1. Introduction
Breast cancer (BC) is the most common neoplasm in women worldwide, and metastatic mammary tumours account for over 40% of all cancer-related deaths in females. Progression of breast cancers from benign to invasive as well as metastatic forms is the main cause of cancer-related mortality in women and metastatic BC remains incurable [
1]. Novel insights into the mechanisms of the invasive spread of BC are therefore imperative.
Chemokine receptors belong to the class A family of G-protein coupled receptors (GPCRs) and together with their respective ligands (chemokines), form a complex network that mediates multiple cellular functions in development, homeostasis, and pathology [
2]. A significant line of evidence has been accumulated regarding the roles of chemokine receptors in cancer and metastasis. Cancer cells manipulate the chemokine system, either through upregulation of specific receptors or through secretion of chemokines to regulate cell migration, proliferation and survival in autocrine or paracrine fashions [
3].
In cancer progression and metastasis, the expression of chemokine receptor 4 (CXCR4) and chemokine receptor 7 (CCR7) is of particular importance [
4]. Since the first demonstration of their tumour-promoting role in breast cancer [
5], the significant contribution of these receptors to the metastatic progression in a number of cancers is now firmly established [
6,
7,
8,
9,
10]. Furthermore, the expression of both CXCR4 and/or CCR7 are of significant prognostic value in multiple other cancers. Most importantly, both CXCR4 and CCR7 have been proposed to be potential therapeutic targets in cancer treatments [
11].
GPCR aggregation has been outlined as a critical determinant of their signalling and function. These seven transmembrane receptors have been shown to form homodimers, heterodimers, and multimeric complexes [
12]. Importantly since the first demonstration of the pathologic association between the angiotensin II AT1 receptor and the bradykinin B2 receptor [
13], GPCR aggregates were shown to strongly contribute to many disease conditions such as atherosclerosis, Alzheimer’s disease, neurodegeneration and others [
14].
Regulation of chemokine receptor activity was also found to be at least partially dependent on the formation of homo- and heterodimers [
15]. Chemokine receptor dimerization affects ligand affinity, downstream signalling and receptor internalisation/recycling [
16]. Notably, accumulating evidence indicates that the receptor heterodimer is an active unit with distinct and unique signalling as well as pharmacological properties [
17,
18,
19,
20] and that dimeric or oligomeric chemokine receptor complexes should be considered as unique signal-transducing units with their own distinct biochemical and functional signatures [
21].
The association of chemokine receptors in heterocomplexes has been well documented. Heterodimerization and the functional outcomes of CXCR4 and CCR7 cross-regulation have been mainly reported in the context of immune cells. In CD4
+ T cells, CXCL12-mediated signalling promotes CXCR4-CCR7 heterodimerization and further augments T cell migration [
22]. During B cell development, CXCR4 and CCR7 association results in deficient activation of the G protein alpha subunits Giα1 and 2 [
23], leading to a differential downstream signalling specific to the heterodimer. The co-stimulatory effect of CXCR4 and CCR5 in primary T cells has been correlated to the formation of CCR5-CXCR4 heterodimers, with distinctive signalling and biological properties [
18]. Furthermore, CXCR4 and ACKR3 co-expression result in constitutive recruitment of β-arrestin-2. It also enhances cell migration, in response to CXCL12 stimulation and lung metastasis of breast cancer cells [
24,
25].
However, despite these previous findings, due to the inherent difficulty in detecting native chemokine receptor dimers, as well as the limitations in the availability of in vivo models, there is minimal evidence in relation to the functional outcomes of chemokine receptor heterodimerization in a pathological context in general, and in cancer in particular [
16,
26].
We have previously demonstrated that it is the functional activation of both CXCR4 and CCR7, as opposed to their expression levels, that correlates with the invasive and metastatic phenotype of breast cancer cells [
27]. We also established a direct connection between the activation of CXCR4 and CCR7, and the inhibition of detachment-induced apoptosis (anoikis) in metastatic breast cancer cells that potentially contributes to the metastatic spread of mammary tumours [
28]. However, a possible physical association between CXCR4 and CCR7 or its functional significance in relation to breast cancer progression has not been previously investigated. In this study, the existence of the CXCR4-CCR7 heterodimers in primary mouse and human mammary tumours is shown for the first time. Moreover, we also demonstrate the significance of the CXCR4-CCR7 complex formation to tumour-promoting receptor function in breast cancer cells. The results described here may thus present new therapeutic opportunities by disrupting the CXCR4-CCR7 hetero-complex in the treatment of advanced breast cancer.
2. Materials and Methods
2.1. Mice
All experimental procedures were approved by the animal ethics committee of the University of Adelaide. Mice were maintained in pathogen-free conditions in the University of Adelaide Animal Services facility. The FVB/NJ MMTV-PyMT mice were purchased from the Jackson Laboratory and were backcrossed for 14 generations to the mice with C57Bl/6 background. The C57Bl/6 background was subsequently confirmed by microsatellite analysis.
2.2. Human Tissues
Ethical approval was granted by the Royal Adelaide Hospital Ethics Committee. Normal breast and carcinoma tissues were obtained from R. Whitfield (Breast Endocrine and Surgical Oncology Unit, Royal Adelaide Hospital). All patients gave written informed consent for use of tissue for medical research prior to surgery. The human breast tissue microarray (TMA-1005) was purchased from Protein Biotechnologies (Ramona, CA, USA).
2.3. Human Gene Expression Analysis
The gene expression dataset used here was METABRIC [
29]. Raw data were obtained from the Oncomine™ platform (Thermo Fisher Scientific, Waltham, MA, USA) for
CXCR4 (GC02M136114) and
CCR7 (GC17M040556). A compound log
2 fold-change gene profile for the two-gene expression was created by taking a mean log
2 fold change of each individual gene [
30]. Statistical analysis was performed using GraphPad Prism software (San Diego, CA, USA).
2.4. Cell Lines
All human cell lines were purchased from the American Type Culture Collection (Manassas, VA, USA): MDA-MB-231 (CRM-HTB-26); MDA-MB-361 (HTB-27); T47D-KBluc (CRL-2865); MDA-MB-453 (CRL-1500); ZR-75-30 (CRL-1504). Cells were grown at 37 °C in 5% CO2, in a humidified atmosphere according to the supplier’s instructions.
2.5. Isolation of Mouse Mammary Epithelial Cells
Total mammary cells were derived as previously described [
31]. Total cell populations were isolated from multiple lesions and pooled from 2–3 mice. Briefly, all mouse mammary glands were dissected, and lymph nodes removed, manually dissociated and then digested in Dulbecco’s modified Eagle’s medium (DMEM) (Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 1 mg/mL collagenase IA, 100 U/mL hyaluronidase (Worthington Biochemical Corporation, Lakewood, NJ, USA) and 2% foetal calf serum (FCS) for three h at 37 °C. The freshly isolated total cell preparations were cultured overnight in non-adherent conditions [1:1 volumes of DMEM and Ham’s F12 nutrient mix (Thermo Fisher Scientific), supplemented with 10 ng mL
−1 epidermal growth factor (EGF), 20 ng mL
−1 basic fibroblast growth factor (bFGF) (both PeproTech, Cranbury, NJ, USA) and 0.5 × B27 (Thermo Fisher Scientific)] to obtain a pure epithelial cell culture. Cells were then cultured adherently in DMEM supplemented with 10% FCS and antibiotic–antimycotic (Thermo Fisher Scientific) at 37 °C in 5% CO
2 in a humidified atmosphere.
2.6. Isolation of Human Mammary Epithelial Cells
Isolation of the human mammary epithelial cells from surgical specimens were performed as previously reported [
32]. Briefly, human specimens were manually dissociated and digested in DMEM supplemented with 20 mM HEPES (Thermo Fisher Scientific, Waltham, MA, USA), 1 mg/mL collagenase IA, 100U/mL hyaluronidase (both Worthington Biochemical Corporation, Lakewood, NJ, USA), 12 U/mL DNase I (Merck, Darmstadt, Germany), 1% penicillin-streptomycin and 0.25 μg/mL fungizone (both Sigma Aldrich, St. Louis, MO, USA). Initial digests were washed with DMEM, lysed of red blood cells and single-cell suspensions were obtained by further 10 min digest in trypsin (Thermo Fisher Scientific) at room temperature. Cells were filtered through a 70 μm nylon mesh (Corning, Somerville, MA, USA) prior to further analysis or culture.
2.7. In Vivo Metastasis Assay
An experimental metastasis experiment was carried out as previously described in [
28]. Briefly, human breast cancer cell lines were engineered by retroviral transduction to stably express GFP. Six to eight-week female CB-17 SCID mice (ARC, Perth, WA, Australia) were injected IV into the tail vein with 5 × 10
5 cells suspended in 200 µL PBS (Thermo Fisher Scientific, Waltham, MA, USA). Ten weeks after cell injection mice were sacrificed, lungs excised, perfused with PBS and bright field and fluorescent images were recorded by stereo microscope Leica MZ16FA (Wetzlar, Germany).
2.8. Ligand Cooperation Assay
Synthetic chemokine ligands were obtained from the Biomedical Research Centre, University of British Columbia (Vancouver, BC, Canada). Cells in suspension were incubated with 5 ng/mL of biotinylated CCL19 alone or in combination with 10 ng/mL of unlabelled CXCL12 for 30 min on ice. Cells were fixed in 4% formaldehyde (Sigma Aldrich, St. Louis, MO, USA), incubated with FITC-conjugated streptavidin (Rockland Immunochemicals, Limerick, PA, USA) and analysed by flow cytometry.
2.9. Flow Cytometry
5 × 104 cells were fixed in 4% formaldehyde (Sigma Aldrich, St. Louis, MO, USA) and immunostained for 45 min on ice in PBS containing 0.5% bovine serum albumin (PBS-0.5% BSA). The antibodies used, were PE-conjugated anti-human CXCR4 (clone 1D9, BD, Franklin Lakes, NJ, USA), biotinylated anti-mouse CXCR4 (clone 2B11, BD, Franklin Lakes, NJ, USA) and APC-conjugated anti-human/mouse CCR7 (clone 3D12, eBioscience, Thermo Fisher Scientific, Waltham, MA, USA). Samples containing biotinylated antibodies were treated with PE-conjugated streptavidin (Rockland Immunochemicals, Limerick, PA, USA) in PBS/0.5% BSA for 30 min on ice. Flow cytometry was carried out using FACSCanto equipment (BD, Franklin Lakes, NJ, USA) with standard settings. Data analysis was performed using FlowJo software (BD, Franklin Lakes, NJ, USA). Positive events were defined above the level of background staining observed using matched isotype control antibodies.
For the FACS-FRET experiments, the bandpass filter settings for detection on FACSCanto equipment were changed to the following excitation/emission windows: PE—488 nm/>556LP + 585 ± 42; FRET—488 nm/>655 nm, and APC—633 nm/660 ± 20. Positive FRET signal was defined as the level of fluorescence above background staining observed using matched isotype control antibodies.
2.10. Immunofluorescence Analysis
For the human tissue analysis, formalin-fixed, paraffin-embedded sections of 4 µm were rehydrated and immersed in a 10 mM citric acid buffer at pH 6.0, boiled for 20 min, then cooled to room temperature. Specimens were blocked in 5% normal goat serum in PBS for 30 min and incubated overnight at 4 °C with the mouse anti-human CXCR4 (clone 44708, R&D Systems, Minneapolis, MN, USA) at 20 μg/mL; rabbit anti-human CCR7 (clone Y59, Epitomics, Burlingame, CA, USA) at 15 μg/mL diluted in 0.5%BSA in PBS. The secondary antibodies used, were anti-mouse Alexa Fluor® 647 and anti-rabbit Alexa Fluor® 488 (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA) at 1:400 dilution and were then incubated for 90 min at room temperature in the dark. Samples were mounted using Vectashield mounting media (Vector Laboratories, Burlingame, CA, USA). Immunofluorescence images were acquired using the Leica SP5 spectral scanning confocal microscope (Wetzlar, Germany).
2.11. Proximity Ligation Assay (PLA)
Following the primary antibody incubation described in immunofluorescence analysis, the PLA procedure was carried out according to the manufacturer’s protocol (OLINK Bioscience, Uppsala, Sweden). Briefly, primary antibodies of two different species are bound to the target proteins. Next, the oligonucleotide-conjugated secondary antibody pairs (PLA probes) are added. The close proximity of bound PLA probes brings together two oligonucleotides from both probes as a template for complementary circular DNA ligation. Finally, DNA polymerase addition initiates rolling circle amplification (RCA) primed by oligonucleotides on one of the PLA probes, and fluorescent oligonucleotides are used to visualize the RCA products. RCA product consists of a single DNA strand with several hundred complements of the DNA circle that is labelled by hybridized fluorophore-conjugated short DNA oligonucleotides (detection oligonucleotides). The bright discrete RCA product consists of a distinct sub-μm signal that allows visualization and enumeration of single molecules by fluorescent microscopy [
33,
34].
Detection reagent Red was used for the amplification step. After the final PLA step, the slides were washed with PBS and further incubated with anti-human EpCAM Alexa Fluor® 488 (Santa Cruz Biotechnology, Dallas, TX, USA) at 1:50 dilution in 0.5% BSA in PBS overnight. Subsequently, samples were washed and mounted with Vectashield mounting media (Vector Laboratories, Burlingame, CA, USA) for TMA samples and Vectashield mounting media containing DAPI for primary normal breast and tumour tissues. Images were acquired using the Leica TCS SP5 confocal microscope (Wetzlar, Germany). The slides were evaluated by sequential scanning. Ten fields of view were acquired for each TMA tissue spot. The number of PLA signals per μm
2 of epithelium defined by the EpCAM positive staining were quantified using particle analysis in FIJI, (
http://fiji.sc/Fiji, accessed on 15 August 2021). The images were processed by blind analysis. Representative images shown are the maximum intensity Z-projections.
For the human cell lines, one day prior to primary antibody staining, cells were plated onto Poly-L-Lysine adhesion slides (Thermo Fisher Scientific, Waltham, MA, USA) using 12-well tissue culture inserts (flexiPERM®, SARSTEDT, Nümbrecht, Germany) at 1 × 10
3 cells per well. The cells were fixed using a 4% formaldehyde solution (Sigma Aldrich, St. Louis, MO, USA). The PLA procedure was then carried out as described above. Five fields of view for each cell line type were acquired. The number of PLA signals per cell were quantified using particle analysis in FIJI (
http://fiji.sc/Fiji, accessed on 15 August 2021). Representative images shown are the maximum intensity Z-projections.
2.12. Forced Heterodimerization
Inducible CXCR4 and CCR7 dimerization system was constructed using iDimerize™ Inducible Heterodimer System (Clontech, Takara Bio Inc, Kusatsu, Shiga, Japan) according to manufacturer’s instructions. The system uses DmrA and DmrC domains altered to specifically bind the A/C heterodimerizer (ACH) ligand. Briefly, to generate the fusion pair, CXCR4 cDNA and Myc-tagged CCR7 cDNA was PCR-amplified and inserted into pHet-Mem1 (pCXCR4-H) and pHet-1 (pCCR7-M) plasmids. Primer sequences and vector maps are available upon request.
2.13. Transient Transfection of Human Cell Lines
MDA-MB-231 and T47D cell lines were transiently transfected with pCXCR4-H and pCCR7-M plasmids using Optifect™ transfection reagent (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s instructions. Following transfection, the cells were cultured at 37 °C for 48 h prior to further analyses.
2.14. AlphaScreen cAMP Assay
cAMP levels were assessed using the AlphaScreen Detection Kit (Perkin Elmer, Waltham, MA, USA). All cells were serum-starved for at least three hrs prior to the analysis. Briefly, cells were resuspended at 5 × 105 cells/ml in AlphaScreen stimulation buffer containing HBSS (Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 0.5 mM IBMX, 5 mM HEPES, 0.1% BSA (all from Thermo Fisher Scientific, Waltham, MA, USA), pH 7.4 in the presence or absence of forskolin (10 μM final concentration) (Sigma Aldrich, St. Louis, MO, USA) and treated or untreated with A/C heterodimerizer (100 mM final concentration). 3000 cells/well were aliquoted in triplicate into 384-well OptiPlate (Perkin Elmer, Waltham, MA, USA) and stimulated with CXCL12 and CCL19 chemokine ligands at indicated concentrations for 15 min at 37 °C. Further, cells were processed according to the manufacturer’s protocol. Alpha Screen signal was recorded using PHERAstar® FSX detection system (BMG Labtech, Ortenberg, Germany) using the AlphaScreen optical module (Ex. 680 nm/Em. 520−620 nm). Data analysis was performed using GraphPrism software (San Diego, CA, USA).
2.15. Matrigel Invasion Assay
MDA-MB-231 and T47D cells were co-transfected with pCXCR4-H and pCCR7-M iDimerize plasmids. 48 hrs post-transfection cells were serum-starved for at least three h and then treated or untreated with A/C heterodimerizer (100mM final concentration) for 1 hour at 37 °C. 5 × 104 cells were then aliquoted in duplicate into Matrigel (BD, Franklin Lakes, NJ, USA) pre-coated 96 transwell assay plates with 5μm pore permeable support inserts (Corning, Somerville, MA, USA). Cells were then treated with CXCL12 at 10 ng/mL and/or CCL19 at 20 ng/mL final concentrations and allowed to migrate for 24 h at 37 °C. Cells were then detached from the underside of the transwell insert with 0.1% trypsin (Thermo Fisher Scientific, Waltham, MA, USA), loaded with Calcein AM/1 (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA) and cell density was estimated by PHERAstar® FSX detection system (BMG Labtech, Ortenberg, Germany) relative to total cell input.
2.16. Statistical Analysis
The statistical analysis was performed using GraphPad Prism software (San Diego, CA, USA). Refer to figure legends for details of the statistical analyses undertaken. P-values were calculated to assess statistical significance with levels of significance * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001 and **** p ≤ 0.0001.
4. Discussion
The fact that class C GPCRs function as heterodimers and that their context-dependent heterodimerization is critical for receptor function is widely accepted. But the significance of the oligomeric status of most numerous class A and B receptors, which constitute >90% of all GPCRs, remains hotly debated. In this manuscript, we have described the first functional CXCR4-CCR7 chemokine receptor, which is specifically expressed in advanced breast cancer cells in vitro in continuous cell lines, ex vivo, in primary mammary mouse tumour cells, human tumour cells and archived human patients breast cancer tissues. Most importantly, we have demonstrated that the function of its protomers is strictly controlled by the formation of the dimer, as neither CXCR4 nor CCR7 activation could be detected in non-invasive mammary tumour cells, where the presence of the heterodimeric CXCR4-CCR7 receptor was not found. It is important to note that even though it is widely accepted that resonance transfer approaches allow identification of interacting units within protein complexes, these techniques do not allow to sufficiently differentiate between closely located protomers versus bona fide protein aggregates. Therefore, we have employed several alternative methods to confirm the CXCR4 and CCR7 association.
Forced dimerization of the CXCR4 and CCR7 in non-invasive cells has led to a partial restoration of the functional response to their cognate ligands suggesting an involvement of other factors in addition to dimer formation in controlling activation of these chemokine receptors. Our observations thus provide the first clear evidence for a specific novel link between the CXCR4-CCR7 heterodimerization, CXCR4 and CCR7 function and the metastatic propensity of breast cancer cells.
CXCR4 protein expression has long been suggested as a survival prognostic marker in numerous cancers [
41]. In contrast, CCR7 expression levels on the surface of cancer cells have not yet been sufficiently analysed to make statistically unbiased conclusions regarding correlations with cancer outcomes. However, neither of these cellular receptors has been conclusively demonstrated as a biomarker for locally invasive or metastatic tumours. Using a tailored PLA approach, we have shown for the first time that the expression of the CXCR-CCC7 dimeric complex significantly correlates with the presence of lymph node metastasis in human mammary tumours which indicates that this chemokine receptor heterodimer may be a novel biomarker of the distant spread in breast cancer. Of particular interest also is the fact that our analysis of the publicly available gene data sets showed that while individual expression levels of
CXCR4 and
CCR7 did not correlate with any breast tumour characteristics,
CXCR4 and
CCR7 co-expression was highly significantly linked to the tumour grade, further emphasizing the role of the interaction between the two receptors in breast cancer progression.
As indicated above, forced dimerization of the CXCR4 and CCR7 demonstrated that simply bringing the receptors together under the conditions employed only partially activated their signal transduction. This finding can be interpreted in a number of ways. First, it could indicate the importance of a specific tertiary structure of the dimeric receptor for its full activity and this required specific heterodimer conformation may not be completely reproduced by just bringing two protomers together through their C-termini forced interaction. Second, our data demonstrating the complete inactivity of CXCR4 and CCR7 receptors in non-metastatic cells paints a more complex picture suggesting that inherent differences in breast cancer cells are likely to be important determinants of CXCR4-CCR7 receptor heterodimerisation in the context of tumour progression. The changes of multiple factors occurring in metastatic cells may likely be required for maximal CXCR4-CCR7 heterodimer activity. These may include the presence or absence of co-factors that change receptor conformation upon binding, expression of specific protein-modifying enzymes that mediate receptor post-translational modifications, differential expression of G-protein subunits that can selectively mediate CXCR4 and CCR7 allosteric changes or numerous other changes in cellular components that have been demonstrated to have an impact on the multimeric GPCRs [
42]. We have previously found that in a panel of human breast cancer cell lines, the coupling of Gα
i and Gβ proteins with CXCR4 varies significantly between cell lines with different invasive properties [
27]. These findings suggest that the composition of the heterotrimeric complex, together with other factors may determine the properties of chemokine receptor heteromers.
Overall, our findings further demonstrate that chemokine receptor activity is regulated at multiple molecular levels with heterodimerization being a very significant and efficient molecular switch mechanism, which likely can be further affected by spatial and temporal protomer and accessory protein expression. This multilayered and multifactorial organisation emphasizes the potential importance of the tight control of the chemokine receptor activity in homeostasis. In pathology, a breakdown in those control mechanisms may unleash strong responses augmenting and even superseding negative regulators to advance disease progression. In cancer, inherent genetic instability selection pressure may lead to the expansion of cells with a more aggressive phenotype, which in part may be characterised as well as driven by functional CXCR4-CCR7 dimers on the cell surface. Future studies should focus on elucidating the CXCR4-CCR7 molecular dimerizer “switch” that could be then targeted for more effective therapies in metastatic breast and potentially other cancers.