ICU patient-on-a-chip emulating orchestration of mast cells and cerebral organoids in neuroinflammation

communications biology Article https://doi.org/10.1038/s42003-024-07313-z ICU patient-on-a-chip emulating orchestration of mast cells and cerebral organoids in neuroinflammation Check for updates Pelin Saglam-Metiner 1,7,SenaYanasik 1,7, Yusuf Caglar Odabasi1, Jennifer Modamio2, Moritz Negwer2, Cigir Biray-Avci3,AyseGuler 4, Ali Erturk2, Ender Yildirim5,6 & Ozlem Yesil-Celiktas 1,6 Propofol and midazolam are the current standard of care for prolonged sedation in Intensive Care Units (ICUs). However, the effects and mechanism of these sedatives in brain tissue are unclear. Herein, the development of an ICU patient-on-a-chip platform to elucidate those effects is reported. The humanized neural tissue compartment combines mast cells differentiated from human induced pluripotent stem cells (hiPSCs) with cerebral organoids in a three-dimensional (3D) matrix, which is covered with a membrane populated with human cerebral microvascular endothelial cells (hCMEC/ D3) that separates the tissue chamber from the vascular lumen, where sedatives were infused for four days to evaluate neurotoxicity and cell-mediated immune responses. Subsequent to propofol administration, gene expressions of CD40 and TNF-αin mast cells, AIF1 in microglia and GFAP/ S100B/OLIG2/MBP in macroglia were elevated, as well as NOS2,CD80, CD40, CD68, IL6 and TNF-α mediated proinflammation is noted in cerebral organoids, which resulted in higher expressions of GJB1, GABA-A and NMDAR1 in the tissue construct of the platform. Besides, midazolam administration stimulated expression of CD40 and CD203c+reactivated mast cell proliferation and compromised BBB permeability and decreased TEER values with higher barrier disruption, whereas increased populations of CD11b+microglia, higher expressions of GFAP/DLG4/GJB1 and GABA-A-/ NMDAR1identities, as well as glutamate related neurotoxicity and IL1B, IFNG, IFNA1, IL6 genes mediated proinflammation, resulting in increased apoptotic zones are observed in cerebral organoids. These results suggest that different sedatives cause variations in cell type activation that modulate different pathways related to neuroinflammation andneurotoxicityintheICU patient-on-chipplatform. Neuroinflammation, a synonym of glial cell reactivity, is a complex cellular response of the immune system to various harmful stimuli in the central nervous system (CNS), which initiates the healing process to protect the brain’s cells and overall function1–3. In the case of neuroinflammation, the integrity of the blood-brain-barrier (BBB) can also be disrupted by the diffusion of proinflammatory cytokines and inflammatory cells into the brain tissue4,5.Inflammatory response facilitates the repairing of damage by eliminating infective agents through innate immune cells such as microglia, mast cells, astrocytes, T-cells and also inflammatory cytokines released from these cells2,6. Mast cells communicate with other CNS cells, extracellular matrix (ECM), and blood vessels, as they are located in the abluminal part (brain side) of the BBB. During development, mast cells circulate in the blood in immature form until they migrate to the brain where they complete their differentiation. Mature mast cells are also capable of migrating from the periphery to the brain. There are important findings supporting that reactivated mast cells disrupt the integrity of the BBB and act as catalysts with the mediators secreted in various circumstances such as stress, inflammation or trauma3,7. Moreover, secreted inflammatory mediators, neurotransmitters, and chemo-attractants interact with neurons and microglia through transgranulation, where they can transfer granules to 1Department of Bioengineering, Faculty of Engineering, Ege University, Izmir, Türkiye. 2Institute for Tissue Engineering and Regenerative Medicine (iTERM), Helmholtz Zentrum München, Neuherberg, Germany. 3Department of Medical Biology, Faculty of Medicine, Ege University, Bornova, Izmir, Türkiye. 4Department of Neuroscience, Faculty of Medicine, Ege University, Bornova, Izmir, Türkiye. 5Department of Mechanical Engineering, Middle East Technical University, Ankara, Türkiye. 6ODTÜ MEMS Center, Ankara, Türkiye. 7 These authors contributed equally: Pelin Saglam-Metiner, Sena Yanasik. e-mail: [email protected] Communications Biology | (2024) 7:1627 1 1234567890():,; 1234567890():,; other cells8,9.Whilemicrogliainitiateinflammatory responses, provide repair of the CNS and secrete proinflammatory mediators, they also respond to stimuli produced by degranulation and reactivation of other cells such as mast cells7. Indeed, some in vitro studies report direct microglial reactivation occurring as a result of mast cell degranulation10,11. Besidestheinflammatory response caused by the reactivation of microglia, astrocytes and mast cells has a protective effect, the persistent and prolonged inflammation inhibit neuronal regeneration leading to many neurodegenerative diseases, while the mechanisms still remain unclear12. Thus, realistic neuroinflammation outputs should be investigated in advanced preclinical models including both resident immune cells, the BBB structure and brain parenchyma construct3. The majority of the recent neuroinflammation studies are based on in vivo rodent models that summarize communication between glia and neurons13. However, while the neuronal migration and maturation is observed in around one week in mice, it occurs in an extended time exceeding 20 weeks in human neurogenesis14.Furthermore,thehuman brain has unique anatomical features such as an enlarged external subventricular region containing basal radial glia cells critical for cortical expansion, whereas mice lack this regionalization15. Thus, animal models cannot completely mimic the complex and unique structure of human brain physiology due to differences in molecular and cellular composition between species16,17.Ontheotherhand,thecommoninvitroneuroinflammation models elevate co-culturing of human primary, immortalized and/or induced pluripotent stem cells (iPSCs) differentiated CNScells in 2D manner. But these models lack a complex brain architecture, loss of tissuespecific mechanical properties with cell-cell interactions and can not emulate the complex communication between glial and neuronal cells in a 3D context18–21. Within the state-of-the-art technology, easy-to-setup 3D neural tissue models established with iPSCs-induced neural/glial cells, natural brain dECM and brain-on-a-chip platforms have gained importance2,22,23. Additionally, 3D unguided, guided and fused brain organoids generated from human iPSCs are utilized as advanced in vitro models that allow the study of realistic brain cellular interactions and structural architecture in recent years24–27. These models have significant potential not only for the development of new therapies28 but also investigation of cellular responses to drug administration at extended durations such as the interventions in ICUs. Sedatives are frequently used in mechanically ventilated ICU patients. While these drugs provide sedation by slowing down the CNS, the administration time, technique and dosage vary according to the clinical requirement of the patient. For instance, long-term infusion is required, especially for sedated patients with head trauma and status epilepticus29.In addition, patients on mechanical ventilation may have infusion periods longer than 48 h30. Propofol, a lipid-based agent used in ICU, is administered intravenously to maintain the hypnotic component, while midazolam, one of the most commonly used sedatives, is the most lipophilic benzodiazepine that rapidly crosses the BBB31,32.Propofol’s good pharmacokinetic properties, such as rapid onset of action and short half-life, make it an important alternative for reducing anxiety and agitation and improving tolerance to mechanical ventilation in critically ill patients admitted to the ICU33,34. Propofol and benzodiazepines enhance and prolong GABA binding to the GABA-A receptor located at postsynaptic sites with selective modulation35. Recent studies have reported that these GABA agonists may cause neurotoxicity and long-term cognitive problems36–38,inadditionto some studies showing protective effects at lower doses and administration times39–41. Besides, most in vitro studies focusing on the effects of sedatives on the CNS, are based on static cultures that lack the mimicry of blood and/ or cerebrospinal fluid circulation and 3D human brain physiology42. Therefore, the effect and mechanism of these sedative substances in the realistic biomimetic brain tissue model has not been fully elucidated. This study highlighted a newly designed ICU patient-on-a-chip platform that is consisted of a vascular chamber with a membrane lined up hCMEC/D3 to recapitulate the BBB, and a neural tissueformedbyco-culturingofhiPSCs differentiated mast cells and cerebral organoids in 3D matrix. We hypothesize that the administration of propofol and midazolam for specific durations to the on-chip platform allows investigation of the alterations in BBB and the cellular response in neural tissue of ICU patients. Results Numerical simulation of velocity and shear stress of ICU patienton-a-chip platform simplifies determination of operating flow condition Here, we present an easy to set multi-layered biomimetic microfluidic platform composed of biocompatible polydimethylsiloxane (PDMS) that closely models the BBB and 3D brain parenchyma, by recapitulating the interaction of cerebral organoids and mast cells (Fig. 1a). The BBB was simulated in the presence of shear stress caused by fluid movement in the vascularization layer formed by hCMEC/D3 cells seeded on the polyethylene terephthalate (PET) membrane. First, we conducted simulation studies to ascertain flow conditions preventing cell damage from shear stress while maintaining endothelium physiology. An oscillatory flow motion takes place in 30 sec on the rocker and the flow rate is 18 µL/s at 1 RPM. In the simulations performed under the fluid flow rate condition derived from the specified time, the maximum shear stress was observed to be 6.88 × 10−4Pa at a maximum velocity of 7.73 × 10−4m/s (Fig. 1b). We analyzed these values by performing simulation studies under different oscillatory flow conditions (5, 10 RPM) (Supplementary Fig. 1). The maximum shear stress is 2.97 × 10−3Pa at 5 RPM speed, whereas 7 × 10−3Pa at 10 RPM speed. Thus, we concluded that the increase in fluid flowrateincreasedtheshearstressonthewallsofthe vascular layer, where elevated levels would have caused damage to the endothelial cells. In order to provide the ideal physiological conditions for cerebral organoids, we carried out optimization studies using 1 RPM as the flow rate. When the 3D matrix was mimicked, we found the maximum velocity to be 9.4 × 10−4m/s yielding a maximum shear stress of 0.06 × 10−3Pa (Fig. 1c), which is ~50and 116-fold lower than the values obtained for 5 and 10 RPM (2.97 × 10−3,6.99×10 −3Pa), respectively. As for the tissue construct emulating brain parenchyma, we observed a homogeneous velocity and shear stress distribution in the matrix in which the organoids and mast cells would be embedded. Additionally, we simulated laminar flow conditions to reveal the difference between flow in a 3D matrix and free flow (Supplementary Fig. 1). In the presence of free fluid flow, the maximum velocity on the organoids was 8.04 × 10−4m/s, leading to an average shear stress of 1.24 × 10−3Pa. Although the shear stress increased due to porosity in the presence of the 3D matrix, its homogeneous distribution on the organoids showed the importance of the 3D extracellular matrix organization in the platform. We also predicted that the maximum velocity observed in the continuous free flow condition would have resulted in the movement of organoids with the flow, which again reveals the necessity of embedding the organoids in the hydrogel matrix. iPSC-derived mast cells that are co-cultured with cerebral organoids recapitulate resident immune cells in ICU patient-ona-chip We differentiated mast cells from iPSCs in order to understand mast cellrelated neural tissue response of an ICU patient to propofol and midazolam administrations. From the first days of the process, we observed mesodermal differentiation from iPSCs to start with changes in desired cell morphology on day 4 after BMP4 treatment. On day 6, immediately after culturing in mast cell (MC) StemPro-1 medium, we observed mesenchymal progenitor floating cells, which we expanded until day 27 with resuspension in MC StemPro-2 medium. From this stage, we cultured the floating cells in monolayer phase with MC monolayer medium for maturation of mast cells, while the progenitor floating cell culture could be continued until day 50 in MC StemPro-2 medium. We continued the proliferative monolayer culture until passage 10, when highly proliferative and specific cell morphologies were maintained (Fig. 2a). We characterized mast cell maturity by measuring expression levels of genes (qRT-PCR) and proteins by western blot (WB) and immunofluorescence staining (IF) (Fig. 2b–d). According to https://doi.org/10.1038/s42003-024-07313-z Article Communications Biology | (2024) 7:1627 2 qRT-PCR results (Fig. 2b, Supplementary Data 1a), both 27 days-old (floating cells) and 45–60 days old (monolayer cells) mast cells lost their stemness identities and gained hematopoietic immune cell character43,with significantly decreased expression levels of Yamanaka genes (SOX2, POU5F1, KLF4 and MYC) and significantly increased expression levels of PTPRC, CD34 and ITGAM genes, compared to (undifferentiated) control iPSCs. The increase over time of specificmaturemastcellsurfaceantigen gene, c-KIT,andhigh-affinity IgE-receptor specific gene, FES (15.4-fold and Fig. 1 | Design and simulation of the ICU patient-on-a-chip. a Layer-by-layer view of the ICU patient-on-a-chip platform. bnumerical simulation of velocity magnitude (m/s) and shear stress (Pa) over BBB at 1 RPM under laminar flow conditions, solved by Navier-stokes equations, and cover brain parenchyma at 1 RPM under 3D biomimetic matrix, solved by Brinkman equations. https://doi.org/10.1038/s42003-024-07313-z Article Communications Biology | (2024) 7:1627 3 Fig. 2 | Mimicry of neural tissue by co-culturing mast cells and cerebral organoids. a Mast cell differentiation process from iPSCs with brightfield microscopic images (scale bar = 100 μm for ×10 magnification, MOTIC AE31E), bHeatmap generated by qRT-PCR analysis of RNA samples isolated from floating cells on 27 days, and monolayer mast cells on 45 and 60 days compared to control iPSCs (independent repeats = 2) cWB analysis for CD203c protein of mature mast cells on day 60 (independent repeats = 2), dIF staining of mast cells on days 30, 45, and 60 of culture with CD203c in the mast cell differentiation protocol (scale bars = 200 µm for ×40 magnification images, independent repeats = 3, Zeiss Axio Vert.A1). eNewly designed ICU patient-on-a-chip platform with differentiated mast cells, µ-platform matured cerebral organoids on day 60 and hCMEC/D3 populated membrane, for the BBB evaluation and cellular response in neural tissue recapitulating the response of a patient in the ICU to sedatives. fBrief characterization of µ-platform matured cerebral organoids with whole confocal imaging after tissue clearing, highlighted by FOXG1, SOX2, TUJ1, SATB2, CTIP2, PAX6, CD11b, GFAP, OLIG2, and DAPI stainings (scale bars = 200 μm for ×5, 50 μm for ×20 and 20 μm for ×40 magnification images, independent replicates = 3, Zeiss LSM 880). https://doi.org/10.1038/s42003-024-07313-z Article Communications Biology | (2024) 7:1627 4 13.2-fold, respectively), also supported the success of the differentiation process. We found three important genes were to increase on day 60 of mast differentiation process: GRM1, a metabotropic glutamate receptor specific gene associated with neurodevelopmental disorders44.GJB1,ajunctional transmembrane protein specific gene associated with degranulation of mast cells45 and GABRA1, a specific receptor gene for controlling the majority of inhibitory signaling in the CNS46, which are related to maturation, reactivation and functionality of mast cells, significantly increased mostly on day 60 as 217.5-fold, 44.5-fold, and 6.2-fold, respectively. We also investigated miRNA expression, as those are key players in differentiation of mature mast cells. We measured an increase in MIR9-1 gene expression over time (4.5-fold on day 27, 13.7-fold on day 45 and 21.9-fold on day 60), which fits with miRNA upregulation over time during mast cell maturation47,48. Additionally, we observed upregulation on day 27 (41-fold) and downregulation on day 45 (6.6-fold) of another ionotropic glutamate receptor subunit specific gene, GRIN1, which has been previously described in mast cells in tendon repair49. According to WB and IF analyses, a specific cell surface protein of active mast cells, as a type II transmembrane protein, CD203c, increased over time throughout the differentiation process mostly on day 60 (Fig. 2c, d). Then, we co-cultured these characterized mast cells with 60 days old µ-platform matured cerebral organoids with FOXG1+/ SOX2+/TUJ1+forebrain, SATB2+/CTIP2+/PAX6+cortical plate and CD11b+/GFAP+/OLIG2+micro-macro glial cell identities (see details in our previous study26), and completed the ICU patient-on-a-chip by inserting PET membrane seeded with hCMEC/D3 endothelial cells (Fig. 2e, f). Sedative administration induces BBB permeability via endothelial cell destruction and cellular immune response In clinical practice, the infusion time of sedative agents may vary according to the patient and their response to treatment. Exposure to propofol for more than 48 h is called prolonged exposure and should not exceed a maximum of 7 days50. In order to emulate long-term repetitive sedative exposure in ICU patients (Fig. 3a) in our experimental setup, we first carried out dose-dependent sedative exposures on hCMEC/D3 endothelial cells. Subsequently, we administered the elicited safety doses of 200 μMpropofol and 25 μM midazolam in the co-culture media (1:1:1 ratio of endothelial cell, mast cell and cerebral organoid growth media) to the vascular lumen of hydrodynamic platform on days 1 and 2, and then harvested the culture supernatants, BBB layer and neural tissue construct of ICU patient-on-achip platform on day 4 for characterization. Specifically, we analyzed identity markers for mediated neuroinflammation associated with mast cells, microglia, macroglia and damaged neurons in neural tissue construct andalsoBBBlayerinICUplatform,intermsofbothgene(Supplementary Table 1) and protein expression levels (CD203c, CD31, ZO-1, vGLUT1, MAP2,CD11b,Ki-67,GABA-A,NMDAR1) (Fig. 3b). Within this context, we first determined the safety-dose to be administered to the ICU patienton-a-chip platform. We applied propofol and midazolam (400, 200, 100, 50, 25, 12.5, 6.25, 3.13 μM) to hCMEC/D3 cells for 4 days, observed them day by day under a brightfield microscope (Supplementary Fig. 2), and analyzed cell viabilities quantitatively on day 4 with MTT assay (Supplementary Fig. 3, Supplementary Data 1b). We elicited the safety doses of propofol and midazolam as 200 μMand25μM, respectively, since the cell morphologies were observed to be in appropriate shape and structure, whereas cell viabilities were above 70%, with significant differences from the highest doses (p< 0.0001 for midazolam and p< 0.05 for propofol). At elicited doses, we quantitatively evaluated barrier integrity with FITC permeability test and TEER measurement as gold standards51. We found a differential effect on BBB permeability, where 25 μM midazolam administration to the on-chip platform caused significant changes (p < 0.01) on BBB permeability with measured FITC value of 4.02 × 10−6cm/s (see the standard curve in Supplementary Fig. 4a, Supplementary Data 1c). In contrast, 200 μM propofol-administration has not yielded a significant difference in FITC permeability (3.4 × 10−6cm/s) compared to untreated controls (3.2 × 10−6cm/s) (Fig. 3c, Supplementary Data 1d). Similarly, representing the barrier integrity, TEER value of the control group was 421 Ωcm2, whereas significant decreases (p < 0.05) were noted in both sedative administered groups, mainly for the midazolam group (p< 0.01, 146 Ωcm2)(Fig.3d, Supplementary Data 1e). We also characterized the effect of sedatives on two important endothelial celljunction complexes, CD31 and ZO-1 markers26.Wefoundadecreasein viable CD31+endothelial cells after sedative exposure, as well as interruptions to the continuity of ZO-1 staining mostly subsequent to midazolam administration, which is in line with our increased permeability and decreased TEER values (Fig. 3e). Midazolam administration stimulates neuroinflammation and neural tissue defect via resident immune cell reactivation In the neural tissue compartment, we first investigated neuroinflammationassociated activation of mast cells. To this end, we harvested mast cells from ICU patient-on-a-chip platform on day 4 of sedative administration and cultured them on plate, followed by IF staining for CD203 and cell counting. We found increased CD203+reactivated mast cell population in the midazolam (average of 5.65 × 104cells/cm2,p< 0.001) administered group compared to the control (average of 3.52 × 104cells/cm2), while there was no significant change in numbers of mast cells after propofol (average of 2.73 × 104cells/cm2,p> 0.05) administration (Fig. 4a, Supplementary Fig. 5a, Supplementary Data 1f). In addition, we used qRT-PCR to evaluate a wide range of proinflammatory, anti-inflammatory, and specifically mast cell related gene expression levels in harvested cells from the ICU platform (Supplementary Table 1). We found that both sedative administrations resulted in increased expression levels of proinflammatory genes, such as CD40 (>50-fold), TNF (>15-fold) and NOS2 (>10-fold), and decreased levels of anti-inflammatory genes, most prominently of IL12B (>75-fold), IL13 (>25-fold) and IL10 (>15-fold). This indicates that sedative administration resulted in a reactivated mast cell-related neuroinflammatory response. While we observed no major changes in non-inflammatory related gene expression levels, GRIN1 gene levels were significantly decreased in both propofol (53.8-fold) and midazolam (15.8-fold) treated mast cells (Fig. 4b, Supplementary Data 1g). As resident immune cells, cellular sensor and effector in the CNS, microglia and mast cells have been mostly involved in neuroinflammation and neurodegeneration. Under these pathological states, they rapidly release excessive amounts of pathogenic immunomodulatory molecules, such as TNF-αand excitatory glutamate2,52.Thus,in order to quantify the extent of activated CNS immune cells in the neural tissue compartment following sedative exposure, we measured elevated TNF-αand glutamate concentrations as indicators of proinflammationmediated neuroinflammation from the culture supernatants using ELISA and specific glutamate kit (Fig. 4c, d). We found that TNF-αcytokine was released at significantly higher concentrations in both sedativeadministered groups (p< 0.05) compared to the control group. Interestingly, propofol showed more (p< 0.01) proinflammatory effects via TNF-α pathway (Fig. 4c, Supplementary Data 1h), which matched our qRPCR results. To investigate the effect of sedatives on immune cell-related glutamate expression, we measured glutamate levels in the culture supernatant with the calorimetric glutamate assay (see the standard curve in Supplementary Fig. 4b, Supplementary Data 1i). In comparison to the control group (24.2 ± 2.4 μM), the groups administered with midazolam (95 ± 4.9 μM, approximately fivefold, p< 0.001) and propofol (76 ± 2.4 μM, approximately threefold, p< 0.01) exhibited notably elevated releases of glutamate, suggesting the occurrence of neuroinflammation due to the reactivation of immune cells in the CNS (Fig. 4d, Supplementary Data 1j). Finally, in order to understand the changes in the brain parenchyma, we examined cerebral organoids harvested from the neural tissue compartment of the on-chip platform by WST-1 viability assay, qRT-PCR, vDISCO tissue clearing, and TUNEL apoptosis assay (Figs. 4e, g and 5). Although these concentrations of sedatives did not show cytotoxic effects (above 70% cell viability) on monolayer endothelial cells (Supplementary Fig. 3), their cytotoxic effect on cells within the main neural tissue in the ICU patient ona-chip platform was measured by decreased colorimetric absorbances in https://doi.org/10.1038/s42003-024-07313-z Article Communications Biology | (2024) 7:1627 5 parallel with cell viabilities, mostly of midazolam (p< 0.05), compared to control group (Fig. 4e, Supplementary Data 1k). These findings were also supported by qRT-PCR results where expression levels of the proliferative cell marker MKI67 gene were significantly decreased in both propofol (12.4fold) and midazolam (9-fold) administered organoids, and expression levels of the apoptosis marker CASP3 gene were specifically increased in the midazolam (5.2-fold) group (Fig. 4f, Supplementary Data 1l). As such, we found that propofol led to an increase in the gene expression levels of the macroglia markers GFAP (10.4-fold), S100B (10.1-fold), OLIG2 (4.4-fold), MBP (4-fold),aswellasthemicrogliamarkerAIF1 (6.2-fold). Midazolamadministration resulted in a similar but comparatively lower increase in the gene expression levels of both GFAP (7.2-fold) (macroglia) and ITGAM (5.7-fold) (microglia). Considering the genes associated with organoid maturation, we found that propofol-administration led to increases in the expression of GJB1 (25.3-fold) and GABRA1 (13.6-fold). midazolam administration exhibited a similar increase in GJB1 (7.2fold) and DLG4 (12.4-fold) mRNA, but a decrease in the expression levels of GRIN1 (73.7-fold), GABRA1(42-fold), and GLS (11-fold). Moreover, we observed increases mostly in the gene expression levels of pro-inflammatory markers NOS2 (19.2-fold), CD80 (4.4)-fold, and CD40 (3.7-fold) in the propofol-administered group. In contrast, with midazolam, we observed increases in IL1B (8.3-fold), IFNG (4.2-fold), and IFNA1 (3.3-fold) mRNA levels, which are also indicative of pro-inflammatory expression. With respect to a general pro-inflammatory shift, we observed significant decreases in anti-inflammatory-related gene expression (IL10, MRC1, and STAT6 genes, except ARG1)inbothsedativegroups(Fig.4f, Supplementary Fig. 3 | Effects on sedative administration permeability and barrier integrity. aSchematic illustration of sedative administration on ICU patient-on-a-chip platform. brecapitulating the interaction between BBB, mast cells, and cerebral organoids inflammatory response by cytokine release from neuronal, glial, and mast cells upon sedative administration; both aand bare generated with Servier Medical Art, provided by Servier, licensed under a Creative Commons Attribution 3.0 unported license. cFluorescein Isothiocyanate (FITC) permeability assay results of BBB construct at the end of 4 days of sedative administration as P (cm/s × 10−6), and dtrans-endothelial electrical resistance (TEER) measurement results at the end of 4 days of sedative administration as Ωcm2(One-way ANOVA Tukey’s multiple comparisons test, ns; p> 0.05, *p< 0.05, **p< 0.01, independent replicates = 3). eIF staining of CD31 and ZO-1 markers (scale bars = 20 μm for ×40 magnification images, independent replicates = 3, Leica Microsystems, Stellaris 5) for BBB construct at the end of 4 days of sedative administration in the ICU patient-on-a-chip. https://doi.org/10.1038/s42003-024-07313-z Article Communications Biology | (2024) 7:1627 6 Data 1l). We next wondered whether there would be spatial differences in neuronal markers across the organoids. To this end, we used wholeorganoid clearing with the vDISCO protocol and imaged vGLUT1 and MAP2, two neuronal markers for synapses and dendrites2,26, respectively, using a confocal microscope (Fig. 4g). Additionally, we performed mean fluorescence intensity analysis using confocal microscopy images to quantitatively corroborate the qualitative data (Supplementary Fig. 5b, Supplementary Data 1m). We observed that the protein expression levels of the Fig. 4 | Neuroinflammatory and neurotoxic effects of sedative administration on neural tissue. a IF staining of specific mast cell marker CD203c (scale bars = 20 μm for ×20 magnification images, independent replicates = 3). bHeatmap generated by fold regulation data obtained as a result of qRT-PCR analysis of RNA samples isolated from mast cells (independent repeats = 2). cTNF-αconcentrations (pg/mL) and dextracellular glutamate concentrations (µM) in the supernatant, and eWST-1 analysis of organoids (one-way ANOVA Tukey’s multiple comparisons test, ns; p> 0.05, *p< 0.05, **p< 0.01, ***p< 0.001, independent replicates = 2). fHeatmap generated by fold regulation data obtained as a result of qRT-PCR analysis of RNA samples isolated from organoids (independent repeats = 2). g3D confocal microscopy images of VGLUT1/MAP2 stained vDISCO tissue cleared cerebral organoids harvested from ICU patient-on-a-chip platform at the end of 4 days of sedative administration (scale bar = 500 μm independent replicates = 3, Zeiss LSM 880). https://doi.org/10.1038/s42003-024-07313-z Article Communications Biology | (2024) 7:1627 7 presynaptic protein vGLUT1 were clearly decreased in both sedativeadministered groups, mostly in propofol (p< 0.05) with a reduced fluorescence intensity of 1.71-fold, compared to the control. Besides, MAP2,a marker for dendrites in maturing neurons, was regionally decreased in both sedative administered groups, again mostly in propofol (p< 0.05) with a reduced fluorescence intensity of 1.34-fold, compared to the control. It is worth noting that a reduction in the expression of vGLUT1, a glutamate transporter located in the membrane of synaptic vesicles in the presynaptic terminal, has been linked to functional impairment following neuroinflammation, neurodegeneration and neuronal damage26,53–55. These results support our hypothesis that prolonged administration of propofol and midazolam in ICU patients might cause abnormal CNS immune cell reactivation-mediated neuroinflammation, resulting in neural cell death, BBB disruption and loss of function in neural tissue, due to increased neurotoxic glutamate release. Additionally, confocal microscopy and quantitative analysis of mean fluorescence intensity and cell counting were performed to examine some of the genes that stood out in the qRT-PCR analysis of cerebral organoids at the protein level (Fig. 5a, b, Supplementary Data 1n). GABA-A+/CD11b+/Ki67+/TUNEL−, which have also been previously characterized in detail at both protein and gene levels26, were found to exhibit an additionally active NMDAR1+cell population with organized tissue structure in the ICU patient-on-a-chip platform. The alterations in the tissue subsequent to sedative administration was characterized by decreased cellular viability with decreases in Ki-67 protein expressions (p< 0.0001, 18.12% less for propofol, 24.84% less for midazolam), in agreement with qRT-PCR results. While NMDAR1+(1.95-fold high, p< 0.001) and GABA-A+(1.2-fold high, p< 0.05) identities clearly increased with highly populated CD11b+ cells (2.32-fold high, p< 0.0001) in the propofol administered group, the increased CD11b+cell population (1.53-fold high, p< 0.001) was interestingly characterized by decreased NMDAR1 (regionally, 1.98-fold less, p< 0.0001) and GABA-A (1.35-fold less, p< 0.01) protein expressions after midazolam administration, compared to control group. On the other hand, the TUNEL fluorescent staining was performed to determine apoptotic regions, cell death and cell survival26 in cerebral organoids harvested from ICU patient-on-a-chip platform after 4 days (Fig. 5c, d, Supplementary Data 1o). DNAse-treated organoid as a positive control displayed DAPI and visible green fluorescence, validating the functionality of the assay. Cell apoptosis was characterized by labeled fragmented DNA breaks in TUNEL +/DAPI+cells at an apoptotic zone of ~106%. Almost no elevated rates of DNA breakage and cell apoptosis were observed in the organoids of the control group (7.5%), exhibiting more complex cytoarchitecture and regionalization. However, the presence of apoptotic cells in the peripheral regions of the organoids that were administered with sedatives, mostly with midazolam (p> 0.05 versus DNAse-treated group), significantly increased (apoptotic zone of 36%, p< 0.01, and 87%, p< 0.0001, for propofol and Fig. 5 | Further characterization of neural tissue from ICU patient-on-a-chip platform by IF and TUNEL after sedative administration. a 3D confocal microscopy images of Ki-67, NMDAR1, GABA-A, and CD11b stained vDISCO tissue cleared cerebral organoids harvested from ICU patient-on-a-chip platform at the end of 4 days of sedative administration (scale bars = 50 μm for ×20 and 20 μm for ×40 magnification images, independent replicates=3, Leica Microsystems, Stellaris 5). bQuantification of the mean fluorescence intensity signal of NMDAR1 and GABA-A stainings and Ki-67 +&CD11b+cell % versus DAPI+cell countings (One-way ANOVA Tukey’s multiple comparisons test, *p< 0.05, **p< 0.01, ***p< 0.001, ****p< 0.0001 independent replicates = 3–5). cTUNEL apoptosis fluorescent staining images of paraffin sectioned cerebral organoids harvested from ICU patient-on-a-chip platform at the end of 4 days of sedative administration. The DAPI-stained cell nucleus (blue), green fluorescent-stained apoptotic zones in the cell nucleus (green) (scale bar = 200 μm, independent replicates = 3, Zeiss Axio Vert.A1). dPercentage of apoptotic zone graph prepared from TUNEL images (independent replicates = 3, one-way ANOVA Tukey’s multiple comparisons test, ns; p> 0.05, **p< 0.01, ****p< 0.0001). https://doi.org/10.1038/s42003-024-07313-z Article Communications Biology | (2024) 7:1627 8 midazolam, respectively) due to neurotoxicity and cell-tissue destructive effects, in parallel with the Ki-67 and CASP3 expressions. On the other hand, we conducted a separate set of experiments in well plates to elucidate how cerebral organoids alone respond to the administration of propofol and midazolam without the presence of BBB and mast cells. We measured TNF-αand glutamate levels in culture supernatants, analyzed organoid viability with WST-1 assay and the changes in specific gene-protein expressions with qRT-PCR and whole confocal imaging after vDISCO clearing (Supplementary Fig. 6, Supplementary Data 1p-t). Although the neuroinflammatory response defined by increased TNF-α levels after propofol administration and elevated glutamate levels with decreased organoid viability after midazolam administration were parallel to the platform results, this phenomenon was approximately twofold higher in our ICU patient-on-chip platform with the presence of endothelial cell function in the BBB barrier and the reactivation of mast cells in the neural tissue construct. Similarly, although there was no statistically significant differences (<2.5-fold) in the up-down regulation of most genes compared to the control group, we found that propofol led to significant increases only in ITGAM (5.8-fold) microglia and NOS2 (2.8-fold) apoptotic gene expressions with significant decreases only in IL10 (3.5-fold) and MRC1 (5.75-fold) anti-inflammatory gene expressions, as expected. Moreover, we observed increases in ITGAM (8.3-fold) microglia and ARG1 (2.6-fold) gene expressions with significant decreases only in IL1B (2.6-fold) gene expression in the midazolam administered group. Confocal microscopy images also support qRT-PCR results, as in addition to regional GABA-A, vGLUT and MAP2 protein expression changes, only an increase in CD11b+ microglia population was noticeable after sedative exposure. This result proved that the response to sedative exposure was different in the presence of mast cells and BBB endothelial cells. Thus, the ICU patient-on-a-chip platform, which we designed to recapitulate in vivo-like brain tissue with the BBB barrier, mast cells and cerebral organoids has provided a physiological integrity that can be used for translational studies. Gene ontology enrichment analysis revealed the molecular pathways of sedative induced neuroinflammation processes For further investigation of propofol and midazolam administrations on both cerebral organoids and mast cells, we conducted Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) Enrichment analyses to qRT-PCR genes. We examined the relationship between genes from qRT-PCR and all three GO terms cellular component (CC), molecular function (MF) and biological process (BP). About 42 GO terms were detected for the propofol administered organoids and 47 GO terms for the midazolam counterparts (Supplementary Data 2a,b). When we examined the top enrichment GO terms of propofol group and related genes, “type 2 immune response”,“immune systems process”, and “nervous system development”GO terms showed IL10, CD40, MRC1, ITGAM, IFNG, IFNB1, ARG1, GFAP, GABRA1, S100B, IL6 and TNF genes related neuroinflammation (Fig. 6a,b).Inthemidazolamgroup,theGO terms as “transmembrane signaling receptor activity”,“signaling receptor activity”and “organic substance transport”,alsothegenesasITGAM, MRC1, CD40, SLC17A7, GRIN1, CD40/80, ARG1, GFAP, GABRA1, IL6 and IFNG stood out (Fig. 6c, d). However, there were no significant KEGG enriched pathways for organoids in the groups administered with both propofol and midazolam (Supplementary Data 2c,d). For the mast cell qPCR results, we found 32 GO terms for propofol administered mast cells and 44 GO terms for midazolam counterparts (Supplementary Data 2e,f). When we examined the top enrichment GO terms of propofol administered mast cells and related genes, positive regulation of both “B cell mediated immunity”and “immunoglobuliıtgamn mediated immune response”were elicited, whereas negative regulation of “inflammatory response”GO terms showed propofol drived neuroinflammation effect and CD40/80, GRIN1, IL12B,IFNG,IFNB1,TNF,STAT6,ARG1,IL1B,IL10,andIL13 genes were related to these GO terms (Fig. 7a, b). With respect to midazolam administered mast cells, the GO terms showed negative regulation of both “defense response”,and“apoptotic signaling pathway”, whereas positive regulation of both “NF-kappaB transcription factor activity”and “T-helper cell differentiation”and the genes IL13,IL10,IL12B,ENPP3,TNF,IFNB1, IL1A,IFNG,BCL2,andCD40/80belonging to these GO terms stood out (Fig. 7c, d). In addition to GO, we found 2 and 6 significant KEGG enriched pathways for propofol and midazolam administered mast cells, respectively (Supplementary Data 2g, h). Among them, “NF-kappaB signaling pathway” for propofol (Supplementary Fig. 7a), and “Necroptosis pathway”for midazolam administered mast cells (Supplementary Fig. 7b) were the most important pathways to associate their relationship with neuroinflammation process. Discussion Physiologically functional mast cells that are characterized mainly with CD203c, CD117, CD34 and FceR1 markers, play important roles in the regulation of almost every part of the neuroinflammatory process by transforming from protective immune cells to pro-inflammatory cells56,57. Mast cells can be differentiated from a variety of stem cells including iPSCs via mesodermal differentiation and be used as a peripheral brain parenchyma tissue-resident immune cell in in vitro models to evaluate neuroinflammation58,59. In our study, we successfully differentiated CD34/cKIT/FES higher expressed CD203c+mature mast cells from hiPSC. We then incorporated those into our easy to set, multi-layered and biomimetic ICU patient-on-a-chip platform that involves cerebral organoids derived from same origin iPSCs and hCMEC/D3 representing BBB and left the platform overnight to reach hemostasis. While sedatives are administered at various concentrations and durations, real-time analysis of the infusion of anesthesia in intensive care patients is not possible. This is important to determine the optimal sedative dose and maintain the continuity of anesthesia. To overcome this challenge, studies have examined the sedative concentrations of blood plasma from human blood. In one of these studies, the concentration of midazolam in blood plasma of a patient was reported to decrease to 0.2 μg/ml at the end of 1 hour, to whom intravenous midazolam was administered at a bolus dose of 0.2 mg/kg for induction of general anesthesia60. Another study focused on the concentrations of propofol in blood plasma and brain tissue. The initial plasma concentration of propofol was approximately 2.5 μg/ml and 3.5 μg/g in the brain. The plasma and brain concentrations decreased gradually with time, but the brain concentration was always higher than the plasma concentration61. Likewise, focusing on propofol concentrations in human plasma and brain tissue, concentrations in brain tissue were reported to be higher than plasma concentrations62. In our study, we administered sedatives at concentrations of 200 μMand25μM for propofol and midazolam, respectively, elicited as safety doses for endothelial cells. We repeated this administration on day 2 to evaluate the effects of prolonged sedative administration and harvested on day 4. The immune cell-mediated potential neuroinflammatory and neurotoxic effect of sedative administration were evaluated by analyzing 4 different samples harvested from our on-chip platforms: (i) Permeability, TEER values and IF analysis of BBB construct, (ii) IF and qRT-PCR analysis of mast cells in the neural tissue construct, (iii) TNF-αand glutamate level analyses in culture supernatants, and finally, (iv) WST-1 tissue viability, qRT-PCR, whole confocal imaging after vDISCO clearing and TUNEL analyses of organoids in the neural tissue construct. The studies have shown the reactivation of mast cells in case of neuroinflammation were the earliest activators of BBB damage and precede both macroglial and microglial cell reactivation63–65. Therefore, we investigated the changes in the BBB integrity in terms of cell adhesion, junctions, permeability, TEER values and proinflammatory state, as well as the reactivation status of mast cells. The impaired barrier integrity with higher FITC permeability and lower TEER values due to decreases in CD31 and ZO-1 junctional complexes in endothelial cells and increase in the TNF-αexpression levels after mainly midazolam administration, reveals the interactive nature between mast cells and BBB construct during mast cell mediated neuroinflammation. On the other hand, propofol exposure has not revealed such a significant alteration on BBB integrity and irregular mast cell reactivation, which is in agreement https://doi.org/10.1038/s42003-024-07313-z Article Communications Biology | (2024) 7:1627 9 with DAPI containing mounting media, and imaged with confocal microscopy (Zeiss LSM 880). WST-1 assay. Cell proliferation analyses of organoids in on-chip platforms were quantitatively performed with the WST-1 test based on formazan crystals released due to mitochondrial activity. Following the 1:10 rule found in the WST 1 analysis, 200 µL of nutrient medium/20 µL of reagent was added to each group. Then, the absorbance values of the organoids incubated at 37 °C for 2 h were obtained with a fluorescent microplate reader. ELISA assay. The TNF-αmediated neurotoxic effect of sedatives in ICU patient-on-a-chip platforms was examined by determining the amount of TNF-αin the supernatant collected from the platform at the end of the 4th day, by following the standard protocols of the human TNF-αELISA kit (EH0302, FineTest). Glutamate assay. Releasing glutamate levels from ICU patient-on-achip platform after sedative administrations were evaluated for the immune cell-mediated neurotoxic effect of sedatives. In this context, after 4 days of treatment, the supernatants of the platforms were collected and analyzed with the colorimetric glutamate kit (MAK330-1K, SigmaAldrich) according to the manufacturer’s instructions. Glutamate levels (mM) were quantitatively analyzed by GraphPad Prism 8.3.0 program via standard curve (Supplementary Fig. 4b, Supplementary Data 1i). qRT-PCR analysis. Total RNAs were extracted from both mast cells and cerebral organoids from on-chip platforms. The qRT-PCR analysis was performed for target genes including proinflammatory; TNF (TNF-α), IL6 (IL-6), NOS2 (Inos), CD40, CD68, CD80, CD86, IL1A (IL-1α), IL1B (IL-1β), IFNA1 (IFN-α), IFNB1 (IFN-β), IFNG (IFN-γ), anti-inflammatory; IL5 (IL-5), IL10 (IL-10), IL12B(IL-12), IL13 (IL-13), MRC1 (CD206), STAT6, ARG1 and non-inflammatory related; ENPP3 (CD203c), SLC17A7, FES, GRIN1, CASP3 (Caspas3), c-KIT, CD34, GLS (Glutaminase), BCL2 (BAX), GJB1, PTPRC, GRM1, GABRA1, MIR9-1, MKI67 (Ki67), CASP3 genes for mast cells, and also neural/glial cell type specific; MAP2, RBFOX3 (NEUN), OLIG2, MBP, GFAP, S100B, ITGAM, AIF1 (IBA1), non-inflammatory releated; TUBB3 (TUJ1), PAX6, EOMES (TBR2), FOXG1, DLG4 (PSD95), GRIN1, GLS, GJB1, SLC17A7, GABRA1, proinflammatory; TNF, IL6, NOS2, CD40, CD68, CD80, IL1A, IL1B, IFNA1, IFNB1, IFNG and anti-inflammatory: IL10, MRC1, STAT6, ARG1, genes for organoids (Supplementary Table S1) as previously described. Tissue clearing. The organoids in the ICU patient-on-a-chip platforms with and without sedative administered groups carried out under dynamic conditions were evaluated with the 3D vDISCO tissue clearing method to monitor changes in protein levels. For this, cerebral organoids embedded in a 3D matrix in the platform were fixed with 4% PFA for 2 h at 4 °C, after cold PBS washing for matrigel removal. Then, washed with PBT solution (0.1% (v/v) Tween-20 in PBS) 3 times for 30 min and incubated overnight at 4 °C. The next day, samples were blocked in iMarika solution (10% FBS, 2% BSA, 0.05% sodium azide (SigmaAldrich, 71290, USA), 0.5% Triton-X 100 and 5% DMSO in PBT) for 6 h at 4 °C. Then, anti-FOXG1 (Abcam, ab18259, 1:500), anti-SOX2 (Millipore, SC1002, 1:500), anti-TUJ1 (Novus, NB100-161, 1:500), anti-SATB2 (Abcam, ab51502, 1:500), anti-CTIP2 (Abcam, ab18465, 1:500), antiPAX6 (Thermo, 426600, 1:500), anti-CD11b (Abcam, ab8878, 1:500), anti-GFAP (Affinity, DF6040, 1:200), anti-OLIG2 (IBL, 18953, 1:500), anti-MAP2 (Abcam, ab5392, 1:500), anti-vGLUT1 (Thermo Scientific, 48-2400, 1:500), anti-NMDAR1 (Affinity, AF6406, 1:200), anti-GABA-A (Affinity, Ab6207, 1:200) and anti-Ki67 (SantaCruz, sc-23900, 1:500) primary antibodies were diluted in the iMarika solution and samples were incubated in it on a shaker at 4°C for 3 days. After washing with iMarika solution three times for an hour at +4 °C, the organoids were labeled with secondary antibodies (Alexa fluor 488, 594 and 647 conjugated) diluted in iMarika solution (1:1000) on a shaker at 4 °C for additional 3 days. Thereafter, the organoids were washed with iMarika solution three times for an hour at +4 °C and were stained with DAPI solution diluted 1:3000 in iMarika solution at RT for an hour. After the washing steps 2 times in iMarika solution and in PBS, samples were left in PBS until the dehydration and clearing steps. For the dehydration step, labeled organoids were first kept in 50% Tetrahydrofuran (THF, Sigma-Aldrich, 186562, USA) solution in water overnight, and they were passed through a series of 50%, 70%, 90% and 100% THF solution in water for an hour each. Then, they were kept in 100% THF again overnight. The next day, dehydrated organoids were incubated in a dichloromethane solution (Sigma-Aldrich, 270997, USA) in corrosion-resistant polypropylene tubes under a hood for ~15 min until collapse. Then, the samples were cleared repeatedly in ethyl cinnamate solution (Sigma-Aldrich, 66761, USA) for 3 h to transparency of the tissue. In the final stage, image analysis of whole organoids (without sectioning) in this solution was performed by confocal microscopy (Zeiss LSM880)102–104. TUNEL apoptosis test. To examine apoptotic regions in sedativeadministered organoids in ICU patient-on-a-chip platform, DeadEnd™ Fluorometric TUNEL System (Promega, G3250) assay was carried out. For this, the paraffin-embedded sections of organoids were washed twice in xylene, 1× PBS, 100% ethanol, ethanol dilutions down to 50%, and 0.85% NaCl for 5 min. After rewashing in PBS, sections were incubated in 4% formaldehyde for 15 min, washed in PBS again and 20 μg/mL Proteinase K was dropped on samples for 10 min. After the 4% formaldehyde step was repeated, the positive control group was treated with 1× DNAse solution. Then, sections were equilibrated with a 100 μL equilibration buffer for 10 min at RT, were labeled with 50 μL of TdT reaction mix, and incubated for 60 min at 37 °C in a humidified chamber in the dark. To stop the reaction, sections were washed in 2× SSC for 15 min and in PBS for 5 min. Finally, to visualize all nuclei in tissue, sections were stained with 1:1000 DAPI for 10 min, washed in PBS again, and mounted with mounting medium, and the localized green fluorescence of apoptotic cells in tissue was detected by fluorescence microscopy (Zeiss, Axio Vert.A1). GO and KEGG enrichment analysis. Functions ‘gseGO and ‘gseKEGG in R package ‘clusterprofiler v4.8.3’were used for GO and KEGG analysis of target genes of qRT-PCR105. All GO terms (CC, MF, and BP) were selected. The parameters minimum genesets size, maximum genesets size and pvalue were set 3, 800 and 0.05, respectively. To visualize GO and KEGG results, functions ‘dotplot and ‘cnetplot and R package pathview’ were used. Statistics and reproducibility. All data are representative of at least two different biological replicates and were statistically analyzed by one-way or two-ways ANOVA, Tukey’s or Dunnett’s multiple comparisons test with ±95% confidence interval and pvalues < 0.05 were considered statistically significant in GraphPad Prism 8.3.0 program. Data was presented as the mean ± standard deviation with individual data points. In order to indicate degree of significance, ns; p> 0.05, *p< 0.05, **p< 0.01, ***p< 0.001, ****p< 0.0001 were used. The Student’sttest was also used for qRTPCR analysis (independent replicates = 2 with RNA isolate pool containing at least two samples from the same batch). Reporting summary Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article. Data availability ThesourcedatabehindthegraphsinthepapercanbefoundinSupplementary Data 1. https://doi.org/10.1038/s42003-024-07313-z Article Communications Biology | (2024) 7:1627 16 Code availability GO and KEGG enrichment analysis: https://github.com/Odabasi-Yusuf/ ICU-GO. Received: 8 March 2024; Accepted: 22 November 2024; References 1. Diaz-Castro, B., Bernstein, A. M., Coppola, G., Sofroniew, M. V. & Khakh, B. S. Molecular and functional properties of cortical astrocytes during peripherally induced neuroinflammation.Cell Rep. 36, 109508 (2021). 2. Saglam-Metiner, P., Duran, E., Sabour-Takanlou, L., Biray-Avci, C. & Yesil-Celiktas, O. Differentiation of neurons, astrocytes, oligodendrocytes and microglia from human induced pluripotent stem cells to form neural tissue-on-chip: a neuroinflammation model to evaluatethetherapeuticpotentialofextracellularvesiclesderivedfrom mesenchymal stem cells. Stem Cell Rev. Rep. 20, 413–436 (2024). 3. Hendriksen, E., van Bergeijk, D., Oosting, R. S. & Redegeld, F. A. Mast cells in neuroinflammation and brain disorders. Neurosci. Biobehav. Rev. 79, 119–133 (2017). 4. Cecen, B. et al. Biosensor integrated brain-on-a-chip platforms: progress and prospects in clinical translation. Biosens. Bioelectron. 225, 115100 (2023). 5. Kempuraj, D. et al. Brain and peripheral atypical inflammatory mediators potentiate neuroinflammation and neurodegeneration. Front. Cell Neurosci. 11, 216 (2017). 6. Kempuraj, D. et al. Neuroinflammation induces neurodegeneration. J. Neurol. Neurosurg. Spine 1, 1003 (2016). 7. Dong, H., Zhang, X. & Qian, Y. Mast cells and neuroinflammation. Med. Sci. Monit. Basic Res. 20, 200–206 (2014). 8. Germundson, D. L. & Nagamoto-Combs, K. Potential role of intracranial mast cells in neuroinflammation and neuropathology associated with food allergy. Cells 11, 738 (2022). 9. Mittal, A., Sagi, V., Gupta, M. & Gupta, K. Mast cell neural interactions in health and disease. Front. Cell Neurosci. 13, 110 (2019). 10. Dong, H. et al. Suppression of brain mast cells degranulation inhibits microglial activation and central nervous system inflammation. Mol. Neurobiol. 54, 997–1007 (2017). 11. Dong, H. et al. Stabilization of brain mast cells alleviates LPSinduced neuroinflammation by inhibiting microglia activation. Front. Cell Neurosci. 13, 191 (2019). 12. Kwon,H.S.&Koh,S.-H.Neuroinflammation in neurodegenerative disorders: the roles of microglia and astrocytes. Transl. Neurodegener. 9,42(2020). 13. Ormel, P. R. et al. Microglia innately develop within cerebral organoids. Nat. Commun. 9, 4167 (2018). 14. Toma, K., Wang, T. & Hanashima, C. Encoding and decoding time in neural development. Dev. Growth Differ. 58,59–72 (2016). 15. Fair, S. R. et al. Electrophysiological maturation of cerebral organoids correlates with dynamic morphological and cellular development. Stem Cell Rep. 15, 855–868 (2020). 16. Sun, N. et al. Applications of brain organoids in neurodevelopment and neurological diseases. J. Biomed. Sci. 28, 30 (2021). 17. Mak, I. W., Evaniew, N. & Ghert, M. Lost in translation: animalmodels and clinical trials in cancer treatment. Am. J. Transl. Res. 6, 114–118 (2014). 18. Cui, X., Hartanto, Y. & Zhang, H. Advances in multicellular spheroids formation. J. R. Soc. Interface 14, 20160877 (2017). 19. Peng, H., Harvey, B. T., Richards, C. I. & Nixon, K. Neuron-derived extracellular vesicles modulate microglia activation and function. Biology 10, 948 (2021). 20. Ao, Z. et al. Tubular human brain organoids to model microgliamediated neuroinflammation. Lab Chip 21, 2751–2762 (2021). 21. Bang, S., Lee, S., Choi, N. & Kim, H. N. Emerging brain‐ pathophysiology‐mimetic platforms for studying neurodegenerative diseases: brain organoids and brains‐on‐a‐chip. Adv. Healthc. Mater. 10, e2002119 (2021). 22. Yaldiz, B., Saglam-Metiner, P., Cam, S. B., Korkusuz, P. & YesilCeliktas, O. Effect of sterilization methods on the mechanical stability and extracellular matrix constituents of decellularized brain tissues. J. Supercrit. Fluids 175, 105299 (2021). 23. Ingber, D. E. Human organs-on-chips for disease modelling, drug development and personalized medicine. Nat. Rev. Genet. 23, 467–491 (2022). 24. Lancaster, M. A. et al. Cerebral organoids model human brain development and microcephaly. Nature 501, 373–379 (2013). 25. Pașca, S. P. et al. A nomenclature consensus for nervous system organoids and assembloids. Nature 609, 907–910 (2022). 26. Saglam-Metiner, P. et al. Spatio-temporal dynamics enhance cellular diversity, neuronal function and further maturation of human cerebral organoids. Commun. Biol. 6,173 (2023). 27. Cho, A. N. et al. Microfluidic device with brain extracellular matrix promotes structural and functional maturation of human brain organoids. Nat. Commun. 12, 4730 (2021). 28. Sabate-Soler, S. et al. Microglia integration into human midbrain organoids leads to increased neuronal maturation and functionality. Glia 70, 1267–1288 (2022). 29. Muhlhofer, W. G. et al. Duration of therapeutic coma and outcome of refractory status epilepticus. Epilepsia 60, 921–934 (2019). 30. Dolinay, T. et al. Ventilator weaning in prolonged mechanical ventilation—a narrative review. J. Clin. Med. 13, 1909 (2024). 31. Griffin, C. E., Kaye, A. M., Bueno, F. R. & Kaye, A. D. Benzodiazepine pharmacology and central nervous system-mediated effects. Ochsner J. 13, 214–223 (2013). 32. Sahinovic, M. M., Struys, M. M. R. F. & Absalom, A. R. Clinical pharmacokinetics and pharmacodynamics of propofol. Clin. Pharmacokinet. 57, 1539–1558 (2018). 33. Slingerland-Boot, R.,Kummerow, M., Arbous, S. M. & van Zanten,A. R. H. Association between first-week propofol administration and long-term outcomes of critically ill mechanically ventilated patients: a retrospective cohort study. Clin. Nutr. 43,42–51 (2024). 34. Singh, A. & Anjankar, A. P. Propofol-related infusion syndrome: a clinical review. Cureus 14, e30383 (2022). 35. Jurd, R. et al. General anesthetic actions in vivo strongly attenuated by a point mutation in the GABA(A) receptor beta3 subunit. FASEB J. 17, 250–252 (2003). 36. Hughes, J. M. etal. The effects of propofolon a human in vitro bloodbrain barrier model. Front. Cell Neurosci. 16, 835649 (2022). 37. Twaroski, D. M., Yan, Y., Olson, J. M., Bosnjak, Z. J. & Bai, X. Downregulation of MicroRNA-21 Is involved in the propofol-induced neurotoxicity observed in human stem cell-derived neurons. Anesthesiology 121, 786–800 (2014). 38. Bosnjak, Z. J., Logan, S., Liu, Y. & Bai, X. Recent insights into molecular mechanisms of propofol-induced developmental neurotoxicity: implications for the protective strategies. Anesth. Analg. 123, 1286–1296 (2016). 39. Zheng, J. et al. Midazolam ameliorates impairment of the blood–brain barrier (BBB) against LPS. Neurotox. Res. 40, 751–762 (2022). 40. Han, X. et al. The protective effect of propofol on ionizing radiationinduced hematopoietic system damage in mice. RSC Adv. 9, 36366–36373 (2019). 41. Li, X., Yao, L., Liang, Q., Qu, H. & Cai, H. Propofol protects hippocampal neurons from hypoxia-reoxygenation injury by decreasing calcineurin-induced calcium overload and activating YAP signaling. Oxid. Med. Cell Longev. 2018,1–12 (2018). 42. Dominguini, D. et al. The effects of anaesthetics and sedatives on brain inflammation. Neurosci. Biobehav. Rev. 127, 504–513 (2021). https://doi.org/10.1038/s42003-024-07313-z Article Communications Biology | (2024) 7:1627 17 43. Lim, W. F., Inoue-Yokoo, T., Tan, K. S., Lai, M. I. & Sugiyama, D. Hematopoietic cell differentiation from embryonic and induced pluripotent stem cells. Stem Cell Res. Ther. 4, 71 (2013). 44. Crupi, R., Impellizzeri, D. & Cuzzocrea, S. Role of metabotropic glutamate receptors in neurological disorders. Front. Mol. Neurosci. 12, 20 (2019). 45. Harcha, P. A. et al. Mast cell and astrocyte hemichannels and their roleinAlzheimer’s disease,ALS,andharmful stress conditions. Int J. Mol. Sci. 22, 1924 (2021). 46. Wen, Y. et al. Glutamate and GABAA receptor crosstalk mediates homeostatic regulation of neuronal excitation in the mammalian brain. Signal. Transduct. Target. Ther. 7, 340 (2022). 47. Mayoral, R. J. et al. MicroRNA-221–222 regulate the cell cycle in mast cells. J. Immunol. 182, 433–445 (2009). 48. Fenger, J. M. et al. Overexpression of miR-9 in mast cells is associated with invasive behavior and spontaneous metastasis. BMC Cancer 14, 84 (2014). 49. Alim, M. A. et al. Increased mast cell degranulation and colocalization of mast cells with the NMDA receptor-1 during healing after Achilles tendon rupture. Cell Tissue Res 370, 451–460 (2017). 50. Fodale, V. & La Monaca, E. Propofol infusion syndrome. Drug Saf. 31, 293–303 (2008). 51. Goksel, O. et al. Comprehensive analysis of resilience of human airway epithelial barrier against short‐term PM2.5 inorganic dust exposure using in vitro microfluidic chip and ex vivo human airway models. Allergy https://doi.org/10.1111/all.16179 (2024). 52. Sandhu, J. K. & Kulka, M. Decoding mast cell-microglia communication in neurodegenerative diseases. Int. J. Mol. Sci. 22, 1093 (2021). 53. Ormel, L., Stensrud, M. J., Bergersen, L. H. & Gundersen, V. VGLUT1 is localized in astrocytic processes in several brain regions. Glia 60, 229–238 (2012). 54. Rodriguez‐Perdigon, M. et al. Down‐regulation of glutamatergic terminals (VGLUT1) driven by Aβin Alzheimer’s disease. Hippocampus 26, 1303–1312 (2016). 55. Xu, J. et al. Early developmental exposure to repetitive long duration of midazolam sedation causes behavioral and synaptic alterations in a rodent model of neurodevelopment. J. Neurosurg. Anesthesiol. 31, 151–162 (2019). 56. Grootens, J., Ungerstedt, J. S., Nilsson, G. & Dahlin, J. S. Deciphering the differentiation trajectory from hematopoietic stem cells to mast cells. Blood Adv. 2, 2273–2281 (2018). 57. Ribatti, D. & d’Amati, A. Hematopoiesis and mast cell development. Int J. Mol. Sci. 24, 10679 (2023). 58. Igarashi, A. et al. Mast cells derived from human induced pluripotent stem cells are useful for allergen tests. Allergol. Int. 67, 234–242 (2018). 59. Luo, Y. et al. A novel approach for studying mast cell–driven disorders: mast cells derived from induced pluripotent stem cells. J. Allergy Clin. Immunol. 149, 1060–1068.e4 (2022). 60. Miyake, W. et al. Electroencephalographic response following midazolam-induced general anesthesia: relationship to plasma and effect-site midazolam concentrations. J. Anesth. 24, 386–393 (2010). 61. Kawata, M.et al. Development ofextended pharmacokinetic models for propofol based on measured blood and brain concentrations. Sci. Rep. 14, 6326 (2024). 62. Van Hese, L., Theys, T., Absalom, A. R., Rex, S. & Cuypers, E. Comparison of predicted and real propofol and remifentanil concentrations in plasma and brain tissue during target‐controlled infusion: a prospective observational study. Anaesthesia 75, 1626–1634 (2020). 63. Wang,Y.etal.Themastcellisanearlyactivatoroflipopolysaccharideinduced neuroinflammation and blood-brain barrier dysfunction in the hippocampus. Mediators Inflamm. 2020, 8098439 (2020). 64. Versele, R., Sevin, E., Gosselet, F., Fenart, L. & Candela, P. TNF-α and IL-1βmodulate blood-brain barrier permeability and decrease amyloid-βpeptide efflux in a human blood-brain barrier model. Int. J. Mol. Sci. 23, 10235 (2022). 65. Takata, F., Nakagawa, S., Matsumoto, J. & Dohgu, S. Blood-brain barrier dysfunction amplifies the development of neuroinflammation: understanding of cellular events in brain microvascular endothelial cells for prevention and treatment of BBB dysfunction. Front. Cell Neurosci. 15, 661838 (2021). 66. Längrich, T. et al. Disturbance of key cellular subproteomes upon propofol treatment is associated with increased permeability of the blood-brain barrier. Proteomes 10, 28 (2022). 67. Iskarpatyoti, J. A. et al. Mast cell regranulation requires a metabolic switch involving mTORC1 and a glucose-6-phosphate transporter. Cell Rep. 40, 111346 (2022). 68. Jin, Y., Silverman, A. J. & Vannucci, S. J. Mast cells are early responders after hypoxia-ischemia in immature rat brain. Stroke 40, 3107–3112 (2009). 69. Bidri, M. et al. Inhibition of mouse mast cell proliferation and proinflammatory mediator release by benzodiazepines. Immunopharmacology 43,75–86 (1999). 70. Xu, D. et al. Midazolam exposure impedes oligodendrocyte development via the translocator protein and impairs myelination in larval zebrafish. Mol. Neurobiol. 59,93–106 (2022). 71. Zhang,X.et al.Activatedbrainmastcellscontributeto postoperative cognitive dysfunction by evoking microglia activation and neuronal apoptosis. J. Neuroinflamm. 13, 127 (2016). 72. Chen, Y. Z. et al. Magnetic resonance imaging of glutamate in neuroinflammation. Radiol. Infect. Dis. 3,92–97 (2016). 73. Sall, J. W., Stratmann, G., Leong, J., Woodward, E. & Bickler, P. E. Propofol at clinically relevant concentrations increases neuronal differentiation but is not toxic to hippocampal neural precursor cells in vitro. Anesthesiology 117, 1080–1090 (2012). 74. Paolicelli, R. C. et al. Microglia states and nomenclature: a field at its crossroads. Neuron 110, 3458–3483 (2022). 75. Wittekindt, M. et al. Different methods for evaluating microglial activation using anti-ionized calcium-binding adaptor protein-1 immunohistochemistry in the cuprizone model. Cells 11, 1723 (2022). 76. Qiao, H. et al. Propofol affects neurodegeneration and neurogenesis by regulation of autophagy via effects on intracellular calcium homeostasis. Anesthesiology 127, 490–501 (2017). 77. Guo, S., Wang, H. & Yin, Y. Microglia polarization from M1 to M2 in neurodegenerative diseases. Front. Aging Neurosci. 14, 815347 (2022). 78. Krishnan, M. L. et al. Integrative genomics of microglia implicates DLG4 (PSD95) in the white matter development of preterm infant. Nat. Commun. 8, 428 (2017). 79. Kantamneni, S. Cross-talk and regulation between glutamate and GABAB receptors. Front. Cell Neurosci. 9, 135 (2015). 80. Kotani, Y., Shimazawa, M., Yoshimura, S., Iwama, T. & Hara, H. The experimental and clinical pharmacology of propofol, an anesthetic agent with neuroprotective properties. CNS Neurosci. Ther. 14, 95–106 (2008). 81. Wang,S.et al.Thechanging ofα5‐GABAA receptorsexpressionand distribution participate in sevoflurane‐induced learning and memory impairment in young mice. CNS Neurosci. Ther. 30, e14716 (2024). 82. Zurek, A. A., Bridgwater, E. M. & Orser, B. A. Inhibition of α5γaminobutyric acid type A receptors restores recognition memory after general anesthesia. Anesth. Analg. 114, 845–855 (2012). 83. Zhang, S. et al. Cerebral mast cells contribute to postoperative cognitive dysfunction by promoting blood brain barrier disruption. Behav. Brain Res. 298, 158–166 (2016). 84. Xu, F. et al. Prolonged anesthesia induces neuroinflammation and complement-mediated microglial synaptic elimination involved in https://doi.org/10.1038/s42003-024-07313-z Article Communications Biology | (2024) 7:1627 18 neurocognitive dysfunction and anxiety-like behaviors. BMC Med. 21, 7 (2023). 85. Helmy, S. A. K. & Al-Attiyah, R. J. The immunomodulatory effects of prolonged intravenous infusion of propofol versus midazolam in critically ill surgical patients. Anaesthesia 56,4–8 (2001). 86. Ots, H. D., Tracz, J. A., Vinokuroff, K. E. & Musto, A. E. CD40–CD40L in neurological disease. Int. J. Mol. Sci. 23, 4115 (2022). 87. Kim,D. Y., Jeoung,D. & Ro,J.Y. Signaling pathwaysin theactivation of mast cells cocultured with astrocytes and colocalization of both cells in experimental allergic encephalomyelitis. J. Immunol. 185, 273–283 (2010). 88. Kim, D. Y., Hong, G. U. & Ro, J. Y. Signal pathways in astrocytes activated by cross-talk between of astrocytes and mast cells through CD40-CD40L. J. Neuroinflamm. 8, 25 (2011). 89. Suurmond, J. et al. Communication between human mast cells and CD4+T cells through antigen-dependent interactions. Eur. J. Immunol. 43, 1758–1768 (2013). 90. Ottum, P. A., Arellano, G., Reyes, L. I., Iruretagoyena, M. & Naves, R. Opposing roles of interferon-gamma on cells of the central nervous system in autoimmune neuroinflammation. Front. Immunol. 6, 539 (2015). 91. Abellanas, M. A. et al. Midbrain microglia mediate a specific immunosuppressive response under inflammatory conditions. J. Neuroinflamm. 16, 233 (2019). 92. Suurmond, J., Dorjée, A. L., Huizinga, T. W. J. & Toes, R. E. M. Human mast cells costimulate T cells through a CD28-independent interaction. Eur. J. Immunol. 46, 1132–1141 (2016). 93. Zhou, X. Y., Chen, K. & Zhang, J. A. Mast cells as important regulators in the development of psoriasis. Front. Immunol. 13, 1022986 (2022). 94. Ali, M. Y., Knight, D. & Howlader, M. M. R. Nonenzymatic electrochemical glutamate sensor using copper oxide nanomaterials and multiwall carbon nanotubes. Biosens. (Basel) 13, 237 (2023). 95. Zeynaloo, E. et al. Design of a mediator-free, non-enzymatic electrochemical biosensor for glutamate detection. Nanomedicine 31, 102305 (2021). 96. Soldatkina, O. V. et al. A novel amperometric glutamate biosensor based on glutamate oxidase adsorbed on silicalite. Nanoscale Res. Lett. 12, 260 (2017). 97. Sun, Y. et al. In vivo glutamate sensing inside the mouse brain with perovskite nickelate–nafion heterostructures. ACS Appl Mater. Interfaces 12, 24564–24574 (2020). 98. Ganesana, M., Trikantzopoulos, E., Maniar, Y., Lee, S. T. & Venton, B. J. Development of a novel micro biosensor for in vivo monitoring ofglutamate releaseinthebrain. Biosens. Bioelectron. 130, 103–109 (2019). 99. Spitz, S., Schobesberger, S., Brandauer, K. & Ertl, P. Sensor‐ integrated brain‐on‐a‐chip platforms: improving the predictive validity in neurodegenerative research. Bioeng. Transl. Med. 9, e10604 (2023). 100. Akbari, S. et al. Robust, long-term culture of endoderm-derived hepaticorganoids for diseasemodeling. Stem Cell Rep. 13, 627–641 (2019). 101. Lancaster, M. A. & Knoblich, J. A. Generation of cerebral organoids from human pluripotent stem cells. Nat. Protoc. 9, 2329–2340 (2014). 102. Ertürk, A. et al. Three-dimensional imaging of solvent-cleared organs using 3DISCO. Nat. Protoc. 7, 1983–1995 (2012). 103. Cai, R. et al. Panoptic imaging of transparent mice reveals wholebody neuronal projections and skull–meninges connections. Nat. Neurosci. 22, 317–327 (2019). 104. Cai, R. et al. Whole-mouse clearing and imaging at the cellular level with vDISCO. Nat. Protoc. 18, 1197–1242 (2023). 105. Wu, T. et al. clusterProfiler 4.0: a universal enrichment tool for interpreting omics data. Innovation 2, 100141 (2021). Acknowledgements Financial support provided by The Research Fund of Ege University, International Cooperation Project under grant number FUA-2020-22187 is highly appreciated. P.S.M. gratefully acknowledges the TUBITAK 2211-A National Graduate Scholarship Program and 2214-A International Doctoral Research Fellowship Program (special thanks to Dr. Onur Basak from UMC Utrecht Brain Center, Translational Neuroscience Department, Utrecht University, for fellowship supervision). S.Y. gratefully acknowledges the TUBITAK 2210-C National Graduate Scholarship Program. This study is partially conducted under the OrChESTRA (Organ-on-a-Chip Focused Strategic Partnership) project, which has received funding from the European Union’s Horizon Europe’s research and innovation program under Grant Agreement no. 101079473. Author contributions P.S.M. and S.Y. equally contributed to this work. P.S.M., S.Y., J.M., M.N., and Y.C.O. designed the experiments, analyzed the data and wrote the paper. O.Y.C., E.Y., C.B.A., A.G., and A.E. wrote, reviewed and edited the paper. O.Y.C. conceived the project, supervised all the experiments and edited the whole paper. All authors have read and agreed to the published version of the manuscript. Competing interests The authors declare no competing interests. Additional information Supplementary information The online version contains supplementary material available at https://doi.org/10.1038/s42003-024-07313-z. Correspondence and requests for materials should be addressed to Ozlem Yesil-Celiktas. Peer review information Communications Biology thanks Jagdeep Sandhu and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Primary HAandling Editor: Christina Karlsson Rosenthal. Reprints and permissions information is available at http://www.nature.com/reprints Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unlessindicated otherwise in a creditline to the material. If material isnot includedinthe article’sCreativeCommonslicenceandyourintended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/bync-nd/4.0/. © The Author(s) 2024 https://doi.org/10.1038/s42003-024-07313-z Article Communications Biology | (2024) 7:1627 19