Full Paper www.small-journal.com Microfluidic In Vitro Platform for (Nano)Safety and (Nano) Drug Efficiency Screening Yvonne Kohl,* Margit Biehl, Sarah Spring, Michelle Hesler, Vladimir Ogourtsov, Miomir Todorovic, Joshua Owen, Elisabeth Elje, Kristina Kopecka, Oscar Hernando Moriones, Neus G. Bastús, Peter Simon, Tibor Dubaj, Elise Rundén-Pran, Victor Puntes, Nicola William, Hagen von Briesen, Sylvia Wagner, Nikil Kapur, Espen Mariussen, Andrew Nelson, Alena Gabelova, Maria Dusinska, Thomas Velten, and Thorsten Knoll* 1. Introduction Microfluidic technology is a valuable tool for realizing more in vitro models capturing cellular and organ level responses for rapid and animal-free risk assessment of The need for new perspectives on in new chemicals and drugs. Microfluidic cell-based devices allow high-throughput vivo-like and animal-free approaches screening and flexible automation while lowering costs and reagent consumption due for chemical and pharmaceutical safety assessment is growing both in the area of to their miniaturization. There is a growing need for faster and animal-free approaches drug development and in risk assessment for drug development and safety assessment of chemicals (Registration, Evaluation, of chemicals Registration, Evaluation, Authorisation and Restriction of Chemical Substances, REACH). The work presented Authorisation and Restriction of Chemical describes a microfluidic platform for in vivo-like in vitro cell cultivation. It is equipped Substances (REACH).[1] with a wafer-based silicon chip including integrated electrodes and a microcavity. A The main points of criticism regarding animal testing are related to not only ethical proof-of-concept using different relevant cell models shows its suitability for label- concerns but also the limited transferability free assessment of cytotoxic effects. A miniaturized microscope within each module to human physiology, due to significant dif- monitors cell morphology and proliferation. Electrodes integrated in the microfluidic ferences in anatomy, organ structure, organ channels allow the noninvasive monitoring of barrier integrity followed by a label-free function, metabolism, and DNA-repair. assessment of cytotoxic effects. Each microfluidic cell cultivation module can be As an ethical alternative to animal experi- operated individually or be interconnected in a flexible way. The interconnection of ments, cellular organ-on-a-chip models represent a promising approach to better the different modules aims at simulation of the whole-body exposure and response imitate the complex organization of organs and can contribute to the replacement of animal testing in risk assessment studies in and tissues, as they allow for 3D cellular compliance with the 3Rs to replace, reduce, and refine animal experiments. organization with microfluidics simulating circulation in the body.[2–16] Dr. Y. Kohl, M. Biehl, S. Spring, M. Hesler, Prof. H. von Briesen, E. Elje, Dr. E. Rundén-Pran, Prof. E. Mariussen, Prof. M. Dusinska Dr. S. Wagner, Dr. T. Velten, T. Knoll NILU-Norwegian Institute for Air Research Fraunhofer Institute for Biomedical Engineering IBMT Department for Environmental Chemistry Fraunhofer-Gesellschaft zur Förderung der angewandten Forschung e.V. Health Effects Laboratory Joseph-von-Fraunhofer-Weg 1, Sulzbach 66280, Germany Instituttveien 18, Kjeller 2007, Norway E-mail: yvonne.kohl@ibmt.fraunhofer.de; thorsten.knoll@ibmt.fraunhofer.de E. Elje Dr. V. Ogourtsov, M. Todorovic Faculty of Medicine Tyndall National Institute Institute of Basic Medical Sciences University College Cork Department of Molecular Medicine Dyke Parade, Cork T12 R5CP, Ireland University of Oslo Dr. J. Owen, Prof. N. Kapur Sognsvannsveien 9, Oslo 0372, Norway Institute of Thermofluids K. Kopecka, Dr. A. Gabelova School of Mechanical Engineering Department of Nanobiology University of Leeds Cancer Research Institute Leeds LS2 9JT, UK Biomedical Research Center of the Slovak Academy of Sciences The ORCID identification number(s) for the author(s) of this article Dubravska cesta 9, Bratislava 84505, Slovakia can be found under https://doi.org/10.1002/smll.202006012. O. H. Moriones, Dr. N. G. Bastús, Prof. V. Puntes © 2021 The Authors. Small published by Wiley-VCH GmbH. This is an Institut Català de Nanociència i Nanotecnologia (ICN2) open access article under the terms of the Creative Commons Attribu- CSIC and BIST tion-NonCommercial-NoDerivs License, which permits use and distribu- Campus UAB, Bellaterra 08193, Barcelona, Spain tion in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. DOI: 10.1002/smll.202006012 Small 2021, 17, 2006012 2006012 (1 of 15) © 2021 The Authors. Small published by Wiley-VCH GmbH www.advancedsciencenews.com www.small-journal.com Standardized cell-based assays do neither mimic the struc- lab equipment and do not allow cultivation and simultaneous tural arrangement of different, sometimes highly specialized, optical analysis under dynamic flow conditions. The rapid pro- cell arrangements within an organ or capture defined cell–cell gress in the development of new micro- and nanotechnologies interactions, which are important for physiological cell func- together with more advanced cell culture models led to the devel- tion. For example, mechanical stimuli in respiration[17] and opment of a modular microfluidic platform for (nano)safety and intestinal peristaltic[6] have a decisive influence on the forma- (nano)drug efficiency screening.[76] This platform innovatively tion and regulation of a microphysiological environment. combines a microfluidic cartridge for cell cultivation with a minia- Organ-on-chip models bring the advantage of reproduc- turized incubator microscope for optical analysis and electrodes tion of these microphysiological conditions in vitro and thus for impedance measurements to determine cytotoxicity. differ clearly from standardized cell-based assays, e.g., for safety testing described in current Organisation for Economic Co-operation and Development (OECD) guidelines.[18–23] 2. Results and Discussion A challenge in the cultivation of complex organ models under physiologically relevant conditions is the maintenance of 2.1. Setup of the Microfluidic Platform their function over a long period of time (days, weeks, months). Maintenance of organ models is enabled by microfluidic tech- Before fabricating the microfluidic platform (a microfluidic car- niques to allow spatial control over fluids in µm-sized channels. tridge combined with a miniaturized incubator microscope) a Microfluidics can be exploited to enhance the physiological computational fluid dynamics (CFD) simulation was performed relevance of 3D cell models. During the last decade, lab-on-a- within Comsol Multiphysics to determine the requirements for the chip devices have been developed and used in various fields of setup. A 3D advection-diffusion model was created and meshed life sciences.[3,5,9,14,24–36] Various in vitro models have already using 2.3 million tetrahedral elements, confirmed to give a mesh- been combined with microfluidic systems (e.g., skin-on-chip, independent solution. The inlet and outlet channels and the cavity lung-on-chip, liver-on-chip, gut-on-chip, kidney-on-chip, retina- region, in which the cell models are cultured, was modeled, as on-chip, and blood-brain-barrier-on-chip).[2,4,5,7,8,11,12,15,16,37–53] shown in the cross-section of the model in Figure 1. Fluid flow A much better mimicry of the human in vivo-like physiology can with a density of 998 kg m−3, dynamic viscosity of 8.9 × 10−4 Pa be achieved by merging different organs on chip.[5,10,26,34,54–59] s, and volumetric inflow rate of 100 µL h−1 was predicted before Microfabrication technologies offer new opportunities in simulating the transport of chlorpromazine, a soluble pharma- the construction of miniaturized cell culture systems and ceutical compound (diffusion coefficient of 3.25 × 10−10 m2 s−1).[77] their applications.[13,60–64] To improve the analytical capability The no-slip and zero flux boundary conditions were applied at of human organ-on-chip systems, various optical[65] or electro- the geometry walls. The normalized concentration, c/co (where chemical biosensors[4,35,66] have been integrated in the devices co is the inlet concentration and c is the local concentration) is to detect cytotoxic agents. Among the sensing technologies, shown on the surfaces of the cavity (Figure 1), demonstrating impedance spectroscopy and transepithelial electrical resistance that the species is transported successfully to the cells within (TEER) measurements became the label-free method of choice 700 s of entering the chip. No undesirable flow effects, such as for noninvasive determination of cell barrier integrity.[67–71] fluid recirculation, were observed within the cavity geometry The need for live-cell imaging during cell cultivation has led that could have disturbed the transport of species to the cells. to the development of systems allowing real-time visualization To confirm that hydrodynamic forces did not have a det- and characterization of cell behavior in parallel. These solu- rimental effect on the adherence of the cells to the surface tions include customized incubators with inverted fluorescence of the microfluidic cavity or exceeded typical values seen in microscopes, compact microscopes that can be placed inside a vitro which can indicate adverse impact on cell physiological laboratory incubator or completely automated cell culture obser- behavior, the shear stress was calculated on the cavity sur- vation systems.[72–75] All these solutions require costly and bulky face using the CFD fluid flow model as shown in Figure 2. A peak shear stress of 0.0012 dyn cm−2 (for the study flow rate of 100 µL h−1) was observed at the mid-point of the cellular sur- O. H. Moriones face, with the variation either side of this caused by the change Universitat Autònoma de Barcelona (UAB) in fluid velocity due to the transition from the entry and exit Campus UAB, Bellaterra, 08193, Barcelona, Spain channels to the cavity. CFD results and analytical calculations Prof. P. Simon, Dr. T. Dubaj for the shear stress (Equation (6)) showed excellent agreement Institute of Physical Chemistry and Chemical Physics (Figure 2d). The shear stress on the cells is shown to be below Faculty of Chemical and Food Technology SUT Radlinskeho 9, Bratislava 812 37, Slovakia the critical values to cause adverse impact on cell behavior as Prof. V. Puntes cell viability with critical values taken as 5 dyn cm−2 (A549), Vall d’Hebron Institut de Recerca (VHIR) 0.33 dyn cm−2 (HepG2), and 2 dyn cm−2 (TH-1)[78–80] and is in Barcelona 08193, Spain line with other studies within the field.[81] Prof. V. Puntes Institució Catalana de Recerca i Estudis Avançats (ICREA) Barcelona 08193, Spain 2.1.1. Microfluidic Cartridge Dr. N. William, Prof. A. Nelson School of Chemistry University of Leeds The core part of the microfluidic cartridge is a custom-designed Leeds LS2 9JT, UK microfabricated silicon chip. It comprises a microcavity with Small 2021, 17, 2006012 2006012 (2 of 15) © 2021 The Authors. Small published by Wiley-VCH GmbH www.advancedsciencenews.com www.small-journal.com Figure 1. CFD simulation of fluid flow and species transport through a microfluidic channel. a) Cross-section of the geometry and the normalized concen- tration of species at the cavity walls after b) 400 s, c) 500 s, d) 600 s, e) 700 s, f) 800 s, and g) 900 s (bottom edge of (b)–(g) is left hand edge of cavity in (a)). a volume of ≈2.5 µL (Figure 3a), in which the cells are culti- on the silicon nitride membrane. The principle of the bar- vated on a thin transparent membrane made of silicon nitride rier model cultivation on the microhole membrane is similar that comprises a regular array of microholes (Figure 3b). With to that of commercially available larger coculture systems, as, these microholes, the microfluidic cartridge can be used as a e.g., the Transwell system. In contrast to cell culture inserts minaturized device for barrier models with dense cell layers used in static cultures, the integrated microfluidic channels Figure 2. CFD prediction of shear stress in the microfluidic channel cavity at flow rates of 50–500 µL h−1. a) Location along the bottom surface of the cavity where shear stress was determined. b) Flow velocity (mm s−1) through the center of the microfluidic cell at a flow rate of 100 µL h−1. c) Predicted shear stress in the cavity, where distance on the x-axis refers to the position along the cavity wall in the direction of flow. d) Comparison of the mean shear stress predicted using CFD with the average shear stress determined by an analytical equation for calculating shear stress in microfluidic channels. Small 2021, 17, 2006012 2006012 (3 of 15) © 2021 The Authors. Small published by Wiley-VCH GmbH www.advancedsciencenews.com www.small-journal.com Figure 3. Setup of microfluidic cartridge and miniaturized incubator microscope platform. a) SEM image of microcavity with microhole array mem- brane (1.6 mm x 1.6 mm). b) SEM image of microhole array with hole distance 10 µm. c) Microfluidic cartridge with connected tubing and cables for impedance measurement. This cartridge is integrated in the miniaturized incubator microscope. d) Incubator microscope with inserted microfluidic cartridge and lid, for bright field imaging and cell cultivation under controlled temperature. e) Schematic illustration of the microfluidic cartridge. The electrodes (yellow) are positioned in the two fluidic channels. The electric current flows between the electrodes through the pores in the membrane (dashed line). f) Schematics of experimental setup with two platform modules for parallel or serial operation. Small 2021, 17, 2006012 2006012 (4 of 15) © 2021 The Authors. Small published by Wiley-VCH GmbH www.advancedsciencenews.com www.small-journal.com allow for the cultivation of the cells under dynamic flow con- (SiO2-FITC-NP), titanium dioxide nanoparticles (TiO2-NP), and ditions. Figure 3c shows the assembled microfluidic cartridge gold NPs functionalized with polyethylene glycol (PEG)-SH with the integrated silicon chip. The gold electrodes in the MW5000 (Au-PEG-NP) were prepared following well-known microchannels are connected with the impedance measure- standard procedures. A summary of the physical and chemical ment setup that detects the impedance spectrum between the characteristics of the NPs used in this study is presented in two channels and through the membrane with the cell layer Figure S2 and Table S1 in the Supporting Information. (Figure 3c). An exploded view drawing of the individual compo- nents of the microfluidic cartridge can be found in Figure S1 in the Supporting Information. 2.3. In Vitro Studies on the Microfluidic Platform 2.3.1. Cell Characteristics on the Microfluidic Platform 2.1.2. Miniaturized Incubator Microscope The suitability of a silicon nitride membranes as a surface The developed miniaturized incubator microscope allows for for cell cultivation has been demonstrated previously.[82] To continuous real-time monitoring of cell characteristics, such verify the suitability of the miniaturized incubator micro- as morphology, cell growth, and cell response in the micro- scope as a microwell for cultivation, three human cell lines fluidic cartridge. The inserted microfluidic cartridge is con- A549 (lung), HepG2 (liver), and TH-1 (kidney) were used as nected with an adapter for temperature regulation (Figure 3d). example models for studying adhesion and cultivation, cell Components for temperature regulation (Peltier elements, morphology (Figure 4a) as well as the percentage of viable Negative temperature coefficient thermistors) and the light and dead cells (Table 1). During cell cultivation, the platform source for bright-field illumination are integrated in the is maintained at a constant temperature of 37 ± 0.5 °C. All fixation lid on the microfluidic cartridge (Figure 3d). The studied cell types successfully adhere on the silicon nitride horizontal assembly protruding from the side of the micro- membrane and proliferate. Cells cultivated on the plat- scope tube comprises illumination components for fluores- form show no differences in morphology, migration, adher- cence imaging. The compact microscope has a field of view ence or proliferation rate compared with cultivation under of 1.4 mm x 1.05 mm and a magnification of 4.4. With the static conditions. A549 and TH-1 cells are characterized by restricted numerical aperture (N.A. = 0.208) of the small objec- a spindle-shaped morphology and HepG2 by a round shape tive lens a point-shaped object fluorescing at a wavelength of morphology. The results of cell viability assessment using 610 nm will theoretically be depicted with a full width at half FDA/PI (fluorescein diacetate/propidium iodide) staining are maximum of 1.5 µm, which is in reasonable accordance with comparable under microfluidic and static conditions, for all the camera pixel size of 1.25 µm. three tested cell types (Table 1). After adhering to the surface of the silicon nitride mem- brane, the cells form a dense cell layer and, depending on 2.1.3. System Integration of the Microfluidic Platform cell type, specific cell junctions (tight junctions) are formed. This cellular changes can be quantified noninvasively via Each module of the microfluidic platform comprises a fluidic impedance measurement using the gold electrodes posi- system with valves, tubing, and pumps that is integrated with tioned in the upper and lower fluidic channel. From the time the miniaturized incubator microscope and the microfluidic of seeding, TH-1 cells proliferate in the cartridge as seen by cartridge. Cell suspension, reagents, and cell culture medium an impedance change of 1061 Ω after 12 h and 1275 Ω after are delivered to and led away from the silicon microcavity via 48 h (Figure 4b). In the case of the A549 cells, the impedance flow channels by means of a syringe pump. During cell cul- change is 1552 Ω after 12 h and 1625 Ω after 48 h. Although, tivation, fresh cell culture medium is constantly delivered HepG2 cells appear to proliferate more slowly in the car- to the cells in the microcavity. The electrodes for impedance tridge, as seen by an impedance change of 710 Ω after 48 h measurement are positioned in the two fluidic channels and (Figure 4c), they still continue to grow and proliferate. For the electric current flows through the pores in the membrane instance, after 5 d (120 h) HepG2 cells show an impedance (Figure 3e). Each platform module, equipped with an imaging change of 990 Ω (Figure 4c). module and fluidic components, can be operated indepen- dently or connected with other platform modules operating in parallel or in series. Control of the imaging modules and 2.3.2. DNA Stability of Cells under Microfluidic Conditions time-lapse imaging is performed by a LabView program. The electrodes of both microfluidic modules are connected with the HepG2 cells were cultured in the closed cartridge under flu- LabView-controlled impedance measurement system for par- idic conditions before being subjected to the enzyme-linked allel recording of the impedance data (Figure 3f). version of the comet assay for analyzing DNA stability (DNA strand breaks and oxidized base lesions) (Table 2). Previous live/dead staining (data not shown) confirms that their cell 2.2. Nanoparticle Synthesis and Characterization viability was high enough for the cells to be used for the comet assay (Figure 2). The basal DNA damage of HepG2 Silver nanoparticles (Ag-NP), fluorescently labeled silica cells in the microfluidic system is slightly higher than for the nanoparticles with fluorescein isothiocyanate (FITC) case where cells were cultured in standard static 2D systems Small 2021, 17, 2006012 2006012 (5 of 15) © 2021 The Authors. Small published by Wiley-VCH GmbH www.advancedsciencenews.com www.small-journal.com Figure 4. Growth behavior of the human cells lines on the miniaturized incubator microscope platform. a) Bright-field imaging of different cell types with different morphologies, cultured under static and fluidic conditions. Human epithelium of lung (A549), liver (HepG2), and kidney (TH-1) were cultured under standard static conditions (T25 flask) and in the developed cartridge under fluidic conditions (medium flow 100 µL h−1). After 24 h the statically cultured cells were imaged via standard microscope (IX81, Olympus) and the cells in the cartridge via miniaturized incubator microscope. No morphological differences were identified between cells cultivated under static compared to fluidic culture conditions. b,c) Impedance analysis of different epithelial cell lines during their culture in the microfluidic cartridge. The human cell lines b) TH-1 and A549 and c) HepG2 were cultured in the developed cartridge under fluidic conditions (medium flow 100 µL h−1). Impedance analysis at 10 kHz frequency was performed in line with the cell cultivation over a period of b) 48 h in case of TH-1 and A549 and c) 120 h in case of HepG2. For all cell types cell proliferation resulted in an increase of the impedance. (96-well plate).[83] Fluidic conditions induce a slight, nonsig- (Table 2). The positive control (H2O2 for DNA SBs and Fpg nificant increase in DNA strand breaks (SBs) but no increase for oxidized base lesions) induce DNA damage as expected in oxidized DNA base lesions (formamidopyrimidine Fpg) (Table 2). Table 1. Proportion of viable and dead cells after 24 h cultured under Table 2. Comet assay results of HepG2 cells cultured under static and static and fluidic conditions. Human cell lines (A549, HepG2, and TH-1) fluidic conditions. The cells were exposed to fresh medium under static were cultured under standard static conditions (T25 flask, incubator, (96-well plate) and fluidic conditions (medium flow 100 µL h−1, 37 °C) 37 °C, 95% humidity, 5% CO2) and in the developed cartridge under flu- for 24 h before measurement of DNA damage via comet assay. H2O2 idic conditions (medium flow 100 µL h−1, 37 °C). After 24 h in culture the (100 × 10−6 m) was used as a positive control for DNA strand breaks viability of the cells was analyzed via live/dead staining. (DNA SB). Results show mean ± standard deviation of three inde- pendent experiments. No net Fpg was determined after H2O2 exposure. Cell line Culture condition Amount of cells [%] Viable Dead DNA damage Culture condition % DNA in tail A549 Static 97.0 ± 1.0 3.0 ± 2.5 Cell medium H2O2 A549 Fluidic 95.7 ± 3.3 4.3 ± 3.3 DNA SBs Static 5.0 ± 2.3a),b) 95.5 ± 0.2c) HepG2 Static 92.0 ± 2.6 8.0 ± 1.6 DNA SBs Fluidic 10.3 ± 5.2 83.3 ± 9.0 HepG2 Fluidic 96.0 ± 4.9 4.0 ± 2.9 Net Fpg Static 3.7 ± 2.1a),b) – TH-1 Static 95.7 ± 1.2 4.3 ± 1.0 Net Fpg Fluidic 11.6 ± 10.7 – TH-1 Fluidic 94.0 ± 3.6 6.0 ± 3.6 a)Elje et al.[83]; b)n = 9; c)Elje et al.[84] Small 2021, 17, 2006012 2006012 (6 of 15) © 2021 The Authors. Small published by Wiley-VCH GmbH www.advancedsciencenews.com www.small-journal.com Figure 5. In vitro effects of reference chemicals and SiO2–FITC-NP under microfluidic conditions. a) Impedance analysis of A549 cells at 10 kHz fre- quency during exposure to cis-Pt (concentration 20 µg mL−1, blue line) in the microfluidic cartridge (flow rate 100 µL h−1). Nonexposed cells (black) and Triton X exposed cells (green) act as negative and positive controls, respectively. b) Impedance change after 48 h exposure of A549 cells with cis-Pt, Triton X, and cell culture medium (control). c) Impedance analysis at 10 kHz frequency of TH-1 cells during their exposure to SiO2–FITC-NP (100 µg mL−1) in the microfluidic cartridge (flow rate 100 µL h−1). Nonexposed cells (black) act as negative control (control). d) Impedance change after 48 h exposure of TH-1 cells with SiO2–FITC-NP and cell culture medium (control). 2.3.3. Validation with Reference Chemicals under Microfluidic 2.3.4. In Vitro Effect of SiO2-FITC-NP under Microfluidic Conditions Conditions Before the platform was used for in vitro studies with nano­ TH-1 cells (2 × 106 mL−1), cultured in the microfluidic plat- materials, the system was validated with reference chemicals. form, were exposed to silica–FITC core labeled nanoparti- For this purpose A549 cells (2 × 106 mL−1), cultured in the micro- cles (SiO2-FITC-NP, 100 µg mL−1). Impedance measurements fluidic platform, were exposed to cis-Platin (cis-Pt, 20 µg mL−1), during the cultivation of TH-1 cells in the cartridge show a cytostatic drug whose effect is based on the inhibition of DNA no significant decrease in impedance after the exposure to replication by crosslinking of two adjacent guanine bases of a SiO2-FITC-NP at 24 h (Figure 5c, red line), compared to non- DNA strand. Impedance measurements during the cultivation exposed cells (Figure 5c, black line). Cytotoxicity studies with of A549 cells in the cartridge show a significant decrease of the SiO2-FITC-NP under static conditions demonstrate no toxic impedance after 48 h exposure to cis-Pt (Figure 5a, blue line), in effects up to a concentration of 15.3 × 10−3 m (922 µg mL−1) over contrast to the nonexposed cells (Figure 5a, black line). Over a a duration of 24 h (data not shown). Under fluidic conditions, period of 2.5 h the impedance is decreased from 1404 to 550 Ω SiO2-FITC-NP induces the same effect as determined under (39.2%) (Figure 5a, blue line). Whereby the nonexposed con- static conditions (Figure 5d). trol remains stable between 1568 Ω (start of the measurement) and 1745 Ω (2.5 h) (Figure 5a, black line). Also in the case of Triton X-100 (positive control), the impedance of the cells 2.3.5. In Vitro Effect of SiO2-FITC-NP on the Microfluidic In Vitro decreases significantly compared to the control (Figure 5a, Platform green line). The value of Triton X is reduced over this period (2.5 h) by 100% from 1463 to 0 Ω. A performance test for hazard screening via the intercon- EC50 value (half-maximal effective concentration) for cis-Pt nected microfluidic in vitro platform (Figure 6a) was per- (24 h) in A549 cells is determined to be 25 µg mL−1 under static formed with SiO2–FITC-NP. HepG2 cells and TH-1 cells were conditions (data not shown). Under fluidic conditions cis-Pt cultured in individual microfluidic cartridges (module 1 and induces the same effect as determined under static conditions module 2), interconnected via microfluidics, simulating in (Figure 5b). vivo-like physiology (Figure 6b). SiO2–FITC-NP were injected Small 2021, 17, 2006012 2006012 (7 of 15) © 2021 The Authors. Small published by Wiley-VCH GmbH www.advancedsciencenews.com www.small-journal.com Figure 6. In vitro effects of SiO2–FITC-NP under microfluidic conditions in the interconnected platform. a) Two interconnected microfluidic platforms. Right module with lid removed to show the microfluidic cartridge. b) Schematic illustration of the fluidics interconnection. c) Impedance analysis of HepG2 cells (module 1, black line) and TH-1 cells (module 2, red line) at 10 kHz frequency during exposure to SiO2–FITC-NP (concentration 100 µg mL−1) on the two interconnected microfluidic platforms (flow rate 100 µL h−1). The diagram displays the impedance over time of exposure. in module 1 (HepG2 cells) and electrical impedance meas- 2.3.6. Transport of Au-NP over Bronchial Epithelial Cell Barrier urements were performed in both modules (Figure 6c). After under Microfluidic Conditions the addition of SiO2–FITC-NP in module 1, the impedance of the HepG2 cells increase constantly due to cell proliferation, The microholes in the silicon nitride membrane enable which does not indicate any cytotoxic effect on the HepG2 the transport of NPs through the membrane to the bottom cells. 12 h after the addition of SiO2–FITC-NP, the imped- channel. This brings the following advantages: a) transfer of the ance in module 2 (TH-1 cells) decreased, indicating the induc- NPs to a second sensing module by interconnecting; b) realiza- tion of a cytotoxic effect in the TH-1 cells (Figure 6c). The tion of transport studies with NPs, but also drugs and chemi- results indicate that the SiO2–FITC-NP were transported from cals; and c) sampling of transported NPs for physicochemical module 1 to module 2 and the effect was cell type-dependent. characterization. The impedance data give information on the It is also possible that the liver cells that interact first with the coverage of the microholes by the cells and the condition of the SiO2–FITC-NPs release a messenger substance, for example cells in case of NP exposure. cytokines, which are transported to module 2, triggering The transport of Au-PEG-NP over a cell barrier was studied a cytotoxic effect in the TH-1 cells. After demonstrating the in the human bronchial epithelial cell line, 16HBE, which was detection of dose-dependent, organ-specific effects, with refer- cultured under both static and fluidic conditions. It is well ence chemicals and NPs, the qualification of the system for known that 16HBE cells form tight junctions and produce a transport studies was studied. dense, cellular barrier in cell culture inserts. TEER serves as an Small 2021, 17, 2006012 2006012 (8 of 15) © 2021 The Authors. Small published by Wiley-VCH GmbH www.advancedsciencenews.com www.small-journal.com Table 3. Au elemental distribution of PEG-Au NPs after transport are commonly performed under static conditions in multiwell studies under static conditions. Via ICP-MS Au elemental distribution plates, were successfully modified and adapted to the dynamic in apical cell culture medium, basolateral cell culture medium, and cell lysates were determined after the exposure of Au-PEG-NPs (5 µg mL−1) conditions of the microfluidic platform. The use of the micro- to human bronchial epithelial cells (16HBE) under static conditions in a fluidic platform is independent of the chosen cell line. Direct Transwell setup. For each elemental part a blank (absence of cells, pres- control of the cells cultivated in the microfluidic cartridge ence of Au-PEG-NP) and a control (presence of cells, absence of Au- was obtained by real-time imaging and real-time impedance PEG-NP) were measured. The amounts of Au-PEG-NPs are expressed measurements. as total µg Au. Most other approaches either do not provide real-time imaging of cells or require the use of a lab incubator to cul- Elemental Au [µg] Exposure time tivate the cells in the microfluidic devices.[86–88] The micro- 4h 24 h 48 h fluidic platform provides both the necessary environment for Static apical cell cultivation and the components for real-time electrical and optical detection of the cell’s growth status. Thus, the platform Control <0.125a) <0.125a) <0.125a) does not require any additional microscope or lab incubator. It Blank 1.9 ± 0.0b) 1.6 ± 0.0 b) 1.65 ± 0.05b) meets the need for flexible technologies in the field of organ- Au-PEG-NP 2.57 ± 0.21 1.93 ± 0.12 1.63 ± 0.05 on-chip devices.[89,90] Static basolateral The innovative flexibility of the microfluidic platform was shown to enable real-time investigations at the single and Control <0.125a) <0.125a) <0.125a) multicellular level, to study different endpoints, as cytotoxicity, Blank 1.5 ± 0.05b) 1.5 ± 0.0b) 1.55 ± 0.0b) genotoxicity, and toxicological pathways. This multimodular Au-PEG-NP <1.2 <1.2 <1.2 approach allows different aspects of one or several processes to be interrogated and analyzed simultaneously. The advantage Static cell lysate of the developed platform is the application as a serial or par- Control <0.125a) <0.125a) <0.125a) allel platform and freedom to select the number of platforms to Blank 0.29 ± 0.07b) 0.17 ± 0.003b) 0.16 ± 0.005b) interconnect with each other. Real-time TEER measurements of the formed cell layers inside the microfluidic cartridge has been Au-PEG-NP 0.32 ± 0.01 0.29 ± 0.03 0.41 ± 0.03 successfully combined with real-time bright-field and fluores- a)Controls were below the detection limit; b)In the case of the blanks, levels of Au cent imaging of the cells in the individual modules. The com- around 0.165 µg were systematically measured, which are indicative enough of the bination of fluidic systems with biosensors, such as impedance presence of Au in the samples. spectroscopy, and microphysiometry as well as physiology iden- tification and quantification of the expression of specific cellular indicator for the formation of tight junctions and the density of markers is the next level of lab-on-chip devices.[69,9,91] The devel- this cellular barrier.[69,85] After 4 d of cultivation in a 24-well cell oped microfluidic platform follows this approach. In contrast to culture insert, the TEER of 16HBE reaches a constant value of other existing microphysiological systems,[86,92–94] the described 688 ± 22 Ω. This is consistent with data from other studies.[85] cartridge fabrication can easily be adapted for higher numbers. In the microfluidic system the impedance value was also con- Most components, such as the cartridge bottom and lid and stant before cell exposure was started. adhesive foils for microchannels are suitable for mass fabrica- After exposure of 4, 24, and 48 h the amount of Au was ana- tion and (semi)automated assembly. This aspect is important lyzed by induced coupled plasma-mass spectroscopy (ICP-MS) for increased acceptance and future use and marketing of those and atomic spectroscopy. The analysis of elemental gold after microphysiological systems beyond the research and develop- static exposure in cell culture insert shows that even after ment area. 48 h most of the particles remain still on the apical side of the 16HBE cellular barrier. The NP concentration is reduced over time. However, the amount of Au on the basolateral side 3. Conclusion was still below the detection limit after 48 h (Table 3). In the cell lysate, only a small increase in the amount of Au could The goal of this study was to develop a more efficient, mod- be detected (0.41 ± 0.03 µg after 48 h exposure of 5 µg mL−1 ular microfluidic in vitro platform for (nano)safety and (nano) Au-PEG-NP). Under fluidic conditions, 16HBE were exposed drug efficiency screening and to provide “how” and “why” with Au-PEG-NPs (5 µg mL−1) (flow rate of 100 µL h−1, answers for the fundamental processes of human response n = 2). As in the static setup, a time-dependent reduction to NPs. For this purpose, an innovative modular microfluidic of the Au amount in the apical medium could be detected platform was developed with the capability of reliably analyzing (3.76 ± 0.26 µg L−1 after 4 h, 3.06 ± 0.02 µg L−1 after 24 h, bio-nano-interactions in real-time. The developed microfluidic 2.58 ± 0.25 µg L−1 after 48 h). Thus, the uptake of Au-PEG-NP in vitro platform allows bright field and fluorescent imaging into the cells could be shown under static and fluidic condi- in parallel with impedance measurements. This enables the tions, which proves that the developed microfluidic cartridge is determination of interwell cell distribution, investigation of suitable for uptake and transport studies. changes in cell morphology and the detection of fluorescent The presented results show that the developed microfluidic dyes, reporter genes or cell markers. The developed microflu- platform can be used for several applications in the field of idic integration provides the infrastructure for the cultivation drug testing and (nano)toxicity testing. Standard assays, which of artificial tissues on-chip under physiological conditions. The Small 2021, 17, 2006012 2006012 (9 of 15) © 2021 The Authors. Small published by Wiley-VCH GmbH www.advancedsciencenews.com www.small-journal.com interconnection of the different modules aims at the simulation du of whole-body exposure and response. The platform is cell-type γ = (5) dy independent and can flexibly be interconnected as a serial or where u is the magnitude of flow velocity (m s−1). parallel system. Due to the array structure of the microholes on For 2D microfluidic channels, the wall shear stress can be estimated the microfabricated transparent membrane the system could using the following expression[81] also be used for studies on a single cell level.[95] After successful validation with a variety of cell lines, future 6 µQ τ= (6) steps in development of this infrastructure include combina- h 2w tion with other cell models, such as human induced pluripo- where Q is the flow rate (m3 s–1), h is the channel height (m) and w is tent stem cells or reporter cells, for example. The modular the channel width (m). platform allows the generation of a more reliable cell-based Microfluidic Cartridge: The microfluidic cartridge was designed as assay data for many types of applications, such as safety an adaptable modular device in microscope slide format. The cavity is analysis (endpoint and toxicokinetic studies) including food fabricated by wet chemical etching of silicon in potassium hydroxide. allergy studies, disease modeling, immune therapy develop- The membrane has a regular array of microholes with a hole diameter < 5 µm that is fabricated by reactive ion etching of the silicon nitride ment, functional food development, and drug efficacy testing layer. The cavity chip with the porous membrane is placed between in the pharma area. two microfluidic channels, each having a cross section of 0.6 mm2. In all fields of application this in vitro platform contributes to The microfabricated silicon-based cavity chip was integrated into a the replacement of animal testing in risk assessment studies in microfluidic cartridge, which was made up from two injection-molded compliance with the 3Rs to replace, reduce, and refine animal polystyrene plates with Luer adapters (Greiner Bio-One), a 0.5 mm experiments. By simulating the in vivo-like physiological con- thick ceramics plate (CoorsTek, USA) and two layers of medical grade pressure-sensitive adhesive (AR 90106, Adhesives Research, Ireland). The ditions the microfluidic in vitro platform meets the need for cartridge has a symmetrical structure with the cavity chip in the center. new perspectives on in vivo-like and animal-free approaches for Two microchannels are defined by the adhesive layer that connects the chemical and pharmaceutical safety assessment. polystyrene plates with the 0.5 mm thick aluminum oxide ceramics plate. This plate has an opening for the insertion of the silicon cavity chip and separates the two microchannels both electrically and fluidically. Thin film gold electrodes for impedance sensing were deposited by sputtering onto 4. Experimental Section two polystyrene parts that cover two flow channels. Miniaturized Incubator Microscope: The miniaturized microscope Fluid Dynamics Simulation: CFD simulations of fluid flow at a flow comprises two subsystems: the imaging module and the heating unit that rate of 100 μL h−1 and species transport through the microfluidic is used for temperature control of the microfluidic cartridge. Based on a channel, using the methodology applied to study flow through CMOS (complementary metal-oxide-semiconductor) camera, the imaging a similar microfluidic geometry by the authors,[96] confirmed no module was specifically designed for the microfluidic cartridge and the undesirable flow characteristics were observed in the microfluidic channel and cell cavity. The concentration of species within the cavity implementation into the microfluidic platform. Components with high reached that of the inflow concentration after approximately 700 s, a accuracy requirements for the optical path were fabricated mechanically by short time compared to the overall timescale of the exposure period computerized numeric control milling and anodized. Other components, (24 h) during an experiment. CFD simulations were completed using including the housing for the light-emitting diode (LED) light source, COMSOL Multiphysics v5.3a to predict fluid flow and chemical species electronics and the socket for the board-level camera, were manufactured transport within the microfluidic cell. The Navier–Stokes equations by 3D-printing. The imaging module comprises a lens (Lensagon for incompressible, steady-state, isoviscous fluid flow through the B10M7224, Lensation, Germany), a color CMOS camera (UI-3592LE-C, microfluidic channel with no gravity effects were solved Image Development Systems, Germany), a fluorescence 4-band-filter-set for 4’,6-diamidin-2-phenylindol, FITC, tetramethylrhodamine, and Cy5 (1) (Semrock, USA) as well as the LEDs. An LED in the amber range (588 nm, ∇ ·u= 0 LUXEON Z LXZ1-PL01, LUMILEDS, USA) was used as a light source for ρ (u · ∇ ) u = − ∇p + µ∇ 2u (2) transmission bright-field imaging. Fluorescence imaging was realized with Köhler incident light illumination using four LEDs (LUXEON UV U where u is the fluid velocity vector (m s−1), p is the pressure (Pa), μ is Line, LUMILEDS, USA) with excitation wavelengths in the UV (385 nm), the dynamic viscosity (Pa s), and ρ is the pressure (kg m−3). blue (470 nm), and lime range (567.5 nm), which were mounted on Time-dependent species transport through the microfluidic channel a metal core printed circuit board and controlled by multi-LED driver was solved using the advection-diffusion transport equation electronics. A LabView program was developed for the optical monitoring and analysis of cells in the microcavity. The software controls all LEDs ∂c = D∇ 2c − u · ∇c (3) and the CMOS camera, thus allowing for video recording and time-lapse ∂t imaging in bright-field and fluorescence mode. where c is the species concentration (mol m−3) and D is the diffusion All components for temperature control, together with bright-field coefficient of the dilute species (m2 s−1). illumination, were integrated into the lid of the microscope. Two Peltier The shear stress in the channel was calculated to determine if the elements, each equipped with a heat sink, are in thermal contact with fluid flow influenced the adherence of the cells to the surface of the the top part of the microfluidic cartridge, and an negative temperature microfluidic cavity. The wall shear stress was calculated along the cavity coefficient-temperature-sensor measures the temperature at the surface wall, shown in Figure 2, using the following equation of the microfluidic cartridge. A commercial Peltier-controller (TLK 33, Sika GmbH, Germany) evaluates the signals from the temperature sensor τ = µ · γ (4) and controls direction and intensity of the current through the Peltier elements. Temperature during cultivation was measured in a test cartridge where τ is the shear stress (Pa), expressed as dyn cm−2 with an integrated Pt-100 temperature sensor. During cell cultivation (1 Pa = 10 dyn cm−2), and γ is the fluid shear rate (s−1) equivalent to the temperature was maintained at a value of 37 ± 0.5 °C. Stability of the velocity gradient perpendicular to channel surface in the y direction pH (between 7.4 and 7.8) and sufficient oxygen concentration in the Small 2021, 17, 2006012 2006012 (10 of 15) © 2021 The Authors. Small published by Wiley-VCH GmbH www.advancedsciencenews.com www.small-journal.com cartridge were also measured in the test setup. Samples taken from the Synthesis of Silica Dioxide –FITC Core Labeled Nanoparticles (SiO2- microchannels at different timepoints were analyzed with a Portamess pH FITC NP): Core labeled silica nanoparticles with the fluorophore FITC meter and oxygen concentration was measured using oxygen nanoprobes were prepared by following and adapting the method reported by van and a fiber-optic oxygen sensor from PyroScience (Germany). Blaaderen et al.[98] First, the silane-dye precursor was prepared. For Fluidic Circuit: Commercially available fluidic and microfluidic components this purpose, in a 100 mL flask, the FITC was measured accurately and were selected to guarantee a maximum flexibility for the platform operation. dissolved in dry ethanol (50 mL). Then, the APTES was added under These commercial components comprise solenoid pinch valves (Bio- stirring and the flask was sealed using septum and left for 15 min under Chem Valves, USA), syringe pumps (New Era Pump Systems, USA), and an argon atmosphere in dark conditions. Finally, the reaction was left special microfluidic lids for Falcon tubes (Elvesys, France), which can be for 24 h under magnetic agitation. Once the precursor was prepared, in used with standard high-pressure liquid chromatography connectors and a flask sealed with the help of a septum and under argon atmosphere, polytetrafluoroethylene tubing. Elastic tubing (C-Flex, Cole-Parmer, USA) dry ethanol (45 mL) was added. Then, after 10 min of stirring, 2.02 mL was used in combination with the pinch valves. (50.7 mmol) of ammonium hydroxide 28% was added and left for 5 min Electrical Impedance Measurement: A commercial impedance analyzer with continuous stirring. Finally, TEOS (1.7 mL, 7.6 mmol) and APTES- (HP4248A) in combination with an electrical switch (Keithley 7001 switch FITC precursor (5 mL) were sequentially added in the ethanol anhydride system) was used for measuring the electrical impedance spectrum solution whilst maintaining a constant solvent volume. The reaction between both electrodes of the microfluidic cartridge in the range was left for 24 h stirring in dark conditions. Dialysis was used to purify between 100 Hz and 100 kHz. Electrical connection of the microfluidic unreacted dye and precursor. All the above experiments were conducted module with the impedance measuring system was carried out by at room temperature. commercial connectors that were soldered onto the contact pads of the Synthesis of Titanium Dioxide Nanoparticles (TiO2-NP): Titanium cartridge. A LabView program controls data acquisition and impedance anatase nanoparticles of mean size ≈ 4 nm in diameter were prepared by data were recorded in time intervals of 10 min. a precipitation method following and adapting the method reported by System Integration of the Microfluidic Platform: All subsystems Pottier et al.[99] The adapted method is described in Elje et al.[83] Samples described form one module of the microfluidic platform. The miniaturized were further purified by three centrifuging cycles and resuspended in incubator microscope with the microfluidic cartridge was integrated an aqueous solution of Tetramethylammonium hydroxide (TMAOH, with the fluidic circuit and connected to the impedance measurement 100 mol L−1). Afterwards they were characterized by transmission system. The flexible modular platform concept allows for the individual electron microscopy (TEM), dynamic light scattering, and UV–vis operation of one single module and for the interconnection of multiple spectroscopy. The former was used to determine the particle size and to modules (Figure 6). Control of the microfluidic module was performed characterize the crystallinity of particles. by a custom made electronic instrumentation. It provides a universal Synthesis of Polyethylene Glycol-Coated Gold Nanoparticles serial bus (USB) communication interface to support microfluidic valves (Au-PEG-NP): Gold nanoparticles of ≈10 nm in diameter were handling by corresponding graphical user interface (GUI) software. prepared by the citrated-stabilized seeded-growth method reported by Reagents for Nanoparticle Synthesis: Silver nitrate (AgNO3), Gold(III) Bastús et al.[100] A solution of 2.2 × 10−3 m sodium citrate in Milli-Q water chloride trihydrate (HAuCl4·3H2O, 99%), trisodium citrate (Na3C6H5O7), (150 mL) was heated with a heating mantle in a 250 mL three-necked and tannic acid (C76H52O46), poly(ethylene glycol) methyl ether thiol round-bottomed flask for 15 min under vigorous stirring. A condenser (mol wt 5000), tetraethyl orthosilicate (TEOS, ≥ 98%), 3-(aminopropyl) was utilized to prevent the evaporation of the solvent. After boiling had triethoxysilane (APTES, ≥ 98%), fluorescein 5(6)-isothiocyanate (FITC, commenced, 1 mL of HAuCl4 (25 × 10−3 m) was injected. The color of ≥ 90%), hydrochloric acid ACS reagent (HCl, 37%), sodium hydroxide the solution changed from yellow to bluish-grey and then to soft pink in BioXtra, ≥ 98% (acidimetric) pellets (anhydrous), tetramethyl- 10 min. The resulting particles (≈10 nm, ≈3 × 1012 NPs mL−1) are coated ammonium hydroxide (TMAOH) solution ACS reagent with a with negatively charged citrate ions and hence are well suspended in concentration 1.0 ± 0.02 m in water and cis-diammineplatinum(II) H2O. Immediately after the synthesis of the Au seeds and in the same dichloride, were purchased from Sigma-Aldrich. Ammonia solution reaction vessel, the reaction was cooled and NPs were conjugated (ACS Reagent, p.a. 28–30%) and dried ethanol (seccoSolv max 0.01% directly with a thiol-PEG MW5000 solution to a final concentration of H2O) were purchased from Merck and titanium(IV) isopropoxide 97%, 2 × 10−6 m and was left overnight with gently stirring, to assure the was purchased from Fluka Chemika. Triton X-100 was purchased from complete coating of the Au NPs. Finally, the obtained PEGylated Au-NP Applichem, FDA and PI from Invitrogen. All chemicals were used as were purified by centrifugation and dispersed in water, to remove the received without further purification. Distilled water passed through unbound PEG and the free citrate molecules. a Millipore system (ρ = 18.2 MΩ) was used in all experiments. All UV–vis Spectroscopy: UV–visible spectra were acquired with an Agilent glassware was first rinsed with acetone and then with Millipore water Cary 60 UV–vis spectrophotometer as described in Elje et al.[83] before use. Transmission Electron Microscopy: TEM diameter of the synthesized Synthesis of Silver Nanoparticles (Ag-NP): Silver nanoparticles of nanoparticles was obtained from the analysis of TEM images (Figure 3) ≈20 nm in diameter were prepared by the seeded-growth method recently acquired with an FEI Tecnai G2 F20 S-TWIN HR(S) TEM equipped with reported by Bastús et al.[97] In a typical experiment, 100 mL volume of an energy-dispersive X-ray spectroscopy detector, operated at an accelerated aqueous solution containing sodium citrate (SC, 5 × 10−3 m) and tannic voltage of 200 kV. Microliters of the samples were prepared by drop- acid (TA, 0.1 × 10−3 m) was prepared and heated up to 100 °C with a casting 10 µL of the sample on a carbon-coated copper TEM grid and heating mantle in a three-neck round-bottomed flask for 15 min under left to dry at room temperature. Scanning electron microscopy (SEM) vigorous stirring. A condenser was used to prevent the evaporation was conducted using an FEI Magellan 400L XHR SEM, in scanning mode of the solvent. After boiling had commenced, AgNO3 (1 mL, operated at 1 kV and in transmission mode operated at 20 kV. Scanning 25 × 10−3 m) of was injected into this solution. The solution became transmission electron microscopy was used for bigger sizes. Average size bright yellow immediately indication the formation of the silver seeds. and size distribution of the samples were measured using ImageJ software Immediately after the synthesis of Ag seeds and in the same vessel, by counting at least 300 particles from different regions of the grid. the synthesized silver seeds were grown by cooling down the solution Dynamic Light Scattering: Dynamic light scattering and laser Doppler to 90 °C, the seed solution was diluted by extracting 20 mL of sample anemometry were used to determine the hydrodynamic size and the and adding Milli-Q-water (17 mL) and then SC (500 µL, 25 × 10−3 m), surface charge of the nanoparticles as reported by Elje et al.[83] TA (1.5 mL, 2.5 × 10−3 m), and AgNO3 (1 mL, 25 × 10−3 m) were Induced Coupled Plasma-Mass Spectroscopy: ICP-MS was performed sequentially injected, again. This process was repeated up to two times, using an ICP-MS Perkin Elmer (model: NexION 300). For ICP-MS progressively growing the size of the Ag-NP until reaching the desired analysis, the nanoparticles were dissolved in concentrated aqua regia, size ≈ 20 nm (≈ 1.8 × 1012 NP mL−1). The obtained Ag-NP were purified which was then heated to ensure complete dissolution of all the metals by centrifugation and stored in a solution containing both TA and SC. and diluted to an optimal concentration for ICP-MS analysis. The Small 2021, 17, 2006012 2006012 (11 of 15) © 2021 The Authors. Small published by Wiley-VCH GmbH www.advancedsciencenews.com www.small-journal.com samples for ICP-MS measurements were prepared as follows: First, Triton X-100, pH 10.4) for at least 1 h. As a control for SBs, separate the colloidal solution containing the nanoparticles was centrifuged at slides were submerged in 100 × 10−6 m H2O2 in PBS for 5 min at 4 °C, 40.000 g for 45 min. The supernatant and the pellet were then analyzed washed twice for 2 min in PBS (4 °C) and then submerged in a separate separately. coplin jar of lysis solution. Cell Lines and Standard Cell Cultivation: All cell culture reagents After lysis, slides for Fpg treatment were washed twice for 8 min in were obtained from Invitrogen (Karlsruhe, Germany), unless stated buffer F (40 × 10−3 m 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid otherwise. The human lung epithelial carcinoma cells A549 (ATCC CCL- (HEPES), 0.1 m KCl, 0.5 × 10−3 m EDTA, 0.2 mg mL−1 bovine serum 185), the human lung bronchial epithelial cells 16HBE (SCC150, Merck, albumin (BSA), pH 8.4), added Fpg diluted in buffer F, and covered with Darmstadt, Germany) and the human hepatocarcinoma cells HepG2 a polyethylene foil and incubated at 37 °C for 30 min in a humid box. As (ATCC HB-8065) were cultured in Roswell Park Memorial Institute (RPMI) positive control for function of Fpg, reference samples were included; 1640 medium supplemented with l-glutamine (4 × 10−3 m), penicillin A549 cells exposed to the photosensitizer Ro 19–8022 (Hoffmann La (100 U mL−1), streptomycin (100 µg mL−1), and 10% (v/v) fetal calf serum Roche, Switzerland), which with light induces oxidized purines, mainly (FCS), as previously described in Kohl et al.[82] The human renal proximal 8-oxoG, which is detected by the Fpg.[104,105] A549 cells were exposed to tubule epithelial (TH-1) cells obtained from Kerafast were cultivated Ro 19-8022 (2 × 10−6 m) and irradiated with visible light (30 cm distance in Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with from cells, 250 W) on ice for 4 min, before freezing of aliquots and penicillin (100 U mL−1), streptomycin (100 µg mL−1), and 10% FCS as embedding into gels on day of experiment. described in Sramkova et al.[101] In cocultivation experiments TH-1 cells After Fpg treatment, slides with cells embedded in gels were placed were cultured in RPMI 1640 medium as all other cell lines. Cells were in the electrophoresis tank, submerged in electrophoresis solution maintained in a 5% CO2 humidified atmosphere at 37 °C. (0.3 m NaOH, 1 × 10−3 m EDTA, pH > 13.4) for 20 min to let the DNA Cell Cultivation and NP Exposure in a Single Microfluidic Platform unwind, before running electrophoresis for 20 min (25 V, 1.25 V cm−1, Module: The cells were trypsinized and diluted in cell culture medium to Consort EV202). The gels were neutralized in PBS, washed in ultrapure 2 × 106 cells mL−1. The cell solution was manually delivered into the cavity water and left to dry overnight. Comets were visualized after staining of the microincubator by a syringe. After 2 h of adherence, the tubing of with SYBR gold (1:2000, S11494, Sigma-Aldrich), and scored in Leica the syringe pump was connected to the microfluidic incubator module DMI 6000 B (Leica Microsystems), equipped with an SYBRphotographic and fresh cell culture medium was constantly pumped through the micro filter (Thermo Fischer Scientific) using the software Comet assay IV incubator at a flow rate of 100 µL h−1. After 24 h of cultivation the cells were 4.3.1 (Perceptive Instruments, Bury St Edmunds, UK). Median DNA tail treated with the test substances. 24 h exposures with cis-Pt, SiO2-FITC-NP intensity was calculated from 50 comets per gel as a measure of DNA and Ag-NP were performed with this experimental setup. SBs, with a total of three gels and 150 comets per chip sample and two Cell Cultivation and NP Exposure in Interconnected Microfluidic Platform gels and 100 comets per reference sample. Modules: Two single modules were combined using one tubing system. Viability Assay (alamarBlue) in 96-Well Plates: Cells were seeded at a The cells were prepared as described above. Just before cell exposure density of 2 × 104 per well into a 96-well plate and cultured for 24 h in the medium syringe of module 1 was replaced by a syringe with the test cell culture medium. Exponentially growing cells were then preincubated, substance (20 µg mL−1 cis-Pt or 100 µg mL−1 SiO2-FITC-NP) and the in the presence of test compounds or untreated (control) cells, for 24 h. microfluidics of the two models guaranteed a constant flow of the test After treatment, cells were incubated with 100 µL of working solution of substance in both modules. 24 h exposures with cis-Pt and SiO2-FITC-NP alamarBlue (Invitrogen, USA) for 4 h according to the manufacturer’s were performed with this experimental setup. protocol. The fluorescence (excitation 530 nm, emission 590 nm) in each Electrochemical Cell Characterization (Impedance Measurement): well was measured on a microplate reader (POLARStar OPTIMA, BMG Impedance spectra were recorded at different frequencies between LABTECH). EC50 values were calculated based on the alamarBlue results. 100 Hz and 10 kHz in order to characterize barrier integrity during Transport Studies under Static Conditions: 16HBE cells (8.0 × 104 per treatment with test substances. Impedance values were measured every well) were seeded into a 12-well cell culture insert (order# 3460, 10 min. Based on previous findings with microhole array membranes, Corning) and cultured for 4 d in cell culture medium. Via CellZScope 10 kHz was selected as an appropriate measurement frequency the impedance was measured continuously. At an impedance value Kurz et al.[102] > 600 Ω/cm2 the cells were exposed to Au-PEG-NPs (5.0 µg mL−1). Live/Dead Staining: The cells were stained with FDA and PI to allow After 4, 24 and 48 h the NP-containing cell culture media on the apical for visualization of live and dead cells, respectively. The cells were and basolateral side was removed and used for ICP MS analysis to incubated in the cartridge with FDA (30 µg mL−1) and PI (40 µg mL−1) for determine the amount of Au in the sample. Before analysis the NPs up to 1 min, before washing with phosphate buffered saline (PBS) and were dissolved in aqua regia, which was heated to ensure complete imaging via the miniaturized microscope. The recorded fluorescence dissolution of all metals. 200 µL of the cell culture media was diluted images were visually analyzed using GIMP. to a final volume of 10 mL. Cells were harvested, digested, and diluted Comet Assay after Experiments in Cartridge: After live/dead staining at to a volume of 5 mL. ICP-MS was performed using an ICP-MS Perkin the end of the culture period, cells from the microfluidic cartridge were Elmer (Model NexION 300). detached by trypsinization (0.25%) for 10 min. The cells were resuspended Transport Studies under Fluidic Conditions: 16HBE cells (4 × 106 mL−1) in culture medium and centrifuged (5 min, 200 g, 21 °C), before the cell were injected in the microfluidic cartridge and cultured for 24 h in cell number was counted in Countess (Invitrogen) with trypan-blue staining, culture medium (flow rate 100 µL h−1). The impedance was measured and the DNA damage was measured by the comet assay. continuously. After the impedance reached a constant value the cells The miniaturized 12-gel modified comet assay was performed as were exposed to Au-PEG-NPs (5.0 µg mL−1, in cell culture medium). described by El Yamani et al.,[103] with the bacterial repair enzyme During the 4 h, 24 h, and 48 h exposure the NP-containing cell culture formamidopyrimidine DNA glycosylase (Fpg, gift from Professor media on the apical side was collected in a tube and prepared for atomic Andrew Collins, University of Oslo, Norway), which converts oxidized spectroscopy to determine the amount of Au in the sample. Before (majority of it is 8-oxoG) or alkylated purine bases to SBs[84,103,104] for analysis the NPs were dissolved in pure water which was heated to detection of oxidized or alkylated bases. Cell cultures were disaggregated ensure complete dissolution of all metals. Throughout the experiments as explained above. As reported by Elje et al.[100] the cell suspension water from NANOpure system (Wilhelm Werner GmbH, Germany) was was mixed in ratio 1:3 with low melting point-agarose (0.8% w/v, used. The atomic absorption spectrometry (AAS) measurements were A9414, Sigma-Aldrich, 37 °C) giving a final agarose concentration of performed at 242.8 nm using a high-resolution continuum source atomic 0.6% w/v, and mini-gels (10 µL) were made on microscope slides absorption spectrometer AA700 (Analytik Jena AG, Jena, Germany). As precoated with 0.5% standard melting point agarose (05066, Sigma- a stock solution for AAS calibration, standard 1.000 g L−1 Au solution Aldrich). The samples were submerged in lysis solution (2.5 m NaCl, was used. The limit of detection was determined to be lower than 0.1 m ethylenediaminetetraacetic acid (EDTA), 10 × 10−3 m Tris, 10% v/v 0.125 µg L−1. The amount of Au in the sample was calculated as µg L−1. Small 2021, 17, 2006012 2006012 (12 of 15) © 2021 The Authors. Small published by Wiley-VCH GmbH www.advancedsciencenews.com www.small-journal.com Supporting Information [7] J. D. Stucki, N. Hobi, A. Galimov, A. O. Stucki, N. Schneider-Daum, C.-M. Lehr, H. Huwer, M. Frick, M. Funke-Chambour, T. Geiser, Supporting Information is available from the Wiley Online Library or O. T. Guenat, Sci. Rep. 2018, 8, 14359. from the author. [8] L. Elberskirch, T. Knoll, A. Moosmann, N. Wilhelm, H. von Briesen, S. Wagner, J. Drug Delivery Sci. Technol. 2019, 54, 101248. [9] I. Maschmeyer, A. K. Lorenz, K. Schimek, T. Hasenberg, A. P. Ramme, J. Hübner, M. Lindner, C. Drewell, S. Bauer, Acknowledgements A. Thomas, N. S. Sambo, F. Sonntag, R. Lauster, U. Marx, Lab T.K. and Y.K. contributed equally to this work. The authors thank Chip 2015, 15, 2688. Frank Bauerfeld for system assembly and electrical testing, Werner [10] E.-M. Materne, A. P. Ramme, A. P. Terrasso, M. Serra, P. M. Alves, Haberer for assembly of the microfluidic cartridges, Axel Brenner for C. Brito, D. A. Sakharov, A. G. Tonevitsky, R. Lauster, U. Marx, the fabrication of the silicon micro cavity chips, Hoffman La Roche for J. Biotechnol. 2015, 205, 36. providing the Ro 19-8022, NorGenoTech for providing the Fpg enzyme, [11] J. Lee, S. Kim, Curr. Drug Metab. 2018, 19, 577. and Karen Steenson for proofreading. This research was funded by the [12] D. T. Phan, R. H. F. Bender, J. W. Andrejecsk, A. Sobrino, S. J. Hachey, European Commission under the Horizon2020 programme (HISENTS, S. C. George, C. C. Hughes, Exp. Biol. Med. 2017, 242, 1669. Grant Agreement No. 685817 and VISION, Grant Agreement No. [13] G. Mazzoleni, D. Di Lorenzo, N. Steimberg, Genes Nutr. 2009, 857381); and by the Norwegian Research Council Norway (RCN) via 4, 13. the European Research Area Network (ERA-NET) EuroNanoMed II [14] R. Mittal, F. W. Woo, C. S. Castro, M. A. Cohen, J. Karanxha, project INNOCENT (RCN 271075) and the ERA-NET EuroNanoMed J. Mittal, T. Chhibber, V. M. Jhaveri, J. Cell. Physiol. 2019, 234, III project Graphene-encapsulated magnetic nanoparticles (RCN 8352. 246672/O70). K.K. received Short Term Scientific Mission Grant [15] M. W. van der Helm, A. D. van der Meer, J. C. T. Eijkel, A. van den (ID 42926) under COST Action CA 17140 “Cancer Nanomedicine from Berg, L. I. Segerink, Tissue Barriers 2016, 4, e1142493. the Bench to the Bedside” supported by COST (European Cooperation in Science and Technology). E.E. received funding from the Norwegian [16] A. Ehrlich, D. Duche, G. Ouedraogo, Y. Nahmias, Annu. Rev. Research Council (272412/F40). O.H.M. received funding from the Biomed. Eng. 2019, 21, 219. Spanish Ministry of Economy and Competitiveness (SEV-2013-0295-17-3). [17] J. Shrestha, S. Razavi Bazaz, H. Aboulkheyr Es, D. Yaghobian Azari, Open access funding enabled and organized by Projekt DEAL. B. Thierry, M. Ebrahimi Warkiani, M. Ghadiri, Crit. Rev. Biotechnol. 2020, 40, 213. [18] Test No. 439: In Vitro Skin Irritation: Reconstructed Human Epi- dermis Test Method, OECD 2020. Conflict of Interest [19] Test No. 492: Reconstructed Human Cornea-like Epithelium (RhCE) Test Method for Identifying Chemicals Not Requiring Classification and The authors declare no conflict of interest. Labelling for Eye Irritation or Serious Eye Damage, OECD 2019. [20] Test No. 431: In Vitro Skin Corrosion: Reconstructed Human Epidermis (RHE) Test Method, OECD 2013. Keywords [21] Test No. 432: In Vitro 3T3 NRU Phototoxicity Test, OECD 2019. [22] Test No. 490: In Vitro Mammalian Cell Gene Mutation Tests Using drug efficiency, in vitro culture-on-chip, microfluidic platform, the Thymidine Kinase Gene, OECD 2015. miniaturized incubator microscope, (nano)safety [23] Test No. 476: In Vitro Mammalian Cell Gene Mutation Test, OECD Publishing 1997. Received: September 25, 2020 [24] C. Giese, U. Marx, Adv. Drug Delivery Rev. 2014, 69–70, 103. Revised: December 18, 2020 [25] J. Hogan, Nature 2006, 442, 351. Published online: January 18, 2021 [26] C. Luni, E. Serena, N. Elvassore, Curr. Opin. Biotechnol. 2014, 25, 45. [27] L. Richter, V. Charwat, C. Jungreuthmayer, F. Bellutti, H. Brueckl, P. Ertl, Lab Chip 2011, 11, 2551. [1] European Comission, Regulation (EC) No 1907/2006 – Registration, [28] P. Yager, T. Edwards, E. Fu, K. Helton, M. R. KristenNelson Tam, Evaluation, Authorisation and Restriction of Chemicals (REACH) – B. H. Weigl, Nature 2006, 442, 412. EU-OSHA, https://osha.europa.eu/de/legislation/directives/regula- [29] R. Daw, J. Finkelstein, Nature 2006, 442, 367. tion-ec-no-1907-2006-of-the-european-parliament-and-of-the-council [30] K. Yum, S. G. Hong, K. E. Healy, L. P. Lee, Biotechnol. J. 2014, (accessed: November 2020) 9, 16. [2] N. Ashammakhi, K. Wesseling-Perry, A. Hasan, E. Elkhammas, [31] B. Zhang, M. Radisic, Lab Chip 2017, 17, 2395. Y. S. Zhang, Kidney Int. 2018, 94, 1073. [32] H. Kimura, T. Ikeda, H. Nakayama, Y. Sakai, T. Fujii, J. Lab. Autom. [3] N. Alépée, A. Bahinski, M. Daneshian, B. De Wever, E. Fritsche, 2015, 20, 265. A. Goldberg, J. Hansmann, T. Hartung, J. Haycock, H. Hogberg, [33] S. N. Bhatia, D. E. Ingber, Nat. Biotechnol. 2014, 32, 760. L. Hoelting, J. M. Kelm, S. Kadereit, E. McVey, R. Landsiedel, [34] E.-M. Materne, I. Maschmeyer, A. K. Lorenz, R. Horland, M. Leist, M. Lübberstedt, F. Noor, C. Pellevoisin, D. Petersohn, K. M. S. Schimek, M. Busek, F. Sonntag, R. Lauster, U. Marx, U. Pfannenbecker, K. Reisinger, T. Ramirez, B. Rothen-Rutishauser, J. Visualized Exp. 2015, 98, e52526. M. Schäfer-Korting, K. Zeilinger, M.-G. Zurich, ALTEX 2014, 31, 441. [35] M. Rothbauer, H. Zirath, P. Ertl, Lab Chip 2018, 18, 249. [4] Q. Ramadan, M. A. M. Gijs, Lab Chip 2015, 15, 614. [36] M. Rothbauer, J. M. Rosser, H. Zirath, P. Ertl, Curr. Opin. Biotechnol. [5] D. Bovard, A. Sandoz, K. Luettich, S. Frentzel, A. Iskandar, 2019, 55, 81. D. Marescotti, K. Trivedi, E. Guedj, Q. Dutertre, M. C. Peitsch, [37] B. Bachmann, S. Spitz, M. Rothbauer, C. Jordan, M. Purtscher, J. Hoeng, Lab Chip 2018, 18, 3814. H. Zirath, P. Schuller, C. Eilenberger, S. F. Ali, S. Mühleder, [6] A. Bein, W. Shin, S. Jalili-Firoozinezhad, M. H. Park, E. Priglinger, M. Harasek, H. Redl, W. Holnthoner, P. Ertl, Biomi- A. Sontheimer-Phelps, A. Tovaglieri, A. Chalkiadaki, H. J. Kim, crofluidics 2018, 12, 042216. D. E. Ingber, Cell. Mol. Gastroenterol. Hepatol. 2018, 5, 659. [38] R. Booth, H. Kim, Lab Chip 2012, 12, 1784. Small 2021, 17, 2006012 2006012 (13 of 15) © 2021 The Authors. Small published by Wiley-VCH GmbH www.advancedsciencenews.com www.small-journal.com [39] D. Mandt, P. Gruber, M. Markovic, M. Tromayer, M. Rothbauer, [69] O. Y. F. Henry, R. Villenave, M. J. Cronce, W. D. Leineweber, S. R. A. Krayz, F. Ali, J. Van Hoorick, W. Holnthoner, S. Mühleder, M. A. Benz, D. E. Ingber, Lab Chip 2017, 17, 2264. P. Dubruel, S. Van Vlierberghe, P. Ertl, R. Liska, A. Ovsianikov, [70] Y. Mermoud, M. Felder, J. D. Stucki, A. O. Stucki, O. T. Guenat, Int. J. Bioprint. 2018, 4, 144. Sens. Actuators, B 2018, 255, 3647. [40] C. Drieschner, S. Könemann, P. Renaud, K. Schirmer, Lab Chip [71] T. Sun, S. Tsuda, K.-P. Zauner, H. Morgan, Biosens. Bioelectron. 2019, 19, 3268. 2010, 25, 1109. [41] A. Artzy-Schnirman, N. Hobi, N. Schneider-Daum, O. T. Guenat, [72] J. Kahle, R. Levin, W. Niles, B. Rasnow, M. Schehlein, C. Shumate, C.-M. Lehr, J. Sznitman, Eur. J. Pharm. Biopharm. 2019, 144, 11. J. Assoc. Lab. Autom. 2010, 15, 355. [42] S. Ehrmann, O. Schmid, C. Darquenne, B. Rothen-Rutishauser, [73] S. R. Heidemann, P. Lamoureux, K. Ngo, M. Reynolds, J. Sznitman, L. Yang, H. Barosova, L. Vecellio, J. Mitchell, R. E. Buxbaum, BioTechniques 2003, 35, 708. N. Heuze-Vourc’h, Expert Opin. Drug Delivery 2020, 17, 463. [74] M. M. Frigault, J. Lacoste, J. L. Swift, C. M. Brown, J. Cell Sci. 2009, [43] L. J. van den Broek, L. I. J. C. Bergers, C. M. A. Reijnders, S. Gibbs, 122, 753. Stem Cell Rev. Rep. 2017, 13, 418. [75] L. Lin, S.-S. Wang, M.-H. Wu, C.-C. Oh-Yang, Sensors (Basel) 2011, [44] A. Bal-Öztürk, B. Miccoli, M. Avci-Adali, F. Mogtader, F. Sharifi, 11, 8395. B. Çeçen, G. Yaşayan, D. Braeken, E. Alarcin, Curr. Pharm. Des. [76] S. K. Mahto, V. Charwat, P. Ertl, B. Rothen-Rutishauser, S. W. Rhee, 2018, 24, 5437. J. Sznitman, Nanotoxicology 2015, 9, 381. [45] A. Grassart, V. Malardé, S. Gobaa, A. Sartori-Rupp, J. Kerns, [77] C. Rassie, J. Van Wyk, L. Wilson, N. Ntshongontshi, A. Jonnas, K. Karalis, B. Marteyn, P. Sansonetti, N. Sauvonnet, Cell Host U. Feleni, R. F. Ajayi, P. Baker, E. Iwuoha, J. Nano Res. 2017, 45, 164. Microbe 2019, 26, 435. [78] A. Jaitovich, S. Mehta, N. Na, A. Ciechanover, R. D. Goldman, [46] M. Maurer, M. S. Gresnigt, A. Last, T. Wollny, F. Berlinghof, K. M. Ridge, J. Biol. Chem. 2008, 283, 25348. R. Pospich, Z. Cseresnyes, A. Medyukhina, K. Graf, [79] R. Baudoin, L. Griscom, J. M. Prot, C. Legallais, E. Leclerc, M. Gröger, M. Raasch, F. Siwczak, S. Nietzsche, I. D. Jacobsen, Biochem. Eng. J. 2011, 53, 172. M. T. Figge, B. Hube, O. Huber, A. S. Mosig, Biomaterials 2019, [80] N. A. Stathopoulos, J. D. Hellums, Biotechnol. Bioeng. 1985, 27, 1021. 220, 119396. [81] L. Kim, Y. C. Toh, J. Voldman, H. Yu, Lab Chip 2007, 7, 681. [47] K. Kulthong, L. Duivenvoorde, H. Sun, S. Confederat, J. Wu, [82] Y. Kohl, G. J. Oostingh, A. Sossalla, A. Duschl, H. von Briesen, B. Spenkelink, L. de Haan, V. Marin, M. van der Zande, H. Thielecke, Nanoscale Res. Lett. 2011, 6, 505. H. Bouwmeester, Toxicol. In Vitro 2020, 65, 104815. [83] E. Elje, E. Mariussen, O. H. Moriones, N. G. Bastús, V. Puntes, [48] M. Gröger, J. Dinger, M. Kiehntopf, F. T. Peters, U. Rauen, Y. Kohl, M. Dusinska, E. Rundén-Pran, Nanomaterials 2020, 10, 545. A. S. Mosig, Adv. Healthcare Mater. 2018, 7, 1700616. [84] E. Elje, M. Hesler, E. Rundén-Pran, P. Mann, E. Mariussen, [49] J. Deng, Z. Chen, X. Zhang, Y. Luo, Z. Wu, Y. Lu, T. Liu, W. Zhao, S. Wagner, M. Dusinska, Y. Kohl, Mutat. Res., Toxicol. Environ. B. Lin, Biomed. Microdevices 2019, 21, 57. Mutagen. 2019, 845, 403033. [50] K. H. Dodson, F. D. Echevarria, D. Li, R. M. Sappington, J. F. Edd, [85] B. Srinivasan, A. R. Kolli, M. B. Esch, H. E. Abaci, M. L. Shuler, Biomed. Microdevices 2015, 17, 114. J. J. Hickman, J. Lab. Autom. 2015, 20, 107. [51] H. J. Kim, D. E. Ingber, Integr. Biol. 2013, 5, 1130. [86] P. Miranda-Azpiazu, S. Panagiotou, G. Jose, S. Saha, Sci. Rep. [52] E. Chin, E. Goh, Methods Cell Biol. 2018, 146, 159. 2018, 8, 8784. [53] T. T. G. Nieskens, M. J. Wilmer, Eur. J. Pharmacol. 2016, 790, 46. [87] Q. Ramadan, F. C. W. Ting, Lab Chip 2016, 16, 1899. [54] K. Achberger, C. Probst, J. Haderspeck, S. Bolz, J. Rogal, [88] A. Skardal, S. V Murphy, M. Devarasetty, I. Mead, H.-W. Kang, J. Chuchuy, M. Nikolova, V. Cora, L. Antkowiak, W. Haq, N. Shen, Y.-J. Seol, Y. Shrike Zhang, S.-R. Shin, L. Zhao, J. Aleman, K. Schenke-Layland, M. Ueffing, S. Liebau, P. Loskill, Elife 2019, 8, A. R. Hall, T. D. Shupe, A. Kleensang, M. R. Dokmeci, S. Jin Lee, e46188. J. D. Jackson, J. J. Yoo, T. Hartung, A. Khademhosseini, S. Soker, [55] L. Boeri, L. Izzo, L. Sardelli, M. Tunesi, D. Albani, C. Giordano, C. E. Bishop, A. Atala, Sci. Rep. 2017, 7, 8837. Bioengineering (Basel, Switzerland) 2019, 6, 91. [89] U. Marx, T. B. Andersson, A. Bahinski, M. Beilmann, S. Beken, [56] T. Satoh, S. Sugiura, K. Shin, R. Onuki-Nagasaki, S. Ishida, F. R. Cassee, M. Cirit, M. Daneshian, S. Fitzpatrick, O. Frey, C. Gaertner, K. Kikuchi, M. Kakiki, T. Kanamori, Lab Chip 2017, 18, 115. C. Giese, L. Griffith, T. Hartung, M. B. Heringa, J. Hoeng, W. H. de Jong, [57] J. H. Sung, Methods Cell Biol. 2018, 146, 183. H. Kojima, J. Kuehnl, M. Leist, A. Luch, I. Maschmeyer, D. Sakharov, [58] I. Wagner, E.-M. Materne, S. Brincker, U. Süssbier, C. Frädrich, A. J. A. M. Sips, T. Steger-Hartmann, D. A. Tagle, A. Tonevitsky, T. Tralau, M. Busek, F. Sonntag, D. A. Sakharov, E. V. Trushkin, S. Tsyb, A. van de Stolpe, R. Vandebriel, P. Vulto, J. Wang, J. Wiest, A. G. Tonevitsky, R. Lauster, U. Marx, Lab Chip 2013, 13, 3538. M. Rodenburg, A. Roth, ALTEX 2016, 33, 272. [59] J. H. Sung, M. B. Esch, J.-M. Prot, C. J. Long, A. Smith, [90] M. Mastrangeli, S. Millet, T. Orchid Partners, J. J. Hickman, M. L. Shuler, Lab Chip 2013, 13, 1201. J. Van den Eijnden-van Raaij, ALTEX 2019, 36, 650. [60] Y. Temiz, R. D. Lovchik, G. V. Kaigala, E. Delamarche, Microelectron. [91] Y. S. Zhang, J. Aleman, S. R. Shin, T. Kilic, D. Kim, S. A. Mousavi Eng. 2015, 132, 156. Shaegh, S. Massa, R. Riahi, S. Chae, N. Hu, H. Avci, W. Zhang, [61] T. H. Park, M. L. Shuler, Biotechnol. Prog. 2003, 19, 243. A. Silvestri, A. Sanati Nezhad, A. Manbohi, F. De Ferrari, A. Polini, [62] Y. Tanaka, K. Sato, T. Shimizu, M. Yamato, T. Okano, T. Kitamori, G. Calzone, N. Shaikh, P. Alerasool, E. Budina, J. Kang, N. Bhise, Biosens. Bioelectron. 2007, 23, 449. J. Ribas, A. Pourmand, A. Skardal, T. Shupe, C. E. Bishop, [63] J. El-Ali, P. K. Sorger, K. F. Jensen, Nature 2006, 442, 403. M. R. Dokmeci, A. Atala, A. Khademhosseini, Proc. Natl. Acad. Sci. [64] S. Petronis, M. Stangegaard, C. Bovöge Christensen, M. Dufva, U. S. A. 2017, 114, E2293. BioTechniques 2006, 40, 368. [92] B. M. Maoz, A. Herland, O. Y. F. Henry, W. D. Leineweber, [65] P. Shah, A. Kaushik, X. Zhu, C. Zhang, C.-Z. Li, Analyst 2014, 139, M. Yadid, J. Doyle, R. Mannix, V. J. Kujala, E. A. FitzGerald, 2088. K. K. Parker, D. E. Ingber, Lab Chip 2017, 17, 2294. [66] H. A. Alhadrami, Biotechnol. Appl. Biochem. 2018, 65, 497. [93] P. M. Misun, J. Rothe, Y. R. F. Schmid, A. Hierlemann, O. Frey, [67] T. Sun, E. J. Swindle, J. E. Collins, J. A. Holloway, D. E. Davies, Microsyst. Nanoeng. 2016, 2, 16022. H. Morgan, Lab Chip 2010, 10, 1611. [94] C. Oleaga, C. Bernabini, A. S. T. Smith, B. Srinivasan, M. Jackson, [68] A. Susloparova, D. Koppenhöfer, J. K. Y. Law, X. T. Vu, W. McLamb, V. Platt, R. Bridges, Y. Cai, N. Santhanam, B. Berry, S. Ingebrandt, Lab Chip 2015, 15, 668. S. Najjar, N. Akanda, X. Guo, C. Martin, G. Ekman, M. B. Esch, Small 2021, 17, 2006012 2006012 (14 of 15) © 2021 The Authors. Small published by Wiley-VCH GmbH www.advancedsciencenews.com www.small-journal.com J. Langer, G. Ouedraogo, J. Cotovio, L. Breton, M. L. Shuler, [101] M. Sramkova, K. Kozics, V. Masanova, I. Uhnakova, J. J. Hickman, Sci. Rep. 2016, 6, 20030. F. Razga, V. Nemethova, P. Mazancova, L. Kapka- [95] M. Hosokawa, A. Arakaki, M. Takahashi, T. Mori, H. Takeyama, Skrzypczak, M. Kruszewski, M. Novotova, V. F. Puntes, T. Matsunaga, Anal. Chem. 2009, 81, 5308. A. Gabelova, Mutat. Res., Toxicol. Environ. Mutagen. 2019, 845, [96] J. Owen, M. Kuznecovs, R. Bhamji, N. William, N. Domenech-Garcia, 403017. M. Hesler, T. Knoll, Y. Kohl, A. Nelson, N. Kapur, Rev. Sci. Instrum. [102] C. M. Kurz, A. Maurer, K. Thees, S. Schillberg, T. Velten, 2020, 91, 025002. H. Thielecke, Sens. Actuators, B 2011, 158, 345. [97] N. G. Bastús, F. Merkoçi, J. Piella, V. Puntes, Chem. Mater. 2014, [103] N. El Yamani, A. R. Collins, E. Rundén-Pran, L. M. Fjellsbø, 26, 2836. S. Shaposhnikov, S. Zienolddiny, M. Dusinska, Mutagenesis 2017, [98] A. Van Blaaderen, A. Vrij, Langmuir 1992, 8, 2921. 32, 117. [99] A. Pottier, S. Cassaignon, C. Chanéac, F. Villain, E. Tronc, [104] A. R. Collins, Methods Mol. Biol. 2017, 1641, 173. J.-P. Jolivet, J. Mater. Chem. 2003, 13, 877. [105] A. Azqueta, J. Slyskova, S. A. S. Langie, I. O’Neill Gaivão, [100] N. G. Bastús, J. Comenge, V. Puntes, Langmuir 2011, 27, 11098. A. Collins, Front. Genet. 2014, 5, 288. Small 2021, 17, 2006012 2006012 (15 of 15) © 2021 The Authors. Small published by Wiley-VCH GmbH