3D Bioprinting for In Vitro Models of Oral Cancer: Toward Development and Validation

Abstract

The tumor microenvironment (TME) of oral carcinomas has highly complex contents and a dynamic nature which is difficult to study using oversimplified two-dimensional (2D) cell culture systems. By contrast, three dimensional (3D) in vitro models such as spheroids, organoids, and scaffold-based constructs have been able to replicate tumors three-dimensionality and have allowed a better understanding of the role of various microenvironmental cues in the initiation and progression of cancer. However, the heterogeneity of TME cannot be fully reproduced by these traditional tissue engineering strategies since they are unable to control the organization of multiple cell types in a complex architecture. 3D bioprinting is an emerging field that can be leveraged to produce biomimetic and complex tissue structures. Bioprinting allows for controllable and precise placement of multicomponent bioinks composed of multiple biomaterials, different types of cells, and soluble factors according to the natural compartments of the target tissue, aiming to reproduce the equivalent of the complex tissue. As such, 3D bioprinting provides a unique opportunity to fabricate in vitro tumor models with a complexity similar to that of the in vivo oral carcinoma. This will facilitate a thorough investigation of cellular physiology, cancer progression, and anti-cancer drug screening with unprecedented control and reproducibility. In this review, we discuss the role of 3D bioprinting in reconstituting oral cancer, the prospects of application to fill the literature gap, and the challenges that need to be addressed in order to exploit this emerging technology for future work in oral cancer research.

1. Introduction

Cancers of the oral cavity and oropharynx are comprising the 8^th most common cancer worldwide, with oral squamous cell carcinoma (OSCC) encompassing more than 90% of all oral malignancies [1]. According to the Global Cancer Observatory (GLOBOCAN) 2020, there were 377,713 new cases of lip and oral cavity cancer in the world, while there were 177,757 deaths. When taken together with the oropharynx, these two sites account for an estimated 476,125 new cases (figure 1) and are corresponding to 2.5% of total cancers [2] with the highest incidence and mortality reported with the pharynx and the tongue[3]. Global oral cancer is expected to increase in the future so that by 2035, the incidence is projected to rise to 856,000 new cases per year [4]. This growing trend presents a serious threat, particularly with the drop of the 5-year survival rates from 80% in localized early diagnosed cancer to 20% if the cancer has spread to regional lymph[5]. In addition to the detrimental health effects, oral cancer treatment carries an enormous financial cost that could be directly incurred by investigation, treatment, and surveillance, or indirectly arising from expenses of cancer therapy complications. In England, for example, the cost of treating oral and oropharyngeal carcinoma was £213 million during the period 2007 to 2011 (Keeping et al. 2018). Likewise, the average total cost of OSCC treatment in Australia was $92,958 throughout 2016/2017 [6].

Figure 1. The last release of the GLOBOCAN 2020 data produced by the International Agency for Research on Cancer (IARC) regarding oral cavity and oropharyngeal cancers.

(A) estimated number of new cases of lip and oral cavity cancer; (B) estimated number of deaths in lip and oral cavity cancer; (C) estimated number of new cases of oropharynx cancer; (D) estimated number of deaths in oropharynx cancer. The data represents estimates of the incidence and mortality of oral cancer and Oropharynx for both sexes and all ages in 185 countries of the world. The incidence and mortality for each sex were estimated with an age-standardized rate (ASR) per 100,000. Data source: (GLOBOCAN) 2020, (https://gco.iarc.fr) [7].

This situation poses an exciting challenge to investigate the complex mechanisms related to oral cancer initiation and progression. A major hurdle, however, in the clinical translation of potential anti-cancer drugs is the discrepancy in the in vitro to in vivo efficacy of drugs. Thus, the improvement in cancer therapeutics is strictly related to the use of physiologically relevant in vitro models that can recapitulate the native cancer microenvironment. Such advanced preclinical models can replace the current lengthy, costly, and risky anti-cancer drug development systems. Nowadays, the process of developing a new drug from the first testing to the final Food and Drug Administration (FDA) approval and marketing may take more than 13 years and cost up to $648 million [8]. Despite these significant investments, 95% of anti-cancer drugs tend to fail in clinical trials due to inefficacy or toxicity, regardless of showing promise during preclinical testing [9]. Advanced cancer screening models can decrease the costs related to the drug development process by reducing the number of animals needed during the preclinical trial phases. More importantly, efficient in vitro models can reduce the clinical trial failure and drug withdrawal during the post-market surveillance since these models provide a more reliable response to anti-cancer agents when compared to 2D cell culture and animal models [10]. Conventional cancer models, both two-dimensional (2D) and three-dimensional (3D), cannot mimic the complexity of tumors as they are unable to reconstitute the heterogeneity of the tumor microenvironment (TME), and are also unable to precisely control the organization of various cellular and non-cellular components. Therefore, the use of these models for drug screening and fundamental cancer biology studies is restricted. 3D bioprinting is an emerging method in which biomaterials containing cells and other bioactive factors can be stacked to form a complex 3D structure, capturing the native living tissue complexity [11].

In this review, we have discussed the significance of 3D bioprinting technologies in oral cancer research and their potential utility in building in vitro constructs that can recapitulate oral TME. In the context of oral cancer modelling, we discussed the key limitations of the current conventional methods as well as the advantages, the future directions, the challenges, and the perspective solutions linked to the implementation of bioprinting oral cancer studies.

2. Heterogeneity of Oral Tumor Microenvironment

TME refers to the local environmental conditions immediately surrounding and influencing a tumor, including the extracellular matrix (ECM), neighboring stromal cells, soluble growth factors, blood vessels, nutrients, and waste[12]. The TME also includes physical factors such as interstitial flow, shear stresses, fluid forces, ECM stiffness, and topography [13]. These biochemical and mechanical elements provide continuous feedback to tumors, driving all hallmarks of cancer including apoptosis, angiogenesis induction, energy metabolism deregulation, resistance to the immune detection, genetic and epigenetic status, invasion, and metastasis [14]. OSCCs are remarkably heterogeneous due to the site-specific diverse stromal composition, epithelial turnover rate, microbial flora, saliva, and close proximity to the external environment. Such conditions have the potential to modulate the pathogenesis of OSCC [15]. Indeed, the presence of different anatomical sites, each of which has a distinct set of mucosal and submucosal tissue composition, is a quite unique feature of oral tissues that leads to cancer heterogeneity. For example, the gingiva and hard palate are characterized by dense collagenous stroma attached to the periosteum of bone. The soft palate is characterized by the presence of minor salivary glands surrounded by dense stroma, while buccal mucosa is comprised of adipose tissue, muscles, and minor salivary glands, and the tongue contains primarily muscle tissues. Therefore, the presence of site-specific inter-tumoral heterogeneity is quite conceivable in OSCC. In this regard, Frohwitter et al.[16] studied the difference between OSCCs of the floor of the mouth, tongue, and other oral sites concerning their molecular background. It was found that the X-linked inhibitor of apoptosis (XIAP) showed a significantly higher expression in OSCCs of the floor of the mouth compared to OSCC of the tongue and other locations within the oral cavity. Also, expression levels of p53, CA IX, beta-catenin, Hif-1-alpha, and c-kit were observed to be inversely related to the SCCs of the floor of the mouth and those of the tongue respectively. Similarly, Sathyan et al. [17] studied the possible differences in the cell cycle regulatory mechanism of OSCC in buccal mucosal and tongue by comparing the expression of major cell cycle regulatory proteins. The combined analysis showed simultaneous downregulation of p16 and p21 in 47% of tongue cancer cases, compared to 28% in buccal carcinoma. It was concluded that tongue and buccal mucosa cancers represent divergent biological sub-entities with potential prognostic therapeutic implications.

Despite progress made with 3D in vitro models, heterogeneous TME models remain difficult to reproduce. In light of the evidence implicating TME as a cancer inhibitor or promotor [18, 19], the ability to model its heterogeneity is crucial for understanding cancer biology, as well as for developing anti-cancer therapy. Fundamentally, there are three essential prerequisites for the in vitro oral cancer model to address the challenge of heterogeneity. The model should be comprising ECM-mimetic materials comparable to the native tissue counterpart, incorporating multicellular components to elucidate paracrine communication, and integrating perfusable network to illustrate vascular-dependent mechanisms such as metastasis or chemotherapeutics delivery.

3. Conventional in vitro oral cancer models

3.1. 2D models

In 2D culture, cells grow as a monolayer in a culture flask or in a flat petri dish. This model has remained the most common method for in vitro cancer studies as it is inexpensive, simple, and reproducible. However, this oversimplified culture method cannot represent the complex, heterogeneous, and dynamic TME due to several limitations such as poor cell differentiation, unrealistic cell proliferation, decreased drugs resistance, and inaccurate response to mechanical or chemical stimuli [20]. These limitations lead to strong discrepancies when compared to the in vivo counterparts in terms of morphogenesis, differentiation, proliferation, migration, genes and protein expression, signal transduction, and responsiveness to stimuli [21,22]. Several studies have shown that the 2D pre-clinical model is considered a suboptimal system and might result in misleading research observations because cells respond differently to anti-cancer drugs when cultured in 2D, or as a 3D construct [23]. Early evidence of this phenomenon was reported in 1985 when Miller and colleagues demonstrated that tumor cells grown as multicellular spheroids in a collagen gel exhibited greater drug resistance when compared to the cells cultured in a monolayer [24]. Similar results were obtained in recent studies which compared the viability of head and neck squamous cell carcinoma (HNSCC) cells in 2D and 3D spheroids after treatment with cisplatin and radiation. The results showed that cells grown in 3D tumor spheroids were more radio-resistant and had higher viability after treatment with increasing doses of cisplatin [25,26]. In an attempt to mitigate drawbacks linked with monolayer cultures, in vitro 3D cancer models have been subsequently developed based mainly on spheroids, organoids, and those based on the utilization of 3D matrices.

3.2. 3D models

3.2.1. Spheroids

Spheroids are self-assembled cultures of tumor cells formed in conditions where cell-cell interactions predominate over cell-substrate interactions. Cell spheroids offer several advantages over 2D cell cultures including better cell-cell interactions and diffusive mass transport. Multicellular tumor spheroids resemble avascular tumor nodules of large solid tumors with respect to their morphological features, volume growth kinetics, cell proliferation, drug access, gradients of nutrient distribution, oxygen concentration, and hypoxic core [27]. The cells present in the spheroid center form a necrotic core due to limited diffusion of nutrients, growth factors, and accumulation of metabolic wastes which causes a sharp decrease in the pH of the cellular microenvironment. This pH gradient can affect the drug release profile as well as the expression of genes controlling multiple drug resistance (MDR), such as MDR-1 which can reduce cellular uptake of hypoxic end-effector therapeutics [28]. In contrast, cells in the outer rim of spheroids (100–300 μm) have higher proliferating potentials due to easy access to oxygen and nutrients [29]. To date, a number of studies demonstrated that HNSCC spheroids are a promising model in reproducing many characteristics of tumor tissue, especially growth state and hypoxia [30–33]. However, the application of such a system is restricted by limited self-renewal and the differentiation capacity of spheroids. Also, spheroids can be used only for mimicking microniches because the control of spheroids during culturing is difficult and reaching larger than a certain size causes core necrosis [34]. More importantly, spheroids lack tumor tissue complexity because they are often composed of a single cell line and have insufficient ECM deposition [27]. The development of stroma-rich spheroids, consisting of epithelial HNSCC cancer and stromal cells that play a key role in tumorigenesis like cancer-associated fibroblasts (CAFs) provide a more relevant tumor model [35].

3.2.2. Organoids

An organoid is an extended cellular spheroid that has a physicochemical environment very similar to the tissue it is representing. In organoids, cells are arranged in a more ordered architecture on specific matrices, compared to spheroids where the cells are arranged in an irregular shape with no external matrix [36]. Organoids can display various biological features seen in vivo, such as tissue organization, regeneration, responses to drugs, and retaining the heterogeneity of original tumors [37]. Recently, Driehuis et al.[38] established human mucosal organoids from 31 patient-derived HNSCC samples, originating from different head and neck areas. The immunohistochemical and genetic characterizations revealed that organoids retained histologic characteristics and molecular alterations of the original tumor specimens. In terms of functionality, differential sensitivity of organoids to a panel of drugs including cisplatin, carboplatin, cetuximab, and radiotherapy in vitro was observed, and indicated that HNSCC organoids can be used as a platform for drug screening. Additionally, organoids displayed levels of atypia, tripolar mitotic figures, nuclear pleomorphism, and muscle invasion upon subcutaneous xenotransplantation into mice which suggests that HNSCC organoids are capable to elicit tumorigenicity in vivo. Collectively, these observations indicated that organoids can recapitulate OSCC genetically, histologically, and functionally in a manner that encourages their use as a potential platform for personalized cancer therapy.

Despite the advantages of organoids, several limitations still need to be resolved. Tumor organoids often recapitulate tumors in a single organ, but they cannot reconstitute cancer metastasis in multi-organ examples. In addition, cancer organoid culture cannot precisely organize the spatial distribution of cellular and non-cellular elements of TME [37]. This results in organoids varying in complexity, size, morphology, and 3D architecture which leads to difficulties in standardizing the culture [39].

3.2.3. Scaffold-based models

Scaffolds made from natural (e.g. collagen, Matrigel, and silk), synthetic (e.g. poly ethylene glycol, and poly lactic-co-glycolic acid), or a combination of materials provide robust support for cell growth and can simulate native ECM in many of their mechanical and biochemical properties [40]. The scaffold-based approach has been utilized for engineering various oral tissues that are used to study the underlying mechanisms governing cancer. For example, human tissue engineered oral mucosal model (TEOMM) has been developed by various research groups via growing keratinocytes seeded onto the fibroblast-populated matrix and cultured at the air-liquid interface [41]. TEOMMs have enabled several investigators to capture different stages of OSSC such as oral dysplasia [42], early invasive OSSC [43], and neoplastic transformation associated with stromal fibroblasts [44]. For example, Sawant et al. [45] developed a TEOMM representing normal, dysplastic, and malignant tongue tissues to study neoplastic progression during the various steps of human tongue tumorigenesis. The findings of histomorphometry, immunohistochemistry, and electron microscopy analyses of the three model types showed that the stratified growth, cell proliferation, and differentiation were comparable between co-cultures and their respective native tissues. Such findings showed that the major steps of oral tumorigenesis can be reproduced in vitro using scaffold-based engineered tissues. However, it must be considered that the differences between conventional 3D models and native TME are not simply a matter of dimensionality. Instead, tumorigenesis is a more complex process where multiple cues synergistically drive the hallmarks of cancer. Accordingly, the more accurate capture of in vivo scenarios dictates more biomimetic in vitro equivalent. Generally, the key limitations of the conventional in vitro 3D models are limited vascularization potential and lack of well-organized spatial distribution of tumor cells and ECM compositions.

4. 3D bioprinting for oral cancer models

4.1. An overview of 3D bioprinting

Bioprinting is a form of robotic 3D fabrication technology in which successive layers of cell-laden biomaterials are incrementally stacked on top of one another to construct complex 3D functional living tissues or artificial organs. Bioprinting modalities comprise of laser-induced bioprinting, inject-based bioprinting, extrusion-based or robotic dispensing bioprinting (Figure 2), which have been described in detail elsewhere [46–48]. Generally, bioprinting can be performed either in a scaffold-free or a scaffold-based manner. In the former approach, cell aggregates are printed on a sacrificial material which is subsequently replaced by ECM components. In the latter approach, by contrast, a bioink comprising of cells encapsulated in hydrogel is used [49]. A comprehensive review conducted by Hospodiuk et al. [50] classified scaffold-based biomaterials into natural or synthetic hydrogels, decellularized matrices, and microcarriers, while scaffold-free bioink materials include tissues spheroids, cell pellets, and tissue strands. The comparative evaluation of the bioink materials revealed a substantial difference among them in terms of bioprintability, cell viability, biomimicry, resolution, scalability, mechanical and structural integrity, post-bioprinting maturation times, tissue fusion, as well as other factors. Taking these variables into consideration leads to the careful selection of the bioink to meet the requirements of the native tumor matrix which is highly complex and unique for each tumor type and stage [51]. Various multilateral bioinks have been used to meet the mechanical and functional requirements that cannot be achieved by a single biomaterial. For example, natural polymers, such as collagen and gelatin, have been widely used because they contain the tripeptide Arginine-Glycine-Aspartate (RGD) motifs important for cell attachment and migration. However, these materials often suffer from low mechanical properties, and thus, other biomaterials and additive elements have been combined to obtain multi-material bioinks with improved properties when compared with conventional single material-based bioinks [52]. For example, in skin bioprinting, gelatin-collagen hydrogel is used to provide appropriate viscosity and tissue thickness, respectively. However, this mixture often results in tissue contraction that lasts 7 days before the addition of keratinocytes. Modulation of this bioink formulation by adding fibrinogen and elastin has allowed for the production of a human-like skin with minimized tissue contraction, improved structural and barrier function, as well as eliminated the need for a long incubation time before the application of epidermal cells [53].

Figure 2. Schematic diagrams of 3D bioprinting technologies.

Laser-induced forward tranfer bioprinting use lasers focused on a laser energy absorbing layer to generate a high gas pressure propelling the bioink onto a collector substrate. Thermal inkjet printers electrically heat the printhead to produce air-pressure pulses that expels droplets from the nozzle, whereas in piezoelectric inkjet printing no heating is used, but a direct mechanical pulse is applied to the fluid by a piezoelectric actuator that forces the bioink through the nozzle. Robotic dispensing or extrusion bioprinting use either pneumatic, piston- or screw-driven dispensing systems to extrude continuous beads of bioink on a building platform. Reproduced with permission from ref. [54].

4.2. Advantages of bioprinting in the reconstitution of oral cancer

4.2.1. Multiple cell populations in a defined spatial architecture

TME contains many distinct cell types including fibroblasts, myofibroblasts, endothelial cells, neutrophils, eosinophils, basophils, mast cells, lymphocytes, macrophages, as well as others [55]. These cells play their role in promoting or rejecting cancer by dependence over signaling cues received from their group or from a different cell group. Cell communications involve direct adjacent communication with nearby cells (juxtacrine), indirect local communication over short distances (paracrine), or large distances via hormones (endocrine) [56]. Several lines of evidence demonstrated that the mutual paracrine effects of OSCC and intratumor stromal cells, in particular CAFs, macrophages, and endothelial cells, play an important role in promoting tumor growth and progression [55, 57, 58]. For example, it was demonstrated that during tumor invasion, OSCC cells induced the transdifferentiation of primary oral normal fibroblasts into myofibroblasts via secretion of transforming growth factor-beta 1 (TGF-beta 1). In turn, myofibroblasts released factors that stimulated OSCC cell proliferation [59].

Bioprinting permits the incorporation of several cancer-associated cell types to better represent the reciprocal cellular interaction in solid cancers (figure 3). Recently, Langer et al. [60] bioprinted tumor tissue containing cancer cells, fibroblasts, and human umbilical vein endothelial cells (HUVECs). These bioprinted tissues replicated solid tumor architecture, in which a core tumor cell bioink was surrounded by a normal stromal cell bioink. Following several days in culture, staining showed the presence of collagen fibers, predominantly in the stromal compartment which indicates that the original bioprinted spatial organization was maintained. Also, increased staining at day 10 versus day 7 demonstrated that the cells within the tissues are depositing their own ECM and maturing over time. Together, these pieces of data show that bioprinted epithelial and stromal cell types can self-organize and interact to form tumor-like tissue.

Figure 3. Three-Dimensional Bioprinted Model of Pancreatic Cancer.

(A) Bioprinted tissue containing a patient-derived xenograft (PDX)-derived cell line from a human pancreatic cancer cell line surrounded by normal human primary pancreatic stellate cells (PSCs) and HUVECs after 7 days in culture. IF for epithelial cancer cells KRT8/18 (green), stromal fibroblasts VIM (red), and endothelial cells CD31 (yellow) with DAPI (blue) nuclear counterstain (Scale bar 500 μm). (B) The tumor-stromal border with VIM-positive cells within the tumor core, and KRT8/18 cells in the stroma (Scale bar 200 μm). (C) IF for KRT8/18 (green) and Ki67 (red) with DAPI (blue) (Scale bar 200 μm). (D) IF for α-SMA in PSCs after 7 days in culture (Scale bar 50 μm). (E) Digital image on bioprinted pancreatic tumor tissue printed with PSCs and HUVECs in the stromal compartment and human pancreatic cancer cell line (HPAFIIs) and HUVECs in the cancer compartment after 10 days of culture (Scale bar 100 μm). (F) Pancreatic tumor tissues printed as in (E) were treated for 6 days with either 10 or 50 μM gemcitabine. IF performed for KRT8/18 (green) and CD31 (yellow) with DAPI (blue). (G) Day 10 quantitation of total flux for bioprinted tissues grown as in (E). (IF: immunofluorescence; KRT8/18: Cytokeratin 8/18; VIM: vimentin; α-SMA: Alpha-Smooth muscle actin). Reproduced with permission from ref.[60].

4.2.2. Spatial control on matrix properties

Spatial distribution of biochemical factors can be incorporated in a 3D bioprinted model to mimic the native tissue (figure 4). Freeform reversible embedding of suspended hydrogels (FRESH) is one of the newer techniques of extrusion-based bioprinting methods for mimicking the complex biological structure of native tissues. FRESH embedded ECM soft proteins and polysaccharide hydrogels within a secondary gelatin hydrogel serve as temporary, thermoreversible, and biocompatible supports to print complex geometries that would otherwise be impossible using biomaterials such as collagen and alginate. Employing the FRESH technique allows the fabrication of 3D multi-material structures with reasonable mechanical integrity and complex internal and external architectures [61]. The integration of 3D printing with shell-core capsules is another applicable method for precise spatiotemporal control over chemical and biomolecular gradients. Gupta et al. [62] demonstrated this technique by printing stimuli-responsive core-shell capsules for the programmable release of multiplexed gradients within hydrogel matrices. These capsules are composed of an aqueous core and a poly (lactic-co-glycolic) acid shell. Notably, the shell can be loaded with plasmonic gold nanorods that permit selective rupturing of the capsule by controlling the wavelength of irradiation. The advantages of this method include highly monodisperse capsules, efficient encapsulation of biomolecules, precise spatial patterning of capsule arrays, programmable reconfiguration of gradients, and versatility for incorporation in hierarchical architectures.

Figure 4. Control of biochemical factors distribution in a bioprinted model.

Structural designs of 3D printed scaffold laden with DPSCs and spiked with BMP-2 and VEGF. Group 1: DPSCs printed structure using 2% collagen. Group 2: DPSC/BMP-2 printed structure using 2% collagen. Group 3: DPSC/dual growth factors of BMP-2 and VEGF using 2% collagen and 10% alginate/10% gelatin blend (DPSCs: Mesenchymal dental pulp derived stem cells; BMP-2: bone morphogenetic protein 2; VEGF: vascular endothelial growth factor). Reproduced with permission from ref.[63].

In addition to control over the biochemical distribution, bioprinted tumor models can adjust serval ECM mechanical properties such as matrix stiffness, pore size and architecture, proteins crosslinking and density, and fiber network configuration [64,65]. For example, the matrix stiffness of neoplastic tumors is considerably higher than neighboring normal tissue, and it is highly correlated with cancer progression, aggressiveness, and metastasis [66]. One of the bioprinting methods used to modulate matrix stiffness is generating hybrid multicomponent hydrogels which are either physically or chemically crosslinked. Beck et al. [67] introduced the composite hydrogel poly-ethylene glycol (PEG) (PEG/Matrigel) as a platform to investigate cancer cell metastasis. In the study, the rigidity of the matrix was tuned by varying the crosslinking density of PEG networks within Matrigel scaffolds, while cell adhesive signals were incorporated into the PEG networks using peptide-conjugated cyclodextrin. It was found that the adhesive PEG networks induced the dissemination of malignant epithelial cells at intermediate values of adhesion and rigidity. Another approach includes using a photocrosslinkable decellularized extracellular matrix (dECM) and digital light processing (DLP)-based 3D bioprinting. This method is capable of arranging dECM with tailorable mechanical properties to serve as a platform for studying the effects of 3D matrix stiffness on cancer progression and invasion. Rapid DLP-based 3D bioprinting technology has enabled the flexible design of physiologically relevant geometries, as well as precise control over the hydrogel’s mechanical properties. By changing the light exposure time, the stiffness can be easily controlled without modifying the hydrogel components. Thus, this technique may eliminate the potentially detrimental effects of different material concentrations or chemical compositions on cell behavior [68].

4.2.3. Vascular networks integration

One key weakness of current 3D models is the lack of vasculature which is essential for cancer cell interactions with endothelial cells during tumor proliferation, angiogenic recruitment, and intravasation [69]. Blood vessels are known to lead to the formation of oxygen diffusion gradients that can promote chemotactic tumor cell invasion [70] and angiogenic sprouting [71], as well as influence the delivery of chemotherapeutics to solid tumors [72]. Therefore, avascularized tissue-engineered tumors may model certain stages of cancer such as oncogenesis, but are incapable to reproduce other steps like metastatic dissemination of cancer cells which occurs primarily via blood vessels [73]. Accordingly, the development of vascular networks in a biomimetic cancer model is essential not only for maintaining tissue viability during the long-term post-printing maturation [74], but also for unraveling the close relationships between tumors and blood vessels. These include tumor-associated angiogenesis, anti-angiogenic treatment, drug delivery, and metastasis [75].

Generally, bioprinting techniques used for vascularized tissue fabrication fall into two broad categories. The first category includes directly bioprinting the vascular wall structures (figure 5), while the second category of bioprinting generates hollow channels within 3D matrices and subsequently, the channels are endothelialized [74] (figure 6). Many vascular models are being designed based on these two approaches. For example, alginate-based tubular structures were prepared by using the inkjet nozzle system. The wall thickness and the inner diameters of tubular structures were adjusted from 35–40 μm and from 30–100 μm, respectively, by altering the diameter of the microgel beads from 10 to 40 μm [76]. Another study showed the fabrication of scaffold-free vessels by using spheroids of various vascular cell types including smooth muscle cells and fibroblasts. The spheroids of 300–500 μm were printed layer-by-layer concomitantly with a molding template of agarose rods. The post-printing fusion of the discrete units resulted in single and double-layered, small-diameter vascular tubes ranging from 0.9 to 2.5 mm. This quick and scalable technique enabled the construction of hierarchical tree shaped vessels with distinct diameters [77].

Figure 5. Bioprinting of vascular constructs.

(A1) Representative images of bioprinted fibroblast tubes made of 2% alginate loaded with 5×10⁶ cells/ml; (A2) left inset: different views of a printed Y-shaped cellular tube bioprinted using 2% sodium alginate solution, and right inset: dyed cells in blue and dyed living cells in green (reproduced with permission from Ref.[78]); (B) Direct cell patterning in collagen hydrogels using a near-infrared femtosecond laser. The confocal microscopy images show tube formation after 14 days in the YZ plane and aligned endothelial cells in the XY plane (reproduced with permission from Ref.[79]). (C) Scanning electron microscopy images showing a complex vascular network in poly(ethylene glycol) diacrylate hydrogel (Reproduced with permission from ref.[80]).

Figure 6. 3D bioprinting of the prevasularized tissue constructs.

(1) (A) Schematic of the bioprinting platform. (B) Bioprinted acellular construct featuring intended channels with gradient widths. (C) Bioprinted cellular construct with HUVECs encapsulated in the channels. (D–F) Fluorescent images demonstrating the bioprinting of heterogeneous cell-laden tissue constructs with uniform channel width. HUVECs (red) are encapsulated in the intended channels and HepG2 (green) are encapsulated in the surrounding area. (G–I) Fluorescent images demonstrating the bioprinting of heterogeneous cell-laden tissue constructs with gradient channel widths. Scale bars 250 μm. (2) Endothelial network formation after 1-week culture of the prevascularized tissue construct in vitro, (A–C) Confocal microscopy images show HUVECs (Green, CD31-positive) and supportive mesenchymal cells (10T1/2, Purple, alpha-smooth muscle actin (α-SMA)-positive) aligned within the patterned channel regions with different vessel sizes. (D) Cross-section view shows the endothelial cells (CD31-positive) form lumen-like structures along the bioprinted channels. (E) 3D view of the endothelial cells lining along the printed microchannel walls by confocal microscopy. Endothelial cells were labeled by fluorescent cell tracker (red) and stained by CD31 (green). Reproduced with permission from ref. [81].

For bioprinting endothelialized hollow channels within a thick matrix, a rapid microscale continuous optical bioprinter was utilized to print prevascularized tissue with well-designed vascular channels without the use of sacrificial materials or perfusion systems. In this technique, gelatin methacryloyl (GelMA)-hyaluronic acid and GelMA-cell laden bioinks were used to print channels and channel-surrounding regions, respectively. HUVECs and mesenchymal stem cells (MSCs) were encapsulated in the hydrogel and printed into the perfusable vascular channel for endothelialization. These predesigned vascular channels represented a platform for high resolution control over the distribution of multiple cell types, where cells formed endothelial networks and lumen-like structures with higher cell viability [81].

4.3. Literature gap and future research

Several authors have investigated OSCCs using conventional 3D bioengineered oral tissues like TEOMM [41], bone [82, 83], and tongue [84]. However, to date, in vitro oral cancer models present two main limitations: the inability to accurately reconstitute TME of the native tissue counterpart, and the absence of a heterogeneous construct encompassing multiple tissue types in a single unit. Such drawbacks limit the validity of the current oral tumor models in translating in vitro findings which are heavily influenced by intra and inter tissue diversity. 3D bioprinting has successfully reproduced several tumor models such as breast tumors, ovarian cancer, pancreatic adenocarcinoma, and hepatocarcinoma[11]. However, the applications of bioprinting in oral cancer studies are still in their infancy, and many studies need to be undertaken in order to fully exploit the advantages of this technology in advancing oral cancer research. It is seen that the successful outcome of engineered tissue models relies on several key factors including mechanical properties, biochemical composition, and complexity of the final tissue. Due to this, the future direction of bioprinting in oral cancer modelling generally requires first fabricating a heterogeneous single oral tissue equivalent, and second, fabricating composite oral tissues constructs.

4.3.1. Fabricating heterogeneous single oral tissue equivalent

Capturing of the TME entails the inclusion of diverse cellular components as well as the modulation of matrix properties to fit with the emulated tumor. For example, skin containing fibroblasts embedded in a collagenous matrix and augmented with a layer of keratinocytes atop the dermis was bioprinted using different bioinks and bioprinting techniques to simulate the epidermis [85]. However, this uniform structure does not recapitulate the complexity of native skin with diverse population of cells, complex distribution of proteins, vasculature, and adnexal structures. The incorporation of additional cells, beyond keratinocytes and fibroblasts, and the adjustment of scaffold features will have an important impact on model validation. Recently, a research group bioprinted a novel 3D skin model with layers of HUVECs networks, dermal fibroblasts, and multilayered keratinocytes using biomaterials for each layer of skin tissue. In this model, the layer of HUVECs was bioprinted with the GelMA 7.5%-alginate2% bioink. Then, 7.5% GelMA hydrogels encapsulating fibroblasts were prepared on the layer of internal vascular networks. Lastly, multilayered keratinocytes have been obtained by multiple steps of seeding keratinocytes on the GelMA, followed by gelatin-coating. This technique modulated matrix stiffness, which played a critical role in regulating production levels of collagen type I alpha 1, and Matrix Metallopeptidase1. The study also showed that optimized gel can maintain the viability of HUVECs for 7 days, leading to the generation of bioprinted internal vessel networks with multiple keratinocytes being seeded with gelatin-coating between seedings, which shortened the culture time to create thicker epidermal layers [86]. Due to structural similarities between skin and oral mucosa, this method can be adopted to generate a more relevant tumor oral mucosa model, including other cell types such as HUVECs or immune cells.

4.3.2. Fabricating composite oral tissue construct

One of the special features of the oral and maxillofacial region is the presence of numerous tissue types such as bone, teeth, lining mucosa, covering skin, muscle, and a rich neural and vascular network within a relatively small area. This proximity often results in multiple tissue being involved with the neoplastic process which eventually carries implications on tumor staging, choice of treatment, prognosis, and quality of life [87]. For example, the invasion-metastasis cascade starts at the primary tumor site where the epithelial cancer cells locally breach the basement membrane to invade the underlying connective tissue stroma [88]. The tumor cells can penetrate lymphatic channels and migrate to regional lymph nodes in the neck (regional metastasis) or move via vessels and extravasate to distant metastatic sites (distant metastasis)[89]. Bone invasion is frequently observed in OSCC and is associated with more aggressive tumors with high recurrence [90]. Mandibular medullary invasion of OSCC, in particular, could be an independent prognostic factor that predicted a poor disease-specific survival [91]. Likewise, vascular and perineural invasion of tongue OSCC cells are considered prognostic factors for local recurrences, regional and distant metastasis, and worse disease-specific survival [92]. The progression of OSCC to diverse tissues poses a special challenge, not only in the construction of composite tissues, but also in assembling these tissues in their normal anatomical relationship. Although the application of bioprinting in the construction of composite tissue is still in the early stages, few promising studies demonstrated the feasibility of using this technology to fabricate hybrid oral constructs. 3D bioprinting is expected to enable the construction of hybrid models comprising of different tissues through the deposition of multicomponent bioinks composed of multiple types of biomaterial and cells, as well as additive materials or biomolecules [52]. For example, 3D bioprinted human periodontal/osteoblastic construct was developed to emulate periodontal ligament (PDL)-alveolar bone (AB) structure. PDL was modeled by using GelMA bioink for bioprinting human PDL fibroblasts. On the side, the alveolar bone was modeled by using a composite bioink comprised of GelMA and hydroxyapatite-magnetic iron oxide nanoparticles (GelMA/HAp-MNPs) used for bioprinting human osteoblasts. After 14 days in culture, characterization of 3D printed GelMA/HAp-MNP construct showed that the maximum strength value (~20 kPa) was twice as much as that of the 3D printed GelMA construct (9.3 kPa). In addition, scanning micrographs confirm that the microtissues preserve their shape without shrinkage or deformation in aqueous conditions. Moreover, it was evident that the bioprinted cells stayed highly viable and localized in their own layers, as revealed by positive stromal cell surface marker-1 and osteocalcin in the PDL layer and osteoblastic layer, respectively [93]. Such techniques can be modified and adopted to bioprint other oral composite constructs such as osteomucosal, osteochondral, musculoskeletal, and other hybrid tissues.

4.4. Current Challenges

Despite the significant progress in the bioprinting domain, several challenges in pre-printing, printing, and post-printing tissue maturation stages need to be overcome for widespread adoption of this technology in tissue modelling [74]. The most important step in pre-printing stage involves isolation, expansion, and differentiation of cells. With the emergence of bioengineered tissue models, the use of stem cells has become common. While promising, it is critical to notice the relevancy of the fabricated model is affected by the protocols employed to control differentiation and maturation toward the desired lineage. Cells differentiated from stem cells are often considered immature, and more resemble fetal cells than primary cells of native tissue. Therefore, using new methods facilitating differentiation pathway to reach mature phenotypes and functionalities are crucial for accurate data of disease processes [46].

Concerning the printing stage, material selection remains another major concern and limitation for bioprinting. The development of printable bioink that meets the requirements of each bioprinting technique is important, particularly for extrusion-based bioprinting due to its practicality and capability to produce scalable 3D tissue constructs. For example, bioink materials used in the extrusion technique should possess various properties, particularly shear thinning, rapid crosslinking, and solidification capabilities. The lack of rapid solidification of hydrogels leads to spreading upon extrusion, and inability to maintain the original bioprinted shape. In drop-based bioprinting, however, novel hydrogels should possess appropriate mechanical, structural, and gelation properties, yet should not possess a viscosity or fibrous microarchitecture that can easily clog the nozzle [50].

In addition to the printability property, the biological properties of hydrogels play a crucial role in successful tissue construction. One of the major drawbacks of the currently available hydrogel materials is that they do not support the differentiation of cells into multiple lineages. As the majority of native oral issues are heterogeneous with multiple cell types interacting in a highly complex microenvironment, the development of smart bioink materials that can facilitate differentiation of stem cells into different lineages is fundamental. In the bioprinting of composite constructs where different tissues of various anatomical and structural features are built-in single models, the development of hydrogels that allow smooth transitions between different parts would be more challenging. For example, the development of osteochondral models is considered problematic. One possible solution could be bioink reinforcement with nanomaterials that can induce differentiation of cells with a seamed transition in a chondral bone model [94]. Finally, the multicellularity of bioprinted models poses yet another challenge in post-printing tissue maturation and culture stage. Different cell types have different nutritional and metabolic requirements and consequently require different kinds of growth media. The development of a standard or universal growth media for cells would be beneficial for the growth and maturation of composite tissue constructs [74].

5. Conclusion

This review discussed the current state of 3D bioprinting to fabricate a biomimetic oral cancer model. Bioprinting is a very advanced fabrication technique that, compared to conventional microfabrication methods, offers unique advantages in the accurate reconstitution of oral TME. The ability of this versatile manufacturing system to offer precise dispensing of multicellular bioink, spatial control of matrix components, and integration of vascular network would facilitate the construction of more physiologically relevant in vitro models. These models could not only represent a single tissue type containing multiple cellular and matrix components, but also reconstruct composite tissues by attaching them together into one unit. Such models would provide a great opportunity to elucidate the poorly understood mechanisms relating to oral carcinoma, as well as to investigate the efficacy of new anti-cancer therapeutics that cannot be otherwise examined by using conventional cell culture methods. However, considering the challenges and limitations associated with different stages of 3D bioprinting is an important step toward achieving full harness of this technology. Nonetheless, an efficient application of 3D bioprinting in oral cancer research is still in the stage of infancy. Further advancements are expected to play a critical role in defining new directions and possibilities that enable the cancer biologists to explain the unclear mechanisms governing oral carcinogenesis.