*3.2. 3D Predictive Models: From Cancer Study to Drug Testing*

The scientific community has recently focused on the design and bioengineering of innovative 3D culture systems able to overcome the well-recognized inadequacy of conventional bi-dimensional (2D) in vitro models in recapitulating the complexity of in vivo microenvironment [82–84]. 3D tools and technologies reproduce cellular heterogeneity, tissue-specific ECM, and biological interactions in a more biomimetic way, providing in vitro platforms which closely resemble native microenvironment. These biomimetic materials can be exploited for basic biological studies, drug screening, and reproduction of viable biological niches for in vivo transplantation. The biomineralization process is an excellent strategy for the development of advanced biomimetic models, including 3D

biomaterials, cellular coatings, and nanoplatforms with flexibility, diversity, and utility of frameworks for a wide variety of applications [85,86].

Ye and co-workers [87] established a rapid biomimetic mineralization approach to obtain a 3D porous and mineralized hydroxyapatite/collagen composite scaffold for bone regeneration. By a custom synthesis process based on self-assembled collagen fibrils as fixed template, the authors created an in vitro 3D bone-like niche seeded with human Umbilical Cord Mesenchymal Stem Cells (hUCMSCs) with high cell viability, adhesion, proliferation, and differentiation into osteoblasts due to the mineralized scaffold. A rabbit femoral condyle defect model was tested to confirm the ability of the viable niche to facilitate bone regeneration and repair over a period of 6–12 weeks. The mineralized collagen scaffold seeded with hUCMSCs successfully promoted the healing of bone defect in vivo; as new bone tissue formed, the scaffold gradually degraded and was absorbed, confirming the promising use of the hUCMSCs-loaded bone-like niche for in vivo transplantation for bone tissue regeneration.

The same concept was exploited by Menale et al. [71] in 2019 for a cell-therapy based strategy. In this case, a biomineralized bone-like scaffold was used as a rationally designed device conceived to be seeded with cells and subsequently transplanted in vivo to restore or replace a missing function that cannot be completely renewed by only cells [88]. The authors used the scaffold as productive factory of bioactive soluble osteogenic Receptor Activator of Nuclear Factor k B Ligand (RANKL) directly secreted by seeded Mesenchymal Stem Cells (MSCs) on the 3D support [89,90]. The scaffold, obtained through direct nucleation of magnesium-doped hydroxyapatite (HA) nano-crystals on self-assembling collagen fibrils (MgHA/Coll) by a pH-driven biomineralization process, showed structural, compositional, and morphological similarities to the native bone ECM. The biomineralized scaffold guarantees the development of an in vitro, viable, bone-like niche able to compensate the RANKL factor deficit in Autosomal Recessive Osteopetrosis (RANKL-ARO) once transplanted in vivo due to the continuous secretion by MSCs; the MgHA/Coll scaffold promoted the differentiation of MSCs towards osteoclasts [91,92], helping to restore the physiological functions of bone cells in a RANKL-/- mice.

Recently, the same MgHA/Coll scaffold was used as bone-like ECM to be seeded with tumor spheroids, called sarcospheres, and parental cells of MG63 and SAOS-2 osteosarcoma cell lines as enriched Cancer Stem Cells (CSCs) models with the final purpose of obtaining a 3D in vitro CSC-niche of osteosarcoma (Figure 4) [73]. The material provided specific physical-chemical and biomechanical stimuli to the critical pluripotent stem cell population, giving birth to a 3D predictive in vitro model of CSC-niche of osteosarcoma with enhanced stemness and niche-related properties compared to those seeded with parental cells. Through an in-depth cellular and molecular characterization of sarcospheres, and an optimization of the scaffold resembling tumor ECM, the authors were able to provide a closely mimetic in vitro platform for tumor studies and CSC-specific drug screening [93].

A novel biomineralization-inspired cancer therapy has recently been developed as proof-of-concept of advanced nanotechnological therapy. Natural mineral accumulation is a significant biological process that, in abnormal cases, causes the excessive deposition of calcium ions in damaged or defective tissues, leading to common pathologies, such as kidney stones and vascular calcification [94]. The anomalous mineralization can be exploited as "biomimetic pathological mineralization" onto some tumor cells, such as human cervical cancer cell line (HeLa), which can selectively assimilate, folate, and concentrate calcium ions by the overexpression of folate receptor in cancer cells, creating a Cancer Cell-Targeting Calcification-based therapy (CCTC) as reported by Zhao and co-workers [95]. On this trajectory, a biomineralization-inspired drug free strategy can be used to promote cell death by creating a calcium phosphate (CaP) mineral cell coating that leads to the agglutination of tumor cell nuclei without inducing normal cell death [95]. This approach also showed promising results on metastasis, where the survival rate of pathological mice improved significantly (up to 80%) due to the suppression of the metastasis by selective calcification-based substitution of the tumor with curable sclerosis. However, the required

concentration of calcium ions exceeds physiological levels, thus identifying an innovative biomineralization-inspired material able to specifically accumulate ions in the target tissue to facilitate calcium mineral nucleation is still a challenge that may be addressed, for example, by the exploitation of specific ligand/antigens interactions on cancer cell membranes [95]. Without a doubt, this concept can be used to eventually create in vitro 3D biomineralized-based scaffolds able to specifically induce tumor cell death.

**Figure 4.** Panel of figures of in vitro 3D model of Cancer Stem Cells (CSCs)-niche of osteosarcoma from [73]. (Under a Creative Commons Attribution 4.0 International License) using biomineralized scaffolds based on collagen and magnesiumdoped hydroxyapatite (MgHA/Coll scaffold) as bone-like ExtraCellular Matrix (ECM). (**A**) Histological analysis after Haematoxylin–Eosin (H&E) staining of the MgHA/Coll scaffold seeded with both cellular phenotypes, parental and spheroidal, of MG63 and SAOS-2 osteosarcoma cell lines. The morphological features and the interaction behavior of the sarcospheres and parental cells with the scaffold is shown, with image enlargements of 200 μm on the right of the figure. (**B**,**C**) Immunofluorescence analysis of the 3D MgHA/Coll models with sarcospheres. Representative image of OCT-4 immunolocalization in SAOS-2 sarcospheres in image (**B**); scale bar 50 μm. SOX-2 immunolocalization in MG63 sarcospheres in image (**C**); scale bar 25 μm. Blue DAPI: cell nuclei; green: OCT-4; and red: SOX-2.

Microcalcifications (MCs) also serve as diagnostic markers for breast cancer; breast cancer screenings (e.g., mammography) frequently rely on MCs, and their chemical composition (e.g., calcium phosphate, apatite, calcium oxalate, etc.) is associated with tumor malignancy [96,97]. However, due to the absence of sufficiently predictive 3D tumor models, little is known about how they form in the body, their effective role in cancer progression, or how cancer cells are involved in the mineralization process. Therefore, Vidavsky and co-workers [98] exploited the role of biological induction of biomineralization [99] for developing in vitro 3D model of breast tumor MCs to study the cellular pathways involved in MCs formation as a function of malignancy potential. Mammary multicellular spheroids were obtained by parent MCF10A benign human breast epithelial cell line, MCF10DCIS.com [100] and MCF10CA1a [101] which derived from MCF10A and possessed ductal carcinoma in situ (DCIS) and invasive tumors characteristics, respectively; together, these three cell lines allow the modeling of varying stages of breast cancer, ranging from non-malignant (MCF10A), pre-cancerous (MCF10DCIS.com), to invasive phenotype (MCF10CA1a) allowing us to investigate the correlation between cell phenotype and MCs formation. To ensure the physiological relevance of the model, the authors cultured cells in ultra-low attachment conditions with media that contained calcium, magnesium, and phosphate concentrations similar to the human body, but lacked any osteogenic agents in order to observe the real malignancy potential of spheroids just by the development of MCs. Obtained spheroids had diameters larger than 300 mm with low cell viability at the core due to limited diffusion of oxygen and nutrients (Figure 5A). Interestingly, no particles are

observed in the MCF10A spheroids (Figure 5B–D). Moreover, apatite MCs were primarily detected within viable cell regions in the shells and their number and size increased with malignancy potential of the spheroids; conversely, alkaline phosphatase (ALP) decreased with malignancy potential, while osteopontin (OPN) increased. These findings support the induction of a mineralization pathway by cancer cells in a manner that is linked to their malignancy potential. This work offers an innovative exploitation of the mineralization process, which allows us to both create more reliable 3D stage-specific cancer models by inducing specific-MCs as indicators of malignancy potential and, consequently, use these platforms to deeply investigate cancer pathways.

**Figure 5.** Panel of figures of in vitro 3D culture model of breast cancer microcalcifications from [98]. (**A**) A schematic description of the proposed mineralization pathways in the 3D in vitro breast cancer model of various tumor stages; while OPN expression levels increase, ALP expression levels decrease with an increase in malignancy potential of cell line. Viable cell region in light blue, necrotic core in gray, and calcification in red. (**B**–**D**) Mineralized particles in pre-cancerous MCF10DCIS.com spheroids core by SEM magnified section in images; EDS spectrum in image C of asterisk-marked area showing the presence of calcium (Ca), phosphorus (P), and sulfur (S). SRF maps of a spheroid section showing S, P, and Ca distribution in image (**E**). Ca K-edge XANES of the particle marked in (**E**) and a hydroxyapatite standard in image (**F**). (**G**–**N**) Characterization of non-malignant MCF10A (**G**,**I**,**K**,**M**) and invasive MCF10CA1a (**H**,**J**,**L**,**N**) spheroids at day 13 of culture; light microscope (**G**,**H**), 3D reconstructed volumes of spheroids (**I**,**J**), H&E histological staining (**K**,**L**), and nanoCT data stained with iodine of the spheroids cross section (**M**,**N**). All the figures of this panel are reproduced with permission from *Studying biomineralization pathways in a 3D culture model of breast cancer microcalcifications* by Vidavsky et al., 2018, Biomaterials, 179, 12 (doi:10.1016/j.biomaterials.2018.06.030). Copyright 2018 by Elsevier.

In conclusion, few studies exploited biomineralization-inspired process for various useful biomedical applications, from cancer modelling [73] to diagnosis markers [98], showing the need to deeply investigate the potential used of this process independently from conventional applications.
