An oxidized dextran and thiolated chitosan-based hydrogel driven biomimetic triple negative breast cancer 3D in vitro model for cancer progression and therapeutic studies

Abstract

In the advancing field of in vitro cancer modeling, three-dimensional (3D) culture systems are increasingly recognized for their ability to recapitulate critical tumor-specific characteristics. Given the aggressive nature and high mortality associated with triple-negative breast cancer (TNBC), there is a pressing need to develop physiologically relevant 3D in vitro models that effectively simulate key tumor promoting factors (TPFs). This study presents a modified dextran–chitosan (MDC) hydrogel with engineered non-fouling properties that supports the formation of MDA-MB-231-derived 3D tumoroids. The hydrogel facilitated upregulated expression of extracellular matrix markers, including COL1A1 (2.29-fold↑) and FN1 (0.84-fold↑). Cell proliferation within 3D cultures was significantly reduced on days 2 (p < 0.001), 4 (p < 0.0001), and 6 (p < 0.001) compared to 2D cultures. Enhanced hypoxic conditions (based on EF5 adducts’ fluorescence; p < 0.0001), epithelial-to-mesenchymal transition (EMT) traits, and stemness marker expression [e.g., NANOG (3.33-fold↑)] were observed in 3D tumoroids. Additionally, the 3D tumor microenvironment showed elevated activity of key TPFs, including IL6, IL10, TNFA, FGF2, BMP2, and active TGFB (p < 0.0001). The MDC hydrogel, with stiffness mimicking breast tissue (∼11 kPa), also promoted mechanotransducive signalling, evidenced by increased YAP1 expression (2.4-fold↑) and a significantly elevated nuclear-to-cytoplasmic YAP1 ratio (p < 0.0001) relative to 2D cultures on TCPS (∼3 GPa). Whole transcriptome sequencing and gene set enrichment analyses further validated the enhanced tumorigenic phenotype of the 3D model. Moreover, the 3D tumoroids exhibited significant resistance (p < 0.001) to combined doxorubicin–paclitaxel treatment. Thus, the MDC hydrogel-based 3D TNBC model emerges as a robust and scalable platform for anticancer drug screening, evaluating precision medicine and investigating cancer biology.

---

1. Introduction

The complex architecture and dynamic environment of tumors poses significant challenges in cancer modelling, particularly in the recreation of the interaction between cancer cells and their surrounding microenvironment.¹ The extracellular matrix of tumors drives metastasis and cancer progression by facilitating biochemical and mechanical cues.^2,3 The factors that aid in the process of tumor growth further leading to cancer progression are considered as tumor promoting factors (TPFs). While foundational in cancer research, the traditional in vitro 2D platform fails to mimic the physiological structure and heterogeneity of the tumor, lacking cell–cell and cell–ECM interactions essential to capture in vivo like tumor properties.^4,5 The lack of dimensional complexity limits their ability to accurately predict therapeutic responses and the underlying biological mechanisms, whereas using in vivo animal models presents ethical concerns, substantial cost,⁶ and discrepancies in the immune responses.⁷ These limitations highlight the need for in vitro models that can better replicate the tumor microenvironment along with pertinent TPFs to provide a reliable platform for studying cancer biology and screening therapeutics.

Recent advancements in 3D cell culture technologies revolve around scaffold-free and scaffold-based 3D cell culture aiding systems, where scaffold-free systems function more on physicochemical cues but lack biomechanical and biochemical cues, which can be aided by the 3D matrix used in scaffold-based systems. The commercially available basement extracts such as Geltrex®, Matrigel®, and Cultrex® possess the limitations of batch-to-batch variations and possible immune reactivity.^8,9 As an alternative, the synthetic polymeric scaffold usage nullifies batch-to-batch variation possibilities, but they do not fully capture the intricate biological complexity of native tissues and often exhibit a decrease in cell survival.^10,11 On the other hand, natural polymer based scaffolds lack mechanical strength. Thus, there is a need to develop innovative polymeric alternatives to address the aforementioned challenges. Also, to successfully establish 3D cancer spheroids/tumoroids using a 3D cell culture system, it is crucial to evaluate important TPFs in it so that a physiologically pertinent 3D in vitro model is achieved and can be relevantly translated for cancer-related in vitro studies. Most of the natural polymeric scaffold-based platforms, which have been reported to avail 3D spheroids, such as agarose based microwell plates,¹² Matrigel®,¹³ 3D printed alginate–gelatin–Matrigel hydrogels,¹⁴ and GelMA-based micropatterned substrates,¹⁵ have explored limited tumor promoting factors before declaring the aided 3D spheroids/tumoroids as 3D in vitro models. Given these constraints, there is growing interest in developing an innovative scaffolding system for 3D cell culture that require simple processing and do not rely on high-end equipments. Such systems aim to integrate key features such as mimicry of extracellular matrix composition, biochemical and mechanical cues, tissue-specific architecture, and cell–ECM interactions.

To address this challenge, the present study reports a modified natural polymeric duo – oxidized dextran (OD) and thiolated chitosan (TC) – capable of forming in situ crosslinking in an aqueous medium at neutral pH to create a macroporous modified dextran–chitosan (MDC) scaffold [in freeze-dried format], which can aid formation of triple negative breast cancer (TNBC) in vitro 3D model. Unlike the hydrogels developed based on natural polymers for 3D cell culture, which often lack comprehensive biological validation of the resulting 3D spheroids/tumoroids, it remains an open question whether they can be considered physiologically relevant 3D in vitro microtumors. The present study aims to evaluate crucial TPFs – extracellular matrix (ECM) deposition, proliferation potency, EMT characteristics, hypoxia and hypoxia-inducible factor (HIF), and cancer stemness in the MDC hydrogel [dipped in aqueous medium PBS/culture media] aided MDA-MB-231 3D tumoroids. To establish germane TPFs, a drug cocktail consisting of doxorubicin (DOX) and paclitaxel (PTX) was tested for its synergistic efficacy on the 3D in vitro TNBC model compared to 2D cultured cells. Importantly, the MDC scaffold and the cell dimensionality-driven mechanotransduction effect imposed on the 3D cultured cells compared to that on the 2D cultured cells have been demonstrated through YAP1 expression and cell localization. Thus, the present study underscores the advantage of the MDC hydrogel in overcoming the limitations of batch-to-batch variations and chemical ambiguity of Matrigel®,¹³ the requirement for equipment like 3D printing and expertise to develop a 3D in vitro model,^16,17 and the need for photopolymerization in the case of gelatin methacryloyl (GelMA) based hydrogels for gelation.¹⁸ The MDC hydrogel uniquely integrates reproducibility, process-simplified fabrication, the use of ECM mimicking modified natural polymers, tunability, macroporosity and scalability, thus serving as an ECM mimicking, biochemical and biomechanical cue delivering hydrogel aiding the formation of a physiologically mirroring in vitro TNBC model with suitable TPFs for efficient anti-cancer drug screening considering the aspect of scaffold-driven cell dimensionality.

2. Materials and methods

2.1. Materials

2.1.1. For in vitro studies. MDA-MB-231, a TNBC cell line, was procured from the National Centre for Cell Science (NCCS), Pune, India. Antibiotic–antimycotic, Dulbecco's modified Eagle's medium with high glucose concentration [DMEM – HG (4.5 g L⁻¹)], 0.25% trypsin-EDTA, and fetal bovine serum (FBS) were purchased from Gibco® and Thermo Fisher Scientific (USA). The cell counting kit (CCK8), 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), doxorubicin hydrochloride and paclitaxel were procured from Merck and Sigma-Aldrich. The NucleoSpin® RNA extraction kit was acquired from Macherey-Nagel. The kit for cDNA synthesis and the SYBR® Green supermix were purchased from Bio-Rad and Invitrogen, respectively. IL6 (Cat. No. BMS213-2), IL10 (Cat. No. BMS215-2), TNFA (Cat. No. BMS223-4), TGFB1 (Cat. No. BMS249-4), and FGF2 (Cat. No. KHG0021) ELISA kits were obtained from Invitrogen. The BMP2 ELISA kit (Cat. No. DBP200) was obtained from RnD Systems.

2.2. Methods

2.2.2. Fabrication of the MDC hydrogel and in vitro tumoroid culture. Chemical modification of dextran and chitosan was performed as mentioned in the previously reported studies.^4,19 Briefly, dextran was oxidized by NaIO₄, whereas chitosan was modified with cysteine amino acid using EDC based crosslinking (refer to S1 for detailed procedures, SI). The modified polymers were characterized by NMR and degree of oxidation analyses to confirm the oxidation of dextran (S2 and S3, SI) and determined thiolation of chitosan through Ellman's assay (S4, SI). Furthermore, using OD and TC [4% w/v of OD and 4% w/v of TC in a 1 : 1 ratio], an in situ crosslinking polymer mixture, referred to as modified dextran and chitosan (MDC), was obtained. A scaffold was eventually fabricated using the freeze-drying method (S5, SI). Moreover, the crosslinking of OD and TC and its incorporation in MDC were physico-chemically characterized by Fourier-transform infrared (FTIR) spectroscopy (S6, SI), differential scanning calorimetry (DSC) (S7, SI) and hydrophobicity measurement (S8, SI).

For biological studies, according to the number of cells required at the end of the target experiment, the number of MDA-MB-231 cells was counted and seeded on the hydrogel (1–2 × 10⁵ per 3D substrate) (see S9 and S10, SI). The same number of cells were seeded on the TCPS substrate as its 2D counterpart. Furthermore, after 6 days of culture on the hydrogel, the cells were trypsinized, harvested, and filtered using a 40 μm nylon cell strainer (HiMedia) to obtain a single-cell suspension and used for further experiments wherever isolated cells were demanded (termed 3D cultured cells).⁴ To isolate 3D tumoroids, the hydrogel containing culture vessel was gently taped and most tumoroids were retrieved using slow pipette suction with a blunt 1 ml tip or flushed with media if needed.

2.2.3. Mechanical stiffness determination of the hydrogel. Following the protocol reported by Liu Y et al., with some modifications, hydrogels (dipped in dPBS) (n = 3) of approximately 1 × 1 cm² size were subjected to uniaxial compression using a mechanical testing machine (Instron 5943) in a confined environment.²⁰ Samples were compressed at a rate of 1 mm min⁻¹ until the samples fractured. Furthermore, Young's modulus was determined from the slope of the compressive stress vs. compressive strain curve from the curve segment between 0.02 and 0.12 mm mm⁻¹. Furthermore, swollen hydrogels were subjected to repetitive cyclic compression for 10 cycles up to 80% compression,²¹ with a 1 h 30 min wait time between cycles at a rate of 2 mm min⁻¹. All the representative graphs were plotted using GraphPad Prism 8 software.

2.2.4. Morphological evaluation of the hydrogel and 3D tumoroids through microscopy. 2 × 10⁵ MDA-MB-231 cells were seeded on the hydrogel and the tumoroid formation was monitored using a bright field microscope (Zeiss Axiovert 40C) and a scanning electron microscope (SEM) (Evo®18 [Carl Zeiss, Germany]). For SEM analysis, after 6 days of culture, the hydrogel consisting of 3D tumoroids was fixed with sodium cacodylate buffer (0.1 M)–glutaraldehyde (2%) with pH 7.2 for 1 h, followed by incubation with gradually increasing concentrations of alcohol series. The air-dried samples were sputter-coated with gold–palladium and analysed at 5 kV at different magnifications.⁴ The SEM micrographs were used to estimate the area of the pores of the MDC hydrogel and the area of tumoroids, morphological analysis of the 3D cultured cells, and the shape characteristics of individual 2D and 3D cultured cells were estimated from the SEM micrographs using ImageJ software opting for ‘Area’ and ‘shape descriptors’ in the ‘set measurements’ parameter under the ‘Analyse’ tool of ImageJ software.²¹

2.2.5. Estimation of tumor promoting soluble factors. IL6, IL10, TNFA, FGF2, BMP2 and TGFB1 released from the 2D and 3D tumoroids were estimated by quantifying the target analyte from the conditioned medium of 2D cultured cells and 3D tumoroids post 6 days of culture using ELISA kits with human specificity. Following the method described previously,^4,22 the numbers of 2D and 3D cultured cells post-harvesting were counted, and an equal volume of conditioned media obtained post-culture was utilized to measure the total protein in the collected cell conditioned medium quantified using the bicinchoninic acid (BCA) colorimetric assay [S11, SI] referring to the bovine serum albumin (BSA) standard curve [S12, SI]. The volume representing equal protein concentration from the 2D and 3D conditioned media was taken as a sample for analyte quantification through ELISA. As per the principle of sandwich ELISA, wells containing human IL6, IL10, TNFA, TGFB1, FGF2 and BMP2 from the conditioned media, avidin-HRP conjugate, further bind to the biotinylated detection antibody together and indicated the presence of analyte with a color change to blue. Lastly, the substrate–enzyme reaction was stopped by adding the stop solution. The color change from blue to yellow directly indicating analyte concentration was estimated by evaluating its absorbance at 450 nm using an H1 Synergy multimode microplate reader (BioTek, USA). The concentration of the analytes was quantified using the standard curve of the respective analytes derived using the given IL6, IL10, TNFA, BMP2, FGF2 and TGFB1 reference standards availed in the respective ELISA kits. The final quantitative level of the analytes was deduced after normalizing with the cell numbers of 2D and 3D cultured cells counted on the 6th day, from which the equal volume of conditioned media was obtained.

2.2.6. mRNA isolation and expression level quantification in 2D and 3D cultured cells. RNA from the 2D and 3D cultured cells was derived after equalising the isolated number of cells in each group. The final cell pellet for RNA isolation was obtained through centrifugation and then lysed using the lysis buffer provided in the NucleoSpin® RNA isolation kit and proceeded to the final RNA isolation step following the manufacturer's protocol. Later, the isolated RNA was quantified using Take3 plates in the BioTek Synergy H1 hybrid multimode plate reader and Take3 session using Gen5 1.11 software. Furthermore, utilizing the Bio-Rad iScript™ cDNA synthesis kit, cDNA from the isolated RNA was obtained. Moreover, the level of gene expression was quantified using quantitative real time PCR (qRT-PCR) [7500 Real-Time PCR system, Applied Biosystems] employing the PowerUp® SYBR Green Master Mix (Invitrogen). The qRT-PCR conditions implemented for the qRT-PCR run and the primer sequences of target genes are specified in S13 and S14 (SI), respectively. The Ct values of genes in each group were normalized with the GAPDH of the respective group and the fold change value was manually calculated using the 2^(−ΔΔCt) formula as described previously^23,24 and represented using GraphPad Prism 8 software.

For methods pertaining to whole transcriptome sequencing and analysis of 2D and 3D cultured cells, refer to S15 (SI).

2.2.7. Hypoxia detection through EF5. The method was inspired from the previous reports by Brader, P. et al., and Kelly, CJ. et al.^25,26 To detect hypoxic cells, the EF5 based hypoxia detection kit (Cat. No. EF5-30C3, Merck) was utilized. 500 μM EF5 working solution was prepared in cell culture media. The 2D cultured cells, dissociated and reseeded 3D tumoroids on a confocal dish, spheroids aided by a bioinert polymer coated μ-dish (Cat. No. 81150, Ibidi®), and cells incubated with cobalt chloride (CoCl₂) [positive control]²⁵ were treated with culture media containing EF5 and incubated for 4 h at 37 °C. Post incubation, cells were washed and fixed with 4% paraformaldehyde for 1 h. Furthermore, the samples were blocked using 1% BSA in PBST buffer to avoid non-specific binding for 1 h. Later, the samples were incubated with 75 μM anti-EF5 antibody (ELK3-51 tagged with cyanin 3) prepared in 1% BSA containing PBST solution for 6 h at 37 °C. Post incubation and a rinse with PBS, nuclear staining was performed using DAPI. Finally, the samples were analysed using Zeiss CLSM 780 with the blue channel – filter set 49 (Exc.: 365 and Em.: 445–450) and the red channel – filter set 20 (Exc.: 546/12 and Em.: 575–640). Furthermore, corrected total cell fluorescence (CTCF) was calculated from the fluorescence images of each group for EF5 and plotted as a bar graph using GraphPad Prism 8 software.

2.2.8. Immunocytochemistry based evaluation of MKI67 and YAP1. The proliferation potential of 2D and 3D cultured cells and mechanotransduction effects were assessed via MKI67 and YAP1 expression analyses, respectively, using immunocytochemistry (ICC). 3D cultured cells were reseeded on coverslips until adherence, fixed with 4% paraformaldehyde for 20 min, and permeabilized with 0.1% Triton X-100 for 15 min. Blocking was conducted with PBST buffer for 1 hour, followed by overnight incubation with MKI67 (MAB4190) and YAP1 (MA535282) primary antibodies at 4 °C.²⁷ Fluorescence tagging was achieved using Cy-5 anti-mouse (715-175-150, Jackson Immunoresearch) for MKI67 and Alexa 488 anti-rabbit (A-11008) for YAP1 for 2 h at RT. MKI67 samples also underwent phalloidin-FITC staining (MAB4190) for F-actin. DAPI was used for nuclear counterstaining. Fluorescence images were acquired using a Zeiss CLSM 780, and MKI67 CTCF was quantified using ImageJ and plotted using GraphPad Prism 8. Phenotypic descriptors (aspect ratio and perimeter) were analysed from phalloidin-FITC-tagged actin images, and YAP1 localization analysis details are available in S16 (SI).²⁸

CTCF = integrated density – (area of measurement × background mean fluorescence)

2.2.9. Proliferation of 2D and 3D cultured cells. Following the reported protocol²⁹ and manufacturer's instructions, the CCK8 assay was conducted to estimate the viable cell number at the 2^nd, 4^th and 6^th days of culture. An equal number of cells seeded on TCPS and the hydrogel were counted on the 2^nd, 4^th and 6^th day post-culture using the CCK8 assay. At the end of each time interval, CCK8 solution containing WST-8 dye stock solution was directly added to the existing culture media at a ratio of 10 μL (CCK8) to 100 μL (culture media). Samples were further incubated in a CO₂ incubator at 37 °C for 4 hours. The absorbance of the culture media was measured using a multimode microplate reader at a 450 nm wavelength. The number of cells in each sample on the 2^nd, 4^th, and 6^th days was determined based on the absorbance values measured using the cell counting kit-8 (CCK-8). These absorbance values were then compared to a standard curve generated by plotting absorbance against a known range of cell numbers. Using the corresponding linear equation, the number of viable cells on each day was calculated. A graph depicting the number of cells for 2D and 3D cultured cells on the 2^nd, 4^th, and 6^th days was plotted using GraphPad Prism 8 software. For the method followed to conduct proliferation analysis of 2D and 3D cultured cells through the MTT assay, refer to S17 (SI).

2.2.10. Cancer stemness related surface marker study through flow cytometry. After 6 days of culture, the 3D cultured cells were isolated from the hydrogel and parallelly an equal number of 2D cultured cells were collected post trypsinization. Then the cells were fixed and permeabilized using 1% paraformaldehyde and 0.1% Triton-X 100 for 20 min and 15 min, respectively. Furthermore, the samples were incubated with the Fc receptor inhibitor.³⁰ Then, the 2D and 3D cultured cells were individually divided into two groups and incubated with anti-human primary antibodies – CD133 (Cat. No. 11-1339-42, eBioscience^TM, Invitrogen) and the CD44-FITC (Cat. No. 555478, BD Biosciences) + CD24-PE (Cat. No. 555428, BD Biosciences) co-staining mixture independently for 1 h with gentle periodic mixing. After two washes, the cells were centrifuged and re-suspended in cell staining buffer until acquisition by flow cytometry [BD FACS Aria^TM Fusion].

2.2.11. Drug efficacy testing on 2D cultured cells and 3D tumoroids. First, different dosages (0.5, 1, 2.5, 5, 7.5 and 10 μM) of doxorubicin, paclitaxel and the drug cocktail (DOX. + PTX., 1 : 1) were individually screened on the 2D cultured MDA-MB-231 cells to estimate the IC₅₀ value. Later, the cocktail efficacy was tested on the 2D cultured cells, 3D tumoroids, and spheroids formed using an ultra-low attachment (ULA) U bottom plate (commercial control – Cat. No. 74925, Nunclon™ Sphera™, Thermo Fisher Scientific). The 2D cultured cells, MDC-aided 3D tumoroids, and ULA-aided spheroids were treated with 1 μM and 5 μM of the drug cocktail (1 : 1) for 72 h. Following the test period, 0.5 mg mL⁻¹ of MTT solution was added, and the mixture was incubated for three hours. Then, dissolution of formazan crystals was conducted using DMSO and absorbance reading was taken using a Synergy H1 multimode plate reader at 570 nm.³¹ Finally, the %cell viability of the treated groups with respect to the control (cells incubated with only culture media) was calculated and depicted using GraphPad Prism 8 software.

2.2.13. Statistical analysis. The statistical analysis was conducted according to the experimental groups and conditions (mentioned in the respective figure legends). All data were statistically evaluated to determine significant differences between the test groups. Data plots are depicted using mean ± standard deviation with error bars. Probability values are presented as follows: p < 0.05 (*), p < 0.01 (**), p < 0.001 (***) and p < 0.0001 (****).

3. Results

To develop an ideal 3D in vitro model, a close mimicry of the tumorigenic properties and the tumor microenvironment (TME) to physiological tumors is desirable. To achieve this, the substrate in this study, composed of oxidized dextran (OD) and thiolated chitosan (TC), was selected for its physicochemical properties, previously shown by our group to aid 3D cell culture through favourable surface chemistry.^4,19 This study investigates scaffold-driven formation of the 3D in vitro TNBC model and verifies the key tumor promoting factors (TPFs).

3.1. Physico-chemical characterization of the MDC hydrogel and 3D tumoroids’ development

The modification of dextran and chitosan to OD and TC was confirmed before preparing the MDC scaffold. The degree of oxidation to dextran was 48.041% (discussed in detail in S2, SI), which was further verified by NMR, where the spectra showed peaks at 4.6–5.8 ppm defining the hemiacetal formation from aldehyde–vicinal hydroxyl interactions (discussed in detail in S3, SI). The grafted cysteine concentration to chitosan was 4 μM g⁻¹ (discussed in detail in S4, SI). The oxidation and thiolation of the dextran and chitosan validated through FTIR (discussed in detail in S6, SI) with the detection of the signature functional groups of aldehyde and thiol, respectively, were adequate to make it an in situ crosslinking polymer at physiological pH in aqueous medium, later aiding the macroporous scaffold on freeze drying (S5, SI). On oxidation, OD acts as a crosslinker and on thiolation the chitosan becomes more soluble in an aqueous medium crosslinking through disulfide and imine bonds as confirmed by the FTIR data (discussed in detail in S6, SI). Furthermore, the crosslinking of OD and TC was verified through changes in the heat flow through the polymers, modified polymers and the freeze-dried MDC performed using DSC, indicating the amorphous nature of MDC (discussed in detail in S7, SI), which may contribute to its wettability as studied through the hydrophobicity index measurement (discussed in detail in S8, SI). Overall, the physico-chemical characterization of OD, TC and MDC revealed a successful modification of the polymers as well as their crosslinking, resulting in the fabrication of the MDC scaffold having surface chemistry that can influence cell adhesion, cell spreading, and aggregate formation when used as a 3D substrate.

The MDC scaffold, intended to be used as a 3D substrate, post freeze drying attained macroporosity with an average pore area of 4721.599 μm² [Fig. 1(A and B)]. The Young's modulus of the MDC hydrogel was ∼11.3 kPa [Fig. 1(C)]. Furthermore, during cyclic compression of up to 80%, the MDC hydrogel did not buckle and reabsorbed the supplied buffer, demonstrating an elastic and inter-crosslinked network of polymeric chains [Fig. 1(D) and S18, SI]. After the 6th day of MDA-MB-231 cell culture, compact three-dimensional micro-tumoroids were observed all over the MDC scaffold [Fig. 1(E) and S19(A and B), SI]. The area of the tumoroids calculated after the 6th day of cell culture from the SEM micrographs indicated that the MDC hydrogel aided tumoroids ranging from 1004.5 μm² to 33 674 μm², where a greater number of tumoroids were found to have an area between 2000 and 10 000 μm² [Fig. 1(F)]. The MDC hydrogel triggered cells to be present in the 3D form, which was characterised by defining their shape descriptors [S20, SI] compared to the MDA-MB-231 2D cultured cells [S21, SI]. The area of the 2D cultured cells was 437.29 ± 263.28 μm², whereas that of 3D cultured cells was 228.48 ± 112.5 μm². Furthermore, the average perimeter and aspect ratio of 2D cells were 92.81 ± 27.3 and 2.92 ± 0.72, whereas in 3D cells the values were 55.92 ± 14.8 and 1.24 ± 0.14, respectively. This was further supported by the circularity, roundness and solidity measurements, which were nearer to 1 for MDC cultured cells defining MDC cultured cells in the 3D form compared to cells cultured over TCPS.

Fig. 1 Results of the physical characterisation of the MDC hydrogel and morphological analysis of MDC hydrogel aided 3D tumoroids. (A) SEM micrograph of the hydrogel showing its porous surface topography (250× magnification, scale bar = 100 μm), (B) histogram showing pore area distribution, (C) stress vs. extension graph deduced from the MDC hydrogel cyclic compression analysis, (D) compressive stress vs. compressive strain graph to calculate the Young's modulus of the MDC hydrogel, (E) SEM images depicting compact 3D tumoroid formation of MDA-MB-231 cells after 6 days of cell seeding at 1000× magnification (scale bar = 10 μm) and (F) histogram showing the area of tumoroids calculated from multiple SEM images of 3D tumoroids (n = 37).

3.2. ECM, proliferation and EMT characteristics

The ECM deposition trait of the MDC aided 3D tumoroids was studied by analysing two crucial ECM proteins: collagen type I and fibronectin 1 at mRNA level. FN1 [Fig. 2(A)] and COL1A1 [Fig. 2(B)] were 0.84-fold and 2.29-fold more expressed in 3D cultured cells as opposed to that in the monolayer cultured cells. Furthermore, the differential expression of ITGB1 in 3D cultured cells was 0.14-fold elevated in contrast to the expression of ITGB1 in the 2D cultured cells [Fig. 2(C)]. Overall, the increased ECM components corroborate with the smooth deposition observed surrounding the 3D tumoroids as observed in Fig. 1(E) and S19(B) (SI).

Fig. 2 ECM component and related protein, proliferation and EMT characteristics of MDC aided 3D tumoroids. Differential (A) FN1, (B) COL1A1, and (C) ITGA1 gene expression in 3D cultured cells in comparison to 2D cultured cells analysed through qRT-PCR. Proliferation attribute evaluation through (D) fluorescence microscopy images observed with a 20× objective lens showing Cy-5 tagged MKI67 expression (red) in 2D and 3D cultured cells with phalloidin stained actin (green) and DAPI stained nucleus (blue) captured using a 20× objective lens, scale bar = 50 μm; (E) calculated CTCF converted to the relative fold change value for 2D and 3D cultured cells; (F) differential MKI67 expression in 3D cultured cells in comparison to that in 2D cultured cells derived using qRT-PCR; (G) graph showing cell proliferation of the 2D and 3D cultured cells assessed by estimating the number of metabolically active cells on the 2^nd, 4^th, and 6^th days using the CCK8 assay. EMT phenotype analysis through FITC-tagged phalloidin stained actin (H) 2D cultured cells – epithelial type (scale bar = 50 μm); (I) 3D cultured re-seeded cells – mesenchymal type enlarged with filamentous protrusion [inset image] (scale bar = 50 μm); (J) phenotypic descriptors measured for 2D and 3D cultured cells. EMT markers’ differential (K) CDH1, (L) CDH2, (M) VIM, and (N) MMP2 levels in 2D and 3D cultivated cells.

At the molecular level, the proliferation potential of 3D cultured cells was evaluated by analysing the MKI67-nuclear proliferation marker. From the immunocytochemistry based results for the expression of MKI67, the expression level of MKI67 was observed to be higher in 2D cultured cells than the 3D cultured cells [Fig. 2(D) and S17(A), SI]. MKI67 was 0.58-fold less expressed in the 3D cultured cells than in the 2D cultured cells (***p < 0.001) [Fig. 2(E)]. Similarly, the differential mRNA expression of MKI67 was 0.64-fold less in 3D cultured cells in comparison to the 2D cultured cells (***p < 0.001) [Fig. 2(F)]. Furthermore, the proliferation potency of cells existing in the 3D tumoroid form and cells cultured as the monolayer was estimated using the CCK8 assay [Fig. 2(G)]. The number of metabolically active cells counted on the 2^nd day post cell seeding for 2D cultured cells was 115 × 10³, whereas the number of cells in 3D tumoroids was 8 × 10³. On the 4^th day, 2D cultured cells proliferated to 210 × 10³, whereas 3D cultured cells proliferated to 15 × 10³. Furthermore, on the 6^th day, the 2D cultured cell population reached 305 × 10³, whereas it reached 16 × 10³ in the cells present in the 3D tumoroid. Thus, a significant difference in proliferation potency was observed on the 2^nd (***p < 0.001), 4^th (****p < 0.0001) and 6^th (*** p < 0.001) days between the 2D and 3D cultured cells. Also, remarkable was that the number of cells on the 4^th to the 6^th day in 3D cultured cells just increased from 15 × 10³ to 16 × 10³, which indicated that the proliferation rate also slowed down in the 3D tumoroids. Thus, the CCK8 results were further supported by the MTT assay [S17(B), SI]. The overall proliferation of 3D cultured cells was notably low compared to the 2D cultured cells. Being part of the 3D tumoroid along with the presence of its own developing TME, its influence on the EMT characteristics of the cells was evaluated. The actin filament staining demonstrated the epithelial-like morphology of 2D cultured MDA-MB-231 cells [Fig. 2(H)]. When reseeded, the 3D cultured cells isolated from the tumoroids showed more cells within the 3D cultured cell population having elongated and enlarged morphology denoted by the yellow dotted line [Fig. 2(I)], clearly indicating epithelial-like and mesenchymal-type mixed phenotypes. Cytoplasmic protrusions extending from the 3D cultured cells were also observed [inset, Fig. 2(I)]. Furthermore, this phenotypic switch was characterised by measuring the aspect ratio and circularity phenotypic descriptors of 2D and 3D cultured cells [Fig. 2(J)]. The aspect ratio of the 2D cultured cells was 1.46, whereas 3D cultured cells had an aspect ratio of 2.82. The higher the aspect ratio, the more the cell elongation. Additionally, the circularity of the 2D cultured cells was 0.75, while that of the 3D cultured cells was 0.44. Hence, the significant differences observed in the aspect ratio (**p < 0.001) and circularity (***p < 0.001) as well as the actin arrangement of the 3D cultured cells showed enhanced EMT phenotypic characteristics of 3D cultured cells compared to 2D cultured cells. This was further validated at the mRNA level. The CDH1 and CDH2 mRNA differential expression was 0.68-fold low (***p < 0.001) [Fig. 2(K)] and 0.54-fold high (**p < 0.01) [Fig. 2(L)], respectively, in 3D cultured cells in comparison to 2D cultured cells. Moreover, the differential MMP2 expression in 3D cultured cells was 1.25-fold more than that in the 2D cultured cells [Fig. 2(M)]. The VIM mRNA expression showed no significant difference (0.06-fold) between 2D and 3D cultured cells [Fig. 2(M)]. Thus, EMT markers’ expression and EMT phenotypic traits indicated more EMT characteristics of 3D cultured cells compared to the 2D cultured cells.

3.3. Tumor promoting secretome and hypoxia

For the closest mimicry of the TME – the key influencer of TPFs – some of the crucial TPGFs and cytokines were quantified from the collected conditioned media of 2D and 3D cultured cells. The IL10 [Fig. 3(A)] and IL6 [Fig. 3(B)] levels measured for 2D cultured cells were 0.116 pg mL⁻¹ and 5.73 pg mL⁻¹, whereas 5.727 pg mL⁻¹ and 62.28 pg mL⁻¹ were measured for the 3D cultured cells. Furthermore, 0.452 pg mL⁻¹ TNFA was secreted by 2D cultured cells in contrast to 3.93 pg mL⁻¹ secreted by the 3D cultured cells [Fig. 3(C)]. Moreover, BMP2 was quantified to be 5.60 pg mL⁻¹ and 68.05 pg mL⁻¹ in 2D and 3D cultured cell conditioned media, respectively [Fig. 3(D)]. Additionally, FGF2 estimated in the secretome of 2D and 3D cultured cells was 2.07 pg mL⁻¹ and 22.27 pg mL⁻¹ [Fig. 3(E)]. Importantly, the total TGFB1 (latent TGFB1 + active TGFB1, where activation was induced to latent TGFB1 with acid treatment) quantified was 17.8 pg mL⁻¹, whereas the total TGFB1 secretion spiked high in the 3D tumoroid secretome, which was 5305.3 pg mL⁻¹ [Fig. 3(F)]. In addition, the presence of active TGFB1 (estimated without acid induction) in the 2D cultured cell conditioned medium was 7.39 pg mL⁻¹, whereas an elevated concentration of active TGFB1 (79.72 pg mL⁻¹) was estimated in the 3D tumoroids’ secretome [Fig. 3(G)]. Furthermore, differential TGFB1 expression in 2D and 3D cultured cells through qRT-PCR showed a 10.46-fold increase in 3D cultivated cells in comparison to 2D cultured cells at the mRNA level [Fig. 3(H)].

Fig. 3 Tumor promoting secretome: (A) IL10, (B) IL6, (C) TNFA, (D) BMP2, (E) FGF2, (F) induced activated total TGFB1 and (G) microenvironment activated TGFB1 protein quantified in 2D and 3D cell cultured conditioned media through ELISA. (H) Differential TGFB1 mRNA expression in 2D and 3D cultured cells through qRT-PCR. Hypoxia evaluation; (I) fluorescence images showing hypoxia determination through EF5 adduct formation cyanin-3 (red) and nucleus staining with DAPI (blue) in 2D cultured cells and 3D cultured cells (captured through a 20× objective lens, scale bar = 50 μm); (J) mean fluorescence intensity of EF5 adducts formed in hypoxic cells tagged with anti-EF5 Cy3 measured in various experimental groups; (K) HIF1A mRNA expression analysed through qRT-PCR between 2D and 3D cultured cells.

Hypoxia, which is a signature marker of tumors and plays a vital role in cancer progression, was assessed in the MDC aided 3D tumoroids using EF5. EF5 selectively binds to the hypoxic cells and forms adducts. The fluorescence microscopy images clearly indicated comparatively more hypoxic cells in the MDC aided 3D tumoroids than in the 2D cultured cells [Fig. 3(I)]. A notably higher fluorescence intensity was observed in the dissociated tumoroids moving towards the centre of the dissociated tumoroids [Fig. 3(I), 3D-EF5 panel]. The EF5 adduct formation indicating hypoxia was also compared with the commercially available bioinert polymer-coated dish [S22, SI] with CoCl₂-treated cells as a positive control. The mean fluorescence intensity was calculated from the fluorescence images, quantitatively representing the hypoxia in the cells cultured in different culture systems [Fig. 3(J)]. The presence of hypoxia compared to 2D cultured cells was highest in MDC aided 3D tumoroids (****p < 0.0001). Moreover, to elucidate the effect of elevated hypoxia in 3D tumoroids at the molecular level, the HIF1A mRNA expression was quantified. HIF1A was 1.4-fold differentially more expressed in 3D cultured cells than in 2D cultured cells [Fig. 3(K)]. Thus, the study revealed elevated levels of tumor promoting secretome and hypoxia in MDC aided 3D tumoroids.

3.4. Cancer stemness and drug resistance

The expression of surface markers – CD133, CD44 and CD24 – representing cancer stemness was investigated in the 2D and 3D cultured cells through flow cytometry. A clear population shift towards the right, representing increased fluorescence intensity, indicated enhanced CD133 expression in the 3D cultured cells compared to that in the 2D cultured cells [Fig. 4(A), the CD133 panel]. The Q2 quadrant, designated as CD133^high cell population, contained only 9.53% of 2D cultured cells, whereas 85.99% of the cell population expressed CD133^high in 3D cultured cells. Additionally, the 2D and 3D cultured cells co-stained with CD44-FITC and CD24-PE [Fig. 4(A), the CD24 + CD44 panel] also clearly exhibited a slight increase in CD44 expression in 3D cultured cells (2.05%) in comparison to the 2D cultured cells (1.88%) in the Q2 quadrant designated for the CD44^high cell population. Cells with CD24^high (Q4) expression were 0.04% in 3D cultured cells and 0.015% in 2D cultured cells. Thus, cells with only the CD24^high trait were observed less in 3D cultured cells in comparison to the 2D cultured cells. Also, the co-stained cells, CD44^high/CD24^high (Q3) cells, did not drastically vary in 2D and 3D cultured cells. Furthermore, the cancer stemness of the 3D cultured cells was investigated at the mRNA level. The ALDH1A1 [Fig. 4(B)] and SOX2 [Fig. 4(C)] expression was 1.83- and 0.42-fold elevated in 3D cultured cells compared to that in the 2D cultured cells. Furthermore, NANOG overexpression was observed to be highest, i.e., 3.33-fold higher than in the 2D cultured cells [Fig. 4(D)]. Thus, cumulatively cancer stemness was verified to be increased in the 3D tumoroids compared to the monolayer cultured cells.

Fig. 4 Cancer stemness surface marker evaluation through flow cytometry. (A) Dot plot showing expression analysis of CD133 and CD24 + CD44 tagged 2D and 3D cultured cells. Differential (B) ALDH1A1, (C) SOX2 and (D) NANOG expression in 3D cultured cells in comparison to that in 2D cultured cells assessed through qRT-PCR. (E) Graph showing %cell viability after 1 and 5 μM drug cocktail treatment (doxorubicin : paclitaxel = 1 : 1) after 72 hours in 2D cultured cells and 3D tumoroids through the MTT assay. (F) Differential ABCB1 expression in 2D and 3D cultured cells aiding in the drug efflux action in 3D tumoroids.

After evaluating varied crucial TPFs, the MDC aided 3D tumoroids were verified to be used as in vitro TNBC models for drug screening. For this, initially, different dosages of doxorubicin (DOX), paclitaxel (PTX) and the DOX : PTX (1 : 1) drug cocktail ranging from 0.5 μM to 10 μM were tested independently on the 2D cultured cells. The %cell viability to control for each drug concentration after 24 h, 48 h and 72 h was analysed using the MTT assay [S23, SI]. The IC₅₀ values deduced from the DOX and PTX concentrations tested for 24 h, 48 h and 72 h [S24, SI] were further utilized as a ground for a drug cocktail study on 3D tumoroids. After 72 h of treatment, the IC₅₀ value for the drug cocktail was lower than the IC₅₀ value for DOX and PTX independently, indicating the synergistic anticancer drug efficacy. Furthermore, the %cell viability after 72 h of treating 3D tumoroids with 1 μM and 5 μM of the drug cocktail to that of the control suggested substantial drug efficacy differences due to two different culture systems: 2D cultured cells and 3D tumoroids [Fig. 4(E)]. At 1 μM and 5 μM drug cocktail, the %cell viability in 2D cultured cells was 51.70% and 10.61%, whereas in 3D tumoroids the %cell viability remained significantly high (***p < 0.001), i.e., 89.47% and 83.11%, respectively. The drug resistance observed in the MDC aided 3D tumoroids was also compared with that of the 3D spheroid aided by the ultralow attachment (ULA) U bottom plates (most commercially used substrate to aid 3D spheroids for drug screening) [S25, SI]. For the 1 μM treated group, the difference in the drug efficacy on MDC aided 3D tumoroids and ULA aided 3D spheroids was marginally different. Furthermore, in the 5 μM treated group, the MDC aided 3D tumoroids showed somewhat higher drug resistance than the ULA aided 3D spheroids (*p < 0.05). Overall, the MDC aided 3D tumoroids showed higher drug resistance than that observed in the ULA aided 3D spheroid and was significantly high compared to the 2D cultured cells.

With the evaluation of the varied TPFs, to explore the overall impact at molecular levels of the dimensionality in which cells are cultured, whole transcriptomic analysis was performed. The analysis revealed that 5363 genes were downregulated and 5765 were upregulated in 3D versus 2D cultured cells [Fig. 5(A)]. Statistically significant gene expression differences are shown in Fig. 5(B). Genes being differentially expressed with high significance appear towards the top of the plot with red dots representing upregulated and green dots representing downregulated genes, while non-differentially expressed genes are shown as black dots. Key genes related to ECM, EMT, hypoxia, cancer stemness, drug resistance, and other few tumor promoting factors are highlighted through a heatmap [Fig. 5(C)]. Relative gene expression is color-coded green to red indicating low to high expression. Thus, summarizing that the downstream result of culturing cell in 2D and 3D, rendered to differentially expressing gene profile which definitely get translated in the observed structural and functional changes in the 2D and 3D cultured cells. This also drives the resultant TPFs observed in 2D cultured cells and MDC aided 3D tumoroids.

Fig. 5 (A) Bar graph showing the total number of upregulated and downregulated genes in 3D cultured cells in comparison to the 2D cultured cells; (B) volcano plot depicting differentially expressed genes with statistical significance, genes with a low p-value appear at the top of the plot [downregulated (green), upregulated (red) and the genes with no significant differential expression (black)]; (C) heat map showing differential expression of target genes involved in ECM synthesis, ECM binding proteins, EMT, cancer stemness, hypoxia, drug efflux, proliferation and key genes involved in tumor progression signalling in 3D cultured cells in comparison to the 2D cultured cells; relatively high expression values are shown in red; (D) heat map showing differential scores for cancer associated key pathway analysis in 3D cultured cells in comparison to the 2D cultured cells (relatively high expression values are shown in light orange decreasing to sky blue color).

Moreover, to further dive into differential signalling pathway analysis as a consequence of the 2D and 3D culture conditions, gene set enrichment analysis (GSEA) was performed. Using signature gene sets taken as input from the transcriptome analysis data, key molecular signalling pathways known to be involved in cancer progression were investigated. The heatmap [Fig. 5(D)] depicts differential enrichment scores derived after performing GSEA by means of the color code, representing up/downregulation of the gene set representing a particular signalling pathway in 2D and 3D cultured cells. The GSEA for key signalling pathways playing a role in tumor development and cancer progression clearly showed variations between the 2D and 3D cultured cells and is discussed in detail in S15 (SI).

3.5. Mechanotransduction aided by the MDC hydrogel to the 3D cultured cells

The mechanotransduction in 2D and 3D cultured cells was studied by exploring YAP (Yes-associated protein) [a mechanosensitive transcriptional activator] expression and its cellular localization in 2D and 3D cultured cells through immunocytochemistry [Fig. 6]. In the case of 2D cultured cells, YAP1 expression was restricted to the nucleus in the majority of the cells [Fig. 6(A) – 2D YAP1], whereas, in the 3D cultured cells, the YAP1 expression was also observed in the cytoplasm along with its presence in the nucleus [Fig. 6(A) – 3D YAP1]. The deduced scatterplot showed maximum pixels falling into quadrant 3 and the slope of intensity inclined more towards quadrant 1, indicating pixels fluorescing for YAP1 colocalized with DAPI signals in the 2D cultured cells [Fig. 6(B)]. On the other hand, the scatterplot for 3D cultured cells [Fig. 6(C)] showed the fluorescence intensity slope inclining more towards quadrant 2. Thus, it is deduced that the YAP1 intensities were high in the fluorescence image for the 3D cultured cells. Also, the pixel set having an independent signal of YAP1- Alexa Fluor 488 away from the signal of DAPI was clearly observed in quadrant 2 of Fig. 6(C) [encircled]. The following observation was supported by the significant difference (***p < 0.001) in the colocalization coefficient deduced for 2D cultured cells (0.864) and 3D cultured cells (0.488) [Fig. 6(D)]. In addition, Mander's overlap coefficient for 2D cultured cells (0.83) and 3D cultured cells (0.78) [Fig. 6(E)] indicated a nonsignificant difference between the pixels showing YAP1- Alexa Fluor 488 overlapping with DAPI signals in both groups. Hence, in 2D and 3D cultured cells, the YAP1 expression site definitely overlapped with that of the nucleus. Moreover, the ratio of the YAP1 intensity in the nucleus to that in the cytoplasm (Nuc./Cyt.) of the 2D and 3D cultured cells was calculated [Fig. 6(F)]. The Nuc./Cyt. ratio of YAP1 in 2D cultured cells (9.47) was significantly higher (****p < 0.0001) than that in the 3D cultured cells (2.58). Additionally, the CTCF calculation indicated a significantly (****p < 0.0001) escalated expression of YAP1 in the nuclei of the 3D cultured cells (2.2-fold) compared to that in the nuclei of the 2D cultured cells [Fig. 6(G)]. Thus, as observed from the results, YAP1 is primarily located in the nucleus in the 2D cultured cells. In contrast, the 3D cultured cells showed YAP1 in both the nucleus and cytoplasm as indicated by a lower Nuc./Cyt. ratio. The higher YAP1 CTCF in the nuclei of the 3D cultured cells suggested increased YAP1 expression than in the 2D cultured cells’ nuclei. At the mRNA level as well, the YAP1 expression was noted to be 2.4-fold higher (***p < 0.001) in the 3D cultured cells compared to that in the 2D cultured cells [Fig. 6(H)]. Thus, the high YAP1 expression corroborated with the overall elevated expression of YAP1 in 3D cultured cells including YAP1 in the nucleus as well as the cytoplasm of the 3D cultured cells compared to that in the 2D cultured cells.

Fig. 6 (A) Fluorescence images showing YAP1 expression (green) and nucleus staining with DAPI (blue) in 2D and 3D cultured cells captured using a 20× objective lens. The inset in the image shows a magnified cell, focusing on the localization of YAP1 in 2D and 3D cultured cells. Scatter plot representing co-localization of the fluorescent signals of the nucleus (TV2) and YAP1 (TV3) in (B) 2D cultured cells and (C) 3D cultured cells. (D) Co-localization coefficient and (E) Mander's overlap coefficient derived from YAP1 and nucleus fluorescence signals in 2D and 3D cultured cells. (F) YAP1 nucleus to cytoplasm mean fluorescence intensity ratio in 2D and 3D cultured cells. (G) YAP1 expression in the nucleus measured in the form of CTCF from 2D and 3D cultured cell immunofluorescence images and (H) differential YAP1 expression analysis in 2D and 3D cultured cells.

4. Discussion

Developing physiologically relevant 3D in vitro models requires balancing the ideal properties of the substrate with the model's functionality, which remains a challenge. The 3D cell culture technique offers tumor models that mimic the spatiotemporal architecture and physiological features including cellular communication, metabolism, gene expression, and morphology mimicry of the tumor's cancer cells.^32,33 Few studies have assessed the full range of TPFs before validating them as relevant in vitro 3D models. Evaluating and ensuring the presence of TPFs is essential but challenging. Additionally, replicating the secretion, activation, and interaction of TPGFs in 3D in vitro models is crucial. Investigating tumor-promoting factors within 3D in vitro models, influenced by scaffold-driven cell dimensionality and mimicking in vivo conditions, can offer new opportunities to explore cancer biology and enhance the precision of chemotherapeutic studies. Thus, to achieve an in vitro 3D model, a substrate composed of oxidized dextran and thiolated chitosan was chosen in this study.

The macroporosity in the 3D substrate to be used for culturing cells plays a crucial role in oxygen diffusion.³⁴ The average pore size of the MDC hydrogel can be considered favourable for developing in vitro model as supported by a reported finding that pore sizes of 4060 ± 160 μm² and larger than that allows more efficient oxygen diffusion compared to the structure having no pores or pores with smaller size than 4060 ± 160 μm².³⁵ Also, the macroporous hydrogel aids in nutrient transportation,³⁶ flushing off the cellular metabolic waste, and allows faster diffusion.^37,38 The macroporous spatially distributed voids also facilitate cells to aggregate and later grow to form 3D tumoroids.^4,19,39 In addition to macroporosity, which mimics the ECM architecture, the substrate stiffness similar to the target tissue is desirable. The stiffness of the normal breast tissue is ∼2.0–3.0 kPa, whereas the stiffness measured in the form of Young's modulus ranges from ∼10 to 40 kPa, which varies depending on the type of BC tissue.⁴⁰ On the other hand, the stiffness of the TCPS frequently used as a substrate for 2D cell culture is ∼3 GPa, which is significantly stiffer than the cancerous breast tissue and can exhibit unrealistic cell fate and characteristics.^41,42 Hence, the MDC hydrogel having a stiffness of ∼11.3 kPa can be considering as having desirable stiffness to be employed as a breast cancer tissue stiffness mimicking 3D substrate. The stiffness of the hydrogel is crucial for availing biomechanical cues to the cultured 3D cells. This may in turn influence the tumorigenic characteristics of the developed 3D tumoroids. As a result of the surface chemistry, topography, macroporosity, and the stiffness of the MDC hydrogel, the formation of 3D tumoroids from the seeded MDA-MB-231 cells was achieved. Moreover, the area of the tumoroids calculated after the 6th day of cell culture through SEM micrographs indicated that the MDC hydrogel aided a greater number of tumoroids having an area of 2000–10 000 μm². The resultant average size of the MDC aided 3D tumoroids [Fig. 1(F)] corroborated with the studies that reported spheroids with such a size range stood appropriate to establish oxygen,⁴³ and nutrient gradients,³⁶ and variable pH conditions intracellular and extracellular of the tumor spheroids.⁴⁴ Thus, the physicochemical characterization of the MDC hydrogel and aided 3D tumoroids indicated the further need to check the TPFs’ resemblance. This can further ensure the MDC aided 3D tumoroids as a physiologically simulating TNBC 3D in vitro model.

As a key TPF, the ECM deposition is markedly elevated in the tumors physiologically via higher production of ECM components – collagen,^45,46 fibronectin,^47,48 glycosaminoglycans (GAG),^49,50 laminins,^51,52 and elastin⁵³ – which play a key role in diseased tissue remodelling.⁵⁴ Enhanced FN1 leads to an increase in ECM–cell adhesion and thus hinders drug penetration.⁵⁵ The existence of cells in 3D form as part of MDC aided 3D tumoroids may cumulatively contribute to elevated FN1 expression [Fig. 2(A)]. Also. FN1 plays a regulatory role in collagen type 1 deposition,⁵⁶ which was also noted to be co-regulating in the 3D tumoroids reflected by the increased COL1A1 expression [Fig. 2(B)], thus showcasing the greater ECM marker expression in the 3D tumoroids than in its 2D counterpart, which may further play a role in resistance in the drug penetration. Furthermore, a tumor has a heterogenous cell population; the cells in the proliferative zone will only proliferate and the other cells present will become part of the quiescent zone and necrotic core due to the accumulation of cell waste and the lack of oxygen.⁵⁷ Our results [Fig. 2(D–G)] were in line with the other reported studies where the proliferation of the 3D cultured cells was reported to be low in comparison to its 2D counterpart.^12,58 Also, the claudin-low BC subtype analysed for MKI67 expression amongst human origin breast tumor, BC cell line, mouse origin breast tumor microarray data sets and patient clinical data clearly presented low expression of MKI67. Henceforth, the claudin-low subtype is a slow proliferating tumor subtype compared to the other BC subtypes,⁵⁹ thus proving that the MDC aided 3D tumoroids more closely resemble the claudin-low proliferative feature observed under biological conditions compared to the 2D cultured cells. Moreover, with EMT being the key signature of TNBC-type breast cancer, the molecular profiling of metastatic cells has clearly depicted EMT as a responsible factor for tumor aggressiveness and its drug resistance.⁶⁰ This hybrid epithelial/mesenchymal state of cells imparts aggressive traits and is linked with the cancer stemness.^61,62 The phenotypic analysis of 2D cultured cells showed epithelial morphology, whereas many of the 3D cultured cells in this study were enlarged and flattened, showcasing a more mesenchymal phenotype with some having an epithelial like mixed phenotype as well [Fig. 2(H–J)]. This reflects epithelial–mesenchymal plasticity under the effect of cell dimensionality and the TME developed cumulatively in the MDC aided 3D tumoroids corroborating the morphological analysis and morphological description of the cell undergoing epithelial-to-mesenchymal transition reported in various studies.^63–65 Furthermore, the EMT features cadherin expression switching – CDH1 expression is lowered, while CDH2 is enhanced when cells transition from an epithelial to a mesenchymal fate.^66,67 Also, MMP2 plays a crucial role in matrix degradation and has been considered a signature contributor to the success of EMT.⁶⁸ For motility of cells, F-actin and vimentin both undergo structural and quantitative changes. Reduction of either of them does not significantly affect the cell motility, but together it can impair the cell motility and growth of MDA-MB-231.⁶⁹ Thus, the low expression of CDH1 and increased expression of CDH2 and MMP2 [Fig. 2(K–N)] imply that the 3D tumoroids had more EMT phenotypic as well as genotypic traits, thus underscoring the resemblance of EMT – a key TPF in the MDC aided 3D tumoroids, which may further play a key role in modulating other TPFs and drug resistance in the developed 3D tumoroids. Moreover, other factors such as hypoxia and soluble tumor promoting factors can induce EMT in the MDC aided 3D tumoroids.

Among the soluble TPFs, the TPGFs in the tumor milieu such as TNF, FGF2, and BMP2 play crucial roles in cancer progression by inducing EMT.⁷⁰ The TNF has been studied to enhance cancer stemness in BC by up-regulating TAZ through a non-canonical NF-kB pathway.⁷¹ FGF2 enhances the ALDH^highCD44⁺ cancer stemness phenotype via FGF2-DACH1 signalling.⁷² BMP2 can induce cancer stemness via CD44.⁷³ Furthermore, incremental secretion of IL6 and IL10 is considered to be an early prognostic marker of invasive BC.^74,75 Thus, increased IL6, IL10, TNFA, BMP2, and FGF2 secretion in the tumoroid milieu by the MDC aided 3D tumoroids [Fig. 3(A–E)] replicates the scenario of enhanced soluble TPF secretion observed under physiological conditions. Another such soluble TPF, TGFB, a master regulator of abundant cellular functions, regulates fibronectin and type I collagen synthesis through TGFB1⁷⁶ and plays a vital role in EMT induction.^77–79 Most studies aimed to elucidate the effect of TGFB on various aspects related to cancer include short-term (24–48 h) exposure to externally added active TGFB,^80,81 while a study suggests that long-term continuous TGFB exposure stabilizes EMT and also influences cancer stemness.⁸² The increased secretion of active TGFB1 in the 3D tumoroids [Fig. 3(F and G)] can be attributed to the enhanced fibronectin I and ITGB1 levels [Fig. 2(C)] as the integrins play a crucial role in converting latent TGFB1 to an activated form of TGFB1⁸³ directing towards a more realistic mimicry of the secretome by the MDC aided 3D tumoroids to that present under biological conditions.

Another crucial TPF-hypoxia, which has been reported to enhance HIF1A, has been involved in directly boosting cancer stemness^84,85 and establishing drug resistance⁸⁶ by influencing the drug efflux system in the tumor.⁸⁷ Also, hypoxic malignant epithelial cells can induce more IL10 production,⁸⁸ which was also observed in the secretome of the 3D tumoroids. The size of the tumoroids can be said to be a contributing factor in elevating hypoxia [Fig. 3(I–K)] in the MDC aided 3D tumoroids. Hypoxia, in turn exhibits its tumor promoting effect by influencing the cancer stemness, which is an important hallmark of cancer.⁸⁹ Cancer stemness surface markers observed in CSCs such as CD133 are crucial BC clinical biomarkers indicating cancer stem cell prognosis.^90,91 CD133 is induced under hypoxic conditions via HIF1A.⁹² Upregulation of CD133 plays a vital role in tumorigenesis, invasive potential and drug resistance.^93–97 Also, CD24 and CD44, independently or together, have been considered markers to detect CSCs in the BC cell population.⁹⁸ The CD44^high/CD24^high cells can retain pro-invasive factors such as IL6⁹⁹ and TGFB.^100–102 Cells having ALDH^highCD44⁺CD133⁺ characteristics show high metastatic potential and tumorigenicity.¹⁰³ The cancer stemness marker ALDH1A1 elevates collagen expression¹⁰⁴ and imparts drug resistance.¹⁰⁵ Additionally, SOX2 acts as an inducer of the CSC population by regulating the cell cycle pathway involving SOX2.¹⁰⁶ SOX2 in 3D spheroids aids an anchorage-independent growth and participates in spheroid growth via AKT signalling.¹⁰⁷ Moreover, NANOG imparts elevated chemotherapy resistance to CSCs via PI3K/AKT pathway activation.¹⁰⁸ The increased cancer stemness markers in the 3D tumoroids in this study [Fig. 4(A–D)] matched the enhanced cancer stemness observed in the 3D colorectal cancer models,¹⁰⁹ 3D spheroids of ovarian cell lines,¹⁰⁶ 3D spheroids of the HNSCC cell lines,¹¹⁰ and BCC¹¹¹ in comparison to their 2D counterparts. The hypoxic microenvironment developed in the 3D in vitro model is considered to be vital to maintain stemness in the CSCs, where HIF1A induces CD133¹¹² and stemness factors NANOG, OCT4, and SOX2.^113,114 Thus, the hypoxic conditions detected in the MDC aided 3D tumoroid encouraged cancer stemness, which further promoted drug resistance – the most important hallmark of aggressive tumors – hampering chemotherapeutic strategies to act upon to inhibit cancer progression.¹¹⁵ Both the drugs tested in this study have different mechanisms of action: DOX intercalates in DNA aiding DNA damage by generating ROS,¹¹⁶ while PTX, being a mitotic inhibitor, arrests cells in the G2/M phase¹¹⁷ and at lower concentrations induces apoptosis in G0 and G1/S phases.¹¹⁸ Thus, longer-term treatment synergistically exhibits two different mechanisms of action: inhibiting cancer cell proliferation and survival. On the other hand, MDC aided 3D tumoroids showed all the key tumorigenic properties such as the variability in the availability of the drug through the heterogenous cell population,^119,120 compactness,¹²¹ ECM deposition,^122–124 the presence of hypoxia^125,126 and augmented cancer stemness^108,127,128 desirable to mimic solid microtumors, which fostered drug resistance and led to less drug cocktail efficacy in the 3D tumoroids [Fig. 4(E)]. The variation observed in the drug efficacy tested on 2D cultured cells, 3D in vitro models, in vivo models¹²⁹ and the clinical trials is also attributed to the trait called drug efflux exhibited by the ATP-binding cassette (ABC) transporter, which imposes multidrug resistance (MDR),¹³⁰ which is dramatically expressed in the tumor.¹³¹ MDC aided 3D tumoroids evidently exhibited the act of drug efflux being also implemented compared to the 2D cultured cells [Fig. 4(F)] via increased expression of ABCB1 corroborating the drug efflux studies performed using 3D spheroids.^93,132 Hypoxia tends to elevate ABCB1, contributing to efflux, and thus the intracellular drug retention becomes low.⁸⁷ With profound higher expression of ITGB1, FN1, COL1A1 and ABCB1 in the 3D tumoroids, it exhibits cell adhesion mediated drug resistance (CAM-DR) properties.^133,134 Adhesion to fibronectin via ITGB1 tunes p27^kip1 aiding CAM-DR.⁵⁵ Also, ITGB1 binding to COL1 aids chemoresistance via ABC transporters, specifically doxorubicin being effluxed by ABCB1 as observed in MDA-MB-231 cells.¹³⁵ Furthermore, the soluble factor mediated drug resistance (SFM-DR) model can also be said to be implemented in the 3D tumoroids with IL6 and TGFB1 secretion. The MDC aided 3D tumoroid consisting of hypoxic conditions induces HIF1A, in turn levelling up CD133, which influences the expression of drug efflux protein, which is also reported in studies of the hypoxia–cancer stemness–drug resistance axis in cancer cells.^136–138 Also, with elevated IL6 levels drug resistance is imparted^139,140via the IL6–STAT3 axis¹⁴¹ and it increases MDR1 by activating the CCAAT enhancer-binding protein (C/EBP) family transcription factor.¹⁴² Synchronously, as observed in the 3D tumoroids, the TGFB1 overexpression effectively contributes to chemoresistance.^143,144 Through mechanotransduction mediated by the MDC hydrogel to the 3D cultured cells, along with the influence of the tumor-promotive milieu; the whole transcriptome sequencing data revealed the overall variations in the expression of signature genes governing the key TPFs between 2D and 3D cultured cells [Fig. 5(A–C)]. Moreover, the GSEA for key signalling pathways playing a role in tumor development and cancer progression clearly showed variations between the 2D and 3D cultured cells [Fig. 5(D)]. A detailed discussion in the differential signalling pathway analysis is provided in section S15.2 of the SI, highlighting the importance of the dimensionality in which the cells are grown. Along with dimensionality, the 3D cells forming 3D tumoroids with relevant TPFs and TME, particularly TPGFs, cytokines and hypoxia, contributed to the resulting differential signalling pathway trend compared to the 2D cultured cells. Also, this result strongly supports the experimental evidence proving MDC aided 3D tumoroids to be more physiologically relevant 3D in vitro models, mimicking tumorigenic properties as well as structural and underneath molecular function to that of the triple negative type metastatic mammary adenocarcinoma.

Along with the tumor promoting milieu, there is changing tumor niche. The diseased tissue region starts stiffening with enhanced ECM deposition,^145–147 affecting the biological function of the cell through mechanotransduction.^148,149 YAP has been identified as a sensor as well as a mediator of ECM rigidity, cell shape or adhesive area, which directs mechanical signals to the nucleus.¹⁵⁰ The talin and linker of the nucleoskeleton and cytoskeleton (LINC) complex mechanically establishes a connection between ECM–focal adhesion–cytoskeleton and the nucleoskeleton. This way, the mechanical force exerted on the cell reaches nuclei and triggers YAP1 translocation.^151–153 This cascade of transferring the stimuli can happen dependent¹⁵⁴ or independent of the hippo signalling pathway¹⁵⁵ through YAP/TAZ. Due to the flattened morphology of cells on the stiffer substrate, the nucleus is also flattened, which may elevate the nuclear pore permeability due to increased nuclear curvature. Thus, nuclear pore exposure on the side of the cytoplasm increases and hence import to the nucleus increases compared to the export.^155–157 Thus, due to the varied stiffness of the substrate in this study – TCPS (∼3 GPa)⁴¹ and MDC hydrogel (∼11.3 kPa), the shape and adhesion fashion of the 2D and 3D cultured cells, the cells have sensed the mechanical cues [Fig. 6(A–H)]. Our result corroborated the findings of Caliari et al., where the 2D cultured cells showed nuclear localization of YAP1 due to the more spread shape of cells, whereas cells cultured on the hydrogel with ∼20 kPa stiffness showed a lower YAP1 nucleus/cytoplasm ratio.¹⁵⁸ Hence, the cell shape, along with cell dimensionality and the prevailing stiffness of the MDC hydrogel, provided mechanical cues that were reflected as biochemical signals in the form of distinct YAP localization patterns observed in 2D and 3D cultured cells. Mechanotransduction mediated through YAP1 in 3D cultured cells may contribute to the enhanced tumorigenic properties observed in 3D tumoroids, as YAP is widely associated with cancer progression through the inhibition of autophagy and the promotion of EMT in cancer cells.^159–161 The MDC [-ECM mimetic biomaterial] aided 3D tumoroid – proved to be capable of mimicking key tumor promoting factors: tumor promoting secretome, ECM, hypoxia, EMT characteristics, proliferation, cancer stemness, drug efflux, and drug resistance resembling the native TNBC milieu. Thus, the MDC hydrogel aided TNBC 3D in vitro model more meticulously recapitulated major TNBC characteristics.

Furthermore, to demonstrate its ability to be translated to a high throughput screening point-of-care testing platform, the MDC hydrogel was transformed into a new design, leading to the development of a ready-to-use 3D in vitro model aiding platform. Maintaining the MDC chemistry, the new design of the scaffold was aimed to avail physical cue in the form of centrifugal force to trigger cell aggregation as well as surface chemistry based induction to achieve a 3D in vitro tumor model. For that, MDC was translated to a multiwell format via the casting process using a mould, which made the process easy and reproducible [S26, SI]; the scaffold was referred to as the multiwell MDC scaffold (MMS). This scaffold has advantages over the ultra-low attachment plates by availing ECM compositional mimicry as well as ECM surface topographical cues by the MMS. Thus, MDC has a promising translatable potential as a high throughput point-of-care testing scaffold aided in vitro 3D microtumor model.

5. Conclusion

The stiffness, macroporosity, surface chemistry and ECM mimicking polymeric composition of the MDC hydrogel mimicked the BC tissue stiffness, physiological ECM composition and topography. Thus, MDC stands out as a relevant 3D substrate having the potential to aid 3D tumoroids. These 3D tumoroids imitated major TPFs including secretome, ECM deposition, EMT characteristics, hypoxia, cancer stemness, and proliferation potential, leading to relevant drug resistance including drug efflux. The MDC hydrogel facilitated mechanotransduction to the 3D cultured cells reflected through YAP1 expression and localization. Overall, this study demonstrated the MDC hydrogel as a desirable 3D substrate aiding MDA-MB-231 based 3D tumoroids, which efficiently mirrored the TPFs and hence the tumorigenic properties of TNBC [Fig. 7]. Also, the 3D cultured MDA-MB-231 more closely mimicked properties close to characteristic features of the metastatic mammary adenocarcinoma cells as compared to its 2D counterpart. Thus posing desirable tumorigenic traits to attain physiologically closely resembling in vitro model of tripple-negative type metastatic mammary adenocarcinoma as MDA-MB-231 cell line was established from the pleural effusion of a woman diagnosed with tripple-negative metastatic mammary adenocarcinoma. The 3D in vitro model constituted by MDA-MB-231 exhibited the properties of metastatic mammary adenocarcinoma, a receptor based triple negative and molecularly claudin low subtype. The MDC aided 3D tumoroids mimicked key properties such as low proliferation, invasive potential with EMT characteristics, high cancer stemness and elevated drug resistance in comparison to its 2D counterpart. Hence, the MDC aided MDA-MB-231 3D tumoroids through this study were well demonstrated as a promising 3D in vitro TNBC model of metastatic mammary adenocarcinoma of aggressive triple negative type. Furthermore, translating the MDC scaffold in the form of a multiwell MDC scaffold (MMS), as a ready-to-use point-of-care testing platform, illustrates its potential for scaling up to high throughput levels and a physiologically relevant 3D in vitro model for in-depth cancer biology studies and drug screening. Hence, this study opens new avenues for deep cancer biology study and drug screening platforms aiding promising 3D in vitro cancer models.

Fig. 7 Schematic diagram compiling the present study, illustrating pertinent tumor promoting factors in MDC hydrogel aided 3D tumoroids in comparison to the 2D cultured cells, further contributing to enhanced drug resistance, thus establishing a promising in vitro 3D TNBC model.

Author contributions

Unnati Modi: conceptualization, investigation, methodology, validation, formal analysis, data curation, visualization, writing – original draft. Pooja Makwana: methodology, validation. Bindiya Dhimmar: methodology, validation. Soundharya Ramu: software, formal analysis. Mohit Kumar Jolly: validation. Rajesh Vasita: conceptualization, validation, data curation, visualization, supervision, resources, project administration, funding acquisition, writing – review & editing.

Conflicts of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). The SI includes data methods and results pertaining modification of the polymers, fabrication of the MDC hydrogel, physicochemical characterization of the MDC hydrogel, qRT-PCR and whole transcriptome sequencing, localization of YAP1, % cell viability post drug treatment on 2D cultured cells, 3D tumoroids and ULA derived spheroids and other drug treatment related studies. Details of the primers' sequences utilized in the study. Supplementary information is available. See DOI: https://doi.org/10.1039/d5tb00812c.

Acknowledgements

The authors would like to acknowledge the Science and Engineering Research Board (SERB), Department of Science and Technology, Govt. of India [TTR/2021/000120 and SIR/2022/000483] for the financial support. UM would like to acknowledge the UGC-Savitribai Jyotirao Phule Single Girl Child Senior Research fellowship. PM and BD would like to thank the Council of Scientific & Industrial Research, India (CSIR) for providing the CSIR-SRF fellowship. Also, the authors extend their thanks to Gujarat Biotechnology Research Centre, Department of Science and Technology, Govt. of Gujarat for flow cytometry analysis. The authors also thank Eurofins Genomics, India for the whole transcriptome sequencing service and related analysis. MKJ was supported by Param Hansa Philanthropies. SR acknowledges support from the Axis Bank PhD Fellowship, Axis Bank Center for Mathematics and Computing, IISc Bangalore, India.

References

1. M. El-Tanani, S. A. Rabbani, R. Babiker, I. Rangraze, S. Kapre, S. S. Palakurthi, A. M. Alnuqaydan, A. A. Aljabali, M. Rizzo, Y. El-Tanani and M. M. Tambuwala, Cancer Lett., 2024, 591, 216894.
2. Y. Liu, X. Yao, Y. Zhao, D. Fang, L. Shi, L. Yang, G. Song, K. Cai, L. Li, Q. Deng, M. Li and Z. Luo, Cell Rep., 2023, 42, 113213.
3. S. J. Conner, H. B. Borges, J. R. Guarin, T. J. Gerton, A. Yui, K. J. Salhany, Jr., D. N. Mensah, G. A. Hamilton, G. H. Le, K. C. Lew, C. Zhang and M. J. Oudin, Cancer Res., 2024, 84, 2761–2775.
4. U. Modi, D. Kedaria and R. Vasita, Macromol. Biosci., 2022, 22, 2200196.
5. C. Wang, Q. Zhao, X. Zheng, S. Li, J. Chen, H. Zhao, F. Chen, L. Cui and W. Li, Acta Biomater., 2023, 158, 132–150.
6. T. Ferreira, T. Azevedo, J. Silva, A. I. Faustino-Rocha and P. A. Oliveira, Expert Opin. Drug Discovery, 2024, 19, 189–207.
7. M. Zeng, Z. Ruan, J. Tang, M. Liu, C. Hu, P. Fan and X. Dai, Cancer Cell Int., 2023, 23, 120.
8. K. L. Sodek, T. J. Brown and M. J. Ringuette, BMC Cancer, 2008, 8, 223.
9. N. C. Talbot and T. J. Caperna, Cytotechnology, 2015, 67, 873–883.
10. M. T. Kozlowski, C. J. Crook and H. T. Ku, Commun. Biol., 2021, 4, 1387.
11. R. Cruz-Acuña and A. J. García, Matrix Biol., 2017, 57–58, 324–333.
12. Z. Huang, P. Yu and J. Tang, OncoTargets Ther., 2020, 13, 5395–5405.
13. M. A. Badea, M. Balas, A. Hermenean, A. Ciceu, H. Herman, D. Ionita and A. Dinischiotu, SLAS Discovery, 2019, 24, 563–578.
14. S. Flores-Torres, O. Peza-Chavez, H. Kuasne, J. G. Munguia-Lopez, J. Kort-Mascort, L. Ferri, T. Jiang, C. V. Rajadurai, M. Park, V. Sangwan and J. M. Kinsella, Biofabrication, 2021, 13, 025001.
15. N. Peela, F. S. Sam, W. Christenson, D. Truong, A. W. Watson, G. Mouneimne, R. Ros and M. Nikkhah, Biomaterials, 2016, 81, 72–83.
16. D. T. Nguyen, Y. Fan, Y. M. Akay and M. Akay, IEEE Trans. NanoBiosci., 2016, 15, 289–293.
17. A. Herreros-Pomares, X. Zhou, S. Calabuig-Fariñas, S.-J. Lee, S. Torres, T. Esworthy, S. Y. Hann, E. Jantus-Lewintre, C. Camps and L. G. Zhang, Mater. Sci. Eng., C, 2021, 122, 111914.
18. K. Da Silva, P. Kumar, S. F. van Vuuren, V. Pillay and Y. E. Choonara, ACS Omega, 2021, 6, 21368–21383.
19. D. Kedaria and R. Vasita, J. Tissue Eng., 2017, 8, 2041731417718391.
20. Y. Liu and B. P. Lee, J. Mater. Chem. B, 2016, 4, 6534–6540.
21. I. V. Gofman and A. L. Buyanov, J. Mech. Behav. Biomed. Mater., 2017, 71, 238–243.
22. P. Makwana, U. Modi, B. Dhimmar and R. Vasita, Biomater. Adv., 2024, 163, 213952.
23. R. Solanki, P. K. Rajput, B. Jodha, U. C. S. Yadav and S. Patel, Sci. Rep., 2024, 14, 2595.
24. P. K. Rajput, J. F. Varghese, A. K. Srivastava, U. Kumar and U. C. S. Yadav, Cell. Signalling, 2023, 107, 110686.
25. F. Lamela, R. Bologna-Molina, F. Parietti, V. Pereira-Prado, M. Millán, A. Silva, J. Llaguno, J. Alonso, A. Fernández, J. Sotelo-Silveira, M. Domingues, M. Arocena and J. Hochmann, Tissue Cell, 2024, 88, 102408.
26. C. J. Kelly, K. Hussien, E. Fokas, P. Kannan, R. J. Shipley, T. M. Ashton, M. Stratford, N. Pearson and R. J. Muschel, Radiother. Oncol., 2014, 111, 72–80.
27. D. Miao, J. Shi, Q. Lv, D. Tan, C. Zhao, Z. Xiong and X. Zhang, Cancer Commun., 2024, 44, 361–383.
28. B. Dhimmar, U. Modi, S. S. Parihar, P. Makwana, C. L. Boldrini and R. Vasita, Biomater. Adv., 2024, 164, 213991.
29. K. Präbst, H. Engelhardt, S. Ringgeler and H. Hübner, in Cell Viability Assays: Methods and Protocols, ed. D. F. Gilbert and O. Friedrich, Springer, New York, NY, 2017, pp. 1–17 DOI:10.1007/978-1-4939-6960-9_1.
30. L. Zhou, X. Wei, L. Cheng, J. Tian and J. J. Jiang, Laryngoscope, 2007, 117, 455–460.
31. R. Supino, in In Vitro Toxicity Testing Protocols, ed. S. O’Hare and C. K. Atterwill, Humana Press, Totowa, NJ, 1995, pp. 137–149 DOI:10.1385/0-89603-282-5:137.
32. Y. Nanki, T. Chiyoda, A. Hirasawa, A. Ookubo, M. Itoh, M. Ueno, T. Akahane, K. Kameyama, W. Yamagami and F. Kataoka, Sci. Rep., 2020, 10, 1–11.
33. F. Jacob, R. D. Salinas, D. Y. Zhang, P. T. Nguyen, J. G. Schnoll, S. Z. H. Wong, R. Thokala, S. Sheikh, D. Saxena and S. Prokop, Cell, 2020, 180(188–204), e122.
34. N. E. Fedorovich, E. Kuipers, D. Gawlitta, W. J. A. Dhert and J. Alblas, Tissue Eng., Part A, 2011, 17, 2473–2486.
35. H. Li, A. Wijekoon and N. D. Leipzig, PLoS One, 2012, 7, e48824.
36. T. M. Pawlik, W. W. Souba, T. J. Sweeney and B. P. Bode, J. Surg. Res., 2000, 91, 15–25.
37. H. Tokuyama, Y. Nakahata and T. Ban, J. Membr. Sci., 2020, 595, 117533.
38. N. S. Patil, J. S. Dordick and D. G. Rethwisch, Biomaterials, 1996, 17, 2343–2350.
39. X. Yang, B. Yang, Y. Deng, X. Xie, Y. Qi, G. Yan, X. Peng, P. Zhao and L. Bian, Adv. Mater., 2023, 35, 2300636.
40. A. Samani, J. Zubovits and D. Plewes, Phys. Med. Biol., 2007, 52, 1565.
41. H. Wang, M. W. Tibbitt, S. J. Langer, L. A. Leinwand and K. S. Anseth, Proc. Natl. Acad. Sci. U. S. A., 2013, 110, 19336–19341.
42. C. Yang, M. W. Tibbitt, L. Basta and K. S. Anseth, Nat. Mater., 2014, 13, 645–652.
43. L. M. Langan, N. J. F. Dodd, S. F. Owen, W. M. Purcell, S. K. Jackson and A. N. Jha, PLoS One, 2016, 11, e0149492.
44. M. Al-Husari, C. Murdoch and S. D. Webb, J. Math. Biol., 2014, 69, 839–873.
45. J. Thorlacius-Ussing, C. Jensen, N. I. Nissen, T. R. Cox, R. Kalluri, M. Karsdal and N. Willumsen, J. Pathol., 2024, 262, 22–36.
46. H. Sun, Y. Wang, S. Wang, Y. Xie, K. Sun, S. Li, W. Cui and K. Wang, Sci. Rep., 2023, 13, 100.
47. R. C. Hall, A. M. Vaidya, W. P. Schiemann, Q. Pan and Z.-R. Lu, Cells, 2023, 12(5), 685.
48. G. Ye, X. Zhang, M. Li, Z. Lin, Y. Xu, H. Dong, J. Zhou, J. Zhang, S. Wang, Y. Zhu, X. Yu and X. Qian, BMC Cancer, 2023, 23, 412.
49. M. Balbisi, S. Sugár, G. Schlosser, B. Szeitz, J. Fillinger, J. Moldvay, L. Drahos, A. M. Szász, G. Tóth and L. Turiák, Sci. Rep., 2023, 13, 6268.
50. K. Dzobo and C. Dandara, Biomimetics, 2023, 8, 146.
51. G. Giannelli and S. Antonaci, Clin. Exp. Metastasis, 2000, 18, 439–443.
52. V. Berno, D. Porrini, F. Castiglioni, M. Campiglio, P. Casalini, S. M. Pupa, A. Balsari, S. Menard and E. Tagliabue, Endocr. Relat. Cancer, 2005, 12, 393–406.
53. J. Li, X. Xu, Y. Jiang, N. G. Hansbro, P. M. Hansbro, J. Xu and G. Liu, BMC Cancer, 2020, 20, 217.
54. W. Chen, T. Gu, Q. Chen, C. Qu, C. Zhang, Y. Hu, R. Xia, Y. Zhang, M. Wang, X. Huang, J. Li, C. Shi and Z. Tian, Cell Biosci., 2023, 13, 122.
55. L. A. Hazlehurst, J. S. Damiano, I. Buyuksal, W. J. Pledger and W. S. Dalton, Oncogene, 2000, 19, 4319–4327.
56. R. J. Owens and F. E. Baralle, EMBO J., 1986, 5, 2825–2830.
57. S. Dini, B. J. Binder, S. C. Fischer, C. Mattheyer, A. Schmitz, E. H. K. Stelzer, N. G. Bean and J. E. F. Green, J. R.Soc., Interface, 2016, 13, 20160649.
58. N. Menon, H. X. Dang, U. S. Datla, M. Moarefian, C. B. Lawrence, C. A. Maher and C. N. Jones, Biomater. Sci., 2020, 8, 2786–2796.
59. A. Prat, J. S. Parker, O. Karginova, C. Fan, C. Livasy, J. I. Herschkowitz, X. He and C. M. Perou, Breast Cancer Res., 2010, 12, R68.
60. Z. Guo, M. Ashrafizadeh, W. Zhang, R. Zou, G. Sethi and X. Zhang, Cancer Metastasis Rev., 2024, 43, 29–53.
61. A. W. Lambert and R. A. Weinberg, Nat. Rev. Cancer, 2021, 21, 325–338.
62. S. Pasani, S. Sahoo and M. K. Jolly, J. Clin. Med., 2021, 10, 60.
63. J. Yang, P. Antin, G. Berx, C. Blanpain, T. Brabletz, M. Bronner, K. Campbell, A. Cano, J. Casanova, G. Christofori, S. Dedhar, R. Derynck, H. L. Ford, J. Fuxe, A. García de Herreros, G. J. Goodall, A.-K. Hadjantonakis, R. Y. J. Huang, C. Kalcheim, R. Kalluri, Y. Kang, Y. Khew-Goodall, H. Levine, J. Liu, G. D. Longmore, S. A. Mani, J. Massagué, R. Mayor, D. McClay, K. E. Mostov, D. F. Newgreen, M. A. Nieto, A. Puisieux, R. Runyan, P. Savagner, B. Stanger, M. P. Stemmler, Y. Takahashi, M. Takeichi, E. Theveneau, J. P. Thiery, E. W. Thompson, R. A. Weinberg, E. D. Williams, J. Xing, B. P. Zhou, G. Sheng and E. M. T. I. A. On behalf of the, Nat. Rev. Mol. Cell Bio., 2020, 21, 341–352.
64. F. Haasters, W. C. Prall, D. Anz, C. Bourquin, C. Pautke, S. Endres, W. Mutschler, D. Docheva and M. Schieker, J. Anat., 2009, 214, 759–767.
65. S. E. Leggett, J. Y. Sim, J. E. Rubins, Z. J. Neronha, E. K. Williams and I. Y. Wong, Integr. Biol., 2016, 8, 1133–1144.
66. M. J. Wheelock, Y. Shintani, M. Maeda, Y. Fukumoto and K. R. Johnson, J. Cell Sci., 2008, 121, 727–735.
67. R. Cappellesso, G. Marioni, M. Crescenzi, L. Giacomelli, V. Guzzardo, A. Mussato, A. Staffieri, A. Martini, S. Blandamura and A. Fassina, Histopathology, 2015, 67, 491–500.
68. Y. Zhu, L. Yan, W. Zhu, X. Song, G. Yang and S. Wang, J. Cell. Physiol., 2019, 234, 15847–15855.
69. S. Kwon, S. J. Han, K. S. Kim, S. Kwon, S. J. Han, K. S. Kim, S. Kwon, S. J. Han and K. S. Kim, Oncol. Rep., 2023, 50, 200.
70. H. Tian, T. Zhou, H. Chen, C. Li, Z. Jiang, L. Lao, S. A. Kahn, M. E. L. Duarte, J. Zhao, M. D. Daubs, Z. Buser, E. J. Brochmann, J. C. Wang and S. S. Murray, J. Orthop. Res., 2019, 37, 1638–1648.
71. W. Liu, X. Lu, P. Shi, G. Yang, Z. Zhou, W. Li, X. Mao, D. Jiang and C. Chen, Sci. Rep., 2020, 10, 1804.
72. V. Bhat, M. Piaseczny, D. Goodale, U. Patel, A. Sadri and A. L. Allan, Clin. Exp. Metastasis, 2024, 41, 717.
73. P. Huang, A. Chen, W. He, Z. Li, G. Zhang, Z. Liu, G. Liu, X. Liu, S. He, G. Xiao, F. Huang, J. Stenvang, N. Brünner, A. Hong and J. Wang, Cell Death Dis., 2017, 3, 17039.
74. N. Ahmad, A. Ammar, S. J. Storr, A. R. Green, E. Rakha, I. O. Ellis and S. G. Martin, Cancer Immunol. Immunother., 2018, 67, 537–549.
75. Hamidullah, B. Changkija and R. Konwar, Breast Cancer Res. Treat., 2012, 133, 11–21.
76. R. A. Ignotz, T. Endo and J. Massagué, J. Biol. Chem., 1987, 262, 6443–6446.
77. V. Tirino, R. Camerlingo, K. Bifulco, E. Irollo, R. Montella, F. Paino, G. Sessa, M. V. Carriero, N. Normanno, G. Rocco and G. Pirozzi, Cell Death Dis., 2013, 4, e620–e620.
78. N. Ghazi, N. Saghravanian, M. Taghi Shakeri and M. Jamali, J. Dent., 2021, 22, 33–40.
79. Y. Hao, D. Baker and P. ten Dijke, Int. J. Mol. Sci., 2019, 20, 2767.
80. M. Flum, S. Dicks, Y.-H. Teng, M. Schrempp, A. Nyström, M. Boerries and A. Hecht, Oncogene, 2022, 41, 1492–1506.
81. C. Dong, K. Wu, S. Gu, W. Wang, S. Xie and Y. Zhou, Cell Cycle, 2022, 21, 1406–1421.
82. Y. Katsuno, D. S. Meyer, Z. Zhang, K. M. Shokat, R. J. Akhurst, K. Miyazono and R. Derynck, Sci. Signal., 2019, 12, eaau8544.
83. J. S. Munger, J. G. Harpel, F. G. Giancotti and D. B. Rifkin, Mol. Biol. Cell, 1998, 9, 2627–2638.
84. Q. Zhang, Z. Han, Y. Zhu, J. Chen and W. Li, Mol. Med. Rep., 2021, 23, 17.
85. O. McGinn, V. K. Gupta, P. Dauer, N. Arora, N. Sharma, A. Nomura, V. Dudeja, A. Saluja and S. Banerjee, Sci. Rep., 2017, 7, 7872.
86. S. Muñoz-Galván, E. M. Verdugo-Sivianes, J. M. Santos-Pereira, P. Estevez-García and A. Carnero, J. Exp. Clin. Cancer Res., 2024, 43, 57.
87. M. Zihlif, T. Hameduh, N. Bulatova and H. Hammad, Biomed. Rep., 2023, 19, 88.
88. W. Dummer, B. C. Bastian, N. Ernst, C. Schänzle, A. Schwaaf and E.-B. Bröcker, Int. J. Cancer, 1996, 66, 607–610.
89. P. M. Aponte and A. Caicedo, Stem Cells Int., 2017, 2017, 5619472.
90. A. M. Hefni, A. M. Sayed, M. T. Hussien, A. Z. Abdalla and A. G. Gabr, Cancer Biomarkers, 2022, 35, 207–215.
91. F. Brugnoli, S. Grassilli, Y. Al-Qassab, S. Capitani and V. Bertagnolo, J. Oncol., 2019, 2019, 7512632.
92. A. Soeda, M. Park, D. Lee, A. Mintz, A. Androutsellis-Theotokis, R. D. McKay, J. Engh, T. Iwama, T. Kunisada, A. B. Kassam, I. F. Pollack and D. M. Park, Oncogene, 2009, 28, 3949–3959.
93. E. Gheytanchi, M. Naseri, F. Karimi-Busheri, F. Atyabi, E. S. Mirsharif, M. Bozorgmehr, R. Ghods and Z. Madjd, Cancer Cell Int., 2021, 21, 204.
94. S. Peickert, J. Waurig, C. Dittfeld, A. Dietrich, Y. Garbe, L. Kabus, M. Baumann, M. Grade, T. Ried and L. A. Kunz-Schughart, Lab. Invest., 2012, 92, 1607–1622.
95. L. Goričan, B. Gole and U. Potočnik, Cell, 2020, 9, 1707.
96. S. T. Nguyen, L. S. Nguyen, T. H. P. Nguyen, P. H. Vo, N. M. Do, K. D. Truong and P. Van Pham, Prog. Stem Cell, 2020, 7, 279–289.
97. J. Sun, C. Zhang, G. Liu, H. Liu, C. Zhou, Y. Lu, C. Zhou, L. Yuan and X. Li, Clin. Exp. Metastasis, 2012, 29, 185–196.
98. A. Jaggupilli and E. Elkord, Clin. Dev. Immunol., 2012, 2012, 708036.
99. Z. Cai, L. Chen, S. Chen, R. Fang, X. Chen and W. Lei, Cell Commun. Signaling, 2023, 21, 292.
100. J.-C. Néel, L. Humbert and J.-J. Lebrun, ISRN Mol. Biol., 2012, 381428.
101. S. K. Gudey, R. Sundar, C.-H. Heldin, A. Bergh and M. Landström, Oncotarget, 2017, 8(58), 97703.
102. S. Ripka, A. König, M. Buchholz, M. Wagner, B. Sipos, G. Klöppel, J. Downward, T. M. Gress and P. Michl, Carcinogenesis, 2007, 28, 1178–1187.
103. A. K. Croker, D. Goodale, J. Chu, C. Postenka, B. D. Hedley, D. A. Hess and A. L. Allan, J. Cell. Mol. Med., 2009, 13, 2236–2252.
104. K. Sterzyńska, A. Klejewski, K. Wojtowicz, M. Świerczewska, M. Nowacka, D. Kaźmierczak, M. Andrzejewska, D. Rusek, M. Brązert, J. Brązert, M. Nowicki and R. Januchowski, Int. J. Mol. Sci., 2019, 20, 54.
105. M. Świerczewska, K. Sterzyńska, K. Wojtowicz, D. Kaźmierczak, D. Iżycki, M. Nowicki, M. Zabel and R. Januchowski, Int. J. Mol. Sci., 2019, 20, 2053.
106. M. Robinson, S. F. Gilbert, J. A. Waters, O. Lujano-Olazaba, J. Lara, L. J. Alexander, S. E. Green, G. A. Burkeen, O. Patrus, Z. Sarwar, R. Holmberg, C. Wang and C. D. House, Cancers, 2021, 13, 262.
107. C. Choe, H. Kim, S. Min, S. Park, J. Seo and S. Roh, OncoTargets Ther., 2018, 11, 6197–6207.
108. J. Smith, M. Field and K. Sugaya, Genes, 2023, 14, 1276.
109. I. Hassani, B. Anbiah, A. L. Moore, P. T. Abraham, I. A. Odeniyi, N. L. Habbit, M. W. Greene and E. A. Lipke, J. Biomed. Mater. Res., Part A, 2024, 112, 231–249.
110. S. Melissaridou, E. Wiechec, M. Magan, M. V. Jain, M. K. Chung, L. Farnebo and K. Roberg, Cancer Cell Int., 2019, 19, 16.
111. H. Dhandapani, A. Siddiqui, S. Karadkar and P. Tayalia, Adv. Healthcare Mater., 2023, 12, 2300164.
112. E. E. Bar, A. Lin, V. Mahairaki, W. Matsui and C. G. Eberhart, Am. J. Pathol., 2010, 177, 1491–1502.
113. J. M. Heddleston, Z. Li, R. E. McLendon, A. B. Hjelmeland and J. N. Rich, Cell Cycle, 2009, 8, 3274–3284.
114. J. Mathieu, Z. Zhang, W. Zhou, A. J. Wang, J. M. Heddleston, C. M. A. Pinna, A. Hubaud, B. Stadler, M. Choi, M. Bar, M. Tewari, A. Liu, R. Vessella, R. Rostomily, D. Born, M. Horwitz, C. Ware, C. A. Blau, M. A. Cleary, J. N. Rich and H. Ruohola-Baker, Cancer Res., 2011, 71, 4640–4652.
115. P. Matarrese, R. Vona, B. Ascione, C. Cittadini, A. Tocci and A. M. Mileo, Cells, 2023, 12, 1048.
116. C. F. Thorn, C. Oshiro, S. Marsh, T. Hernandez-Boussard, H. McLeod, T. E. Klein and R. B. Altman, Pharmacogenet. Genomics, 2011, 21.
117. J. J. Manfredi and S. B. Horwitz, Pharmacol. Ther., 1984, 25, 83–125.
118. N. C. Kampan, M. T. Madondo, O. M. McNally, M. Quinn and M. Plebanski, BioMed Res. Int., 2015, 2015, 413076.
119. E. H. Maleki, A. R. Bahrami and M. M. Matin, Genes Dis., 2024, 11, 189–204.
120. V. Salemme, G. Centonze, L. Avalle, D. Natalini, A. Piccolantonio, P. Arina, A. Morellato, U. Ala, D. Taverna, E. Turco and P. Defilippi, Front. Oncol., 2023, 13, 1170264.
121. G. Gunay, H. A. Kirit, A. Kamatar, O. Baghdasaryan, S. Hamsici and H. Acar, Gynecol. Oncol., 2020, 159, 563–572.
122. T. Li, Y. Li, H. Wu, C. Peng, J. Wang, S. Chen, T. Zhao, S. Li, X. Qin and Y. Liu, Mechanobiol. Med., 2023, 1, 100023.
123. C. Wu, T. Rakhshandehroo, H. I. Wettersten, A. Campos, T. von Schalscha, S. Jain, Z. Yu, J. Tan, E. Mose, B. G. Childers, A. M. Lowy, S. M. Weis and D. A. Cheresh, Nat. Cell Biol., 2023, 25, 309–322.
124. M. Cortini, F. Macchi, F. Reggiani, E. Vitale, M. V. Lipreri, F. Perut, A. Ciarrocchi, N. Baldini and S. Avnet, Cancers, 2023, 15, 1221.
125. H. A. Bakshi, M. Mkhael, H. L. Faruck, A. U. Khan, A. A. A. Aljabali, V. Mishra, M. El-Tanani, N. B. Charbe and M. M. Tambuwala, Cell. Signalling, 2024, 113, 110911.
126. N. Srivastava, S. S. Usmani, R. Subbarayan, R. Saini and P. K. Pandey, Front. Oncol., 2023, 13, 1199105.
127. L. López, L. Fernández-Vañes, V. N. Cabal, R. García-Marín, L. Suárez-Fernández, H. Codina-Martínez, S. L. Lorenzo-Guerra, B. Vivanco, V. Blanco-Lorenzo, J. L. Llorente, F. López and M. A. Hermsen, J. Pers. Med., 2023, 13, 1504.
128. N. Ebrahimi, M. Afshinpour, S. S. Fakhr, P. G. Kalkhoran, V. Shadman-Manesh, S. Adelian, S. Beiranvand, F. Rezaei-Tazangi, R. Khorram, M. R. Hamblin and A. R. Aref, Crit. Rev. Oncol./Hematol., 2023, 182, 103920.
129. M. N. Brodeur, K. Simeone, K. Leclerc-Deslauniers, H. Fleury, E. Carmona, D. M. Provencher and A.-M. Mes-Masson, Sci. Rep., 2021, 11, 18183.
130. D. C. Belisario, M. Akman, M. Godel, V. Campani, M. P. Patrizio, L. Scotti, C. M. Hattinger, G. De Rosa, M. Donadelli, M. Serra, J. Kopecka and C. Riganti, Cells, 2025, 14, 622.
131. M. Spitzwieser, C. Pirker, B. Koblmüller, G. Pfeiler, S. Hacker, W. Berger, P. Heffeter and M. Cichna-Markl, Oncotarget, 2016, 7, 73347.
132. A. Filipiak-Duliban, K. Brodaczewska, A. Kajdasz and C. Kieda, Int. J. Mol. Sci., 2022, 23, 1166.
133. M. B. Meads, R. A. Gatenby and W. S. Dalton, Nat. Rev. Cancer, 2009, 9, 665–674.
134. L. A. Hazlehurst, S. A. Enkemann, C. A. Beam, R. F. Argilagos, J. Painter, K. H. Shain, S. Saporta, D. Boulware, L. Moscinski, M. Alsina and W. S. Dalton, Cancer Res., 2003, 63, 7900–7906.
135. F. Baltes, V. Pfeifer, K. Silbermann, J. Caspers, K. Wantoch von Rekowski, M. Schlesinger and G. Bendas, Biochim. Biophys. Acta, Mol. Cell Res., 2020, 1867, 118663.
136. P. Krishnamurthy, D. D. Ross, T. Nakanishi, K. Bailey-Dell, S. Zhou, K. E. Mercer, B. Sarkadi, B. P. Sorrentino and J. D. Schuetz, J. Biol. Chem., 2004, 279, 24218–24225.
137. M. He, H. Wu, Q. Jiang, Y. Liu, L. Han, Y. Yan, B. Wei, F. Liu, X. Deng, H. Chen, L. Zhao, M. Wang, X. Wu, W. Yao, H. Zhao, J. Chen and M. Wei, Mol. Oncol., 2019, 13, 403–421.
138. W.-J. Wang, H. Sui, C. Qi, Q. Li, J. Zhang, S.-F. Wu, M.-Z. Mei, Y.-Y. Lu, Y.-T. Wan, H. Chang and P.-T. Guo, Oncol. Rep., 2016, 36, 428–440.
139. V. S. Jones, R.-Y. Huang, L.-P. Chen, Z.-S. Chen, L. Fu and R.-P. Huang, Biochim. Biophys. Acta, Rev. Cancer, 2016, 1865, 255–265.
140. J. Han, S. Oh, H.-H. Hoang, D. T. T. Nguyen, W. Lim, T. H. Shin, G. Lee and S. Park, J. Ind. Eng. Chem., 2018, 62, 352–361.
141. C. Liu, Y. Zhu, W. Lou, Y. Cui, C. P. Evans and A. C. Gao, Prostate, 2014, 74, 201–209.
142. D. Conze, L. Weiss, P. S. Regen, A. Bhushan, D. Weaver, P. Johnson and M. Rincon, Cancer Res., 2001, 61, 8851–8858.
143. J. Wang, Y. Chen, F. Xiang, M. Li, H. Li, J. Chi and K. Ren, Artif. Cells, Nanomed., Biotechnol., 2018, 46, 1505–1512.
144. N. Oshimori, D. Oristian and E. Fuchs, Cell, 2015, 160, 963–976.
145. N. F. Boyd, L. J. Martin, J. Stone, C. Greenberg, S. Minkin and M. J. Yaffe, Curr. Oncol. Rep., 2001, 3, 314–321.
146. Y.-P. Guo, L. J. Martin, W. Hanna, D. Banerjee, N. Miller, E. Fishell, R. Khokha and N. F. Boyd, Cancer Epidemiol., Biomarkers Prev., 2001, 10, 243–248.
147. P. P. Provenzano, K. W. Eliceiri, J. M. Campbell, D. R. Inman, J. G. White and P. J. Keely, BMC Med., 2006, 4, 38.
148. D. E. Discher, P. Janmey and Y.-L. Wang, Science, 2005, 310, 1139–1143.
149. L. Trichet, J. Le Digabel, R. J. Hawkins, S. R. K. Vedula, M. Gupta, C. Ribrault, P. Hersen, R. Voituriez and B. Ladoux, Proc. Natl. Acad. Sci. U. S. A., 2012, 109, 6933–6938.
150. S. Dupont, L. Morsut, M. Aragona, E. Enzo, S. Giulitti, M. Cordenonsi, F. Zanconato, J. Le Digabel, M. Forcato, S. Bicciato, N. Elvassore and S. Piccolo, Nature, 2011, 474, 179–183.
151. A. Elosegui-Artola, R. Oria, Y. Chen, A. Kosmalska, C. Pérez-González, N. Castro, C. Zhu, X. Trepat and P. Roca-Cusachs, Nat. Cell Biol., 2016, 18, 540–548.
152. X. Zhang, G. Jiang, Y. Cai, S. J. Monkley, D. R. Critchley and M. P. Sheetz, Nat. Cell Biol., 2008, 10, 1062–1068.
153. M. L. Lombardi, D. E. Jaalouk, C. M. Shanahan, B. Burke, K. J. Roux and J. Lammerding, J. Biol. Chem., 2011, 286, 26743–26753.
154. H. Kwon, J. Kim and E.-H. Jho, FEBS J., 2022, 289, 5798–5818.
155. A. Elosegui-Artola, I. Andreu, A. E. M. Beedle, A. Lezamiz, M. Uroz, A. J. Kosmalska, R. Oria, J. Z. Kechagia, P. Rico-Lastres, A.-L. Le Roux, C. M. Shanahan, X. Trepat, D. Navajas, S. Garcia-Manyes and P. Roca-Cusachs, Cell, 2017, 171(1397–1410), e1314.
156. S. S. Patel, B. J. Belmont, J. M. Sante and M. F. Rexach, Cell, 2007, 129, 83–96.
157. Y. Li, Z. Zhong, C. Xu, X. Wu, J. Li, W. Tao, J. Wang, Y. Du and S. Zhang, Nat. Sci. Rev., 2023, 10, nwad165.
158. S. R. Caliari, S. L. Vega, M. Kwon, E. M. Soulas and J. A. Burdick, Biomaterials, 2016, 103, 314–323.
159. L. Jin, Y. Chen, D. Cheng, Z. He, X. Shi, B. Du, X. Xi, Y. Gao and Y. Guo, Cell Death Dis., 2021, 12, 457.
160. W. Choi, J. Kim, J. Park, D.-H. Lee, D. Hwang, J.-H. Kim, H. Ashktorab, D. Smoot, S.-Y. Kim, C. Choi, G. Y. Koh and D.-S. Lim, Cancer Res., 2018, 78, 3306–3320.
161. Y. Xia, T. Chang, Y. Wang, Y. Liu, W. Li, M. Li and H.-Y. Fan, PLoS One, 2014, 9, e91770.

---