1. Introduction
Metastasis is critical in the progression of breast cancer and is frequently associated with poor prognosis. Although strategies that inhibit metastasis will increase progression-free survival (PFS), the identification of therapeutic druggable targets that prevent metastasis remain in its infancy [
1]. Approaches for preventing metastasis require drugs that have
cytostatic rather than
cytotoxic properties, and are principally aimed at suppressing progression along the multistep metastasis pathway [
2].
The repurposing of beta-adrenergic receptor antagonists (beta-blockers) as an adjuvant therapy for the treatment of breast cancer has been proposed on the basis of their anti-metastatic properties [
3,
4,
5]. In vitro and in vivo models have demonstrated propranolol-induced inhibition of cancer cell signalling pathways decreases cell adhesion, migration, invasion, extravasation and colonisation in distant tissues including bone [
6,
7,
8], thereby leading to reduced metastasis [
9]. These pathways are triggered by catecholamine hormones such as norepinephrine acting on beta (β-) adrenergic G-protein coupled receptors (GPCR) expressed on breast cancer cells. Epidemiology studies observing the therapeutic potential of beta-blockers for treating breast cancer have revealed an association between (coincidental) beta-blocker usage and survival benefits [
10,
11,
12]. The clinical evaluation of propranolol as a neoadjuvant or perioperative treatment for breast cancer is on-going [
13,
14,
15,
16]. However, a recent contradictory study has reported no benefit between prescribed beta-blockers and survival [
17], whereas a different study using the basal-type MDA-MB-231 breast cancer cell line model showed that beta-adrenergic receptor (ADRβ2) agonism (rather than antagonism) inhibited tumour proliferation [
18]. Further studies are required to explain these discordant findings, which could result from variance in (a) in vitro cell line models; (b) patient cohort selected in pre-clinical studies; (c) pharmacologic selectivity of prescribed beta-blockers.
In this study, adrenoceptor expression and β2-adrenoceptor-mediated metastasis-associated cell behaviour were examined in three frequently used in vitro cell line models of ‘stress-induced’ triple-negative basal-type breast cancer and compared to a popular oestrogen-positive cell line model. β2-adrenoceptor-induced proteomic changes were assessed to better understand ADR-mediated cancer pathways, and provide biomarker and therapeutic treatment target identification. The study reveals complex and distinct differences between the cell lines and also identified a link between ADRβ2 signalling and LYPD3; revealing LYPD3 as a potential key mediator in ADRβ2 driven metastasis.
3. Discussion
Early population studies reported an association between beta-blocker use and survival in patients with breast cancer [
10,
11,
12]. The proposed physiological mechanism involved inhibition of β-adrenoceptor-activated cancer cell signalling pathways [
8,
19,
23,
24,
25,
26,
27,
28,
29,
30,
31], resulting in a reduction of metastasis [
9,
32,
33]. Specific mediators implicated downstream of this process have, to our knowledge, not been identified. Furthermore, the influence of adrenergic ligands on cell migration in vitro have been controversial [
7,
8,
9]. We found that two of the non-stimulated triple-negative basal-type breast cancer models, MDA-MB-231 and MDA-MB-468, showed significantly raised levels of β2-adrenoceptor gene and protein expression compared to the oestrogen responsive cell line. Although a challenge with a non-selective ADRβ2 agonist confirmed functional adrenoceptors in all four cancer cell lines, the strongest response was seen in MDA-MB-231 and MDA-MB-468, as evidenced by increased cAMP.
Although laboratory evidence supporting a clinical use for the beta-blocker propranolol in breast oncology is strong, it has not been universal, in that propranolol is reported by some to increase the proliferation of basal-type MDA-MB-231 breast cancer cells [
34,
35,
36,
37], but not by others [
19,
38]. We did not demonstrate a significant increase in survival in the breast cancer cells in response to a selective β2-adrenoceptor antagonist (ICI-118,551). A recent orthotopic mouse model of basal-type MDA-MB-231 breast cancer concurred that propranolol does not reduce primary tumour growth, but instead reduces the number and size of metastatic tumours [
9]. In summary, beta-blockers appear not to be cytotoxic
per se, but when combined with paclitaxel chemotherapy they appear to enhance its effectiveness, as demonstrated in a mouse basal-type (MDA-MB-231) breast cancer model [
27].
Herein, the association between ADRβ2 expression and key characteristics involved in metastasis was assessed in four breast cancer models, including cell migration, adhesion and invasion. Migration assays confirmed that the agonist norepinephrine increased the migration of MDA-MB-468, BT-549 and MCF-7 cells, which was abrogated by the ADRβ2 antagonist ICI-118,551. In contrast, norepinephrine reduced the migration of MDA-MB-231 cells, and ICI-118,551 increased their migration, as has been reported by Carrie and Sebti [
19]. However, when grown on collagen-1 or fibronectin protein-coated plastic or synthetic membranes coated with basement membrane protein, MDA-MB-231 cells showed a similar migratory behaviour to the other models (data not shown), thereby supporting the findings of others [
8,
39,
40]. In addition, two basal-type cell lines (MDA-MB-231, MDA-MB-468) and the ER-positive MCF-7 cells exhibited increased norepinephrine-induced invasive behaviour that was abrogated by the ICI-118,551 antagonist. These findings suggest ADRβ2 mediated cellular pathways could be influenced by cell adhesion molecules [
41], and may explain anomalous behaviour reported for the MDA-MB-231 cancer model when grown on different surfaces. This is further supported by the identification of proteins involved in cell adhesion, proliferation and migration following a mass spectrometry analysis of MDA-MB-468 cell lysates and secretome after ADRβ2 agonism/antagonism. The differences in migration observed between the MDA-MB-468 and MDA-MB-231 cells could, in part, be explained by differences in signalling arising because MDA-MB-468 cells are of epithelial origin (i.e., pre-epithelial-mesenchymal transition EMT), whilst MDA-MB-231 cells possess a mesenchymal-like phenotype (post-EMT). In MDA-MB-468 cells, norepinephrine could be signalling to initiate the acquisition of a more motile cell phenotype that would allow the cancer cells to spread, whereas in the MDA-MB-231 cells the presence of norepinephrine in the tissue microenvironment would signal that there is a favourable environment present for the cancer cells to successfully colonise and establish a secondary tumour/metastatic lesion [
42,
43,
44,
45]. Literature also demonstrates that catecholamines can both protect cancer cells from apoptosis [
33] as well as providing an advantageous environment whereby secondary tumours often become established at catecholamine-producing tissues such as the brain and adrenal glands [
46]. Furthermore, we hypothesise that the opposing influence of adrenergic stimulation on cell migration of MDA-MB-468 and MDA-MB-231 cell lines could be linked to ADR heterodimerisation. Lavoie et al. showed that heterodimerisation of ADRβ1 and ADRβ2 prevented agonist-induced internalisation of ADRβ2 [
47] and in our study we demonstrated that MDA-MB-468 cells expressed higher levels of the ADRβ1 receptor than MDA-MB-231 cells. Furthermore, pharmacological diversity could be introduced via heterodimerisation leading to differential desensitisation, changes in cell surface receptor expression and alterations to the functionality of ADR subtypes [
48,
49,
50,
51,
52,
53].
Clearly there are many conflicting results in the literature for the two-dimensional migration of MDA-MB-231 cells and its variants [
7,
8,
11,
36,
42]. There are a number of plausible reasons behind the differing results observed between the reports in the literature; including, the differing passage numbers of cells, different sources of the cell lines, the different methods and conditions used in performing the assays as well as time spans of the experiments. In addition, other reasons could be: (i) differing levels of β2-adrenoceptor signalling capacity in relation to the differing levels of β2 adrenoceptor expression and (ii) differing dose-dependent downstream signalling behaviours of the cell lines, including cAMP signalling. Not unsurprisingly, when the GO biological processes associated with the treatment of MDA-MB-231 cells with norepinephrine was investigated in our proteomic study, a number of genes associated with the following pathways were identified; proliferation, migration, adhesion, negative regulation of the apoptotic process and membrane organisation (
Supplementary Materials Table S1B). Furthermore, Madden et al. has also shown that MDA-MB-231 cells can possess an impaired cAMP signalling, with cAMP levels remaining high after stimulation by forskolin, which could have profound effects and could possibly be a reason for the observed differences with this cell line [
28]. In addition, Pon et al. demonstrated that the parental MDA-MB-231 cell line does not possess the feedforward Ca
2+/cAMP loop and therefore has low levels of cAMP production and a delayed response to stress [
25].
Furthermore, Kim et al. and Choy et al. used different variants of the parental MDA-MB-231 cell line, which could have profound influence on the in vitro behaviour of the cells in response to external stimulants such as isoproterenol [
8,
42]. Indeed, Pon et al. showed that the highly-metastatic variant of the MDA-MB-231 cell line, MDA-MB-231
HM are more responsive to β-adrenergic signalling than the parental MDA-MB-231 cell line and this is reflected in the limited cAMP signalling observed in the parental MDA-MB-231 cells compared to the high metastatic MDA-MB-231
HM variant [
25]. The two-dimensional migratory activity of MDA-MB-231 cells in response to isoproterenol has also been assessed on different extracellular matrix (ECM) protein surfaces where again, opposing effects were observed, even on the same surface, adding to the conflicting results of this cell line in the literature [
26,
43].
We report that the strongest association between high ADRβ2 expression and cell behaviours indicative of tumour aggression exists in two triple-negative basal-type breast cancer models (MDA-MB-231, MDA-MB-468). Moreover, we have identified distinctly different protein profiles between the two models in response to ADRβ2 activation, with differential expression of the pro-metastasis protein LYPD3 observed in MDA-MB-468 cells following ADR stimulation and/or ADRβ2 antagonism/inverse agonism. Stimulation of MDA-MB-468 cells with norepinephrine or isoproterenol increased the levels of LYPD3 within the cell lysate, and was also secreted in the media. Treatment with the antagonist ICI-118,551 reduced LYPD3 expression to a lower level than that observed following norepinephrine/isoproterenol stimulation both in the lysate and secreted into the media. LYPD3 is a glycosyl-phosphatidyl-inositol (GPI) anchored glycoprotein whose expression, we have confirmed, is highly restricted in normal tissues [
54,
55]. Studies have also demonstrated a strong association with a poor prognosis [
20,
22,
56,
57,
58] and even though its expression has been observed in breast cancer both in this study and in other published studies [
21,
59], its influence on tumourigenesis is yet to be elucidated although our patient tumour microarray (TMA) data suggests that LYPD3 is exclusively expressed in primary breast cancer and metastatic cases, with no expression observed in normal breast tissue. Furthermore, our in silico analysis has revealed that LYPD3 may be a therapeutic target in multiple cancer types, some of which have not been reported in the literature (testicular germ cell tumours and thymoma). Interestingly, upregulation of LYPD3 has been observed following cellular stress [
60], however, this is the first study showing that treatment of breast cancer cells with the endogenous stress hormone norepinephrine, can also lead to elevated LYPD3 levels.
In MDA-MB-468 cells, treatment with the non-selective ADRβ agonist, isoproterenol, unregulated LYPD3, whereas ICI-118,551, a selective ADRβ2 antagonist, reduced norepinephrine-induced LYPD3 expression and, when administered alone, reduced LYPD3 levels to that below the basal expression. This would suggest, through a mechanism yet to be elucidated, that LYPD3 is regulated via the ADRβ2 signalling pathway. In oesophageal cancer, LYPD3 is regulated via CREB (cAMP response element binding protein) transcription co-activator signalling [
61], and therefore, it is postulated that norepinephrine regulates LYPD3 through an ADRβ2/cAMP/PKA/CREB/LYPD3 effector pathway [
45]. In this pathway a conformational change in ADRβ2 would be induced following binding of norepinephrine, mediating activation of the G
sα protein subunit. Following activation, G
sα can then stimulate the adenylyl cyclase-induced conversion of adenosine triphosphate (ATP) into cAMP, resulting in protein kinase A (PKA) activation. Downstream this would enable phosphorylation of CREB via PKA, thereby inducing the transcriptional upregulation of LYPD3, and hence, increasing cell migration [
45,
61]. In this study we have demonstrated the importance of LYPD3 in breast cancer cell migration by successfully knocking down the expression of LYPD3 using LYPD3-specific shRNA. We found that cell migration, measured using transwell migration assays, was significantly reduced following shLYPD3 compared to shControl.
In the secretome of MDA-MB-468 cells, increased levels of LYPD3 and decreased levels of LAMC1 (Laminin Subunit Gamma 1) were also observed following treatment with norepinephrine and isoproterenol. This is an interesting observation because it is thought that LYPD3 can be cleaved from the cell surface by ADAM Metallopeptidase Domain (ADAM)-10 and -17 and this shedding can be induced by hypoxia [
62]. To speculate, once LYPD3 has been shed, it may still be able to bind laminin, via associating with the α6β4 integrin and matrix metallopeptidase-14 (MMP-14) and contribute towards its fragmentation and the observed decrease in LAMC1 that we observed in our proteomic analysis [
60,
63]. Alternatively, like its structural homologue urokinase-type plasminogen activator receptor (uPAR), shed-LYPD3 could function as a chemoattractant [
64].
In MDA-MB-231 cells, LYPD3 was not detected in the library of proteins generated from either the lysate or the secretome and this observation was confirmed by gene expression analysis of LYPD3 in a range of breast cell lines illustrating that MDA-MB-231 cells do not express LYPD3 at the mRNA level. LYPD3 was expressed at negligible levels in the BT-549, which is also a cell line of mesenchymal origin. Higher levels of expression of LYPD3 were observed in the epithelial-derived cell lines suggesting that LYPD3 expression may correlate with EMT. Furthermore, Harner-Foreman et al. published a spontaneous model of prostate cancer [
65] and unpublished mass spectrometry proteomic profiling data from this study revealed that LYPD3 was downregulated post-EMT. This is further supported by observations made by Oshiro et al., where significant associations were made between LYPD3 and EMT in both colorectal cancer cell lines and in clinical samples [
66].
4. Materials and Methods
4.1. Cell Lines
The following breast cancer cell lines were purchased from the American Tissue Culture Collection (ATCC): BT-549 (HTB-122™), ductal carcinoma; MCF-7 (HTB-22™), adenocarcinoma; MDA-MB-231 (HTB-26™), adenocarcinoma; MDA-MB-468 (HTB-132™), adenocarcinoma (Atcc.org, 2013). Breast cancer cell lines were cultured in DMEM (Dulbecco’s Modified Eagle Medium) (Lonza, Slough, UK) containing 10% v/v FCS (foetal calf serum) (GE Healthcare Life Sciences, Buckinghamshire, UK). All cells were grown at 37 °C, in a humidified atmosphere with 5% v/v CO2. After washing in PBS (phosphate-buffered saline), breast cancer cells were harvested using Trypsin and EDTA (ethylenediaminetetraacetic acid)l (Lonza, Slough, UK).
4.2. RNA Extraction, cDNA Synthesis and qRT-PCR
RNA was extracted using an RNeasy mini kit (QIAGEN, Manchester, UK), then quantified on a NanoDropTM 8000 Spectrophotometer. cDNA synthesis was performed using MMLV-reverse transcriptase (Promega, Southampton, UK) and oligo-dT primers (Promega, Southampton, UK). In brief, 1 μL oligo(dT)15 primer (Promega, Southampton, UK) was annealed to 2 μg of RNA. After being denatured for 5 min at 70 °C, a master mix solution was added, which contained: 0.7 μL Recombinant RNasin® Ribonuclease Inhibitor (Promega, Southampton, UK); 1 μL Moloney-Murine Leukemia Virus (M-MLV) reverse transcriptase (Promega, Southampton, UK); 5 μL M-MLV 5× reaction buffer (Promega, Southampton, UK); 1 μL Deoxynucleotide Triphosphate Solution Mix (Sigma, Dorset, UK) and 2.3 μL nanopure water. The samples were incubated for 60 min at 37 °C and then heated for 5 min at 95 °C. cDNA was then stored at −20 °C.
Semi-quantitative real-time PCR was performed using SYBR
® Green (Bio-Rad, Watford, UK) chemistry and gene-specific primers (MWG Eurofins, Ebersberg, Germany) (
Supplementary Materials Table S2) on a Rotor-Gene 6000 real-time PCR cycler (QIAGEN, Manchester, UK). PCR reactions were performed in 0.1 mL strip tubes containing a 12.5 μL mixture of: SYBR
® Green (Bio-Rad, Watford, UK); nanopure water; gene-specific primers (MWG Eurofins, Ebersberg, Germany) at a concentration of 5 pmol; and either 0.5 μL of cDNA or nanopure water (control). The following cycling conditions were used (35–40 cycles): initial 5 min at 95 °C for enzyme activation, followed by denaturation at 95 °C for 10 s, annealing for 15 s at the primer–specific Tm (
Supplementary Materials Table S2), and extension at 72 °C for 20 s. Following each PCR the melt curves were examined prior to data analysis. Using the Rotor-Gene Q Software and a threshold of 0.08, each transcripts Ct (cycle threshold) value was determined in triplicate. The relative expression of each target gene was then semi-quantified using the 2
−ΔΔCT method.
4.3. Flow Cytometry
Unconjugated antibodies were fluorescently labelled using Lightning-Link
® (Innova Biosciences, Cambridge, UK), according to the manufacturer’s protocol: ADRα1B antibody with PE-Cy7™; ADRα1D antibody with APC (Allophycocyanin) and β2-adrenoceptor antibody with RPE (R-Phycoerythrin). Harvested cells (1 × 10
5) were washed, pelleted and re-suspended in medium. Cells were then treated with an Fc receptor blocking reagent (Miltenyi Biotec, Bergisch Gladbach, Germany) diluted in PBS (10 min, 4 °C). Cells were incubated in the dark (30 min, 4 °C) with conjugated ADR antibodies (at a pre-optimised concentration) (
Supplementary Materials Table S3) and viable cells were identified using LIVE/DEAD™ fixable violet dead cell stain (Thermo Fisher Scientific, Leicester, UK). Cells (minimum 10,000) were analysed by flow cytometry (Beckman Coulter Gallios™, Kaluza™ software).
4.4. cAMP Signalling
Cells were treated with the non-selective β-agonist isoproterenol (1 µM, 10 min) and cAMP production measured using the cAMP ParameterTM ELISA kit (R&D Systems, Minneapolis, MN, USA). IBMX was incorporated to prevent cAMP degradation and phosphodiesterase activity on cAMP production.
4.5. Cellular Survival
Cells were seeded at 5000 cells/well in 100 µL of advanced DMEM containing 2% FCS and 4 mM L-glutamine. The cells were incubated at 37 °C, 5% CO2 in a humidified atmosphere for 24 h. After 24 h the media was carefully removed and the cells washed once with PBS. Fresh serum-free advanced DMEM was added to each well and the cells serum-starved for 24 h. After 24 h, the media was removed and the cells washed with PBS. 100 µL of advanced DMEM containing 2% FCS and 4 mM L-glutamine was adding to the wells with or without the addition of ICI-118,551 (β2-ADR selective antagonist, Sigma, Dorset, UK) (concentration range 1 pM–10 µM). Cells were treated with antagonists for 30 min prior to the addition of 10 µM norepinephrine (Sigma, Dorset, UK) for 72 h at 37 °C in a humidified atmosphere containing 5% CO2. After 72 h, 100 μL of 2× detection reagent was added to each well and the plate incubated for 60 min. The plate was then read using a Tecan Ultra fluorescent plate reader (Tecan Ultra, Mannedorf, Switzerland; excitation 485 nm, emission 535 nm).
4.6. Adhesion Assays
Cells were resuspended in 10 mL of serum-free media at 200,000 cells/mL in a falcon tube and were incubated for 30 min to allow recovery from detachment. 10 μM ICI-118,551 was added to the cells for 30 min prior to the addition of norepinephrine (at pre-optimised concentrations). 50 μL of serum-free media containing norepinephrine or media alone was added to each well of the 96-well plate and the plate was incubated for 30 min. After 30 min, 10,000 cells/well of cell suspension was added to each well of the 96-well plate [96-well plates pre-coated with human fibronectin (1.0 μg/well), human vitronectin (0.5 μg/well) (R&D Systems, Minneapolis, MN, USA), or collagen I (3 mg/mL) (Thermo Fisher Scientific, Leicester, UK)]. Untreated cells were added to wells containing 50 μL of pre-incubated media containing DMSO. Treated cells were added to wells containing 50 μL of pre-incubated media containing 10 µM norepinephrine. The plate was then incubated for 3 h at 37 °C in a humidified atmosphere containing 5% CO2. The media was removed from each well and the cells washed carefully twice with PBS. The final PBS wash was aspirated, the cells fixed with 4% paraformaldehyde in PBS and the plate incubated at room temperature for 20 min. The wells were washed twice with PBS. Cells were stained by adding 50 μL of crystal violet/cell stain solution and incubated at room temperature for 15 min. Wells were washed twice with deionised water and the last wash aspirated. The wells were allowed to fully dry at room temperature. Once dry, the wells of the plates were scanned using a C.T.L. ELISPOT plate reader and the number of remaining cells counted using ImmunoSpot® software (ImmunoSpot, Bonn, Germany).
4.7. Cultrex® Cell Migration Assay
Cells were serum-starved for 24 h prior to performing the assay. 1 × 106 cells/mL in serum-free media (MDA-MB-231, BT-549) or 0.5% FCS containing media (MDA-MB-468, MCF-7) were added to a 1.5 mL microtube along with 10 μM ICI-118,551. 50 μL/well of cell suspension (50,000 cells) with or without antagonists was added to the top chamber of the plate followed by the addition of 150 μL of DMEM containing 10% FCS to the bottom chamber of each well. The cells were incubated for 30 min prior to the addition of norepinephrine (10 μM). The plate was incubated for 24 h at 37 °C in a humidified atmosphere containing 5% CO2. After incubation, the top chamber was inverted and carefully shaken to remove the culture medium and transferred to the black receiver plate. Each well of the top chamber was washed with 100 μL of warm 1× wash buffer and the top chamber inverted and carefully shaken to remove excess wash buffer and placed back into black receiver plate. To the bottom chamber of each well is added 100 μL of 1000× cell dissociation solution/ Calcein AM solution and the plate incubated at 37 °C in a humidified atmosphere containing 5% CO2 for 60 min. The plate was read on a fluorescent plate reader at 485 nm excitation, 520 nm emission (Tecan Ultra, Mannedorf, Switzerland).
4.8. CultreCoat® Medium BME Cell Invasion Assay
Cells were serum-starved for 24 h prior to performing the assay. 1 × 106 cells/mL in serum-free media (MDA-MB-231, BT-549) or 0.5% FCS containing media (MDA-MB-468, MCF-7) were added to a 1.5 mL microtube along with 10 μM ICI-118,551. 25 μL/well of cell suspension (25,000 cells) were added to the top chamber of the plate followed by the addition of 150 μL of DMEM containing 10% FCS to the bottom chamber of each well. The cells were incubated for 30 min prior to the addition of 10 μM norepinephrine. The plate was incubated for 24 h at 37 °C in a humidified atmosphere containing 5% CO2. After incubation the top chamber was washed with 100 μL of warm 1× wash buffer and placed into a black receiver plate. To the bottom chamber of each well 100 μL of 1000× cell dissociation solution/Calcein AM solution is added and the plate incubated at 37 °C in a humidified atmosphere containing 5% CO2 for 60 min to fluorescently label and dissociate cells from the membrane. After 60 min, the top chamber was removed and fluorescence was measured using a plate reader (Tecan Ultra, Mannedorf, Switzerland; 485 nm excitation, 520 nm emission).
4.9. Proteomic Analysis
MDA-MB-231 (1.3 × 106) and MDA-MB-468 (1.2 × 106) cells were seeded and grown to 90% confluency. On the day of treatment 1 mM stocks of norepinephrine, isoproterenol and ICI-118,551 were made up in serum free DMEM and passed through a 0.22 µm filter prior to serial dilution. Media was removed from each flask, and cells were washed three times with PBS. Untreated cells received 25 mL of serum free DMEM, and treated cells received 25 mL of DMEM containing 10 μM of the appropriate treatment condition. After 30 min incubation 10 µM norepinephrine or 10 µM isoproterenol was added into the relevant flasks. Flasks were then incubated at 37 °C, 5% CO2 for 24 h. Twenty four hours after treatment secretome samples were obtained by removing the media and centrifuging (300× g, 5 min), filtered (0.22 µm) and concentrated using Amicon Ultra-15 Centrifugal Filter Units (Merck, Kenilworth, NJ, USA). Cell lysates were prepared using 9.5 M urea (Melford, Stowmarket, UK)/2% v/v dithiothreitol (Melford, Stowmarket, UK)/1% v/v n-octyl-beta-glycopyranoside (Apollo Scientific Limited, Stockport, UK) containing protease inhibitor (Sigma, Dorset, UK). The lysates were harvested, chilled on ice for 5 min, sonicated for 5 min and this process was repeated three times before centrifugation for 10 min at 12,000× g and storage at −80 °C.
Cell lysate (100 µg) and secretome (100 µg) were diluted in 50 mM tri-ethyl ammonium bicarbonate (TEAB) before reduction (5 mM DTT at 56 °C for 20 min) and alkylation (15 mM iodoacetamide at room temperature for 15 min) and then digested for 16 h using Trypsin/Lys-C (Promega, Southampton, UK) at 37 °C at a 20:1 protein:protease ratio (
w/w) in a thermomixer (650 rpm) [
67]. Next, samples were cleaned up using HyperSep C
18 cartridges (Thermo Scientific, Leicester, UK) for solid phase extraction. A vacuum concentrator was then used to concentrate the samples before resuspension in 5% acetonitrile + 0.1% formic acid and subsequent analysis of the peptides on an AB Sciex TripleTOF 5600+ MS/MS instrument in both SWATH (Sequential Window Acquisition of All Theoretical Mass Spectra) and IDA (information dependent acquisition) acquisition modes (Sciex, Framingham, MA, USA).
4.10. Mass Spectrometry
Each sample was analysed on a Sciex TripleTOF 5600+ mass spectrometer coupled in line with an Eksigent ekspert nano LC 425 system running in micro flow (5 µL/min) mobile phase B (100% acetonitrile + 0.1% formic acid) over mobile phase A (0.1% formic acid). In brief, 8 µg of sample was injected and trapped onto a YMC Triart-C18 pre-column (5 mm, 3 µm, 300 µm ID) at a flow rate of 10 µL min for 2 min. The sample was then eluted off the pre-column and onto a YMC Triart-C18 analytical column (15 cm, 2 µm, 300 µm ID) that was in line with the Sciex TripleTOF 5600+ Duospray Source using a 50 µm electrode in positive mode, +5500V. The following linear gradients were used: for SWATH, mobile phase B increasing from 3% to 30% over 38 min, 30% to 40% over 5 min, 40% to 80% over 2 min for wash and re-equilibration (total run time 57 min). For IDA, mobile phase B increasing from 3% to 30% over 68 min, 30% to 40% over 5 min, 40% to 80% for column wash and re-equilibration over 2 min (total run time 87 min). Data independent acquisition was performed using 100 variable SWATH windows (optimised on sample type) (TOFMS m/z 400-1250) 25 ms accumulation time; 2.6 s cycle and IDA with a top 30 ion fragmentation (TOFMS m/z 400-1250) followed by 15 s exclusion using rolling collision energy, 50 ms accumulation time; 1.8 s cycle.
Spectral library generation, alignment and fold change analysis were performed as described previously [
68]. In brief, IDA data were searched using ProteinPilot 5.0 (iodoacetamide alkylation, biological modifications emphasised in a thorough search) against the Swiss-Prot human database (June 2017). The Sciex OneOmics software was used to analyse the SWATH data following extraction against the locally generated library with the following parameters; 12 peptides per protein, six transitions per peptide, XIC width 30 ppm, 6 min retention time window.
4.11. LYPD3 Protein Expression
Mass spectrometry and curated protein profiling of the two basal-type cell lines (MDA-MB-231 and MDA-MB-468) identified increased LYPD3 expression, a pro-metastasis protein. LYPD3 protein expression in breast cancer tissue was confirmed by immunostaining three wax-embedded TMA slides (BC081120b, BC10010d, BR1201) comprising 260 cases of invasive ductal adenocarcinoma, 50 cases of metastatic adenocarcinoma and 10 cases of normal breast tissue (U.S Biomax, Rockville, MD, USA).
Sections were immunostained using a monoclonal rabbit anti-human LYPD3 antibody (
Supplementary Materials Table S3) on a BenchMark ULTRA stainer (Ventana Medical System, Inc, Oro Valley, AZ, USA) with ultraView Detection Kit (Ventana Medical, Oro Valley, AZ, USA).
The intensity of LYPD3 immunostaining was microscopically assessed for cytoplasmic staining using a five-point scoring technique, where a score of 0 represented nil staining; a score of 1: weak; 2: moderate; 3: strong; 4: strong cytoplasmic staining with additional cell membrane staining.
The association between LYPD3 expression and clinical variables in malignant adenocarcinoma, metastasis and adjacent normal breast tissue was statistically tested using SPSS (Version 24, IBM, UK). Immunohistochemistry scores were dichotomously categorised (0, 1 = negative; 2, 3, 4 = positive) for Chi-square tests. Significance levels were p = 0.05 or less.
4.12. Western Blot
Total cell lysates were used from the proteomic analysis. In brief, 50 µg of total protein from each sample was prepared with sample reducing buffer (60 mM Tris-HCl (pH 6.8), 2% SDS (sodium dodecyl sulphate), 10% glycerol and 0.01% bromophenol blue) at a ratio of 1:3 sample vs. reducing buffer. The sample was resolved on an SDS gel (10% resolving gel, 5% stacking gel) with Tris/glycine/SDS gel running buffer (Geneflow) at a constant voltage of 150 V. After separation, samples were transferred onto nitrocellulose membranes using Tris/glycine/methanol transfer buffer at a constant current of 180 mA for 75 min at 4 °C. Membranes were blocked in 10% Marvel™ dried skimmed milk powder for 1 h before being probed with rabbit anti-LYPD3 antibody (1:1000, ab151709, Abcam, Cambridge, UK), rabbit anti-beta 2 adrenergic receptor antibody (1:1000, ab182136, Abcam, Cambridge, UK) and rabbit anti-beta actin antibody (1:5000, ab8227, Abcam, Cambridge, UK) overnight at 4 °C. The membranes were then washed and goat anti-rabbit IgG HRP-linked antibody was added (1:1000, Cell Signalling Technology, London, UK). Membranes were washed and exposed to the Clarity Western ECL Substrate (1:1) and imaged using a Syngene G:Box and Genesys v1.5.4.0 software (Syngene, Cambridge, UK).
4.13. Generation of LYPD3 Knockdown Cell Line
Lentiviral shRNA plasmids and packaging mix (SHP001) were purchased from Sigma (Dorset, UK): shLYPD3 (catalogue number: SHC204) and shControl (catalogue number: SHCLNV-NM_133743). MCF-7 cells (2.5 × 105) were transfected with shRNAs and hexadimethrine bromide (Sigma, H9268) at a final concentration of 8 µg/mL for 18 h. Resistant colonies were selected using media containing 2 µg/mL puromycin.
4.14. In Silico Gene Expression Profiling
LYPD3 gene expression profiling was performed using patient gene expression profiles generated through The Cancer Genome Atlas (TCGA) [
69] and the Genotype-Tissue Expression (GTEx) project [
70] and data was assessed via the Gene Expression Profiling Interactive Analysis (GEPIA) [
23].