Correction: A multi-organ-on-chip to recapitulate the infiltration and the cytotoxic activity of circulating NK cells in 3D matrix-based tumor model
Year: 2026
Type: erratum
Abstract: Immunotherapies represent one of the current most promising challenges in cancer treatment. These are based on the concept of boosting the host's immune system toward the elimination of cancer and include several strategies such as the use of monoclonal antibodies (mAb) targeting tumor-associated molecules or blocking immune checkpoints, anti-cancer vaccines and cell-based therapies (Ringquist et al., 2021;Shelton et al., 2021). The success of immunotherapeutic approaches aimed to unleash the activity of cytotoxic cells such as CD8 + T lymphocytes or Natural Killer (NK) cells strictly depends on their interaction with cancer cells and other immune cells. Such interaction might be profoundly affected by the highly complex niche of the tumor microenvironment (TME), which is populated by different type of cells (i.e., stromal cells, tumor cells, immune cells), interconnected within a complex threedimensional vascularized matrix. Given these premises, the development of reliable preclinical human models has become crucial for the assessment of the best immune therapeutic approaches (Boucherit et al., 2020;Ando et al., 2021;Ringquist et al., 2021).To date, preclinical safety and efficacy assessment of immunotherapies are carried out through 1) extensive in vitro cultures, addressing the cellular and molecular basis of immune responses, cancer initiation and development, and its interactions with immune cells 2) in vivo xenografts and genetically engineered animal models, for a necessary systemic contextualization. However, both approaches show limitations. Conventional in vitro models use 2D cultures that underrepresent the complexity of the TME (three-dimensionality, shear stresses due to fluidic exposure, irroration affecting molecule distribution) (Rodrigues et al., 2021). On the other side, the animal models classically used in oncology often rely on immunodeficient mice xenografted with human cell lines or patient-derived tumor cells (PDX). These models, although useful for testing anti-tumor drugs, do not allow to obtain information regarding the anti-tumor activity of the human immune system. Other in vivo approaches are based on syngeneic mice with a fully competent immune system with results that not always recapitulate the human setting (Franklin et al., 2022). The more sophisticated and informative animal models to date are the so called "humanized mice" having a human HSC-derived immune system which can be engrafted with human tumors. However, some limitations are related to the often incomplete reconstitution of the human immune system, in terms of cell lineage development, wrong proportion of the various immune cell subtypes and degree of their activity and maturation. These defects are due to the presence of a mousespecific microenvironment lacking human stroma and growth factors (Cogels et al., 2021). Moreover, such models are very expensive, time-consuming, and hardly usable in personalized medicine.Therefore, 3D, human in vitro tumor models, including spheroids and organoids, as well as microfluidics approaches integrating the human immune components, are increasingly being developed and adopted (Rodrigues et al., 2021;Vitale et al., 2022). 3D tumor models have advantages over conventional 2D cultures, such as an increased cellular and architectural cancer complexity, like the presence of a biomimetic TME with the proper chemical and biomechanical features (Cavo et al., 2016), beside to the possibility of coculturing stromal, endothelial and cancer cells where cells can migrate and infiltrate in a 3D space. Moreover, from a technical point of view, the adoption of 3D cancer models allows to scale up cell yields, with a sample size compatible with a wide variety of standard downstream analysis, such as molecular and cytofluorimetric analysis, beside immunohistochemistry maintaining a possible anisotropic cells behavior (Ballester-Beltran et al., 2015).The arising microfluidic platforms display the potential to recapitulate the physiological blood flows affecting the survival of circulating tumor cells (Marrella et al., 2021a), the intra/ extravasation of circulating cells (tumor cells or activated immune cells), besides a reliable drug distribution (Marrella et al., 2021b). However, some microfluidic models used to coculture immune and tumor cells in physically separated compartments are currently adopted in static conditions (Lee et al., 2018). Moreover, the over-miniaturization of some organ on chips (OOC) allow to host only few cells and very small amount of tumor samples, thus limiting the downstream analytical approaches.From a manufacturing point of view, the conventional OOC models and microfluidic models under development are typically fabricated using the polydimethylsiloxane (PDMS) elastomer, in which UV lithography is utilized to create an overall chip architecture with microscale fluid channels across the compartments where few microliters of media circulate without any sampling/injection port (Amin et al., 2016;Lee et al., 2018). Besides suffering the difficulty of mimicking the complex structures of the microenvironment in vivo (Amin et al., 2016), these PDMS-based devices lead to the adsorption of small hydrophobic compounds on the chip, causing the reduction of their bioavailability, finally resulting in issues in terms of cellular responses and/or bias in biochemical analysis.In this paper, we present a unique and promising approach aimed at recapitulating the immune cell infiltration and activation against cancer cells in a humanized, fluid-dynamic and 3D environment. An organ-on-chip technology (MIVO ® ) has been recently adopted by authors for recapitulating the systemic administration of anticancer drugs and for carrying out efficacy assays in comparison with the standard xenograft model, demonstrating the high predictability of this in vitro 3R approach (Marrella et al., 2021b). Similarly, the same technological platform has been adopted for culturing aggressive breast cancer models and recapitulating the cancer cells migration and infiltration in the fluid flow circuit, which represents the first physio-pathological step towards the metastatic onset (Cavo et al., 2018). Here, the generation of a humanized and immunocompetent in vitro cancer model (Marrella et al., 2019), where tumor cells are cultured in a compartment physically separated through a porous permeable membrane from the fluid flow compartment, relies on the capability to emulate the microcirculation as well as the circulatory behavior of immune cells within the TME. The possible access to both the tumor and the circulating compartments allows to monitor and quantify the changes that occur in the TME (soluble molecules, cell death, tumor cell invasion), in circulating immune cells, and potentially in additional compartments physically connected each other through the circulating fluid flows (i.e., evaluation of the metastatic site). The flexibility of this approach carries the important potential of better recapitulating a clinical scenario, opening the way for a more reliable platform for 1) a personalized investigation of the specific migratory and infiltrative capacity of immune cells, 2) the analysis of the anti-tumor activity of both drug-based and cell-based therapies, 3) the investigation of the effects of tumor-immune cell cross-talk often leading to the onset of resistant tumor variants (Bottino et al., 2021).Assessment of natural killer cell viability in standard 2D culture vs. 3D alginate embedding High-risk neuroblastoma (NB) is an aggressive, metastatic pediatric cancer difficult to treat and still characterized by poor overall survival. We have recently developed an alginate-based 3D NB culture as in vitro model characterized by a more physiopathological setting showing only a partial overlap with the standard 2D culture in terms of expression of immune-related molecules (Marrella et al., 2019). We also demonstrated that the NB cell viability and proliferation are preserved after embedding NB cells within the 3D alginate matrix.The aim of the present work is the implementation of such three-dimensional culture with features mimicking 1) the dynamic microcirculation within the tumor microenvironment and 2) the circulatory behavior of immune cells. Then, a reliable dynamic culture, consisting in the 3D NB model cultured with circulating NK cells, has been generated as a suitable tool for analyzing immune cell migration and infiltration, as well as tumor cell killing.Given the hypothesis that circulating NK cells could be able to specifically infiltrate the alginate-based NB culture, we wanted to verify whether the alginate matrix per se could be detrimental for the NK cell viability. NK cells were cultured in standard conditions (suspended culture within a 96-well plate) or embedded in the alginate-based 3D matrix; the cells were then collected from the standard culture by pipetting or recovered from the 3D culture through alginate dissolution, stained with a Viability Dye, and analyzed by flow cytometry.In the NK 3D culture, a clear population of unstained (thus viable) cells is shown beside to a clear stained population of dead cells (Figure 1A, lower histogram); conversely, in the 2D culture the presence of a mid-stained population might be indicative of early cell death/structural degradation of the cell, enriching the population of non-viable cells (Figure 1A, upper histogram). According to the clearly viable population, a significant difference was in cell viability the standard NK culture and the 3D embedding is one of the most adopted approaches for and assessment of the of TME. the aim to a reliable in vitro model, to some preclinical in a we wanted to flow could be to cell recovered from the alginate as a model of a complex 3D We by analyzing NK NK cells were cultured in standard 2D conditions or in 3D embedding and the cells from the cell or the We the expression of the being that such is highly in NK cells (Figure as we high of + cells significant were the 2D and 3D culture setting in terms of (Figure of cells vs. then with NK and the NB cell to the of the of we with different and The NB was specifically with the aim to to in vivo where few NK cells infiltrate the most aggressive including NB et al., et al., et al., then by flow the of cells, on cells. in the NK of the cells show as we + cells in the coculture and + cells in the verify NK cells could infiltrate the alginate matrix and with tumor cells, we NK cells and NB at different and embedded in the alginate NK cells and tumor cells are in a the of NK cells to within the matrix for the tumor cells, we a NK cells were with the cell and then with the tumor cells. Then, the 3D was in and stained with a specific for the We several NK cells to with NB cells (Figure to a significant tumor cell by a high NK cell viability (Figure cells analysis that tumor cell increased NB cells were with NK cells in 3D models, that NK cells are able to and tumor cells in 3D the of dead tumor cells in the presence of NK cells was also increased in 2D culture, at T was Moreover, NK cells recovered from a high demonstrating that 3D alginate are suitable models for testing NK we have analyzed the of 3D cultures in static we carried out the dynamic by using ® The dynamic culture has been specifically to allow the of NK cells tumor cells which are in a standard the ® (Figure has been connected to a system, which the of a fluid NK cells the thus mimicking the circulatory system within the tumor The is to any molecule and immune cells at the was not of have been show that within ® from to which are of blood the flow is (Figure then the of NK cells to migrate (thus against the an with on the fluid flow per this additional 1) a of cell migration 3D alginate without tumor cells, in culture with as a and 2) a alginate with culture without any The tumor were by 3D alginate cultures of cells, in standard conditions culture with the cells for in a dynamic the the were and the cells were by of a shown in we a significant of cell migration in the tumor as well as in the or NK cell migration was in the demonstrating an NK cell The of the is by the aimed at fluid NK cells the not cells from the towards the of the fluid as shown in the NK cells were to infiltrate the 3D tumor the presence of the alginate and to with tumor cells, we the same dynamic culture the cells for under the 3D alginate cultures were with NK cells stained for the et al., and under a shown in cells were the of the is indicative of the of NK cells not only to specifically migrate toward the tumor culture, also to infiltrate the alginate matrix.The same was by NK cells with the cell the of the and then their toward the NB culture also as a of concept that the cells within the the dynamic culture were thus NK tumor cells the alginate a bias in the of the of immune cells not dynamic the 3D alginate were and stained for the tumor and under with the aim to possible cell shown in NK cells were the tumor within the alginate NK cell is for their anti-tumor with NK cells being related to a high cytotoxic However, of the tumor the NK cells which are characterized by these premises, we to verify dynamic culture could recapitulate the of a specific NK cell in the and NK cell We then a dynamic culture, 3D with NK cells Then, we collected 1) the 3D tumor 2) the NK cells, the and 3) the circulating NK cells, which were analyzed for and expression through flow The were as and the cell was analyzed the NK cells the tumor shown in a although cell population was present within the the on the of the NK cells to infiltrate the 3D tumor a that and only the cells were to the lineage (Figure in the in as well as in the proportion of NK cells is in the and with the circulating which is to the standard NK might a of NK cell population, recapitulating in different et al., et al., et al., we aimed at analyzing the of in tumor cells, after coculturing with circulating NK cells. a significant in the of on tumor cells that the NK cells are able to specifically tumor cell death, thus at in their cytotoxic of preclinical cancer models represents the basis for the of the development of more and personalized therapeutic and the reduction of clinical of oncology drugs, due to the of both 1) complex in vitro models that better recapitulate the physio-pathological features in et al., et al., et al., and 2) fully humanized animal models et al., et al., et al., 2022). and therapies a fully human microenvironment to be the the different immune cell tumor cells, and other cells within the TME. Moreover, the the of the tumor and the presence of within the human TME be for both drug-based and cell-based from a across the development of three-dimensional cultures and the current of microfluidic the implementation of a with the of fluid flows mimicking the dynamic environment. this the organ-on-chip technology represents a generation of in vitro models, which in the of cell cultures the physiological and et al., and different cellular compartments which can through channels and/or porous et al., the of cancer cells and immune cells in separated connected through within the same chip, allows the of a biochemical for immune cell from the first toward the tumor cells. However, most of these are used in static conditions et al., et al., et al., et al., et al., et al., et al., et al., et al., et al., or through flow et al., or with very fluid flow being from physiological in a model, et T cells a fluid flow of to et al., in an immune recently developed by et al., immune cells were cultured through a flow of to et al., these devices could be with microfluidic more reliable cell culture models with shear stresses and with fluid flows mimicking the of blood as well as the and biochemical that rely dynamic flows of immune cells, drug Here, we a fluid-dynamic technology for dynamic with a flow to a physiological blood flow (Marrella et al., et al., for mimicking the microcirculation of the TME. NK cells were by the fluid flow the tumor model, which was in cultured in a 3D matrix. NK cells and the tumor model were physically separated in different compartments through a porous permeable membrane the adoption of different culture media growth NK cells were to the by tumor cells and to a toward the tumor cells NK cells were able to migrate against the and by the fluid demonstrating that their migration is specifically by factors by tumor cells, as or migration has been in the tumor the of molecules the of the cells, leading to a cell et al., that the cells to towards the is also the generated by the presence of tumor cells to a migration of immune cells that was that with increased in the demonstrating for the first a migration of immune cells in a organ-on-chip NK cells and tumor cells were in the dynamic setting in a standard et al., et al., et al., et al., 2021). However, we that only of circulating NK cells were able to specifically in the upper compartment, in with the of NK cells in tumor et al., et al., et al., only this cellular the potential to with tumor cells, the limitations of the current static where often high are used NK cells the leading to a possible of their efficacy in tumor technical of the current approaches might lead to a in the preclinical evaluation of in which the activation and of specific of immune cells as well as of often engineered cells cell-based represent the crucial first step for their this at in the high success of immunotherapies at the preclinical which is not by the same success the step for activated immune cells to be against the tumor is their capability to infiltrate the tumor an activation without being affected by the from the TME. for the of cell for being for such cell-based has not a for et al., et al., 2021). this show beside the of the extravasation as indicative of immune cell is to the immune cell infiltration within three-dimensional tumor cultures, where the and of the matrix better the in the over-miniaturization of the microfluidic beside in recapitulating the and clinical features of to the of the tumor the Moreover, this also carries some technical limitations related to downstream biochemical assays and to small et al., et al., the use of very small cell et al., 2019), et al., and/or very small cells et al., are not always suitable for standard analytical such as and flow the to a different cell culture with a with to the standard cultures and a of analytical (i.e., often to cell these microfluidic devices not to in conventional On the the approach 3D cultures were with more cells with NK circulating cells. the increased sample size to represent the for the use of this platform with more complex 3D cultures (i.e., based on and/or with different cell to better the as well as with The could have a on personalized for based on the better or for the efficacy of cell-based demonstrated the of alginate-based cultures for different (Cavo et al., et al., et al., and in for NB cultures, where the to therapies and the tumor cells were and (Marrella et al., 2019). Here, we demonstrated that the alginate matrix is suitable also for culturing NK cells, without affecting their viability or their Moreover, NK cells and NB cells were in alginate we that NK cells their to with tumor cells and to of their behavior within the NK cells and the 3D tumor in dynamic we also the cytotoxic of NK we a significant of on tumor cells, we demonstrated that NK cells also their capacity within the alginate we analyzed NK cell in terms of within the circulating NK cells), the and after of dynamic we a clear in NK cells within the and + cells represent a of NK cells cytotoxic activity that by the and 2021). the infiltration and of NK cells have been with and to in NB et al., et al., is crucial to of a preclinical model that the human this model represents a for the of in vitro models that can be for testing immunotherapies also in different in a personalized Here, by using a human cell highly recapitulating the most aggressive NB et al., we a humanized and immunocompetent platform the standard in vitro and animal its by the of cells and molecules the this model could be more utilized for or the to the and the amount of human NK cell infiltration in et al., et al., the present model this infiltration could on the activity of factors such as or shown to be highly by cells et al., these could be in personalized strategies to the efficacy of neuroblastoma (NB) cell was by et al., and cultured in with and The cells were cultured in a ) at and were used to 3D tumor blood cells were from blood of by NK cells were by using the and were cultured on in with and NK cells were for analyzing NK cell et al., within ® was to 1) the fluid within the and 2) the fluid NK cells the analysis was by using the model of 1) a flow 2) an fluid et al., et al., 2022). The to be are the for the of and the for of + + is the fluid and the across the The of the ) and the was as at flow of was as to the to the fluid as the was as a was an was used to the the for fluid flow was to the model to the and of the NK cells in the fluid as a of The model was used to the of the in the fluid Such model is based on the of + is the of the its and the and the The was through the is is the NK cells as to et al., The was is the the to et al., and the fluid on the was that the of the that the is NB models were generated as (Marrella et al., 2019). cells were in with and and with a alginate to obtain a alginate cell was a spheroids The were then with and in a 96-well cultured in with and and the of NK cell 3D cultures, after NK cells were recovered for and then cultured for additional in the standard 2D setting plate) or embedded within the alginate matrix. the 3D culture was NK cells in with and and with a alginate The 3D were the The cells were then analyzed by flow in terms of cell viability through with a Viability cell 3D cells or NK cells were in with and or with and and then at different The 3D were the and in culture analyzing the expression of the NK through flow dynamic were by using the ® with the aim to recapitulate the complexity of a 3D, dynamic TME. NK cells were stained with in with and and within the ® to an of with tumor cells. The of NK cells was by a system, through the of a fluid flow of 3D NB were cultured with the physically separated through a permeable porous extravasation after dynamic both circulating the and the tumor NK cells were and by of a NK cells were also cultured the or an alginate NB cells with as a 3D dynamic or static NK cells were used or with proliferation NK cells were with and at in the same was at the of and for at in The was by culture and cells were the NK cell or NB 3D cultures, from dynamic or static were with a with and in with with NK cells, after with blocking in the were stained for the et al., for by a were with The were on a with a and then under a with NK cells, after with blocking in the were stained for the by using a were with The 3D were on a clear and on the using a or a on a was using a for a for a for was with or a of the NK tumor in 3D cultures, NB and NK cells were cultured or in cell cells were stained with and antibodies to NB cells from NK cells + Then, cells were stained with Viability and analyzed by flow flow NK cell cells were recovered from the 2D standard plate) culture or from the 3D was by dissolution, through in an alginate for in a were then with and and stained with a Viability NK cell infiltration within the 3D tumor culture, as well as their the cells were recovered from the alginate as were also recovered from the the within the NK and from the circulating compartment circulating NK The cells were then with and and with the for the cells were stained with an and an the analysis of tumor cell after dynamic the cells were recovered from the alginate and stained with an and an with the the cells were and through a were analyzed with the analysis was with the were by where or by analysis of was or by was have been with a of technical
Institutions: University of Milano-Bicocca, National Academies of Sciences, Engineering, and Medicine, Institute of Electronics, Computer and Telecommunication Engineering, National Research Council, University of Genoa, University of Milan, Ospedale Policlinico San Martino, Istituto Giannina Gaslini
Language: English
Subfield: Biomedical Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $3,801
OpenAlex ID: W7125399180
Cites:
Cited by:
Related to: