Background
Breast cancer is a highly heterogeneous disease with subtypes based on hormone receptors, oestrogen or progesterone receptors (ER/PR) and HER2 overexpression. More recently, gene expression profiling led to identification of five main molecular subtypes of breast cancer: HER2 overexpression (ER−/PR−/HER2+), basal-like (ER−/PR−/HER2−/basal marker+), luminal A (ER+/PR+/HER2−/KI67-), luminal B (ER+/PR+/HER2−/KI67+ or ER+/PR+/HER2+/KI67+) and normal-like (ER+/PR+/HER2−/KI67-) [
1]. Further subtypes have also been identified based on integrative analysis of gene expression and copy number, suggesting increased complexity of breast cancer heterogeneity [
2]. Despite major breakthroughs in the treatment of breast cancer over the last twenty years, there is still a significant number of patients who do not respond, develop resistance to therapy, or experience tumour recurrence; late relapse in ER+ breast cancer continues to be a particular issue. There is now a plethora of evidence to suggest that cancer stem cells (CSCs) are responsible for the incidence of metastatic disease which is the main cause of death in patients with breast cancer [
3]. Triple negative breast cancer or basal-like subtype constitutes around 20% of breast cancer cases and it is highly metastatic with limited therapeutic options [
4]. Chemotherapy remains the only treatment option for this disease subtype. The chemotherapy resistant CSC population has increased metastatic potential in triple negative breast cancer through activation of oncogenic pathways such as STAT3, therefore there is an urgent need for new therapeutic options which target CSCs [
5,
6]. On the other hand, the most common type of breast cancer, ER+ (luminal A/B or normal-like), is treated with endocrine therapy in both the adjuvant and metastatic settings [
7]. Tumour recurrence in endocrine-resistant breast cancer patients leads to a more aggressive type of breast cancer with enhanced metastatic ability [
8]. In patients treated with neoadjuvant letrozole, CD44
+/CD24
− mammosphere forming cells, representative of CSCs, were increased and the remaining tumour cells appeared to have a mesenchymal phenotype consistent with the more aggressive, basal-like type of breast cancer [
9]. This acquired endocrine therapy resistance has been attributed to the activation of survival pathways such as the epidermal growth factor receptor (EGFR) pathway and, more recently, the Notch pathway [
10]. The Notch 4 receptor, in particular, regulates breast CSC activity [
11] and it is also implicated in endocrine therapy resistance in women treated with tamoxifen [
12,
13]. Furthermore, tumour and plasma levels of the Notch 1, 2, or 4 receptors and DLL4 ligand were positively correlated with nodal and distant metastases in breast cancer and shorter disease-free or overall survival compared to patients with high DLL4 levels [
14,
15]. In relevant in vitro and in vivo cancer models, DLL4 has also been implicated in chemoresistance [
16], tumour angiogenesis [
17] and CSC activity [
18]. Therefore, all of these studies suggest that DLL4 and Notch 4 are viable therapeutic targets for both triple negative and ER+ breast cancer treatment.
FK506-binding protein like (FKBPL) is a novel anti-tumour protein that belongs to the family of immunophilins, but is a divergent member lacking peptidyl prolyl isomerase activity [
19]. Immunophilins orchestrate protein-protein interactions therefore regulating many cellular processes including cell signalling, differentiation, cell cycle progression, metabolic activity and apoptosis [
20]. FKBPL has diverse anti-tumour roles both as an intracellular and extracellular protein. Intracellular FKBPL regulates ER signalling and, as such, has prognostic value in terms of breast cancer survival. This was demonstrated using publically available datasets [
21] and in a meta-analysis of five independent breast tissue microarray (TMA) cohorts [
21,
22]. In this cohort of 3277 patients, FKBPL was a significant and independent predictor of breast cancer specific survival (BCSS), with low FKBPL expression being associated with shorter BCSS (HR = 1.31, 95% CI 1.15–1.50,
p < 0.001). Likewise, in a cohort of 2365 ER+ breast cancer patients, low FKBPL expression had also a significantly shorter BCSS compared to high FKBPL expression (HR = 1.34, 95% CI 1.13–1.58,
p < 0.001) [
21]. Similarly, RBCK1, an E3 ubiquitin-protein ligase, which regulates FKBPL levels, also demonstrated a potential role as a prognostic and predictive biomarker of response to endocrine therapy in breast cancer patients in terms of BCSS [
23].
In addition to this intracellular role, FKBPL’s extracellular anti-angiogenic and anti-CSC roles were identified, potentially through its ability to target CD44 [
5,
24,
25]. Upregulation of CD44 is associated with angiogenesis, stemness, tumourigenicity and cell migration [
26]. The ‘first-in-class’ FKBPL-based peptides, AD-01 (24-amino acid pre-clinical therapeutic candidate) and ALM201 (23-amino acid clinical therapeutic candidate which has successfully completed a Phase Ia clinical trial [EudraCT 2014–001175-31]) [
27], have also demonstrated strong anti-angiogenic and anti-CSC effects [
5,
24,
25]. The anti-CSC activity of AD-01 led to downregulation of stem cell markers, Nanog, Oct4 and Sox2 in breast cancer cell lines while the intratumoural knockdown of FKBPL in a ZR-75 breast cancer xenograft mouse model increased the expression of Nanog/, Oct4 and Sox2 [
5]; Sox2 has been implicated in both metastasis and endocrine therapy resistance [
28‐
30]. Therefore, since FKBPL and its peptides have demonstrated inhibitory effects on angiogenesis [
24,
25], CSC signalling [
5] and ER signalling [
21,
22], we hypothesised that FKBPL could also inhibit metastasis and endocrine therapy resistance driven by CSCs in breast cancer. Here, we show for the first time, that FKBPL and its therapeutic peptides reduce metastatic burden in a triple negative breast cancer model and inhibit endocrine therapy resistant CSCs, thereby reducing tumour initiation, in ER+ disease. Furthermore, we elucidate additional targets of FKBPL such as DLL4 and Notch 4, which in addition to CD44, are potentially involved in the multiple anti-tumour effects of FKBPL and its therapeutic peptide derivatives.
Methods
Cell culture
All cells were obtained from the American Type Culture Collection, authenticated by short-tandem repeat (STR) profiling carried out by the suppliers, and verified as mycoplasma-free. MDA-MB-231 CD44 stable knockdown (KD) cells were a gift from Prof. David Waugh (QUB) [
31]. The MCF-7 and MDA-MB-231 cell lines were cultured in Dulbecco’s modified Eagle medium (DMEM; Life Technologies, UK) supplemented with 10% foetal calf serum (FCS; GE Healthcare, UK). Cells stably overexpressing FKBPL (D2 from parental cell line, MCF-7, and A3 from parental cell line, MDA-MB-231) were selected using 750 μg/mL G418 (Sigma, UK) and grown in the presence of 375 μg/mL (3.1D2) or 750 μg/mL (A3) G418 (Sigma, UK) as previously described [
5]. Cell culture experiments were carried out at 37°C in a humidified atmosphere of 95% O
2/5% CO
2.
Boyden chamber assays
A Boyden chamber assay was used to examine cell migration and invasion. MDA-MB-231 cells were treated with AD-01 (1 nM) for 24 h. Following 24 h cells were trypsinized and re-suspended, (1.0 × 104 cells in 200 μl RPMI-1640 medium) and then placed into the uncoated (for migration) or Matrigel coated (for invasion) upper chambers (8-mm pore size; Millipore, USA). The lower chambers were filled with 600 μl complete medium with 10% FBS. After incubation for 12 h at 37 °C, non-invading cells were removed from the top of the chamber with a cotton swab. The invaded cells on the lower surface of the inserts were fixed and stained with 0.1% crystal violet, and five random fields for each insert were counted at 200× magnification.
Primary samples
Solid breast tumour mastectomy samples or core biopsies treated in the neoadjuvant setting with letrozole were collected from patients with fully informed consent (NIB14–0117; Northern Ireland Biobank), cut into small pieces (1 mm), and digested for 2 h on a rotating platform in RPMI (Gibco, UK) containing 10% collagenase/hyaluronidase (Stem Cell Technologies, UK). Following tissue digestion, filtration through 70 μm and 40 μm cell strainers (BD Technologies, UK) was carried out and 500 cancer cells per cm
2 were seeded in the mammosphere medium DMEM-F12 (Gibco, UK), containing B27 minus vitamin A (Life Technologies, UK), 20 ng/ml EGF (Roche, UK), PenStrep (Invitrogen, UK) ± ALM201 (100 nM) as previously described [
5]. Frozen pleural effusion samples collected from the patients with metastatic breast (
n = 3) with fully informed consent (COREC# 05/ Q1403/25 and 05/Q1403/159; Division of Cancer Sciences, Manchester, United Kingdom) were defrosted, cells counted and seeded in the mammosphere assay for 72 h ± ALM201 (100 nM) as previously described [
5].
Treatments
1.2 × 104 MCF-7 cells were plated in a monolayer in complete medium for 24 h. The medium was replaced by DMEM-F12 containing 10% charcoal-stripped serum medium and 17β-estradiol (100 nM; Sigma, UK) was added to all wells except for the control well. Tamoxifen (1 μM; Sigma, UK) and ALM201 (1 nM) were added alone or in combination for 72 h and cells incubated at 37°C in a humidified atmosphere of 95% O2/5% CO2. In a separate experiment, MCF-7 and MDA-MB-231 cell monolayers were treated with AD-01 (100 nM) or ALM201 (100 nM) before being used in mammosphere assays, western blotting or quantitative real-time polymerase chain reaction (qRT-PCR).
Mammosphere assay
A single cell suspension was prepared following enzymatic (0.125% Trypsin-EDTA (Invitrogen, UK)) and manual disaggregation and 500 cells/cm
2 were seeded in low adherent culture 6-well plates (VWR, UK) coated with 1.2% poly-HEMA (Sigma-Aldrich, UK) in mammosphere medium at 37°C in a humidified atmosphere of 95% O2/5% CO
2 for 5–7 days as described previously [
5].
Flow cytometry
MCF-7 and MDA-MB-231 were grown in a cell monolayer or as mammospheres for 72 h before cells were disaggregated and incubated with pre-conjugated primary antibodies BEREP4-FITC (1:10; Dako), CD44-APC (1:20; BD Pharmigen), and CD24-PE (1:10; Beckman Coulter) as previously described [
11]. Fluorescence was measured using BD FACSCanto II and analyzed by WinMDI 2.9.
Clonogenic assay
MCF-7, 3.1D2, MDA-MB-231 and A3 cells were plated at a density of 50 or 100 cells/cm2 per well in a six well plate containing DMEM + 10% FCS medium and incubated for 10 days at 37°C in a humidified atmosphere of 95% O2/5% CO2. Following incubation the medium was removed, colonies were fixed with 1% crystal violet/70% ethanol and holoclones/meroclones/paraclones counted manually.
Western blotting
MDA-MB-231 or MCF-7 cells were treated with ALM201 or AD-01 (100 nM) for 24 h before cells lysates were prepared using Laemmli buffer (Sigma, UK) and subjected to western blotting as reported previously [
25]. Primary antibodies used included: DLL4 (Abcam, UK, cat: ab7280; 1:500), Notch4-ICD (Abcam, UK, cat:ab33163; 1:400), FKBPL (Proteintech, USA cat: 10060–1-AP; 1:1,000), CD44H (R&D Systems, USA, cat: BBA10; 1:1,000), GAPDH (Sigma, UK; cat: G9545; 1:10,000). HRP-linked secondary antibodies were either anti-mouse or anti-rabbit (GE Healthcare, UK; 1:10,000). Densitometry was performed using ImageJ software (NIH, USA) and adjusted to GAPDH.
Quantitative real-time PCR
Following treatment of the adherent cells, as described above, RNA was extracted using GeneJET RNA purification kit (Fisher Scientific, UK) according to manufacturer’s instructions and RNA was quantified using a Nanodrop spectrophotometer (Thermo Fisher Scientific, Basingstoke, UK). Complimentary DNA (cDNA) was produced using Transcriptor first stand cDNA synthesis kit (Roche, Herefordshire, UK) according to manufacturer’s instructions. qRT-PCR was performed using the Lightcycler 480 PCR machine (Roche, UK). All Taqman primer probe sets were supplied by Roche (DLL4, cat:100073803; GAPDH, cat: 100065048; β-Actin, cat: 100063228).
In one set of experiment, 8–12 week old in-house bred female SCID mice (C.B-17/IcrHsd-PrkdcscidLystbg) were selected at random and pre-treated subcutaneously (s/c) once daily (a.m.) with AD-01 (0.003 or 0.3 mg/kg/day, n = 5) or PBS (n = 6) for one week prior to injection with 5 × 105 MDA-MB-231-lucD3H1 cells, followed by continuation of treatment with AD-01 or PBS for a further 28 days. Lung cell load was assessed following i.p. injection of luciferin (150 mg/kg) on day 0 when mice were inoculated with cells, then lung metastatic colonization was assessed at day 28,), using non-invasive bioluminescence of total photon flux. In the second experiment, MDA-MB-231-LucD3H1 cells were grown in a monolayer and treated with AD-01 (1 nM) for 1 day before 8–12 week old female SCID mice were inoculated intravenously with 4 × 105 pre-treated or mock (PBS) treated MDA-MB-231-LucD3H1 cells. Following inoculation, mice with detectable lung metastasis deposits were treated with control (PBS, n = 5) or AD-01 (0.3 mg/kg/day, n = 5 and 0.003 mg/kg/day, n = 5) for 26 days via i.p. injection. On day 26, primary experimental outcome i.e. lung metastatic colonization was assessed using non-invasive bioluminescence of total photon flux. At the end of the experiment, mice were euthanized by the carbon dioxide method. One-way ANOVA with post-hoc Dunnett’s multiple comparisons statistical test was used to compare the metastatic burden between control and the two treatment mice groups. All animals were of a similar weight (approx. 20 g) at the start of the experiments; weight and animal wellbeing was monitored at least twice weekly. Mice were housed in a group of up to 5 per cage in special SPF cages which included autoclaved bedding material. All in vivo procedures were carried out at the Biological Resource Unit at Queen’s University Belfast.
Limiting dilution in vivo assay
MCF-7 cells (5 × 106) were implanted intradermally into 8–12 week old in-house bred female SCID mice bearing oestrogen pellets (0.25 mg). Once MCF-7 xenografts were established (100–150 mm3), the following treatments were administered to randomly selected mice once daily (a.m.): 1) vehicle control via oral gavage (100μl) and PBS s/c (100 μl; n = 6), 2) tamoxifen citrate (Sigma, Cambridge, UK) via oral gavage (250 μg/100 μl; n = 4), 3) ALM201 s/c (0.3 mg/kg/day; n = 4) and 4) tamoxifen citrate via oral gavage (250 μg/100 μl) and ALM201 s/c (0.3 mg/kg/day; n = 4). The treatments were administered for the duration of 21 days and tumours were measured every 3 days. Following three weeks of treatment, mice were euthanized using the carbon dioxide method, tumours were excised, disaggregated and used for ex vivo mammosphere assays or intradermal re-implantation into secondary (untreated) female SCID mice at 5 × 105 cell concentrations per mouse (control, n = 16; tamoxifen only, n = 15; ALM201, n = 7; tamoxifen plus ALM201, n = 6). The primary experimental outcome, i.e. time taken for tumour initiation, was recorded. The secondary experimental outcome was the number of mammospheres formed from tumours ex vivo from each group. One-way ANOVA with post-hoc Dunnett’s multiple comparisons statistical test was used to compare tumour initiation and mammosphere content between control and the three treatment groups.
Statistical analysis
Data presented are a mean of at least 3 independent experiments ± SEM. Primary sample data are from one patient; statistics were performed on 3–6 replicates. One-way ANOVA or t-test were used to assess differences between control and treatment groups. For multiple comparisons post-hoc Dunnett’s multiple comparison test was used. Statistical significance was determined by the P values less or equal to 0.05; *, P < 0.05; **, P < 0.01; ***, P < 0.001.
Discussion
We have previously demonstrated a role for FKBPL in ER signalling, endocrine therapy response, angiogenesis and CSC differentiation [
5,
22,
24]. To date, the mechanism of action has been attributed to a potential role in the CD44 pathway and stabilisation of p21 [
5,
22,
36]. In addition to this, we have shown that high FKBPL levels are associated with a positive prognosis in breast cancer [
21]. In this study, for the first time, we assessed the pre-clinical activity of novel systemic anti-cancer therapeutic peptides, ALM201 & AD-01, in the metastatic setting, and highlighted their impact on endocrine therapy resistant cancer stem cells; both areas of unmet clinical need. These effects were demonstrated using a range of experiments with cell lines, primary breast cancer samples and in vivo models.
In triple negative breast cancer using MDA-MB-231 cells, we demonstrated FKBPL-mediated differentiation of CSCs to more “mature” cancer cells, no cytotoxic effect and inhibition of cell migration, invasion and metastasis. In our in vivo lung metastasis model we demonstrated that pre-treatment with AD-01 prevents lung colonization of breast cancer cells which is, likely, through prevention of engraftment of the tumour cells given AD-01’s inhibitory effect on cell migration and invasion. In ER+ breast cancer, using MCF-7 cells and ER+ breast cancer samples, we also demonstrated FKBPL-mediated CSC differentiation, inhibition of CSCs resistant to endocrine therapy and delay in tumour initiation. Interestingly, ALM201 in combination with tamoxifen appeared even more effective at inhibiting CSC population than ALM201 alone while tamoxifen shows no effect on CSCs. Furthermore, FKBPL appears to downregulate DLL4 and Notch 4 levels, which has not been previously reported. Therefore, we identified a novel role for FKBPL in reducing the metastatic burden which could be linked to the inhibition of CSCs and the regulation of CD44, DLL4 and Notch 4. This is very important since other anti-angiogenic agents show increased metastatic potential [
34]. CSCs have been implicated in cancer metastases, as the primary cells likely to migrate and populate metastatic sites, due to their strong migratory and pluripotent potential [
37]. High Notch activity has been implicated in cancer pathogenesis and Notch 4 is specifically active within breast CSCs [
11,
38]. Moreover, both Notch and CD44 have been implicated in hypoxia-driven enrichment of CSC population, tumour recurrence and enhanced metastatic phenotype after treatment with anti-angiogenic agents or hypoxia inducible factors [
39‐
41]. Our data suggests that FKBPL-based peptides in addition to their well-established anti-angiogenic [
24,
25] and anti-CSC activity [
5] via CD44, are able to inhibit lung metastasis, possibly by modulating the Notch pathway members, DLL4 and Notch 4, within breast cancer, giving these agents a potential competitive advantage. Further studies would be required to elucidate the role of FKBPL/ALM201/AD-01 in Notch 4 and DLL4 signalling.
Furthermore, our in vivo data in relation to tamoxifen treatment confirms that tamoxifen does not target CSCs or inhibit tumour initiation. Conversely, ALM201 alone or in combination with tamoxifen demonstrated a substantial delay in tumour initiation and reduced the proportion of the CSC-like population assessed by
ex vivo mammosphere assay, which correlates with the content of CD44
+/CD24
− CSC population. The combination of tamoxifen and ALM201 had a more pronounced inhibitory effect on tumour initiation and the CSC-like population compared to ALM201 alone, thus suggesting that this combination might be advantageous clinically. Notch inhibitors have already demonstrated activity in combination with tamoxifen, and Notch 4, in particular, has been implicated as a viable target to prevent metastasis in tamoxifen-resistance breast cancer [
42,
43]. Nevertheless, correlation between the activity of Notch ligands, receptors and target genes is complex and elucidating the functional role for individual Notch receptors and ligands in mediating resistance to therapy, tumour recurrence or metastasis in breast cancer is necessary [
44,
45]. Our data suggests that FKBPL can specifically downregulate DLL4 and intracellular Notch 4, however the effects on other important members of the Notch pathways and Notch signalling needs to be investigated further.
In summary, based on the results obtained in this study and previously published studies, while the novel FKBPL-based anti-cancer therapeutic peptides, ALM201 and AD-01, are not cytotoxic, these agents have multiple synergistic anti-tumour activities including anti-angiogenic, anti-CSC and anti-metastatic involving CD44, and possibly, DLL4 and Notch 4 which gives them a clinical advantage over other anti-angiogenic agents.