Although the contribution of the microbiome to the development of cancer has long been proposed, recent evidence has demonstrated that the microbiota plays multifaceted roles in cancer development, progression, and treatment (Fu et al., 2023; Zhao et al., 2023; Battaglia et al., 2024). These recent discoveries have sparked significant interest in understanding the type, localization, mechanisms of action, and interactions of different types of microorganisms with the tumor microenvironment (TME) and its components (Lythgoe et al., 2022; Li et al., 2024). Many tumor-resident microbes (TRMi) are opportunistic, facultative anaerobes and intracellular pathogens that induce DNA damage (Espinoza and Minami, 2018; Gu X. et al., 2020), produce genotoxins (Cao et al., 2022; Joo et al., 2024), and cause inflammation (Lu et al., 2024; Zhang et al., 2024). TRMi can inactivate anti-cancer drugs and modulate the immune system, leading to poorer prognoses for therapies (Jiang et al., 2023; Sfanos, 2023; Kang et al., 2024). The recent reports of the involvement of certain microorganisms with some tumors such as colorectal cancer (Zepeda-Rivera et al., 2024) has attracted the attention of health authorities and clinicians, urging elucidation of their roles in the induction of malignant processes, dissecting intra- and extracellular effects, and clinical outcomes.

A substantial body of work has demonstrated, through comprehensive 16S rRNA analysis on several solid tumor types (Nejman et al., 2020) and spatial profiling combined with RNA sequencing (Galeano Nino et al., 2022), that TRMi interact with cancer cells, immune cells, and the extracellular matrix. This leads to the recruitment of immune cells and transcriptional changes in cancer cells and cancer-associated fibroblasts (Galeano Nino et al., 2022; Narunsky-Haziza et al., 2022; Wong-Rolle et al., 2022; Zepeda-Rivera et al., 2024). However, elucidating the exact mechanistic events linking the presence of microorganisms and activity with cancer is challenging due to the need for reliable pre-clinical models that emulate patient conditions (Rodrigues et al., 2021;Vitale et al., 2024). In vivo models of host-microbiota interactions remain the most reliable tools for understanding cancer biology (Ladaycia et al., 2021; Rodrigues et al., 2021; Li G. et al., 2023), and are well-established in the literature. Nevertheless, these models cannot accurately capture the extensive heterogeneity of native microorganisms observed in the human microbiome, which is shaped by unique genetic and immune composition (Aizpurua et al., 2023; Cao et al., 2024). Furthermore, traditional cell culture and co-cultures of mammalian cells and microorganisms often inadequately represent the spatial organization and ratios inherent in the tumor microenvironment (Asghar et al., 2015; Kapalczynska et al., 2018).

Alternative models such as organotypic cultures that combine engineering and biology provide a real-time simulation of critical elements of the interactions between TRMi and the TME in vitro (Radhakrishnan et al., 2020; Neufeld et al., 2022). These systems include 3D extracellular matrices (Yang et al., 2024), vasculature (Zervantonakis et al., 2012), and the interplay among stromal, immune, and cancer cells with microorganisms (Jouybar et al., 2024; Shin et al., 2024). Tumors-on-a-chip, for instance, emerged as attractive in vitro models that can tackle several challenges associated with accurately replicating the unique properties of the tumor microenvironment (Sontheimer-Phelps et al., 2019; Maulana et al., 2024). These microdevices incorporate microfluidic channels and manipulable fluid flow to model multiple hallmarks of the disease, including shear stress (Ortega Quesada et al., 2024), diverse cell types, an extracellular matrix, control of hypoxic gradients, and metabolic exchange between microorganisms and mammalian cells (Shah et al., 2016; Thacker et al., 2020; Shin et al., 2024), among other features. These platforms may enable scientists to decipher the interfaces between TRMi and different phases of cancer development by modulating patient-derived tissues in chips and capturing critical biomechanical elements of the TME.

In this review, we describe opportunities and challenges associated with using tumor-on-a-chip models to investigate interactions between the TME and TRMi. We discuss the requirements for designing chips that represent the specificities of the TME and exemplify some applications of these platforms in tumors directly known to harbor microorganisms.

Despite using microfluidic devices to study tumor-microbe interactions represents a relatively new development, the relationship between microorganisms and healthy tissues has been modeled for at least 20 years (Liu and Zhu, 2005; Kim and Sung, 2024). In the organs on-a-chip field, the gut was one of the first targets for chip modeling (Huh et al., 2013), and gut-on-a-chip devices have become very popular among researchers (Kim and Sung, 2024). These models have been improving in incorporating characteristics of native tissues and their interfaces, facilitating the growth of native-like microbial species (Ballerini et al., 2025). As a result, they have significantly enhanced the diversity and heterogeneity of microbial models, enabling the inclusion of bacteria, fungi (Srinivasan et al., 2011), viruses (Bein et al., 2021), and parasites (Nikolaev et al., 2020). Such advancements are critical for studying microbial species that may directly or indirectly contribute to dysbiosis, inflammation, and tumor establishment (Figure 1).

In the context of tumor initiation, microfluidic technologies hold significant potential to elucidate the causative mechanisms underlying the association between dysbiosis on epithelial barriers and cancer development (Poceviciute and Ismagilov, 2019). However, reproducing multispecies cultures in vitro presents challenges due to the complex interactions of these microorganisms on their native tissue, which occur under varying oxygen conditions (e.g., skin, colorectal, and oral cavities) and dynamic metabolic interactions (Moossavi et al., 2022). Static systems can promote microorganism overgrowth, which may overwhelm the rate of mammalian/microorganism cell interactions (Jalili-Firoozinezhad et al., 2019; Shin et al., 2024). Microfluidic devices can emulate tight junction formation, epithelial layers, mucus, and peristaltic movements related to shear stress to adopt microbial diversity and interactions with epithelial cells during early cancer development (Sibilio et al., 2024). These devices also serve as excellent platforms for studying bacterial-carcinogenic signaling pathways (Valiei et al., 2023). For instance, a PDMS chip was designed to mimic the gut, featuring multiple rows of microgut culture chambers (Lee et al., 2024). The microgut was generated using Caco cell layers in collagen gels, creating a villus structure subjected to shear stress, hypoxia (1% oxygen to facilitate bacterial growth for 1 day), and flow to support epithelial tissue formation. After 6 days, the monolayer exhibited a well-differentiated brush border, various intestinal cells (including goblet cells, enterocytes, and Paneth cells), defined tight junctions, and mucin production. The system supported the growth of both enterotoxigenic (ETBF8 and ETBF9) and non-toxigenic (NTBF) strains of Bacteroides fragilis. In this model, only the ETBF9 strain significantly stimulated pre-oncogenic signaling pathways, such as WNT-β-catenin and p-STAT3, compared to the microorganism-free group. Introducing Lactobacillus rhamnosus into the microgut significantly inhibited ETBF9 growth. These findings were previously observed only in vivo and demonstrate that microfluidic devices can replicate substantial biomimetic physiological and pre-carcinogenic events (Owens et al., 2021).

Another important area of exploration in recent years is the crosstalk between microbiota and the immune system in cancer (Jayakrishnan et al., 2024). Dysbiosis can lead to pathobiont proliferation and inflammation (Biragyn and Ferrucci, 2018). Opportunistic microorganisms, such as F. nucleatum and Candida albicans, proliferate under dysbiosis conditions (Gu P. et al., 2020; Li et al., 2022). Their interaction with epithelial and cancer cells triggers inflammatory cytokines (e.g., interleukins IL-6, IL-17, IL-8) and pathways (e.g., NF-kB, STAT3), promoting epithelial-to-mesenchymal transition, immune chemoattractants, and cancer cell proliferation (Narunsky-Haziza et al., 2022; Wang and Fang, 2023; Wang et al., 2023b). Additionally, microorganisms interact with the immune system in cancer through toll-like receptors (TLR2 and TLR4) and myeloid-derived suppressor cells, triggering an immunosuppressive environment (Yang et al., 2017; Wong and Yu, 2023). Some of these interactions have therapeutic consequences, especially in immunotherapies. Recent evidence has shown that F. nucleatum can produce succinic acid which neutralizes the effect of anti-PD1 antibodies (Jiang et al., 2023).

Combining all elements of the epithelium, the microbiome, and cancer cells together with immune cells in vitro is challenging. However, microfluidic devices have recapitulated some of these complexities by providing an architecture that supports immune cell infiltration, differentiation, and communication with microorganisms, which can be directly applied to the TME (Figures 1A–D). Researchers have developed a microfluidic device to understand the immune-responsive human microbiota-intestine axis (De Gregorio et al., 2022) (Figures 1E–G). This device features an open central chamber to accommodate a trans-well membrane separating the lumen and serosal sides. This chamber is housed within a poly(methyl methacrylate)-derived cap to provide N2 flow for an anaerobic luminal environment to support the growth of L. rhamnosus and Bifidobacterium longum. The microorganisms studied exhibited a protective immunoregulatory effect in the presence of lipopolysaccharide (LPS, derived from E. coli). Peripheral blood mononuclear cells (PBMCs) introduced to the chip produced higher levels of anti-inflammatory cytokines (such as IL-10) and lower levels of inflammatory cytokines (such as IL-2), demonstrating that the immunomodulatory role of probiotics can be recapitulated in complex models (De Gregorio et al., 2022) (Figures 1E–G).

Considering the immune system’s complexity, which requires antigen presentation and interaction with other systems, such as the circulatory and lymphatic systems, leaves several opportunities remain to explore host-microbe interactions in cancer. However, as blood and lymphatic vessels can be engineered in organs-on-a-chip, integrating multiple aspects of the microbiome, epithelial tissues, and immune cells with the vasculature, will provide robust tools to predict cancer development and progression.

Due to the diversity and heterogeneity of microbes across various tissues, there is a compelling need to simulate their interactions within specific physiological environments in the human body (Shin et al., 2024). For instance, the gastrointestinal tract and oral cavity can be inhabited by anaerobic pathobionts (Zhao et al., 2023; Liao et al., 2024), while the mammary glands can harbor aerobic species (Tzeng et al., 2021). Furthermore, numerous uncertainties persist regarding how local and tissue-specific microbes might influence the initiation, progression, and resistance of distant tissue tumors (Fu et al., 2023). To address many of these biological questions, organ-on-a-chip technologies enable the investigation of microbe interactions with specific tissues (such as the tumor microenvironment) and their spread to other tissues (metastasis). In this section, we elucidate the applications of these advanced models in studying various cancer types, such as colorectal, breast, and head and neck cancers (Figure 2).

The microbiome is relevant to most, if not all, cancers, beyond those outlined here, including malignancies of the skin, lung, urogenital system, and bone marrow. Therefore, there is a need to develop models that can mimic the physiology of these tissues and their connections with the microbiome (Ramirez-Labrada et al., 2020; Martin et al., 2022; Voigt et al., 2022; Woerner et al., 2022). Despite documentation of interactions between the microbiome and cancer development, many opportunities remain to further explore the potential of organ on-a-chip modeling within the tumor microenvironment, such as their mechanistic interactions with the immune system in the TME, immunosuppression, metastasis, and immunotherapies. Furthermore, the connection of tissue-specific microbiota with distant cancers and metastasis should be explored, primarily because the gut microbiome has been correlated with metastasis to several tissues (Fu et al., 2023).

While the outlook for host-microbes on-a-chip models in cancer studies is promising, several challenges are still to be overcome. Current models often struggle to capture the full complexity of the microbiome, including the diversity and density of microorganisms that accurately reflect local and systemic niches. Variability in environmental factors such as temperature, oxygen levels, growth media, and shear stress further complicates the standardization of experimental conditions, reducing reproducibility and comparability across studies. Furthermore, achieving the precise organization of tissues, including accurate cell-cell interactions and extracellular matrix composition, remains a significant technical challenge. Finally, the integration of advanced technologies, such as artificial intelligence and high-throughput systems, is still in its early stages, limiting the predictive power and scalability of these platforms.

Overcoming these challenges will require a multi-disciplinary approach that integrates expertise in tissue engineering, microbiology, and computational sciences. Addressing the need for microbiome diversity, optimizing growth conditions for heterogeneous microorganisms, and refining tissue engineering techniques are essential steps toward enhancing the fidelity and translational of these models. Additionally, incorporating cutting-edge technologies will not only improve data analysis but also enable the development of more robust and scalable systems. Once these limitations are addressed, organ-on-a-chip models have the potential to transform our understanding of microorganisms in cancer and develop specific immune and microbiological therapies. These platforms could serve as “live biopsies” or personalized tissue avatars, allowing clinicians to evaluate patient-specific responses to treatments in real-time or to assess microbiota profiles and correlate them with disease progression. Also, these platforms can be used to test engineered microorganisms by synthetic biology to compete with TRMis and attract, polarize, and stimulate tumor-killing immune cells. Such applications will generate several opportunities for more precise therapeutic interventions, ultimately improving treatment outcomes and enhancing the quality of life for cancer patients.

In conclusion, while the path to fully realizing the potential of organ-on-a-chip models in microbes-cancer research is linked with technical and biological challenges, their promise opens multiple doors to answer unique biological questions. These innovative platforms represent a significant step forward in understanding and treating cancer through a microbiome-informed perspective, offering a future where precision medicine can address the complexities of cancer biology accurately.