Harnessing Next‐Generation 3D Cancer Models to Elucidate Tumor‐Microbiome Crosstalk

Abstract

The tumor microenvironment (TME) is a complex and dynamic ecosystem increasingly recognized for its interplay with the microbiome. In colorectal, breast, lung, liver, and brain cancers, bacterial communities and their metabolites are shown to influence tumor progression, immune responses, and therapeutic outcomes. To study these interactions in physiologically relevant contexts, advanced 3D in vitro models have emerged, including spheroids, organoids, microfluidic organ‐on‐a‐chip platforms, and 3D‐bioprinted constructs. These systems provide spatial organization, mechanical cues, and co‐culture capabilities that facilitate investigation of host–microbiome–tumor cross‐talk. Incorporation of live bacteria, their metabolites, and immune components into these platforms has yielded new insights into how the microbiome shapes cancer behavior, inflammation, and drug resistance. This review outlines recent advances in 3D model development for studying tumor–microbiome interactions, highlighting organ‐specific applications, extracellular matrix‐mimicking hydrogels, and biofabrication strategies. It also addresses key challenges, including maintaining microbiome viability, modeling temporal dynamics, and integrating immune complexity. Overcoming these limitations requires interdisciplinary approaches that merge bioengineering, microbiology, and oncology. Evolving 3D platforms offer powerful tools for microbiome‐informed cancer modeling and hold significant promise for advancing therapeutic screening and precision oncology.

Centralizes the microbiome within 3D tumor‐microbiome model platforms, including spheroids, organoids, 3D‐bioprinted constructs, and microfluidic chips, each enabling structured host‐tumor‐microbe studies. These systems support bacterial colonization, facilitating investigation of microbial impacts on tumor growth, immunity, and therapy. The microbiome anchors a systems‐level perspective of the tumor‐microbiome‐host triad, underscoring the need for integrated dynamic models. Created with BioRender.

1. Introduction

As the microbiome is increasingly understood to be a crucial modulator of cancer biology, impacting tumor initiation, progression, and therapeutic outcomes, the relationship between cancer and the microbiome has taken center stage in modern‐day oncology.^{[ 1} , ² , ^{3 ]} The TME, once thought to be sterile, is now recognized to contain various bacterial communities. These bacteria interact with distinct components of the TME, including immune cells, stromal cells, extracellular matrix (ECM), and vasculature, with each type of tumor exhibiting a distinct microbiome profile.^{[ 4} , ⁵ , ^{6 ]} By influencing immune cell recruitment, inducing immune evasion, and altering drug metabolism, these intra‐tumoral bacteria can affect the course of cancer^{[ 7 ]} (Figure  1 ).

Figure 1.

Tumor‐microbiome interaction mechanisms. This figure illustrates the multifaceted interactions between the TME and the microbiome, and how these interactions can influence tumor evolution through distinct mechanisms. Mechanisms 1‐3 represent pro‐tumorigenic effects, while mechanisms 4–7 represent anti‐tumorigenic effects. Created with BioRender.

Dysbiosis, a disturbance in the composition and function of microorganisms, has been linked to carcinogenesis through processes including the production of genotoxic metabolites, chronic inflammation, and epigenetic changes in host cells.^{[ 4 ]} For instance, Helicobacter pylori (H. pylori) is a known risk factor for gastric cancer because of its inflammatory effects and virulence factors like CagA,^{[ 5 ]} while Fusobacterium nucleatum (F. nucleatum) is a known risk factor for breast and colorectal cancers (CRC), enhancing tumor growth and metastasis by inducing M2 macrophage polarization and suppressing T cell activity.^{[ 6 ]} By rearranging the actin cytoskeleton, intratumoral bacteria can also increase the tumor cells' resistance to mechanical stress and promote metastasis.^{[ 7 ]}

By influencing immune cell differentiation, cytokine production, and inflammation, or by causing DNA damage and metabolic reprogramming that can result in T cell exhaustion and immune suppression, microbiome metabolites like lactate, colibactin, ammonia, and short‐chain fatty acids (SCFAs) also play a variety of roles in tumor biology.^{[ 8} , ^{9 ]} Growing interest in how altering the microbiome may improve cancer survival and treatment effectiveness has been spurred by these discoveries.

Additionally, the gut microbiome is crucial in regulating how the body reacts to immunotherapy, radiation, and chemotherapy.^{[ 10 ]} For example, some bacteria in pancreatic tumors can metabolize gemcitabine and other chemotherapeutic agents, making them inactive and causing resistance to treatment.^{[ 11} , ^{12 ]} On the other hand, commensal bacteria like Bifidobacterium species and Akkermansia muciniphila (A. muciniphila) have been linked to enhanced antitumor immunity and improved responses to immune checkpoint inhibitors (ICIs), such as anti‐programmed cell death protein 1 (anti‐PD‐1) and anti‐cytotoxic T‐lymphocyte‐associated protein 4 (anti‐CTLA‐4) treatments.^{[ 13} , ^{14 ]} Although antibiotic‐induced microbiome imbalances, dysbiosis, have been linked to decreased efficacy of immunotherapy,^{[ 15 ]} certain bacterial metabolites can regulate immune checkpoint pathways like PD‐1/PD‐L1, further influencing immunotherapy effectiveness.^{[ 8 ]}

Importantly, the microbiome influences host metabolism and systemic immune responses in addition to local tumor biology. Through immune modulation, metabolic cross‐talk, and bacterial translocation, host–microbiome interactions can have a substantial impact on the course of cancer and the effectiveness of treatment.^{[ 16} , ^{17 ]} For instance, in germ‐free or antibiotic‐treated animal models, the effectiveness of chemotherapeutic agents such as cyclophosphamide and cisplatin is reduced, highlighting the function of commensal bacteria in mediating therapeutic responses.^{[ 18 ]}

Additionally, clinical research has demonstrated that the diversity and composition of the microbiome may function as predictive biomarkers for the results of immunotherapy and chemotherapy.^{[ 16} , ¹⁸ , ^{19 ]} In CRC, high levels of F. nucleatum are linked to resistance against oxaliplatin and 5‐fluorouracil (5‐FU) via autophagy activation and inhibition of apoptosis. In contrast, a higher abundance of butyrate‐producing bacteria, such as Roseburia and Dorea, correlates with better chemotherapy response, while their depletion predicts poor prognosis.^{[ 20 ]} In bladder cancer, the presence of drug‐metabolizing bacteria has been linked to lower intravesical concentrations of gemcitabine and worse outcomes. These results suggest that targeted microbiome modification could improve therapeutic efficacy and support the potential of microbiome profiling as a non‐invasive biomarker to guide personalized cancer therapy.^{[ 21 ]}

Many studies are being conducted to investigate these interactions, as the microbiome plays a crucial role in determining tumor biology and therapeutic outcomes (Figure  2 ). The implementation of experimental models that faithfully capture the intricate and ever‐changing characteristics of the TME, including the microbiome, is still a significant obstacle (Table  1 ).

Figure 2.

A Timeline of landmark developments in tumor modeling technologies and microbiome–cancer insights. The top tier presents breakthroughs in tumor modeling systems, from early tumor‐derived spheroids^{[ 23 ]} to advanced platforms, such as lung‐on‐a‐chip, long‐term epithelial organoid culture,^{[ 24} , ^{25 ]} 3D‐bioprinted tumoroids with perfusable vessels,^{[ 26 ]} and Tumor‐on‐a‐chip models.^{[ 27} , ^{28 ]} These innovations enabled dynamic, physiologically relevant in vitro studies of tumor biology. The bottom tier highlights key microbiome–cancer milestones: the first link between H. pylori and gastric cancer,^{[ 29 ]} the Human Microbiome Project,^{[ 24 ]} gut‐on‐a‐chip models (HuMiX),^{[ 30} , ^{31 ]} bacterial‐mediated chemoresistance,^{[ 12 ]} and AI‐based prediction of tumor drug response.^{[ 32 ]} The ONCOBIOME Network was established to systematically investigate microbiome signatures across diverse cancer types in large patient cohorts.^{[ 33 ]} Most recently, it has been demonstrated that intratumoral bacteria promote metastatic colonization by enhancing tumor cell survival under stress.^{[ 34 ]} Together, these milestones form the foundation for integrative tumor–microbiome research. Created with BioRender.

Table 1.

A summarizing table comparing 2D to 3D to animal models in regard to microbiome‐inclusive TME.

Model Type	Key Features	Limitations	Relevance to Microbiome Studies	References
2D Cell Cultures	Easy to use; reproducible	Lack 3D structure, gradients, ECM; not compatible with microbiome co‐culture	Poor representation of TME and host‐microbiome interactions	[39, 40]
Animal Models	Systemic context; intact immune system	Species‐specific differences in immunity, metabolism, microbiome; high cost; low throughput; ethical concerns	Useful but limited in predictive value and microbiome relevance	[41, 42]
3D Models (Spheroids, Organoids, Bioprinted)	Mimic 3D architecture, oxygen/nutrient gradients, ECM stiffness; support both aerobic and anaerobic microbiome	Complex to fabricate and maintain	Enable mechanistic studies of host‐microbiome‐tumor crosstalk; drug and immune response profiling	[43, 44]

The spatial and biochemical complexity of the TME cannot be adequately recapitulated by conventional 2D cell culture models, despite their widespread utility due to their ease of use and reproducibility. They are devoid of oxygen/nutrient gradients, mechanical stiffness, and 3D architecture, all of which are necessary for preserving physiological cell behavior. Furthermore, because they do not promote bacterial growth conditions or physiologically relevant interactions with host tissues, they are inappropriate for microbiome co‐culture.^{[ 22 ]}

Although intact immune systems and a systemic context are advantages of animal models, there are substantial translational challenges. Their predictive value may be limited by variations in immune signaling, metabolism, and microbiome composition between model organisms (mainly immunocompetent rodent models for murine cancers or immunodeficient for human cancer xenografts) and humans. Furthermore, their widespread use in microbiome‐related research is hampered by low scalability, high costs, and ethical concerns.^{[ 35 ]}

On the other hand, 3D cancer models, such as spheroids, organoids, and 3D‐bioprinted constructs, have become viable tools for bridging the gap between intricate in vivo models and basic in vitro systems.^{[ 36} , ^{37 ]} These systems are designed to support microbiome colonization in both aerobic and anaerobic environments, and they can replicate multicellular architecture, oxygen, and nutrient gradients, and ECM stiffness. According to research, 3D models are ideal for investigating the effects of systemic and localized bacteria on tumor behavior, immunological responses, and medication effectiveness.^{[ 38 ]}

This review highlights the development of 3D cancer models as effective tools for studying tumor–microbiome interactions, discusses advanced 3D fabrication strategies that integrate biomaterials with engineering principles, and presents key studies demonstrating how these models capture the complexity of host‐microbiome dynamics and advance cancer research and therapy. In this review, the term “microbiome” refers specifically to the bacterial microbiome associated with tumors, encompassing its genetic material, metabolites, and interactions with the TME. Other microbial domains, such as viruses, fungi, and archaea, are beyond the scope of this work.

2. 3D Cancer Model Fabrication Techniques

As aforementioned, the microbiome significantly influences cancer development and progression through its complex interactions with the TME. Microbiome populations can modulate inflammation, immunity, and metabolism, thereby affecting tumor growth.^{[ 45 ]} Traditional 2D culture systems lack the complexity to model these interactions effectively. In contrast, 3D models offer a more physiologically relevant platform, enabling the study of cell–cell and cell‐ECM interactions, tumor invasion, migration, and heterogeneity.^{[ 46} , ⁴⁷ , ^{48 ]} This section outlines the various approaches for creating 3D cancer models, including spheroids, organoids, microfluidic systems, and 3D‐bioprinted models.

2.1. Spheroid Models

Spheroids are self‐assembling 3D cell aggregates that closely mimic the complex conditions of living tissues, enabling the study of key aspects of tumor biology.^{[ 49 ]} Their geometrical configuration parallels that of native tissue cells, promoting enhanced tissue‐specific gene expression and influencing vital cellular functions such as proliferation and differentiation. Additionally, spheroids maintain a dense ECM, which serves as a protective barrier against external stimuli and helps preserve cell viability. However, the core of larger spheroids may suffer from nutrient and oxygen depletion, as well as accumulation of metabolic waste, which can induce necrosis.^{[ 50 ]} To mitigate these challenges, various strategies for spheroid fabrication have been employed, such as the use of biocompatible scaffolds (Figure  3 A). The most widely used and simple technique for spheroid generation is the hanging drop method, where cell‐laden droplets are suspended on an inverted surface, and gravity drives cell aggregation. Surface tension, hydrophobicity, and humidity control are key to maintaining stable droplets, while additives like methylcellulose can enhance spheroid formation and uniformity.^{[ 51 ]} A more advanced method using low‐adhesion plates enables reliable, size‐controlled spheroids by preventing cell attachment to the culture surface, allowing cells to freely aggregate.^{[ 52} , ^{53 ]} Another spheroid fabrication method is the spinner flask technique, which uses gentle rotational motion to promote uniform cell aggregation while minimizing shear stress. This system supports efficient nutrient and oxygen exchange, enabling long‐term culture and reproducible spheroid size. Its adaptability makes it valuable for applications like drug screening and tissue engineering.^{[ 52 ]} Beyond the mechanical forces governing spheroid formation, biomaterials play a critical role in supporting 3D cell aggregation, nutrient diffusion, long‐term viability, and simulated ECM properties, which will be further discussed in detail in the next Sections 2.4 of this review.

Figure 3.

Experimental platforms for modeling tumor–microbiome interactions in 3D models. A) Spheroids: Tumor cells and microbiome components can be co‐cultured using different techniques such as hanging drops, low‐attachment plates, or spinner flasks to generate tumor spheroids. B) Organoids: Tumor‐derived organoids cultured in gel matrices can be microinjected with bacteria into their lumen, enabling the establishment of complex 3D tumor–host‐microbiome organoids. C) Microfluidic systems: Tumor‐microbiome co‐cultures integrated into organ‐on‐chip devices allow perfusion through the side channels, with inlet/outlet reservoirs enabling the delivery of immunotherapies, nanoparticles, peripheral blood mononuclear cells (PBMCs), or metabolites under dynamic flow conditions. D) Bioprinted constructs: Tumor cells, stromal components, immune cells, and microbiome elements incorporated in bioinks can be processed using extrusion, droplet, laser‐assisted, or photolithography bioprinting. Crosslinking can yield spatially defined 3D tumor‐microbiome constructs.

Incorporating microbiome components into 3D spheroid models represents a transformative approach to examining the intricate interplay between TMEs and cancer cells. Recent biofabrication advancements have enabled the development of 3D models that incorporate both tumor cells and microbiome elements, enhancing the physiological relevance of in vitro studies.^{[ 54 ]} These models begin with scaffolds or matrices that support the growth of cancer cells alongside the microbiome, which can actively shape tumor behavior. Research indicates that a specific microbiome can alter tumor cell phenotypes and enhance the secretion of tumor‐promoting factors. For instance, co‐culturing intestinal bacteria with CRC cells has been shown to induce morphological changes and increase the production of molecules that influence ECM composition, fostering an environment conducive to tumor progression.^{[ 55 ]} To better analyze these interactions, recent advancements in 3D spheroid models have further enhanced our understanding of the complex relationship between the gut microbiome and CRC. These models incorporate cancer‐associated fibroblasts, endothelial cells, and immune cells, better replicating tumor heterogeneity and improving drug response evaluations.^{[ 56 ]} A key advantage of spheroids for bacterial incorporation is their natural oxygen gradient, from an oxygen‐rich periphery to an oxygen‐deficient core, which creates favorable conditions for anaerobic bacteria. For example, integrating CRC‐associated anaerobic bacteria, such as F. nucleatum, into spheroid models has provided critical insights into bacterial‐driven TME alterations and their potential influence on cancer progression.^{[ 57 ]} Comparative studies between 3D spheroid and organoid models further highlight their superiority over conventional 2D cultures, particularly in drug screening and personalized medicine applications.^{[ 58 ]} Collectively, these advancements emphasize the essential role of 3D spheroid models in bridging the gap between in vitro research and clinically relevant tumor behavior, thereby driving oncological research and therapeutic innovations.

However, studying the microbiome within spheroid models presents several challenges that complicate comprehensive analysis. Unlike traditional monolayer cultures, spheroids create heterogeneous oxygen and nutrient gradients, influencing microbiome colonization in unpredictable ways and making results difficult to standardize. Their lack of vascularization further limits their ability to replicate systemic microbiome effects, such as immune modulation and metabolite exchange. Additionally, their small size and limited lifespan constrain long‐term studies, while technical hurdles, such as extracting bacterial DNA or distinguishing bacterial populations within dense cellular structures, add another layer of complexity to microbiome investigations.^{[ 50 ]}

2.2. Organoid Models

Organoid models provide an advanced and physiologically relevant tool for investigating tumor‐microbiome interactions. Specifically, patient‐derived organoids are established from patient tumor biopsies, conserving genetic mutations, and cellular heterogeneity of the original tumor tissue.^{[ 59} , ⁶⁰ , ^{61 ]} This makes organoid models extremely relevant for developing personalized cancer treatment, including studying how individual microbiomes influence tumor behavior and therapeutic response.^{[ 62} , ⁶³ , ^{64 ]}

Fabricating organoids is centered around using compatible and supportive biomaterials that mimic the ECM found in the tumor. Matrigel, an animal‐based ECM hydrogel, has been the gold standard as it is rich in growth factors and adhesion molecules that make it attractive for organoid formation.^{[ 62 ]} There are some limitations when dealing with Matrigel, such as batch‐to‐batch variability and potential immunogenicity, which can limit its reproducibility and translational potential.^{[ 60} , ^{65 ]} As such, the development and use of synthetic matrices has become more popular, like polyethylene glycol (PEG)‐based hydrogels, which can be tailored to fit the mechanical properties of interest, as well as be functionalized with bioactive peptides, and have high reproducibility.^{[ 65} , ^{66 ]} Recent innovations also include the use of nanocellulose‐based hydrogels, which are highly hydrated, modifiable, and represent a sustainable alternative for organoid culture, supporting both mouse and patient‐derived tumor organoids.^{[ 63 ]}

To study microbiome interactions with the TME, there are several co‐culture techniques that have been established. One method involves injecting bacteria into the organoid lumen, which allows controlled and physiologically relevant exposure of the epithelium to bacteria (Figure 3B).^{[ 67 ]}

The integration of bio‐fabrication methods, such as microfluidics and 3D‐bioprinting, is opening new avenues for engineering organoid‐based tumor models with unprecedented spatial and cellular complexity. For example, extrusion‐based bioprinting has been used to assemble multicellular tumor models, demonstrating enhanced tumor‐stromal interactions and drug resistance compared to traditional cultures.^{[ 39} , ⁶⁸ , ^{69 ]} The spontaneous formation of tumor spheroids within 3D‐bioprinted hydrogels further enhances the physiological relevance of these models, supporting the study of cell migration, matrix remodeling, and bacteria‐induced tumorigenesis.^{[ 70 ]}

Despite these advances, significant limitations remain. Most organoid models still lack all the TME components, such as immune and stromal cells, which are critical for capturing the intricate interplay between tumors, microbiome, and host immunity. Efforts are underway to address these gaps by incorporating additional cell types into synthetic matrices and leveraging microfluidic systems to enable long‐term, stable co‐cultures with precise control over environmental parameters.^{[ 71} , ⁷² , ^{73 ]} Another ongoing challenge is the control of organoid function and maturation after assembly, which is being tackled through the development of compositional hydrogels with tunable mechanical and biochemical properties. Specifically, co‐culturing gastrointestinal organoid models with anaerobic bacteria remains a significant challenge due to the opposing oxygen requirements of host epithelial cells and obligate anaerobes.^{[ 74 ]} The need for a hypoxic or anoxic environment to support bacterial viability often compromises epithelial cell health, making long‐term co‐culture difficult. To address this, researchers have developed dual‐compartment systems, such as diffusion chambers, that maintain separate yet interactive microenvironments.^{[ 75 ]} For example, a two‐chamber transwell setup was established, in which anaerobic bacteria (e.g., Bifidobacterium adolescentis (B. adolescentis), Bacteroides fragilis (B. fragilis)) were cultured in an anaerobic upper chamber, while human colon epithelial cells were maintained in an oxygenated lower chamber.^{[ 76 ]} This configuration allows biochemical communication between the two compartments while preserving the viability and function of both microbial and host components.

In short, organoid models, especially when combined with innovative biomaterials and biofabrication techniques, represent a powerful and versatile platform for dissecting tumor‐microbiome interactions. These advances are enabling more physiologically relevant, patient‐specific studies that can inform precision medicine approaches and accelerate the development of new cancer therapies.

2.3. Microfluidic Systems and Organ‐on‐a‐Chip Platforms

Microfluidic systems, often referred to as “Organ‐on‐a‐Chip” (OoC) systems, are an emerging 3D cell culture technology that replicates physiological characteristics of tissues and organs, including mechanical forces (e.g., shear stress), biochemical signals, and gradients of pH, nutrients, and oxygen.^{[ 77} , ⁷⁸ , ⁷⁹ , ⁸⁰ , ^{81 ]} OoC models can be tailored to increase biological complexity by incorporating vascularization and multi‐organ models in combination with spheroids and organoids.^{[ 82} , ⁸³ , ⁸⁴ , ^{85 ]} OoC design enables the connection of multiple OoC units in sequential order to simulate the transition of biochemical signals and circulating cells (such as immune cells or circulating cancer cells).^{[ 86} , ⁸⁷ , ⁸⁸ , ^{89 ]}

The design of microfluidic OoCs addresses the biological elements, such as relevant cell types, ECM or tissue‐mimicking hydrogel, and physiological parameters (e.g., mechanical forces, flow, shear stress).^{[ 80} , ⁸¹ , ^{89 ]} OoC platforms typically include two or more compartments,^{[ 90} , ⁹¹ , ^{92 ]} filled with cell‐laden hydrogels and perfusable media or an air compartment (channels). Compartmentalization is achieved using membranes or hydrodynamic features like micropillar arrays, narrow gaps, or rail structures. (Figure  3C).^{[ 90} , ⁹³ , ^{94 ]} The hydrogels used to embed cells are critical to phenotype and function. Collagen (notably type I), fibrin, Matrigel, and synthetic matrices like PEG or polyacrylates allow the tuning of mechanical and biochemical environments.^{[ 95} , ⁹⁶ , ⁹⁷ , ^{98 ]} ECM components like laminin, entactin, integrins, and collagen IV enhance tissue mimicry.^{[ 99} , ^{100 ]} Hydrogels also serve both structural and biochemical roles, enabling the formation of gradients and providing vascularization cues.^{[ 94} , ¹⁰¹ , ¹⁰² , ^{103 ]} Controlling the mechanical and chemical properties of the gels in OoC systems allows for the better recapitulation of the TME, making such models suitable for mimicking healthy and tumorous tissues.^{[ 35 ]}

Of particular interest is the use of OoCs in modeling tumor–microbiome interactions,^{[ 78} , ⁹² , ^{104 ]} especially by connecting gut‐on‐a‐chip systems^{[ 82} , ¹⁰⁵ , ¹⁰⁶ , ¹⁰⁷ , ^{108 ]} containing microbiome^{[ 35} , ^{107 ]} to tumor‐on‐a‐chip organs, such as lung,^{[ 109 ]} kidney, liver,^{[ 84} , ¹¹⁰ , ^{111 ]} brain,^{[ 112} , ^{113 ]} pancreas,^{[ 114 ]} breast,^{[ 115 ]} or bone.^{[ 116 ]} Microfluidic systems offer significant advantages for cancer modeling by replicating cell–cell interactions and organ‐to‐organ communication, both of which are essential for studying tumor progression and immune responses.^{[ 82} , ^{107 ]} One of the key advantages of OoCs is their ability to monitor tumor‐microbiome interactions in real time under dynamic conditions, including metabolic changes, immune response, and the microbiome's effect on tumor behavior. The integration of sensors, transparent fabrication materials compatible with imaging tools, enables the tracking of dynamic processes such as immune cell infiltration, tumor growth, or metabolic flux.^{[ 117} , ¹¹⁸ , ^{119 ]} Platforms incorporating transepithelial/transendothelial electrical resistance (TEER) measurements or electrochemical sensors support barrier function analysis in tissues such as the gut or brain.^{[ 82} , ^{120 ]} Additionally, OoCs scalability and precision make them suitable for high‐throughput studies.^{[ 83} , ¹²¹ , ^{122 ]} Despite their potential, challenges remain in replicating native 3D tissue architecture, standardizing fabrication, and integrating with clinical workflows.^{[ 123} , ¹²⁴ , ¹²⁵ , ¹²⁶ , ¹²⁷ , ^{128 ]} One of these challenges in the co‐culturing of human cells with anaerobic bacteria is due to their conflicting oxygen requirements. Microfluidic systems are particularly well‐suited for overcoming this barrier, as they enable precise control of oxygen gradients and compartmentalized environments.^{[ 133 ]} Recent efforts have focused on perfusion‐based designs that maintain separate yet interactive conditions for host cells and microbes. One notable example is the HumiX system, a microfluidic platform that supports cross‐talk between human epithelial cells and Bacteroides caccae under simulated gastrointestinal conditions.^{[ 31 ]} Another model, the mucosal anoxic‐oxic interface chip, which recreates the oxygen gradient across colonic epithelium, mimicking the in vivo gut environment.^{[ 129 ]} Additionally, a gut microbiome‐on‐a‐chip system was developed to support anaerobic B. fragilis co‐culture with epithelial layers, further demonstrating the potential of these platforms for host‐microbe interaction studies.^{[ 130 ]}

Ongoing innovations, including the development of automation,^{[ 86} , ^{131 ]} AI‐driven analysis,^{[ 116} , ^{132 ]} and patient‐derived cell integration, are moving these platforms toward high‐throughput, precision medicine applications.^{[ 133} , ¹³⁴ , ^{135 ]}

2.4. 3D‐Bioiprinted Models

3D‐bioprinting is a powerful tool that allows the mimicry of live tissues with micrometer‐scale precision. 3D‐bioprinting adopts principles of traditional 3D‐printing, with integration of living cells, biomaterials, and growth factors to recreate the complexity of native biological structures. To date, three principal bioprinting modalities have been established: extrusion‐based, droplet‐based, and laser‐assisted bioprinting, each suited for different bioinks/materials and models with altered structural requirements.^{[ 136} , ¹³⁷ , ¹³⁸ , ¹³⁹ , ^{140 ]}

2.4.1. 3D‐Bioprinting Techniques

Extrusion‐Based Bioprinting: Versatility and Structural Complexity

Extrusion‐based bioprinting is currently the most prevalent and versatile approach to 3D bioprinting. This technique extrudes bioinks through a nozzle using pneumatic pressure, mechanical pistons, or screw‐driven systems. It accommodates highly viscous materials and cell‐dense hydrogels, enabling the fabrication of stable, volumetric 3D constructs.^{[ 136} , ¹³⁸ , ¹³⁹ , ^{141 ]} However, resolution is relatively limited (typically 200–500 µm), and the extrusion process imposes mechanical shear stress that may impair cell viability during deposition.^{[ 140} , ^{142 ]} To overcome these limitations and expand its structural and functional versatility, several sub‐modalities have been developed. One such approach is FRESH (Freeform Reversible Embedding of Suspended Hydrogels), which utilizes a thermo‐reversible support bath composed of various materials, such as gelatin, Pluronic^® F‐127, gellan gum, κ‐carrageenan, and nanosilica particles. The support bath provides mechanical support to the extruded bioink during deposition, while permitting smooth nozzle movement through the medium. Such support enables the precise fabrication of complex and delicate geometries that would otherwise collapse under gravity if printed in air. The bioinks extruded into the bath typically include soft hydrogels such as collagen, alginate, or decellularized extracellular matrix (dECM). Following printing, the support bath is removed, typically by heating, cooling, or chemical means, leaving the printed construct intact and undamaged.^{[ 136} , ^{143 ]}

Another advancement in 3D‐bioprinting is microfluidic extrusion bioprinting, which integrates microchannel networks into the printhead to precisely control the composition, flow, and deposition pattern of multiple bioinks in real time. This allows spatially defined co‐printing of various cell types or matrix components in complex, heterogeneous tissue constructs.^{[ 138 ]}

In coaxial bioprinting, concentric nozzles are used to extrude core‐shell structures, which are particularly suited for creating perfusable tubular constructs that mimic vascular geometries. This configuration also enables in situ crosslinking, by co‐extruding alginate with calcium chloride, allowing for immediate stabilization of the printed shell layer.^{[ 70 ]}

Despite its many advantages, extrusion‐based bioprinting is not without challenges. Shear stress during extrusion may reduce cell viability, and the method is prone to nozzle clogging and relatively slow print speeds. Fabricating delicate or free‐standing features remains difficult without additional support strategies.^{[ 136} , ¹³⁸ , ^{141 ]} Nevertheless, its compatibility with a wide range of biomaterials, capacity for high cell density, and modular integration with multi‐material systems establish extrusion as the foundational platform in modern bioprinting. Ongoing innovations in bioink formulation, nozzle design, and embedded support systems continue to expand their potential, particularly for the development of 3D cancer‐microbiome interaction models.^{[ 136} , ¹³⁷ , ^{138 ]}

Droplet‐Based Bioprinting: High‐Throughput Precision at the Microscale

Droplet‐based bioprinting involves the controlled ejection of picoliter‐sized droplets of bioink using thermal, piezoelectric, or electrostatic actuation‐ most notably in inkjet and electrohydrodynamic jetting (EHDJ) systems.^{[ 140 ]} In these approaches, each droplet is deposited individually, enabling very high spatial resolution, precise placement of single cells, and remarkably fast printing speeds (Figure 3D). These features make droplet‐based techniques particularly attractive for applications requiring high‐throughput arrays and fine‐scale cellular patterning.^{[ 137} , ¹³⁸ , ^{140 ]} In inkjet bioprinting, droplets are expelled in response to short thermal or mechanical impulses. Typically, the bioink is briefly heated (to 200–300 °C for µs), forming a vapor bubble that generates sufficient pressure to eject a droplet. Inkjet systems are relatively inexpensive and rapid and are well‐suited for single‐cell printing. However, the transient thermal exposure may negatively impact cell viability, particularly for heat‐sensitive cell types.^{[ 137} , ¹⁴⁰ , ^{144 ]} EHDJ, by contrast, employs a strong electric field to drive the formation of ultra‐fine droplets. It offers greater control over droplet size, ejection frequency, and spatial accuracy, often exceeding that of inkjet systems. Nonetheless, it requires more advanced equipment and is sensitive to environmental fluctuations.^{[ 137} , ^{138 ]}

Droplet‐based printing is best suited for low‐viscosity bioinks (1–10 mPa·s) and is less compatible with viscous or highly cell‐laden formulations. Due to the lack of mechanical support during printing, the method is generally unsuitable for building thick, load‐bearing 3D structures, limiting its use to relatively simple or 2D models.^{[ 144 ]} Moreover, concerns remain regarding potential thermal, electrical, or mechanical stressors that may adversely affect sensitive cell types.^{[ 138} , ^{140 ]}

Nevertheless, droplet‐based bioprinting offers unique advantages for studies involving complex biological systems, such as cancer and the microbiome. It facilitates the controlled deposition of distinct cell populations or signaling molecules (e.g., bacteria, cytokines, or immune cells) in defined spatial arrangements.

Laser‐Assisted Bioprinting: Single‐Cell Precision and No‐Contact Deposition

Laser‐assisted bioprinting (LAB) is one of the most advanced methods, with high spatial resolution and near single‐cell precision (Figure 3D). The leading technique in this category is laser‐induced forward transfer (LIFT), where a focused laser pulse is directed onto an energy‐absorbing donor layer, typically positioned atop a transparent substrate. The laser pulse generates localized pressure that propels a microscale volume of bioink toward a receiving substrate‐without any mechanical contact between components.^{[ 138} , ^{140 ]}

This contactless configuration offers several key advantages. First, it eliminates the need for nozzles, thereby avoiding clogging issues that commonly affect extrusion‐based systems. Second, it minimizes shear stress during deposition, preserving the viability of delicate cell types. Additionally, LAB accommodates a wide range of bioink viscosities, from dilute solutions to dense hydrogels, expanding its material compatibility. Finally, its high spatial precision allows for controlled placement of individual cells or microparticles, making it particularly valuable for studying cell‐cell and cell‐bacteria interactions at micrometer resolution.^{[ 136} , ¹³⁷ , ^{138 ]}

Despite these advantages, LAB systems also pose significant technical and biological challenges. The equipment is costly and complex, requiring meticulous calibration and environmental control. Additionally, the long‐term effects of repeated laser exposure on cell viability remain insufficiently understood. Another limitation is the relatively low throughput, which restricts the method's applicability to small‐scale or highly localized constructs, rather than the fabrication of large volumetric tissues.^{[ 138} , ^{140 ]}

In the context of cancer and microbiome research, laser‐assisted bioprinting offers powerful capabilities for engineering precise biological microenvironments. For example, it enables high‐resolution placement of tumor cells alongside stromal or immune cells, or the spatially defined incorporation of bacteria within host‐derived matrices. Although it is less suited for fabricating entire tissue constructs, LAB provides a valuable platform for high‐fidelity in vitro models and mechanistic studies conducted under tightly controlled conditions.^{[ 138} , ^{140 ]}

Photolithography‐Based Bioprinting: High‐Resolution Light‐Driven Structuring

Among the existing bioprinting technologies, photolithography‐based methods, namely, stereolithography (SLA) and digital light processing (DLP), offer a combination of micrometer‐scale resolution, rapid fabrication, and compatibility with complex structures. DLP operates by projecting patterned UV/Visible light onto photocross‐linkable bioinks (e.g., gelatin methacrylate (GelMA) or PEG diacrylate (PEGDA)) in a layer‐by‐layer manner (Figure 3D). This process polymerizes the hydrogel selectively, yielding geometrically intricate, cell‐laden constructs with high fidelity.^{[ 136} , ¹⁴¹ , ^{145 ]} The choice and concentration of photoinitiators play a key role in balancing crosslinking efficiency, cytocompatibility, and resolution. Recent advances in light modulation and grayscale patterning have enabled the fabrication of graded or compartmentalized tissue models, including vascularized constructs, tumor‐on‐a‐chip platforms, and cardiac or hepatic microtissues.^{[ 136} , ¹⁴¹ , ¹⁴⁵ , ^{146 ]}

Nevertheless, DLP bioprinting presents several technical limitations. The method requires bioinks that are both optically clear and photoreactive, limiting material options. Additionally, DLP often involves high light intensities that may damage cells or cause uncontrolled polymerization. Moreover, the optical setup in DLP systems is often complex and costly, requiring careful calibration for consistent performance.^{[ 136} , ¹⁴¹ , ¹⁴⁶ , ¹⁴⁷ , ^{148 ]} Despite these challenges, the ability of DLP to generate biologically functional, high‐resolution, and geometrically intricate constructs, at relatively high speed, makes it one of the most promising modalities.

2.4.2. ECM‐Like Hydrogels

At the core of 3D‐bioprinting lies the bioink formulation of biomaterials mimicking the ECM, which enables the spatial deposition of living cells into structurally and functionally relevant architectures. Bioink must provide printability, mechanical support, support cell viability, lineage‐specific differentiation, and maintain long‐term tissue function. These requirements are particularly important when modeling systems such as tumors or host–microbiome interfaces, where cellular and microbiome components coexist under tightly regulated biochemical and biophysical gradients.^{[ 136} , ¹³⁸ , ^{141 ]} Bioinks used in these advanced applications must be highly tunable in terms of viscosity, crosslinking kinetics, oxygen permeability, and bioactivity, to accommodate the disparate demands of cells and microorganisms. Aerobic‐compatible formulations typically exhibit high water content, porosity, and oxygen diffusivity, making them suitable for eukaryotic cell cultures such as collagen or GelMA‐alginate blends. Conversely, anaerobic‐tolerant bioinks tend to be denser, less oxygen‐permeable, such as dense fibrin–collagen or dextran–based hydrogels, and tailored to support bacterial viability under hypoxic or anoxic conditions.^{[ 149} , ¹⁵⁰ , ¹⁵¹ , ^{152 ]} A growing class of bioinks demonstrates dual compatibility, enabling the co‐culture of microbiome and host cells within the same construct. This is achieved through precise modulation of physicochemical parameters such as temperature, polymer concentration, degree of crosslinking, and the inclusion of biologically active moieties (e.g., ECM proteins). Examples include methacrylated hyaluronic acid (HAMA), bacterial cellulose‐based hydrogels for dynamic constructs, and composite matrices mimicking tumor‐microbiome niches. These systems enable the formation of gradients and spatial compartmentalization‐ key features of the tumor‐microbiome axis.^{[ 141} , ¹⁵¹ , ^{153 ]}

Among the most promising bioinks for tissue engineering and regenerative medicine applications are dECM‐based hydrogels. Derived from native tissues stripped of their cellular components, dECM retains an organ‐specific repertoire of collagens, glycoproteins, proteoglycans, and growth factors, offering unparalleled biomimicry of the natural ECM niche. While inherently low in viscosity and mechanically fragile, dECM bioinks can be effectively combined with structural polymers to yield hybrid matrices that recapitulate both the biological identity and the architectural fidelity required for functional 3D constructs.^{[ 136} , ¹³⁷ , ¹⁴¹ , ^{154 ]} Recent studies have demonstrated the use of intestinal and colonic dECM in co‐cultures with anaerobic microbiome, such as Lactobacillus and F. nucleatum, providing physiologically relevant models of the gut and tumor‐microbiome interface. These bioinks, often hybridized with alginate or GelMA, support bacterial viability under hypoxic conditions while preserving mammalian cell functionality and spatial organization.

The most widely used materials that are used in mimicking tumor and microbiome studies are: alginate, gelatine, GelMA, cellulose, hyaluronic acid (HA), collagen, fibrin, pectin, PEGDA, polycaprolactone (PCL), and Pluronic^® F‐127. Their material properties, biological compatibility, and 3D‐bioprinting applications are summarized in Table  2 .

Table 2.

Bio‐inks for 3D‐bioprinting.

Bioink	Key Advantages	Limitations	Compatible Printing Techniques	Representative Applications	References
Alginate	Rapid ionic gelation, excellent shape fidelity, biocompatibility, low immunogenicity	Bioinert; lacks cell‐adhesive motifs; often blended with ECM components	Extrusion, Photolithography (modified alginate)	Structural frameworks, co‐printing with bioactive hydrogels	[136, 141, 146, 155, 156, 157, 158]
Gelatin	Integrin‐binding sites support cell adhesion; reversible thermal gelation	Poor mechanical stability; temperature‐sensitive; needs crosslinking or blending	Extrusion (temperature‐controlled)	Soft tissue models, cell‐laden scaffolds	[136, 141, 146, 155, 156]
GelMA	Photocrosslinkable; high‐resolution patterning; cell‐adhesive	Requires photoinitiators and light control; potential phototoxicity	Extrusion, Photolithography, Laser‐assisted	Vascular, dermal, and neural constructs	[136, 141, 146]
Cellulose & Derivatives	Enhances viscosity, mechanical support; tunable rheology	Biologically inert; requires blending with bioactive materials	Extrusion	Shape‐stable scaffolds, composite bioinks	[137, 138, 157]
HA	Excellent biocompatibility; supports cellular signaling; hydrophilic	Low mechanical strength; requires chemical modification (e.g., HAMA)	Extrusion, Photolithography	Cartilage, dermal, and neural tissue engineering	[136, 141, 146, 159]
Collagen	Bioactive; supports cell adhesion, proliferation, and differentiation	Low viscosity and mechanical strength; sensitive to pH and temperature; poor print fidelity alone	Extrusion (e.g., FRESH), Laser, Light‐based	Skin, vascular models, soft tissue mimicry	[136, 138, 141, 146, 160]
Fibrin	Biocompatible, bioactive, supports cell adhesion and migration, enzymatic gelation under mild conditions	Low mechanical strength, rapid degradation, limited shape fidelity	Extrusion, Photolithography	Soft tissue modeling, co‐culture with bacteria, wound‐like or inflamed tissue constructs	[141, 149]
Pectin	Mild gelation with calcium; natural and biocompatible	Lacks cell‐binding motifs; mechanically weak; limited to soft, non‐load‐bearing applications	Extrusion	Co‐bioinks for soft tissues, additive to blends	[137, 161]
PEGDA	Highly tunable; photopatternable; consistent batch quality; excellent optical clarity	Bioinert; requires functionalization (e.g., RGD peptides) to support cell adhesion	Photolithography, Inkjet/Droplet‐based	High‐resolution tissue models, vascular microchannels	[136, 141, 146, 148, 162]
PCL	Excellent mechanical strength; slow degradation; suited for load‐bearing constructs	Not cell‐compatible in native form; printed at high temperature; no direct cell encapsulation	Melt Extrusion	Bone, muscle, orthopedic scaffolds	[136, 141, 146, 163]
Pluronic® F127	Thermoresponsive behavior; reversible sol–gel transition; supports high print fidelity and removal	Bioinert; not suitable for long‐term cell or microbiome encapsulation	Extrusion‐based	Sacrificial layers; perfusable channels; spatial patterning of microbiome niches	[141, 151, 164]

2.4.3. Bioinks for Host‐Microbiome Co‐Culture

The co‐culture of mammalian and bacterial cells within engineered 3D environments presents a unique and growing frontier in the bioprinted models. These systems aim to recapitulate complex host‐microbiome interactions as they occur in native tissues such as the gastrointestinal tract, skin, or TME.^{[ 141} , ¹⁴⁹ , ¹⁵⁰ , ¹⁵¹ , ^{152 ]} However, mammalian and bacterial cells differ substantially in their physiological requirements, including oxygen tension, pH tolerance, nutrient demands, and susceptibility to mechanical stress, posing substantial challenges for integrated culture within a single construct.^{[ 136} , ¹³⁸ , ^{141 ]}

To address these constraints, bio‐inks must be carefully designed to support both cell types simultaneously and sustainably. This includes tuning of rheological properties, cross‐linking mechanisms, diffusivity, and biochemical compatibility. Bioinks must also maintain spatial organization, enable distinct microenvironments, and permit metabolic exchange without disrupting structural integrity.^{[ 136} , ¹⁵³ , ^{157 ]} This section reviews 3D‐bioprinted models of host cells and bacteria co‐cultures, which are summarized in Table  3 .

Table 3.

Bioinks and Formulations for Host‐Microbiome Co‐Culture.

Formulation	Printing Method	Printed Model	Tissue Source / Cell Type	Microbiome Type	Co‐culture Duration	Reference
Alginate‐Gelatin	Extrusion	Gradient oxygen models	Colon epithelial + stromal cells	Anaerobic gut bacteria	3–7 days	[160, 166]
Alginate‐Fibrin	Extrusion	Infection models	Colon epithelial cells	Pathogenic bacteria (e.g., E. coli)	< 3 days	[141, 166]
Alginate‐Methylcellulose (AlgMC)	Extrusion (Coaxial)	Coaxial bacterial encapsulation	Intestinal epithelial + bacterial cores	Anaerobic bacteria (F. nucleatum, B. fragilis)	3–7 days	[156, 158, 161, 162]
GelMA‐ HA	Photopolymerization	Tumor/inflammation	Colorectal tumor cells + immune cells	Tumor‐associated bacteria (F. nucleatum)	7–14 days	[136, 141, 167]
PEGDA + dECM or RGD peptides	Photopolymerization	Spatial compartmentalization	Hepatocytes + fibroblasts	Variable; dependent on ink composition	> 14 days	[141, 151, 163, 168]
Pluronic® F‐127 (Sacrificial)	Extrusion (sacrificial)	Perfusable voids/microchannels	Nutrient/metabolite transport support	Facilitates microbiome niche formation	Not for long‐term culture (sacrificial)	[141, 167]
GelMA+ embedded hollow channels	Extrusion/Embedded printing	Metabolic modeling	Host cells with metabolite turnover needs	Dynamic interaction model	7–14 days	[167, 170]
Alginate‐Collagen	Extrusion	Gut/tissue modeling	Epithelial cells, fibroblasts	Anaerobic bacteria	3–7 days	[141, 158]
Gelatin‐Collagen	Extrusion	Inflammation and stromal models	Mammalian stromal and immune cells	Supportive for facultative bacteria	1–5 days	[141, 171]
Alginate (alone)	Extrusion	Microbiome encapsulation	Bacteria or epithelial cells	Anaerobic bacteria	< 3 days	[157, 160]
Nanocellulose‐Alginate	Extrusion	Mechanically stable host‐bacteria scaffolds	Skin epithelial + dermal fibroblasts	Facultative bacteria	7–14 days	[162, 167, 168, 172]
GelMA‐Nanoclay	Extrusion / light‐assisted	Tumor–microbiome co‐culture	Tumor and stromal cells	Relevant for tumor‐associated bacteria	3–7 days	[167]
HA	Extrusion / Photopolymerization	Inflammatory gut models	Colon epithelial cells	Anaerobic gut microbiome	1–5 days	[136, 141, 155]
Fibrin	Extrusion	Wound/infection models	Dermal fibroblasts, immune cells	Skin pathogens (e.g., S. aureus)	< 3 days	[141, 166]
GelMA	Photopolymerization	Skin/tumor‐microbiome interaction	Skin keratinocytes, immune cells	Commensal skin bacteria (e.g., Staphylococcus spp.)	3–7 days	[136, 141, 167]
PEGDA	Photopolymerization	Perfused liver/gut constructs	Hepatocytes, fibroblasts	Controlled, limited viability	7–14 days	[136, 168]
dECM Hydrogel	Extrusion / Photopolymerization	Tissue‐specific host‐bacteria models	Colon organoids, stromal cells	Anaerobic bacteria (B. fragilis)	3–7 days	[141, 165, 168]
PEGDA+ Microparticles	Photopolymerization	Tumor models with nutrient modulation	Colon cancer cells	F. nucleatum, E. coli	> 14 days	[141, 151, 170]

Natural hydrogels, including alginate, gelatin, fibrin, and HA, are widely used due to their biocompatibility and mild gelation conditions.^{[ 136} , ^{141 ]} Alginate offers fast ionic cross‐linking and oxygen permeability, making it suitable for bacterial encapsulation, but it lacks cell‐adhesive motifs. Its oxygen permeability has also been successfully leveraged in co‐culture systems with anaerobic probiotics in the colonic epithelium.^{[ 156 ]} Gelatin and GelMA both support mammalian cell adhesion and remodeling, however, their interaction with bacteria can be less optimal due to their relatively high degradation rates, which may interfere with long‐term bacterial viability.^{[ 136} , ¹⁴¹ , ¹⁵¹ , ^{153 ]} Fibrin, formed enzymatically from fibrinogen and thrombin, is highly bioactive but mechanically weak and rapidly degraded.^{[ 141} , ^{149 ]} HA provides viscoelasticity and cellular signaling but often requires methacrylation (e.g., HAMA) for printability.^{[ 136} , ^{153 ]} Modified HA has been increasingly utilized in bacterial co‐cultures to facilitate immune signaling and host–bacteria interactions in systems designed to mimic inflammatory or tumor‐associated environments.^{[ 136} , ^{153 ]}

Synthetic materials such as PEGDA and PCL contribute mechanical strength and architectural fidelity.^{[ 136} , ¹⁴¹ , ¹⁵¹ , ^{153 ]} PEGDA is photopolymerizable and tunable but bioinert without functionalization.^{[ 136} , ^{153 ]} Incorporating dECM into PEGDA hydrogels has been shown to enhance bioactivity while preserving mechanical fidelity, making them particularly suitable for compartmentalized co‐culture systems.^{[ 165 ]} PCL, used in melt extrusion, provides long‐term support in hybrid constructs. Thermoresponsive Pluronic^® F‐127 is commonly used as a sacrificial ink, enabling the formation of flow‐permissive voids or microbiome niches by being selectively removed post‐printing under mild thermal conditions.^{[ 141} , ^{151 ]}

To simultaneously support the divergent physiological needs of host and microbiome, bioink formulations are engineered to integrate complementary natural and synthetic components, each contributing distinct biochemical and mechanical functions.^{[ 141} , ¹⁴⁹ , ¹⁵⁰ , ¹⁵¹ , ^{152 ]} Natural and synthetic components with complementary properties. These combinations are designed to balance oxygen diffusion, cell adhesion, mechanical stability, and printability, while creating spatially distinct but metabolically interactive compartments.

For example, alginate‐gelatin blends are frequently used in co‐culture constructs. Alginate provides structural support and is highly permeable to oxygen, making it ideal for encapsulating bacteria. In contrast, gelatin supports mammalian cell adhesion through integrin‐binding motifs, creating a hospitable environment for epithelial or stromal cells. The resulting matrix supports both populations while allowing the formation of oxygen gradients essential to many host‐bacteria interactions.^{[ 141} , ^{149 ]} An example of this combination is in colon epithelial models where anaerobic gut bacteria (F. nucleatum) were successfully encapsulated within the alginate‐gelatin blend, creating a microenvironment for studying gut microbiome‐host interactions. The model was printed using extrusion‐based 3D printing and maintained bacterial viability for 3–7 days.^{[ 160} , ^{166 ]}

Another widely used combination is alginate‐fibrin, which supports rapid cross‐linking while combining biological activity with mechanical support. Fibrin enhances initial cell attachment and promotes tissue‐like remodeling, while alginate contributes to structural stability and slows degradation. This formulation is particularly useful for short‐term host‐microbiome interactions or infection models.^{[ 141} , ^{149 ]} In a study of infection models, alginate‐fibrin formulations were used to support colon epithelial cells co‐cultured with pathogenic bacteria such as Escherichia coli (E. coli) for less than 3 days. This 3D‐bioprinted model was printed using an extrusion‐based method, and the bacteria remained viable throughout the short‐term co‐culture period.^{[ 141} , ^{166 ]}

Alginate‐methylcellulose is also effective in extrusion‐based approaches, offering improved printability and shape fidelity. While methylcellulose contributes to viscosity and stability, the alginate component supports bacterial survival in the core, particularly in coaxial designs.^{[ 149} , ^{150 ]} In coaxial designs using F. nucleatum and B. fragilis bacteria, the formulation supported anaerobic bacterial survival (3–7 days) while allowing for better printability in the 3D‐bioprinting process.^{[ 156} , ¹⁵⁸ , ¹⁶¹ , ^{162 ]}

GelMA‐HA systems allow light‐based crosslinking and enable tuning of stiffness and degradation profiles. HA contributes to the viscoelastic properties of the matrix and can support immune‐related signaling pathways, making this combination relevant for inflammatory co‐culture models or tumor‐microbiome systems.^{[ 136} , ¹⁴¹ , ^{151 ]} This combination also facilitates the fine‐tuning of matrix stiffness and degradation rates, while the presence of HA enhances viscoelasticity and supports immune‐related signaling features, particularly valuable in inflammatory co‐cultures or tumor–microbiome models. An example of its use is in tumor‐associated bacteria models, where F. nucleatum was introduced into TME to study the interaction between tumor cells and gut microbiome. The model was created using photopolymerization and was maintained for 7–14 days.^{[ 136} , ¹⁴¹ , ^{167 ]}

Advanced co‐culture systems have also utilized coaxial core–shell printing strategies to create spatially partitioned but interactive compartments, enabling hepatocyte–fibroblast co‐culture, offering promise for host–microbiome separation within perfusable constructs.^{[ 149 ]} An example of this is a co‐culture system where hepatocytes were 3D‐bioprinted alongside fibroblasts in a perfusable construct, creating distinct but interacting compartments to simulate liver‐tumor interactions. The model was 3D‐bioprinted using coaxial extrusion and was maintained for longer‐term culture (>14 days).^{[ 141} , ¹⁵¹ , ¹⁶³ , ^{168 ]}

These combinatory strategies enable the fabrication of constructs that are able to maintain stability, long‐term viability, biological activity, and sustain spatially controlled, metabolically interactive host‐microbiome ecosystems.^{[ 141} , ¹⁵⁰ , ^{151 ]}

In 3D cancer models, this approach is used to study the effects of microbiome metabolites on tumor growth and immune response. Precise nutrient availability and distribution are critical for maintaining functional co‐cultures in 3D constructs, particularly in systems where host and microbiome exhibit distinct metabolic demands and consumption rates. It has been shown that oxygen concentration and bacterial strain identity influence the expression of genes involved in nutrient uptake, such as amino acid and sugar transporters.^{[ 152 ]} To overcome these challenges, engineering strategies include embedding microchannels, using sacrificial inks (e.g., Pluronic^® F‐127), and tuning hydrogel mesh size via cross‐linking density. These approaches improve passive diffusion and can direct nutrient delivery to specific compartments. For example, soft GelMA matrices with embedded hollow channels help maintain metabolite turnover and avoid localized nutrient depletion.^{[ 152 ]}

Additional solutions include integrating glucose reservoirs or degradable microparticles for controlled release. In tumor‐like constructs, such approaches are particularly relevant given the presence of hypoxic, acidic, and nutrient‐poor regions that affect both microbiome and host cell viability.^{[ 141 ]}

Given the dynamic nature of host‐microbiome interactions in disease contexts such as cancer, static biomaterials are insufficient.^{[ 141} , ^{151 ]} Bioinks must respond to environmental cues, including microbiome metabolites, pH shifts, or inflammatory mediators, to capture the evolving behavior of living systems.^{[ 141} , ¹⁴⁹ , ^{169 ]} This convergence of material science and biology sets the stage for 4D‐bioprinting, which is addressed in the following section.

2.5. 4D‐Bioprinting for Dynamic Tumor Modeling

While conventional 3D‐bioprinted tumor models have advanced cancer research, their static nature limits their ability to replicate the dynamic processes of tumor evolution, such as hypoxia cycling, matrix remodeling, inflammation, microbiome‐induced shifts, and treatment resistance.^{[ 169 ]} Given that cancer is a temporal disease shaped by changes over time, modeling systems must be equally dynamic.^{[ 173 ]}

4D‐bioprinting addresses this need by integrating stimuli‐responsive materials that alter their structure or function. These adaptive behaviors reflect how the TME, including the microbiome, continuously evolves and influences cancer progression. Recent innovations highlight 4D‐bioprinting's potential in capturing these dynamics. Shape‐shifting hydrogels have been developed to release therapeutics under tumor‐like or microbiome conditions,^{[ 149 ]} while bioprinted smart bacterial systems can detect changes such as acidity or microbiome signaling and respond with targeted drug delivery.^{[ 169 ]}

Moreover, microbiome‐host cross‐talk can be modeled through perfusable 4D architectures, such as coaxial hydrogel systems that support live anaerobic microbiome adjacent to epithelial layers.^{[ 151} , ^{170 ]} These platforms enable real‐time diffusion of microbiome metabolites and immune‐modulating factors, simulating the spatiotemporal dynamics of tumor‐microbiome interactions.

Together, these approaches demonstrate that 4D tumor models, driven by material transformation, environmental responsiveness, and microbiome influence, offer a powerful means to study the evolving tumor ecosystem.^{[ 158} , ¹⁶⁷ , ^{174 ]}  Looking ahead, combining smart materials, microbiome niches, and computational modeling will enable personalized, microbiome‐aware tumor avatars. These could be used to test therapies on patient‐derived cells while accounting for microbiome‐driven variability in drug response,^{[ 170} , ¹⁷⁴ , ¹⁷⁵ , ^{176 ]} advancing the path toward precision oncology.

3. 3D Modelling of Organ and Human Microbiome Interactions

3.1. Host–Microbiome Signaling Axes in the Immuno‐Oncology Landscape

The advancements in 3D modeling of cancers discussed so far have opened opportunities to explore how the microbiome modulates anti‐tumor immunity. As aforementioned, these models integrate tumor heterogeneity, stromal components, and microbial communities to better replicate the in vivo TME. Microbiome composition and function can shape T cell activation, drive exhaustion phenotypes, and influence responsiveness to immune ICIs. Specific taxa such as Akkermansia muciniphila, Bifidobacterium longum, and Faecalibacterium prausnitzii have been linked to enhanced anti‐PD‐1/PD‐L1 efficacy, whereas dysbiosis can foster immunosuppressive TME and resistance to therapy.^{[ 177} , ¹⁷⁸ , ^{179 ]} These findings have important clinical relevance, particularly in the context of widely used PD‐1/PD‐L1 inhibitors such as pembrolizumab, nivolumab, and atezolizumab, as well as in combination regimens with CTLA‐4 blockade (e.g., ipilimumab)^{[ 180} , ¹⁸¹ , ^{182 ]} where microbiome composition may act as a determinant of therapeutic success. Microbiome‐derived metabolites, including SCFAs (e.g., butyrate, propionate), inosine, and tryptophan catabolites, can either bolster CD8⁺ T cell proliferation and cytotoxicity or promote functional exhaustion through upregulation of inhibitory receptors such as PD‐1, TIM‐3, and LAG‐3.^{[ 183} , ¹⁸⁴ , ¹⁸⁵ , ^{186 ]} Cutting‐edge immunocompetent 3D co‐culture systems are now being used to dissect these interactions with precision. For instance, a 3D ex vivo tumor‐immune co‐culture (3D‐HyGTIC) model has demonstrated progressive upregulation of exhaustion markers on CD8⁺ T cells and functional impairment under tumor‐induced stress‐mimicking in vivo exhaustion dynamics‐and showed restored function following anti‐PD‐1 treatment.^{[ 187 ]} Similarly, new platforms integrating organoids with immune cells (such as microfluidic, and bioprinted systems) enable mechanistic studies of how microbial signals and ICIs influence immune activation and dysfunction.^{[ 188 ]}

Studies using immunocompetent 3D co‐cultures incorporating tumor cells, autologous or allogeneic immune cells, and live bacteria are demonstrating that such platforms can monitor exhaustion marker expression, cytokine release, and T cell function in response to microbiome cues and ICI treatment.^{[ 188} , ¹⁸⁹ , ^{190 ]} Therefore, by capturing these immune–microbiome–tumor interactions under physiologically relevant conditions, these models can help identify bacterial signatures predictive of ICI response or resistance and inform strategies to reverse T cell exhaustion in the clinic.

Importantly, these immune effects are supported by specific host‐microbiome signaling pathways. Bacterial ligands such as LPS, flagellin, and CpG DNA activate Toll‐like receptors (TLRs) on tumor or immune cells, triggering MyD88 or TRIF‐dependent cascades that converge on NF‐κB and interferon regulatory factors, thereby shaping inflammatory cytokine profiles and antigen presentation.^{[ 191 ]} Activation of NF‐κB through TLRs, cytokine receptors, or NOD‐like receptors sustains transcription of IL‐6, TNF‐α, and chemokines, which can either recruit cytotoxic immune cells or reinforce immune suppression depending on the tumor context.^{[ 192 ]} Inflammasome complexes such as NLRP3 detect bacterial toxins or metabolic danger signals, leading to caspase‐1‐mediated maturation of IL‐1β and IL‐18, with dual roles in promoting anti‐tumor immunity or chronic inflammation.^{[ 193} , ^{194 ]} In some cases, bacterial genotoxins like colibactin and cytolethal distending toxin (CDT) directly induce DNA damage in tumor or immune cells, potentially altering their function.^{[ 195 ]}

These mechanistic frameworks provide the basis for designing advanced 3D model systems capable of reproducing and interrogating microbiome‐tumor‐immune interactions across diverse cancer types.

3.2. Gastrointestinal Cancers (Gastric, Colorectal, and Pancreatic)

Incorporating the gut microbiome in gastrointestinal (GI) tumor 3D‐modeling is crucial due to the gut microbiome's significant role in the development, progression, and treatment response of GI cancers. The gut microbiome refers to the ≈40 trillion microorganisms that inhabit the human gut. Specifically, the colorectum harbors about 30 trillion bacteria that constantly cross‐talk with the intestinal epithelium, immunological cells, and mucosal barrier, while affecting each other. Its influence is mediated through complex interactions involving metabolism, immune modulation, inflammation, and the production of microbiome metabolites. This connection is also affected by the composition of the GI microbiome.^{[ 196 ]} The microbiome can significantly influence the effectiveness of various treatments for GI cancers, including chemotherapy, immunotherapy, and radiotherapy. The importance of the intestinal microbiome is becoming increasingly evident through drug discovery and clinical trials, as it has been found to affect drug pharmacokinetics and treatment outcomes. Dysbiosis, can promote the development of cancer, however, certain bacteria may provide protective effects against it or improve the response to treatments.^{[ 197 ]} For example, Enterobacteria, Porphyromonas gingivalis, and F. nucleatum are more abundant in esophageal cancer patients, correlating with disease progression. Additionally, elevated levels of Bacteroides, particularly Bacteroidaceae and Bacteroidales, in esophageal squamous cell carcinoma (ESCC) patients are linked to tumor invasion through inflammatory pathways. It was found that overall, there was a reduction in the bacterial species of esophageal cancer patients compared to the normal healthy esophageal microbiome.^{[ 198 ]}

Additionally, various environmental factors can affect the microbiome, which can influence cancer progression. Acid reflux caused by environmental pollution could lead to esophageal inflammation and mucosal damage, causing changes to the esophageal microbiome in the distal part. This process allows the columnar epithelium to replace the original squamous epithelium, which could lead to the proliferation of harmful bacteria that may progress to esophageal adenocarcinoma.^{[ 199 ]} An example of how the microbiome can increase the risk of gastric cancer is infection with H. pylori. It is strongly associated with gastric carcinogenesis through mechanisms like cytotoxin production (cytotoxin‐associated gene A and vacuolating cytotoxin A), chronic inflammation, reactive oxygen species (ROS) activation, and atrophic gastritis. Moreover, it reduces gastric acidity, enabling other carcinogenic bacteria to thrive.^{[ 200 ]}

CRC patients show reduced bacterial diversity and increased abundance of pathogenic bacteria. For example, Enterotoxigenic B. fragilis (ETBF) is a pathogenic strain of B. fragilis that is correlated with colon tumorigenesis. Its virulence factor, fragilysin, triggers a pro‐carcinogenic inflammatory cascade by activating NF‐κB signaling in colonic epithelial cells, leading to the infiltration of pro‐tumoral myeloid cells. Additionally, Hungatella hathewayi, enriched in CRC, is significantly associated with the hypermethylation of the promoter regions of tumor suppressor genes like Caudal Type Homeobox 2 (CDX2) and MLH1, potentially driving intestinal tumorigenesis.^{[ 201 ]}

The microbiome can also be utilized to improve cancer patients' outcomes. Probiotics such as Lactobacillus, Bifidobacterium, Clostridium butyricum, and A. muciniphila show strain‐specific anti‐cancer effects, such as inducing apoptosis in cancer cells, enhancing cytotoxic T lymphocyte activity, and producing anti‐inflammatory SCFAs. Other probiotics reduced postoperative inflammation, enhanced immunity, and restored gut microbiome composition in gastric cancer patients.^{[ 201} , ^{202 ]}

Although the link between microbiome and GI cancers, such as esophageal, gastric, and CRC, is well established, its role in pancreatic ductal adenocarcinoma (PDAC) progression remains less clear. Emerging evidence, however, suggests that the gut microbiome may contribute to PDAC pathogenesis. For instance, microbiome have been shown to inhibit the infiltration and cytotoxic activity of natural killer (NK) cells within the TME. Beyond NK cells, microbiome metabolites can further shape the immunosuppressive landscape characteristic of PDAC, potentially influencing disease progression.^{[ 203} , ^{204 ]} Given that PDAC is one of the leading causes of cancer‐related death worldwide, the microbiome represents a promising, yet underexplored, therapeutic opportunity. As research continues to evolve, we anticipate increased attention to the microbiome's role in PDAC biology and treatment response.

Efforts are being made to model the PDAC TME in vitro using 3D models. One such study recreated the PDAC ECM using components like collagen type I, fibronectin, laminin, and hyaluronan, providing a physiologically relevant scaffold for evaluating drug responses, specifically to gemcitabine/nab‐paclitaxel. Due to its effectiveness in assessing treatment efficacy, this model could be further optimized by incorporating a patient‐specific microbiome, enabling more comprehensive studies of host‐microbiome‐tumor interactions.^{[ 205 ]}

The GI microbiome significantly influences the efficacy of chemotherapy, immunotherapy, and radiotherapy for GI cancers through microbiome‐mediated modulation of drug metabolism, immune responses, and tissue homeostasis. Certain bacterial populations have been shown to enhance chemotherapy responses. For example, H. pylori and other gut bacteria can suppress the activity of glutathione peroxidases (GPX‐1 and GPX‐2) in intestinal epithelial cells, enzymes responsible for peroxide detoxification. In murine models, the downregulation of Gpx1 and Gpx2 genes increased tumor sensitivity to chemotherapy.^{[ 206 ]} A similar effect was demonstrated by Lactobacillus plantarum in vitro. It amplified the cytotoxic effects of 5‐FU in CRC cells by decreasing the population of stem cell‐like cancer cells.^{[ 207 ]}

Conversely, certain bacterial species contribute to chemoresistance. F. nucleatum promotes resistance to 5‐FU by upregulating baculoviral IAP repeat‐containing 3 (BIRC3), an anti‐apoptotic gene, as demonstrated in both in vitro and in vivo studies. In clinical settings, a high abundance of F. nucleatum has been associated with poor response to 5‐FU‐based adjuvant chemotherapy in advanced CRC patients post‐surgery.^{[ 208 ]}

There is growing evidence indicating that specific bacterial species and their metabolites can modulate both innate and adaptive immune pathways. These interactions influence the TME and significantly affect the therapeutic outcomes of ICIs. Fecal microbiome transplantation (FMT) combined with pectin supplementation has been shown to significantly improve immune responses in both healthy individuals and CRC patients. This combination increased microbiome diversity, enhanced T cell infiltration, and notably boosted the response to anti‐PD‐1 monoclonal antibodies.^{[ 209 ]} In addition, the presence of specific bacteria, such as A. muciniphila and Bifidobacteria, has been correlated with improved clinical outcomes in patients treated with PD‐1 inhibitors like pembrolizumab and nivolumab.^{[ 210} , ^{211 ]} Butyrate‐producing bacteria have been linked to enhanced efficacy of CTLA‐four blockade, promoting the activation of macrophages and dendritic cells, as well as stimulating regulatory T cell (Treg) function. In a separate study, combining Lactobacillus acidophilus lysates with CTLA‐4‐targeting antibodies significantly improved anti‐tumor immunity in a murine colon cancer model. This synergistic effect was associated with increased infiltration of CD8⁺ T cells, expansion of effector memory T cells, and a reduction in immunosuppressive populations, including Tregs and M2 macrophages, within the TME.^{[ 212} , ^{213 ]} Therefore, there is a strong need to accurately model the TME and include GI microbiome components to deepen our understanding of the mechanisms involved in cancer progression and expand the horizons for new therapeutic options.

3.2.1. 3D Tumor Modeling of the GI‐Microbiome Interaction

Spheroids and organoids are among the most extensively studied 3D models of GI cancers (Figure  4 A), owing to their relative ease of fabrication and their effectiveness in recapitulating tumor–microbiome interactions. For example, a 3D CRC spheroid model was used to demonstrate the anti‐cancer potential of Lactobacillus fermentum. The Lactobacillus cell‐free supernatant (LCFS) of this probiotic inhibited cancer cell proliferation and induced apoptosis by suppressing NF‐κB signaling. These findings suggest that LCFS may serve as a multitarget agent that induces cell death.^{[ 214 ]} In another study, co‐cultured spheroids were used to explore the impact of the intratumoral microbiome on cellular and spatial heterogeneity in CRC and oral squamous cell carcinoma. Combining spheroids with spatial profiling, single‐cell RNA sequencing, and omics tools was the basis for studying processes like vascularization, metastasis, and immune‐microbiome‐stromal interactions. Notably, F. nucleatum induced neutrophil swarming and enhanced cancer epithelial cell migration, underscoring its pro‐tumorigenic role.^{[ 197 ]}

Figure 4.

Schematic illustration of advanced 3D models of GI cancers. A) Spheroids and organoids 3D models culture systems. Bright field and hematoxylin and eosin (H&E) staining show differences in growth among human gastric cancer organoids that are infected with lentivirus carrying either overexpressed or downregulated ONECUT2, a member of the ONECUT transcription factor family. Reproduced under the terms of the CC‐BY 4.0 license.^{[ 215 ]} 2024, published by Cell Death & Disease. Additionally, apoptotic morphology of HCT‐116 spheroids, a human CRC cell line, has been increased after treatment with Lactobacillus cell‐free supernatant (LCFS). Reproduced under the terms of the CC‐BY license.^{[ 214 ]} 2019, Biomolecules, MDPI. B) Transmembrane and flipwell systems. Tight junction formation (ZO‐1, in green) and mucin expression (MUC2, in red) of 3D intestinal model in e‐Transmembrane devices. Reproduced under the terms of the CC‐BY 4.0 license.^{[ 225 ]} 2024, published by Small Science, Wiley. A 3D multicellular co‐culture system designed to replicate the gut mucosal microenvironment, incorporating a blend of the colon epithelial cell line, Caco2, and the mucus‐secreting cell line, HT‐29. Reproduced under the terms of the CC‐BY 4.0 license.^{[ 224 ]} 2023, published by Scientific Reports. C) OoC models. A device consists of a cell culture chamber and side channels for gradient generation, which could be utilized to co‐culture bacteria and cancer cells and study their interactions in static and dynamic environments. Utilizing this model, the researchers demonstrated that the bacterial stimulus promotes the growth of the CRC cell line, HCT‐116, over time. Reproduced under the terms of the CC‐BY license.^{[ 219 ]} 2023, Sensors, MDPI D) 3D‐bioprinted models. Printed tumor tissues with patient‐derived, tumor‐laden, tissue‐specific bioinks. Researchers developed a 3D‐bioprinted GC model for preclinical chemotherapy using extrusion‐based 3D bioprinting technology, printed in a 24‐well plate. Reproduced under the terms of the CC‐BY 4.0 license.^{[ 226 ]} 2025, Advanced Science, Wiley. Fabrication process of silk‐based porous scaffolds for 3D human intestine engineering. Staining was conducted to confirm the morphology and phenotype of the cells inside the human 3D scaffolds. Reproduced under the terms of the CC‐BY 4.0 license.^{[ 227 ]} 2015, Scientific Reports. Created with BioRender.

In gastric cancer research, Lin et al. utilized a 3D gastric cancer organoid model to track the pathway of H. pylori infection. They demonstrated that H. pylori upregulates ONECUT2 transcriptional activity via the NF‐κB pathway, subsequently reducing AKT and β‐catenin phosphorylation. This highlights the role of ONECUT2 in gastric cancer progression and its potential as a therapeutic target.^{[ 215 ]}

For the CRC models and their interaction with the gut microbiome, several research groups have developed intestinal organoids. Leslie et al. used 3D organoids to study the pathogenesis of the obligate anaerobe Clostridium difficile. Heo et al. combined Cryptosporidium infection with small intestine organoids and allowed observation of its complete life cycle.^{[ 216} , ^{217 ]}

Despite the advantages of spheroids and organoids, these models still lack dynamic physiological features, such as blood flow and mechanical stimulation, that are particularly relevant in GI cancers, where peristalsis, fluid shear, and nutrient flow play critical roles in tumor progression and microbiome interactions. To bridge this gap, OoC systems have been developed. One notable example of the use of an OoC model is the gut‐liver axis chip, composed of two chambers, gut epithelium and liver spheroids, separated by a porous membrane. This platform was used to investigate how microbiome‐derived metabolites and extracellular vesicles influence hepatocyte function.^{[ 218 ]} Another microfluidic co‐culture model used HCT116, a CRC cell line, in a collagen matrix, with side channels introducing Bacillus bacteria and lipopolysaccharide (LPS) as Gram‐positive and Gram‐negative stimuli, respectively. Both agents significantly enhanced cancer cell proliferation, demonstrating the tumor‐promoting role of specific bacterial stimuli (Figure 4C).^{[ 219 ]} In a different model, the enterotoxigenic ETBF strain was incorporated into a gut‐microbiome‐on‐a‐chip model to study its pathogenic behavior. Indeed, it disrupted the epithelial barrier and activated pro‐tumorigenic signaling. Penarete‐Acosta et al. developed a microfluidic device combining colonocyte spheroids and colorectal microbiome, specifically F. nucleatum, enabling studies of host‐microbiome interactions under variable environmental conditions. This setup can be further expanded to include immune and stromal components for advanced microbiome engineering research.^{[ 220 ]} Additionally, Cremer et al. designed a microfluidic system that mimics intestinal peristalsis using membrane valves to simulate colonic wall contractions. Using E. coli strains, they demonstrated how mechanical forces shape microbiome density and composition, highlighting the importance of physical dynamics in gut ecology.^{[ 221 ]}

Moreover, given the importance of maintaining intestinal barrier integrity in the context of tumor‐microbiome interactions, advanced platforms have been developed to enable real‐time monitoring of epithelial function. For example, to examine the impact of microbiome on gut barrier integrity, Moysidou et al. developed an electronic transmembrane (e‐Transmembrane) device. This bioelectronic platform integrates an electroactive scaffold that serves both as a tissue support and as an electrode, enabling real‐time, label‐free monitoring of barrier tissue function using electrochemical impedance spectroscopy. By co‐culturing the scaffold with live bacteria, the system captures distinct patterns over 24 h and correlates these with microscopic and molecular analyses of host tissue integrity. While the model primarily investigates microbiome effects on intestinal health and immune signaling, and does not focus specifically on cancer, its framework offers a valuable tool for screening microbiome‐derived therapies and understanding host‐bacteria interactions in contexts that could include cancer in future applications.^{[ 222 ]}

Another advancement in gut modeling is the 3D flipwell system, which enables the co‐culture of stratified mucosal components, including bacteria, mucus, epithelial cells, and immune cells. This model simulates the mucosal barrier architecture and facilitates the study of cellular crosstalk among gut components, all of which are essential to understanding pathological processes such as inflammation and cancer progression. This layered system is particularly well‐suited for investigating how the gut microbiome influences cancer‐relevant mechanisms, such as epithelial transformation, immune modulation, and barrier dysfunction. By reconstituting the multicellular, multilayered nature of the gut mucosa, the 3D Flipwell system offers a unique platform for dissecting the microbiome‐mucosa‐immune axis in colorectal and other GI cancers (Figure 3B).^{[ 223} , ²²⁴ , ^{225 ]}

Despite significant advances in 3D‐bioprinted GI cancer models, they remain one of the most challenging platforms for accurately modeling the TME, particularly when incorporating the microbiome alongside patient‐derived tumor tissues. While several studies have successfully 3D‐bioprinted tumor constructs that preserve epithelial, stromal, and even immune components, the inclusion of live commensal or pathogenic bacteria remains an unmet challenge. One example is a 3D‐bioprinted model that uses patient‐derived gastric cancer tissues encapsulated in dECM hydrogels, preserving the native tissue architecture. This model maintains key stromal elements, particularly fibroblasts, and enables the reconstruction of cell–matrix and stromal interactions that are critical for accurate drug response prediction. Although the current model includes both cancer and stromal cells, future directions include incorporating patient‐derived immune components, such as CD4⁺ T helper cells, to further mimic the tumor immune microenvironment. Another promising avenue, though not yet explored, is the inclusion of patient‐specific microbiome, which would allow the model to better recapitulate the complexities of the tumor‐microbiome ecosystem (Figure 4D).^{[ 226 ]}

In a related effort, Chen et al. developed a bioengineered 3D intestinal model featuring a hollow lumen like that of the human gut. Built on a silk fibroin scaffold, the model represents healthy intestinal tissue and, although it does not currently include microbiome, holds potential as a functional, physiologically relevant platform for studying tumor progression, drug interactions, and future host‐microbiome co‐cultures.^{[ 227 ]}

Developing 3D‐bioprinted GI tumor models that incorporate a functional microbiome would represent a significant leap forward, enabling a more faithful recreation of the TME and providing a powerful platform for microbiome‐targeted therapies, drug testing, and immunomodulation studies.

3.3. Lung Cancer Models

Recent evidence shows that a variety of bacteria reside within lung tumors, highlighting the protective role of airway microbiome.^{[ 228} , ^{229 ]} For example, Pseudomonas aeruginosa (P. aeruginosa) and Staphylococcus aureus (S. aureus) have been detected as tumor‐associated bacteria in a high proportion of lung cancers (up to 85% of elderly patients).^{[ 230 ]} To emulate these interactions in vitro, non‐small cell lung cancer (NSCLC) spheroids were co‐cultured with a biofilm‐forming P. aeruginosa in a microfluidic 3D model.^{[ 231 ]} The biofilms on the spheroids' surface secreted iron‐scavenging siderophores that protected tumor cells from ferroptotic cell death, leading to significantly enhanced tumor viability and growth compared to sterile spheroids. This mechanistic finding underscores how incorporating the microbiome into 3D cancer models can reveal mutualistic interactions that would be missed in traditional 2D cultures.

Lung tumor models that incorporate airway microbiome have illuminated how dysregulated lung commensals can accelerate tumor progression by engaging inflammatory, immune, metabolic, and ECM remodeling pathways.^{[ 232 ]} In a 3D lung cancer organoid model, microbiome‐derived butyrate from Roseburia intestinalis (R. intestinalis) significantly enhanced tumor invasiveness, driving collective invasion fronts that penetrated type I collagen.^{[ 233 ]} This was accompanied by increased secretion of matrix metalloproteinases (MMP1 and MMP9), which degrade the ECM and facilitate invasion. Mechanistically, butyrate suppressed HDAC2 activity and increased H3K27 acetylation at the H19 promoter, upregulating the lncRNA H19, which in turn promoted epithelial–mesenchymal transition (EMT) markers and cytoskeletal remodeling, enabling cells to migrate through dense ECM. Additionally, butyrate derived from R. intestinalis induced polarization of tumor‐associated macrophages toward an anti‐inflammatory phenotype, creating an immunosuppressive microenvironment that supports tumor progression and metastasis. Although R. intestinalis is an intestinal bacterium, this study underscores the importance of employing 3D lung cancer models to investigate microbiome‐related tumor progression.

To accurately model these complex interactions, the choice of the hydrogel scaffold is critical. Hydrogels for lung cancer models must replicate the soft, elastic mechanics of alveolar tissue while supporting bacterial viability and host cell function. GelMA and PEG‐based hydrogels allow tunable stiffness and stretch, supporting epithelial barrier formation, surfactant production, and fibroblast–epithelial interactions under physiologically relevant conditions.^{[ 234} , ²³⁵ , ^{236 ]} For example, a 3D “inverse‐opal” GelMA scaffold with interconnected 200 µm pores enabled the formation of ≈7000 alveolar‐like sacs in just a few mm³, where primary alveolar epithelial cells formed tight ZO‐1‐positive barriers and secreted surfactant proteins more effectively than on flat substrates.^{[ 239 ]} Similarly, PEG‐norbornene hydrogels co‐cultured with iPSC‐derived alveolar cells and fibroblasts demonstrated that “soft” lung‐like gels supported healthy epithelial function, while “stiff” fibrotic‐like gels promoted fibroblast activation and pathological differentiation.^{[ 237 ]}

Natural ECM‐based hydrogels, including lung dECM and collagen, further enhance cell maturation and preserve key biochemical cues.^{[ 238 ]} For instance, dECM hydrogels supported rapid surfactant protein expression and alveolosphere formation from iPSC‐derived AT2 cells, outperforming Matrigel or 2D cultures.^{[ 239 ]} These matrices not only enable host–bacteria co‐culture but also facilitate the study of how bacterial signals influence tumor behavior within a responsive, tissue‐like microenvironment.

Patient‐derived lung cancer assembloids, created by encapsulating tumor organoids with autologous stromal cells within microfluidic droplets, effectively preserved tumor heterogeneity, accurately reflected in vivo tumor biology, and successfully predicted individual patient responses to chemotherapy and targeted therapies (Figure  5 A).^{[ 240 ]} Multilayered lung cancer spheroids cultured on TransWell, at the air‐liquid interface (ALI), develop hypoxic cores and exhibit enhanced chemoresistance, providing a physiologically relevant model for evaluating inhaled chemotherapeutics.^{[ 241} , ^{242 ]} Furthermore, ALI cultures facilitate oxygen exposure, crucial for bacterial growth, enabling accurate replication of lung microenvironments and host–bacteria interactions.^{[ 243} , ^{244 ]} Incorporating bacteria into TransWell models could further enhance their physiological relevance by modeling tumor‐microbiome interactions and evaluating their influence on tumor progression and treatment responses (Figure 5B).

Figure 5.

Advanced 3D lung cancer models incorporating microbiome interactions. A) Spheroids and organoids. Representative images of patient aggregates driven from lung cancer‐associated tissue are shown. Immunofluorescence staining demonstrates heterogeneous TME. Reproduced under the terms of the CC‐BY 4.0 license.^{[ 240 ]} 2024, Nature Communications. B) Air–liquid interface (ALI) culture model enabling direct exposure of epithelial cells to microorganisms, widely used for studying host‐microbiome interactions. Reproduced under the terms of the CC‐BY 4.0 license.^{[ 244 ]} 2020, Scientific Reports. C) Lung‐on‐a‐chop model. The integrated superhydrophobic microwell array chip (InSMAR‐chip) includes a superhydrophobic microwell array that enables controlled droplet formation for 3D lung cancer organoid culture. Lung cancer organoids (LCOs) grown on‐chip showed comparable growth and viability to those cultured in conventional plates, maintaining tumor morphology over 14 days. Reproduced under the terms of the CC‐BY 4.0 license.^{[ 245 ]} 2021, Nature Communications.

3D‐bioprinting and microfluidic lung‐on‐a‐chip technologies have substantially enhanced our ability to recapitulate the spatial, mechanical, and cellular complexity of lung tumors.^{[ 245 ]} Moreover, lung‐on‐chip devices that simulate cyclic mechanical strain mimicking pulmonary breathing motions demonstrated enhanced innate immune responses to bacterial challenge, emphasizing biomechanics as modulators of host‐bacteria interactions.^{[ 246 ]} The lung‐on‐a‐chip platform, although currently utilized primarily for modeling infectious diseases such as early Mycobacterium tuberculosis infection, demonstrates significant potential for lung cancer research.^{[ 247 ]} By directly incorporating bacteria into alveolar epithelial channels under physiological breathing conditions, this model effectively captures the complexities of host–bacteria interactions. A recent lung tumor‐on‐a‐chip model utilized an ALI cultured EGFR‐mutant NSCLC cells alongside stromal components to evaluate resistance to EGFR‐TKIs under physiologically relevant conditions.^{[ 248 ]} Integrating this ALI platform with the airway microbiome would enable investigation of how bacterial factors influence tumor progression, immune modulation, and therapeutic resistance within a controlled, multicellular lung cancer microenvironment (Figure 5C).^{[ 247} , ^{248 ]}

3.4. Breast Cancer Models

Breast cancer is a heterogeneous disease deriving from the ductal epithelium of mammary tissue and represents the most prevalent form of cancer among women.^{[ 249 ]} The TME is composed of diverse cell populations, including myoepithelial cells, stromal fibroblasts, immune cells, and adipocytes, which are part of the ECM.^{[ 250 ]}

The complex relationship of the TME, specifically the microbiome, and breast cancer plays a critical role in tumor progression and therapeutic response. Various studies confirmed that the microbiome secretes cytostatic metabolites that inhibit the proliferation and migration of cancer cells. For example, it was shown that indolepropionic acid (IPA), a bacterial tryptophan metabolite, had cytostatic properties and selective targeting to breast cancer cells, without affecting non‐transformed, primary fibroblasts. IPA reduced the amount of cancer stem cells and the metastasis of cancer cells in vivo and in vitro models.^{[ 251 ]}

Additionally, it was observed that the presence of different bacterial communities in healthy versus breast cancer tissue could provide a diagnostic and predictive biomarker. The microbiome colonies were studied both in healthy and breast cancer tissues, and different types of bacterial colonies were detected. Due to this fact, microbiomes from breast cancer tissue have the potential to complement existing diagnostic and therapeutic methods. For example, the researchers discovered that neoadjuvant chemotherapy reduces the diversity of the microbiome in tumor tissue.^{[ 252 ]} Genus‐level analysis revealed higher levels of Pseudomonas and lower levels of Prevotella in tumor samples from patients who received neoadjuvant chemotherapy.^{[ 253 ]}

To investigate interactions within breast cancer and its TME, 3D models such as spheroids and organoids have been used as tools for exploring the relationship between the microbiome and breast cancer, particularly in the context of tumor progression and treatment response (Figure  6 A).^{[ 254} , ²⁵⁵ , ^{256 ]} For example, researchers facilitated the metabolite exchange between gut bacteria and the tumor by guiding the flow in the chips from the microbiome to the breast tumor spheroids. The outcome was that these spheroids cultured with Estradiol‐Glow‐supplemented media showed a high expression of GREB1 (estrogen‐responsive gene transcript) as compared to the vehicle control group (Figure 6B).^{[ 257 ]} In addition, recent studies have shown that bacterial populations in the gut may significantly impact breast cancer progression and therapy response. Microfluidic chips that simulate host‐bacteria‐cancer interactions, enabling a real‐time analysis of how microbiome metabolites influence breast cancer spheroids. This system seems to be a solid tool of how the microbiome modifies cancer cell behavior, immune responses, and drug resistance. This system offers a robust understanding of tumor progression and therapeutic resistance.^{[ 43 ]}

Figure 6.

Illustration of 3D breast cancer models, which include organoid, 3D‐bioprinted, and OoC platforms. A) 3D breast cancer cells spheroids. Paraffin‐embedded, fixed BT‐474 breast cancer spheroid cross sections show increased calcium deposition in the central regions after exposure to supplemented calcium and bacterial extracellular vesicles in the culture medium. Reproduced under the terms of the CC‐BY license.^{[ 256 ]} 2025, iScience, Elsevier. B) OoC. The 3D model tissues and supporting matrices are loaded into the chip. The hydrogel and media compartments of the single‐channel tumor chip (SCTC) are separated by a porous membrane to avoid flow blockage. Confocal microscopy imaging allows visualization of epithelial and junctional markers (E‐cadherin, ZO‐1, β‐catenin) in collagen‐embedded MCF7, human breast cancer spheroids. Reproduced under the terms of the CC‐BY 4.0 license.^{[ 257 ]} 2024, Advanced Science, Wiley. C) 3D‐bioprinted tumor‐stroma breast cancer model. On the right, tumor core (blue), embedded in dECMs that was served as a bioink. In the middle and bottom left images, live‐cell confocal imaging of breast cancer stem cells (green) surrounded by human breast fibroblasts (pink). Top left, images of H&E staining after 7 days in culture. Even though microbiome components were not included in this 3D model it could be further utilized to study tumor‐host‐microbiome interactions. Adapted under the terms of the CC‐BY‐NC‐ND 4.0 license.^{[ 262 ]} 2024, ACS Applied Materials & Interfaces, American Chemical Society. Created with BioRender.

Researchers using breast cancer models containing low and high collagen amounts showed that bacteria penetrate and extensively colonize low collagen‐containing homotypic tumor spheroids. The bacteria remain adhered to the peripheral part of the physiologically relevant high collagen‐content tumor spheroids. The reduced bacterial colonization leads to a reduced antitumor effect of the bacteria. Further, quantitative evaluation of the spatial distribution of bacteria, collagen, and apoptotic cells showed a robust correlation between high collagen content, reduced bacterial colonization, and reduced regional tumor cell death.^{[ 258 ]} Understanding the effect of collagen on bacteria‐based cancer therapy performance may lead to making more efficacious strains, capable of overcoming this barrier to effective colonization of primary tumors and metastases.

In parallel, ECM components such as collagen I and laminin play key roles in shaping the breast TME. Increased collagen I density and alignment have been correlated with poor prognosis and resistance to HER2‐targeted therapies.^{[ 259 ]} PEG‐based synthetic hydrogels functionalized with collagen I or laminin demonstrated that collagen I‐rich matrices enhanced cancer cell spreading, motility, and invasion, while laminin‐rich environments promoted more stable phenotypes.^{[ 260 ]} Laminin also influences cancer stem cell self‐renewal through activation of the Hippo transducer TAZ, creating a feedback loop that sustains tumor growth. These findings underscore the importance of incorporating specific ECM cues into in vitro breast cancer models to accurately study tumor progression and therapeutic response.

3D‐bioprinted breast cancer models significantly enhance the ability to study the TME, including microbiome influences. Recent 3D‐bioprinted applications included models that facilitate the interaction between breast cancer cells and bone cells, utilizing GelMa hydrogels with nanocrystalline hydroxyapatite, which promotes enhanced cancer cell growth. Furthermore, 3D models using bioinks like gelatin‐alginate‐fibrinogen demonstrated higher drug resistance in breast cancer cells when compared to 2D cultures.^{[ 259 ]} The flexibility of 3D‐bioprinting, including techniques like FRESH allows better control over pore size and fiber density, influencing cellular responses and resulting in an improved mimic of breast cancer microenvironments by incorporating proteins like collagen and fibronectin (Figure 6C).^{[ 261} , ^{262 ]} 3D‐bioprinted systems provide better physical and biochemical signals and therefore hold significant potential for personalized cancer research and drug testing.^{[ 260 ]}

Another application of 3D‐bioprinted models arose from the need for a reliable in vitro model that recapitulates the complexity of the breast cancer stem cells (CSC) TME to enhance drug discovery. This research introduced a 3D‐bioprinted breast CSC tumor‐stroma model comprising triple‐negative breast CSCs (TNB‐CSCs) and stromal cells, within a breast‐derived dECM bioink. It was shown that the model can outline the invasive potential of TNB‐CSC. As a proof‐of‐concept, researchers conducted high‐throughput drug testing analyses to assess the efficacy of CSC‐targeted therapy in combination with conventional chemotherapeutic compounds. The results concluded that tumor‐stroma models offer an advantage as a drug‐screening platform. These models may provide valuable platforms for evaluating the impact of microbiome‐targeted therapies on breast cancer progression and metastasis.^{[ 263 ]} Therefore, the implementation of microbiomes in such 3D‐bioprinted models will be beneficial in the study of the microbiome's impact on tumor progression and therapeutic efficacy.

3.5. Skin Cancer and Melanoma Models

Creating skin models requires advanced 3D models that are essential for capturing the intricate architecture and cellular dynamics of human skin. Among these, organotypic skin models (OSMs) stand out as sophisticated in vitro systems that replicate the structural, cellular, and functional complexity of native skin, providing a more human‐relevant alternative to traditional 2D cultures and animal models. These models typically consist of stratified keratinocytes forming an epidermal layer atop a fibroblast‐populated dermal matrix embedded in collagen or fibrin hydrogels, simulating the connective tissue layer of the skin. Culturing keratinocytes at the air–liquid interface, with the dermal side submerged and the epidermal side exposed to air, induces stratification and cornification, forming a stratum corneum (the outermost layer of skin) that mimics the barrier function of native skin.^{[ 264 ]} Integration of melanocytes enables studies of pigmentation and melanoma progression,^{[ 265} , ^{266 ]} while the incorporation of immune cells such as macrophages or dendritic cells allows modeling of inflammatory responses and tumor–immune cross‐talk (Figure  7C).^{[ 266} , ^{267 ]} OSMs provide a powerful platform to investigate tumor–stroma interactions, EMT, immune evasion, and therapeutic response under patho‐physiologically relevant conditions. Advanced iterations employ dECM scaffolds, 3D‐bioprinting, and microfluidic skin‐on‐chip technologies that support spatial patterning of different cells, vascular mimicry, and dynamic perfusion‐features critical for drug screening, metastasis modelling, and personalized therapy development.^{[ 268} , ²⁶⁹ , ^{270 ]}

Figure 7.

Overview of advanced 3D skin models and melanoma constructs. A) Spheroids: Stepwise generation of melanoma multicomponent spheroids from human melanoma cells (orange), endothelial cells (blue), and dermal fibroblasts (green). Confocal images show progressive tumor outgrowth over time. Adapted under the terms of the CC‐BY‐NC‐ND license.^{[ 274 ]} 2024, Star Protocols, Elsevier. B) Assembloids: Fusion of spheroids composed of human dermal microvascular endothelial cells (hDMECs), human dermal fibroblasts (hDFbs), human adipose‐derived stem cells (hASCs), and melanoma cells. Time‐lapse images demonstrate structural integration, compartmentalization, and invasive capacity of melanoma assembloids. Reproduced under the terms of the CC‐BY license. 2024, Acta Biomaterialia, Elsevier.^{[ 277 ]} C) Organotypic models: Schematic of organotypic culture generation. A fibroblast‐containing collagen plug is seeded with melanocytes and neonatal human epidermal keratinocytes (NHEKs). Air‐liquid interface culture induces epidermal stratification. H&E staining and immunofluorescence confirm epidermal architecture and melanocyte localization. Reproduced under the terms of the CC‐BY 4.0 license.^{[ 266 ]} 2022, Current Protocols. D) 3D‐bioprinted skin constructs. Models were 3D‐bioprinted by the extrusion method using dECM‐based bioinks. Top row: Confocal projections on day 7 and day 14 showing progressive invasion of melanoma cells (magenta) into fibroblast‐laden dermal dECM (green). Bottom left: Macroscopic view of a layered 3D‐bioprinted construct (scale bar 0.5 cm). Bottom middle: High‐resolution confocal images illustrating melanoma (magenta) invasion into fibroblast‐populated dermal dECM (green) with orthogonal views. Bottom right: Transversal sections showing melanoma infiltration across the dECM layers. Reproduced under the terms of the CC‐BY 4.0 license.^{[ 288 ]} 2024, Wiley advanced. Created with BioRender.

Despite challenges such as donor variability and limited vascularization, OSMs have proven valuable in recapitulating key features of the melanoma TME and have become instrumental in elucidating skin‐microbiome interactions. By enabling colonization with defined microbiome consortia, these models offer unparalleled opportunities to explore how commensal and pathogenic bacteria modulate skin immunity, barrier integrity, and tumor progression. Recent work has demonstrated the feasibility of sustaining long‐term microbiome stability on OSMs, offering insight into microbiome metabolite signaling, immune modulation, and potential contributions to treatment resistance and cancer progression.^{[ 271 ]} The convergence of tumor biology, immunity, and microbiome in OSMs enables holistic modeling of the skin tumor–immune–microbiome axis, opening new avenues for targeted and personalized treatments.

3D melanoma spheroids are widely utilized as in vitro models that recapitulate essential features of solid tumors, including cellular organization, oxygen and nutrient gradients, and both cell‐cell and cell‐matrix interactions. These structures offer a physiologically relevant system to investigate melanoma progression, invasion, drug resistance, and immune evasion.^{[ 272} , ^{273 ]} Spheroids, while widely used in cancer research, are often homotypic, composed primarily of tumor cells. Although they can be adapted into co‐culture systems by incorporating additional cell types such as fibroblasts or immune cells,^{[ 274 ]} their structural simplicity and limited spatial organization restrict their ability to fully mimic the complexity of TME (Figure 7A).

To overcome these limitations, assembloids have emerged as more advanced multicellular 3D constructs. By incorporating tumor cells along with stromal and immune components—such as fibroblasts, endothelial cells, macrophages, dendritic cells, or T cells assembloids more accurately recreate the heterogeneity and cellular cross‐talk within the TME. These models enable dynamic assessments of tumor–immune interactions, immunosuppressive mechanisms, and therapeutic efficacy, particularly in the context of immune checkpoint blockade.^{[ 275} , ²⁷⁶ , ²⁷⁷ , ^{278 ]}

Furthermore, both spheroid and assembloid systems have recently been adapted to include microbiome components, providing new insights into the role of the skin microbiome in modeling and modulating melanoma biology. Co‐culture studies with selected bacterial strains have shown effects on cytokine signaling, immune cell recruitment, and cancer cell behavior, highlighting the microbiome's capacity to reshape the immune landscape of tumors and potentially alter responses to immunotherapy (Figure 7B).^{[ 188} , ²⁷⁶ , ^{277 ]}

Advanced 3D‐bioprinted skin constructs are created with bioinks composed of skin‐relevant cells such as keratinocytes, fibroblasts, melanocytes, endothelial cells, and biomimetic ECM components. These constructs replicate the compartmentalized architecture of human skin, including epidermal, dermal, and, in advanced models, hypodermal and vascularized layers. Unlike traditional 2D cultures, 3D‐bioprinted models enable spatially defined multicellular co‐culture, making them ideal platforms for investigating melanoma‐specific processes such as tumor–stroma interactions, immune evasion, and drug resistance.^{[ 279} , ^{280 ]}

Recent innovations have yielded patient‐derived melanoma constructs that integrate cancer stem cells, fibroblasts, mesenchymal stem cells, and endothelial cells within hydrogel scaffolds, allowing for the recreation of TME heterogeneity and personalized therapeutic testing. The use of dECM based bioinks has further improved physiological relevance by preserving native biomechanical and biochemical signals (Figure 7D).^{[ 281 ]} Vascularized, immunocompetent skin equivalents generated via 3D‐bioprinting have enabled more realistic modeling of melanoma progression and immune cell dynamics in response to treatment.^{[ 282 ]}

While most current models focus on replicating the cellular and stromal complexity of the TME, emerging approaches are beginning to incorporate microbiome to study their influence on skin physiology and disease. Although such microbiome‐integrated constructs have not yet been widely applied to melanoma, they hold significant potential to unravel the role of the skin microbiome in tumor development, immune modulation, and therapeutic efficacy. The interplay between the host‐associated microbiome and the immune system has emerged as a critical axis in melanoma biology, influencing tumor development, progression, and responsiveness to immunotherapy. The skin microbiomes function as immunologically active ecosystems that engage in complex crosstalk with host cells, shaping local and systemic immune landscapes.^{[ 283 ]}

The skin microbiome, long underestimated in oncology, is now recognized as a dynamic regulator of cutaneous immune homeostasis. Commensals such as Staphylococcus epidermidis and Cutibacterium acnes enhance antitumor immunity through TLR‐mediated signaling, modulation of dendritic cell maturation, and cytokine release, including IL‐17A and IFN‐γ, cytokines known to support immune surveillance.^{[ 284} , ^{285 ]} In contrast, skin dysbiosis, particularly characterized by the enrichment of Staphylococcus aureus, has been linked to chronic inflammation, impaired barrier function, and local immunosuppression, collectively fostering a tumor‐permissive niche. These bacterial influences are substantiated by spatial metagenomic studies identifying viable, tumor‐specific intracellular bacteria within melanoma lesions, implicating the microbiome in tumor progression and immune evasion mechanisms.^{[ 22} , ^{286 ]}

To better recapitulate the tumor‐microbiome‐immune axis, advanced immunocompetent 3D models such as organotypic skin cultures and 3D‐bioprinted melanoma constructs are being developed. These platforms integrate live microbiome or microbiome metabolites to capture physiologically relevant interactions, offering a more predictive tool for mechanistic and translational studies.^{[ 287} , ^{288 ]}

3.6. Other Cancer Types (Prostate, Liver, and Brain)

While the microbiome's impact on colorectal, pancreatic, and breast cancers has been widely characterized, its role in less‐studied cancers such as prostate, liver, and brain is only beginning to be uncovered. Advanced 3D tissue models are crucial for the understanding of these complex tumor‐microbiome interactions.^{[ 289} , ^{290 ]}

Prostate cancer research has increasingly focused on the role of urinary and gut microbiome in shaping local inflammation and contributing to tumorigenesis. Bacterial colonization by species such as Propionibacterium acnes (P. acnes) has been implicated in driving chronic inflammation within the prostate, thus promoting tumor‐supportive conditions.^{[ 291 ]} Although microbiome co‐culture models remain limited, patient‐derived prostate cancer organoids developed by Gao et al. retain essential tumor characteristics and present a promising platform for microbiome‐related studies. These systems could be used to explore how the urinary microbiome influences epithelial cell behavior, inflammatory cytokine expression (e.g., IL‐6, TNF‐α), and immune signaling pathways relevant to prostate cancer development.^{[ 292 ]}

In prostate cancer research, organoids have been key in modeling tumor behavior and interactions with the surrounding microenvironment. They support long‐term co‐culture with immune cells and bacteria, enabling the study of complex interactions over extended periods. For instance, co‐culturing prostate cancer organoids with immune cells allows researchers to investigate immune responses and potential immunotherapies. Additionally, introducing microbiome‐derived factors into these cultures helps in understanding how microbiome metabolites influence tumor progression and treatment responses. The ability to maintain these co‐cultures over time makes organoids a valuable tool for studying the dynamic interactions between tumors, the immune system, and the microbiome.^{[ 61} , ^{293 ]}

Previously, the prostate gland was considered to be free of bacterial colonization. However, recent evidence has revealed the presence of a diverse urinary microbiome that may play a significant role in prostate health. Shifts in microbiome composition have been linked to chronic inflammation, which is a well‐established contributor to prostate cancer development, suggesting that the microbiome could influence both tumor initiation and progression.^{[ 294} , ^{295 ]}

Organotypic 3D models such as patient‐derived prostate organoids provide new opportunities to explore these interactions in a biologically relevant context. These organoids replicate glandular architecture, maintain key oncogenic signaling pathways like androgen receptor activity, and support long‐term culture conditions necessary for host‐microbe studies.^{[ 296 ]} The landmark study by Pamarthy et al. demonstrated that patient‐derived organoids from both localized and advanced prostate cancer retain molecular features of the originating tumors and are amenable to drug screening and genomic profiling. The study highlighted the ability of patient‐derived organoids to reflect inter‐patient heterogeneity and treatment responses, establishing them as a promising platform for personalized medicine (Figure  8 A).^{[ 297 ]}

Figure 8.

3D Cancer Models for Prostate, Liver, and Brain Tumors. A) Prostate tumor organoids. This figure illustrates the workflow for deriving prostate cancer organoids from patient tumor biopsies, modeling tumor‐specific genetic alterations in vitro. Reproduced under the terms of the CC‐BY 4.0 license.^{[ 297 ]} 2021, Molecular Cancer. B) Brain tumor spheroids. This panel shows the process for generating glioblastoma spheroids in ultra‐low attachment (ULA) plates. Reproduced under the terms of the CC‐BY license.^{[ 315 ]} 2015, Journal of Biotechnology, Elsevier. C) Liver tumor 3D‐bioprinted scaffolds. This approach involves 3D‐bioprinting of hydrogel scaffolds embedded with HCC cells. The platform is formed layer‐by‐layer and crosslinked using photoinitiated chemistry to mimic the liver's ECM architecture. Reproduced under the terms of the CC‐BY 3.0 license.^{[ 303 ]} 2023, Biomaterials Science, Royal Society of Chemistry. Created with BioRender.

Focusing on this, recent research has introduced strains such as E. coli and P. acnes (commonly detected in urine or prostate tissue) into organoid co‐cultures. These bacteria have been shown to trigger upregulation of inflammatory pathways, disrupt epithelial polarity, and modulate stem cell behavior, potentially promoting tumorigenic microenvironments.^{[ 295 ]}

Moreover, the urinary microbiome has been implicated in modifying responses to prostate cancer therapies, such as androgen deprivation, potentially through microbiome metabolism of host compounds or modulation of local immune activity.^{[ 291 ]} This has led to an increased focus on using prostate organoid models not only to study tumor progression but also as platforms for testing microbiome‐targeted therapies and immunomodulatory interventions.^{[ 298 ]}

Hepatocellular carcinoma (HCC), the most common primary liver cancer, is increasingly understood to be shaped by the gut‐liver axis, a dynamic system through which gut‐derived microbiome metabolites and endotoxins, such as short‐chain SCFAs, LPS, and secondary bile acids, reach the liver via the portal vein.^{[ 299} , ³⁰⁰ , ^{301 ]} These microbiome components influence hepatic inflammation, immune responses, hepatocyte proliferation, DNA damage, and cytokine production, contributing to tumor progression and HCC pathogenesis.^{[ 292 ]} Gut dysbiosis exacerbates chronic liver injury and fosters a pro‐carcinogenic microenvironment.

To investigate these interactions in a physiologically relevant setting, scaffold‐based 3D liver models have been developed to replicate the structural and functional complexity of hepatic tissue.^{[ 300} , ^{302 ]} These models typically use ECM‐based biomaterials such as collagen, GelMA, alginate, or decellularized liver matrix to support hepatocyte growth and maintain liver‐specific functions over extended culture periods.^{[ 300 ]} Importantly, such platforms enable compartmentalized co‐culture systems that simulate the gut‐liver axis, allowing microbiome metabolites to act on hepatocyte layers via semi‐permeable membranes or microfluidic interfaces.

A recent review by Paradiso et al. emphasized the potential of these engineered liver tissues as next‐generation platforms that outperform traditional 2D and animal models in terms of long‐term hepatocyte functionality and responsiveness to drugs and toxins (Figure 8C).^{[ 303 ]} Biofabrication approaches, including layered scaffolds and 3D‐bioprinted liver constructs, have been used to assess not only drug metabolism but also microbiome‐mediated modulation of liver function. For instance, several studies have applied microbiome metabolites or conditioned media from gut commensals such as B. fragilis and Clostridium species to these 3D hepatic systems to evaluate their roles in hepatic inflammation and metabolism.^{[ 304 ]}

Together, these scaffold‐based 3D liver models offer powerful platforms to dissect microbiome‐driven mechanisms in HCC and to evaluate microbiome‐targeted therapies in preclinical research.

Despite the immune‐privileged status of the central nervous system (CNS) and the protective function of the blood‐brain barrier (BBB), the systemic influence of the gut microbiome on brain tumor progression is increasingly recognized.^{[ 305} , ^{306 ]} Microbiome‐derived metabolites, such as SCFAs, indole derivatives, and secondary bile acids, as well as immune signals like IL‐1β and IFN‐γ, can modulate BBB permeability, neuroinflammation, and glioma invasion.^{[ 307} , ³⁰⁸ , ^{309 ]} These signals reach the brain either by crossing the BBB or by altering peripheral immune cells that interact with the brain TME.^{[ 310 ]}

To study these complex interactions, researchers have developed advanced 3D in vitro models, particularly BBB‐on‐a‐chip systems and tumor spheroids. These platforms combine human‐derived endothelial cells, pericytes, astrocytes, and tumor cells under perfused conditions to recreate the semi‐permeable vascular interface of the BBB and simulate key features of the brain TME, such as hypoxia, drug penetration barriers, and immune cell trafficking.^{[ 305} , ³¹¹ , ^{312 ]} Notably, they allow controlled introduction of microbiome metabolites and immune modulators, enabling mechanistic investigations into how the microbiome influences tumor progression and therapeutic resistance.^{[ 307} , ^{308 ]}

One prominent application of OoC technology is to evaluate tumor cell transmigration and BBB remodeling during metastasis and drug delivery.^{[ 313} , ^{314 ]} These systems have revealed how metastatic tumor cells compromise BBB integrity, and how microbiome immune interactions further exacerbate BBB disruption and immune cell infiltration.^{[ 309 ]} For example, Ivanov et al. developed a co‐culture model that integrated medulloblastoma spheroids with human neural stem cells (hNSCs), demonstrating enhanced drug resistance and long‐term viability, which are key features for assessing how microbiome signals may modulate therapeutic efficacy (Figure 8 B).^{[ 315 ]}

Moreover, brain organoids and BBB‐chip platforms exposed to microbiome secretomes or microbiome‐conditioned immune cells have shown alterations in the activation state of microglia and tumor‐associated macrophages, as well as changes in BBB permeability and inflammatory signaling.^{[ 308 ]} These findings support the idea that both pediatric and adult brain tumors are susceptible to microbiome‐mediated modulation, especially through gut–brain axis signaling.

Looking forward, computational modeling tools are enhancing these platforms by simulating metabolite diffusion and drug transport across the BBB, offering new predictive capabilities for personalized medicine.^{[ 306 ]} Altogether, these integrated technologies provide a powerful framework for dissecting the microbiome‐immune‐brain tumor axis and hold promise for developing microbiome‐targeted therapies and improving treatment strategies for glioblastoma and other CNS malignancies.

4. Future Perspectives and Challenges

4.1. Matched Patient‐Derived 3D Cancer‐Microbiome‐Immune Platforms for Clinical Translation

Using matched or paired patient‐derived samples‐tumor tissue, autologous microbiome communities, and PBMCs to build personalized 3D model systems represents a promising future direction in TME modelling. These “holistic” ex vivo platforms preserve each patient's unique tumor‐immune‐microbiome interactions, allowing functional testing under conditions that mimic physiology.^{[ 316} , ^{317 ]}

At the individual patient level, these models could guide personalized therapy by testing both standard‐of‐care and experimental treatments, including repurposed small molecules, personalized immunotherapies such as tumor‐infiltrating lymphocytes (TIL), chimeric antigen receptor (CAR) T cells, therapeutic cancer vaccines, and antibody‐drug conjugates (ADCs). By monitoring cytotoxicity, cytokine release, immune infiltration, and microbiome changes, clinicians could identify the most effective therapy for each patient before starting in vivo treatment.

From a discovery standpoint, detailed molecular profiling of these matched systems could reveal biomarkers that predict response or resistance, ranging from bacteria and metabolite signatures^{[ 318 ]} to immune checkpoint expression patterns.^{[ 319 ]} Such biomarkers could assist in patient stratification in clinical trials and support the development of companion diagnostics.

In drug development, pharmaceutical pipelines could utilize these platforms for lead compound selection and indication prioritization. For example, when multiple drug candidates are available, patient‐matched models could help identify the compound with the most favorable tumor‐immune‐microbiome activity profile. Likewise, testing across various patient‐derived tumor types could pinpoint the best cancer indication for first‐in‐human trials, reducing risks in early‐stage clinical development.

By combining advances in precision oncology, immunotherapy, and microbiome research, matched patient‐derived 3D cancer‐microbiome‐immune platforms offer a versatile and translational approach for next‐generation cancer care.

4.2. Challenges in Standardization and Reproducibility

While the models discussed highlight the remarkable potential of 3D systems to mimic host‐microbiome interactions, their translation into robust and broadly adaptable platforms is still hindered by significant challenges, with reproducibility being at the forefront; despite impressive advances, variations in culture conditions, biomaterials, and bacterial handling protocols often result in outcomes that are difficult to compare across laboratories. A well‐known example is the widespread reliance on Matrigel, as discussed previously, which introduces lot‐to‐lot variability and complicates benchmarking.^{[ 320 ]}

Parallel efforts have focused on the standardization of microbiome co‐culture protocols. For instance, Nature Protocols published a stepwise framework for intestinal organoid‐microbe co‐culture that details methods such as lumen microinjection and apical exposure, while also incorporating quality‐control checkpoints: “We provide detailed protocols for characterizing the coculture with regard to bacterial and organoid cell viability and growth kinetics”.^{[ 321 ]} Such standardized guidance is crucial for controlling variables like multiplicity of infection, oxygen exposure, and co‐culture duration, all of which profoundly affect biological readouts. Complementary to this, JoVE has contributed visualized protocols for organoid‐pathogen and gut‐on‐a‐chip systems,^{[ 107} , ³²² , ^{323 ]} which provide bill‐of‐materials, timing, and handling details that help minimize user‐specific variation.

Reviews increasingly stress the importance of the specification of technical parameters, from matrix composition and media sourcing to handling timings and analysis pipelines. Journals such as Nature Protocols have begun to set the standard by including detailed reagent lists and reporting summaries, which can serve as templates for harmonized practice.^{[ 180} , ³²⁰ , ^{321 ]}

Taken together, these efforts highlight that improving reproducibility will require both technological innovations, such as chemically defined matrices and automated devices, and cultural shifts toward more rigorous and transparent reporting. As methods become more standardized and accessible through detailed written and visual protocols, the ability to compare findings across laboratories will strengthen, accelerating the translation of 3D host‐microbiome models into both basic research and clinical applications.

5. Conclusion

Understanding the multifaceted interactions between tumors and the microbiome requires patho‐physiologically relevant models that replicate the dynamic and spatial complexity of the TME. Over the past decade, advanced 3D in vitro models, including spheroids, organoids, microfluidic systems, and 3D‐bioprinted constructs, have transformed our ability to study tumor–microbiome interactions across multiple cancer types. These platforms allow for spatial control, compartmentalized co‐cultures, dynamic mechanical cues, and host‐microbe metabolic exchange, enabling more accurate interrogation of how bacterial factors influence tumor progression, immune modulation, and therapy resistance.

This review highlights how tumor‐type‐specific models are revealing new insights into the role of microbiome metabolites, immune signaling, and matrix composition in cancer. The incorporation of tissue‐relevant ECM components, such as collagen I and laminin, as well as the use of scaffold‐based systems that simulate the gut‐liver or gut–brain axis, are pushing the frontier of microbiome‐TME modeling. In parallel, 3D‐bioprinted systems have demonstrated advantages in tailoring architecture, integrating stromal components, and enabling high‐throughput screening of personalized cancer constructs, with recent studies beginning to co‐culture live bacteria or integrate microbiome metabolites into these platforms.

However, despite these advances, several limitations persist. Maintaining long‐term bacterial viability without overgrowth, reproducing physiological oxygen and nutrient gradients, and co‐culturing diverse immune and bacterial cell types remain technical challenges. Additionally, many models lack standardization, exhibit limited scalability, and struggle to mimic temporal evolution, a core aspect of tumor biology. Incorporating microbiome complexity, such as anaerobic conditions, community dynamics, and metabolite flux, into cancer models remains an ongoing hurdle.

To address these challenges, 4D‐bioprinting has introduced the fourth dimension‐time‐as a functional design parameter. This allows constructs to respond to stimuli, such as microbiome metabolites, cytokines, or enzymatic activity, through smart materials that morph, release drugs, or remodel ECM. Looking ahead, a major advancement would involve the integration of personalized patient data, including genomics, proteomics, and metabolomics, into the fabrication process, enabling constructs that mirror an individual tumor's molecular fingerprint.^{[ 172} , ¹⁷⁵ , ³²⁴ , ^{325 ]} Additionally, integration of machine learning, real‐time biosensing, and robotic automation allows adaptation of bioinks, geometry, and print parameters in response to environmental changes such as pH, oxygen, or cytokine levels.^{[ 172} , ¹⁷⁵ , ³²⁵ , ^{326 ]}

Together, these computationally augmented platforms open the door to fully personalized, adaptive tumor avatars that evolve with disease progression, enabling more predictive therapeutic testing, refined immuno‐oncology strategies, and real‐time patient stratification. While still in early development, they offer a powerful path toward next‐generation cancer models that can capture both the temporal plasticity of cancer and its interpatient heterogeneity.

As the field advances, integrating bioengineering, artificial intelligence, and systems biology will be essential for building holistic tumor‐microbiome models that not only replicate but respond to biological complexity, ushering in a new era of precision oncology and microbiome‐informed therapy design.

Conflict of Interest

R.S.‐F. is a board director at Teva Pharmaceutical Industries Ltd. and receives unrelated research funding from Merck KGaA. R.S.‐F. is a cofounder and officer with an equity interest in Selectin Therapeutics Inc. All other authors declare no conflict of interest.

Biographies

Marina Green Buzhor is a Research Associate in the Department of Physiology and Pharmacology at Tel Aviv University, in the laboratory of Prof. Ronit Satchi‐Fainaro. She earned her PhD in Organic Chemistry from Tel Aviv University, focusing on enzyme‐responsive polymeric nanocarriers. She then completed a postdoctoral fellowship at ETH Zurich in the Institute of Pharmaceutical Sciences, where she developed 3D‐printed capsules for oral delivery of live bacteria. Her current research integrates tissue engineering, drug delivery, and microbiome science to create advanced 3D cancer models that elucidate tumor‐microbiome interactions.

Giuseppe Longobardi is a Postdoctoral Research Fellow in the Department of Physiology and Pharmacology at Tel Aviv University, in the laboratory of Prof. Ronit Satchi‐Fainaro. He earned his PhD in Pharmaceutical Sciences from the University of Naples Federico II. His research focuses on advancing cancer therapies through nanomedicine, aiming to improve precision and reduce side effects compared to conventional treatments. Bridging pharmaceutical formulation and biological insights, he develops 3D tumor models that replicate the tumor microenvironment to better understand cancer progression and evaluate therapeutic outcomes.

Or Kandli is an M.Sc. student at the Gray School of Medical Sciences at Tel Aviv University, supervised by Prof. Ronit Satchi‐Fainaro. She holds a B.Sc. in Chemistry with a specialization in organic chemistry. Her research focuses on developing 3D‐bioprinted capsules for the co‐delivery of live bacteria and small‐molecule immunotherapies. This platform is designed to modulate the gut microbiome and immune system, thereby improving the efficacy of immunotherapies. By integrating material engineering, microbiology, and drug delivery, she aims to establish a versatile system that could also be applied to other types of cancer.

Anne Krinsky is an MD‐PhD student at Tel Aviv University, under the supervision of Professor Ronit Satchi‐Fainaro. Her research focuses on developing 3D‐bioprinted cancer models for target discovery and personalized therapy for gastrointestinal patients. Beyond her studies and research, she is actively involved in advancing HealthTech in Israel. She has been leading volunteering efforts for the 8200Bio Association, collaborating between healthcare companies and organizations to strengthen the local HealthTech ecosystem. She is also a co‐founder of Nucleate's Israeli branch, a global organization promoting life science entrepreneurship, and she has been leading the Tel Aviv University branch.

Opal Avramoff is a direct‐track Ph.D. student in Prof. Ronit Satchi‐Fainaro's laboratory at Tel Aviv University. She holds a B.Sc. in Medical Laboratory Sciences from Ben‐Gurion University of the Negev and has been awarded several fellowships, including the TEVA BioInnovation Fellowship. Her research focuses on an ongoing clinical trial at Sheba Medical Center, where she develops patient‐specific drug testing platforms. These model platforms closely recapitulate clinical responses, advancing patient‐specific therapies, drug discovery, and bridging preclinical research with clinical applications in glioblastoma.

Anshika Katyal completed her integrated B.Tech.M.Sc. in Biotechnology at Thapar Institute of Engineering and Technology and Tel Aviv University (2021) and is currently pursuing her Ph.D. at Tel Aviv University under the supervision of Prof. Ronit Satchi‐Fainaro. Her research focuses on developing patient‐derived 3D‐bioprinted perfusable cancer models integrated with AI‐driven analytics to transform drug discovery and bridge the gap between preclinical models and clinical translation, revolutionizing personalized medicine. She collaborates with leading partners across academia, hospitals, and industry.

Koren Salomon is a biologist and PhD student in Prof. Ronit Satchi‐Fainaro's laboratory at Tel Aviv University. His research focuses on lung cancer brain metastasis, with emphasis on tumor‐microglia interactions. Using advanced 3D models, spatial imaging, and multi‐omics approaches, he investigates mechanisms that drive metastatic progression and explores therapeutic strategies to target the brain tumor microenvironment. His interdisciplinary work combines molecular biology, immunology, and computational analysis to uncover new insights into cancer biology.

Adan Miari is a Ph.D. candidate in Medical Sciences in Professor Ronit Satchi‐Fainaro's laboratory at Tel Aviv University, where she investigates neuro‐oncology. Her research focuses on the interactions between high‐grade glioma cells and neurons, with a particular emphasis on pediatric brain tumors. She earned both her B.Sc. and M.Sc. in Pharmacology and Physiology at Tel Aviv University. Hands‐on experience at Tikcro Technologies and BiomX, combined with her M.Sc. research on brain metastases, has provided her with strong expertise in molecular and cellular biology and fuels her drive to advance cancer therapy.

Dana Venkert is a graduate student in the Adi Lautman Interdisciplinary Program at Tel Aviv University, pursuing her M.Sc. in the Sagol School of Neuroscience. Her research in Prof. Ronit Satchi‐Fainaro's laboratory focuses on developing 3D organoid‐based models of the bloodbrain barrier to study cancer immunotherapy. She has conducted research at LMU Munich, Helmholtz Pioneer Campus, and Goethe University Frankfurt on cancer dormancy, regenerative medicine, and vascular modeling. She is a Forbes Israel 30 Under 30 honoree and recipient of multiple national and international awards.

Tania T. Barnatan is a Ph.D. student in Prof. Ronit Satchi‐Fainaro's lab. Tania is interested in BRCA1‐mutated ovarian cancer, and, more specifically, in identifying different avenues of inducing synthetic lethality and how it can change the tumor microenvironment. Previously, Tania completed an M.Sc., where she studied RNA‐binding proteins and their role in mammalian autophagy. Prior to this, Tania has been involved in translational research since she began her Bachelor's in 2016 at Penn State University.

América García Alvarado is an M.Sc. student in Medical Sciences at Tel Aviv University under the supervision of Prof. Ronit Satchi‐Fainaro. Her research focuses on P‐selectin‐targeted PLGA nanoparticles for pediatric low‐grade gliomas, using 3D microfluidic tumor models. She holds a B.Sc. in Chemistry and Nanotechnology Engineering from Tecnológico de Monterrey, where she developed amino acid‐based pharmaceutical formulations for pregnancy‐associated diseases.

Shahar Greenberg is a master's student in Prof. Ronit Satchi‐Fainaro's laboratory at Tel Aviv University. She joined her M.Sc. studies through the ELITE excellence program, which enabled her to accelerate her B.Sc. degree and transition early into graduate research. Her work focuses on breast cancer brain metastasis, investigating molecular mechanisms driving this process. She explores therapeutic strategies using RIBOTACs and ProTACs molecules to target key regulators of metastasis and develops 3D chip‐based platforms to test these interventions, aiming to identify novel approaches to prevent or disrupt the formation of brain metastases.

Ronit Satchi‐Fainaro is a Full Professor at Tel Aviv University, where she is Head of the Gray School of Medical Sciences, Director of the Cancer Research & Nanomedicine Laboratory, and the TAU Kahn 3D BioPrinting Initiative. She also serves as the Chair of the Cancer Biology Research Center and holds the Kurt and Herman Lion Chair in Nanosciences and Nanotechnologies. She previously served as Chair of the Department of Physiology & Pharmacology and as President of the Israeli Controlled Release Society. She is a board director at Teva Pharmaceutical Industries Ltd. and cofounder of Selectin Therapeutics Inc.

Green Buzhor M., Longobardi G., Kandli O., et al. “Harnessing Next‐Generation 3D Cancer Models to Elucidate Tumor‐Microbiome Crosstalk.” Adv. Healthcare Mater. 15, no. 1 (2026): e03198. 10.1002/adhm.202503198