Advances in engineering immune–tumor microenvironments on-a-chip: integrative microfluidic platforms for immunotherapy and drug discovery

Abstract

The design and application of microfluidic immune system-on-a-chip (ISOC) technology have played a critical role in cancer immunology and drug discovery over the past decades. The system provides a highly controlled and physiologically relevant platform for studying immune responses and therapeutic interventions. Emerging trends in 3D bioprinting, organoid fusion, and multi-organ systems-on-a-chip further expand the capabilities of ISOC by enabling systemic immune interactions and modeling of the tumor microenvironment. Despite these advances, scalability, standardization, and long-term immune cell viability remain significant challenges that must be addressed to fully harness the potential of ISOC in clinical applications. Overall, ISOC represents a transformative tool in cancer research, offering innovative solutions for immunotherapy trials, biomarker discovery, and precision medicine. Therefore, in this study, the role of ISOC in cancer immunotherapy was investigated, focusing on its ability to recapitulate primary and secondary immune functions, model immune-tumor interactions, and enhance screening and optimization of immune-based therapies. Device design and modeling strategies were also discussed, demonstrating how ISOC platforms simulate dynamic immune cell activity, cytokine signaling, and antigen presentation to improve drug efficacy assessments. The application of ISOC technology in drug discovery and its potential to accelerate clinical trials and develop personalized immunotherapy were further explored.

Graphical Abstract

Introduction

Researchers often use in vitro (cell models) and in vivo (animal models) studies during drug design, development, and testing of drugs and immune system function in the face of cancer [1–4]. However, these techniques have limitations in accurately predicting human disease pathways, understanding personalized drug sensitivities in specific patient groups, and assessing off-target drug toxicity [5]. Adverse drug reactions, such as cardiac, hepatic, and renal toxicities, can occur, highlighting the inefficiencies and uncertainties surrounding drug mechanisms of action [6]. To address these challenges, innovative methods and approaches are essential for drug discovery and health research. One potential solution is microfluidic systems, which leverage advancements in tissue engineering and microfluidics and improved techniques for extracting, culturing, and maturing human cells [6] (Fig. 1).

Fig. 1.

In vivo, ex vivo, and in vitro models have disadvantages. In vivo models are challenged by non-human physiology, complexity, and primarily phenotypic data. Ex vivo models are limited by non-systemic approaches, short life spans, and reliance on specific conditions. In vitro models often oversimplify biological systems, lack systemic interactions, and have restricted immune components. On the other hand, microfluidic models offer advantages such as precision, low cost, high throughput, integration, and versatility. However, when using these systems, paying special attention to factors such as complexity, scalability, interfacial effects, clogging issues, and a limited viscosity range is essential. Figure generated in BioRender (BioRender.com)

The primary aim of microfluidic systems technology, encompassing organ-on-a-chip and cell-on-a-chip developments, is to accurately emulate the intricate features of specialized microenvironments and tissue architectures on a miniaturized platform [7]. These microsystems are engineered to replicate the complex physiological interactions and conditions of living organisms through sophisticated design methodologies. This innovation fosters the advanced modeling of biological processes and disease mechanisms, facilitating more profound insights into pathophysiology and improving drug development and testing [8]. Immune system cells on a chip have been used to predict, prevent, and treat various diseases, including cancer, by investigating processes such as angiogenesis, invasion, migration, and adhesion [9]. Microfluidic chips offer significant advantages as they can simulate multi-organ systems by integrating multiple modules. Additionally, they are compatible with online analytical methods, allowing for real-time monitoring of organoid status [10, 11].

Given the rising costs associated with drug discovery and the limited predictability of current models, human-on-a-chip or multi-organ systems have gained increasing interest as models for drug discovery [12, 13]. These systems can examine potential drugs’ pharmacodynamics and pharmacokinetics, reduce costs, and enhance efficiency throughout development [14]. Human-on-a-chip platforms for drug discovery aim to replicate the human immune system, including circulating immune cells, to more accurately simulate immune surveillance processes. When circulating immune components are omitted, potential immunosuppressive effects from mild inflammation to cytokine storm cannot be assessed [15–17]. Collecting immune response data alongside functional data and non-invasive biomarkers can significantly improve the scalability of human-on-a-chip systems, enhancing their applicability for both broad and targeted drug discovery initiatives [16].

Cancer is understood as a systemic disease, and tumor immunology has focused significantly on local immune responses within the tumor microenvironment (TME) [18–23]. Tumors have a complex microenvironment consisting of a dense extracellular matrix (ECM), diverse stromal and immune cells, irregular blood vessels, and limited nutrient perfusion, each of which can significantly impact the effectiveness of treatments [24, 25]. The development of cancer organoids has led to the creation of organoid banks to explore cancer transformations [26, 27]. Combining various physiological components, such as blood vessels, allows cancer-on-a-chip models to examine the relationships between cancer and other organs thoroughly [26, 28].

In recent years, cancer-on-a-chip has emerged as an innovative tool in cancer research, offering an advanced alternative to traditional animal models and in vitro cell culture systems [29]. This technology is based on microfluidic platforms specifically designed to simulate the tumor microenvironment, enabling more accurate and scalable models for studying cancer biology, tumor progression, and therapeutic responses. These advanced systems integrate biological cells, biomaterials, and microfluidic technology to create a dynamic, miniaturized platform that replicates the complex biological processes of human tumors [29].

Historically, cancer research and drug development have heavily relied on animal models, which, despite their value, come with significant limitations, including species-specific differences, high costs, and ethical concerns [30]. In contrast, cancer-on-a-chip models utilize human cancer cells, providing a more precise representation of human disease. These microfluidic chips create an environment in which cancer cells can grow, interact with one another, and respond to stimuli in ways that closely mimic in vivo conditions. Cancer-on-a-chip systems typically consist of a network of micro-channels where tumor cells are exposed to various physiological conditions such as fluid flow, mechanical stress, and nutrient gradients. This setup allows researchers to study cellular behavior in response to factors like hypoxia, ECM components, and drug treatments, and to simulate critical aspects of tumor physiology such as metastasis, angiogenesis, and drug resistance [30].

The applications of cancer-on-a-chip technology are vast, including drug screening, personalized medicine, and the development of novel cancer therapies [31]. By providing a more accurate representation of tumor behavior and drug responses, this technology has the potential to accelerate drug discovery, reduce reliance on animal testing, and lead to more effective, personalized cancer treatments. Moreover, it holds promise for advancing our understanding of cancer heterogeneity, enabling researchers to study the differences between tumor subtypes and the factors that drive metastatic potential [31].

However, despite its immense potential, cancer-on-a-chip faces several challenges [32]. These include the complexity of fully recapitulating tumor biology, integrating multiple cell types to simulate tumor-stroma interactions, and scaling these systems for high-throughput drug screening. Nevertheless, with ongoing advancements in microfluidic technology, biomaterials, and cellular biology, the future of Cancer-on-a-Chip is promising and could revolutionize cancer research and therapy [32].

Consequently, this study undertook a detailed exploration of the pivotal role of immune system-on-a-chip (ISOC) in cancer immunotherapy, focusing on its remarkable ability to replicate both primary and secondary immune functions. It aimed to model the intricate interactions between the immune system and tumors while also seeking to enhance the screening and optimization processes for immune-based therapies. The research delved into innovative device design and sophisticated modeling strategies, illustrating how ISOC platforms can effectively simulate the dynamic activities of immune cells, the intricate signaling pathways of cytokines, and the critical process of antigen presentation, all of which are essential for improving assessments of drug efficacy. Furthermore, the study investigated the application of ISOC technology in the drug discovery landscape, highlighting its promising potential to not only accelerate the timeline of clinical trials but also to pave the way for the development of personalized immunotherapy tailored to individual patient needs.

Classifying the immune system: distinguishing innate and adaptive responses

The human immune system is intricately designed to defend the body against a wide variety of threats, including pathogens, toxic biological agents, and environmental factors [33]. Its core function is based on rapid, specific and protective responses to ensure the survival and health of humans. The outcome of these immune processes is the resolution of infectious diseases, the detection and elimination of tumors, the rejection of transplanted tissues or organs, the regulation of autoimmune disorders, and allergies [33]. The effectiveness of the immune response is not only dependent on the function of individual immune cells but also on the coordination and regulation between the innate and adaptive systems. This complex interplay ensures that the body can mount an appropriate defense against a vast array of threats while maintaining tolerance to its own cells and tissues. The immune system operates through a sophisticated network of cells, molecules, and signaling pathways, each with a specialized function. The integration of innate and adaptive responses allows for both immediate and long-term defense against pathogens and the ability to remember previous encounters with antigens forms the foundation of vaccination strategies [33].

The immune system can be divided into innate and adaptive components (Fig. 2) [34]. Innate lymphoid cells (ILCs) are a group of immune cells that lack antigen receptors and are distributed during early development, migrating to non-lymphoid tissues to reside. These cells play critical roles in immune defense and are also involved in tissue growth and regeneration processes. Innate immune cells are categorized into natural killer (NK) cells, ILC1, ILC2, ILC3, and ILCreg groups. In addition to ILCs, certain B and T cells are also considered part of the innate immune system [34]. These include B cells such as B1a and B1b cells, marginal-zone B cells (which derive from B-2 cells through transitional B cells), and newly identified subsets of B cells. Furthermore, innate T cells comprise γδ T cells, CD1-restricted natural killer T (NKT) cells, mucosal-associated invariant T (MAIT) cells, and intestinal intraepithelial lymphocytes (CD8αα + IELs). These cells express B cell receptors (BCR) and T cell receptors (TCR), respectively, and play essential roles in the body’s immune responses [34].

Fig. 2.

Examining Innate and Adaptive Immune Responses: Key Insights into Host Defense Mechanisms. The observed innate and adaptive immune responses highlight how the body’s immune system recognizes and defends against pathogens. The innate response offers rapid, nonspecific defense through cells like macrophages and NK cells, while the adaptive response provides targeted, long-lasting protection by activating T and B cells to generate specific immunity [34]. NK: Natural killer, TSLP: Thymic stromal lymphopoietin, IL: Innate lymphoid, TGF-ß: Transforming growth factor-β, IFNγ: Interferon gamma, TNF: Tumor necrosis factor, G-CSF: Granulocyte colony-stimulating factor, CXCL8: Interleukin-8, Treg: Regulatory T cells, ILC: Innate lymphoid cells, ILCreg: Regulatory innate lymphoid cell, and Th: T helper cells. Figure generated in BioRender (BioRender.com)

The innate and adaptive immune systems share several key features that enable them to work synergistically in protecting the body [34, 35]. Specifically, ILCs exhibit characteristics similar to CD4 helper (Th) and CD8 cytotoxic T cells (CTLs) of the adaptive immune system. For instance, ILC1 produces IFN-γ and utilizes the T-bet transcription factor (similar to Th1 cells), while ILC2 produces IL-5 and IL-13 and expresses the GATA-binding protein 3 (GATA-3) transcription factor akin to Th2 cells. Similarly, ILC3 correlates with Th17 cells as both express the RORγt transcription factor [34].

Furthermore, ILCreg cells express the Id3 transcription factor and produce IL-10 suppressing the activation of other ILCs such as ILC1 and ILC3. This resembles the function of T regulatory (Treg) cells (including Tr1 and other subsets) which inhibit the immune responses of Th1, Th2, and Th17 cells [34, 35]. Conventional NK cells of the innate system derive from EILPs (early IL progenitors) and acquire transcription factors like T-bet and Eomes, which endow them with cytotoxic characteristics similar to adaptive CD8 + T cells (such as cytotoxicity and the production of cytokines like IFN-γ, TNF, and GM-CSF). ILCs have the remarkable ability to undergo transdifferentiation. ILC2 can shift to ILC1 or ILC3 phenotypes and alter cytokine production, thereby contributing to immune responses [34, 35].

Despite these similarities, there are distinct mechanisms of recognition and response to threats that differentiate the two immune systems. B and T cells in the adaptive immune system possess unique BCR and TCR, which are genetically distinct for each individual due to DNA rearrangements during development. BCRs function in both membrane-bound and soluble forms, allowing them to interact with a broad range of molecules, including proteins and polysaccharides. In contrast, TCRs are membrane-bound and bind to antigenic peptide-MHC complexes or non-conventional MHC complexes on antigen-presenting cells (APCs) [34, 35].

Adaptive immune cells express species-specific accessory molecules that assist BCR and TCR functions. These include CD4 and CD8, which facilitate TCR binding to specific ligands, enhancing T cell-APC interactions. Other molecules (CD28, CTLA-4, and PD1) regulate the activation and function of B and T cells by upregulating or downregulating their activities, which can be measured by the production of specific proteins or secreted effector molecules [36]. Adaptive immune cells also rely on species-specific intracellular factors that regulate gene expression, including nuclear transcription factors, epigenetic modifications (DNA methylation and histone modifications), and non-coding RNA transcription. These factors control key processes like proliferation, migration, homing, differentiation, and trans-differentiation [37, 38].

In microfluidic systems, it is possible to precisely simulate both innate and adaptive immune responses, particularly in immune cell co-culture models [35, 39]. In these systems, macrophages and NK cells can effectively respond in rapid and non-selective manners similar to the behavior of the innate immune system when encountering pathogens [40]. T and B cells in co-culture models within these systems can be utilized to simulate adaptive immune responses, where they specifically recognize and target pathogens [41]. By combining both innate and adaptive immune responses in a single microfluidic platform, complex interactions that occur in immune responses to tissue damage or infections can be more accurately and scalable modelled [42]. These capabilities offer more effective analysis in fields such as immune therapies and drug response evaluation [42].

A multiorgan ISOC model has been created which includes recirculating THP-1 immune cells, cardiomyocytes, skeletal muscle cells, and liver cells (contained within separate compartments using a serum-free medium) [10]. A key challenge when utilizing the THP-1 cells is the potential for non-physiological responses when cultured under controlled conditions. However, this issue can be addressed by facilitating interactions with somatic cells that mimic the natural environment of the cells, something a multiorgan model can provide [43]. This in vitro platform simulates both a targeted immune response to specific tissue damage and a broader inflammatory immune response triggered by exposure to inflammatory compounds [44]. In the targeted immune response scenario, fluorescently labeled THP-1 monocytes selectively infiltrate a cardiac module damaged by amiodarone. In contrast, a general exposure to lipopolysaccharide (LPS) and IFN-γ induces widespread damage across all compartments. Biomarker analysis reveals an upregulation of proinflammatory cytokines (TNF-α, IL-6, IL-10, MIP-1, MCP-1, and RANTES) in response to LPS + IFN-ꝩ, suggesting the induction of an M1 macrophage phenotype while amiodarone treatment elevates IL-6 levels which is indicative of an M2 macrophage phenotype. This system provides an alternative to humanized animal models for studying both the direct immune effects of biological therapeutics and the indirect effects of cytokine release as a result of drug pharmacokinetics and pharmacodynamics [10].

In another investigation, the activation of T cells and the secretion of cytokines are critical for understanding cell-mediated immune responses [45]. To address this, researchers introduced a novel droplet-based single-cell immunoassay platform (Drop-SCIA) that combines surface-enhanced Raman spectroscopy with homogeneous-phase immunoassays that allow for highly sensitive, multiplexed detection of cytokines at the single-cell level, offering superior enrichment efficiency when compared to traditional ELISpot assays that rely on interfaces [45]. Using this platform, the activation of Jurkat T cells and their secretion of IL-2 and IFN-γ were examined after co-culturing them with normal breast epithelial cells (MCF-10A) and different breast cancer cell subtypes (MCF-7, MDA-MB-231). The Drop-SCIA platform provides a reliable and robust approach for single cell immune profiling of Jurkat cells allowing for a precise evaluation of their activation status. Significant variations in IL-2 and IFN-γ secretion were observed under different co-culture conditions, suggesting that distinct breast cancer subtypes influence T cell activation and function in unique ways. With merging microdroplet microfluidics with SERS-based detection, Drop-SCIA represents a significant technological advancement in studying CMI, especially in the context of tumor microenvironments. This approach holds immense potential for enhancing our understanding of T-cell immune function, with applications in cancer immunotherapy, vaccine development, and the diagnosis and treatment of viral and infectious diseases [45].

A scalable intracellular delivery method utilizing vortex microfluidics significantly enhances the performance of chimeric antigen receptor T (CAR-T) cells [46]. CAR-T therapy has been approved for hematologic cancers and is currently being explored for solid tumors, autoimmune diseases, cardiovascular issues, and aging. CAR-T and other engineered cellular therapies face challenges in production efficiency and safety. Conventional methods like lentiviral transduction and electroporation often lead to random genetic integration or considerable cellular damage, potentially limiting the safety and effectiveness of these therapies [46]. Hydroporation has been identified as a promising alternative that offers a Gentler, more efficient method for intracellular delivery. Compared to electroporation, hydroporation increases CAR-T yield by 1.7 to 2 times while Maintaining superior cell viability and recovery. Cells processed through hydroporation demonstrate improved CAR-T yield over five days post-transfection, along with rapid proliferation, potent target cell lysis, and enhanced secretion of both pro-inflammatory and regulatory cytokines. This technique is capable of processing up to 5× 10^8 cells in less than 10 s, making it an ideal solution for high-yield CAR-T production and offering the potential to improve therapeutic outcomes in clinical applications [46].

Mechanisms and strategies in cancer immunology and immunotherapy

TME refers to a tumor’s surroundings including nearby blood vessels, immune cells, fibroblasts, signaling molecules, and the ECM (Fig. 3) [47]. The cancer and its surrounding microenvironment are constantly interacting [48]. Tumors can affect this microenvironment by releasing extracellular signals inducing angiogenesis (forming new blood vessels) and establishing local immune tolerance. These tumor-host interactions help shape regional and systemic immunity ultimately promoting tumor development and creating an immunosuppressive environment [47].

Fig. 3.

The tumor microenvironment is a complex network of various stromal and immune cells organized within irregular blood vessels and collagen structures. The limited blood flow and the dense packing of glycolytic tumor cells result in decreased oxygen availability, the creation of an acidic environment, and insufficient nutritional supplies. This environment also leads to a buildup of anti-inflammatory cytokines, chemokines, and metabolic byproducts. Figure generated in BioRender (BioRender.com)

Immunotherapy can have different effects on cancer patients because the TME is heterogeneous and is directly related to the type of cancer, the stage of cancer progression, and the patient's physiological conditions [49, 50]. Therefore, a complete cellular understanding of the TME is critical to increasing the effectiveness of immunotherapy in patients. One of the factors that plays an essential role in shaping the TME is the physical barriers, inhibition of immune cells, and recruitment of immunosuppressive cells created by the injection of cancer cells [49–52]. Cancer cells skillfully secrete immunosuppressive cytokines such as transforming growth factor-β (TGF-β), IL-10, and vascular endothelial growth factor (VEGF) stimulate the expression of inhibitory receptors and ligands (such as programmed death-ligand 1/2 (PD-L1/2) and cytotoxic T-lymphocyte antigen 4 (CTLA-4)), and ultimately lead to the reduction of tumor-specific primary histocompatibility complex class I (MHC-I) antigens. Another route cancer cells take is to produce weak and chronic antigen signals to exhaust T cells and disable the immune response [49–52].

Tumour stromal cells, including carcinoma-associated fibroblasts (CAFs), tumor-associated endothelial cells, and components of the ECM, engage in dynamic interactions with malignant cells, facilitating angiogenesis [53]. CAFs prevalent in approximately 80% of pancreatic and breast tumors are crucial in enhancing myeloid cell recruitment to the TEM. Additionally, CAFs contribute to the immune evasion mechanisms of tumors by promoting the differentiation of T cells into regulatory T cells (Tregs) thereby inhibiting the activity of effector T cells and NK cells [54]. They support tumor progression and immune suppression through the induction of tolerogenic dendritic cells (DCs) and by activating immunosuppressive M2 macrophages via the secretion of factors such as TGF-β and IL-10. Furthermore, CAFs facilitate the replacement of antigen-presenting cells (APCs) with tumor-associated endothelial cells, which function as non-professional APCs, thereby undermining the adaptive immune response within the TEM [54, 55].

The immune system employs two primary mechanisms to regulate the activity of cancer cells: targeting tumor-specific antigens unique to cancer cells and responding to tumor-associated antigens. Research on murine models indicates that immune responses against tumors largely hinge on mutated proteins that serve as tumor-specific antigens [56]. Notably, such mutations have been identified in genes related to human papillomavirus (specifically, E6 and E7) which are implicated in cervical cancer as well as in Epstein–Barr nuclear antigen 1, the Epstein-Barr virus nuclear antigen associated with Burkitt lymphoma and nasopharyngeal carcinoma [57, 58]. Monoclonal antibodies are commonly employed to detect tumor-associated antigens [59]. A pertinent example is melanoma-associated antigen 1, a melanoma-specific antigen that can activate human T cells in vitro and effectively prime the immune system to target tumor antigens [60]. Additionally, the proteasome pathway plays a critical role degrading proteins into shorter peptides within both normal and neoplastic cells [61]. These peptides are then presented by MHC-I molecules located on APCs to cytotoxic CD8 T cells (Fig. 4a) [61].

Fig. 4.

a Peptides derived from three normal self-proteins are shown in yellow, blue, and green on the major histocompatibility complex (MHC) molecules found on the surface of cancer cells. These indicate normal self-peptides (yellow, blue, and green). When a mutation occurs, the tumor cell's inability to repair DNA damage may change a normal protein, resulting in the presentation of mutated peptides (black) on the surfaces of tumor cells. Additionally, a typical protein (yellow) can be overexpressed in the tumor cell due to mutations or regulatory factors influencing its expression, causing its peptides to be presented at unusually high levels on the cell surface. Regarding post-translational modifications, a typical protein can undergo abnormal processing (such as complexing, glycosylation, phosphorylation, or lipidation) after translation, leading to an atypical array of peptides on the tumor cell surface [61]. b Several modalities of T cell-based immunotherapy are employed in cancer treatment [62]. One approach involves isolating tumor-infiltrating lymphocytes (TILs) directly from the patient's tumor tissue, then expanding the population of tumor-reactive T cells ex vivo before reinfusing them into the patient. Another method generates antigen-presenting cells (APCs) from peripheral blood monocytes. These APCs are loaded with tumor-specific antigens, identified through mutational and immunopeptidomic profiling, and co-cultured with autologous T cells to select and expand those with anti-tumor activity. In TCR-engineering strategies, T cell receptors specific to tumor antigens are cloned and introduced into patient donor-derived T cells, which are expanded in vitro before infusion. T cells can be isolated from patients or allogeneic donors and genetically modified to express chimeric antigen receptors (CARs) that specifically recognize and target malignant cells [62]. Figure generated in BioRender (BioRender.com)

Furthermore, peptides derived from mutated gene products or aberrantly expressed proteins in cancer cells can be directly administered to T cells [63, 64]. DCs can also be infused with these peptides, proteins, or entire tumor cells sourced from tumors and then cocultured with T cells replicating the natural interactions in vivo [65].

The pathway these cells follow is associated with suppressing immunosuppressive cytokines (TGF-β and IL-10) and inhibiting cancer cell escape from elimination by immune cells [66]. Increased levels of regulatory T cells have been observed in patients with head and neck cancer, for example [67]. Thus, it can be said that regulatory T cells can control the TEM [68]. The cells, therefore, escape destruction by immune cells [68]. The immunosuppressive effects of the tumor can also be systemic [69]. Regulatory T cells have increased in the peripheral blood of patients with head and neck cancer [68, 69]. For example, studies have shown that in patients with colorectal cancer or pancreatic tumors, the number of activated granulocytes and myeloid-derived suppressor cells increased with cancer progression, which was directly related to the suppression of T3 cells [70, 71].

Immunotherapy of cancer cells is achieved through monoclonal antibodies, cancer vaccines, T cells, immune checkpoint inhibitor therapy, and chimeric antigen receptor (CAR) T-cell therapy. Monoclonal antibodies have shown significant effectiveness against tumor antigens in various cancers. For example, trastuzumab targets human epidermal growth factor receptor (HER2)-positive breast cancer [72], Rituximab is effective for B-cell lymphomas [73] Cetuximab is used against certain cancers of the head and neck, lung, and colorectal regions that express the epidermal growth factor receptor [74–76].

Active-specific immunotherapy, often known as cancer vaccines, serves as an alternative to injecting existing tumor-specific antibodies or T cells [77]. Examples of these vaccines include those targeting the HER2 antigen in breast cancer, the immunoglobulin tumor marker in B-cell lymphoma, the mucin1 (cell surface-associated) antigen in lung cancer, and DCs exposed to tumor peptides or killed tumor cells in melanoma [78–81]. The administration of anti-cytotoxic T-lymphocyte associated antigen four antibodies alongside cancer vaccines can enhance the selective activation and expansion of T cells that directly respond to the vaccine [82]. This combination improves tumor immunity while minimizing the risk of autoimmunity.

Additionally, the intravenous infusion of autologous T cells explicitly targeting tumors has proven an effective form of immunotherapy (Fig. 4b) [62]. Several modalities of T cell-based immunotherapy are employed in cancer treatment [62]. During the early phases of T cell activation, inhibitory T cells increase the expression of CTLA-4 and, later, programmed cell death protein 1 [83–85]. These interact with their corresponding ligands, B7-1 and B7-2 for CTLA-4, and PD-L1 or PD-L2 for PD-1, which are presented by tumor cells, regulatory T cells (Tregs), myeloid cells, and APCs [86–88]. Through this interaction, cytotoxic T-cell activation is reduced, and this change plays a key role in inhibiting the immune response that would otherwise allow cancer cell proliferation and tumor progression to occur. Administration of immune checkpoint inhibitors is one way to address this reduction, allowing primary cytotoxic T cells to kill cancer cells [88].

Adoptive cellular therapy is recognized as a three-part method that includes the infusion of tumor-infiltrating lymphocytes (TILs), genetically modified T-cell receptor therapies, and CAR-modified T cells [89]. The theory behind immune checkpoint inhibitor therapy is that T cells have negative regulatory markers, known as “checkpoints,” that are essential for controlling their activation [90, 91].

Exploring advanced on-a-chip models: from cells to full human systems

In recent years, miniaturization in biological and medical applications has become a leading trend. One of the most common solutions for enhancing laboratory research through cell culture is the use of cell-on-a-chip systems [92]. These systems refer to microfluidic platforms where live cells are placed at a microscale to simulate conditions similar to their natural environment. This technology allows researchers to precisely measure and study cellular behavior, cell interactions, and physiological responses at the micro level. These systems are typically used to investigate cellular reactions to drugs, toxins, or other environmental conditions. The development of complex and stable systems that can simulate cellular, tissue, and organ interactions remains a significant challenge in this field. Simple and reproducible laboratory models (in vitro) that can accurately and reproducibly replicate disease states are still lacking [92].

Organ-on-a-chip refers to microfluidic systems designed to replicate the physiological and functional conditions of specific organs. These models typically incorporate microchannels in which cells from various organs (heart, liver, lungs, and kidneys) are cultured [93, 94]. The purpose of organ-on-a-chip systems is to assess biological responses, evaluate drug toxicity, and study drug biological interactions. These systems offer a more accurate and efficient means of drug development and medical treatments. Organ-on-a-chip platforms can also function as high-throughput tools, potentially bridging the gap between 2D cell cultures and animal testing [93, 94].

The ISOC is a microfluidic model designed to simulate the complex interactions between immune cells and other components of the immune system, enabling researchers to study immune responses to infections, vaccines, drugs, and cancer cells under inflammation and immune responses. Compared to 2D cell cultures, models like cancer-on-a-chip and immune-on-a-chip offer advantages in mimicking realistic physiological conditions but are still far from competing with in vivo animal models in predicting clinical outcomes [94–96]. Advancements in 3D in vitro models show potential to replace animal models in drug response and disease modeling. The combination of cancer-on-a-chip and immune-on-a-chip systems could reveal crucial insights into immunotherapy, including immune cell infiltration into tumors, aiding the development of next-generation cancer treatments. As tissue engineering progresses, immune-on-a-chip technologies offer the potential to create personalized immune organs for transplantation, potentially transforming treatments for autoimmune diseases, cancer, and infections by regenerating the immune system in patients or enhancing immune defenses [94–96].

Human-on-a-chip is an integrated system that combines multiple organ-on-a-chip models into a single microfluidic platform to simulate the biological and physiological responses of the human body [92–94]. This model typically includes organs (liver, lungs, heart, and kidneys) and the immune system are working together. It enables researchers to perform accurate simulations of human physiological processes, with applications in drug development, immunotherapy, and disease studies. Although a single-organ system replicates a single organ to assess drug effects on that specific organ, in contrast a multi-organ approach involves two or more organs to examine interactions between organs via drug metabolites or soluble signaling molecules. The multi-organ-on-a-chip system is the most complex and least documented but microfluidic technology allows for the integration of multiple chips, facilitating real-time drug monitoring and multi-organ interaction analysis [92–94].

Microfluidic fabrication techniques: methods and considerations

In recent years, significant advancements have been made in the development of organ-on-a-chip models and microfluidic devices for simulating human organs in the laboratory [97]. These technologies offer researchers the ability to replicate complex biological systems and their responses at the micro-scale. The choice of fabrication method depends on various factors including physical precision, process complexity, cost, and the specific research objectives. The methods for fabricating microfluidic devices are categorized into low-volume production techniques (casting, lamination, laser ablation, and 3D printing) and high-volume production techniques (hot embossing, injection molding, and film/sheet processes) [97].

Microfluidic fabrication methods can be broadly classified into mechanical, energy-assisted, and MEMS-based techniques [97]. Traditional machining approaches (CNC microcutting), can be costly and time-consuming due to the high capital expenses of the equipment and the complexity of the processes. A more cost-effective alternative is the replication approach and subsequent devices are duplicated from this master. This method reduces the overall manufacturing cost, as only the mould inserts, not the entire mould, need to be re-fabricated [97].

The master typically consists of two parts (a tool that provides structural support and a mould insert that defines the micro/nano features of the device). These mould inserts are often fabricated using conventional methods like photolithography, wet/dry etching, or other micro-engineering techniques. In energy-assisted methods, concentrated energy (lasers or electron beams) is used to etch or ablate the material, enabling precise fabrication of microstructures. While these techniques can also be used for direct polymer fabrication, they are typically more suitable for creating masters due to the high cost of serial processing [97].

Low-volume production

Casting

Casting is a key manufacturing method for silicone-based elastomers used primarily for molding or as a stamp to create microstructures or nanostructures [98]. This process often referred to as soft lithography includes several variants (microcontact printing, replica molding, microtransfer molding, micromolding in capillaries, and solvent-assisted micromolding). These techniques are advantageous for small-scale prototyping because they are low-cost, simple, and offer high precision [97, 98]. Typically, PDMS (polydimethylsiloxane) is used for casting, although other elastomers like polyurethanes and polyimides have been explored.

PDMS is popular in microfluidic applications due to its several mechanical, chemical, and optical properties. It can easily deform under external forces so allowing for replication of non-planar surfaces [97, 99, 100]. The Material is Durable, non-toxic, and provides good gas permeability, and useful for cell culturing. PDMS is transparent down to 280 nm, making it suitable for optical observation. However, its major limitations include swelling when exposed to nonpolar organic solvents, especially for high aspect ratio structures [97, 99, 100].

Casting PDMS on an SU-8 master mold is a cost-effective process for developing microfluidic devices. The process begins with spin-coating a negative SU-8 resist on a substrate, followed by a baking process to prepare the resist for UV exposure. After exposing the resist to UV light and baking it again the design is developed to create the microfluidic architecture. The PDMS mixture is then poured onto the SU-8 mold, cured (typically at 40–70 °C) and demolded. Degassing the PDMS mixture is necessary to remove air bubbles [97, 99, 100]. Once the PDMS structure is formed, it can be sealed using oxygen plasma or corona discharge to create closed microfluidic systems. PDMS casting on an SU-8 mold is effective for creating microfluidic devices with multiple layers allowing complex operations such as Quake-type valves and integrated pumping systems. These multi-layer devices are essential for applications like point-of-care diagnostics, though alignment and bonding between layers can be challenging [99, 100].

Lamination

Laminate manufacturing involves structuring individual layers [101]. Typically, this process is used for devices requiring between 4 to 12 layers, offering a variety of material options, including polymers like polycarbonate, polymethyl methacrylate (PMMA), cyclic olefin copolymer (COC), glass, and adhesive tapes. Each layer can be cut with a knife, plotter, or laser, with knife plotting being cheaper but offering lower resolution [97]. The layers are bonded using various methods (adhesive, thermal, or chemical bonding). Adhesive transfer tape is commonly used for prototyping devices that perform functions like mixing, particle separation, and high-temperature reactions. Also, polyester sealing films have been used to create multilayer devices that are biocompatible and offer wide temperature ranges [97]. In laminate-based devices, channel sizes typically exceed 100 μm, and the overall thickness of the device ranges from 2.4 to 5.2 mm. Laser cutting of PMMA layers, followed by bonding with ethanol and heat; has been used to construct multi-layer structures. Other approaches include custom-made holders and heating plates to align and bond the layers under controlled conditions [101, 102].

Laser ablation

Lasers can be used for direct scribing on workpieces or through a mask offering precision for fabricating microfluidic devices [103]. Lasers are especially effective for Machining small-volume devices with complex 3D geometries (including channels, holes, and components like valves). UV and femtosecond lasers offer superior precision, while CO₂ lasers, though less precise, are more cost-effective when using a femtosecond laser combined with micromilling [104]. UV lasers have also been used for cutting grooves in PVC films to create microfluidic devices for isolating cancer cells. CO₂ lasers, while less precise, have been employed for creating gas-actuated microvalves, peristaltic pumps, and other essential components within PDMS and PMMA substrates [103].

3D printing

3D printing or additive Manufacturing is a process that builds 3D objects by adding material layer by layer [104, 105]. This approach is quicker, more cost-effective, and simpler to implement than conventional Manufacturing techniques. It has gained widespread adoption for producing prototypes and Low-volume production of intricate parts with diverse mechanical and physical properties. Advanced 3D printing methods enable the creation of precise features at a significantly reduced cost compared to traditional cleanroom-based approaches [104, 105]. Several 3D printing technologies are available (selective laser sintering (SLS), fused deposition modeling (FDM), and inkjet-based systems). For example, FDM uses a heated nozzle to extrude molten polymer, forming the object layer by layer. Selective laser sintering uses a laser to fuse powdered Material, creating solid objects. 3D printing is particularly advantageous for microfluidic devices because it can rapidly prototype devices with complex Geometries. Inkjet printing systems can create parts with resolutions of around 100 microns or less. Photopolymer systems, especially those based on stereolithography (SLA) and digital light processing (DLP), use UV light to cure resin, enabling the creation of fine features and complex structures [104, 105]. The most common 3D printing methods for microfluidics are SLA and DLP. SLA uses a laser to trace each cross-section, whereas DLP cures entire layers at once using a UV light source (23,24). SLA offers higher resolution and better print quality but is slower compared to DLP (which is faster but less precise). Although these advanced techniques are still being refined, they hold promise for fabricating microfluidic devices with intricate internal structures. The combination of 3D printing with nanoimprinting can increase the resolution and scalability of microfluidic devices for high-volume production [104, 105].

High-volume production

Hot embossing

Hot embossing is a process that transfers intricate features onto a polymer substrate by applying high temperature and pressure [106]. During this process, the polymer is heated to a pliable state and a patterned mold is pressed onto the surface imprinting the desired features onto the material. This technique is commonly used for producing microstructured surfaces with high precision. It involves heating the polymer film and moulds in a vacuum, pressing the moulds to transfer the features, and then cooling the polymer. A vacuum environment is needed to avoid air traps, uniform temperature is critical to minimize warping and low friction between the mould and polymer is essential for smooth demoulding. Moulds are typically made of silicon or metal. Hot embossing is suitable for high aspect ratios and fine features (nanometres), and its process time varies from 4 to 30 min. Factors like applied pressure, temperature, and polymer chain orientation affect accuracy [106].

Micro-injection moulding

Micro-injection moulding involves injecting molten polymer at high pressure [107]. It's ideal for mass production of small-high-precision parts. The process allows high-quality surface finishes and various techniques exist (reaction injection and injection compression moulding). Micro-injection moulding is highly suitable for microfluidic applications due to its automation and scalability. It can result in imperfections like shrinkage, warping, and micro-level defects. Key parameters like mould pressure, temperature, injection speed, and cooling time need to be optimized for better part replication. Advanced techniques like statistical process control and design of experiments (DoE) are essential for refining the process [107].

Film or sheet operations

A growing trend in disposable microfluidic devices is the use of thin-flexible foil substrates [108]. These are advantageous for applications requiring quick heat transfer, reduced stiffness, or low Mass, such as PCR, centrifugation, and vibrating membranes. Foils also offer protective functions and are cost-effective for Mass production. Polymers, metals, and papers are commonly used for foil-based devices. Roll-to-roll Manufacturing enables scalable and cost-efficient production of such devices. Polymer-based foils thinner than 500 μm are typically used for fabrication with techniques like roller embossing and microthermoforming [108].

Roller embossing

Roller embossing involves imprinting thin foils using heated rollers [97]. The foil is heated above its glass transition temperature, patterned by rollers, and then cooled. This continuous process allows for high throughput and integration with other operations like printing or lamination. The challenge is transferring microstructure from flat silicon molds to a large continuous roll. Key parameters affecting embossing quality include pressure, temperature, speed, and pre-heating temperature of the foil [97].

Leveraging 3D bioprinting for Enhanced Tumor Microenvironment (TME) models

3D-printed microfluidic devices: fabrication, benefits, and challenges

The advantageous features of 3D-printed microfluidic devices include fast and user-friendly fabrication, enabling the creation of complex devices in just a few simple steps without the need for mold-making or complicated processes like soft lithography. The design flexibility is another key benefit that allows for the printing of any desired structure (including complex channels, threaded connections, O-rings, and compatibility with standard fittings) [109]. Precise flow control is achievable with the integration of complex components like pumps, valves, and mixers, which can be easily fabricated to enhance fluid manipulation within the device. The integration of detectors is simplified as electrodes can be easily added, removed, and repositioned, offering reusability and accurate placement [109].

3D-printed microfluidic devices can be designed for high-throughput analysis by fitting into standard 96-well plate formats. Modular systems can also be developed that enable easy replacement of faulty components without discarding the entire device. The ability to support advanced cell culture, with continuous nutrient supply and waste removal, enhances the versatility of the devices, and interchangeable inserts provide greater flexibility. CLIP and 3D-bioprinting can hold the potential to overcome current limitations and further enhance the capabilities of microfluidic systems.

However, there are several Limitations to current 3D-printed microfluidic devices [109]. One of the main challenges is resolution (as most 3D printers cannot create channels smaller than 200 microns) making them unsuitable for applications like cell gating or electrophoresis. The removal of support materials is another issue that can affect flow patterns or lead to contamination. The roughness of printed channel surfaces is another concern that includes adsorption of molecules and inconsistencies in surface modifications, potentially affecting the performance of the device [109]. Biocompatibility is also an issue, making them unsuitable for cell culture without further testing [110]. The lack of gas permeability in these materials restricts their use for long-term cell culture without additional gas-exchange features. The high cost of precision printers is another limiting factor [110].

Newer technologies Like 3D bioprinting are emerging as powerful solutions to overcome Many of the Limitations associated with traditional microfluidic fabrication methods, Like soft Lithography and standard 3D printing. 3D bioprinting allows for the direct printing of cell-laden Materials, Making it possible to create complex, biologically relevant environments that were difficult to achieve with older methods. In contrast to conventional 3D printing, which typically uses Materials Like plastics or acrylates, 3D bioprinting utilizes bioinks that can contain living cells, growth factors, and ECM materials. This advancement makes it particularly useful for cell integration, allowing for more realistic in vitro models for drug testing and disease modelling and even organ-on-a-chip systems [111, 112].

A significant advantage of 3D bioprinting over traditional 3D printing Lies in its resolution. While traditional 3D printers have struggled with high-resolution applications, particularly for small channels or fine details (such as those required for cell culture or single-cell analysis), 3D bioprinting has made strides in producing high-resolution structures with better precision at the micron scale. This allows for the creation of smaller channels and more complex geometries and intricate tissue-like structures that are closer to in vivo conditions [113].

Most traditional 3D printing Materials have limited or unknown interactions with living cells that can lead to toxicity or failure in Long-term cell culture. In contrast, bioinks used in 3D bioprinting are specifically designed to be biocompatible, supporting cell adhesion, growth, and differentiation, thus improving the overall performance of cell culture applications [114, 115]. 3D bioprinting also helps overcome the transparency issue found in conventional 3D-printed devices. Because bioinks are often formulated to be transparent or semi-transparent, 3D bioprinted structures can allow for better optical imaging and live-cell monitoring which is crucial for many biological experiments and analyses [116].

Another important advantage of 3D bioprinting is its ability to print structures that are gas permeable. Many traditional 3D-printed materials are not suitable for long-term cell culture because they lack the permeability needed for gas exchange. In bioprinting, researchers can design and print materials that support oxygen and nutrient flow through the structure and creating more viable microenvironments for cells and enabling long-term culture without the need for additional gas exchange components [117, 118]. 3D bioprinting also overcomes the complexity of fabrication associated with older methods. 3D bioprinting simplifies this by allowing for the direct fabrication of multi-material, multi-cell structures in a single print run, drastically reducing the time and complexity involved [119].

Advancements in 3D bioprinted Tumor Microenvironment (TME) models

A bioprinted TME model offers several unique advantages over traditional PDMS-based devices in addressing biological questions in areas that require more complex, dynamic, and physiologically relevant environments [120]. In a bioprinted TME, researchers can incorporate both cancer cells and various stromal components (such as fibroblasts, immune cells, endothelial cells, and ECM). This allows for the investigation of dynamic interactions between the tumor and its surrounding microenvironment. PDMS models are often limited in their ability to recapitulate these complex cellular interactions and cannot easily simulate the spatial and biochemical gradients that exist within a true TME. The ability to incorporate multiple cell types in their natural spatial arrangement allows researchers to explore how tumors influence their environment and vice versa, addressing questions about immune evasion, angiogenesis, and ECM remodeling [121].

3D cell culture within a bioprinted model better mimics the in vivo structure of a tumor, with heterogeneous regions that include areas of hypoxia, nutrient gradients, and cellular heterogeneity. Standard PDMS models even with microfluidic channels ften fail to create the type of 3D heterogeneity seen in actual tumors because of their simpler and less variable geometry. The ability to print the architecture of tumors (such as the creation of multicellular spheroids or mimicking necrotic regions) gives researchers deeper insights into how the tumor's internal structure influences drug response, metastasis, and the response to treatments [122].

The ECM plays a pivotal role in tumor progression, but PDMS models lack the capacity to easily replicate the complex. In bioprinting, the ECM components (such as collagen, fibronectin, hyaluronic acid, and matrigel) can be precisely printed in defined patterns to mimic the tumor's natural environment more closely [123].

Tumors in vivo experience hypoxic zones due to abnormal vasculature. PDMS devices often lack the capacity to reproduce these gradients because they typically rely on a consistent fluidic flow. Bioprinted models allow for more precise control over the spatial arrangement of cells and can simulate the creation of hypoxic regions that affect cell behavior and treatment responses [124]. PDMS-based models, immune cell infiltration and immune-tumor interactions are often poorly mimicked. By using bioprinted TME models, it is possible to introduce immune cells into the tumor environment, allowing for the study of immune evasion or immune checkpoint inhibitors in a more physiological context [125].

PDMS models often cannot replicate the complexity of blood vessels within a tumor. Bioprinting allows the creation of vascularized structures including the ability to print endothelial cells and form microvascular networks within the TME. This enables the study of how tumors form their vasculature (angiogenesis), as well as the role of vascular architecture in drug delivery and metastasis [125].

In PDMS-based models, drugs May not be distributed as they would be in real tissues Due to the absence of proper cellular arrangement and gradients. Bioprinted models can more accurately replicate tumor heterogeneity and drug resistance mechanisms that are seen in vivo. By incorporating 3D structures, nutrient gradients, and variable ECM stiffness bioprinted models offer a better simulation of how drugs diffuse within the tumor (how they are metabolized by cells and how drug resistance develops) [126]. Bioprinted models provide a better platform for live-cell imaging and real-time monitoring due to their customizable transparency and ability to replicate in vivo like environments. PDMS models often require special coatings or treatments to allow for optical observation and May not allow for continuous monitoring of cellular dynamics in 3D cultures [127, 128].

Immune-system-on-a-chip platforms for immune profiling and therapeutic development

Recreating immune organs such as bone marrow, thymus, Lymph nodes, lymphatic system, and spleen in vitro has become difficult due to the inherent microorganisms and compartmentalization in native organs [129]. Therefore, microfluidic systems with unique experimental designs developed in recent years have provided researchers with valuable information, some of which is mentioned below.

A research project developed a vascularized human bone marrow niche model utilizing a microfluidic system comprised of two hexagonal chambers linked by three bilateral ports [130]. The geometry of the ports was designed based on the structure of a capillary burst valve, which prevents the leakage of fibrin hydrogel while permitting the diffusion of soluble signalling molecules and the migration of cells [130]. COMSOL Multiphysics 5.2 simulated molecular transport with a steady state fluid flow modelled through a porous medium. A 70 kDa dextran was used to simulate soluble factor movement, integrating transport modules and pressure heads based on experimental conditions. Key physical constants were used, such as the fibrin gel’s diffusion coefficient and porosity [130]. The system was fabricated using SU-8 on a silicon wafer with soft photolithography. A 100 μm SU-8 layer was spin-coated, patterned, and silanized before PDMS was mixed with a curing agent and poured onto the mold. After curing, the PDMS was punched, bonded to glass, and sterilized with UV radiation before use [130].

The selection of a no-slip boundary condition for the walls and the use of acellular fibrin gel in the tissue chambers were key decisions that allowed for accurate simulation of fluid flow and hydrostatic pressure. One challenge in this model was the need to match pressure conditions with experimental data, which was achieved by adjusting specific experimental conditions and modifying physical parameters (diffusion coefficient of dextran and fibrin gel porosity). Simulate biological processes like soluble molecule transport within the tissue environment and to analyze hydrostatic changes due to different liquid levels were happened [130]. For device preparation, the use of soft photolithography for fabricating the device's master mold and PDMS as the primary material for the chambers was based on the sterilizability and appropriate transparency of these materials for microfluidic experiments. The choice of PDMS was logical due to its favorable flexibility and transparency properties and its compatibility with microfluidic needs in similar projects. However, challenges such as the need for specialized techniques to prepare surface features and properly treat materials were also considered [130].

Previous models designed in other studies did not fully capture all the features of human bone marrow. The use of murine HSPCs, the absence of an endothelial cell model or vascular system or the failure to simulate bone marrow pathologies were limitations in earlier models. Several key features of bone marrow were successfully replicated (perfusable 3D vascular network, perivascular and endosteal niches, and the maintenance of HSPCs) [130].

While the small volume of the tissue chambers provides the advantage of higher throughput, it also introduces limitations related to the number and variability of the seeded hematopoietic stem and progenitor cells (HSPCs). Although the small chamber size boosts throughput this characteristic also limits the number of HSPCs that can be seeded in each chamber. The device design may require further optimization compared to other models [130].

In this study, the designers opted to exclude features such as bone innervation and mineralized bone (Figs. 5). According to the results, the device was capable of responding appropriately to both doxorubicin and granulocyte colony-stimulating factor (G-CSF), simulating changes in the number of migrating cells. Since the primary focus was on hematopoiesis and cell migration in the model, the design decisions were made to maintain accuracy and simplicity while providing a foundation for studying diseases and chemotherapy drugs. The boronate combined metal oxide affinity chromatography (BMoaC) model can aid in studies of breast cancer metastasis or blood disorders such as leukemia and myelodysplastic syndromes [130].

Fig. 5.

Modeling the Bone Marrow Niche and Characterization of Stromal Cells. a Endothelial cells were combined with either hFOB or BMSC within a fibrin hydrogel to Generate vascular networks in 3D tissue structures in separate chambers. b Between days 4 and 7, vascular networks formed, with hFOB and BMSC predominantly localizing within their respective niches, as confirmed by fluorescence imaging. c Osteopontin (green) was detected in both (I) the endosteal and (II) perivascular niches. Endothelial cells (red) were counterstained with CD34. d SDF-1 was localized near endothelial cells (red), with enhanced expression in the (I) perivascular niche compared to the (II) endosteal niche. e Leptin expression was observed in stromal cells adjacent to endothelial cells, both in the (I) endosteal and (II) perivascular niches, with nuclei counterstained using DAPI (blue). Scale bars: 500 μm for low Magnification, 50 μm for high magnification [130]

Another microfluidic system was created to investigate the interaction between neutrophils and cancer cells and the metastasis of human vascular breast cancer to bone (Figs. 6a-d) [131]. The inlet and outlet channels were designed to measure 4 mm for the culture medium and 1 mm for the hydrogel channel. The chip proved effective, featuring three distinct circular hydrogel region that were enclosed by two channels and two lateral media channels. The inner walls separated the two channels and guided the flow through the circular hydrogel regions, thus establishing an environment for examining neutrophil cancer cell interactions [131]. Perfusable microvascular network model was created for connecting a bone-like microenvironment to a pre-established breast cancer metastasis in bone (BMtB) [131]. This system facilitated the real-time injection and monitoring of circulating neutrophils, providing insights into their interactions with cancer cells (CCs). This BMtB-on-a-chip model allowed for the analysis of how the metastatic microenvironment influenced neutrophil migration and subsequent interactions with CCs [131]. This approach was particularly focused on replicating a bone-specific microvascular environment, providing an opportunity to analyze neutrophil-CC interactions in a bone-mimicking context [131].

Fig. 6.

a A microfluidic system designed to investigate human vascularized breast cancer metastasis to the bone and (b) the mold fabricated using a high-resolution 3D printer and photopolymerizable resin. c Confocal imaging on day 4 of culture reveals connectivity between green fluorescent protein-expressing endothelial cells in the hydrogel channel and red fluorescent protein-expressing endothelial cells in a circular region. d Microvascular networks connecting hydrogel channels to metastatic regions facilitate the flow of 70 kDa tetramethylrhodamine isothiocyanate-dextran, visualizing vascular permeability [131]. e The neutrophil-tumor interaction model on a chip replicates interactions between neutrophils and 3D tumor spheroids embedded in collagen hydrogel, comparing two scenarios separated and contacted scenarios (f) Epifluorescence images (10X magnification) show tumor spheroids (unstained) and neutrophils (yellow, CMRA) at t = 0 h in both “separated” (i) and “contacted” (ii) conditions (scale bar: 100 μm). g A 3D rendering of 20X confocal images illustrates tumor spheroids (blue, CMAC) and neutrophils (red, CMRA) fixed at t = 6 h, highlighting differences in neutrophil infiltration between the “separated” (i) and “contacted” (ii) scenarios [132]

2D models were typically unable to accurately reflect the dynamics of neutrophil interactions with cancer cells as they lacked the three-dimensional structure and precise tumor microenvironment conditions [131]. After neutrophil interaction with CCs, the CCs could be easily recovered and analyzed through quantitative methods (such as flow cytometry and single-cell imaging), allowing for a detailed characterization of their phenotype. This model enabled not only qualitative analysis but also quantitative results through flow cytometry offering a comprehensive view of cellular interactions [131]. Biofabrication technology was employed to create the bone microenvironment and metastatic cell seeds for providing a precise platform for simulating bone metastases and facilitating more accurate findings. Short lifespan was on of the challenges in this study [131].

The neutrophil-tumor interactions-on-a-chip was developed as a microphysiological system to study the neutrophil-mediated attenuation of pancreatic tumor progression through C-X-C motif chemokine receptor two inhibition [132]. The chip design includes a central channel for loading differentiated human promyelocytic leukemia cells (DHL-60) cells or culture medium, a tumor gel channel for loading hydrogel with PANC-1 tumor spheroids ± DHL-60 cells, an empty gel channel for Loading empty hydrogel, and two medium channels with reservoirs for culture mediu. The five parallel channels have a height of about 164 µm and are Linked by four sets of ten trapezoidal posts, each measuring 300 µm and 185 µm at their bases, with a height of 100 µm and spaced 100 µm apart. These posts facilitate the confinement of hydrogel through surface tension and capillary forces. The design of the Master mold was created in AutoCAD and transferred to a chrome-on-glass photomask, That was used to pattern a silicon wafer coated with an SU-8 50 photoresist. After UV exposure, the wafer was developed, and the features’ height was measured with a profilometer [132]. Neutrophil tumor interactions on a chip were produced using soft lithography. PDMS and a curing agent were mixed, degassed, and poured over the master mold [132]. A human cell-based microphysiological system was engineered to study the interactions between neutrophils and tumor spheroids in both separated and contact conditions (Figs. 6e-g) [132]. The findings indicated that neutrophils promote tumor spheroid invasion through the secretion of soluble factors as well as direct contact with cancer cells. The CXCR2 inhibitor AZD-5069 was found to reduce both the invasion and proliferation of tumor spheroids by blocking direct interactions between neutrophils and cancer cells. CXCR2 inhibition also resulted in a decrease in neutrophil migration toward tumor spheroids [132].

A human cell-based NTI-chip was developed to explore the mechanisms by which neutrophils mediate tumor progression and how CXCR2 inhibition can suppress these processes, thus offering a potential neutrophil-based immunotherapy for pancreatic cancer [132]. The flexibility to position neutrophils either in direct contact with or separated from cancer cells revealed that both direct contact and soluble factors play key roles in neutrophil-tumor interactions. Through this system, quantitative multi-parametric data were collected regarding neutrophil migration (e.g., forward migration index, velocity, directionality, and displacement), spheroid invasion (area and circularity), and neutrophil spheroid contact (frequency and duration). AZD-5069 (is a CXCR2 antagonist), was chosen for testing due to its demonstrated ability not to negatively impact neutrophil migration from the bone marrow to peripheral circulation or interfere with key antimicrobial immune functions in healthy volunteers. AZD-5069 has shown good tolerability in Phase II clinical trials for respiratory conditions and is currently under investigation for treating prostate cancer, head and neck squamous cell carcinoma, and pancreatic ductal carcinoma [132].

The treatment with AZD-5069 successfully reduced neutrophil migration toward pancreatic tumor spheroids. In the contact scenario, neutrophils were embedded in collagen gel, whereas in the separated scenario, they were suspended in culture medium that leaves the impact of the collagen gel on neutrophil phenotypes unclear. The gel-liquid interface was essential for maintaining the spatial separation in the current experimental design, future iterations could allow neutrophils to be embedded in collagen gel while remaining spatially distinct from tumor spheroids. The decision not to include endothelial cells was intentional [132]. This study relied on immortalized human cell lines, which limits its broader clinical applicability [132].

A sophisticated multicompartment organ-on-chip system has been meticulously engineered to bridge the gap between 3D ovarian cancer tissue constructs and liver cell models [133]. This innovative setup emulates the systemic administration of the chemotherapy drug cisplatin allowing for a comprehensive evaluation of its therapeutic efficacy while assessing the potential for liver toxicity. By integrating these distinct biological systems the research aims to provide valuable insights into the drug's impact on tumor responses and hepatic health fostering a deeper understanding of its therapeutic profile in a more physiologically relevant environment [133]. This model allows physiological communications among organs and simulates dynamic cisplatin systemic administration through an imposed capillary-like fluid flow (Fig. 7a) [133]. The integration of biologically inspired 3D matrices within a fluidic-based platform offers the possibility of developing next-generation tools for drug research. This approach allows for the thorough assessment of novel drug candidates and a more detailed understanding of drug responses [133]. In this study, a 3D hydrogel matrix was employed to replicate in vivo-like environments for ovarian cancer, optimized for cell viability, migration, and molecule diffusion. HepG2 cells were utilized to model liver tissue despite some metabolic limitations [133].

Fig. 7.

a (1, 4) A cutting-edge multi-organ-on-chip system was developed to facilitate the co-culture of HepG2 cell monolayers in one chamber and (2, 6) 3D hydrogel-based ovarian cancer models in a separate chamber. (3, 5) These compartments are fluidically connected via an external circuit enabling cisplatin flow, effectively simulating systemic drug administration [133]. b An illustration of a microfluidic ex vivo immuno-oncology model showcases its dynamic tumor biopsy environment and control system. c A confocal z-stack image maps tumor cell death across a single plane. d A time-dependent perimeter analysis of tumor-infiltrating lymphocyte penetration highlights the advancing front of tumor-infiltrating lymphocyte migration into the tumor. e A conceptual diagram of the human-microbial cross-talk device depicts three co-laminar microchannels: one for medium perfusion, one for human epithelial cell culture, and another for microbial culture. f Flow cytometry results compare the viability of CD4 + T cells cultured alone versus (g) those co-cultured for 24 h with Lactobacillus rhamnosus GG [134]

In 2D models, drug uptake (e.g., cisplatin) reaches its peak rapidly, leading to unrealistic results. In 3D models, especially under dynamic fluidic conditions, drug diffusion occurs more gradually and continuously, more effectively targeting cancer cells and providing a more in vivo-like representation of the drug delivery process. The use of a multi-compartmental organ-on-a-chip system for systemic cisplatin administration and simultaneous evaluation of therapeutic efficacy and hepatotoxicity demonstrated improved performance compared to static single-organ models. Specifically, a reduction in liver toxicity and enhanced anticancer effects of cisplatin were observed [133]. The results of this study highlight the potential of combining 3D cancer models with fluid flow stimulation and multi-organ organ-on-a-chip systems to provide more accurate preclinical predictions. These models to reducing animal testing, offer valuable contributions to the development of personalized therapies and a deeper understanding of the biological mechanisms underlying systemic diseases [133]. Also, there are some examples of microfluidic platforms for assessing the cell–cell interactions that are presented in Fig. 7b [134].

The role of microfluidic platforms in cancer immunotherapy

One of the successful methods in treating cancer patients is immunotherapy technology [135]. Different studies are presented in Table 1. Immunotherapy has shown considerable promise in treating various cancers by enhancing the body's natural defenses. However, the TME presents significant challenges that hinder its full therapeutic potential [143]. These challenges include immune suppression and tumor immune evasion. These obstacles contribute to the limited effectiveness of immunotherapies, making it difficult to achieve sustained tumor regression in many patients. Addressing these issues remains a critical focus of ongoing research in cancer immunotherapy [143].

Table 1.

Modeling of the immune system-on-a-chip in enhancing immune responses

Organ Model	Role	Cancer or Health Cell Types	Immune Cell	Chip Fabrication Methods	Drug Testing Applications	Results	Category	Ref
Bone marrow	Haematopoiesis and niche formation	MDA-MB-231 breast cancer	CD34 + progenitors	Standard soft photolithography	Doxorubicin and granulocyte-colony stimulating factor	This model recapitulated bone marrow function (maintenance and differentiation of CD34 + hematopoietic stem/progenitor cells, neutrophil (CD66b +) elution, and responses to doxorubicin and granulocyte colony-stimulating factor)	Bone Marrow-on-a-Chip Models	[130]
Lymph node	Immune complexes' interaction and trafficking	Lymphoid follicles	Peripheral blood mononuclear cells	Standard soft photolithography	-	This model enhanced antibody responses to split-virion influenza vaccination compared to 2D cultures, showing increased plasma cells, anti-hemagglutinin IgG production, and cytokine secretion that closely mirrors vaccinated human responses at clinically relevant time points	Lymph node-on-a-Chip Models	[136]
Lymph node	Immune complexes' interaction and trafficking	Human lymphatic fibroblasts	Dendritic cells and T cells	Standard soft photolithography	-	This model developed an engineered stromal network to enhance chemokine secretion, conduit formation, and applications in immunotherapy evaluation, antigen presentation, and adaptive immune response studies	Lymph node-on-a-Chip Models	[137]
Lymph node	Cancer metastasis	PANC-1, LS174T, and B16F10	THP-1 monocytes	Standard soft photolithography	-	Inflammation-driven remodeling and monocytic cells may jointly regulate cell adhesion during lymphatic metastasis	Lymph node-on-a-Chip Models	[138]
Spleen	Blood filtration	Red blood cells	THP-1 monocytes	photolithographic and soft-lithographic techniques	-	Infected reticulocytes from P. yoelii were more deformable than non-infected reticulocytes, similar to findings in P. vivax infection	Spleen-on-a-Chip Models	[139]
Spleen	Sickle-cell disease	Red blood cells	THP-1 monocytes	Standard fabrication processes	-	Oxygen-regulated spleen-on-a-chip (S-Chip and M-Chip) offers a platform for studying how the spleen balances the retention and processing of altered red blood cells. Disruption of this balance can lead to splenomegaly, as retained red blood cells accumulate undestroyed, causing organ swelling	Spleen-on-a-Chip Models	[140]
Vasculature	Blood and lymph vessel- Immune complexes' interaction	Human umbilical vein endothelial cells	T cells and Peripheral blood mononuclear cells	OrganoPlate 3-lane 64 tissues in a standardized 384-titerplate format	-	Exposure to TNFα, INF-γ, and human peripheral blood mononuclear cells led to decreased trans-endothelial electrical resistance, increased intercellular adhesion molecule-1 expression, and changes in endothelial morphology, indicating disrupted barrier function in the human umbilical vein endothelial cells tubules under inflammatory conditions	Vasculature-on-a-Chip Models	[141]
Vasculature	Blood and lymph vessel- Immune complexes' interaction	Melanoma A375 cells, HMEC-1 ECs	T cells	OrganoPlate® 3‐Lane is based on a 384-well microtiter plate and contains 40 individual microfluidic chips	-	Primary human T cells adhered to endothelial walls and underwent transendothelial migration in response to TNFα, CXCL12 gradients, or melanoma cells. T cell migration varied with activation states	Vasculature-on-a-Chip Models	[142]

ISOC platforms in immunotherapy research

The ISOC is an advanced microfluidic platform that replicates key immune responses within a physiologically relevant in vitro environment [144, 145]. By integrating immune cells, tumor cells, ECM components, and biosensors the ISOCs provide a dynamic and controllable system to study immune interactions. This integration enables real-time monitoring of immune cell behavior, drug responses, and tumor-immune interactions providing a more accurate representation of the in vivo environment. ISOCs are invaluable tools for enhancing immunotherapy strategies, optimizing drug screening processes, and advancing personalized medicine by providing insights into how individual patients' immune systems interact with cancer cells and potential therapies [144, 145]. The key features of ISOC platforms include microfluidic precision control, recreation of the TME, real-time biosensing, and monitoring of the immune response [146]. Microfluidic precision control allows ISOCs to regulate the flow of immune cells, cytokines, and therapeutic agents, mimicking the immune-tumor interactions that occur in vivo. This level of control enables the simulation of various in vivo processes, such as immune cell infiltration, blood and lymphatic flow, and interactions with tumor cells [146].

Furthermore, real-time biosensing and immune response monitoring in ISOCs enable continuous measurement of key biomarkers, such as cytokine production (e.g., IL-2, IFN-γ, TNF-α), immune cell behavior, and immune checkpoint interactions [147, 148]. These systems allow for the assessment of critical immune checkpoints like PD-1/PD-L1 inhibition efficacy, providing invaluable insights into the effectiveness of immunotherapy. By monitoring these parameters in real-time, ISOCs offer a dynamic platform to evaluate how immune cells respond to treatment and how tumors evade immune surveillance, ultimately helping to refine therapeutic strategies and improve patient-specific treatment outcomes [147, 148].

The immunotherapeutic high-throughput observation chamber system utilized a miniaturized array of bioreactors to evaluate the effects of anti-PD-1 antibodies on cancer spheroids (MDA-MB-231, PD-L1 +) and T cells (Jurkat). Real-time monitoring of T cell inhibition and reactivation was enabled by measuring metrics such as tumor infiltration and interleukin-2 (IL-2) secretion. The system featured micropillar arrays for sensitive cytokine detection, yielding results comparable to standard ELISA assays. Key observations included the ability to inhibit immune cell function in the presence of cancer cells, reactivate immune cells using checkpoint inhibitors, enhance cytokine secretion following immune activation, and promote efficient immune cell penetration into tumors [149]. The optical transparency of the device and its capacity to observe T cell infiltration via fluorescence microscopy made it a potent tool for monitoring cancer-immune dynamics [149].

Advancements in organ-on-a-chip models for tumor-immune and CAR-T interactions

By combining microfluidic control, TEM modeling, and real-time biosensing, ISOCs represent a cutting-edge technology that enhances cancer immunotherapy research, accelerates drug discovery, and facilitates precision medicine approaches [150].

Using microfluidic systems to study the behavior of immune cells in these models is essential. Immuno-organ systems on a chip are divided into primary immune organs-on-a-chip (bone marrow-on-a-chip and thymus-on-a-chip) and secondary immune organs-on-a-chip (lymph node and lymphatic vessels-on-a-chip, spleen-on-a-chip, immune cell-on-a-chip, neutrophils-on-a-chip, monocytes and macrophages-on-a-chip, T cells-on-a-chip, and multiple immune cells-on-a-chip) categories [129].

In recent advancements, researchers have developed a microfluidic system designed to study the establishment and maintenance of the hematopoietic niche by co-culturing mesenchymal stem cells (MSCs) and hematopoietic stem and progenitor cells (HSPCs) [151]. This system integrates bone marrow-on-a-chip technology and offers a more physiologically relevant model for investigating stem cell interactions in vitro. The device operates with a passively perfused configuration using a poly(ethylene terephthalate) (PET) membrane that supports the sustained co-culture of MSCs and HSPCs for up to three days. Unlike traditional models that rely on gels or scaffolds, this microfluidic system promotes stromal formation driven by MSCs themselves. One key finding from this research was the comparative analysis between mesenchymal stem cell growth on PDMS (polydimethylsiloxane) membranes and PET membranes. The PDMS membranes demonstrated uneven cell coverage, likely due to fluorocarbon (CFx) residues, whereas the PET membranes provided more uniform growth [151]. Mesenchymal stem cells were seeded two days before the placement of hematopoietic stem and progenitor cells. Within three days, both cell types proliferated uniformly within the 3D stromal tissue demonstrating the efficacy of this setup in mimicking the in vivo conditions of bone marrow. The results indicate that the bone marrow-on-a-chip device has considerable potential for generating hematopoietic bone marrow organoids and can serve as a powerful platform for studying hematopoiesis and stem cell biology [151].

The human multiple myeloma-on-a-chip (hMM-on-a-chip) system is an interests model that features a five-channel design that models the bone marrow microenvironment, specifically the endosteal and perivascular niches [152]. It includes separate channels for fibroblasts, mesenchymal stem cell-based endosteal membrane culture, and hydrogel for the perivascular niche, creating a spatially distinct environment. Pericyte cells, visible through fluorescent lectin staining, are incorporated into the system to better simulate bone marrow vasculature. This model has shown enhanced proliferation of multiple myeloma (MM) cells in the perivascular niche. MM cells, tagged with fluorescent Markers, survived and proliferated for up to 12.5 days within the system, mimicking the in vivo conditions of the bone marrow. Multiple myeloma (MM) is the second most common hematological malignancy. Despite therapies like CAR-T cells, relapse remains nearly universal. The bone marrow microenvironment plays a crucial role in MM cell survival and resistance to treatment [152]. This new 3D microvascular model mimics the bone marrow's endosteal and perivascular niches, allowing for the study of MM interactions with the stroma and responses to CAR-T cell therapy. The system demonstrated the prolonged survival of both cell-line-based and patient-derived MM cells. The model also facilitated the perfusion of donor-matched CAR-T cells, enabling the study of T cell survival, differentiation, and cytotoxicity against MM cells. This MM-on-a-chip model could shed light on how the bone marrow microenvironment contributes to MM survival and therapy resistance, ultimately informing the development of more effective treatments [152].

Advancements in tumor immune microenvironment-on-a-chip models

Immune organ-on-a-chip models

The thymus-on-a-chip model has been studied to a limited extent and the research has explored its functionality using artificial scaffolds and decellularized ECM in combination with thymic epithelial cells [153]. In the lymphatic system, the functional units of the lymphangion constitute the lymphatic vessels. Spontaneous depolarization of the muscle cells' pacemaker cells generates an action potential to contract the lymphangion, increase intravascular pressure, close the inlet valve, open the outlet valve, and allow lymph to flow [153]. Dexamethasone improved drainage dysfunction caused by acute inflammation but did not impact chronic inflammation. Lee et al. explored the role of Rho-associated protein kinase in lymphatic endothelial cells, finding that Rho-associated protein kinase inhibition improved drainage by loosening tight junctions in lymphatic endothelial cells while tightening junctions in blood endothelial cells (ECs) [153]. A Rho-associated protein kinase2 inhibitor in a murine lymphedema model reduced tail swelling indicating a reversal of lymphedema. Though lymph nodes play a crucial role in immune trafficking no lymph nodes-on-a-chip model integrates endothelial cells comprehensively. Lymphatic endothelial cells in the lymph nodes express markers differently complicating modeling. Future lymph nodes-on-a-chip models could enhance understanding of the lymphatic system [153].

Lymphatic and spleen models-on-a-chip models

A research study explored lymphangiogenesis and angiogenesis using a microfluidic device that included a chamber filled with collagen gel (Fig. 8) [154]. VEGF-A, B, and C antagonists could act as drugs to combat lymphangiogenesis and angiogenesis by placing lymphatic endothelial cells and human umbilical vein endothelial cells in microchannels. It was noted that angiogenesis affects lymphangiogenesis while lymphatic capillaries developed button-like junctions after treatment with dexamethasone. VEGF antagonists were utilized to assess anti-angiogenic agents with a VEGF-R3 inhibitor specifically targeting lymphatic angiogenesis in the absence of blood vessels. Vascular angiogenesis enhances lymphangiogenesis by increasing the secretion of matrix metalloproteinases from endothelial cells. This model holds potential for drug screening applications related to corneal diseases and tumor development. Lymphangiogenesis, vital for growth, wound healing, and various diseases was replicated using an in vitro perfusion culture. Both mechanical and chemical stimuli contributed to the maturation of lymphatic structures rendering this model significant for investigating tumor formation and corneal inflammation [154].

Fig. 8.

a Lymphatic endothelial cells (LECs) adhered to microchannels within 12 h, leading to the development of lymphatic vessels within 24–36 h. b The process of lymphatic morphogenesis was influenced by VEGF-C in a dose-dependent manner (10 and 50 ng/ml) at days 3, 5, and 7, with podoplanin marking angiogenic sprouting. c After 10 days of perfusion culture, a well-defined lumen structure was observed in the lymphatic vessels. d and e Lymphatic sprouting and individual cell migration were visualized using podoplanin and Prox-1 staining. f Relative Gene expression analysis at days 1 and 10 (VEGF-C 10 and 50 ng mL.⁻¹) revealed that higher VEGF-C concentrations significantly upregulated VEGF-R3, Prox1, and ORAI1 expression (*p < 0.01, error bar = ± SD). g Immunostaining of podoplanin (red), F-actin (green), and nuclei (blue) in lymphatic (LA) and vascular (VA) structures at day 7 under VEGF-C concentrations of 10 and 50 ng/ml. h LA and VA formation were enhanced in a dose-dependent manner with VEGF-C (1, 5, 10, and 50 ng/ml) at day 7 (n = 5). i Single-cell migration of HUVECs and LECs at day 7 (n = 5). Statistical significance: *p < 0.01, **p < 0.05; error bars represent ± SD. j VE-cadherin staining highlights cell–cell junctions in lymphatic vessels (LV), blood vessels (BV), blood capillaries (BC), and lymphatic capillaries (LC). k In the presence of VEGF-C, VE-cadherin was strongly expressed at cell boundaries in BV, LV, BC, and LC, forming characteristic “zipper-like” junctions. l When dexamethasone was introduced alongside VEGF-C, VE-cadherin remained localized at BV and LV cell junctions, with zipper-like formations in BC. Meanwhile, LC exhibited an “oak-leaf” pattern, transitioning into distinct button-like structures [154]

The spleen, filters senescent, damaged, or infected red blood cells (RBCs), a function facilitated by its unique microcirculation system within the red pulp [155]. To replicate the spleen’s filtering mechanism in an in vitro model, researchers have developed a microengineered spleen-on-a-chip that simulates the spleen's hydrodynamic forces and the physical constraints of its fundamental structural and functional unit [139]. This multilayered microfluidic device is specifically designed to mimic key splenic functions including the closed-fast and open-slow microcirculations the reticular mesh where hematocrit levels are regulated, and the interendothelial slit where RBC deformability is tested [139].

Additionally, this platform provides a physiologically relevant model to study the filtration of mature RBCs and reticulocytes within the red pulp and offering new insights into erythrocyte clearance, hematological disorders, and malaria pathophysiology [139]. Hematocrit levels increased within the spleen's reticular mesh. This structure further restricts RBC flow leading to a progressive rise in cell density (δ) until equilibrium is established [139].

In another study, a Microfluidic model of retention and elimination of abnormal RBCs by human spleen with implications for sickle cell disease was assessed [140]. This study focused on sickle cell disease and investigated the impact of hypoxia on splenic RBC retention and clearance. the results demonstrate that RBC retention within intrasplenic endothelial slits and RBC macrophage adhesion occur more rapidly in blood samples from sickle cell disease patients compared to healthy individuals with these effects becoming significantly more pronounced under hypoxic conditions [140].

Additionally, sickled RBCs under low oxygen conditions demonstrate different phagocytic mechanisms than non-sickled RBCs in either low or normal oxygen environments. When reoxygenated there is a significant decrease in RBC retention at the intrasplenic endothelial slits leading to the quick unsickling and fragmentation of the engulfed sickled RBCs within macrophages. These observations offer a vital mechanistic understanding of how the spleen maintains the balance of RBC retention and removal, emphasizing how disturbances in this balance lead to anemia, splenomegaly, and acute splenic sequestration crises in cases of sickle cell disease [140].

A biological model of RBC-macrophage signaling interactions was integrated with a biophysical model of macrophage engulfment, further enhanced by in vitro phagocytosis experiments using spleen-on-a-chip technology [156]. This computational-experimental framework effectively predicts RBC phagocytosis dynamics under various disease conditions reveals distinct patterns between normal and sickle RBCs and identifies key molecular regulators involved in RBC clearance. A major focus of the model is the CD47/SIRPα signaling axis and SHP1 (Src homology 2 domain-containing protein tyrosine phosphatase-1) which are potential therapeutic targets for modulating sickle RBC clearance in the spleen. The model simulates crucial RBC-macrophage signaling mechanisms, such as the “eat me” and “do not eat me” signals [156]. For “eat me” signals, binding of phosphatidylserine (PS) to PS-binding proteins and Band-3 to opsonized FcγRs triggers activation of nonmuscle myosin IIa, promoting RBC engulfment. Increased RBC rigidity further enhances macrophage phagocytosis. For “do not eat me” signals, CD47 binds to SIRPα, activating SHP1, which inhibits myosin activation and prevents RBC clearance. Notably, altered CD47 (CD47_alt) in aged or stressed RBCs binds to TSP1 and SIRPα, triggering myosin activation and facilitating phagocytosis. The model also predicts more efficient phagocytosis of opsonized RBCs when CD47 is inhibited, and shows enhanced clearance of aged RBCs with TSP1 stimulation. Under hypoxic conditions, the model indicates faster erythrophagocytosis in sickle RBCs, suggesting that hypoxia accelerates clearance. The study identifies CD47/SIRPα signaling and macrophage receptor expression as key targets for therapeutic intervention. The biophysical effects of flow velocity on sickle RBC clearance provide insights into drug development and experimental design [156].

Tumor immune microenvironment-on-a-chip models

A cell-on-a-chip model utilizing nanoplasmonics for in situ observation of PD-L1⁺ exosome mediated immune modulation was developed (Fig. 9) [157]. An integrated nanoplasmonic immunoassay cell-on-a-chip was created to enable the co-culture of cancer and immune cells without direct contact. This setup allows for the on-chip exchange of exosomes, real-time detection of PD-L1⁺ exosomes and the measurement of immune cytokine secretion. The platform demonstrated that PD-L1⁺ exosome-mediated immunosuppression characterized by decreased IFN-γ and IL-2 from T cells, occurred only when exosomes were secreted by neighboring cancer cells. These results suggest that the NIIC model could enhance the understanding of exosome-driven immunosuppression and help predict responses to anti-PD-1/PD-L1 immunotherapies [157].

Fig. 9.

a On-Chip Cell Proliferation and Characterization of the Nanoplasmonic Digital Sensing Module. Proliferation rates of MDA-MB-231 cells (red) and Jurkat T cells (green) cultured separately in microchambers over 24, 48, 72, 96, and 120 h (n = 6). b Confocal images tracking the growth of MDA-MB-231 and Jurkat T cells at the same time points. Co-culture of MDA-MB-231 (red) and Jurkat T cells (green) on-chip: (c) bright-field microscopy, (d-e) confocal microscopy, and (f) merged imaging. g On-Chip Cell Proliferation and Nanoplasmonic Digital Sensing Characterization, growth rates of MDA-MB-231 cells (red) and Jurkat T cells (green) cultured separately in microchambers over 24, 48, 72, 96, and 120 h (n = 6). h Confocal microscopy images showing the proliferation of MDA-MB-231 and Jurkat T cells at the same time points. On-chip co-culture of MDA-MB-231 (red) and Jurkat T cells (green): (i) bright-field imaging, (j-k) confocal imaging, and (l) merged visualization. Characterization of capture antibody immobilization within the sensing chamber using (m) fluorescence microscopy, (n) atomic force microscopy (AFM), (o) near-field amplitude Mapping at 1660 cm⁻¹ via s-SNOM, and (p) near-field phase Mapping at 1660 cm⁻¹ via s-SNOM. q Dynamic light scattering (DLS) analysis comparing size distributions of 80 nm AuNSs (solid curve) and AuNS-Dab conjugates (dashed curve). Structural imaging of gold nanostructures: (r) TEM image of 80 nm AuNSs, (s) TEM image of AuNS-Dab conjugates, and (t) SEM image of AuNS-Dab conjugates capturing PD-L1⁺ exosomes [157]

The antitumor activity of NK cells was diminished during indirect coculture with cancer-associated fibroblasts in a pancreatic TIME-on-chip model [158]. 3D co-cultures of PANC-1 tumor spheroids (TSs) activated pancreatic stellate cells (aPSCs) and NK-92 cells embedded in a collagen matrix were used to study cellular interactions and cytokine profile differences in conditioned media through microchannel chips. PANC-1 TSs and aPSCs were indirectly cocultured in this setup while NK-92 cells infiltrated the TS channel via medium flow. The presence of aPSCs supported the growth of PANC-1 TSs and reduced the antitumor cytotoxicity of NK-92 cells. A reciprocal inhibition of cellular activities was observed between aPSCs and NK-92 cells though their migration capabilities remained unaffected. The decreased cytotoxicity of NK-92 cells was associated with a reduced expression of granzyme B. TIME-on-chip model was created using PANC-1 TSs, aPSCs, and NK-92 cells. This model provides valuable insights into the mechanisms of NK cell dysregulation and may help identify potential therapeutic strategies to restore NK cell function in the TEM. A microfluidic chip-based PDAC tumor model was developed in which PANC-1 TSs, aPSCs, and NK-92 cells were cocultured to enable both soluble factor interactions and direct contact between TSs and NK cells within a 3D setup [158]. This innovative TIME-on-chip model was designed to explore the anticancer effects mediated by NK cells. The anticancer activity of NK-92 cells was diminished in the presence of aPSCs due to the impaired cytotoxic function of NK cells. NK cell infiltration into the tumor parenchyma remained unchanged. The TIME-on-chip model is a valuable tool for investigating NK cell-mediated anticancer effects and testing NK-based immunotherapies in vitro [158].

In another study, an immunocompetent human model mimicking the TEM of HER2 + breast cancer was created using a micro immune response on-chip where CAFs, ECM, and immune cells were arranged in a spatial configuration typical of solid tumors [144]. Dynamic imaging of immune cell movement in MIRO revealed that the stromal barrier is critical in guiding immune cells and facilitating their removal from cancerous tissues. The anticancer immune response triggered by anti-HER2 monoclonal antibodies (mAbs; Trastuzumab) can be reactivated in immune-refractory HER2 + breast cancer through IL-2 administration. The significant improvement in treatment effectiveness achieved by combining anti-HER2 mAbs with IL-2 is mainly driven by changes in immune cell mobility and cytotoxicity rather than by altered stromal barrier permeability. IL-2-induced immunomodulation boosts immune cell movement and dispersal allowing them to overcome stromal immunosuppression and revitalize the anticancer response in tumors resistant to treatment [144].

A research project aimed to explore the interaction between Macrophages and T cells within breast TEMs and their influence on tumor progression using a 3D in vitro tumor-on-a-chip organotypic model [159]. The study found that macrophages promote the mobility and growth of cancer cells. However, the presence of both T cells and macrophages reduced the mobility and growth of the cancer cells. Cytokine analysis highlighted the importance of molecules like RANTES and leptin in influencing these effects. Evaluations of cell morphology particularly the presence of hybrid cell populations and macrophage polarization predominantly toward the pro-inflammatory M1 phenotype underscored the dynamic nature of the TEM and its crucial role in cancer progression [159]. The main limitation of this study lies in the use of commercially available cell lines to model macrophages and T cells. These cell lines may not fully capture the biological diversity and complexity of immune cells in the human body that could limit the generalizability of the findings to real-world clinical scenarios. Using patient-derived immune cells could overcome this limitation leading to more accurate and relevant insights into macrophage-T cell interactions within the TIME [159].

Application of ISOC in drug discovery processes

Revolutionizing drug discovery with ISOC technology

ISOC technology has revolutionized drug discovery by providing a relevant physiological platform for investigating immune responses, drug efficacy, and toxicity within a controlled microfluidic environment (Fig. 10a) [160]. By integrating immune cells, tumor cells, ECM components, and biosensors, ISOCs replicate dynamic interactions between the immune system and tumors, enabling real-time monitoring of immune activation, cytokine signaling, and therapeutic effects [161]. This technology has proven particularly beneficial for evaluating immunotherapies as it allows researchers to test checkpoint inhibitors, CAR-T cells, and cancer vaccines in conditions that closely resemble the human physiological environment [162]. Additionally, ISOCs derived from patient samples have advanced personalized medicine approaches, helping predict unique immune responses to treatments and refine drug therapies [163]. Beyond efficacy evaluation, ISOCs have also been critical for safety assessments revealing safety-related toxicities and identifying potential adverse effects during clinical trials [163]. ISOC technology has expedited the development of next-generation immunotherapies and enhanced the precision and efficiency of drug discovery by offering a more accurate and ethical alternative to traditional clinical models [164].

Fig. 10.

a The schematic demonstrates how an immune-system-on-a-chip platform contributes to drug discovery by replicating immune responses based on various drug delivery methods. The left side illustrates intravenous drug delivery, where the medication is directly administered into the bloodstream through an intravenous. The right side shows oral drug administration, where the medication is taken and absorbed via the digestive tract. In the center, immune cells are emphasized as critical elements in the immune-system-on-a-chip model, allowing for real-time monitoring of immune responses. The lower-magnified view displays the microfluidic system, simulating physiological conditions by incorporating immune cells and therapeutic agents to assess drug effectiveness and interactions. b A microfluidic endothelial barrier model was constructed using a porous membrane to allow for basal chemokine stimulation. This enabled controlled activation of endothelial cells and assessment of how tumor necrosis factor (TNF-α) influences breast cancer cell adhesion during metastasis. Significantly higher cancer cell attachment was observed on TNF–α stimulated endothelium. c Breast cancer cell invasion into bone tissue was recapitulated using human umbilical vein endothelial cells (HUVECs) cultured in channels adjacent to 3D collagen gels embedded with osteogenic cells differentiated from mesenchymal stem cells (MSCs). Cancer cell migration into the bone-mimetic compartment was observed. d To study epithelial–mesenchymal transition (EMT), lung cancer spheroids were embedded in micropatterned 3D matrices adjacent to endothelial-lined channels. Fluorescence-based microanalysis was employed to monitor spheroid dispersion and EMT progression within this spatially controlled microenvironment [165]. Figure generated in BioRender (BioRender.com)

Microfluidic models for immune cell migration and chemotaxis in drug development

In parallel to ISOCs, other microfluidic models have been used extensively to simulate cell movement and study the effects of drugs and vaccines. For example, DCs figure naturally migrate toward secondary lymphoid organs in response to chemotactic signals at infection sites [166] with key chemokines such as CCL19, CCL21, and CXCL12 directing this migration [167]. Microfluidic platforms have been employed to expose mouse bone marrow-derived DCs to controlled allowing for precise analysis of their migration patterns [168]. In these studies, CCL19 was found to be significantly more potent than CCL21 or CXCL12 in directing migration demonstrating a 10–100 times higher efficacy in guiding DC movement [169]. These platforms have further advanced the study of immune cell migration with additional studies showing how actin and myosin inhibitors affect crawling speed without impairing directional movement while pertussis toxin interferes with directed migration while leaving velocity unaffected [170].

Microfluidic devices have also provided insights into T cell migration, particularly in secondary lymphoid tissues (SLT). Chemokines such as CCL19 and CCL21 play crucial roles in guiding T cell recruitment and positioning within SLTs [171]. CCL19 gradients induce repulsion [172–174]. Mathematical modeling suggests that this repulsion results from competition between CCL19 and CCL21 for CCR7 signaling and their distinct abilities to desensitize CCR7. These findings offer a novel perspective on how CCL19 and CCL21 cooperate to regulate T cell migration and positioning [175].

Models of lymph nodes and the spleen have also become focal points in drug development due to their critical roles in immune system coordination and pathogen filtration. Lymph node models are essential for studying chemotaxis, immune cell interactions, and the response of immune cells to vaccines and drugs. As the site where adaptive immune responses are initiated these models are instrumental in developing targeted immunotherapies [172, 176].

Applications of bone marrow, lymphatic, and spleen models in drug discovery processes

Spleen models simulate the blood filtration process that is vital for removing pathogens and pharmaceutical drugs from the body. In a study, a simple strategy for spleen-targeted delivery of H2S donors was developed using one of the most widely used drug delivery systems PEGylated liposomes. The optimal PEG size and ratio for the spleen-targeted H2S donor-loaded liposome (ST-H2S lipo) were identified by comparing stability drug loading efficiency in vitro H2S levels and in vivo biodistribution. The ST-H2S lipo induced M2 phenotypic differentiation in vitro enhanced in vivo splenic Treg differentiation in both normal and DSS-induced mouse colitis models and showed significant protective effects in the DSS-induced colitis model. The ST-H2S lipo demonstrated a higher protective effect compared to the conventional long-circulating liposomes (LC-H2S lipo) loaded with H2S likely due to the greater efficiency of ST-H2S lipo in improving systemic immune homeostasis [177].

Understanding how the spleen processes drugs and eliminates foreign substances is crucial for optimizing drug safety and efficacy [178]. Both of these organ models provide valuable insights into immune cell behavior and drug interactions advancing the development of more effective treatments for diseases like cancer and autoimmune disorders (Figs. 10b –d) [165].

Bone marrow models are equally important in drug development due to the bone marrow's central role in immune cell production and the maturation of B lymphocytes [179]. These models help researchers understand hematopoiesis and immune cell maturation processes by simulating the hematopoietic niche. Studies have shown that BM-on-a-chip platforms, which incorporate nestin + cells crucial for hematopoietic stem cell maintenance and pluripotency can sustain hematopoietic functions in vitro for extended periods [180].

In a study bone marrow-on-a-chip model was developed utilizing microfluidic strategies to reconstitute and sustain a functional living bone marrow in vitro. Traditional tissue engineering methods involve implanting materials or cells into living organisms without constraints and this microfluidic platform enables the delivery of nutrients, chemicals, and soluble signals to maintain the viability and function of the engineered organ. By leveraging in vivo tissue engineering techniques the model successfully replicates the complex structural, physical, and cellular microenvironment of natural bone marrow producing the necessary factors to support the hematopoietic system without the need for expensive growth supplements. The engineered bone marrow preserves the normal proportions of hematopoietic stem cells (HSCs) and progenitor cells maintaining their spatial organization within a fully formed three-dimensional bone marrow niche. This approach allows the model to mimic complex tissue-level responses to radiation toxicity and therapeutic agents like G-CSF. Importantly bone marrow-on-a-chip model offers a promising alternative to animal models as it allows for the manipulation of individual hematopoietic cell populations or the introduction of other cell types (e.g., tumor cells) before assessing their response to clinical challenges such as radiation or pharmaceuticals. These systems have proven useful for investigating drug-induced myeloerythroid toxicity and the effects of ionizing radiation, enabling real-time monitoring of growth factors, cytokine secretion, and drug interactions. While these models offer significant advantages, they still fall short of fully replacing animal models for comprehensive bone marrow studies [180].

Lymphatic vessel models have also gained attention due to their role in drug transport between the circulatory and lymphatic systems [181, 182]. These models simulate the microcirculation within lymphatic vessels allowing researchers to investigate how drugs exit the bloodstream and enter the lymphatic system [183]. This is especially important for optimizing drug delivery and ensuring that drugs are efficiently distributed while minimizing side effects [184]. These models help to better understand how drugs interact with lymphatic tissues.

Applications of gut-on-a-chip model in drug discovery processes

Gut-on-a-chip have been developed to study the immune system’s interactions with microbes in the gastrointestinal tract. These platforms allow for the co-culture of human epithelial cells under aerobic conditions and microbial cells under anaerobic conditions while simultaneously maintaining primary human CD4 + T cells [185]. Changes in the human gastrointestinal microbiome have been associated with various diseases and inferring causality requires experiments in representative models. Widely used animal models exhibited limitations so to overcome these a modular microfluidics-based device was developed called HuMiX (human–microbial crosstalk) which allowed the establishment of a model of the gastrointestinal human microbe interface. The device consisted of three co-laminar microchannels (a medium perfusion microchamber, a human epithelial cell culture microchamber, and a microbial culture microchamber). Each microchamber had a dedicated inlet and outlet for inoculating cells and precisely controlling physicochemical parameters through the perfusion of laminar streams of dedicated culture media. Dedicated outlets also allowed the collection of eluates from individual chambers for downstream characterisation. By juxtaposing human and microbial cell populations at a distance of 0.5–1 mm across a separatory nanoporous membrane HuMiX faithfully represented a healthy intact epithelial barrier. The model integrated oxygen sensors (optodes) for real-time monitoring of dissolved oxygen concentrations and a specially designed version of HuMiX permitted the insertion of a commercial chopstick-style electrode (STX2; Millipore) to monitor transepithelial electrical resistance (TEER), aiding in the characterisation of cell growth and differentiation. It was demonstrated that HuMiX successfully recapitulated in vivo transcriptional, metabolic, and immunological responses in human intestinal epithelial cells when co-cultured with the commensal Lactobacillus rhamnosus GG (LGG) grown under anaerobic conditions. HuMiX provided a platform to investigate host-microbe molecular interactions and offered valuable insights into fundamental research questions regarding the relationship between the gastrointestinal microbiome, human health, and disease [185].

Gut-on-a-chip models have shown the potential to integrate immune components as demonstrated by the co-culture of immune-responsive U937 cells with human epithelial cell lines like Caco-2. These models offer a promising approach to studying immune modulation within the intestinal microenvironment and evaluating drug responses in a more physiologically relevant setting [186]. Microfluidic models have been utilized to simulate the human intestinal barrier with dual-layered chips that incorporate Caco-2 epithelial cells and U937 immune cells. This setup mimics physiological flow and supports tight junction formation enabling the study of drug absorption, disease mechanisms, and therapeutic interventions. Exposure to inflammatory stimuli such as LPS and TNF-α has been shown to increase monolayer permeability and modulate pro-inflammatory cytokine expression, further highlighting the utility of these models for drug testing [186].

Recent advancements and emerging trends

Recent advancements in immune-cell-on-a-chip technology have transformed traditional 2D models into complex 3D microfluidic platforms that better mimic the in vivo TEM. These systems now integrate ECM components, stromal cells, and immune cells, enabling more accurate studies of immune interactions with cancer cells [187]. Combining patient-derived tumor organoids with immune-on-a-chip models has further enhanced personalized immunotherapy research by allowing real-time monitoring of immune cell infiltration, cytokine release, and tumor-killing activity [188]. Additionally, integrating nanoplasmonic biosensors and advanced imaging techniques has improved real-time detection of immune responses facilitating non-invasive tracking of immune checkpoint interactions and immune cell activity [189].

Another major advancement is the combination of immune-system-on-a-chip models with other organ-on-a-chip systems such as liver, lung, and gut, to study systemic immune responses [190]. This multi-organ approach provides insights into immune-mediated toxicities, drug metabolism, and tumor-immune interactions across different tissues. Artificial intelligence and machine learning are also being incorporated into these platforms to analyze large datasets identify immune response patterns and refine immunotherapy strategies. Advances in cell isolation techniques now allow the incorporation of patient-derived T cells and NK cells into immune-on-a-chip models making it possible to test personalized immunotherapies, including CAR-T cell therapy and checkpoint inhibitors before patient administration [191].

Emerging trends in immune-on-a-chip research include microbiome-immune interaction models that explore how gut microbes influence cancer immunity and response to immunotherapies. Immune checkpoint inhibitors (anti-PD-1 and anti-CTLA-4) are being tested on these platforms to analyze tumor immune evasion mechanisms and optimize combination therapy approaches [192]. The integration of CRISPR gene-editing technology has further enabled precise modifications of immune cells allowing researchers to study genetic factors that regulate immune responses and improve engineered cell-based therapies [193, 194]. Advances in microfluidic chip design now support long-term culture systems making it possible to study chronic immune-tumor interactions, immune exhaustion, and tumor evolution under immunotherapy over extended periods [193, 194].

Furthermore, next-generation 3D bioprinting techniques are being used to fabricate physiologically relevant immune-on-a-chip systems with precise spatial organization of immune cells, tumor cells, and vascular structures, enhancing the fidelity of in vitro tumor-immune studies [195]. As this technology matures regulatory bodies are beginning to recognize its potential for drug testing leading to standardization efforts and collaborations between academia, industry, and healthcare regulators to accelerate its adoption in immunotherapy development. These advancements in immune-cell-on-a-chip technology are revolutionizing cancer research by offering a more precise, scalable, and personalized approach to studying immune responses. As these innovations continue to evolve they hold immense potential to improve immunotherapy outcomes, streamline drug development, and pave the way for more effective cancer treatments.

3D bioprinting technology offers a promising approach for accurately placing biological materials, living cells, and growth factors to create bioengineered structures through computer assisted transfer and fabrication techniques [196]. This process involves the layer-by-layer assembly of biomaterials and Living cells into a defined 3D arrangement. Automated 3D printers are used to print bioinks into cellular and tissue structures. Recently, 3D bioprinting aimed at organoids or cell masses has progressed significantly and holds great promise for addressing the challenge of producing large tissue structures needed to develop intricate biomimetic tissues and organs in vitro [196].

Bioinks are biomaterials designed for 3D bioprinting, essential for constructing tissues and organoids [196]. They must be biodegradable, bioactive, and non-toxic to cells while possessing good printability and mechanical stability. Printability is influenced by factors such as viscosity, surface tension, and the ability to self-crosslink. High viscosity can cause excessive pressure and shear stress potentially damaging cells whereas low viscosity compromises structural integrity. Mechanical strength is crucial for cell viability, gene expression, and maintaining tissue geometry. Both natural and synthetic polymers are used in bioinks. Natural polymers including agarose, alginate, collagen, and hyaluronic acid, are preferred for their similarity to human ECM. Matrigel derived from mouse sarcoma is widely used in organoid culture and often combines with other bioinks. Collagen-based bioinks, while beneficial are prone to premature setting so they are commonly blended with alginate or hyaluronic acid for improved handling [196]. Alginate extracted from brown algae is frequently used in microextrusion printing due to its hydrogel properties [196]. Other natural bioinks commonly include gelatin, fibrin, and methacrylated gelatin (GelMA) that enhance cross-linking and bioactivity. To improve the mechanical properties of natural bioinks chemical cross-linkers such as polyethylene glycol and thiolated gelatin are often incorporated. Synthetic polymers including polycaprolactone (PCL), polyethylene glycol, and pluronic, offer strong mechanical properties but can have biocompatibility issues and toxic degradation byproducts. Hybrid bioinks combining natural and synthetic materials can optimize both biological and mechanical characteristics.

For example a study demonstrated that a hybrid of alginate and PCL successfully supported cartilage regeneration by first printing a PCL framework and then filling gaps with cell-laden alginate. Overall, advancements in bioink formulations are improving the functionality and structural integrity of bioprinted tissues with potential applications in regenerative medicine and organoid research [196].

Inkjet bioprinting typically uses small nozzles (around 50 µm in diameter) and is best suited for printing bioinks with Low viscosity and cell densities of approximately 1× 10⁶ cells/ml. The technique provides high resolution (20–100 µm) at a relatively low cost. However, its limitation lies in its inability to handle high-viscosity bioinks, as clogging may occur restricting its effectiveness in printing complex tissue structures and organoids with dense cell compositions [196].

Multi-organ-on-a-chip (MOC) systems represent a cutting-edge advancement in tissue engineering and biomedical research aiming to replicate the physiological interactions between different organs within a controlled microenvironment [197, 198]. These systems integrate multiple organ-specific microfluidic compartments allowing for the study of dynamic inter-organ communication, metabolism, and systemic responses to drugs, toxins, or diseases. One of the most significant applications of MOC technology is modeling the immune system's interactions with various tissues. The immune system plays a crucial role in maintaining homeostasis, responding to infections and regulating inflammatory processes. By incorporating immune cells into MOC platforms researchers can simulate immune surveillance, inflammatory responses, and immune-mediated diseases in a more physiologically relevant manner than traditional in vitro models [197, 198].

MOC platforms can integrate lymphoid tissues, circulating immune cells, and organ-specific immune components to study immune surveillance mechanisms [199]. This allows researchers to investigate how immune cells interact with tissues under normal and pathological conditions, such as cancer or autoimmune diseases. By incorporating immune cells, cytokines, and inflammatory mediators, MOC systems enable the study of chronic inflammatory diseases like rheumatoid arthritis, inflammatory bowel disease, and multiple sclerosis. These models can help in identifying potential therapeutic targets and understanding disease progression [199].

Integrating TEMs with immune components in MOC systems allows for assessing immune checkpoint inhibitors, CAR-T cell therapy, and other immunotherapies. These platforms help predict patient-specific responses to treatment and improve drug development strategies [200]. MOC systems with immune components accurately represent drug metabolism, immune-mediated toxicity, and adverse reactions. This is particularly important for developing biologics, vaccines, and immunomodulatory drugs, reducing reliance on animal testing and improving drug safety assessments [201]. While MOC systems have shown immense potential, several challenges remain. Refining microfluidic designs and incorporating patient-derived cells will further enhance the clinical relevance of these platforms [200].

As research progresses MOC systems are expected to play a pivotal role in personalized medicine, drug development, and disease modeling. They will provide a deeper understanding of systemic immune interactions and improve therapeutic outcomes [201].

A microfluidic multi-organ-on-a-chip model was developed to study non-alcoholic fatty liver disease by simulating interactions between intestinal, liver, and immune cells [202]. This system incorporated Caco-2 (intestinal), HepG2 (hepatic), and RAW264.7 (immune) cells within microfluidic channels to mimic blood flow and organ crosstalk. The model successfully replicated oxidative stress, inflammation, and apoptosis associated with non-alcoholic fatty liver disease progression. When lipopolysaccharides or dextran sulfate sodium weakened the intestinal barrier palmitic acid influx exacerbated inflammation leading to higher IL-8 levels. This system effectively demonstrated the role of oxidative stress and inflammation in non-alcoholic fatty liver disease pathogenesis [202].

A multi-organ tissue chip was developed to interconnect mature human heart, liver, bone, and skin tissues through vascular flow enabling organ-to-organ communication. Endothelial barriers played a crucial role in regulating paracrine signaling, drug transport, and immune cell activity with their impact becoming more pronounced as biological complexity increased [203].

A multi-compartment organ-on-chip system was developed to Link 3D ovarian cancer tissues with liver cell models emulating systemic cisplatin administration for concurrent evaluations of drug efficacy and hepatotoxicity. Computational fluid dynamics was utilized to simulate capillary-like blood flow and forecast cisplatin diffusion. Cytotoxicity assays demonstrated a consistent decrease in the viability of SKOV-3 ovarian cancer cells and HepG2 Liver cells as cisplatin concentrations increased. In contrast, the 3D ovarian cancer models showed more resistance to the drug than the 2D models. Notably the organ-on-chip system fluid dynamics and multi-organ interactions yielded the most predictive results for toxicity and efficacy underscoring its potential as a dependable preclinical alternative [133].

Challenges and future directions

The creation of microfluidic immune-system-on-a-chip technology faces multiple obstacles that must be resolved to improve its efficacy in cancer immunology and drug discovery. A primary challenge is the accurate simulation of the intricate interactions within the immune system which encompasses a variety of cell types, signaling pathways, and inflammatory responses [204]. Although existing models emphasize innate immunity, integrating adaptive immune elements like T and B cells poses a technical challenge. Maintaining cell viability and longevity in microfluidic settings is also problematic and particularly for immune cells that depend on specific microenvironments to sustain their activity and functionality over time [164, 204].

Another significant issue is the lack of standardization and reproducibility fabrication methods and experimental conditions frequently result in variable outcomes complicating the comparison of results from different studies [205]. The replication of vascularization and lymphatic flow crucial for the movement of immune cells and the transport of drugs has not yet been fully achieved in current models [206]. Drug delivery and immune responses might not accurately mirror physiological realities without functional blood and lymphatic vessels [207]. Moreover, scalability and throughput continue to be challenged as microfluidic chips provide detailed insights at a small scale but are not yet optimized for high-throughput drug screening [208].

To address these challenges, future studies should concentrate on advanced 3D tissue engineering and organoid combination to produce more physiologically relevant models [209]. Chips customized for individual patients utilizing induced pluripotent stem cells (iPSCs) or cells derived from patients could aid in creating personalized models for precision medicine [210]. Enhancing vascularization and integrating the lymphatic system will be crucial for replicating actual physiological conditions enabling more accurate immune reactions and drug interactions [211].

Automation and AI-assisted data analysis can improve reproducibility, refine experimental conditions and facilitate extensive drug testing. Additionally, forming connections among multiple organs, such as linking an immune-system-on-a-chip with a tumor-on-a-chip, liver-on-a-chip, or gut-on-a-chip can offer a comprehensive perspective on immune responses, drug metabolism, and toxicity across different organs [95, 212].

Lastly, regulatory approval and industry acceptance remain significant hurdles to the practical application of this technology in clinical and pharmaceutical settings [213]. Establishing standardized protocols conducting validation studies and creating regulatory frameworks will be essential for obtaining the endorsement of the biotechnology and pharmaceutical sectors. By tackling these challenges microfluidic immune-system-on-a-chip platforms promise to transform cancer immunology, precision medicine, and drug discovery developing safer and more effective immunotherapies [213].

Despite significant advancements in microfluidic systems several challenges and limitations persist that hinder their full potential. These obstacles affect the overall functionality reproducibility and scalability of the devices, making it difficult to translate laboratory results into practical applications. The complexity of system design, coupled with issues related to long-term stability and precise control over biological environments poses considerable difficulties. Overcoming these limitations is essential for improving the accuracy, reliability, and overall performance of microfluidic platforms in various biomedical applications.

Recent advancements and challenges in organ-on-a-chip models for studying complex cellular interactions and biological processes have been analyzed. Organ-on-a-chip models have shown remarkable promise due to their ability to precisely replicate tissue microenvironments overcoming Many Limitations associated with traditional 2D and animal models.

One of the main challenges highlighted in the literature is the limited ability to replicate complex biological environments. For example, earlier studies on bone marrow-on-a-chip (BMoaC) models failed to incorporate certain intricate features [177]. While the review addressed the replication of endosteal and perivascular niches, future developments should consider incorporating additional bone-specific characteristics to more accurately reflect the complexity of the in vivo environment. This improvement could lead to a more comprehensive understanding of diseases and better simulate the effects of therapies targeting the bone marrow microenvironment [177].

A second challenge is the reliance on non-primary cell lines which are frequently used in organ-on-a-chip models for drug testing. While human primary cells like cord blood-derived HSPCs or patient-derived cancer cells can offer a more accurate model, they are often not as widely available or have limited functionality in in vitro systems [177]. The use of immortalized human cell lines (e.g., HepG2 for liver modeling) can sometimes lead to discrepancies in replicating true human biology as they lack the complex metabolic and immune interactions seen in primary cells [214]. To address this limitation future organ-on-a-chip models should prioritize the use of patient-derived or primary cells particularly when simulating complex diseases like cancer or blood disorders to improve the physiological relevance of the results [214].

A third challenge relates to the technical limitations of microfluidic systems particularly in terms of accurately modeling fluid dynamics and interstitial flow in tissue chambers. For example in the modeling of cisplatin treatment in ovarian cancer adjusting the physical parameters such as diffusion coefficient and hydrostatic pressure was crucial for ensuring that experimental conditions closely matched in vivo scenarios [214]. These modifications can sometimes introduce variability particularly in models with smaller chamber sizes. The trade-off between high throughput and variability in cell seeding density was identified as an area of concern. While smaller tissue chambers increase throughput they can also lead to higher experimental variability which may impact the reliability of drug response assays. Future iterations of organ-on-a-chip platforms should focus on optimizing seeding densities and minimizing experimental variability to improve the consistency of results.

Lastly, scalability and the combination of multi-organ models remain critical hurdles in the field. Although multi-compartmental organ-on-a-chip systems such as those used for modeling cisplatin administration or breast cancer metastasis in bone offer valuable insights into drug efficacy and toxicity, the complexity of integrating multiple organ systems in a fluidic platform remains a challenge. Integrating vascular networks, tumor microenvironments, and immune interactions in a single platform can provide a more accurate depiction of systemic drug effects. Future research should focus on enhancing the interoperability and combination of these models to create more robust systems for studying drug responses and disease progression.

In conclusion, while organ-on-a-chip models represent a major advancement in biomedical research there are still several hurdles to overcome. Future directions should focus on improving the physiological fidelity of these models by integrating more complex tissue architectures, reducing variability in experimental setups, and using patient-specific models to better replicate in vivo conditions. Addressing these challenges will ultimately pave the way for more reliable and reproducible organ-on-a-chip systems that can accelerate drug discovery and improve understanding of human diseases [214].

Conclusion

In conclusion, ISOC technology represents a groundbreaking advancement in cancer immunology and drug discovery offering a controlled and physiologically relevant platform for studying immune responses and therapeutic interventions. The combination of ISOC into cancer research has enabled the modelling of immune-tumour interactions, enhancing the screening and optimization of immune-based therapies and providing valuable insights into tumour microenvironment dynamics. Furthermore, emerging trends Like 3D bioprinting, organoid fusion, and multi-organ systems-on-a-chip expand ISOC capabilities, enabling more complex and systemic immune interactions. Despite progress, challenges such as scalability, standardization, and long-term immune cell viability remain, requiring continued innovation and refinement. As ISOC technology evolves, it holds immense potential to revolutionize personalized immunotherapy, accelerate drug discovery, and optimize cancer treatment strategies, contributing to more effective and precise cancer therapies.

Authors’ contributions

Farnaz Dabbagh Moghaddam: Writing – original draft, review and editing, supervision, and figure design and preparation. Ali Anvar: Writing – original draft, Writing – review & editing. Ehsan Ilkhani: Writing – original draft. Delara Dadgar: Writing – original draft. Maedeh Rafiee: Writing – original draft. Najmeh Ranjbaran: Writing – original draft. Pejman Mortazavi: Writing – review & editing, and writing – original draft. Seyed Majid Ghoreishian: Writing – original draft. Yun Suk Huh: Writing – review & editing, Supervision. Pooyan Makvandi: Writing – review & editing, Supervision.

Funding

P Makvandi acknowledges support of the Leading Talent Award from Zhejiang Provincial Talent Program (Medical & Health Talents). This work was also supported by the Basic Science Research Program through the National Research Foundation of Korea (RS-2024-00406723 and NRF-2021R1A2C3011585) funded by the Ministry of Science, ICT and Future Planning (MSIP), Republic of Korea.

Data availability

No data were generated in the article.

Declarations

Ethics approval and consent to participate

Not applicable.

Competing interests

The authors declare no competing interests.