Several major pathological conditions, such as cardiovascular diseases and cancer display minerals or organic compound depositions [1
]. Tissue calcifications are defined as the deposition of calcium salts, together with smaller amounts of iron, magnesium, and other mineral salts [1
In breast tissues, microcalcifications play a crucial role in early cancer diagnosis [2
]. Indeed, approximately 50% of non-palpable breast cancers are detected by mammography exclusively through microcalcification patterns [3
], revealing up to 90% of ductal carcinoma in situ [4
]. They can be classified according to their appearance on a mammogram based on the Breast Imaging Reporting and Data system [5
], or by their physical and chemical properties [6
]. In a recent paper, we demonstrated, for the first time, the presence of magnesium-substituted hydroxyapatite (Mg-HAp), which was frequently noted in breast cancer but never found in benign lesions [7
]. In our experience, calcium oxalate (CO) calcification is often associated with benign lesions, whereas hydroxyapatite (HA) is related both to benign and malignant lesions [7
]. Recent evidence suggests that the morphological appearance of mammographic microcalcifications is associated with patient prognosis. In fact, patients harboring small breast tumors with casting type calcifications in the mammograms have a poor survival rate for this tumor size category [8
Despite the extensive investigation of breast microcalcifications, the mechanisms leading to their formation are still not defined. Recently, we suggested that ectopic mineralization in pathological conditions might be triggered by the epithelial to mesenchymal transition (EMT) phenomenon and regulated by mechanisms similar to those occurring under physiological conditions [8
]. Ours and other studies [8
] allows postulating an overlap between breast microcalcification production and bone mineralization. In particular, Maria Morgan’s group [10
] investigated the molecular mechanisms of the microcalcification process in breast cell cultures demonstrating that the mineralization process, related to alkaline phosphatase activity, could be similar to that observed in osteoblast cells. Altogether, these pieces of evidence suggest an active role of microcalcifications in breast cancer occurrence and progression. Thus, this study aims to investigate: (a) the putative association between the presence of microcalcifications in breast lesions and the expression of both EMT and bone biomarkers, (b) the possible role of microcalcifications made of CO in the Breast Osteoblast-Like Cells (BOLCs) formation by macrophage-related EMT, and (c) the putative association between elemental composition of microcalcifications and breast cancer progression.
Historically, the identification of microcalcifications during mammographic exams is considered a sign of breast disease, both in benign and breast lesions [15
]. This makes the mammographic screening the current gold standard clinical methods for the early detection of breast lesions [16
]. However, despite the relevant role of microcalcifications in the management of breast cancer patients, and their potential prognostic value, the cellular and molecular mechanisms involved in their formation are largely unknown. In this context, pioneering in vitro studies about osteomimicry of mammary cells have been performed by Maria Morgan and colleagues [11
], who demonstrated that bioengineered 3D scaffolds made of collagen glycosaminoglycan support the growth and mineralization of mammary cell lines due to their capability to simulate the bone microenvironment [11
]. In line with these pieces of evidence, very recent studies identified molecules involved in breast cancer osteomimicry. In our laboratory, we highlighted some ex vivo data about the microcalcification formation. In fact, we described, for the first time, breast cancer cells capable of producing microcalcifications in a process similar to bone mineralization [7
]. Specifically, we showed that breast cancer cells that acquire both morphological and molecular characteristics of mesenchymal cells by the EMT could then assume an osteoblast-like phenotype under the induction of molecules of the bone morphogenetic proteins family [7
]. These cells, called BOLCs, show the capability to produce and secrete breast microcalcifications composed by HA [7
]. Starting from these pieces of evidence, this study aims to investigate: (a) the putative association between the presence of microcalcifications in breast lesions and the expression of both EMT and bone biomarkers, (b) the possible role of microcalcifications made of CO in the BOLCs formation by macrophages-related EMT, and (c) the putative association between elemental composition of microcalcifications and breast cancer progression.
For as much as the study of the putative association between the presence of microcalcifications in breast lesions and the expression of both EMT and bone biomarkers is concerned, we performed ex vivo study on 174 breast biopsies. Immunohistochemical analysis of vimentin and CD44 confirmed the evidence that in the presence of microcalcifications, both benign and malignant breast lesions are characterized by numerous breast cells with a mesenchymal phenotype. Thus, the presence of microcalcifications could be considered a negative prognostic factor regardless of the type of breast lesion. In this study, we did not carry out analysis of other EMT in situ biomarkers, such as the loss of e-cadherin or the acquisition of the n-cadherin since the data associated with these molecules in breast cancer are still controversial. Indeed, despite numerous studies reporting on the loss of e-cadherin during EMT [18
], Hollestelle et al. showed that the loss of e-cadherin is not a necessity for EMT in human breast cancer lines [20
]. The loss of e-cadherin is well described for a special type of breast cancer—the lobular carcinomas [19
]. In addition, Canas-Marques et al. reported several pitfalls in the interpretation of e-cadherin by immunohistochemistry [21
]. As reported above, in our case section, no lobular carcinomas were present. Similarly, despite some studies reported the expression of n-cadherin during breast cancer progression, its expression seems to be associated with special types of breast cancer such as lobular carcinomas and micropapillary carcinomas [22
]. Lastly, in our experience, the number of n-cadherin-positive ductal breast cancer cells is often negligible if evaluated by immunohistochemistry.
More important, ex vivo immunohistochemical analysis showed the prevalence of BOLCs in the malignant breast lesions with microcalcifications when compared to lesions without calcifications. Specifically, we tested the expression of some known osteoblast biomarkers such as RANKL, OPN and VDR [24
]. Among these RANKL and OPN-positive breast cells were higher in lesions with microcalcifications rather than lesions without microcalcifications. Specifically, OPN and RANKL-positive cells were often close to microcalcifications allowing to hypothesize a role of these cells in the production of calcium crystals. In agreement with this, it is known that OPN is a protein involved in the early phases of hydroxyapatite formation [25
], whereas RANKL is a biomarker of real osteoblasts involved in the regulation of bone metabolism by the RANK/OPG/RANKL system [28
]. More importantly, our data showed a similar trend in the expression of osteoblast differentiation markers, BMP-2, BMP-4, and PTX3 [29
]. Consequently, the expression of BMP-2, BMP-4, and PTX3 could explain both the origin of BOLCs and the formation of microcalcifications. Indeed, molecules of the BMP family, as well as PTX3, are known for their capability to induce both osteoblast differentiation and activity. In particular, both Mantovani’s and our group demonstrated the involvement of PTX3 in the deposition of bone matrix [31
]. Immunohistochemical data were also supported by ultrastructural investigation in which we described the morphological characterization of breast cells next to microcalcifications. In fact, the morphology of these cells was similar to that of real osteoblasts (large cytoplasm, a huge rough reticulum and cytoplasm’s rich in vesicles containing electron-dense granules composed of HA).
These and previous ex vivo data have been used to develop an in vitro model for the study of the role of microcalcifications in breast cancer development. In particular, our in vitro model was based on the evidence that (a) CO microcalcifications are more frequently associated with benign breast lesions rather than malignant ones [7
], (b) the presence of CO can induce the macrophage-mediated EMT in epithelial systems [34
], and (c) BOLCs can originate from breast epithelial cells under EMT stimuli [7
]. Thus, to demonstrate a possible active role of CO microcalcifications in breast cancer occurrence and progression, we developed a co-culture system in which breast cancer cell lines (MDA-MB-231) were incubated with both CO and human monocytes. Remarkably, already after 10 days, electron microscopy and EDX analysis displayed the presence of breast cancer cells with osteoblast-phenotype and, most important, the presence of HA crystals. No HA crystals were instead observed in cultures of MDA-MB-231/CTRL or MDA-MB-231-MΦ/CO. Immunohistochemical and western blot analysis of cell cultures also demonstrated that in the presence of CO and monocytes, breast cancer cells undergo to EMT becoming able to produce HA crystals. In agreement with this, in the presence of macrophages and CO, we found an increase in the vimentin-positive MDA-MB-231 cells, a decrease of mutated p53, and metabolic changes toward oxidative metabolism (lower expression of PKM2). These preliminary in vitro data allowed us to hypothesize that CO microcalcifications can participate to breast cancer occurrence and development through the recruitment and activation of macrophages. Specifically, in our hypothesis, CO microcalcifications present in the benign breast lesions could be related to macrophage recruitment and consequently, to the expression of transforming growth factor(TGF) β. As known, TGFβ is the most important EMT inducer able to activate the molecular events responsible to the transformation of breast epithelial cells into mesenchymal-like cells [35
]. In the presence of osteoblast differentiation factors such as BMPs and PTX3, these mesenchymal-like cells can then differentiate into BOLCs. Thus, the presence of CO in benign breast lesions could trigger both breast carcinogenesis and the formation of calcifications made of HA.
To further corroborate our hypothesis, we retrospectively collected breast biopsies from 30 patients that developed a malignant lesion within five years from a previous diagnosis of benign lesions (for each patient we collected biopsies of both benign and malignant lesions). In more detail, there were microcalcifications in only 15 benign lesions out of the 30 samples. Noteworthy, in all patients with benign lesions associated with microcalcifications made of CO, the subsequent malignant lesion developed in the same quadrant. Moreover, these malignant lesions were characterized by the presence of microcalcifications made of HA. Malignant lesions of patients with previously benign lesions composed of CO calcifications were poorly differentiated carcinomas. Noteworthy, lesions with CO calcifications were rich in M2 macrophages (CD163 positive cells). No similar results were observed for patients without calcifications in benign lesions.
Altogether our data shed new light about the role of microcalcifications in breast cancer progression and occurrence. In particular, in this study, we proposed a model for breast cancer carcinogenesis based on the capability of CO calcifications to induce macrophage-EMT. Last but not least, this study further emphasizes the biological similarities between bone and breast metabolism.
4. Material and Methods
4.1. Breast Sample Collection
From December 2010 to December 2014, we collected 176 breast samples in total: breast lesions and 124 malignant breast lesions. From each biopsy, paraffin serial sections were obtained to perform histological classification and immunohistochemical analysis. In addition, 1 mm3 tissue fragments were used to perform ultrastructural (transmission electron microscopy) and microanalytical (EDX-microanalysis) investigations. This study protocol was approved by the “Policlinico Tor Vergata” Independent Ethical Committee (reference number # 129.18, 26 July 2018). Written patient consent was obtained for each patient.
After fixation in 10% buffered formalin for 24 h, breast tissues were embedded in paraffin. Three-micrometer thick sections were stained with hematoxylin and eosin (H&E), and the diagnostic classification was blindly performed by two pathologists [36
4.3. Immunohistochemistry of the Paraffin Sections
Briefly, antigen retrieval was performed on 3-μm-thick paraffin sections using EDTA citrate pH 7.8 or Citrate pH 6.0 buffers for 30 min at 95 °C. Sections were then incubated for 1 h at room temperature with the following primary antibodies, anti-vimentin, anti-CD4, anti-OPN, anti-RANKL, anti-VDR, anti-PTX3, anti-BMP2, and anti-BMP-4 (for details see Table 2
). Washings were performed with PBS/Tween20 pH 7.6. Reactions were revealed by the HRP-DAB Detection Kit (UCS Diagnostic, Rome, Italy).
4.4. Transmission Electron Microscopy (TEM) of Breast Tissues
Small pieces of breast tissue (1 mm3
) from surgical specimens were fixed in 4% paraformaldehyde (PFA) and post-fixed in 2% osmium tetroxide [37
]. After washing with 0.1 M phosphate buffer, the sample was dehydrated by a series of incubations in 30%, 50%, and 70%, ethanol. Dehydration was continued by incubation steps in 95% ethanol, absolute ethanol, and propylene oxide; then, samples were embedded in Epon (Agar Scientific, Stansted Essex CM24 8GF UK) [37
]. Eighty µm ultra-thin sections were mounted on copper grids and observed with Hitachi 7100FA transmission electron microscope (Hitachi, Schaumburg, IL, USA).
4.5. Energy Dispersive X-ray (EDX) Microanalysis
All breast samples underwent EDX microanalysis. Six-micrometer-thick paraffin sections were embedded in Epon resin as previously described [38
], followed by the identification of microcalcifications. Briefly, sections were deparaffinized, hydrated, osmium tetroxide-fixed, dehydrated in ethanol and propylene oxide, and infiltrated in Epon. The embedding capsules were positioned over areas containing previously-identified microcalcifications. Unstained ultra-thin sections of approximately 100-nm-thick were mounted on copper grids for microanalysis. EDX spectra of microcalcifications were acquired with a Hitachi 7100FA transmission electron microscope (Hitachi, Schaumburg, IL, USA) and an EDX detector (Thermo Scientific, Waltham, MA, USA) at an acceleration voltage of 75 KeV and magnification of 12,000×. Spectra were semi-quantitatively analyzed by the Noran System Six software (Thermo Scientific, Waltham, MA, USA) using the standardless Cliff–Lorimer k-factor method [38
]. EDX microanalysis apparatus was calibrated using an x-ray microanalysis standard (Micro-Analysis Consultants Ltd., Cambridgeshire, UK).
4.6. Calcium Oxalate Synthesis
CO was synthesized as described by Grases et al. [39
]. Briefly, 7 L of distilled water was placwd in a crystallizer and heated to 70 °C. Solutions of Na2
(7.5 × 10−3
M) and CaCl2
(7.5 × 10−3
M) were dropped simultaneously at the same speed (250 mL per h). The slurry was filtered, and the crystals were washed with water and ethanol and then dried at 50 °C with a vacuum for 24 h.
4.7. Cell Culture
MDA-MB-231 cells originally derived from a pleural effusion were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA) and maintained by the Cell and Tissue Culture Core, Lombardi Cancer Center (Reservoir Rd. NW, Washington, DC, USA). Cells were routinely cultured in DMEM high glucose (Sigma-Aldrich, St. Louis, MO, USA) supplemented with 10% fetal bovine serum (FBS).
4.8. Monocyte Isolation
Human peripheral blood monocytes (PBMCs) for co-culture experiments were purified by density gradient (Lympholyte-H, Cedarlane) medium following manufacture’s instruction.
After purification, PBMCs were grown in a 24-well (BD Falcon) cell culture plates previously collagenated (rat-tail collagen type I 0.1 mg/mL, BD) in Roswell Park Memorial Institute (RPMI) 1640 (Euroclone, Pero, Milan, Italy) with 10%. FBS, 2 mM L-Glutamine, Penicillin (100 units/mL)/Streptomycin (100 mg/mL). Approximately 40,000 PBMCs were cultured in each well in 700 µL of media. The cells were incubated at 37 °C, and the next day, 20 ng/mL of macrophage colony-stimulating factor (MCSF, Sigma-Aldrich M 9170) was added to induce monocyte activation. After 6–7 days, co-culture experiments were set up.
The coating plate protocol required 0.1 mg/mL collagen type I of rat-tail high-concentration solution (BD Pharmingen, San Jose, CA, USA, cat# 354249) added to the 24 multi-well plates and incubated for 1 h at 37 °C. After incubation, the plates were washed three times with PBS and kept at +4 °C until use.
4.9. In Vitro Model for the Development of “Osteoblast-Like Cells”
For each experiment, about 40,000 PBMCs and 10,000 MDA-MB-231 were cultured. The experimental scheme was MDA-MB-231 + calcium oxalate (MDA-MB-231/CO), MDA-MB-231 + calcium oxalate + activated monocytes (MDA-MB-231-MΦ/CO), MDA-MB-231 + hydroxyapatite + activated monocytes (MDA-MB-231-MΦ/HAP). As controls, we used MDA-MB-231 + activated monocytes (MDA-MB-231-MΦ) and MDA-MB-231 alone (MDA-MB-231/CTRL). After 10–12 days, the cells were used for all subsequent experiments.
4.10. Protein Extraction and Western Blot Analysis
Cells were homogenized directly into following buffer: Tris 50 mM, NaCl 150 mM, EDTA 10 mM, Triton-X 1%, and centrifugated at 10,000 g for 2 min. Protein concentrations were determined by the Bradford assay. Proteins were resolved by 12% SDS-PAGE, electrotransferred on PVDF membranes (Amersham™ Hybond™, GE Healthcare Life Science, Pittsburgh, PA, USA, cat# 28906837) and blocked with 5% (v/v) milk /0.1% (v/v) TBS-T. The blots were probed with the following primary antibodies: mAb mouse anti-beta actin 1:10,000 (Sigma-Aldrich, St. Louis, MO, USA, cat# A5541), mAb mouse anti-p53 1:500 (Santa Cruz Biotechnology, Santa Cruz, TX, USA cat# sc-126) polyAb rabbit anti-PKM2 1:1000 (Cell Signaling Technology, Danvers, MA, USA, cat# 3198).
Membranes were then incubated with the appropriate horseradish peroxidase-conjugated donkey anti-mouse secondary antibody (Jackson Immuno Research, West Grove, PA, USA cat# 715-035-151) 1:15,000 for p53 and 1:20,000 for beta-actin respectively, goat anti-rabbit secondary antibody (Jackson Immuno Research, St. Thomas’ Place, Cambridgeshire, UK, cat# 111-036-047) 1:15,000. Immunodetection was performed by the enhanced chemioluminescence system Western lighting Plus ECL (Perkin Elmer, Waltham, MA, USA, cat# NEL105001EA).
4.11. Cell Culture Immunohistochemistry
Cells were plated on poly-l-lysine coated slides (Sigma-Aldrich, cat #P4707) in 24-well cell culture plates and fixed in 4% paraformaldehyde. After pre-treatment with EDTA citrate at 95 °C for 20 min and 0.1% Triton X-100 for 15 min, cells were incubated 1 h with the rabbit-monoclonal anti-vimentin antibody (see Table 2
). Washings were performed with PBS/Tween20 pH 7.6. Reactions were revealed by HRP-DAB Detection Kit (UCS Diagnostic, Rome, Italy) and Mayer’s Hematoxylin to stain the nucleus.
4.12. TEM and EDX Analysis of Cell Cultures
Cells were fixed in 4% paraformaldehyde, post-fixed in 2% osmium tetroxide and embedded in Epon resin for morphological studies. After washing with 0.1 M phosphate buffer, the sample was dehydrated by a series of incubations in 30%, 50%, and 70%, ethanol. Dehydration was continued by incubation steps in 95% ethanol, absolute ethanol, and hydroxypropyl methacrylate, then samples were embedded in Epon (Agar Scientific, Stansted Essex, UK).
4.13. Statistical Analysis
Statistical analysis was performed using GraphPad Prism 5 Software (San Diego, CA, USA). Immunohistochemical data were analyzed by the Kruskal–Wallis test (p < 0.0001) and by Mann––Whitney test (p < 0.0005).