This April, the European Union (EU) prohibited selling new cosmetic products tested on animals without any exemptions (1). Accordingly, it is forbidden to place a new product on the market that contains even one new ingredient tested on animals in or out of the EU, even if the final product was not tested on animals (2). It is also well-known that pharmaceutical drug development is extremely time-consuming and expensive; moreover, about 90% of drugs fail after preclinical animal testing in human safety and efficacy trials. Additionally, in patients, ∼1/6 newly marketed drugs are withdrawn or discontinued due to serious adverse effects (e.g., hepatic, cardiovascular, hematologic, neurologic, and carcinogenic) (3). Potentially beneficial drugs can also fail and never be placed on the market if these are categorised as toxic or ineffective during preselection based on animal tests (e.g., in recent conditions, aspirin could fail regarding its toxicity test results on rats and rhesus monkey embryos). Thus, certain patients could not benefit or even be omitted due to the limitations of animal models (4). Both U.S. Food and Drug Administration (FDA) and European Medicine Agency (EMA) modernisation acts aim to replace and decrease animal testing. Therefore, the bioengineering research area has rapidly improved in the last decade. Moreover, in vitro human model systems have undergone enormous developments nowadays. In the last pandemic situation, the development of COVID vaccines showed that faster and more efficient technologies are necessary for crisis and medical developments, and we have to leap over animal experiments and use clinical tests as fast as possible (5). Growing evidence propagates those alternative technologies (relevant human cell-based—e.g., organ-on-chips or biofabricated models—or artificial intelligence-combined technologies) could help test and predict human response and toxicity in medical research more accurately (6, 7).

In vitro disease model developments have great efforts to create and serve the need of reducing, and replacing animal experiments and establishing human cell-based in vitro test systems for research use, innovations, and drug tests. We know and accept the quote: “ALL models are wrong but some are useful”—(George E.O. Box, one of the greatest statisticians of the 20th century). We need human cell-based test systems for disease models and experimental cancer research; therefore, in vitro three-dimensional (3D) models have a renaissance, and the development of these technologies is growing ever faster.

Aristotelian doctrine about spontaneous generation describes those non-living substances (water, stones, and salts) that have some additional potential to spontaneously create complex systems and organisms. Regarding these explanations, insects, and flies are developed from mud and inorganic matter (8). The revolution of cell biology and the birth of cellular biology would not have been possible without one great invention—the microscope, after the invention of convex lenses by Janssen and the telescopes created by Galilei. Among many other early medical descriptions (9) Hook and van Leeuwenhoek made their unusual discovery of the invisible microscopic world (life) at the beginning of the 17th century (10). van Leeuwenhoek handcrafted lenses and constructed microscopes. In 1665, Hook published his outstanding findings in Micrographia; he described and illustrated many biological entities as well as defined microscopic units (e.g., “cells” or “pores”). The term “cell” directly comes from this work. The spontaneous generation doctrine was experimentally disproved and laid by Pasteur proving that life arises from pre‐existing living organisms (it was demonstrated that microorganisms are present even in the air) (11). The developments in microscopic technologies helped to improve and even carry out more detailed studies. Virchow, Schwann, and Kölliker examined cells and tissues and also observed that “the elementary parts of all tissues are formed from cells” and suggested that “there is one universal principle of development for the elementary parts of organisms…and this principle is in the formation of cells” (12). Schwann and his colleagues described that the “cells arise inside and near other cells by differentiation of a homogenous primary substance” (13). Virchow and other scientists presented the view that cells are formed via the scission of pre-existing cells. Finally, in 1858, Virchow defined the cell as the fundamental unit of life, and he also created a pathogenic concept—accordingly, diseases are the results of changes in normal cells (“cells with bad behaviour”) (14). In addition, he was the one who laid the base of cellular pathology in 1863 (15) (Figure 1).

The first registered successful in vitro cultures were performed by Roux. He could maintain living cells (from chick embryos) outside the body for a few days (16). In parallel, Loeb could culture skin fragments from guinea pigs in agar culture (17). Then, Harrison developed the first, later termed hanging drop technology from small pieces of frog embryonic tissues, as well as described and introduced the aseptic method which could maintain sterile cell specimens in vitro for weeks (18). In New York, Carrel and his co-workers established cell cultures in vitro using embryonic and adult tissues of many species maintained in special culture media; they were able to culture their cells and tissues for several months (19). They started to work with cancer tissues and also introduced the term “tissue culture.” In 1912, the first “cell line” derived from an explanted chicken embryo heart was also established by Carrel. In 1938, based on his observations and studies, he published a book entitled “The Culture of Organs”. In the 1920s, the introduction of the tissue trypsinization method had a significant impact on cell culture development (20). Rous and his colleague were able to produce single-cell suspension from tissues, detach and subculture adherent cells obtaining homogenous cell strains which were successfully performed with this innovation.

After establishing some animal cell lines, the first human cell line (HeLa derived from cervix cancer) was established by Gey in 1951. In 1961, Hayflick inspired by Carrel’s observations expected that cancer cells have special cancer‐like properties (immortalised and gain limitless proliferation potential), but normal human fibroblasts have restricted growth potential (limited division capacity), and some other normal cells did not grow any longer (21). In parallel, cancer research development, tumorigenesis theories, and developing knowledge about spontaneous or induced transformation allowed viral and cell fusions, to immortalise isolated normal cells, and produce antibodies or recombinant proteins using hybridoma, selection, and molecular biology technologies (22).

In 2006, after the development and isolation of stem cells, Yamanaka and his colleagues described that mouse tail‐tip adult fibroblasts can be reprogramed to stem cells by the simultaneous induction of four transcription factors—Oct3/4, Sox2, Klf4, and c‐Myc. These and the developed induced pluripotent stem cell (iPSC) technologies provide the capacity to form tissues of all three germ layers for tissue cultures (23). Additionally, Takahashi’s pioneer works contributed to the establishment of pluripotent stem cell culturing technologies in many laboratories in the last decades. These results have great importance in medical science, transplantation, oncology, and regenerative medicine, as well.

The history of cancer model system development started at the beginning of the last century in spite of the fact that human tumours are as “old” as human life. There were several prehistorical findings, and the first written documents can be found on Imhotep’s papyrus with 48 described surgical cases, including some breast and other cancers (24). In the ancient ages, Hippocrates had many observations about malignant tumours which reminded him of the moving crabs, therefore, he named the disease cancer. Celsus highlighted the tumours’ invasive behaviour, and in the Middle Ages, Fallopius distinguished benign and malignant tumours, as well. Cancer epidemiology studies were started by Ramazzini and Hill (breast cancers in nuns and testicular cancers in chimney sweeps) in the 17th–18th centuries (25). However, the first Cancer Hospital was established in Reims (France) in 1779, and there were many described and treated cases, the first experimental models on tumours were performed only at the beginning of the last century with chickens and rabbits ( Rous, Yamagiwa, and Ichikawa) (26). After these first models, the number of cancer research experiments increased and the fast development of cell culturing technologies began. The American Type Cell Collection (ATCC) was founded in 1925 and 60 years later, the European one (European Collection of Authenticated Cell Cultures - ECACC) was also established. These developments contributed to the discovery of the first oncogene related to the first viral carcinogenesis experiments with Rous sarcomas in chicken (27). Carcinogenesis-related experiments and implantation of human tumours in immunosuppressed animals were developed, however, all these models have some limitations which need to be considered (28,29,30).

For decades, researchers’ studies relied on a combination of cell culture and animal models for studying cellular mechanisms that lead to human diseases. These models have limited ability to recapitulate the complex tissue microenvironment, organ and body structures. The recognition of the cellular properties and role of tissue microenvironment have motivated the development and use of new techniques/3D cell culturing technologies/biofabrication in more complex in vitro models (31). However, the clinical translation ability of animal models to human disease has been questioned, in correlation with low success rates of clinical phase trials after many promising animal experiments (32). Furthermore, the high cost, the strict ethical regulations and concerns in animal testing are initiated to find less expensive, more predictive and human cell- and extracellular matrix (ECM)-based alternatives. 3D in vitro models are proposed to be bridges between cell cultures and in vivo animal models and even human trials. Early 3D culturing studies involved explanted host tissue, slice cultures and different spheroid cultures using hanging drop cultures, ultra-low attachment plates (ULA), and natural biopolymers (e.g., collagen, cell-extracted native extracellular matrix, matrigels, or polyacrylamide, agarose gels).

Regarding the increasingly well-known and renewed characteristics of various tumour tissues (33), as well as the enormously intensive therapeutic developments, their cost/benefit ratio, and many unsatisfactory clinical test results clear that new models switch to 3D technological platforms are necessary to better understand the background of the resistance mechanisms and tumour evolution. Mimicking the native cellular environment as precisely as possible is the first fundamental step to developing 3D human disease model systems (34). Recent newly developed 3D cell culturing methods (even combined with organoid or stem cell technologies) could be better and hopefully help more in drug screening in the preclinical phase and personalised treatments, as well (35).

The organoid cultures re-appeared in the late 1950s (e.g., Lasfargues established mammary organoids), and after the emerging technologies, they have been widely applied to support pluripotent stem cell culturing and differentiation studies since the late 2010s. Regarding the results of many 3D culturing systems, the tissue-like structures of 3D cell cultures except organoids are far from in vivo or in situ environment. Additionally, there are certain in vivo models in which human tumours such as xenografts in SCID (severe combined immunodeficiency) mice or initiated tumorigenesis could also be investigated. Although it is getting more and more challenging to get permission to perform animal experiments; moreover, there is a need to make efforts to apply the 3R rules (replacement, reduction, refinement) as a consequence, in vivo studies are more and more difficult to perform. Besides, interpreting the observations made in animal experiments also poses a challenge since animal model systems could neither represent the human cellular microenvironment nor the ECM (35, 36). Comparing different options for 3D culturing and even the in vivo models, the advantages and disadvantages of the given experimental system have to be taken into account (Figure 2). Accordingly, the limitations of model systems have to be considered during data evaluation. 3D models are more applicable for reproducing the mechanical and biochemical characteristics, e.g., cell-cell/matrix connections, tissue stiffness, and the gradient distribution of certain factors in tumour tissues (37–39).

Among the new cutting-edge solutions, 3D bioprinted models and the continuous technological developments of tumour models could make novel opportunities for more effective pharmaceutical tests and even for the testing of personalised therapeutic alternatives (35). Combining new cell culturing technologies with applying more complex organoid cultures is expected to result in significantly better disease models. In 1992, the Episkin technique initiated the use of synthetic tissue-like cultures with the application of different cell layers mimicking the epidermis-like structures.

A further breakthrough in tissue engineering and biofabrication is the establishment of 3D bioprinting - Following the first stereolithographic (SLA) printer designed by Hull in the 1980s, the initial developmental steps of bioprinting were formed rapidly, and these coincided with the advancement of 3D printing. The primary bioprinting process, only in two-dimensional (2D), can be attributed to the research group of Thomas Boland (2003) (40). Soon after that, the first 3D bioprinter was also completed, which was able to print more complex multi-layered structures. It is well-known that the Hungarian-born Gábor Forgács and the company named Organovo had a significant role in the development of the 3D bioprinting platform (41). Furthermore, the various bioprinters that have been widespread since 2010 have established the conditions for the creation of 3D printed mini-organs, or even complete organs, with which the printing of “beating hearts”, future transplantable hearts, or other replaceable tissues can be started (42–44). In parallel with these developments, various living tissues, prostheses, and other bioprinted structures and devices have appeared. Such devices are not only being developed, but are already in use in regenerative medicine and dentistry (45, 46), and there are also considerable efforts in the development of testing active drug ingredients (organ-on-chip technology for drug toxicity tests) or drug formulation (e.g., 3D bioprinted pills) and the improvement of various disease models. Additive manufacturing (47), 3D printing and thus 3D bioprinting technology can enable the designed structure to be generated rapidly layer-by-layer by applying computer-aided design (CAD).

Several methods - Have been developed which can be classified based on the printing technology with cell-contained bioinks: 1) drop-based (inkjet- or laser-based bioprinting), 2) extrusion-based, and 3) digital light processing (DLP) or SLA printing. In contrast with other printing methods, the benefits of the most common extrusion-based printing technology are the choice of resolution (fibre thickness)—which of course depends on the material –, fast execution, and relative cost-effectiveness (48). Traditionally, the materials used in 3D printing for medical purposes were inert and cell-free, like plastic (49). As a result of the developments, other types of materials are now also used during 3D bioprinting as biocompatible or biodegradable materials for creating implants or soft tissue reconstruction (e.g., bone replacements, reconstructive plastic surgery) (50, 51). One of the newest areas that have developed from 3D printing is 3D bioprinting with live cells, during which so-called bioinks consisting of cells mixed in a matrix material are used for printing. Applying these, we can create a living tissue-like structure in the course of their long-term in vitro culturing (52). The requirements for bioinks were summarised by Groll et al. as follows: bioink is a cellular preparation that contains biologically active components and biomaterials and can be used in automated production technology (53). The primary criteria for an ideal scaffold or bioink are to provide a suitable environment for cell adhesion, proliferation, differentiation, and migration, as well as cell-matrix interactions. Tumour cells form and develop in a rather complex, multicellular-originated, heterogeneous environment which is made up of various cell types and extracellular matrix components.

Different bioinks - Used in the field of tumour modelling are biomaterials that consist of mixtures of hydrogels and primarily tumour cells and tumour-associated normal cells. Numerous bioinks are biomaterials that are combinations of hydrogels and cells. Bioinks can be made from naturally occurring ingredients, but synthetic versions are also available (54). It is required that the substances of bioinks have appropriate mechanical and biodegradable characteristics. For example, printability and later shape retention are important, so the bioink must meet certain mechanical conditions, e.g., easy application and a high degree of shape fidelity. It is not negligible that the materials used for printing are not allowed to cause cell death, thus these have to be biocompatible or biodegradable even after the printing process (55, 56). The most frequently used components of bioinks are the hydrogels, of which the main ingredient is water, imitating the natural cellular environment (57).

Natural hydrogels differ from synthetic ones; bioinks belonging to the former type have a limited mechanical force, however, these degrade rapidly. In contrast, non-natural bioinks have less biocompatibility, but these can be characterised by great printability and appropriate mechanical properties. Consequently, certain researchers apply a mixture of natural and synthetic gels, a so-called hybrid bioink exploiting their advantageous features. Bioink-selection depends on multiple factors: 1) the type of bioprinting technology, 2) the characteristics of the model tissue (e.g., stiffness, elements in the microenvironment), 3) the necessity of shape-preserving, and 4) the appropriate crosslinking process and application, the influencing effects on cell proliferation, differentiation, and survival. Crosslinking stabilises the 3D bioprinted structure, preserving the viability of the cellular elements of the bioink. This causes physical and chemical modifications in the bioink which ensures that the printed layers remain together. There are several options for crosslinking: e.g., enzymatic (fibrinogen + thrombin = fibrin); ionic (alginate—CaCl₂); chemical (alginate—horseradish peroxidase); physical (gelatine-methacrylate—UV), or thermal (gelatine—high temperature) which can be performed either before or during or even after the printing process is finished (58). 3D bioprinting is a very effective tool, however, the standardisation of 3D bioprinting protocols is essential, and additionally, multi-faceted improvements are also required: 1) developing printing protocols, 2) standardising the materials used as bioinks, 3) creating novel biomaterials which have more sophisticated physical and biological properties, 4) improving the usability of 3D printed structures, as well as establishing test systems required for these (59, 60).

Living cell 3D bioprinting technologies require some major tools - An appropriate number of cells, bioinks, bioprinters, the specific printing design, and pre/post-processing before and after bioprinting. Additionally, biological tests and other biochemical, molecular biology or morphology studies also need special handling with the printed 3D materials. The development of this technology has required newly printable and biocompatible (adequate cell-compatible) materials. 3D bioprinting is easier with cancer cells in correlation with their unlimited cell growth supporting the production of a huge number of cells for bioinks. Additionally, we could combine cancer cell bioinks with cancer-associated and/or “normal” cells, and finally, the layer-by-layer bioprinted materials can form living cancer tissues during longer in vitro culturing.

Applying traditional 3D culturing so-called spheroid cultures, e.g., where cells are maintained in non-adherent plates or hanging drops (without applying matrix-embedded technology, only with the cells’ own matrix production) created by manual seeding of the cells, results in non-uniform cell distribution. Spheroids developed in these types of culturing methods could have various shapes, moreover, final analyses show high statistical deviation in the experimental datasets. For modelling the heterogeneity and the real in vivo situation of cancer tissues, new biofabricated models are required considering to achieve better complexity and standardisation. New approaches—forming small concave wells using a bioprinter, casting forms (grids) for cells or printing similar drops with bioinks (with homogenous cell concentration)—were started to be used in several tumour types (e.g., glioma, sarcoma, breast or cervical cancer cells) (61). The widespread extrusion-based and droplet bioprinting technologies provide more homogenous, controllable size, cell number and shape distributions, as well as potentially “real” tissue formations for tumour biology studies. One of the first layer-by-layer cell printing applications described better cell seeding uniformity and long-term viability (>90%, 14 days) of the printed primary cells (62). Additionally, after 2010, in some studies, magnetic levitation of the tumour cells and fibroblasts was applied aiding the formation of tumour spheres (breast cancer) with defined cellular composition and density by Leonard’s method (63). Others started to print 3D printed and in vitro cultured models with different cancer cell lines (64). In this decade, several functional tests were reconsidered, e.g., bioengineered 3D bone-mimetic structures have been started to be used for bone metastasis models (65), as well. These bone-mimetic and osteogenic niches are developing and are useful to test the effects of new compounds in vitro with potential inhibitory effects in breast tumour cell colonisation in the bone (66). In 2017, researchers started to fabricate multi-cellular bioprinted models for drug penetration and toxicity tests to investigate tumour tissue-like and liver tissue models (67, 68). Organovo and some other companies developed new 3D bioprinted liver tissues for drug tests (69). Several tissue-mimetic structures have been developed combining stem cells with organoid, spheroid cultures or 3D bioprinting in biofabrication. Organoid technologies have recently been exponentially developing; however, these technologies are very expensive and need a special supplement and careful handling. These technologies and different types of cancer cells (e.g., glioma, neuroblastoma, breast, kidney, colorectal, and lung cancer models) have been applied. Furthermore, rare tumours have to also be represented in these innovative approaches (e.g., chondrosarcoma, pelvic carcinomas) (70).

Analysing the increasing number of 3D bioprinting-related publications - About 500 publications are available regarding cancer research and 3D bioprinting shows that about 1/3 of the publications are reviews or not cancer-specific papers (only mention cancer research as a potential another research area where 3D bioprinting technologies are spreading or could help develop new cancer models). Among these ∼500 publications, there are about 180 original experimental papers mentioning cancer models and 3D bioprinting. In these studies, breast- and lung cancer, and brain tumours are represented in higher numbers (∼40 breast cancers or ∼20 each of lung cancers and central nervous system malignancies), but in these studies mainly only one tumour type was used in correlation with the presented research projects until the end of 2021 (Figure 3).

After several pioneer 3D cell culture studies, the development of new breast cancer models - Was started to be tested, therefore, we summarise the recent history of breast cancer model development. Different 3D cultures, 3D basement membranes, and 3D hydrogel scaffolds with breast cancer cells highlighted the importance and alterations of many different behaviours (metabolic alterations, adherence, migration, sensitivity etc.) and the role of cancer microenvironments in vitro models and suggested engineering for 3D tumour models (71, 72). After creating manually prepared breast tumoroids in matrigel, Swaminathan and colleagues investigated directly bioprinted breast epithelial spheroids with different bioinks and combined these spheroids with HUVEC cells as co-cultures. They validated the remaining viability and analysed the morphology of these spheroids (73). Reid et al. described and demonstrated that using a small number (defined concentration) of tumour cells in bioprinted collagen gels, forms tumoroids and could mimic lumen formation when cancer cells are co-printed with “normal breast epithelial cells” (74). In further works, Mollica et al. tested different hydrogels and decellularised breast tissues with human breast cancer cells using 3D bioprinting and showed that using these technologies they could maintain and culture large 3D bioprinted organoids and tumoroids (75). They described the direct bioprinting technology of 3D multi-cellular breast spheroids with endothelial cells (76). In another breast cancer model work, tumour cells and adipocytes were used and monitored the viability of cells after printing in a 10-day culturing course. They described that direct printing of these co-cultures resulted in morphology, localisation, and distribution changes of cells in the printed tissue-mimetic structures (77). An important topic of the bioprinting technique is the vascularised structure development which could increase the complexity of the generated models. There are many attempts with magnetic-ring and coaxial bioprinting to build some tubular structures between spheroids covered with endothelial cells, in vitro forming tissues (78,79,80,81). Cancer-associated adipocytes have a special role in the tumour microenvironment, these cells are involved in some new models using co-printed, co-cultured breast cancer cells with adipocytes to increase the similarity to real tissue environment (82). Besides, there are efforts to build more complex tissue-mimetic structures with stromal cells (fibroblasts), breast cancer cells, and endothelial cells to increase the phenotypic similarities in vitro in 3D bioprinting protocols (83).

The central nerve system (CNS)-derived 3D models - Have also been developed regarding the need for new and more effective targets and therapeutic treatments. These models require the reconstitution of the complexity and heterogeneity of glioblastoma and neuroblastoma tissues, and additionally, the special tumour-stroma interactions and blood-brain barriers for accelerating potential therapeutic interventions (84). These models represent scaffold-free 3D bioprinted spheroid cultures or co-printed, co-cultured glioblastoma or other CNS malignancy-derived tumour cells (e.g., neuroblastoma) cells with the associated macrophages, stromal cells to build the special microenvironment (85,86,87,88). To develop more reliable models in these diseases, patient-derived newly isolated glioma cells are preferred instead of traditional glioma cell lines which were long cultured in 2D cultures (89), and another research direction is to combine these with 3D bioprinted blood-brain barrier models (90, 91).

There are new efforts to validate bioprinted organoids/spheroids in 96-well plates drug screening with several cancer cell lines and patient-derived tumour cells including carcinomas, glioblastomas, sarcomas, and melanomas, respectively (92, 93). Additionally, many new bone metastatic behaviour tests or cancer tissue-associated tumour antigenicity tests have been developed using 3D bioprinters and breast cancer models (94–96). These models and their applications combined with microfluidics and chips, such as tumour-on-chip, will revolutionise both drug and patient-derived tests in the near future (34,97–99).

Overall, the feasibility and complexity of biofabricating multi-cellular, cell-laden bioprinted tissue-mimetic models with the real human microenvironment of breast tumours have started to be developed in many laboratories, however, standardising these is necessary for future applications in different high-throughput drug tests and patient-derived tumour model developments (100, 101).

To complete this short review about the developing 3D bioprinted cancer models, we show our new 3D bioprinted breast cancer model and some new experiences with its establishment and suggestions for others who start applying this cutting-edge technology shortly. In our previous work, we described the metabolic characteristics and differences of the already used breast cancer models and highlighted that 3D bioprinted models are closer to the in vivo situation than the others. In this review, we supplement these data with the already followed long-term growing capacity and the in situ expression patterns and differences of some previously not studied proteins and their heterogeneity in the 3D bioprinted breast cancer model. Additionally, differences in doxorubicin and rapamycin sensitivity among 3D bioprinted, traditional 2D cell culture systems and in vivo xenograft models using luminal B subtype human breast cancer cells could also be highlighted.

ZR75.1 (ATCC-CRL1500), luminal B subtype human breast cancer cell line was used in our experiments. Cells were grown in 10% foetal bovine serum (FBS, Biosera), glutamine (2 mM) and gentamycin-contained RPMI-1640 media (Biosera—Nuaille, France) at standard cell culture conditions. Different treatments were applied in 96-well plates, tissue culture flasks, and 3D bioprinted tissue-mimetic scaffolds. Before the treatments, the bioprinted scaffolds were maintained for 7 days, while the cells in 2D cell cultures were incubated for 24 h. After media refreshment, the 72-h treatments were carried out in 96-well plates, tissue culture flasks, and on scaffolds (minimum 6 parallels maintained in every well of 6-well plates) for sensitivity tests. mTORC1 inhibitor rapamycin (Rapa; 50 ng/mL; Focus Biomolecules, Plymouth Meeting, PA, United States), the natural anthracycline antibiotic and chemotherapeutic agent doxorubicin (Doxo; 50 ng/mL; TEVA, Debrecen, Hungary) and their combinations were applied regarding our previous mTOR inhibitor combination sensitivity studies (102).

For printing, two types of bioinks were used: a. cell-containing-gel: 3% alginate and 1% gelatine (Merck-Sigma-Aldrich, Darmstadt, Germany) bioinks were mixed with cells (1 × 10⁷/mL) immediately before printing; b. cell-free-gel, more rigid gel, 6% alginate and 11% methylcellulose (Merck-Sigma-Aldrich). The scaffold layout (6 layers alternately) was designed with GeSiM Robotics software and performed by an extrusion-based bioprinter (Bioscaffolder 3.2, GeSiM, Radeberg, Germany). The printing conditions were the following: radius and height (2.5–5 mm, 0.5–1 mm); interlayer angle (90°); the distance of infill (1.5 µm); printing speed (10 mm/s); needle diameter and height (110 µm for cell-free-gel and 50–50 µm for cell-gel); pressure (400 kPa for cell-free-gel and 20 kPa for cell-gel). The scaffolds were post-processed by CaCl₂ crosslinking (200 mM, 2 min) and washed twice then maintained in culture media (103).

For quantitative analysis of cell viability and proliferation, Alamar Blue (AB) and Sulforhodamine B (SRB) assays were used regarding mainly to standard protocols—the 3D scaffolds were transferred into new 96-well plates (1 scaffold/well/100 µL media) directly before the measurements. The fluorescence (change) of AB (Thermo Fisher Scientific) was measured after a 4-h incubation, the detected signs were evaluated as relative fluorescence units (RFU) using a fluorimeter (570–590 nm; Labsystems International; Ascent software v. 2.6—Vantaa, Finland) culture system independently. In SRB assays, 10% trichloroacetic acid (60 min; 4°C; Merck-Sigma-Aldrich) fixation, washing steps, and overnight drying were applied before SRB (Merck-Sigma-Aldrich, 0.4 m/V % diluted in 1% acetic acid; 50 µL/well; 15 min for 2D and 1 h for 3D plate; RT) staining. The washing steps were performed carefully with 1% acetic acid then plates were left to dry overnight again (the overnight drying has special importance in the case of scaffolds). The bound SRB was re-dissolved in Tris base solution (10 mM; 150 µL/well, Merck-Sigma-Aldrich) and measured by LabSystems Multiskan RC/MS/EX Microplate Reader (570 nm; Labsystems International; Transmit Software Version 4.5—Vantaa, Finland). Relative cell proliferation was calculated in the percentage of control cells. The proliferation assays were performed with six parallels in three independent experiments.

The in vivo experiments were performed in the Animal Care Facility unit located at the Department of Pathology and Experimental Cancer Research Institute (permission—No# PEI/001/1733-2/2015), the experimental protocols were approved by the Institutional Ethical Review Board (permission—No# PE/EA/801-7/2020). To create ZR75.1 human xenograft mouse models, 2.5 × 10⁶ cells in 100 µL RPMI-1640 media were implanted subcutaneously into the breast region of 8-week-old female SCID mice. The size of the tumours and the body weight were registered in a 3-week treatment course. Afterwards, the mice were sacrificed and tumours were removed, formalin-fixed, paraffin-embedded (FFPE) and sectioned or freshly frozen and lysed for further analyses. Tumour volume calculation was performed using the following equitation: π/6 × (2×shorter diameter + longer diameter³)³.

Immunohistochemistry was carried out using FFPE scaffolds, xenografts, and 2D cell culture slides (gel-embedded spheroids—10⁵ cells/mL) or cytospins. After deparaffinization, antigen retrieval was applied on FFPE materials (citric acid, pH 6, 30 min, pressure cooker). The fresh culture slides and cytospins were fixed in ethanol. After endogenous peroxidases and aspecific immunoreactions blocking, the used primary antibodies were the following: anti-ALDH1 (Cell Signaling; #54135; 1:200), anti-cleaved-caspase-3 (Cell Signaling; #9661; 1:1000), anti-COXIV (Cell Signaling; #4850; 1:2000), anti-LDHA (Cell Signaling; #3582; 1:400), anti-phospho-histone-H3 (Cell Signaling; #9701; 1:100). To visualize the reaction, Novolink™ Polymer Detection Systems (Leica Biosystems, Wetzlar, Germany) were used with 3,3′-diaminobenzidine (Dako, Carpinteria, CA, United States) and haematoxylin counterstaining. The stained slides were evaluated using Pannoramic Viewer Software (3D Histech). Quantitative protein expression from the cell, tissue, and scaffold lysates was investigated with WES^TM Simple (ProteinSimple 004-600; Minneapolis, MN, United States) a fully-automated Western blot system regarding the instructions of manufacturer using Anti-Rabbit or Anti-Mouse Detection Kits (ProteinSimple DM-001, DM-002). The samples were processed in a 12–230-kDa Separation Module (ProteinSimple SM-W004). To dissolve the 3D bioprinted scaffolds, sodium citrate was added to each sample (0.1 M, 30 min, RT), then the samples were centrifuged and lysis buffer (Tris—50 mM, pH 7.5, glycerol 10%, NaCl 150 mM, Nonidet-P40 1%, NaF 10 mM, phenylmethylsulfonyl fluoride 1 mM, Na₃VO₄ 0.5 mM) was added to the pellet. To evaluate protein content, Bradford reagent was used (Bio-Rad, Hercules, CA, United States). The primary antibody dilutions were 1:50 (applied primary antibodies: anti-COXIV, anti-LDHA, anti-Rictor (Cell Signaling; 2140)) and β-actin (Cell Signaling; #4970) was the normalisation control in our measurements. The electropherograms were analysed with Compass software 6.1.0 (San Jose, CA, United States). The original WES^TM Simple graphs were attached to Supplementary Documentation.

To calculate standard deviations (SD) and mean values, the results of three independent experiments with three or more parallels were evaluated. Statistical analysis was performed using PAST (version 3.24) software. Data evaluation of in vitro experiments was performed using Student’s t (two-tailed) test. Statistical significance was defined as p < 0.05.

The included figures (Figures 1–3.) were created with BioRender (https://biorender.com) covered by the Department of Pathology and Experimental Cancer Research, Semmelweis University, institutional license.