Discovery of potential serum and urine-based microRNA as minimally-invasive biomarkers for breast and gynecological cancer

Abstract

BACKGROUND:

Deregulated microRNAs (miRNAs) in breast and gynecological cancer might contribute to improve early detection of female malignancies.

OBJECTIVE:

Specification of miRNA types in serum and urine as minimally-invasive biomarkers for breast (BC), endometrial (EC) and ovarian cancer (OC).

METHODS:

In a discovery phase, serum and urine samples from 17 BC, five EC and five OC patients vs. ten healthy controls (CTRL) were analyzed with Agilent human miRNA microarray chip. Selected miRNA types were further investigated by RT-qPCR in serum (31 BC, 13 EC, 15 OC patients, 32 CTRL) and urine (25 BC, 10 EC, 10 OC patients, 30 CTRL) applying two-sample t-tests.

RESULTS:

Several miRNA biomarker candidates exhibited diagnostic features due to distinctive expression levels (serum: 26; urine: 22). Among these, miR-518b, -4719 and -6757-3p were found specifically deregulated in BC serum. Four, non-entity-specific, novel biomarker candidates with unknown functional roles were identified in urine (miR-3973; -4426; -5089-5p and -6841). RT-qPCR identified miR-484/-23a (all $p\leqslant$ 0.001) in serum as potential diagnostic markers for EC and OC while miR-23a may also serve as an endogenous control in BC diagnosis.

CONCLUSIONS:

Promising miRNAs as liquid biopsy-based tools in the detection of BC, EC and OC qualified for external validation in larger cohorts.

Keywords

miR-484 miR-23a-3p liquid biopsy miRNA endometrial cancer ovarian cancer breast cancer screening

1. Introduction

Among all cancer types in women, breast and gynecological cancers classify as some of the most frequent malignancies. According to an assessment of the World Health Organization (WHO) in 2018, the incidence of breast cancer (BC) in women is 2,088,849 with 626,679 BC related deaths annually worldwide [110]. Endometrial cancer (EC) is the sixth most common cancer in women with 382,069 women being diagnosed every year [110]. However, most EC subtypes show relatively good prognosis rates. Mortality accounts for 89,929 women each year allocating EC the 14 ${}^{\rm th}$ rank in cancer-related deaths among women in the whole world [110]. Ovarian cancer (OC) claims the eighth rank in women’s malignancies with an incidence of 295,414 new cases per year [110]. Late diagnosis in advanced stages due to asymptomatic onset leads to a high yearly death rate of 184,799 women globally [110].

These still high numbers of cancer-related deaths caused by BC, EC and OC illustrate the unmet demand for sufficient cancer management. Facing the limitations in anti-cancer therapy, early detection of circumscribed malignant changes could represent one of the most compelling efforts to improve clinically implemented diagnosis and treatment strategies to significantly reduce BC, EC and OC-related mortality as well as morbidity. However, to date solely BC guidelines provide elaborate screening recommendations. A vast lack of evidence persists for sufficient and non- or minimally-invasive standard methods for early and subtype-specific EC and OC detection [7, 25, 81, 101, 109].

Mammography forms the centerpiece of BC screening. There is a consensus not to screen women under the age of 40, whereas the details of international screening programs vary [83]. A meta-analysis conducted by Nelson et al. attested an overall benefit of BC screening for average and high-risk women, in particular at the age of 60–69 with 21.3 prevented deaths in 10,000 patients (CI, 5.9 to 19.5) [77]. However, Miller et al. did not find a significant benefit by mammographic BC screening in women between 40–59 [71]. In summary, the value of mammographic screening remains questionable due to variations in diagnostic features caused by practitioner-based interpretations, individual breast tissue densities and false positive findings leading to over-diagnosis and subsequent unnecessary biopsies [83]. Hence, further, ideally non-invasive, screening means should be introduced aiming to narrow the gaping benefit-harm ratio of BC screening.

In contrast to the known positive impact of BC screening on successful therapy, EC routine screening studies did not reveal beneficial effects in asymptomatic women [10, 23, 73, 99]. The diagnostic features of endometrial malignancies comprise initial transvaginal ultra-sonography and subsequent endometrial biopsy or hysteroscopic sampling of the endometrium [89]. These biopsy methods represent an invasive procedure that needs general anesthesia and carries both, surgical and anesthetic risks. Laboratory biomarkers facilitating the detection of EC are not implemented in routine clinical EC management to date [10].

The screening for OC does not only comprise imaging-based procedures but also serum biomarkers, namely cancer antigen 125 (CA 125). However, German guidelines [24] as well as the U.S. Preventive Services Task Force recommendations [74] negate a significant benefit on mortality and morbidity by the performance of both screening methods neither in average nor in high risk groups [70, 119]. The significance of CA 125 as biomarker in OC is not completely understood with serum CA 125 being significantly elevated in only 50% of early stage OC [4, 119]. Moreover, varying individual expression levels of CA 125 complicate the definition of an appropriate threshold for OC detection. In spite of its limitations as a biomarker, CA 125 still outplaces upcoming biomarkers like human epididymis protein 4 (HE4) and mesothelin in terms of discriminatory potential [4, 41]. The discussed guidelines and recommendations demonstrate the lack of sufficient screening methods that provide satisfying sensitivity, easy accessibility and acceptable complication rates.

Table 1
Characteristics of BC patients stratified by tumor molecular subtype and healthy controls

	Number of patients: validation; (discovery phase)
BC subtype	Lum A*				Lum B Her2-*				Her2+*				TNBC*
Cohort	Serum		Urine		Serum		Urine		Serum		Urine		Serum		Urine
Total number	9	(5)	8	(5)	11	(5)	12	(5)	6	(2)	3	(2)	5	(5)	2	(5)
Age (years)	64	(65)	64	(65)	61	(64)	56	(65)	60	(63)	76	(63)	55	(63)	65	(63)
Histological subtype
Invasive ductal carcinoma	6	(3)	8	(3)	7	(5)	11	(5)	4	(2)	1	(2)	5	(5)	2	(5)
Invasive lobular carcinoma	0	(2)	0	(2)	4	(0)	1	(0)	0	(0)	1	(0)	0	(0)	0	(0)
Other	3	(0)	0	(0)	0	(0)	0	(0)	2	(0)	1	(0)	0	(0)	0	(0)
Initial tumour stage (T)
pT1	9	(3)	8	(3)	6	(5)	9	(5)	1	(1)	1	(1)	2	(3)	1	(3)
pT2	0	(2)	0	(2)	5	(0)	3	(0)	5	(1)	2	(1)	2	(1)	1	(1)
pT3	0	(0)	0	(0)	0	(0)	0	(0)	0	(0)	0	(0)	0	(1)	0	(1)
pT4	0	(0)	0	(0)	0	(0)	0	(0)	0	(0)	0	(0)	1	(0)	0	(0)
Initial nodal status (N)
pN0	9	(5)	7	(5)	7	(5)	10	(5)	3	(2)	2	(2)	4	(3)	2	(3)
pN1	0	(0)	1	(0)	1	(0)	2	(0)	2	(0)	1	(0)	0	(2)	0	(2)
pN2	0	(0)	0	(0)	2	(0)	0	(0)	0	(0)	0	(0)	1	(0)	0	(0)
pN3	0	(0)	0	(0)	1	(0)	0	(0)	1	(0)	0	(0)	0	(0)	0	(0)
Initial tumour grade (G)
G1	5	(1)	5	(1)	0	(5)	1	(5)	0	(0)	3	(0)	0	(0)	0	(0)
G2	4	(4)	3	(4)	9	(0)	10	(0)	4	(1)	0	(1)	4	(1)	1	(1)
G3	0	(0)	0	(0)	2	(0)	1	(0)	2	(1)	0	(1)	1	(4)	1	(4)
Lymphovascular invasion (L)
L0	9	(5)	7	(5)	7	(5)	10	(5)	2	(2)	2	(2)	1	(3)	2	(3)
L1	0	(0)	1	(0)	4	(0)	2	(0)	2	(0)	0	(0)	2	(1)	0	(1)
n.a.	0	(0)	0	(0)	0	(0)	0	(0)	2	(0)	1	(0)	2	(1)	0	(1)
ER (%)	82	(82)	81	(82)	83	(93)	93	(93)	59	(15)	95	(15)	–		–
PR (%)	68	(77)	59	(77)	51	(57)	75	(57)	47	(0)	87	(0)	–		–
Proliferation rate (Mib) (%)	8	(10)	8	(10)	32	(26)	28	(26)	47	(45)	31	(45)	51	(54)	30	(54)
Menopausal status
Premenopausal	0	(0)	0	(0)	2	(0)	3	(0)	0	(1)	0	(1)	1	(1)	2	(1)
Postmenopausal	8	(5)	6	(5)	9	(5)	6	(5)	4	(1)	3	(1)	3	(4)	0	(4)
n.a.	1	(0)	2	(0)	0	(0)	3	(0)	2	(0)	0	(0)	1	(0)	0	(0)
Healthy controls	CTRL				CTRL U 40				CTRL O 40
Cohort	serum		urine		serum $+$ urine				serum $+$ urine
Total number	32		30		(5)				(5)
Age (years)	51		48		(27)				(54)

BC: breast cancer; Lum A: luminal A molecular subtype; Lum B: luminal B molecular subtype; Her2+: Her2 enriched molecular subtype; TNBC: triple negative molecular subtype; * defined by St Gallen Consensus 2012; n.a.: no information available; ER: estrogen receptor expression; PR: progesterone receptor expression; CTRL U40: healthy women at the age of $\leqslant$ 40 years; CTRL O40: healthy women at the age of $>$ 40 years. All patients showed no metastasis at the time of diagnosis (M0).

Being involved in overall physiological and pathophysiological pathways, microRNAs (miRNAs) are crucial regulators of carcinogenic processes [1, 14, 88]. In that context they can either act as tumor suppressors or exert oncogenic functions [120]. Growing evidence underlines the involvement of miRNAs in BC [21, 59, 98], EC [43, 104]and OC [9, 22, 63, 90]. Furthermore, both circulating miRNAs in the blood stream as well as secreted miRNAs in urine might serve as early biomarkers in screening, diagnostics and therapy response evaluation of various tumor entities [30]. The promising potential of miRNAs as novel, minimally-invasive biomarkers for the detection of BC has been shown by an increasing number of studies [6, 112, 124]. Before unraveling the auspicious potential of using urine as a diagnostic matrix for detecting BC specific miRNA patterns, the feasibility of reliably detecting miRNAs in urine first needed to be determined. A pilot study of Erbes et al. identified four urinary miRNAs that could distinguish between BC patients and healthy controls [29]. Hence, miRNA signatures might serve as innovative biomarkers for BC diagnosis. miRNA expression pattern might even serve as a prospective tool for classifying BC into molecular subtypes and as useful parameters in tumor grading [49]. In gynecological tumors such as EC and OC, miRNA patterns were found to be altered in different body fluids, respectively [32, 72]. However, evidence of miRNA dysregulation in cancer is conflicting and depends to a large degree on isolation methods, applied analytic approach and particularly on the identification of stable endogenous controls in definite settings.

Table 2

Characteristics of EC and OC patients

	Number of patients: validation; (discovery phase)
Tumor entity	EC serum		EC urine		OC serum		OC urine
Total number	13	(5)	10	(5)	15	(5)*	10	(5)*
Age (years)	58	(66)	62	(66)	63	(61)	68	(61)
Histological subtype
Endometrioid adenocarcinoma	11	(3)	8	(3)	1	(0)	0	(0)
Serous adenocarcinoma	1	(1)	1	(1)	11	(4)	8	(4)
Mucinous carcinoma			1	(0)	0	(0)
Other	1	(1)	1	(1)	2	(0)	2	(0)
Initial tumour stage (T)
pT1	9	(5)	7	(5)	5	(0)	2	(0)
pT2	2	(0)	2	(0)	1	(0)	1	(0)
pT3	2	(0)	1	(0)	9	(4)	7	(4)
Initial nodal status (N)
pN0	13	(5)	10	(5)	8	(2)	4	(2)
pN1	0	(0)	0	(0)	7	(1)	5	(1)
n.a.					0	(0)	1	(0)
Initial tumour grade (G)
G1	6	(1)	3	(1)	0	(0)	0	(0)
G2	4	(1)	4	(1)	2	(0)	0	(0)
G3	3	(3)	3	(3)	12	(4)	10	(4)
n.a.					1	(0)	0	(0)
FIGO staging
I	9	(4)	7	(4)	4	(0)	1	(0)
II	2	(0)	2	(0)	1	(0)	1	(0)
III	2	(1)	1	(1)	7	(3)	4	(3)
IV	0	(0)	0	(0)	3	(1)	3	(1)
n.a.					0	(0)	1	(0)
Lymphovascular invasion (L)
L0	13	(4)	10	(4)	5	(1)	2	(1)
L1	0	(1)	0	(1)	8	(1)	6	(1)
n.a.	0	(0)	0	(0)	2	(2)	2	(2)
Menopausal status
Premenopausal	0	(0)	0	(0)	1	(0)	0	(0)
Postmenopausal	8	(5)	7	(5)	11	(3)	9	(3)
n.a.	5	(0)	3	(0)	3	(1)	1	(1)

EC: endometrial cancer; OC: ovarian cancer; n.a.: no information available; FIGO: Fédération Internationale de Gynécologie et d’Obstétrique. All patients showed no metastasis at the time of diagnosis (M0). * one patient diagnosed with a benigne ovarian lesion in retrospect.

The present study focuses on the discovery of blood- and urine-based miRNA biomarkers for screening and early detection of BC, EC and OC. Such easily accessible, minimally-invasive biomarkers could be beneficial for both patients and clinicians and could help to reduce health care costs. In the conducted discovery phase, we first investigate altered miRNA expression signatures in BC, EC and OC as reflected in body fluids using a state-of-the-art microarray chip analysis. Thereby, we aim to narrow the number of miRNA types as defined by miRBase V21 [35] to identify potential, robust biomarkers. An additional approach focuses on the identification of stably expressed miRNAs to be established as robust comparative values for quality control and data normalization in a qPCR setting. Subsequently, selected miRNAs identified in the discovery phase are validated via qPCR.

2. Materials and methods

2.1 Patient cohort and sampling procedures

The initial cohort (discovery phase) comprised of 27 patients, the validation cohort comprised of 82 patients undergoing therapy for BC, EC or OC at the University Hospital Freiburg. Ten (discovery phase) and 62 (validation) proven healthy women attending routine screening procedures such as ultrasound, palpation or mammography were recruited as controls. Patients and controls were enrolled from December 2015 to November 2018 after the study was approved by the local Ethics Committee (#386/16; #607/16) and written informed consent was obtained from each participant. BC subtypes were defined according to St. Gallen Consensus 2012 [33]. For patient characteristics see Tables 1 and 2.

Collection of blood and urine samples was performed by physicians prior to treatment (surgery or neoadjuvant chemotherapy) following standard protocols. Whole blood was incubated for 20 min at room temperature allowing coagulation and then centrifuged (10 min, 3500 rpm, 4 ${}^{\circ}$ C). Following centrifugation, serum was visually inspected for any sign of hemolysis to assure high sample quality, transferred to a sterile 15 ml test tube and 200 $\mu$ l of serum were aliquoted. Hemolytic samples were excluded from further processing. Urine was centrifuged (15 minutes, 3500 rpm, 4 ${}^{\circ}$ C) and cell-free supernatant was transferred to 15 ml test tubes, respectively. Serum and urine specimen were stored at – 20 ${}^{\circ}$ C until further processing.

2.2 RNA isolation

Total RNA was isolated from 200 $\mu$ l serum samples employing Norgen Total RNA Purification Kit #17200 (Norgen Biotek Corp., Thorold, Canada) according to the manufacturers’ protocol. 15 ml of urine were applied for isolation of microvesicles by filtration (Amicon Ultra centrifugal filter units, MWCO 100 kDa, Sigma-Aldrich ${}^{\@setsize{\scriptsize}{8pt}{\viipt}{\@viipt}\textregistered}$ , Taufkirchen, Germany) and RNA was obtained using the above mentioned kit. RNA concentration was assessed by NanoDrop ${}^{\rm TM}$ ND-1000 Spectrophotometer (PEQLAB, Erlangen, Germany). For RT-qPCR validation, 5 fmol of synthetic ath-miR-159a and cel-miR-39 were spiked into the samples prior to RNA isolation.

2.3 miRNA microarray chip analysis

For miRNA microarray chip analysis, RNA from both, serum and urine samples, was pooled by tumor entity. The EC and OC pool consisted of five patients each. The entirety of the BC group contained 17 patients diagnosed with definite BC molecular subtypes. While the BC group was formed by five luminal A (Lum A), luminal B (Lum B) and triple-negative BC (TNBC) patients, respectively, it contained only two patients with Her2 enriched (Her2+) molecular subtype. BC samples were applied in four pools grouped by molecular subtype. Comprising five healthy women each, two control pools (CTRL) were set up separated by age ( $\leqslant$ vs. $>$ 40 years). The pools were generated by merging equal shares of isolated RNA.

The following steps were conferred and processed by OakLabs (OakLabs GmbH, Hennigsdorf, Germany) [80]. From each pool, 100 ng of total RNA was compiled. Human miRNA microarray chip analysis (Agilent-070156 Human miRNA V21.0 Microarray 046064, 8 $\times$ 60K [2]) was performed, whereas quality control of samples was assessed with 2100 Bioanalyzer (Agilent Technologies, Santa Clara, USA) using RNA 6000 Pico Kit. Quantity control was performed via photometrical measurement on Nanodrop ${}^{\rm TM}$ 2000 spectrophotometer and evaluated by RNA integrity number (RIN) algorithm. Expression levels of 2549 human miRNAs specified by the latest version of miRBase database (Release 21) [35] were measured.

Table 3
Deregulated miRNAs in serum of breast and gynecological cancer patients discovered by microarray chip analysis. Log ${}_{2}$ (FC) and $p$ -values

miRNA	log ${}_{2}$ (FC)* (T group)	$p$ -value**	log ${}_{2}$ (FC)*** (BC group)	$p$ -value**	log ${}_{2}$ (FC)**** (EC/OC group)	$p$ -value**
miR-143-3p	$-$ 2.53	0.0008	$-$ 2.53	0.0278
miR-186-3p					3.66	0.0286
miR-195-3p					3.74	0.0016
miR-205-3p	$-$ 1.50	0.0002	$-$ 1.50	0.0181
miR-451b					$-$ 1.97	0.0102
miR-484					$-$ 0.94	0.0102
miR-500a-3p					2.70	0.0102
miR-518b			$-$ 3.10	0.0276
miR-520f-5p	$-$ 2.39	0.0003	$-$ 2.39	0.0199
miR-548ao-3p	$-$ 1.93	0.0011	$-$ 1.93	0.0322
miR-670-5p					3.56	0.0340
miR-770-5p	$-$ 2.34	0.0003	$-$ 2.34	0.0199
miR-888-3p	$-$ 3.29	$<$ 0.0001	$-$ 3.29	0.0001	$-$ 3.29	0.0446
miR-939-3p					1.96	0.0443
miR-1272	$-$ 1.46	0.0011	$-$ 1.46	0.0322
miR-1468-5p	$-$ 2.59	0.0322	$-$ 2.82	0.0266
miR-1972					1.67	0.0102
miR-3153	$-$ 2.62	$<$ 0.0001	$-$ 2.62	$<$ 0.0001	$-$ 2.62	0.0011
miR-3610	0.86	0.0069	0.81	0.0322
miR-3921	$-$ 2.09	0.0003	$-$ 2.09	0.0199
miR-4640-3p	$-$ 2.04	0.0020	$-$ 2.14	0.0001
miR-4719			$-$ 1.20	0.0199
miR-6722-3p	1.14	0.0499
miR-6757-3p			$-$ 3.85	0.0373
miR-6859-5p	3.60	0.0289	3.54	0.0261
miR-7844-5p					2.87	0.0102

FC: foldchange; T: tumor (BC $+$ EC $+$ OC); BC: breast cancer; EC: endometrial cancer; OC: ovarian cancer; CTRL: healthy controls. * FC $=$ T/CTRL; *** FC $=$ BC/CTRL; **** FC $=$ (E/O)/CTRL. ** Benjamini-Hochberg-adjusted $p$ -value.

2.4 Realtime RT-PCR (RT-qPCR)

Applying RT-qPCR, miR-484 (serum) and -205-5p (urine) were further analyzed for their diagnostic potential for EC and OC and miR-10b-5p (urine) as a non-specific biomarker for BC, EC and OC. Additionally, miR-23a-3p was investigated for its housekeeping potential. Poly(A) tailing-based RT was performed in a reaction mix containing 4 $\mu$ l RT-buffer (5x), 1 $\mu$ l 5 $\mu$ M poly A adapterprimer (5’-CAGGTCCAGTTTTTTTTTTTTTTTVN-3’), 1 $\mu$ l 5 mM dNTPs (Jena Bioscience, Germany), 0.25 $\mu$ l Maxima ${}^{\rm TM}$ reverse transcriptase (Thermo), 0.25 $\mu$ l SUPERase In ${}^{\rm TM}$ RNase inhibitor (Thermo), 0.5 $\mu$ l 10 mM ATP (New England Biolabs ${}^{\@setsize{\scriptsize}{8pt}{\viipt}{\@viipt}{\textregistered}}$ Inc., Frankfurt Germany), 0.25 $\mu$ l poly A polymerase (New England Biolabs ${}^{\@setsize{\scriptsize}{8pt}{\viipt}{\@viipt}{\textregistered}}$ Inc.) and 5 $\mu$ l of RNA sample. The reaction was carried out at 42 ${}^{\circ}$ C on a thermal cycler (Eppendorf, Hamburg, Germany) and was stopped after 30 min with 10 min incubation at 85 ${}^{\circ}$ C.

Figure 1.

Workflow of data processing: Agilent-070156 Human miRNA V21.0 Microarray (046064, 8 x 60K) was performed assessing 2549 miRNAs. Dark grey blocks represent serum data set, light grey blocks represent urine data set. Data processing includes the removal of miRNAs at the level of detection (LOD $=$ 0.1), restriction to miRNAs stable in the combined control (CTRL) and the testing of three hypotheses by Welch’s two-tailed $t$ -test. T: tumor (T $=$ BC $+$ EC $+$ OC); BC: breast cancer; EC $=$ endometrial cancer; OC: ovarian cancer. Furthermore, normalization candidates (N) were assessed. Venn-diagrams indicate number of statistically significant miRNA biomarker candidates in serum (left) and urine (right) specimen. An overlap between groups can be observed. miR-484; -23a-3p; -10b-5p and -205-5p were chosen for RT-qPCR validation in the particular matrix and context.

1 $\mu$ l of cDNA (5 ng/ $\mu$ l) was applied in a duplicate analysis on LightCycler ${}^{\@setsize{\scriptsize}{8pt}{\viipt}{\@viipt}{\textregistered}}$ 480 Instrument II (Roche, Basel, Switzerland) with pre-incubation at 95 ${}^{\circ}$ C for 2 min and a two-step amplification of 40 cycles at 95 ${}^{\circ}$ C for 5 sec and 60 ${}^{\circ}$ C for 30 sec. The master mix contained 1 $\mu$ l qPCR buffer (10x), 0.5 $\mu$ l 5 mM dNTPs (Jena Bioscience), 0.5 $\mu$ l Sybr ${}^{\@setsize{\scriptsize}{8pt}{\viipt}{\@viipt}{\textregistered}}$ Green (#PCR-378, Jena Bioscience), 6.45 $\mu$ l nuclease-free water (Analytic Jena, Jena, Germany), 0.05 $\mu$ l HotStart Taq polymerase (Jena Bioscience) and 0.5 $\mu$ l primer pair mix. Specific sense (s) and antisense (as) primers were designed applying Peter K Busk’s software “miRprimer” [11]: miR-484: 5’-GCAGTCAGGCTCAGTC-3’ (s), 5’-GGTCCAGTTTTTTTTTTTTTTTATCG-3’ (as); miR-23a-3p: 5’-AGATCACATTGCCAGGGAT-3’ (s), 5’-GGTCCAGTTTTTTTTTTTTTTTGGAA-3’ (as); miR-10b: 5’-CAGTACCCTGTAGAACCGAAT-3’ (s), 5’-GGTCCAGTTTTTTTTTTTTTTTCAC-3’ (as); miR-205-5p: 5’-CCTTCATTCCACCGGAGT-3’ (s), 5’-GGTCCAGTTTTTTTTTTTTTTTCAGA-3’ (as). Relative quantification was performed according to Marabita et al. [67] monitoring recovery efficiency with help of the spike-in miRNAs ath-miR-159a and cel-miR-39 and by normalizing to the geometric mean of miR-16-5p and -26b-5p. Ath-miR159a: 5’-GCGCA GTTTGGATTGAAG-3’ (s), 5’-AGGTCCAGTTTTTT TTTTTTTTTAGAG-3’ (as); cel-miR-39-3p: 5’-GTCA CCGGGTGTAAATCAG-3’ (s), 5’-GGTCCAGTTTT TTTTTTTTTTTCAAG-3’ (as); miR-16-5p: 5’-CGCA GTAGCAGCACGTA-3’ (s), 5’-CAGTTTTTTTTTTT TTTTCGCCAA-3’ (as); miR-26b-5p: 5’-CGCAGTTC AAGTAATTCAGGAT-3’ (s), 5’-GGTCCAGTTTTTT TTTTTTTTTACCT-3’ (as).

2.5 Data processing and statistical analysis

Expression levels of 2549 miRNAs in total (miRBase Release 21 [35]) were assessed by microarray chip analysis. In a first essential step, raw data obtained by OakLabs was processed by excluding miRNAs that exclusively exhibit expression values of 0.1 (assay limit of detection, LOD) in all investigated pools. An expression value of 0.1 can represent either a miRNA that is not present in the investigated specimen or one that shows a rather weak expression and is therefore not suitable to be assessed reliably by this analysis. This analytical approach reduced the serum data set to 1951 miRNAs and the urine data set to 2136 miRNAs, respectively. Raw data of the remaining targets was subsequently quantile normalized and log ${}_{2}$ -transformed to achieve an approximate normal distribution. The transformed data set was further restricted by combination of control pools. In order to investigate miRNA expression in an age-independent manner, the two control pools were merged. Only miRNAs which at most exhibited a doubling or halving of the mean expression levels between the two control pools were further investigated. This resulted in a total number of 1350 miRNAs in serum and 1391 miRNAs in urine analyses. Welch’s two-tailed $t$ -tests were applied for the remaining targets. Three hypotheses were tested, an overall tumor group (BC Lum A $+$ BC Lum B $+$ BC Her2+ $+$ TNBC $+$ EC $+$ OC), a BC group (Lum A $+$ Lum B $+$ Her2+ $+$ TNBC) and a gynecological tumor group (EC $+$ OC) were tested against a combined control group (CTRL U 40 and CTRL O 40). For workflow of data processing see Fig. 1. Obtained $p$ -values were adjusted for multiple testing using Benjamini-Hochberg correction [115, 116] on 1350 (serum) and 1390 (urine) independent tests. A result is considered significant if the corresponding Benjamini-Hochberg-adjusted p-value is smaller than 0.05 thereby controlling the false discovery rate (FDR) at 0.05. To validate results, the two-sample t-test under the assumption of equal variances was repeated on individual-level measurements. For normalization candidates we additionally calculated separate $t$ -tests for EC and OC against the combined control group. Statistical significance was defined as $p$ -value $<$ 0.017 (0.05/3), which corresponds to a Bonferroni-correction. If at least one of the null hypotheses was rejected a normalization candidate was dropped.

Figure 2.

miR-484 expression levels in serum of breast (BC), endometrial (EC) and ovarian cancer (OC) patients vs. healthy controls (CTRL). (A) Expression levels of miR-484 were significantly elevated in combined EC and OC vs. CTRL ( $p<$ 0.001; two-sample t-test, significance level: 0.017). (B) miR-484 was reliably detected by RT-qPCR in 81.3% (26/32) of CTRL; 78.1% (25/32) of BC; 69.2% (9/13) of EC and 80.0% (12/15) of OC serum samples, respectively. Horizontal lines mark pairwise tumor vs. CTRL comparisons with $p<$ 0.05.

3. Results

3.1 Discovery phase: miRNA biomarker candidates in serum

In serum, 14 miRNAs were discovered by microarray analysis that discriminated between cancer patients and healthy women. Furthermore, the expression levels of 16 miRNAs were altered in the BC group vs. CTRL, eleven in the EC $+$ OC group, respectively (Table 3). The three performed tests disclosed overlaps within the identified miRNAs. Of the potential biomarker candidates, miR-484 was further investigated by RT-qPCR in serum of an independent cohort of 31 BC, 13 EC, 15 OC patients and 32 CTRLs due to a reliable detectability in these samples and as a promising novel liquid biopsy marker candidate for EC and/or OC diagnosis. For miRNA sequences and expression data of significantly regulated miRNAs see supplementary Table S1.

Table 4
Deregulated miRNAs in urine of breast and gynecological cancer patients discovered by microarray chip analysis. Log ${}_{2}$ (FC) and $p$ -values

miRNA	log ${}_{2}$ (FC)* (T group)	$p$ -value**	log ${}_{2}$ (FC)*** (BC group)	$p$ -value**	log ${}_{2}$ (FC)**** (EC/OC group)	$p$ -value**
miR-10b-5b	$-$ 1.52	0.0367
miR-132-5p	$-$ 2.25	0.0001	$-$ 2.25	0.0088
miR-205-5p					$-$ 4.61	0.0377
miR-382-5p	$-$ 1.95	0.0001	$-$ 1.93	0.0115
miR-490-3p	$-$ 1.91	0.0024	$-$ 1.99	0.0079
miR-509-3p					5.04	0.0256
miR-510-3p					2.63	0.0018
miR-526b-3p			$-$ 1.97	0.0478
miR-580-3p					2.27	0.0018
miR-885-5p					$-$ 1.47	0.0324
miR-1206					2.82	0.0187
miR-3529-5p					1.30	0.0256
miR-3591-5p	$-$ 1.12	$<$ 0.0001	$-$ 1.12	$<$ 0.0001	$-$ 1.12	0.0047
miR-3973	$-$ 1.19	0.0001	$-$ 1.19	0.0088
miR-4261					5.17	0.0417
miR-4426	$-$ 1.97	0.0161
miR-4773	$-$ 3.95	0.0001	$-$ 3.95	0.0088
miR-4796-5p					2.74	0.0187
miR-5089-5p	$-$ 1.39	0.0461
miR-5582-5p	$-$ 3.25	0.0044
miR-6825-5p					2.73	0.0377
miR-6841-3p					$-$ 2.69	0.0187

Figure 3.

miR-10b-5p and -205-5p expression levels in urine of 25 breast (BC), 10 endometrial (EC) and 10 ovarian cancer (OC) patients vs. 30 healthy controls (CTRL). (A) No expression level alterations of miR-10b-5p between CTRL and tumor (T). T $=$ BC $+$ EC $+$ OC. (C) Trend of reduced expression levels of miR-205-5p in combined EC and OC vs. CTRL ( $p=$ 0.033; two-sample $t$ -test, significance level: 0.017). (B/D) Horizontal lines mark pairwise tumor vs. CTRL comparisons with $p<$ 0.05. RT-qPCR.

3.2 Discovery phase: miRNA biomarker candidates in urine

Statistical testing of 1391 miRNAs expressed in urine revealed ten miRNAs with tumor-specific attributes. While seven miRNAs discriminated BC from CTRL, only miR-526b-3p did not show overlapping traits with the overall tumor group. In the gynecological tumor group (EC $+$ OC) twelve miRNAs were significantly deregulated (Table 4). MiR-3591-5p, which was identified in all three tests, appears to be a more general, non-tumor entity-specific biomarker for cancer. Since miR-10b-5p and -205-5p were reliably detectable in all urine samples tested by RT-qPCR, these were selected for validation in an independent cohort of 25 BC, 10 EC, 10 OC patients and 30 CTRLs. For miRNA sequences and expression data of significantly regulated miRNAs see supplementary Table S2. Figure 1 illustrates all statistically relevant results as described in the section above. Between serum and urine specimens, no overlaps of deregulated miRNAs in BC, EC and OC have been observed.

3.3 Discovery phase: Promising miRNAs for normalization

Stably expressed miRNAs are of great benefit for sample quality assessment and for data normalization when a methodological approach is restricted to a relative expression analysis (e.g. in RT-qPCR). In the present study, a miRNA type was considered a normalization candidate if foldchanges (FC) between all applied pools did not exceed a value of 2. In addition, a cut-off for quantile normalized expression values was set at a threshold of 0.2, which equals twice the LOD of the applied measurement technique.

Data analysis revealed 60 miRNA normalization candidates in the serum set and 67 in the urine set (Supplementary Table S3). Of these, 14 miRNAs showed consistent expression levels for both compartments, respectively (miR-3656; -3960; -6087; -6807-5p; 6752-5p; -188-5p; -6089; -4763-3p; -3194-5p; -3940-5p; -1249-3p; 6757-5p; -6088; -6756-5p). Since it has been proposed as a suitable miRNA for data normalization in other contexts, miR-23a-3p was chosen for validation in serum [8, 34, 94].

3.4 Validation phase

MiR-484 showed significant differences via RT-qPCR in serum of 13 EC, 15 OC and 32 BC patients vs. 32 CTRLs (Fig. 2A and B). In alignment with the discovery a two-sample $t$ -test of miR-484 for a combined group of EC and OC vs. CTRL was significant ( $p<$ 0.001; CI: $-$ 0.034, $-$ 0.014) (Fig. 2A). However, the microarray chip analysis detected reduced miRNA expression levels of miR-484 in EC and OC (log2(FC) $=$ $-$ 0.94) while these were found elevated compared to the CTRL in the RT-qPCR analysis.

In urine, miR-10b-5p and -205-5p were analyzed in 25 BC, 10 EC, 10 OC patients and 30 CTRL samples (Fig. 3). However, miR-10b-5p could not be confirmed as a non-tumor entity-specific biomarker in this context ( $p=$ 0.299; CI: $-$ 2.001; 0.627) (Fig. 3A). In fact, a tendency of miR-10b-5p upregulation in EC and OC was observed (Fig. 3B). miR-205-5p was tested in a combined group of EC $+$ OC vs. CTRL but missed statistical significance ( $p=$ 0.033; CI: 0.209–4.748) (Fig. 3C). However, a trend of miR-205-5p downregulation in urine samples of gynecological cancer patients could be observed, especially in the EC group (Fig. 3D).

Furthermore, miR-23a-3p was analyzed in the serum validation cohort to determine its potential as endogenous control. While miR-23a-3p showed a relatively stable expression in the discovery cohort (maximum FC $=$ 1.92), its expression varied between the EC, OC and CTRL group. Compared to the CTRL group, miR-23a-3p was downregulated in the EC group ( $p=$ 0.001; CI: 0.239–0.923) and the OC group ( $p<$ 0.001; CI: 0.366–0.890), respectively. No deregulation was detected between the BC and CTRL group ( $p=$ 0.266; CI: $-$ 0.108–0.386) (Fig. 4).

Figure 4.

miR-23a-3p expression levels in serum of 32 breast (BC), 13 endometrial (EC) and 15 ovarian cancer (OC) patients vs. 32 healthy controls (CTRL). Expression levels of miR-23a-3p were significantly decreased in EC ( $p=$ 0.001) and OC ( $p<$ 0.001) vs. CTRL (two-sample $t$ -test, significance level: 0.017). RT-qPCR.

4. Discussion

The present study contributes to the knowledge about miRNA expression alterations in serum and urine of BC, EC and OC patients compared to healthy women. Thereby, it underlines the immanent potential of miRNAs to serve as non- and minimally-invasive biomarkers in the field of breast and gynecological cancer detection. However, our findings also accentuate the complexity of biomarker research and unravel methodological limitations in miRNA biomarker screening.

4.1 Discovery of potential serum miRNA biomarkers for BC, EC and OC

In serum, 14 miRNAs were discovered by microarray analysis that discriminated between breast and gynecological cancer patients and healthy women (Table 1). Furthermore, 16 BC-specific liquid biopsy biomarker candidates were discovered. While 13 of these miRNAs overlapped with the overall tumor group, three miRNAs (miR-518b, -4719, -6757-3p) were found to be specifically deregulated in BC serum. These three miRNAs have not yet been set into context with BC and could be interesting new players in BC blood screening. Testing for combined EC $+$ OC resulted in 11 miRNA biomarker candidates. As a promising candidate, miR-484 was further investigated via RT-qPCR. While were not able to validate miR-484 via RT-qPCR, we observed elevated levels in EC and OC. These findings accentuate the complexibility of biomarker screening studies in regard to the comparibility of results obtained by different analytical procedures. Of course, in the case of the present study, sample pooling during the discovery phase complicates its results. Therefore, a validation step with more sensitive analytical methods was performed. Since we consider RT-qPCR as the more precise detection method, we assume elevated miR-484 serum levels to be correlated with EC or OC. As miR-484 is secreted by its cells of origin it can be detected in the blood stream [46, 60, 117]. At moderate expression levels, it could be reliably detected in 81.3%, 69.2% and 80.0% of our CTRL, EC and OC serum samples, respectively. To date, only limited data about the role of miR-484 in OC exists [105] while it has not been connected to EC, yet. In OC tissue, it could predict chemoresistance and interacted with angiogenic factors. Specificly, Vecchione et al. reported decreased miR-484 levels in tumors of non-responding patients and proved that miR-484 targets VEGFB (vascular endothelial growth factor B) in OC as well as VEGFR2 (vascular endothelial growth factor receptor 2) in associated endothelial cells [105]. These findings suggest a tumor-suppressive role of miR-484 in OC. In another gynecological cancer entity – cervical cancer – miR-484 was also found downregulated on tumor tissue and cell culture level. By targeting ZEB1 (zinc finger E-box binding homeobox 1) and SMAD2 (mothers against decapentaplegic homolog 2), miR-484 suppressed cell proliferation and epithelial – mesenchymal transition (EMT) [39]. Tumor-suppressive functions of miR-484 were discussed for other carcinomas, too [69, 97].

Contrariwise, miR-484 was reported to exert oncogenic functions by targeting MAP2 (microtubule-associated protein 2) and consecutively activating ERK1/2 (extracellular-signal regulated kinases) leading to a poor prognosis in glioma patients [114]. It was furthermore found to be upregulated in non-small-cell lung cancer tissue and promoted proliferation and migration in this carcinoma by downregulating APAF-1 (apoptotic protease activating factor 1) [53]. A correlation of high miR-484 expression and unfavorable clinicopathological characteristics was reported for various tumor entities [46, 85]. Interestingly, miR-484 was predicted to play a role in BC as well, with increased expression levels being associated with a worse prognosis [48, 95]. While we did not observe different expression levels between BC patients and CTRL in serum or urine specimen, Zearo et al. found elevated levels of miR-484 in serum of early BC patients compared to healthy women [117].

While a clear involvement of miR-484 in carcinogenic processes is certain – its role being highly context-dependent – this miRNA seems to interact in further pathological conditions (e.g. tuberculosis, carotid plaque development and axial spondyloarthritis) as well [3, 15, 31]. On the other hand, miR-484 was proposed as a housekeeping miRNA by some authors in liquid and tissue samples [40, 82, 87, 91, 127].

A number of miRNAs (miR-451b; -548ao-3p; -3921; -6757-3p; -6859-5p; -7844) identified during the discovery phase in serum samples have not been described in the literature, yet. As their physiological and pathophysiological impact remains unknown, their aberrational expression in serum of gynecological cancer patients suggests a discrete involvement in carcinogenic processes. Various miRNAs identified by this study have been investigated in the context of several malignancies and other diseases before and will raise the question of specificity. MiR-143-3p represents a concrete example. This miRNA seems to play a role in numerous cancer entities [27, 37, 125] and other medical conditions such as acute ischemic stroke [102], amyotrophic lateral sclerosis [107] and autoimmune diseases like multiple sclerosis [86]. A growing body of literature attributes a tumor-suppressive role to miR-143-3p which is downregulated in malignant tissues [13, 27, 37, 125]. Furthermore, it was drastically downregulated in serum of OC and estrogen receptor positive BC patients [20, 47]. These findings are congruent to our own conclusions as miR-143-3p was not detectable in serum of all gynecological cancer patients but an expression was observed in both control pools (Supplementary Table S1). Hence, an association between reduced expression levels and the existence of a malignancy can be expected. However, since miR-143-3p seems to be deregulated in various diseases, it lacks specificity in regard to the distinction of tumor entities. Consequently, miRNAs such as miR-143-3p should be combined with entity-specific biomarkers within the framework of liquid biopsy.

4.2 Discovery of potential urinary miRNA biomarkers for BC, EC and OC

Ten miRNAs with general tumor specific characteristics were identified (Table 2). Of these, miR-10b-5p was further analyzed via RT-qPCR due to a reliable detection in urine samples tested, but could not be confirmed as a urine-based non-specific diagnostic marker for BC, EC or OC. There seems to be an overall consensus on miR-10b-5p displaying oncogenic functions in BC by targeting, among others, HOXD10 (homeobox D10), KLF4 (kruppel-like factor 4) and PTEN (phosphatase and tensin homolog) [5, 96]. These functions, however, seem not to be represented in urine. Despite being described as an oncomiR, miR-10b-5p was also found to be downregulated in BC tissue in some studies [16, 55]. Furthermore, clinicopathological parameters could be correlated with miR-10b-5p expression in BC [19, 57, 123]. Focusing on EC, the role of miR-10b-5p is insufficiently understood. On the one hand, it was found downregulated in an invasive form of EC cells and in endometrial EC vs. normal endometrial tissue while no differential expression was observed in plasma [26, 103]. On the other hand, Chen et al. described elevated expression levels in EC tissue compared to CTRL [17]. Our RT-qPCR analysis indicated elevated miR-10b-5p levels in urine samples from EC patients compared to healthy women. In OC, the knowledge about miR-10b-5p function is even more scarce. An oncogenic role is proposed by the targeting of HOXD10 and a subsequent increase of MMP14 (matrix metalloproteinase-14) and RHOC (ras homolog family member C) [75]. While no studies focusing on miR-10b-5p expression in urine of BC, EC or OC patients could be identified for comparison with our own results, literature review revealed few serum-based diagnostic studies. Therein, miR-10b-5p showed potential to discriminate BC patients from CTRLs [16, 66]. To date, miR-10b-5p expression analysis in urine was described to help detect bladder cancer and acute rejection of renal transplants [58, 76]. Based on our own results, we could not confirm miR-10b-5p as a urine-based non-specific diagnostic marker for BC, EC or OC.

While six of the deregulated miRNAs discovered in the BC group overlapped with the overall tumor group comparison, miR-526b-3p remained as an exclusive candidate. It has been reported as a tumor supressor in colon cancer [126] and was found to be elevated in BC cell lines by COX-2 [65]. Downregulation in urine from BC patients as observed in this study suggests tumor-suppressive effects of miRNA-526b-3p.

Furthermore, twelve miRNAs were identified as potential markers for combined detection of EC and OC (Table 2). In analogy to our findings in serum, a subset of these miRNAs has been associated with various diseases and malignancies while another subset was found related to EC and OC by other investigators before [80, 81, 82, 83, 84, 85, 86]. A widely studied example is miR-205-5p that was found to be upregulated in OC tissue [50, 78], to act as an OC invasion promotor [51], to be associated with OC tumor stage, grade and poor survival [50, 78], to trigger EMT in EC cells [42] and to be linked to a poor prognosis in EC [64]. Researchers seem to agree on an oncogenic function of miR-205-5p in EC and report predominantly elevated levels in EC tissue [38, 61, 103]. The same assumption accounts for OC, while few studies also suggest tumor-suppressive traits [18, 52]. Targets of miR-205-5p include tumor suppressors such as FOXO1 (forkhead box O1), PTEN and ZEB1 [50, 61, 78, 121]. In our discovery microarray analysis, miR-205-5p showed reduced expression levels in the EC and OC groups. Consequently, miR-205-5p was chosen for validation. Although this miRNA failed to reach statistical significance in a two-sample $t$ -test during the validation process, a trend of downregulation in urine samples of EC and OC patients vs. CTRL was observed. Since urinary miRNAs might originate from the tumor itself, these could also derive from cells of the immune system as a form of response to the presence of a malignancy. This fact and and the fact that tumor-derived miRNAs in urine will inevitably be diluted by the presence of various miRNAs secreted from non-malignant cells and cells from the urinary tract, might explain the different directions of miR-205-5p deregulation in distinctive sample matrices of EC and OC. No precedent studies analyzing the diagnostic potential of miR-205-5p in urine of EC and OC patients could be identified. In plasma, however, epithelial OC could be discriminated from CTRLs [128] via miR-205-5p, while Kapetanakis et al. found expression differences between OC and benigne ovarian lesions [44]. Furthermore, miR-205-5p was upregulated in endometrioid EC vs. CTRL [103]. While our study did not detect a BC-specific deregulation of miR-205-5p – neither in urine nor in serum – it was reported to be downregulated in BC serum vs. CTRL [55, 122] and proposed as a tumor suppressor in BC [106]. Controversially, miR-205-5p upregulation in BC serum has also been observed [93].

In contrast, literature search did not account for any known functional role for miR-3973; -4426; -5089-5p and -6841 offering novel perspectives to a potential applicability as liquid biomarker candidates for urine-based BC, EC or OC diagnosis.

4.3 Identification of stably expressed miRNAs

Regarding the identification of novel miRNA candidates for normalization and quality control, the present study provides promising results. Considering RT-qPCR, stably expressed miRNAs are extremely important to obtain reliable comparative values when relative quantification is applied. Through normalization to so-called housekeeping miRNAs or other “stable genes” as endogenous controls, biological fluctuations are being eliminated. To date, only a limited panel of housekeeping miRNAs is recommended by scientists in the field [68, 87]. There is a shortage of implemented choices of stable normalizers for the especially demanding compartments serum and urine. Working with biological materials, many variables influence sample quality and subsequent results of downstream applications [45, 118]. It is widely accepted that sample quality control constitutes the foundation of any analysis in the field of liquid biopsy [45]. Therefore, expression levels of endogenous controls should be assessed irrespective of the chosen analytical method and there is a great demand for reliable and robust housekeeping candidates. However, both, high percentage of GC bases in many of the here identified miRNAs and repetitive sequences bear methodological challenges, e.g. primer design and effective amplification rates, at current state-of-the-art RT-qPCR. Due to variable miRNA expression patterns in biofluids concerning distinct malignant conditions, an individual set of endogenous controls according to the investigated issue might need assessment and a subsequent validation.

In the conducted discovery phase, miR-23a-3p was identified as a potential endogenous control due to a rather stable expression in patient’s and control’s serum samples. Supporting this finding, miR-23a-3p was proposed for RT-qPCR data normalization by other researchers before, focusing on plasma, tissue and extracellular vesicle analysis [8, 34, 94]. Additionally, a published method for the assessment of hemolysis in serum samples suggests evaluation of miR-451a/-23a-3p ratio by RT-qPCR. Herein, miR-23a-3p serves as normalizer while miR-451a expression levels change according to the degree of hemolysis due to its enrichment in red blood cells [92]. Validation of miR-23a-3p by RT-qPCR in an independent cohort, however, resulted in the rejection of this miRNA as an endogenous control in serum of breast and gynecological cancer patients. In specific, we detected a downregulation of miR-23a-3p expression levels in EC and OC patients compared to CTRLs. In this context, miR-23a-3p expression levels were found decreased in EC vs. normal endometrium and in an invasive form of EC cells, as well [26, 56]. Targeting SMAD3 (mothers against decapentaplegic homolog 3), miR-23a-3p represses EMT and might function as a tumor suppressor in EC [56]. To the best of our knowledge, miR-23a-3p deregulation was not investigated specifically in serum of EC patients, yet. In regard to OC, miR-23a-3p was described to promote cell migration and EMT leading to malignant progression [113, 129]. Beyond these oncogenic functions, its expression in OC tissue was associated with clinicopathological parameters and survival [84, 100]. While our study identified lower miR-23a-3p expression levels in serum of OC patients vs. CTRLs, Pan et al. did not find a difference in plasma expression levels between a group of epithelial OC patients and healthy women. However, they were able to show decreased levels in ovarian cystadenoma vs. epithelial OC patients [82]. Microarray miRNA chip analysis as well as RT-qPCR validation of serum specimen did not detect expression level alterations of miR-23a-3p between BC patients and CTRLs. In the narrower sense, it might be possible to apply miR-23a-3p as an endogenous control for BC serum analysis. Still, literature review revealed an upregulation of miR-23a-3p in BC tissue and Wu et al. even reported elevated serum levels obtained by next generation sequencing [28, 54, 111]. Focusing on its functional role in BC, miR-23a-3p was described to activate the oncogenes FOXM1 (forkhead box protein M1) and HRG (histidine-rich glycoprotein), promoting EMT by targeting E-cadherin and causing tumor migration via SPRY2 (sprouty homolog 2) repression [28, 54, 62]. Generally speaking, miR-23a-3p manifested to be deregulated in several carcinomas, therein either contributing to the malignancy or exerting tumor-suppressive functions [108]. Summarizing, by means of the results of this study, we propose miR-23a-3p as a diagnostic serum-based biomarker for EC and OC while it might serve as an endogenous control in BC serum studies.

4.4 Methodological limitations

Microarray chip analysis is a convenient and efficient way to screen for novel biomarkers as it enables simultaneous investigation of vast numbers of targets. In the field of miRNA research we are faced with serveral challenges when applying this method. As described above, some housekeeping candidate and discovery phase miRNAs contain challenging repetitive sequences. Due to a short length of approximately 18–22 nucleotides [36], such sequences can not be fully avoided when designing probes or primers. These demanding sequences could result in artifacts masking results from microarray analysis through cross reactions of probes with similar miRNA types. One could assume this as one of the reasons leading to incongruent results of different studies [12]. Diverse analytical methods such as NGS, RT-qPCR or microarrays can lead to aberrational findings [118]. Even within one performed analytical method a risk of differing results exists due to slight variations in assay protocols and sample handling [12, 79]. These variable parameters create challenges for data comparability and might explain why the results of this work differ partly from those of other studies as outlined above. Hence, standardized methods and normalization procedures are urgently needed.

We acknowledge some limitations of the present study. Results of the discovery phase could have been impacted due to the small patient cohort and the pooling of samples. The OC pool contained one patient diagnosed with a benigne ovarian lesion in retrospect. This fact could have influenced the results from the OC cohort.

Another restriction of this study could be a high percentage of LOD measurements in the microarray chip analysis. Approximately 57% of all meassurements exhibit expression values at the LOD. Noteworthy, miRNAs identified by the performed t-tests characterize as being either present in samples of control probands while being absent or weakly expressed in patients specimen (LOD meassurements) or vice versa. In case of the overall tumor and the BC group, miRNAs discovered by this study could for the most part not be detected both in serum and urine of patients but were identified in samples of the control group. This fact indicates a downregulation of these biomarker candidates in tumor patients and their absence could possibly detect the existence of a carcinoma. Contrariwise, testing for combined EC and OC identified miRNAs that were predominantly expressed in patients specimens but not in healthy women.

Regarding the age distribution of our control group as described in the ‘Materials and methods’ section, the CTRL O 40 pool (women $>$ 40 years) mirrors women eligible for common mammography screening. As this study aimes to identify miRNA biomarker candidates suitable for age-independent liquid biopsy screening tests, e.g. performed in the framework of routine patient visits at their gynecologist, an CTRL U 40 pool (women $\leqslant$ 40 years) was included to preferably rule out miRNAs whose expression is influenced by age. Taking these limitations into account, the authors of this study conducted a validation step of selected biomarker candidates applying RT-qPCR.

5. Conclusions

miRNA liquid biopsy research has proven its potential in clinically challenging aspects such as screening, diagnostics and therapy response prediction of various tumor entities. Supporting this aspect, the present study proposes several liquid biopsy-based biomarker candidates to identify female malignancies while highlighting at the same time the complexibility of this approach. These as well as other miRNAs discovered in this work should be validated in larger cohorts to create a clearer picture of their eligibility as minimally invasive diagnostic tools in breast and gynecological cancer screening.

Footnotes

Conflict of interest

The authors declare no potential conflicts of interest.

Abbreviations

References

Adams

B.D.

Kasinski

A.L.

and Slack

F.J.

, Aberrant regulation and function of microRNAs in cancer, Curr Biol 24 (2014), R762–76.

Agilent, Human miRNA Microarray Slide (Release 21.0), https://www.agilent.com/en/product/mirna-microarray-platform/mirna-microarrays, accessed: 2019 June 24.

Alipoor

S.D.

Tabarsi

Varahram

Movassaghi

Dizaji

M.K.

Folkerts

Garssen

Adcock

I.M.

and Mortaz

, Serum exosomal miRNAs are associated with active pulmonary tuberculosis, Dis Markers 2019 (2019), 1907426.

Anderson

G.L.

, Ovarian cancer biomarker screening: Still too early to tell, Womens Health (Lond) 6 (2010), 487–90.

Bahena-Ocampo

Espinosa

Ceballos-Cancino

Lizarraga

Campos-Arroyo

Schwarz

Maldonado

Melendez-Zajgla

and Garcia-Lopez

, miR-10b expression in breast cancer stem cells supports self-renewal through negative PTEN regulation and sustained AKT activation, EMBO Rep 17 (2016), 648–58.

Bahrami

Aledavood

Anvari

Hassanian

S.M.

Maftouh

Yaghobzade

Salarzaee

ShahidSales

and Avan

, The prognostic and therapeutic application of microRNAs in breast cancer: Tissue and circulating microRNAs, J Cell Physiol 233 (2018), 774–786.

Bevers

Bibbins-Domingo

Oeffinger

K.C.

and Smith

M.L.

, Controversies in breast cancer screening strategies, J Natl Compr Canc Netw 14 (2016), 651–3.

Borges

D.P.

Cunha-Neto

Bocchi

E.A.

and Rigaud

V.O.C.

, Impact of delayed whole blood processing time on plasma levels of miR- 1 and miR-423-5p up to 24 Hours, Microrna 7 (2018), 115–119.

Braga

E.A.

Fridman

M.V.

and Kushlinskii

N.E.

, Molecular mechanisms of ovarian carcinoma metastasis: key genes and regulatory microRNAs, Biochemistry (Mosc) 82 (2017), 529–541.

10.

Braun

M.M.

Overbeek-Wager

E.A.

and Grumbo

R.J.

, Diagnosis and management of endometrial cancer, Am Fam Physician 93 (2016), 468–74.

11.

Busk

P.K.

, A tool for design of primers for microRNA-specific quantitative RT-qPCR, BMC Bioinformatics 15 (2014), 29.

12.

Butz

and Patocs

, Technical aspects related to the analysis of circulating microRNAs, EXS 106 (2015), 55–71.

13.

Cabrita

M.A.

Bose

Vanzyl

E.J.

Pastic

Marcellus

K.A.

Pan

Hamill

J.D.

and McKay

B.C.

, The p53 protein induces stable miRNAs that have the potential to modify subsequent p53 responses, Gene 608 (2017), 86–94.

14.

Calin

G.A.

Dumitru

C.D.

Shimizu

Bichi

Zupo

Noch

Aldler

Rattan

Keating

Rai

Rassenti

Kipps

Negrini

Bullrich

and Croce

C.M.

, Frequent deletions and down-regulation of micro- RNA genes miR15 and miR16 at 13q14 in chronic lymphocytic leukemia, Proc Natl Acad Sci USA 99 (2002), 15524–9.

15.

Caparosa

E.M.

Sedgewick

A.J.

Zenonos

Zhao

Carlisle

D.L.

Stefaneanu

Jankowitz

B.T.

Gardner

Chang

Y.F.

Lariviere

W.R.

LaFramboise

W.A.

Benos

P.V.

and Friedlander

R.M.

, Regional molecular signature of the symptomatic atherosclerotic carotid plaque, Neurosurgery (2018).

16.

Chan

Liaw

C.S.

S.M.

Tan

H.H.

Wong

C.Y.

Thike

A.A.

Tan

P.H.

G.H.

and Lee

A.S.

, Identification of circulating microRNA signatures for breast cancer detection, Clin Cancer Res 19 (2013), 4477–87.

17.

Chen

Fan

Chen

Wei

Fang

and Feng

, miR-10b inhibits apoptosis and promotes proliferation and invasion of endometrial cancer cells via targeting HOXB3, Cancer Biother Radiopharm 31 (2016), 225–31.

18.

Chen

Wang

Matyunina

L.V.

Hill

C.G.

and McDonald

J.F.

, Overexpression of miR-429 induces mesenchymal-to-epithelial transition (MET) in metastatic ovarian cancer cells, Gynecol Oncol 121 (2011), 200–5.

19.

Chen

Cai

Zhang

Barekati

and Zhong

X.Y.

, The level of circulating miRNA-10b and miRNA-373 in detecting lymph node metastasis of breast cancer: potential biomarkers, Tumour Biol 34 (2013), 455–62.

20.

Chung

Y.W.

Bae

H.S.

Song

J.Y.

Lee

J.K.

Lee

N.W.

Kim

and Lee

K.W.

, Detection of microRNA as novel biomarkers of epithelial ovarian cancer from the serum of ovarian cancer patients, Int J Gynecol Cancer 23 (2013), 673–9.

21.

De Leeneer

and Claes

, Non coding RNA molecules as potential biomarkers in breast cancer, Adv Exp Med Biol 867 (2015), 263–75.

22.

Deb

Uddin

and Chakraborty

, miRNAs and ovarian cancer: An overview, J Cell Physiol (2017).

23.

Deutsche Krebsgesellschaft, Leitlinienprogramm Onkologie (Deutsche Krebsgesellschaft, Deutsche Krebshilfe, AWMF): Diagnostik, Therapie und Nachsorge der Patientinnen mit Endometriumkarzinom, Langversion 1.0, 2018, AWMF Registernummer: 032/034-OL, http://www.leitlinienprogramm-onkologie.de/leitlinien/endometriumkarzinom/, access- ed: 2019 June 24.

24.

Deutsche Krebsgesellschaft, Leitlinienprogramm Onkologie (Deutsche Krebsgesellschaft, Deutsche Krebshilfe, AWMF): S3-Leitlinie Diagnostik, Therapie und Nachsorge maligner Ovarialtumoren, Langversion 2.0 2016, AWMF-Registernummer: 032/035OL, http://leitlinienprogramm-onkologie.de/Ovarialkarzinom.61.0.html, accessed: 2019 June 24.

25.

Deutsche Krebsgesellschaft, Leitlinienprogramm Onkologie (Deutsche Krebsgesellschaft, Deutsche Krebshilfe, AWMF): S3-Leitlinie Früherkennung, Diagnose, Therapie und Nachsorge des Mammakarzinoms, Version 4.0, 2017 AWMF Registernummer: 032-045OL, http://www.leitlinienprogramm-onkologie.de/leitlinien/mammakarzinom/, accessed: 2019 June 24.

26.

Dong

Ihira

Xiong

Watari

Hanley

S.J.

Yamada

Hosaka

Kudo

Yue

and Sakuragi

, Reactivation of epigenetically silenced miR-124 reverses the epithelial-to-mesenchymal transition and inhibits invasion in endometrial cancer cells via the direct repression of IQGAP1 expression, Oncotarget 7 (2016), 20260–70.

27.

Dong

Cheng

and Yin

, MiR-143 regulates the proliferation and migration of osteosarcoma cells through targeting MAPK7, Arch Biochem Biophys 630 (2017), 47–53.

28.

Eissa

Matboli

and Shehata

H.H.

, Breast tissue-based microRNA panel highlights microRNA-23a and selected target genes as putative biomarkers for breast cancer, Transl Res 165 (2015), 417–27.

29.

Erbes

Hirschfeld

Rucker

Jaeger

Boas

Iborra

Mayer

Gitsch

and Stickeler

, Feasibility of urinary microRNA detection in breast cancer patients and its potential as an innovative non-invasive biomarker, BMC Cancer 15 (2015), 193.

30.

Falcone

Felsani

and D’Agnano

, Signaling by exosomal microRNAs in cancer, J Exp Clin Cancer Res 34 (2015), 32.

31.

Fogel

Bugge Tinggaard

Fagny

Sigrist

Roche

Leclere

Deleuze

J.F.

Batteux

Dougados

Miceli-Richard

and Tost

, Deregulation of microRNA expression in monocytes and CD4(+) T lymphocytes from patients with axial spondyloarthritis, Arthritis Res Ther 21 (2019), 51.

32.

Gasparri

M.L.

Casorelli

Bardhi

Besharat

A.R.

Savone

Ruscito

Farooqi

A.A.

Papadia

Mueller

M.D.

Ferretti

and Benedetti Panici

, Beyond circulating microRNA biomarkers: Urinary microRNAs in ovarian and breast cancer, Tumour Biol 39 (2017), 1010428317695525.

33.

Goldhirsch

Winer

E.P.

Coates

A.S.

Gelber

R.D.

Piccart-Gebhart

Thurlimann

Senn

H.J.

and Panel

, Personalizing the treatment of women with early breast cancer: highlights of the St Gallen International Expert Consensus on the Primary Therapy of Early Breast Cancer 2013, Ann Oncol 24 (2013), 2206–23.

34.

Gouin

Peck

Antes

Johnson

J.L.

Vaturi

S.D.

Middleton

de Couto

Walravens

A.S.

Rodriguez-Borlado

Smith

R.R.

Marban

and Ibrahim

A.G.

, A comprehensive method for identification of suitable reference genes in extracellular vesicles, J Extracell Vesicles 6 (2017), 1347019.

35.

Griffiths-Jones

, miRBase Release 21, http://www.mirbase.org/, accessed: 2019 June 24.

36.

and Kim

V.N.

, Regulation of microRNA biogenesis, Nat Rev Mol Cell Biol 15 (2014), 509–24.

37.

Liu

Chen

Han

Jin

Chen

and Song

, MiR-143-3p functions as a tumor suppressor by regulating cell proliferation, invasion and epithelial-mesenchymal transition by targeting QKI-5 in esophageal squamous cell carcinoma, Mol Cancer 15 (2016), 51.

38.

Hiroki

Akahira

Suzuki

Nagase

Ito

Suzuki

Sasano

and Yaegashi

, Changes in microRNA expression levels correlate with clinicopathological features and prognoses in endometrial serous adenocarcinomas, Cancer Sci 101 (2010), 241-9.

39.

Xie

Liu

and Tang

, miR-484 suppresses proliferation and epithelial-mesenchymal transition by targeting ZEB1 and SMAD2 in cervical cancer cells, Cancer Cell Int 17 (2017), 36.

40.

Dong

Wang

L.E.

Liu

Zhao

Tang

Chen

Dai

Wei

Zhang

and Shen

, Serum microRNA profiling and breast cancer risk: the use of miR-484/191 as endogenous controls, Carcinogenesis 33 (2012), 828–34.

41.

Huang

and Sood

A.K.

, Prognostic biomarkers in ovarian cancer, Cancer Biomark 8 (2010), 231–51.

42.

Jin

and Liang

, miR-205 promotes epithelial-mesenchymal transition by targeting AKT signaling in endometrial cancer cells, J Obstet Gynaecol Res 41 (2015), 1653–60.

43.

Kanekura

Nishi

Isaka

and Kuroda

, MicroRNA and gynecologic cancers, J Obstet Gynaecol Res 42 (2016), 612–7.

44.

Kapetanakis

N.I.

Uzan

Jimenez-Pailhes

A.S.

Gouy

Bentivegna

Morice

Caron

Gourzones-Dmitriev

Le Teuff

and Busson

, Plasma miR-200b in ovarian carcinoma patients: distinct pattern of pre/post-treatment variation compared to CA-125 and potential for prediction of progression-free survival, Oncotarget 6 (2015), 36815–24.

45.

Khan

Lieberman

J.A.

and Lockwood

C.M.

, Variability in, variability out: best practice recommendations to standardize pre-analytical variables in the detection of circulating and tissue microRNAs, Clin Chem Lab Med 55 (2017), 608–621.

46.

Kjersem

J.B.

Ikdahl

Lingjaerde

O.C.

Guren

Tveit

K.M.

and Kure

E.H.

, Plasma microRNAs predicting clinical outcome in metastatic colorectal cancer patients receiving first-line oxaliplatin-based treatment, Mol Oncol 8 (2014), 59–67.

47.

Kodahl

A.R.

Lyng

M.B.

Binder

Cold

Gravgaard

Knoop

A.S.

and Ditzel

H.J.

, Novel circulating microRNA signature as a potential non-invasive multi-marker test in ER-positive early-stage breast cancer: a case control study, Mol Oncol 8 (2014), 874–83.

48.

D.H.

, Network-based ranking methods for prediction of novel disease associated microRNAs, Comput Biol Chem 58 (2015), 139–48.

49.

Lehmann

T.P.

Korski

Gryczka

Ibbs

Thieleman

Grodecka-Gazdecka

and Jagodzinski

P.P.

, Relative levels of let-7a, miR-17, miR-27b, miR-125a, miR-125b and miR-206 as potential molecular markers to evaluate grade, receptor status and molecular type in breast cancer, Mol Med Rep 12 (2015), 4692–702.

50.

Gong

Zhu

Wang

Liu

and Wu

, Upregulation of MiR-205 transcriptionally suppresses SMAD4 and PTEN and contributes to human ovarian cancer progression, Sci Rep 7 (2017), 41330.

51.

Gong

Chen

Liu

Wang

Liu

and Wu

, The role of miR-205 in the VEGF-mediated promotion of human ovarian cancer cell invasion, Gynecol Oncol 137 (2015), 125–33.

52.

Yang

Zhang

and Chen

, Tanshinone IIA effects on ovarian cancer cell line, J Pharm Pharmacol 70 (2018), 1369–1377.

53.

Ding

Z.L.

Zheng

Y.L.

and Wang

, MiR-484 promotes non-small-cell lung cancer (NSCLC) progression through inhibiting Apaf-1 associated with the suppression of apoptosis, Biomed Pharmacother 96 (2017), 153–164.

54.

Liu

Zhou

Gao

Jiang

Lobie

P.E.

and Zhu

, c-MYC-regulated miR-23a/24-2/27a cluster promotes mammary carcinoma cell invasion and hepatic metastasis by targeting Sprouty2, J Biol Chem 288 (2013), 18121–33.

55.

Liu

Mao

Liu

Hao

Zhang

and Zhang

, Analysis of miR-205 and miR-155 expression in the blood of breast cancer patients, Chin J Cancer Res 25 (2013), 46–54.

56.

Liu

Wang

Zhang

and Lao

, MicroRNA-23a regulates epithelial-to-mesenchymal transition in endometrial endometrioid adenocarcinoma by targeting SMAD3, Cancer Cell Int 16 (2016), 67.

57.

Liu

Guan

Wang

and Niu

, MicroRNA-10b expression in node-negative breast cancer-correlation with metastasis and angiogenesis, Oncol Lett 14 (2017), 5845–5852.

58.

Lorenzen

J.M.

Volkmann

Fiedler

Schmidt

Scheffner

Haller

Gwinner

and Thum

, Urinary miR-210 as a mediator of acute T-cell mediated rejection in renal allograft recipients, Am J Transplant 11 (2011), 2221–7.

59.

Mao

Shi

Zhang

and Wang

, MicroRNAs in the prognosis of triple-negative breast cancer: A systematic review and meta-analysis, Medicine (Baltimore) 96 (2017), e7085.

60.

and Lu

, The significance of detection of serum miR-423-5p and miR-484 for diagnosis of colorectal cancer, Clin Lab 61 (2015), 187–90.

61.

Yao

and Jiang

, miR-205-5p contributes to paclitaxel resistance and progression of endometrial cancer by downregulating FOXO1, Oncol Res (2019).

62.

Liu

Lei

Ren

Pang

and Qian

, MiR-23a promotes TGF-beta1-induced EMT and tumor metastasis in breast cancer cells by directly targeting CDH1 and activating Wnt/beta-catenin signaling, Oncotarget 8 (2017), 69538–69550.

63.

and Lu

, MicroRNA and Long Non-Coding RNA in Ovarian Carcinoma: Translational Insights and Potential Clinical Applications, Cancer Invest 34 (2016), 465–476.

64.

Y.J.

C.F.

Bai

Z.M.

H.N.

Xiong

and Jiang

, Overexpression of microRNA-205 predicts lymph node metastasis and indicates an unfavorable prognosis in endometrial cancer, Oncol Lett 12 (2016), 4403–4410.

65.

Majumder

Landman

Liu

Hess

and Lala

P.K.

, COX-2 Elevates Oncogenic miR-526b in Breast Cancer by EP4 Activation, Mol Cancer Res 13 (2015), 1022–33.

66.

Mar-Aguilar

Mendoza-Ramirez

J.A.

Malagon-Santiago

Espino-Silva

P.K.

Santuario-Facio

S.K.

Ruiz-Flores

Rodriguez-Padilla

and Resendez-Perez

, Serum circulating microRNA profiling for identification of potential breast cancer biomarkers, Dis Markers 34 (2013), 163–9.

67.

Marabita

de Candia

Torri

Tegner

Abrignani

and Rossi

R.L.

, Normalization of circulating microRNA expression data obtained by quantitative real-time RT-PCR, Brief Bioinform 17 (2016), 204–12.

68.

McDermott

A.M.

Kerin

M.J.

and Miller

, Identification and validation of miRNAs as endogenous controls for RQ-PCR in blood specimens for breast cancer studies, PLoS One 8 (2013), e83718.

69.

Mei

Xue

Yan

Guo

Sun

and Han

, Methylation-induced loss of miR-484 in microsatellite-unstable colorectal cancer promotes both viability and IL-8 production via CD137L, J Pathol 236 (2015), 165–74.

70.

Menon

Griffin

and Gentry-Maharaj

, Ovarian cancer screening – current status, future directions, Gynecol Oncol 132 (2014), 490–5.

71.

Miller

A.B.

Wall

Baines

C.J.

Sun

and Narod

S.A.

, Twenty five year follow-up for breast cancer incidence and mortality of the Canadian National Breast Screening Study: randomised screening trial, BMJ 348 (2014), g366.

72.

Montagnana

Benati

Danese

Giudici

Perfranceschi

Ruzzenenete

Salvagno

G.L.

Bassi

Gelati

Paviati

Guidi

G.C.

Franchi

and Lippi

, Aberrant MicroRNA Expression in Patients With Endometrial Cancer, Int J Gynecol Cancer 27 (2017), 459–466.

73.

Morice

Leary

Creutzberg

Abu-Rustum

and Darai

, Endometrial cancer, Lancet 387 (2016), 1094–1108.

74.

Moyer

V.A.

and Force

U.S.P.S.T.

, Screening for ovarian cancer: U.S. Preventive Services Task Force reaffirmation recommendation statement, Ann Intern Med 157 (2012), 900–4.

75.

Nakayama

Shibazaki

Yashima-Abo

Miura

Sugiyama

Masuda

and Maesawa

, Loss of HOXD10 expression induced by upregulation of miR-10b accelerates the migration and invasion activities of ovarian cancer cells, Int J Oncol 43 (2013), 63–71.

76.

Nekoohesh

Modarressi

M.H.

Mowla

S.J.

Sadroddiny

Etemadian

Afsharpad

Zolfaghari

Barzegari

Saffari

Oskooei

V.K.

Motevaseli

and Ghafouri-Fard

, Expression profile of miRNAs in urine samples of bladder cancer patients, Biomark Med 12 (2018), 1311–1321.

77.

Nelson

H.D.

Cantor

Pappas

Daeges

and Humphrey

, Effectiveness of Breast Cancer Screening: Systematic Review and Meta-analysis to Update the 2009 U.S. Preventive Services Task Force Recommendation, Ann Intern Med 164 (2016), 244–55.

78.

Niu

Shen

Zhang

Zhao

and Lu

, MiR-205 promotes motility of ovarian cancer cells via targeting ZEB1, Gene 574 (2015), 330–6.

79.

O’Brien

K.P.

Ramphul

Howard

Gallagher

W.M.

Malone

Kerin

M.J.

and Dwyer

R.M.

, Circulating MicroRNAs in Cancer, Methods Mol Biol 1509 (2017), 123–139.

80.

OakLabs, https://www.oak-labs.com/, accessed: 2019 June 24.

81.

Oeffinger

K.C.

Fontham

E.T.

Etzioni

Herzig

Michaelson

J.S.

Shih

Y.C.

Walter

L.C.

Church

T.R.

Flowers

C.R.

LaMonte

S.J.

Wolf

A.M.

DeSantis

Lortet-Tieulent

Andrews

Manassaram-Baptiste

Saslow

Smith

R.A.

Brawley

O.W.

Wender

and American Cancer

, Breast Cancer Screening for Women at Average Risk: 2015 Guideline Update From the American Cancer Society, JAMA 314 (2015), 1599–614.

82.

Pan

Stevic

Muller

Oliveira-Ferrer

Pantel

and Schwarzenbach

, Exosomal microRNAs as tumor markers in epithelial ovarian cancer, Mol Oncol 12 (2018), 1935–1948.

83.

Peairs

K.S.

Choi

Stewart

R.W.

and Sateia

H.F.

, Screening for breast cancer, Semin Oncol 44 (2017), 60–72.

84.

Prahm

K.P.

Hogdall

Karlsen

M.A.

Christensen

I.J.

Novotny

G.W.

and Hogdall

, Identification and validation of potential prognostic and predictive miRNAs of epithelial ovarian cancer, PLoS One 13 (2018), e0207319.

85.

Prior

Perez-Gracia

J.L.

Garcia-Donas

Rodriguez-Antona

Guruceaga

Esteban

Suarez

Castellano

del Alba

A.G.

Lozano

M.D.

Carles

Climent

M.A.

Arranz

J.A.

Gallardo

Puente

Bellmunt

Gurpide

Lopez-Picazo

J.M.

Hernandez

A.G.

Mellado

Martinez

Moreno

Font

and Calvo

, Identification of tissue microRNAs predictive of sunitinib activity in patients with metastatic renal cell carcinoma, PLoS One 9 (2014), e86263.

86.

Regev

Healy

B.C.

Khalid

Paul

Chu

Tauhid

Tummala

Diaz-Cruz

Raheja

Mazzola

M.A.

von Glehn

Kivisakk

Dupuy

S.L.

Kim

Chitnis

Weiner

H.L.

Gandhi

and Bakshi

, Association Between Serum MicroRNAs and Magnetic Resonance Imaging Measures of Multiple Sclerosis Severity, JAMA Neurol 74 (2017), 275–285.

87.

Rice

Roberts

Rai

S.N.

and Galandiuk

, Housekeeping genes for studies of plasma microRNA: A need for more precise standardization, Surgery 158 (2015), 1345–51.

88.

Romero-Cordoba

S.L.

Salido-Guadarrama

Rodriguez-Dorantes

and Hidalgo-Miranda

, miRNA biogenesis: biological impact in the development of cancer, Cancer Biol Ther 15 (2014), 1444–55.

89.

S. G.

,Clinical Practice Endometrial Cancer Working Group,Endometrial cancer: a review and current management strategies: part I, Gynecol Oncol,134(2014),385–92.

90.

Samuel

and Carter

D.R.

, The Diagnostic and Prognostic Potential of microRNAs in Epithelial Ovarian Carcinoma, Mol Diagn Ther 21 (2017), 59–73.

91.

Seclaman

Narita

Anghel

Cireap

Ilina

Sirbu

I.O.

and Marian

, MicroRNA Expression in Laser Micro-dissected Breast Cancer Tissue Samples – a Pilot Study, Pathol Oncol Res 25 (2019), 233–239.

92.

Shah

J.S.

Soon

P.S.

and Marsh

D.J.

, Comparison of Methodologies to Detect Low Levels of Hemolysis in Serum for Accurate Assessment of Serum microRNAs, PLoS One 11 (2016), e0153200.

93.

Shaker

Maher

Nassar

Morcos

and Gad

, Role of microRNAs -29b-2, -155, -197 and -205 as diagnostic biomarkers in serum of breast cancer females, Gene 560 (2015), 77–82.

94.

Shen

Wang

Wan

and Xie

, Identification of miR-23a as a novel microRNA normalizer for relative quantification in human uterine cervical tissues, Exp Mol Med 43 (2011), 358–66.

95.

Shi

Dong

Jiang

Wang

Tan

Yang

Guo

Ming

and Huang

, Construction of prognostic microRNA signature for human invasive breast cancer by integrated analysis, Onco Targets Ther 12 (2019), 1979–2010.

96.

Singh

Pochampally

Watabe

and Mo

Y.Y.

, Exosome-mediated transfer of miR-10b promotes cell invasion in breast cancer, Mol Cancer 13 (2014), 256.

97.

Slotta-Huspenina

Drecoll

Feith

Habermehl

Combs

Weichert

Bettstetter

Becker

and Langer

, MicroRNA expression profiling for the prediction of resistance to neoadjuvant radiochemotherapy in squamous cell carcinoma of the esophagus, J Transl Med 16 (2018), 109.

98.

Smith

Agarwal

and Bhowmick

N.A.

, MicroRNA applications for prostate, ovarian and breast cancer in the era of precision medicine, Endocr Relat Cancer 24 (2017), R157–R172.

99.

Smith

R.A.

Andrews

Brooks

DeSantis

C.E.

Fedewa

S.A.

Lortet-Tieulent

Manassaram-Baptiste

Brawley

O.W.

and Wender

R.C.

, Cancer screening in the United States, 2016: A review of current American Cancer Society guidelines and current issues in cancer screening, CA Cancer J Clin 66 (2016), 96–114.

100.

and Liu

, Correlation analysis on the expression levels of microRNA-23a and microRNA-23b and the incidence and prognosis of ovarian cancer, Oncol Lett 16 (2018), 262–266.

101.

The Canadian Task Force on Preventive Health Care, Recommendations on screening for breast cancer in average-risk women aged 40–74 years, CMAJ 183 (2011), 1991–2001.

102.

Tiedt

Prestel

Malik

Schieferdecker

Duering

Kautzky

Stoycheva

Bock

Northoff

B.H.

Klein

Dorn

Krohn

Teupser

Liesz

Plesnila

Holdt

L.M.

and Dichgans

, RNA-Seq Identifies Circulating miR-125a-5p, miR-125b-5p, and miR-143-3p as Potential Biomarkers for Acute Ischemic Stroke, Circ Res 121 (2017), 970–980.

103.

Tsukamoto

Miura

Mishima

Abe

Kaneuchi

Higashijima

Miura

Kinoshita

Yoshiura

and Masuzaki

, Identification of endometrioid endometrial carcinoma-associated microRNAs in tissue and plasma, Gynecol Oncol 132 (2014), 715–21.

104.

Vasilatou

Sioulas

V.D.

Pappa

Papageorgiou

S.G.

and Vlahos

N.F.

, The role of miRNAs in endometrial cancer, Epigenomics 7 (2015), 951–9.

105.

Vecchione

Belletti

Lovat

Volinia

Chiappetta

Giglio

Sonego

Cirombella

Onesti

E.C.

Pellegrini

Califano

Pignata

Losito

Canzonieri

Sorio

Alder

Wernicke

Stoppacciaro

Baldassarre

and Croce

C.M.

, A microRNA signature defines chemoresistance in ovarian cancer through modulation of angiogenesis, Proc Natl Acad Sci USA 110 (2013), 9845–50.

106.

Vimalraj

Miranda

P.J.

Ramyakrishna

and Selvamurugan

, Regulation of breast cancer and bone metastasis by microRNAs, Dis Markers 35 (2013), 369–87.

107.

Waller

Goodall

E.F.

Milo

Cooper-Knock

Da Costa

Hobson

Kazoka

Wollff

Heath

P.R.

Shaw

P.J.

and Kirby

, Serum miRNAs miR-206, 143-3p and 374b-5p as potential biomarkers for amyotrophic lateral sclerosis (ALS), Neurobiol Aging 55 (2017), 123–131.

108.

Wang

Tan

H.Y.

Feng

Y.G.

Zhang

Chen

and Feng

, microRNA-23a in Human Cancer: Its Roles, Mechanisms and Therapeutic Relevance, Cancers (Basel) 11 (2018).

109.

Wilt

T.J.

Harris

R.P.

Qaseem

and P. High Value Care Task Force of the American College of, Screening for cancer: advice for high-value care from the American College of Physicians, Ann Intern Med 162 (2015), 718–25.

110.

World Health Organization, Cancer Today, https://gco.iarc.fr/today/home, accessed: 2019 June 24.

111.

Wang

Guo

and Ge

, Analysis of serum genome-wide microRNAs for breast cancer detection, Clin Chim Acta 413 (2012), 1058–65.

112.

Xie

Wang

Liu

Wang

Qiao

Xie

and Zhang

, Diagnostic significance of circulating multiple miRNAs in breast cancer: a systematic review and meta-analysis, Biomark Med 10 (2016), 661–74.

113.

Yang

Wang

X.L.

Bai

Liu

W.Y.

Liu

and Tang

, miR-23a promotes IKKalpha expression but suppresses ST7L expression to contribute to the malignancy of epithelial ovarian cancer cells, Br J Cancer 115 (2016), 731–40.

114.

Feng

Yang

Huang

Liao

and Luo

, miR-484/MAP2/c-Myc-positive regulatory loop in glioma promotes tumor-initiating properties through ERK1/2 signaling, J Mol Histol 49 (2018), 209–218.

115.

Yoav

and Hochberg

, Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing, Journal of the Royal Statistical Society 57 (1955), 289–300.

116.

Yoav

and Hochberg

, On the adaptive control of the false discovery rate in multiple testing with independent statistics, Journal of Educational and Behavioral Statistics 25 (2000), 60–83.

117.

Zearo

Kim

Zhu

Zhao

J.T.

Sidhu

S.B.

Robinson

B.G.

and Soon

, MicroRNA-484 is more highly expressed in serum of early breast cancer patients compared to healthy volunteers, BMC Cancer 14 (2014), 200.

118.

Zendjabil

Favard

Tse

Abbou

and Hainque

, The microRNAs as biomarkers: What prospects? C R Biol 340 (2017), 114–131.

119.

Zhang

Cai

F.F.

and Zhong

X.Y.

, An overview of biomarkers for the ovarian cancer diagnosis, Eur J Obstet Gynecol Reprod Biol 158 (2011), 119–23.

120.

Zhang

Pan

Cobb

G.P.

and Anderson

T.A.

, microRNAs as oncogenes and tumor suppressors, Dev Biol 302 (2007), 1–12.

121.

Zhang

Hou

and Zhao

, MiR-205 inhibits cell apoptosis by targeting phosphatase and tensin homolog deleted on chromosome ten in endometrial cancer Ishikawa cells, BMC Cancer 14 (2014), 440.

122.

Zhang

Zhao

and Chang

, The expression and clinical significance of serum miR-205 for breast cancer and its role in detection of human cancers, Int J Clin Exp Med 8 (2015), 3034–43.

123.

Zhang

Yang

Zhang

Sun

and Zhang

, MicroRNA-10b expression in breast cancer and its clinical association, PLoS One 13 (2018), e0192509.

124.

Zhang

Jin

Wang

Zhao

Chen

Wang

Cao

Xie

and Liang

, A circulating miRNA signature as a diagnostic biomarker for non-invasive early detection of breast cancer, Breast Cancer Res Treat 154 (2015), 423–34.

125.

Zhang

Feng

Liu

and Yang

, MiR-143 inhibits cell proliferation and invasion by targeting DNMT3A in gastric cancer, Tumour Biol 39 (2017), 1010428317711312.

126.

Zhang

Zhao

Wang

and Jia

, miR-526b-3p functions as a tumor suppressor in colon cancer by regulating HIF-1alpha, Am J Transl Res 8 (2016), 2783–9.

127.

Zheng

Wang

Zhang

Yang

Wang

Liu

and Wang

, Identification and validation of reference genes for qPCR detection of serum microRNAs in colorectal adenocarcinoma patients, PLoS One 8 (2013), e83025.

128.

Zheng

Zhang

Zhao

Yang

Song

Wen

Hao

Zhang

and Chen

, Plasma miRNAs as diagnostic and prognostic biomarkers for ovarian cancer, PLoS One 8 (2013), e77853.

129.

Zhuang

R.J.

Bai

X.X.

and Liu

, MicroRNA-23a depletion promotes apoptosis of ovarian cancer stem cell and inhibits cell migration by targeting DLG2, Cancer Biol Ther 20 (2019), 897–911.

Discovery of potential serum and urine-based microRNA as minimally-invasive biomarkers for breast and gynecological cancer

Abstract

BACKGROUND:

OBJECTIVE:

METHODS:

RESULTS:

CONCLUSIONS:

Keywords

1. Introduction

Table 1 Characteristics of BC patients stratified by tumor molecular subtype and healthy controls

2.1 Patient cohort and sampling procedures

2.2 RNA isolation

2.3 miRNA microarray chip analysis

Table 3 Deregulated miRNAs in serum of breast and gynecological cancer patients discovered by microarray chip analysis. Log 2 (FC) and p -values

3.1 Discovery phase: miRNA biomarker candidates in serum

Table 4 Deregulated miRNAs in urine of breast and gynecological cancer patients discovered by microarray chip analysis. Log 2 (FC) and p -values

3.3 Discovery phase: Promising miRNAs for normalization

3.4 Validation phase

4.1 Discovery of potential serum miRNA biomarkers for BC, EC and OC

4.2 Discovery of potential urinary miRNA biomarkers for BC, EC and OC

4.3 Identification of stably expressed miRNAs

4.4 Methodological limitations

5. Conclusions

Footnotes

Conflict of interest

Abbreviations

References

Table 1
Characteristics of BC patients stratified by tumor molecular subtype and healthy controls

Table 3
Deregulated miRNAs in serum of breast and gynecological cancer patients discovered by microarray chip analysis. Log ${}_{2}$ (FC) and $p$ -values

Table 4
Deregulated miRNAs in urine of breast and gynecological cancer patients discovered by microarray chip analysis. Log ${}_{2}$ (FC) and $p$ -values