3D modeling of normal skin and cutaneous squamous cell carcinoma. A comparative study in 2D cultures, spheroids, and 3D bioprinted systems

The current cancer research and drug testing are primarily based on 2D cell cultures and animal models. However, these methods have limitations and yield distinct drug response patterns. This study addressed this gap by developing an innovative in vitro human three-dimensional (3D) normal skin model and a multicellular model of human cutaneous squamous cell carcinoma (cSCC) using 3D bioprinting technology. Comparative analyzes were performed between bioprinted 3D-cSCC model, consisting of HaCaT keratinocytes, primary normal human dermal fibroblasts and A431 cancer cells (tricellular), bioprinted 3D-A431 model composed of A431 cancer cells only (monocellular), A431 cancer cell spheroids, and conventional 2D models. The models were structurally characterized by light microscopy, immunofluorescence (LIVE/DEAD assay, confocal microscopy) and immunohistochemistry (hematoxylin/eosin, p63, vimentin, Ki67, epidermal growth factor receptor stainings). The spatial arrangement of the 3D models was analyzed using the ARIVIS scientific image analysis platform. All models were also functionally assessed by cetuximab (CTX) response testing with the MTS assay. 3D-cSCC models were maintained for up to 16 weeks. Morphological and histological examinations confirmed the presence of skin-like layers in the bioprinted 3D models of normal skin, and the intricate and diverse features of the bioprinted skin cancer model, replicating the critical in vivo characteristics. In both mono- and tricellular bioprinted tumor constructs, there was a gradual formation and continuous growth of spheroid-like clusters of cancer cells, significantly influencing the morphology of the models. Cancer cells in the 3D bioprinted constructs showed reduced sensitivity to CTX compared to spheroids and 2D cultures. This study underscores the potential of 3D multicellular models in elucidating drug responses and gaining a better understanding the intricate interplay of cellular components within the tumor microenvironment. Developing the multicellular 3D tumor model paves the way for new research critical to advancing fundamental cancer research and future clinical applications, particularly drug response testing.


Introduction
Accumulating evidence underscores the pivotal role of the tumor microenvironment in driving cancer progression and recurrence.The tumor microenvironment not only fosters tumor growth and invasion but also hampers immune responses and influences drug efficacy [1-3].Unfortunately, current approaches to cancer research and preclinical drug testing predominantly rely on monolayer cell cultures (2D) and animal models.Preclinical testing presents a significant challenge, as approximately 90% of potential drug candidates entering clinical trials do not demonstrate the expected therapeutic activity [4].A primary reason for this failure is the inability to accurately replicate the complexities of the human tumor microenvironment in preclinical studies.
Although tumor cell spheroids have served as valuable three-dimensional (3D) models, they still fall short of fully recapitulating the intricate tumor microenvironment.To address this limitation and study the complex interplay between cells and interactions with the extracellular matrix (ECM), 3D models comprising different cell types organized within a supportive matrix are required.Innovative advances in 3D bioprinting technology now offer the possibility of modeling the complex composition and architecture of the tumor microenvironment in vitro [5][6][7].
Cutaneous squamous cell carcinoma (cSCC) is commonly treated with surgical interventions.However, this approach may not prevent the occurrence of frequent and sometimes inoperable local recurrences.In cases where surgical removal is not feasible, alternative treatments such as radiotherapy, laser therapy, and topical 5-Fluorouracil (5-FU) are administered.Systemic chemotherapy, including cisplatin, carboplatin, 5-FU, and taxanes, is reserved primarily for metastatic patients [7,8].Furthermore, cetuximab (CTX, Erbitux ® , Merck), an anti-epidermal growth factor receptor (EGFR) antibody, has emerged as a systemic treatment option for cSCC and head and neck squamous cell carcinoma (HNSCC).However, as shown in colorectal cancer, the response to CTX varies considerably between patients, with only 10%-20% experiencing a favorable response [8,9].CTX has also been approved for concurrent administration with radiotherapy to treat locally advanced HNSCC, offering an alternative treatment for patients who cannot tolerate the toxicities associated with non-CTX therapies [10].
Furthermore, in the phase II clinical trial (www.clinicaltrials.gov/study/NCT00240682),CTX proved to be a promising drug in advanced non-resectable cSCC.Retrospective evaluation of clinical outcomes after six weeks of CTX as a single-agent treatment demonstrated a disease response rate of 69% [11].The overexpression of EGFR in cSCC has been correlated with patient survival, and systemically administered CTX has shown potential for tumor shrinkage, enabling subsequent surgical intervention [12].
This study aimed to harness the capabilities of 3D bioprinting technology to create an in vitro model of cSCC that might more accurately reproduce the complex microenvironment of human skin cancer compared to conventional cell culture models.As a foundation, a human 3D normal skin model (3Dskin), consisting of normal human dermal fibroblasts (NHDF) and HaCaT cells, was bioprinted.The 3D bioprinted skin cancer models included a monocellular model consisting of A431 cells (referred to as 3D-A431), and a tricellular model constructed with NHDF, HaCaT, and A431 cells (referred to as 3D-cSCC), thus mimicking the more complex microenvironment of cSCC.The models were characterized for their structural and functional properties and compared with the spheroids and 2D models.To test the hypothesis that drug responses differ significantly depending on tumor microenvironment, the effects of CTX were evaluated in mono-and tricellular 3D models, as well as in spheroids and 2D models.
This study addresses a significant gap in current research and holds the potential to open new frontiers in preclinical cancer research and individualized treatments.
Printed 3D constructs, spheroids and 2D cultures were incubated in the regular medium under standard conditions, at 37 • C in a humidified atmosphere containing 5% CO 2 .

2D models
NHDF as well as HaCaT, and A431 cells at passage 7 were seeded into 96 well plates (Nunclon Delta-Treated, Flat-Bottom Microplate, Nunc™), at 5 × 10 3 density in the regular medium under standard conditions until they reached 70% confluence.The prepared cell cultures were evaluated both immunohistochemically (sections 2.9 and 4.2) and functionally, by assessing drug response (sections 2.7 and 4.5).Each experiment was conducted four times.

Biofabricated 3D models
The architecture of the bioprinted models, 3Dskin (dicellular normal skin), 3D-A431 (monocellular skin cancer), and 3D-cSCC (tricellular skin cancer), were designed using SLICER 4.0 3D Software.Each of the designed 3D models had dimensions of 5 × 5 × 1 mm.The 3D-skin model consisted of two distinct layers: primary NHDF and cells of a keratinocyte HaCaT cell line.The 3D-A431 model construct consisted of three distinct layers of A431 skin cancer cell line.The 3D-cSCC consisted of three distinct cell layers of primary NHDF, HaCaT keratinocytes, and A431 cells.The models were printed with a BIOX bioprinter (Sweden, Cellink) and cultured in vitro in the regular medium under standard conditions (section 2.1).Bioprinted constructs were cultured for 1-16 weeks (7-112 d).All models were evaluated morphologically, immunohistochemically, and functionally, as described below.All analyzes were independently performed five times and the mean values were used for further calculations.Bioprinting workflow is shown in supplementary S1.

3D-skin model consisted of NHDF and HaCaT cells
A three-layer rectilinear pattern was printed in 24well plates (Thermofisher, Costar), at a temperature of 22 • C-25 • C. The printing process utilized a 22 G nozzle and a pressure ranging from 14-25 kPA.The dermal and epidermal layers were extruded from the 22 G nozzle at a thickness of 0.4 mm with 60% rectilinear infill.For the NHDF suspension, hydrogels Cellink SKIN and Cellink Bioink were loaded in a 2:1 ratio.The HaCaT cell suspension was mixed with the SKIN hydrogel (Cellink) in a 1:1 ratio.The dermal and epidermal layers were extruded using cell suspensions in a 1:2 ratio, respectively, with their corresponding hydrogel carriers Cellink SKIN and Cellink Bioink.The fabricated constructs were crosslinked with a thrombin at a dilution of 10 U ml −1 , which was added to the medium and incubated overnight under standard culture conditions.On the following day, the medium with thrombin was replaced by the regular medium, and the constructs were incubated under standard culture conditions.

3D-A431 model consisted of A431 cells
A three-layer high rectilinear pattern was printed at a nozzle of 22 G, and a pressure of 12-28 kPa.The layers were extruded using cell suspensions of A431 in the BIOINK hydrogel (Cellink) in a ratio of 1:1, respectively.To stabilize the constructs, they were crosslinked using a crosslinking agent (Cellink) containing a 50 mM CaCl 2 solution for 5-10 min.Following this step, the regular medium was added, and the constructs were incubated under standard culture conditions.The next day, the medium was replaced with fresh regular medium, and the constructs were continued to be incubated under standard culture conditions.

3D-cSCC model consisted of NHDF and HaCaT and A431 cells
A three-layer rectilinear pattern was printed with a 22 G nozzle at a pressure of 12-28 kPa.The layers were extruded using cell suspensions of A431:HaCaT:NHDF in a 3:2:1 ratio, respectively.The hydrogel carriers used for the skin layers were Cellink SKIN (ref.IKC207000001, Cellink, Sweden), and BIOINK hydrogel (Cellink) for A431 cancer cells, applied as described in sections 2.4.1 and 2.4.2, respectively.
To stabilize the constructs, they were crosslinked using a crosslinking agent (Cellink) containing a 50 mM CaCl 2 solution for 5-10 min.Following this step, the regular medium was added, and the constructs were incubated under standard culture conditions.The next day, the regular medium was replaced with fresh regular medium, and the constructs continued to be incubated under standard culture conditions.

Monitoring the models
During culture, all models were periodically visualized in a bright field phase-contrast inverted microscope (Olympus CKX53, Optical Co. Ltd, Tokyo), and a light microscope with a camera (Olympus BX53M, Optical Co. Ltd, Tokyo), to assess the morphological features of the cell models, such as confluency, cell morphology, and, in 3D models, also the integrity of cell layers.Additionally, the live cell imager CELLCYTE X™ (Cytena) system was used to visualize the spheroids.Microscopic observations were conducted for up to 6-8 d for the 2D models, 7 d for the spheroids, and over 1-8 weeks for 3D-A431, 3D-skin, and 3D-cSCC models.
2.6.Live/dead staining and confocal imaging of 3D models 3D bioprinting constructs were washed with 1x Hank's balanced salt solution (HBSS, Gibco), and subsequently stained with 1x Live/Dead working solution (cat.R37601, Invitrogen) for 1-1.5 h at 37 • C.After staining, the constructs were washed twice with HBSS and observed using the Zeiss LSM 800 confocal microscope (green FITX/488 and red TexasRed/570 channels) with the pinhole opened to the maximum extent.These observations were conducted over a period of 1-8 weeks for 3D-skin, 3D-A431, and 3D-cSCC models.The spatial arrangement of the 3D models was analyzed using the ARIVIS scientific image analysis platform (Zeiss Zen 3.8).

Cetuximab treatment
CTX is a chimeric IgG1 monoclonal antibody produced from Sp2/0 cells, with an elimination half-life ranging from approximately 70-100 h at the target dose.For this study, CTX was dissolved in the regular medium, as outlined in section 2.1.
In preliminary studies conducted in the 2D cell cultures of NHDF, HaCaT, and A431 cells, as described in section 2.2, metabolic activity was evaluated through the MTS assay at a range of CTX concentrations, from 0.003 mg ml −1 to 4 mg ml −1 .The half-maximal inhibitory concentration (IC50) was determined to be 2 mg ml −1 .Consequently, for subsequent experiments and analyzes involving all cell models, CTX concentrations of 1 mg ml −1 , 2 mg ml −1 , and 3 mg ml −1 were selected.These CTX concentrations were further confirmed through clonogenicity testing (see supplementary, S2).
CTX was added to the 2D cultures of NHDF, HaCaT, and A431 cells, cultured as described in section 2.2, to four-day-old spheroids of NHDF, HaCaT, and A431 cells obtained as detailed in section 2.1, and to the bioprinted 3D-A431 and 3D-cSCC constructs, obtained as outlined in sections 2.4.2 and 2.4.3.In the case of 3D-A431 models, CTX treatment was administered for a period ranging from 1 to 6 weeks.For 3D-cSCC constructs, CTX treatment was applied on the 3rd day, 1 week, 2 weeks, 4 weeks, 5 weeks, 12 weeks, and 16 weeks.
The drug response was assessed after 72 h of exposure, by a colorimetric MTS assay (Promega, USA) that measures metabolic activity, as described below.

Viability and proliferation measurement (MTS assay)
The viability and proliferation of NHDF, and HaCaT and A431 cells were determined by measuring their metabolic activity using the CellTiter 96 ® AQueous one solution cell proliferation assay (Promega, USA), following the previously established protocol [13].Briefly, both treated and untreated cells grown in 2D, spheroids and 3D bioprinted constructs were transferred to a 96 well plate (Nunclon Delta-Treated, Flat-Bottom Microplate, Nunc™) with 20 µl of the MTS substrate and 100 µl of the regular medium.The cells were then incubated for 1-2 h under standard culture conditions.The absorbance of each well was measured at λ = 490 nm using Wallac 1420 Multilabel Counter (Perkin Elmer).
The growth kinetics of 3D bioprinted skin constructs was evaluated using the MTS assay over a period of 1-16 weeks.All analyzes were independently performed six times and the mean values were used for further calculations.

Histological and immunohistochemical analyzes
In the 2D cell culture system, cells were washed twice with phosphate-buffered saline (PBS 10x, D1408, Sigma Aldrich) and fixed with 70% ethanol for 2 h.Subsequently, the cell preparations were stained with hematoxylin (DAKO CS70030-2) and eosin (DAKO CS70130-2).Prior to immunohistochemical staining with specific primary antibodies against Ki67 (confirmation of proliferation activity), vimentin (confirmation of mesenchymal differentiation), and p63, cytokeratin and EGFR (confirmation of squamous cell/epithelial differentiation) (table 1), cell preparations were heated at 96 • C-97 • C in Target Retrieval Solution with Citrate pH 6.0 (DAKO S1699) for 20 min, to increase antigen exposure.
The 3D-A431 and 3D-cSCC constructs were histologically evaluated after 4 weeks in culture.The 3D constructs were washed in 1x Hank's balanced salt solution (HBSS, Gibco) by incubation for 15 min under standard conditions.Next, the 3D constructs were fixed in 4% formaldehyde supplemented with 50 mM CaCl 2 solution (Cellink) for 1 h at room temperature and then dehydrated overnight.The dehydrated samples were embedded in paraffin, cut into 4 µm slices (Microtome Leica RM 2245) and stained with hematoxylin and eosin (HE).Immunohistochemical stainings with antibodies against Ki67, Vimentin, p63, and EGFR were also performed.

Clonogenic assay
The clonogenicity of A431 cells was measured to assess the cytoreductive effect of CTX.Tumor cells were suspended in the regular culture medium and seeded onto 35 mm plastic Petri dishes at a density of 5 × 10 2 cells/dish.Four hours after cell plating, CTX was added to the medium at concentrations of 0.25 mg ml −1 , 0.5 mg ml −1 , 1 mg ml −1 , 2 mg ml −1 , and 3 mg ml −1 and cells were cultured for 10 d.The number of colonies was counted after fixing in methanol and staining with crystal violet (Chempur).Colonies comprising at least 20 cells were counted under the microscope at 40× magnification.The results represent the mean of at least three independent experiments.

Morphological characterization of the 3D bioprinted normal skin model
The Live/Dead assay indicated high viability of the bioprinted fibroblasts and keratinocytes when cultured individually, as assessed after 3 d and 5 weeks (figure 1(A)).However, when these two cell types were combined in the bioprinted skin model, there was a decrease in cell viability after 3 d of culture.During the following 2-5 weeks of culture, we observed proliferating cell layers, long densely packed fibroblast fibers, and noted high cell viability (figure 1(B), see supplementary video 1).The 3D bioprinted skin model displayed characteristics resembling human skin, including stratified epithelial differentiation (figure 1

(C)).
Growth kinetics, assessed by the MTS assay, showed a slight decrease in cell count after 1 week of culture in the bioprinted model that mimicked skin tissue, consisting of fibroblasts and keratinocytes (figure 1(D)).However, this initial decrease was followed by an increase in cell proliferation after 5 weeks of culture, and the increased proliferation continued for up to 16 weeks.
Light microscope analysis revealed that the bioprinted constructs, measuring 5 × 5 × 1 mm, resembled human skin with three layers: the stratum corneum, spinosum, and dermis.Within the spinosum layer, a honeycomb-like pattern of keratinocytes with dark nuclei was observed, transitioning into larger, flatter, and brighter cells in the stratum corneum, indicating a layered differentiation of the epithelium, resembling that of human skin (figure 1(C)).The spatial arrangement of the 3D-skin model revealed that the cells were evenly distributed across the printed layers.After 2 weeks of culture, a typical spindle-shaped morphology of fibroblasts was observed, indicating excellent cell adaptation in the hydrogel environment (see supplementary, S3).

Histological characterization of 2D and 3D bioprinted skin models
In the 2D model, NHDFs demonstrated vimentin expression, confirming their mesenchymal characteristics.This was further supported by the presence of Ki-67, indicating cell proliferation.There was no expression of p63 and EGFR in NHDF cells (figure 2(A)).In the HaCaT cell culture, the staining revealed the expression of p63 and EGFR, suggesting their role in epithelial differentiation and proliferation.Ki-67 staining also confirmed active cell proliferation.Remarkably, vimentin expression was absent in HaCaT cells, indicating their epithelial nature (figure 2(B)).Hematoxylin and eosin (HE) staining conducted on NHDF and HaCaT cells, confirmed the skin-like structure (figures 2(A) and (B)).
Histological examinations confirmed the presence of skin-like layers in the bioprinted 3D normal skin model (figure 2(C)).In particular, epidermis maturation within these constructs was observed, marked by significant keratinization and the presence of well-defined layers, including the stratum corneum, stratum lucidum, stratum granulosum, stratum spinosum, and stratum basale (figures 2(C) and (D)).The expression of vimentin, a protein responsible for orchestrating fibroblast proliferation and keratinocyte differentiation, was also observed.Moreover, the expression of EGFR, critical for tissue homeostasis, involved in regulating cell survival, proliferation, migration, and differentiation, known to be expressed in normal keratinocytes, was confirmed.The presence of Ki67, a marker reflecting the extent   of cellular proliferative activity, and the expression of p63 in the basal cells of the epidermis were notable.All these markers validated the skin-like layers in the 3D-bioprinted skin model.

Morphological characterization of the 3D bioprinted monocellular vs. tricellular skin cancer models (3D-A431 vs. 3D-cSCC)
Phase-contrast microscopy observations showed that the bioprinted 3D-A431 model contained proliferating cell aggregates that increased in size over time.The Live/Dead assay indicated high cell viability throughout the culture period from 2 to 8 weeks.In particular, larger structures that resembled fused spheroids.
Additionally, there were cell aggregates surrounding the central core of dead cells within these 'tumorlike' structures (figure 3(A); see supplementary videos 2-4).
The 3D-cSCC model revealed complex morphological structures (figure 3(B)).After 2 weeks, noticeable irregular cell shapes and structural alterations of the model were observed, and small 'microspheres' appeared.After 6 weeks, numerous protrusions connected with cells were formed, creating a spatial network.Additionally, stratified epithelial differentiation was observed, mirroring characteristics found in the human skin, together with the presence of cancer cell aggregates resembling 'mini-tumors' with heterogeneous shapes distributed throughout the bioprinted construct.These tumor spheroids not only proliferated within the normal epidermis, but also disrupted it, causing craters in the epidermis and extending into the dermis.After eight weeks, micronodules (acini) formed, reminiscent of the in vivo characteristics of skin cancer infiltrating the surrounding tissues.The Live/Dead assay confirmed the complex reorganization of tumor cell structures, highlighting the irregular shapes that developed during long-term culture (see supplementary videos 5-7).The spatial arrangement analysis of the 3D-A431 and 3D-cSCC models confirmed cell aggregations and the complexity of the resulting structures, contributing to the formation of a unique architecture in the bioprinted tumor construct (figures 3(B) and (E)).None of these features were observed in the 2D and spheroid models.

Immunohistochemical evaluation of the 2D-A431, 3D-A431 and 3D-cSCC cancer models
In the 2D-A431 model (figure 4(A)), A431 cells exhibited a flat structure.The expression of EGFR, Ki67, and p63 was confirmed, while vimentin expression was not observed.Within the 3D-A431 model (figure 4(B)), consisting of printed A431 cancer cells alone, the development of cell aggregates closely resembling spheroids was observed, and stronger expression of p63 and EGFR than in 2D cultures was noted.Additionally, the expression of the Ki67 marker of cell proliferation was predominantly observed in the

Cetuximab response of skin cancer cells in 2D cultures, spheroids, and 3D-A431 and 3D-cSCC models
CTX exhibited significant inhibitory effects on the proliferation of A431 cells and NHDF in 2D cell cultures (figure 5), as well as in spheroids (figure 6) at concentrations of 2 mg ml −1 and 3 mg m −1 .HaCaT cells, on the other hand, showed a significant response to CTX at a concentration of 3 mg ml −1 but not at 1 and 2 mg ml −1 (figure 5).
A431 spheroids exhibited slight variability in all morphometric parameters and formed loosely organized aggregates during the 72 h growth period (figure 6).There were differences in shapes between spheroids of normal cells, NHDF, and HaCaT cells.These spheroids maintained a consistent size and shape compared to A431 spheroids.
Based on the results from the 2D and spheroid models (figures 5 and 6), only the 3 mg ml −1 concentration of CTX was employed in the 3D-A431 model, consisting exclusively of cancer cells (figure 7(A)).In all three models, 2D, spheroids, and 3D-A431, a noticeable response to CTX treatment at 3 mg ml −1 was observed after 7 d of culture.This response was characterized by a substantial decrease in metabolic activity, suggestive of cell apoptosis.Furthermore, in the 3D-A431 model cultured for over a longer time, a statistically significant CTX response was observed also after 14 d of culture.However, after 21 d of culture a lack of response to CTX was observed, consistent with an increase in cellular metabolic activity.Examining of the metabolic activity in both the 3D-A431 and 3D-cSCC models (figure 7(A)) revealed some contrasting trends.In the 3D-A431 model, a decline in metabolic activity was observed, while in the 3D-cSCC model, a progressive increase in metabolic activity was discerned over time.Notably, a lack of response to CTX treatment was evident after 21 d of culture in the 3D-A431 model, mirroring a similar unresponsiveness observed after 35 d in the 3D-cSCC model.Analyses of the dose-response to CTX of 3D-cSCC, shown as a relative percentage of viable cells in a control (untreated) (figure 7(B)), demonstrate a progressive increase in live cells over time, extending up to day 56 of culture.However, CTX resistance developed from day 28-56 at a concentration of 3 mg ml −1 and from day 28-112 for 2 mg ml −1 .No response to the CTX at 1 mg ml −1 was observed throughout the entire culture period up to day 112.

Discussion
Advances in bioprinting technology have opened new avenues for cancer studies, providing an essential tool for creating in vitro 3D tumor models for basic and clinical cancer research.However, many existing 3D tumor models lack diversity in cell types, limiting their ability to represent the complex tumor microenvironment accurately.Currently available models of cSCC include 2D and 3D tissue cultures, syngeneic and transgenic mouse models, as well as xenotransplants derived from cell lines and patient-derived xenografts [14].To date, only one 3D tricellular cSCC model has been described, posing an important limitation owing to the lack of comparisons to other models [15,16].This study, focusing on cSCC, incorporated essential cellular components into the 3D bioprinted model.
To address a notable gap in the existing literature, this study directly compares 3D bioprinted models, spheroids, and 2D cultures and drug responses in those models.This approach was designed to create a more physiologically and pathophysiologically relevant platform for studying drug effects and exploring cell behavior and interactions.
The tumor stroma of cSCC includes fibroblasts and keratinocytes.To assess the feasibility of incorporating diverse cell types, we initially developed a bioprinted normal skin model and characterized its morphology and growth kinetics.In the subsequent stage, layers of tumor cells, fibroblasts, and keratinocytes were printed to mimic the cSCC microenvironment, examining how the complexity of the model influenced its morphology and drug response.The printed tricellular cSCC model (3D-cSCC), was compared to the bioprinted 3D model consisting exclusively of tumor cells (3D-A431), as well as to spheroids and 2D models.
Under normal physiological conditions, keratinocytes and fibroblasts typically undergo approximately 40-60 divisions before entering a state of controlled senescence [17].Keratinocytes undergo a cyclic renewal process that limits their division capacity while ensuring the continuous production of skin cells.This process involves gradual migration toward the skin surface, accompanied by differentiation and eventual cell death, all contributing to the formation of the stratum corneum.This continuous skin renewal process persists throughout an individual's life [18].This phenomenon was also observed in the printed 3D-skin construct that mimics native skin.Despite a viability decrease in the initial stage, the keratinocytes and fibroblasts in the 3D-skin model showed high viability and growth even up to 16 weeks of culture, which provides a unique observation not reported in other available studies.Histopathological analysis of the constructs proved a close resemblance of the 3D-skin model to normal skin tissue, including stratified epithelial differentiation.Additionally, the presence of proliferating cell layers, densely packed fibroblast fibers, and skin-like changes in cell organization were observed.Furthermore, the maturation of the immortalized HaCaT keratinocytes was observed, demonstrating their retained capacity for differentiation and keratinization.Previous reports have highlighted the limitations of traditional monolayer models of keratinocytes, as they lack crucial features, such as stratification, vertical cell contacts, and differentiation [19].It is important to note that the growth patterns of keratinocytes can vary depending on their origin and the culture methods employed.Moreover, in the absence of fibroblasts, HaCaT skin keratinocytes tend to form a thin layer of epithelium and subsequently lose their proliferative capacity, as demonstrated in a study by Stark et al [20].Therefore, the approach presented here contributes significantly to the development of in vitro 3D models for studying normal skin development, including the formation of the epidermis and the dermal layer, and regeneration, and investigating tumor cell organization and drug response [14,21,22].
An important observation of the bioprinted 3D structures was the tendency of tumor cells to form spheroid-like clusters, exhibiting an increasing diameter over time.This phenomenon was evident in constructs consisting solely of tumor cells (3D-A431) and in tricellular tumor models (3D-cSCC).Over time, cellular aggregates appeared and merged into larger clusters, in line with previous research findings [6, [23][24][25].The transition from smaller to larger clusters, encompassing the tumor core, was particularly noticeable in tricellular 3D-cSCC constructs.In the initial phase of 3D cultures, the constructs exhibited the designed cellular arrangement that evolved in shape and spatial organization throughout the cell culture period.Over time, in addition to the formation of proliferative cancer cell clusters, densely packed fibroblast fibers and keratinocyte shifts in cell layer organization were documented.Research by Bera et al [26] demonstrated that cancer cell clusters can attract different types of cells, including fibroblasts.In 3D models, cell migration might be facilitated by a hydrogel serving as the ECM, providing a non-adhesive environment for cells.The parameters of a hydrogel, including its viscosity, significantly influence cell migration and can alter the structure of the cell-hydrogel interface [26,27].The hydrogels used in this study exhibiting a viscosity ranging from ⩾7 kPa•s at 0.01 s −1 to ⩽3 Pa•s at 200 s −1 , supported structural integrity within the printed constructs while minimizing cell harm in the initial stages.This viscosity provides an ideal milieu for cell migration, as recent data [26] indicate that migration speed increases with increasing extracellular viscosity, reaching a peak at 5-8 mPa.In the model presented here, changes in cell organization became evident after three weeks, suggesting a programmed alignment of cells according to their functions.The created cell culture conditions seemed to facilitate this transformation.Initially, cells underwent undisturbed growth, while over time, maturation set in, accompanied by the cyclic renewal of cells.What sets the current study apart is the exceptional longevity of the bioprinted cellular constructs, which remained viable for nearly 112 d (16 weeks), while the common observation period in most existing studies was usually limited to 14 d or less [6,25].Cells that do not survive contribute to the ECM [28], which could also explain the reduction in the number of keratinocytes and fibroblasts during prolonged periods of construct culture.
To faithfully recreate the ECM in the bioprinted 3D-cSCC tumor model, a lab-made bioink formulation was used.This formulation, consisting of a combination of fibrinogen, methylcellulose, and alginian, was applied to support cell attachment and migration and to allow the growth of several cell types.Furthermore, it maintained cell viability and proliferation for up to 16 weeks without additional interference.Similar combinations are often used to generate a matrix for skin cultures [27].
It is also essential to address the issue of cellular competition between different cell types, as it can significantly impact results by favoring betteradapted cells [29].This underscores the importance of optimizing bioink physicochemical properties, cell densities, and cell proportions.Within the developed 3D multicellular model, diverse cell division patterns and growth rates among cell types could result in tumor aggregates outcompeting normal cells that might have reached their Hayflick limit [30].Additionally, the mechanical stresses induced by the bioprinting process can contribute to morphological changes.The above factors may potentially influence the final cellular composition and the ability of cells to thrive and interact within the bioprinted environment.
The 3D models were functionally assessed by CTX treatment.Drug responses may be significantly influenced in vivo by the tumor microenvironment and in vitro by the type of experimental model and culture conditions.The underlying hypothesis was based on the premise that drug response within the tricellular 3D model would differ from that observed in 2D cultures and spheroids.A comparative analysis was conducted across various models to test this hypothesis.The analysis did demonstrate that the response to CTX treatment varied depending on the model and culture duration.In 2D cultures and spheroids, the highest concentrations of CTX induced A431 cancer cell death within a few days.In the bioprinted 3D-A431 and 3D-cSCC tumor constructs, a significant response was observed over the initial weeks of culture.However, as time progressed, a loss of drug sensitivity became evident.The difference in drug sensitivity already between 2D and 3D models could be linked to the observed morphological alterations of cells in a non-adherent environment, which have been associated with potential drug resistance [9, 23,25,[30][31][32].Interestingly, while comparing A431 spheroids with 3D-A431, a similar response to CTX at 3 mg ml −1 , i.e. a decrease in A431 cell proliferation, was observed.However, after 21 d, resistance became apparent in the 3D-A431 model.This resistance could be related to the formation of small cancer cell clusters, which might limit drug access to cells.However, as discussed further below in the context of the tricellular model, this may be more critical when larger clusters are formed.On the other hand, the absence of cell-cell contacts, which may trigger apoptosis, may play a role in the evidenced decrease in proliferation.This is further supported by the absence of proliferating cells (Ki67 staining) in the core of 3D-A431.
In the tricellular 3D-cSCC model, the reduction in cancer cell sensitivity to CTX was accompanied by an increase in cell proliferation.After an initial increase in apoptosis at the onset of CTX treatment, resistance developed as the tumor microenvironment matured and underwent morphological changes.This phenomenon may be attributed to several mechanisms.Drug accessibility may be limited due to the observed formation of larger cellular aggregates.Others have also reported that large tumor clusters showed greater drug resistance than small ones [20,28].Thus, aggregated A431 cells treated with anti-EGFR might evade apoptosis.The documented morphological changes in the complex structure of the tricellular 3D-cSCC model may also play a role, as suggested by Smida Rezgui et al [33], who have highlighted significant morphological changes, such as the formation of filopodia of surviving tumor cells, and described cell aggregation.Furthermore, Peuhu et al [34] suggested that such cell changes enable tumor spread and the formation of secondary tumors by overcoming the most significant barrier, the basement membrane.Apoptotic signals might also be suppressed by intercellular contacts with fibroblasts and keratinocytes, and due to the formation of fibronectin fibers.Kataoka et al [35] reached similar conclusions in lymphoma by demonstrating that communications between lymphoma cells and normal stromal cells inhibit cancer cell apoptosis.Likewise, Hu et al [36] reported that cancer cells, by actively inducing fibroblast migration towards cancer cell clusters, contribute to their limited drug response.The 3D cSCC model developed in this study reaffirms these mechanisms and demonstrates that including additional cellular components not only influenced cell behavior and morphological changes but also significantly impacted drug response.Previous studies, such as those by López de Andrés et al [15], have also shown that multiple cell types actively reshape the tumor microenvironment, consequently affecting cell responses to drugs.As demonstrated by Ayuso et al [9], cancer cells cultured with cancer-associated fibroblasts in a co-culture or in spheroids presented a notable increase in drug resistance, which was not the case when the same cells were cultured in a hydrogel.This highlights the complex nature of drug resistance mechanisms [25,[37][38][39][40][41], and the importance of a cell carrier, such as a hydrogel, that mimics the ECM [6, 26,32].
This investigation reveals compelling evidence that the pattern of cetuximab response observed in the clinical setting, characterized by an initial positive response followed by the emergence of resistance, can be modeled in vitro.
Therefore, this study not only validates the credibility of the developed 3D-cSCC model but also holds promising prospects for developing new and personalized treatment strategies.

Conclusion
This study successfully generated highly accurate and reproducible 3D bioprinted skin and cSCC models that faithfully replicated in vivo structures, as confirmed in morphological and immunohistological analyses.The 3D-cSCC model mirrors clinical cetuximab response patterns, which validates the potential application of the model for treatment personalization.Three-dimensional multicellular bioprinted tumor models surpass traditional 2D monolayer cell cultures and spheroids and offer invaluable insights into cell behavior and organization.Understanding complex cell interactions in a tumor is crucial for advancing cancer therapies.and editing of the manuscript, Funding acquisition.All authors accepted the final version of the manuscript.

Figure 1 .
Figure 1.Characterization of 3D NHDF and HaCaT cells bioprinted separately, and of the bioprinted 3D-skin model consisting of NHDF and HaCaT cells.(A)-Fluorescent images of the LIVE/DEAD assay on the 3D bioprinted NHDF and HaCaT cells after 3 d and 5 weeks of culture.Live (green), dead (red).Scale bar: 50 µm.(B)-Fluorescent images of the LIVE/DEAD assay on the 3D-skin model.Scale bar: 50 µm.(C)-Light microscopy images illustrating morphological appearance and structure changes after 3 d and 5 weeks of culture.Scale bar: 200 µm.(D)-Growth kinetics of the 3D-skin model, assessed by MTS.

Table 1 .
Primary antibodies used for immunohistochemical stainings.
a For negative control, phosphate-buffered saline (PBS) was used instead of the primary antibody; no staining was detected when the primary antibody was omitted.b Ready to use.