disease monitoring. This was detected in our study, where the expression of miR-145 significantly varied between grade II and grade III breast cancer patients. Regarding miR-382, we detected a significant downregulation compared to the control group. However, the diagnostic capability of this miRNA was not significant in differentiating between the various grades of breast cancer patients.

In our study, to facilitate a comprehensive analysis, we opted to evaluate the levels of two protein biomarkers alongside our chosen panel of miRNAs. We included CA15-3, a clinically established biomarker with a proven track record in breast cancer monitoring[18]. We also measured the level of 8-OHdG, a recently emerged marker demonstrating promising potential as a novel diagnostic and prognostic tool for breast cancer. The 8-OHdG serves as a biomarker for oxidative DNA damage, and measuring its levels in serum offers a non-invasive way to evaluate the oxidative stress burden in patients. This provides potential utility as both a diagnostic and prognostic biomarker[21,50]. The 8-OHdG has been found to be a promising discriminatory biomarker in the early diagnosis of breast cancer. Moreover, 8-OHdG expression has been associated with clinical stage, tumor size, and postoperative survival in various cancer types[22,51,52]. Although our findings showed that both markers CA15-3 and 8-OHdG differed significantly between patient samples and controls, only CA15-3 showed significant differences across patients with various stages, in addition to a significant diagnostic performance discriminating grade II and grade III breast cancer patients (AUC = 0.768, P > 0.001). Although the diagnostic performance of CA15-3 was more prominent than 8-OHdG, our selected panel of miRNAs exceeds both CA15-3 and 8-OHdG as diagnostic or grade-specific biomarkers.

The findings of our study also underscore the efficacy of combined biomarkers. For differentiating breast cancer patients (grade II) from healthy individuals, the highest sensitivity and specificity were achieved by combining miR-145 with 8-OHdG or with a panel of miR-382 and miR-1246, suggesting a synergistic effect between these markers, potentially leading to a more robust identification of grade II breast cancer patients. Interestingly, the optimal combination for discriminating grade III patients shifted towards miRNAs alone, with miR-382 and miR-221 demonstrating the most robust diagnostic performance. This finding suggests a potential shift in the optimal biomarker panel depending on the breast cancer grade, highlighting the emerging need for grade-specific biomarker panels for breast cancer diagnosis.

Emerging evidence suggests that a multi-analyte panel of biomarkers can improve diagnostic sensitivity by capturing signals from different pathological processes or cellular origins[53-55]. Combining circulating miRNAs with conventional protein or oxidative stress biomarkers could improve diagnostic accuracy by capturing additional aspects of tumor biology[16,56]. For instance, miR-21 is known to regulate oncogenic pathways such as the phosphatase and tensin homolog deleted on chromosome 10/phosphatidylinositol 3-kinase/protein kinase B axis, contributing to cell proliferation, invasion, and chemoresistance in breast cancer[57]. Similarly, miR-1246 is implicated in the p53 pathway[58] and Wnt/β-Catenin Pathway[59], often reflecting tumor aggressiveness and metastatic potential. On the other hand, CA15-3 reflects tumor burden[17], and 8-OHdG serves as a marker of oxidative DNA damage and indicates elevated cellular stress levels often associated with carcinogenesis, which is elevated in malignancies with high metabolic activity[19,21]. The integration of molecular (miRNAs) and protein/oxidative biomarkers thus offers a multi-dimensional view of the disease, improving sensitivity and specificity over single-marker approaches. In our study, this combined approach yielded improved diagnostic accuracy, suggesting its promise in clinical applications for non-invasive detection and tumor grade classification in breast cancer patients.

While our study focused on the diagnostic utility of a chosen panel of circulating biomarkers for identifying breast cancer patients and differentiating between tumor grades, we acknowledge that comprehensive clinical management of breast cancer requires further stratification based on several factors. These include clinical stage (TNM classification), molecular subtype (such as Luminal A/B, HER2-enriched, and Triple-negative), and predicted risk of recurrence and overall prognosis[60]. This kind of categorization would provide more accurate biomarker-based classification systems by aiding in identifying whether miRNA signatures differ across various molecular and anatomical disease settings[61]. Building on our current findings on grading, a critical direction for future research is to thoroughly evaluate this panel of circulating biomarkers for their ability to correlate with different TNM stages, accurately classify molecular subtypes, and assess their predictive power for recurrence risk and overall survival to enhance the clinical value of these circulating biomarkers. Additionally, as the field progresses, current research strongly emphasizes broader approaches for biomarker discovery. High-throughput 'omics' technologies, such as comprehensive circulating miRNAs sequencing, proteomics, and metabolomics applied to larger patient cohorts, offer the potential to identify entirely new candidate biomarkers or panels that could provide deeper insights into tumor biology and progression[62]. Furthermore, developing predictive models with machine learning algorithms could analyze the complex relationships between multiple markers more effectively, achieving optimal diagnostic accuracy and classifying patients into clinically relevant subgroups beyond simple grading, or even predicting treatment response[63].

CONCLUSION

Our study offers a comprehensive analysis of a panel of circulating miRNAs and protein tumor biomarkers to demonstrate their significant potential, either individually or in combination, as diagnostic, non-invasive tools in breast cancer. The distinct expression patterns and the high sensitivity and specificity of miR-21, miR-221, miR-1246, miR-145, and miR-382, particularly when combined with CA15-3 and 8-OHdG, provide promising avenues for early detection and tumor grading. The ability to classify patients based on tumor grade using circulating markers allows for more personalized therapeutic strategies. Therefore, miRNA-based grading could guide clinical decisions regarding treatment, especially with higher-grade disease. Additionally, the dynamic and non-invasive nature of circulating miRNAs supports their use as a liquid biopsy tool for monitoring treatment response or detecting early relapse. Our findings strongly support the development of multi-analyte liquid biopsy approaches in the diagnostic process and personalized management of breast cancer patients.