Background
Triple-negative breast cancers (TNBCs), which lack expression of the estrogen receptor alpha (ERα), progesterone receptor, and human epidermal growth factor receptor 2 (HER2), are particularly aggressive and clinically challenging to manage due to the absence of these common therapeutic targets [
1]. TNBC is responsible for approximately 25 % of all breast cancer deaths, although it represents only 15 % of patients diagnosed with invasive breast cancer. TNBC patients have a poorer clinical outcome, and experience a higher rate of distant recurrence (34.0 % versus 20.4 % in other breast cancers) [
2]. In consequence, they have a lower 5-year overall survival rate than patients with other breast cancer subtypes [
3]. Younger women and women of African-American descent are at a particularly high risk of TNBC [
4]. Therefore, there is a critical need to understand the molecular mechanisms promoting the aggressive phenotype of TNBCs and to identify additional prognostic markers to improve early clinical diagnosis of primary or recurrent disease [
5].
MicroRNAs (miRNAs) are a class of endogenously expressed, evolutionarily conserved, small noncoding RNAs, which play crucial regulatory roles in a variety of normal cellular processes [
6]. Their aberrant expression in cancer has been implicated in various signaling pathways affecting tumor initiation, growth, invasion, and metastasis [
7]. For example, miR-10b has been shown to promote metastasis of 4T1 cell line-derived breast tumors in a mouse mammary model [
8]. In contrast, miR-let-7 suppresses breast cancer cell migration and invasion, in part through downregulation of C-C chemokine receptor type 7 [
9], while in lung cancer, low miR-let-7 levels lead to higher expression of RAS protein [
10]. Blenkiron et al. showed that miR-155 is differentially expressed in ERα– versus ERα + tumors, and is overexpressed in breast tumors compared to normal breast tissue [
11]. Differential expression of six miRNAs (miR-142-3p, miR-505*, miR-1248, miR-181a-2*, miR-25* and miR-340*) was found to accurately discriminate between tumors from BRCA1/2 mutation carriers and noncarriers [
12]. Furthermore, miRNAs are stable in both primary tumors and in the circulation, leading researchers to propose their use as biomarkers for cancer diagnosis and prognosis. The importance of circulating miRNAs as biomarkers for breast cancer has been documented in several studies [
13‐
16]. For example, Zhu and coworkers showed that miR-155 may be differentially expressed in the serum of women with hormone-sensitive versus hormone-insensitive breast cancer [
15]. Madhavan et al. demonstrated that breast cancer patients that are positive for circulating tumor cells (CTCs) versus patients negative for CTCs had significantly higher levels of miR-141, miR-200a, miR-200b, miR-200c, miR-203, miR-210, miR-375, and miR-801 in their plasma [
13]. While considerable effort has been invested in understanding how miRNAs regulate various genes, much less is known about the regulation of miRNA expression.
ADAM8 (a disintegrin and metalloproteinase 8) is a transmembrane protein that belongs to the ADAM family of proteins that mediate cell adhesion, cell migration, and proteolysis of a variety of substrates in the extracellular matrix [
17]. ADAM8 is synthesized with a signal sequence along with five domains, namely the prodomain (PRO), and the metalloproteinase (MP), disintegrin (DI), cysteine-rich (CRD), and epidermal growth factor (EGF)-like (ELD) domains. It also has a transmembrane region and a cytoplasmic tail. The MP domain is catalytically active [
18,
19] and can shed various cytokines and their receptors. The DI and CRD domains of ADAM proteins have been suggested to bind integrins and other receptors, and mediate cell adhesion. This adhesion can sometimes occur via an RGD sequence [
20]; however, ADAM8 lacks this sequence and Stone et al. [
21] have proposed that the three-dimensional (3D) structure of the disintegrin loop, and not its sequence, may play an important role in mediating these interactions. ADAM8 is synthesized as a 120-kDa proform, which can dimerize or multimerize, and autocatalytically clip off its prodomain, leaving an active membrane-associated metalloprotease of 90-kDa [
22]. Active ADAM8 can be further processed by the release of the MP domain into the extracellular matrix, leaving behind a 60-kDa membrane-associated remnant form. The cytoplasmic tail of ADAM8 is relatively long and has a conserved potential SH3 ligand domain, similar to ADAM9 [
23]. The natural ligands of the ADAM8 cytoplasmic domain have not yet been identified, but it is likely that the tail has signaling potential.
Recently, our laboratory has shown that ADAM8 is highly expressed in breast tumors, especially in TNBC, compared to normal tissue and its level correlates with poor patient outcome [
24]. Furthermore, ADAM8 was detected by immunohistochemistry in 48 % of all breast cancer-derived metastases. Knockdown of ADAM8 in TNBC cells decreased their ability to migrate, to invade through Matrigel in a Boyden chamber assay, and to form branched colonies in 3D-Matrigel outgrowth assays in vitro. In an orthotopic mouse model, tumors derived from human TNBC cells with ADAM8 knockdown failed to grow beyond a palpable size due to impaired angiogenesis, and showed greatly reduced ability to metastasize [
24]. Mechanistic studies identified two major ADAM8 functions: (1) promoting angiogenesis through release of VEGF-A and other pro-angiogenic factors; and (2) activating β1-integrin on the cancer cells needed for intravasation and extravasation allowing for tumor dissemination and metastasis. Significantly, treatment with an anti-ADAM8 antibody targeting its extracellular MP and DI domains reduced primary tumor burden and metastases in mice [
24]. Given the key role ADAM8 plays in promoting invasion and metastasis of TNBCs, here we tested the hypothesis that ADAM8 mediates the aggressive phenotype of TNBC cells through regulation of specific miRNAs. Using ADAM8 knockdown strategies, 68 miRNAs were identified in MDA-MB-231 TNBC cells, including miR-720 which is overexpressed in several cancers [
25] and secreted from TNBC cells. miR-720 was shown to be induced by ADAM8 via a β1-integrin to ERK signaling cascade, and to mediate signals that promote the invasive and migratory phenotype of TNBC cells in culture. In an orthotopic mouse model, miR-720 was detectable in the serum of mice bearing ADAM8-expressing tumors. Importantly, miR-720 levels were elevated in serum samples of TNBC patients with high ADAM8 expression. Overall, these studies suggest that miR-720 plays an essential role in the aggressive phenotype of ADAM8-positive TNBCs and may serve in a group of miRNAs as a biomarker for early detection of recurrence and treatment efficacy in ADAM8-positive TNBC patients.
Methods
Antibodies and inhibitors
The ADAM8 antibody (B4068) used for Western blotting was purchased from LifeSpan Biosciences. Antibodies against β-actin (AC-15) and β-tubulin (TUB 2.1), and the ERK inhibitor FR180204 (SML0320) were obtained from Sigma-Aldrich. The antibody to detect specific ERK 1/2 phosphorylated forms (pERK1/2) (9101) was obtained from Cell Signaling Technology. The β1-integrin antibody (552828) and its control isotype matched IgG2A (SC3883) were purchased from BD Biosciences and Santa Cruz Biotechnology, respectively. The anti-ADAM8 antibody MAB10311 and its control isotype-matched IgG1 (MAB002) were from R&D Systems.
Cells and culture conditions
The MDA-MB-231 and Hs578T human TNBC cell lines and the human umbilical vein endothelial cell (HUVEC) line were purchased from the American Type Culture Collection (ATCC) and maintained in media recommended by ATCC. Stable clones of ADAM8 short hairpin RNA (shRNA) (shA8-20) and control shRNA (shCtrl-3) MDA-MB-231 cells expressing green fluorescent protein were isolated as described previously [
24], and kindly provided by Joerg Bartsch (Philipps University, Marburg, Germany). The identity of the MDA-MB-231 cell line and the shA8-20 and shCtrl-3 MDA-MB-231 clones was authenticated using short tandem repeat analysis (Genetica DNA Laboratories), which showed 100 % identity with the MDA-MB-231 cell line of ATCC. The inflammatory breast cancer cell line SUM-149 was kindly provided by Stephen Ethier (University of Michigan Medical School, Ann Arbor, MI, USA) and maintained as previously described [
26]. HEK-293 cells were kindly provided by Nader Rahimi (Boston University School of Medicine, Boston, MA, USA). All cell cultures were confirmed to be free of mycoplasma contamination using polymerase chain reaction (PCR) (VenorGeM Mycoplasma Detection Kit, Sigma). To test for the effects of inhibiting ADAM8 activity, cells were treated either with siADAM8 (see below) or with 20 μg/ml anti-ADAM8 antibody MAB10311, or control isotype IgG1 MAB002. To test for the effects of inhibition of β1-integrin signaling, cells were treated with 10 or 20 μg/ml of either antagonist β1-integrin antibody or control isotype (IgG2A rat). To inhibit ERK signaling, cells were treated with 25 μM FR180204.
Cloned DNA and plasmid transfection
The human full-length ADAM8 (hADAM8) cDNA (MGC:134985; Genbank:BC115404.1) and the remnant form ADAM8 cDNA were kindly provided by Joerg Bartsch as described previously [
27]. The enzymatically inactive mutant of hADAM8 was prepared as described previously [
28]. For transient transfection of shA8-20 cells, cultures were incubated for 48 h in the presence of 4 or 8 μg total DNA in six-well or P60 plates, respectively, with Lipofectamine 2000 (Invitrogen) transfection reagent. For transfection of HEK-293 cells, 1 or 2 μg of total DNA in Lipofectoamine 2000 was used, as indicated. Empty pcDNA3 vector (EV) or pcDNA3.1 myc-his vector (EV-3.1) DNA were used as negative controls.
siRNA knockdown analyses
Transient RNAi-mediated ADAM8 knockdown was performed with the following short interfering RNAs (siRNAs) (Qiagen):
AllStar negative control siRNA (Qiagen) was used as a non-silencing control siRNA (siCtrl). siRNAs (10 nM) were introduced in MDA-MB-231 or SUM-149 cells using Lipofectamine RNAi Max Transfection Reagent (Invitrogen) by reverse transfection according to the manufacturer’s protocol. Transfected cells were used 48 or 72 h later in functional assays.
miRNA microarray analysis and validation studies
MDA-MB-231 cells were transiently transfected with siADAM8-2 or siCtrl. Total RNA was isolated 72 h after transfection using TRIzol reagent (Invitrogen), according to the manufacturer’s protocol and quantified using a Take3 plate reader (BioTek Synergy HT). We chose 72 h post-transfection for functional assays to decrease effects introduced as a result of the transfection protocol. RNA integrity was assessed using the Agilent 2100 Bioanalyzer (Agilent) and RNA 6000 Nano LabChip kit (5065–4476). RNA with an RNA integrity number (RIN) >9.0 was considered good quality. Array analysis was performed using the TaqMan Array Human MicroRNA A + B cards v3.0 (Life Technologies, 4444913) according to the manufacturer’s protocol. Briefly, 1 μg total RNA was amplified and cDNA prepared using Megaplex Reverse Transcriptase (RT) Human Pool Set v3.0 primers (Life Technologies, 4444745) and TaqMan microRNA reverse transcription kit (Invitrogen, 4366596). Subsequently, 6 μl of the cDNA preparation was diluted with TaqMan Universal PCR Master Mix (450 μl) and nuclease free water (444 μl) and loaded on the 384-well TaqMan low-density array card. The cards were then centrifuged to distribute the cDNA samples in the reaction wells using a refrigerated Sorvall and Heraeus bucket centrifuge at 1200 rpm (331 × g) for two 1-min runs. The plate was then sealed and the real-time PCR was carried out on an Applied Biosystems 7900HT Real-Time PCR system. The data were analyzed using SDS software. Relative miRNA expression was calculated by comparing MDA-MB-231 cells with ADAM8 knockdown to cells with the control siRNA. This experiment was performed in duplicate and miRNAs that showed greater than twofold change in both replicates were selected for study.
Mammary fat pad mouse model
All animal work was performed in accordance with a protocol approved by the Institutional Animal Care and Use Committee of Tufts University and Tufts Medical Center. Blood was collected from 6-week-old female nonobese diabetic/severe combined immunodeficient (NOD/SCID) mice (Jackson Laboratory) using submandibular bleeding and plasma isolated and processed as described below. The next day, 2.5 × 106 shCtrl-3 MDA-MB-231 cells in 40 μl 50 % Matrigel (BD Biosciences, CB-40230A) solution (1:1 dilution of Matrigel with DMEM medium) were implanted in the fourth inguinal mammary fat pad. Primary tumor growth was monitored by caliper measurement twice a week. Tumor volumes were calculated as (length × width2)/2. At the indicated times, blood was collected and processed as above. Mice were sacrificed after 3–4 weeks when tumors derived from shCtrl-3 cells had reached a volume of ~1 cm3. Tumors were dissected, photographed, weighed and flash frozen.
Quantitative RT PCR analysis of cell RNA
RNA was isolated from cells using TRIzol reagent and DNA prepared using TaqMan MicroRNA Reverse Transcription Kit (Invitrogen, 4366596) according to the manufacturer’s protocol. Expression of miRNAs was assessed by quantitative reverse transcription PCR (RT-qPCR), using U6 snRNA (Invitrogen, 4427975) as the control. Single tube TaqMan assays (Invitrogen, 4427975) were obtained for all miRNAs of interest (hsa-miR-30d*, hsa-miR-181a-2*, hsa-miR-29c, rno-miR-29C*, hsa-miR-93*, hsa-miR-520c-3p, hsa-miR-130b*, hsa-miR-720, hsa-miR-106*b, hsa-miR-98, and hsa-miR-20a*) and qPCR was carried out as follows: 95 °C for 10 min, then 40 cycles of 95 °C for 15 s and 60 °C for 60 s. All analyses were performed in triplicate and the data were normalized to U6 snRNA. Average fold-change ± SD in miRNA levels relative to those in control untreated cells (set to 1) are presented.
Following isolation of mouse blood via submandibular bleeding, 25 μl of a 10-mM EDTA solution was added to individual samples to prevent coagulation. The samples were stored on ice and centrifuged at 1300 g for 20 min at 4 °C. Supernatants were collected and lysed as recommended in the manufacturer’s protocol for the miRneasy serum/plasma kit (Qiagen, 217184). Subsequently, 3.5 μl of C. elegans miR-39 (Qiagen, 219610) (160 nM) was added as a spike in each sample to control for miRNA recovery. RNA was then isolated as per protocol. Reverse transcription was carried out using the miScript II RT kit (Qiagen, 218161). miRNA expression was assessed by qPCR, and values normalized to the control C. elegans miR-39 (Ce_miR-39_1, MS00019789, Qiagen). The miScript primer assay was used for qPCR of miR-720 (Hs_miR720_1, MS00014833, Qiagen) as follows: 95 °C for 15 min, then 45 cycles of 94 °C for 15 s, 55 °C for 30 s and 70 °C for 30 s. Average fold-change ± SD in normalized miR-720 miRNA levels relative to those in control untreated mice are presented.
Exosome isolation and RNA extraction
Exosome isolation was performed essentially as published previously [
29]. Briefly, cell lines were cultured in ten P100 plates each until they reached 50–70 % confluency. The culture media for MDA-MB-231, shA8-20 and shCtrl-3 cells were then replaced with serum-free media, whereas, for SUM-149 cells which require serum for their viability, exosome-depleted FBS media was used. After 72 h, supernatants were collected, centrifuged at 2000 × g for 20 min to remove debris and filtered using a 0.22-μm filter. Exosomes were isolated by ultracentrifugation at 100,000 × g for 70 min. The exosome pellet was washed with PBS and RNA isolated using the miRCURY RNA isolation kit-cell and plant (Exiqon, 300110). Isolated RNA was subjected to Agilent 2100 Bioanalyzer (Agilent) analysis using a RNA 6000 Nano LabChip kit (5065–4476), which confirmed the lack of rRNA in the miRNA samples. Levels of miR-720 expression, determined by RT-qPCR, are presented ± SEM from two independent experiments.
miRNA knockdown and overexpression
MDA-MB-231 and SUM-149 cells were transfected in six-well plates with 50 nM miR-720 mirVana miRNA inhibitor (Invitrogen, 4464084 (hsa_miR_720, assay id: MH13574)) or negative-control (Invitrogen, 4464076) oligonucleotides using Lipofectamine RNAi Max (Invitrogen). After 16 h, fresh media was added and the cells were collected 48 h after transfection for functional analysis. To overexpress miR-720, ADAM8 knockdown MDA-MB-231 cells (shA8-20 clone) were plated a day prior to transfection. When cells reached 80 % confluency, cultures were transfected overnight with 200 pmol miR-720-mimic (Invitrogen, 4464066 (assay id: MC13574)) or the negative control mimic (Invitrogen, 4464058) using Lipofectamine 2000. The transfection media were replaced with fresh media and cells harvested 48 h after transfection for functional analysis.
Western blot analysis
Whole-cell extracts (WCEs) were prepared using RIPA buffer (50 mM Tris pH 7.6, 150 mM NaCl, 1 % NP-40, 0.1 % SDS, 5 mM EDTA, 1 % sodium sarkosyl) supplemented with protease and phosphatase inhibitors [
30], and 5 mM EDTA and 10 mM phenanthroline to inhibit the autocatalytic activity of ADAM8. Lysates were sonicated and centrifuged at 16,000 g for 15 min. Protein concentration in the supernatants was determined using DC Protein Assay Reagent (BioRad). Samples (25 μg) were subjected to immunoblotting as previously described [
27]. Either β-actin or β-tubulin was used as the loading control.
ATP assay
As a measure of cellular metabolism and therefore cell growth, ATP levels were assessed using an ATPlite luminescence ATP detection assay system (Perkin Elmer), as described previously [
31]. Briefly, cells (3000 cells/100 μl for MDA-MB-231 and shA8-20, or 8000 cells/100 μl for SUM-149) were plated in 96-well plates. After 24 h, an equal volume of APTlite 1Step luminescence reagent was added and luciferase activity measured. Average ATP levels, for triplicate samples, are presented relative to the control condition set to 1 (± SD).
Matrigel outgrowth, migration/invasion and transendothelial migration assays
Matrigel outgrowth assays were carried out as described previously [
32], using 5 × 10
3 cells plated in duplicate in 24-well plates. Cultures were incubated for 11–15 days and photographed at 10× magnification. Migration and invasion assays were performed in triplicate using polycarbonate filter Transwells (Costar) with 8-μm diameter pores, without precoating or precoated with growth factor-reduced Matrigel (BD Biosciences, 356231), respectively. For transendothelial migration, Transwells were coated with a confluent layer of HUVECs instead of Matrigel. Suspensions of 1 × 10
5 tumor cells were layered in the upper compartment of the Transwells and incubated at 37 °C. After 24 h, cells that migrated or invaded to the lower side of the filter were quantified by crystal violet staining and OD
570nm determination. The negative control or control mimic condition was set to 100 % and the mean ± SD from three independent experiments is given.
ELISA and RNA analysis of serum from TNBC patients
All clinical investigation was conducted in accordance with the principles outlined in the Declaration of Helsinki. Serum samples used in this work were provided by the Institut Régional du Cancer Nantes-Atlantique tumor bank, funded by the Institut National du Cancer (INCa) and approved by the French Minister of higher education and research (n°. AC-2008-141). Informed consent was obtained from patients to use their surgical specimens, serum, and clinicopathological data for research purposes, as required by the French Committee for the Protection of Human Subjects (CCPPRB). Ouest IV – Nantes CCPPRB approved use of serum samples for this study (6 May 2013: n°. 357/2013). This study did not need additional ethical approval.
Blood samples were collected in BD Vacutainer red-top tubes (BD Biosciences, 367837) from 37 consenting women diagnosed with TNBC in the Nantes Cancer Center (ICO, France). At the time of sampling, patients had not received any treatment in the form of surgery, chemotherapy, radiation, or endocrine therapy. Blood samples were allowed to clot for 1 h and centrifuged at 400 g for 10 min. Sera were stored at –80 °C within 2 h of being taken. Detailed patient clinicopathological characteristics are listed in Additional file
1 (Table S1). Sera were also collected from 15 consenting healthy female individuals. ADAM8 protein was measured in the serum using an enzyme-linked immunosorbent assay (ELISA) (R&D Systems) according to the manufacturer’s instructions.
Serum samples from patients (150 μl) were thawed to room temperature and processed as recommended for the NucleoSpin miRNA plasma kit (Invitrogen, 740981). Subsequently, 3.5 μl C. elegans miR-39 (Qiagen, 219610) (160 nM) was added as a spike in each sample to control for miRNA recovery. RNA was then isolated as per protocol. Reverse transcription was carried out using the TaqMan MicroRNA Reverse Transcription Kit (Invitrogen, 4366596). miRNA expression was assessed by qPCR, and values were calibrated to the control C. elegans miR-39 (Cel-miR-39-3p). Single tube TaqMan assays (Invitrogen, 4427975) for hsa-miR-720 were carried out as follows: 95 °C for 10 min, then 40 cycles of 95 °C for 15 s and 60 °C for 60 s. All analyses were performed in triplicate and the data were normalized to hsa-miR-16 (Invitrogen, 4427975). Serum levels of miR-720 in patients versus healthy individuals are presented as average fold-change ± SD.
Discussion
We show for the first time that ADAM8 signaling regulates a miRNA subset and that miR-720 specifically plays a critical role in the ability of this transmembrane protein to promote an aggressive phenotype in TNBC cells. In MDA-MB-231 and SUM-149 TNBC cells in culture, knockdown of ADAM8 led to decreased levels of miR-720 while ADAM8 ectopic overexpression increased miR-720 expression. Surprisingly, the induction of miR-720 did not require the metalloproteinase activity of ADAM8, as ectopic expression of either an ADAM8 mutant with a catalytically inactive MP domain or the remnant form lacking the MP domain was essentially as effective as full-length ADAM8 in promoting miR-720 expression. An antibody against the CRD/ELD domain that inhibits the DI activity of ADAM8 [
24] or an antagonist antibody against β1-integrin significantly reduced miR-720 levels, mapping the activity to the DI/CRD/ELD domains of ADAM8. The addition of an ERK-specific inhibitor significantly decreased the levels of miR-720. Together, these findings implicate activation of β1-integrin signaling by the DI/CRD/ELD region of ADAM8, and downstream activation of the ERK pathway in miR-720 induction. Significantly, knockdown of miR-720 using an antagomiR decreased the migratory and invasive phenotype of TNBC cells, whereas ectopic miR-720 expression restored these properties. We confirmed that miR-720 is secreted from TNBC cells, consistent with observations made in other breast cancer lines [
45], and increased miR-720 levels were detected in the blood of mice 7 days following orthotopic implantation of ADAM8-positive TNBC cells, when tumors were barely palpable. Higher levels of miR-720 were detected in the blood of TNBC patients with elevated amounts of circulating ADAM8. Thus, miR-720 is an essential mediator of ADAM8 in the promotion of the aggressive phenotype of TNBC cells, and as a secreted factor has the potential to function as a biomarker for early detection of ADAM8-positive recurrent TNBCs.
While miRNAs have been shown to play a vital role in breast cancer development and to regulate the functions of a number of critical genes [
33,
59,
60], very little is known about the stimuli and processes regulating the biogenesis of miRNAs themselves. Butz et al. [
61] and Wang et al. [
62] elucidated complex pathways downstream of TGF-β1 and c-MYC signaling, respectively, that modulated miRNA processing. For miR-720, work by Ragusa et al. [
63] suggested that its expression may be induced downstream of a MAPK/ERK pathway in colorectal cancer cells, although the mechanism of activation was not elucidated. Our results agree with and extend these findings, showing for the first time that ADAM8 initiates the induction of miR-720 via activation of the β1-integrin/ERK signaling cascade to promote migration and invasion in breast cancer cells. These findings shed light into the complex mechanism by which ADAM8 promotes aggressive phenotype of TNBCs. Many studies have implicated miR-720 in transformed phenotype or suggested its use in cancer diagnosis or as a prognostic marker for cancer [
46,
64,
65]. As discussed above, Lerebours et al. [
25] identified miR-720 in a set of five miRNAs that can serve as predictive markers of poor prognosis in patients with IBC, which we have found also frequently express ADAM8 (data not shown). Similarly, Park and coworkers reported upregulation of miR-720 in blood of patients with metastatic ER+/HER2- breast cancer [
66]. Consistently, miR-720 is upregulated in a variety of other tumors including colorectal and bladder cancers, malignant melanoma, renal cell carcinoma and multiple myeloma [
46,
63,
65,
67‐
71]. In contrast, Li et al. [
72] reported that miR-720 prevented a more aggressive phenotype of breast cancer cells, specifically via repression of synthesis of the EMT marker TWIST1 in MDA-MB-231 breast cancer cells. While potentially interesting, the expression levels of TWIST1, miR-720, and N-cadherin in the stock of MDA-MB-231 cells used in this study are inconsistent with NIH criteria established for this line, as judged by analysis of the NCI-60 cell line panel [
73,
74].
In addition to miR-720, many of the 68 miRNAs modulated by ADAM8 have been found to be upregulated in breast cancer or previously implicated in tumorigenesis (such as miR-19a, miR-106b, miR-181a-2, miR-30a, miR-93, miR-30d, and miR-10b) [
37,
40,
75‐
78]. Out of these 68 miRNAs, two (miR-324 and miR-7) were found to be downregulated by ADAM8. Interestingly, miR-7 is a potent tumor suppressor in breast cancer cells [
79] consistent with the observed ability of ADAM8 to repress its expression. Altogether, these data suggest that ADAM8, through β1-integrin and ERK activation, may regulate a large network of oncomiRs, which likely plays additional roles in promoting invasion and metastasis of TNBC tumors expressing high ADAM8 levels.
Stability of miRNAs in both primary tumors and in the blood (plasma and serum samples) makes them attractive potential biomarkers for non-invasive monitoring of cancer recurrence and for evaluating treatment efficacy. Recent studies have shown the importance of miRNAs as biomarkers in breast cancer [
80‐
87]. In particular, the level of miR-210 was found to be a good indicator of the sensitivity of breast cancer patients to trastuzumab [
80]. We have shown that levels of miR-720 are increased in the blood of mice bearing barely palpable ADAM8-positive tumors. Furthermore, elevated levels of miR-720 were seen in the serum of TNBC patients with high amounts of soluble ADAM8 protein. Although the miR-720 levels in serum of TNBC patients versus healthy individuals were statistically significant, they were not as elevated as the ADAM8 protein. This finding suggests that miR-720 needs to be used along with other tumor markers for detection of disease, and that miR-720 might be a better marker for recurrence where its levels in the serum under a disease-free state for a patient can be established. Notably, other miRNAs found modulated by ADAM8 in this study have been recently reported to help predict breast cancer risk and tumor relapse in TNBC patients [
88,
89]. For example, miR-18b, miR-20a, and miR-30d have been reported to be highly expressed in the serum of relapsing TNBC patients [
88]. Also miR-195 was upregulated in the serum of breast cancer patients in the study by Mishra et al. [
82], and served within a signature for early detection. Lastly, a recent endpoint study of patient serum samples showed that increased miR-720 levels can be detected in the whole blood of breast cancer patients with metastatic disease [
66], consistent with our data showing high ADAM8 expression in almost half of all metastases in breast cancer patients [
24]. Taken together with our data, these findings suggest that miR-720 holds potential for use as a biomarker along with a group of miRNAs regulated by ADAM8 and associated with metastasis for early diagnosis of recurrent TNBC or as a pharmacodynamic marker for treatment efficacy.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
SGD conceived the studies, designed experiments, carried out the miRNA profiling, performed the molecular and mouse studies, data analysis and interpretation, and drafted the manuscript. MR performed in vivo studies, helped with experimental design, data analysis, and drafting of the manuscript. NDM performed in vivo studies, helped with data interpretation, and writing and revision of the manuscript. SBN, PJ, and MC helped design the human study, collected blood samples from TNBC patients and healthy individuals, assembled clinicopathological data, and helped with interpretation of results and revision of the manuscript. GES contributed to the conception and design of the studies, data interpretation, and writing and revision of the manuscript. All authors read and approved the final manuscript.