Automated Droplet-Based Microfluidic Platform for Multiplexed

Each droplet represented an individual enzymatic assay carried out in a ... secretions of reactive oxygen species, proteases, or cytokines in microflu...
0 downloads 0 Views 803KB Size
Subscriber access provided by University of Sussex Library

Article

An automated droplet-based microfluidic platform for multiplexed analysis of biochemical markers in small volumes Diana Fabiola Cedillo-Alcantar, Yong Duk Han, Jonghoon Choi, Jose Luis Garcia-Cordero, and Alexander Revzin Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.8b05689 • Publication Date (Web): 05 Mar 2019 Downloaded from http://pubs.acs.org on March 6, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

An automated droplet-based microfluidic platform for multiplexed analysis of biochemical markers in small volumes Diana F. Cedillo-Alcantar1,2‡, Yong Duk Han2‡, Jonghoon Choi2, Jose L. Garcia-Cordero1*, and Alexander Revzin2* 1Unidad

Monterrey, Centro de Investigación y de Estudios Avanzados del Instituto Politécnico Nacional, Vía del Conocimiento 201, Parque PIIT, Apodaca, Nuevo León CP 66628, México 2Mayo Clinic, Rochester, MN, 55905, United States ABSTRACT: The ability to detect multiple analytes in a small sample volume has significance for numerous areas of research, including organs-on-chip, small animal experiments and neonatology. The objective of this study was to develop an automated microfluidics platform for multiplexed detection of analytes in microliter sample volumes. This platform employed computercontrolled microvalves to create laminar co-flows of sample and assay reagent solutions. It also contained valve-regulated crossjunction for discretizing sample/reagent mixtures into water-in-oil droplets. Microfluidic automation allowed us to control parameters related to frequency of droplet generation, the number of droplets of the same composition as well as the size of droplets. Each droplet represented an individual enzymatic assay carried out in a sub-nanoliter (0.8 nL) volume reactor. An enzymatic reaction involving target analyte and assay reagents produced colorimetric or fluorescent signals in droplets. Importantly, intensity of optical signal was proportional to the concentration of analyte in question. This microfluidic bioanalysis platform was used in conjunction with commercial “mix-detect” assays for glucose, total bile acids and lactate dehydrogenase (LDH). After characterizing these assays individually, we demonstrated sensitive multiplexed detection of three analytes from as little as 3 µL. In fact, this volume was sufficient to generate multiple repeat droplets for each of the three biochemical assays as well as positive control droplets confirming quality of assay reagents and negative control droplets to help with background subtraction. One potential application for this microfluidic bioanalysis platform involves sampling cell-conditioned media in organ-on-chip devices. To highlight this application, hepatocyte spheroids were established in microfluidic devices, injured on-chip by exposure to lipotoxic agent (palmitate) and then connected to the bioanalysis module for daily monitoring of changes in cytotoxicity (LDH), energy metabolism (glucose) and liver function (total bile acids). Microfluidic in-droplet assays revealed increased levels of LDH as well as reduction in bile acid synthesis – results that were consistent with hepatic injury. Importantly, these experiments highlighted the fact that in-droplet assays were sufficiently sensitive to detect changes in functional output of a relatively small (~100) number of hepatocyte spheroids cultured in a microfluidic device. Moving forward, we foresee increasing multiplexing capability of this technology and applying this platform to other biological/medical scenarios where detection of multiple analytes from a small sample volume is desired.

Miniaturization and microfluidics are increasingly being applied to cultivation of cells for organ-on-chip applications.1-5 There have been a number of reports of microfluidic cultures populated with cells and designed to mimic such organs as liver,6,7 lung,8,9 gut10,11 and pancreas.12,13 While microfluidic cell cultures offer numerous advantages, from smaller consumptions of cells/reagents to interesting biological phenomena related to high cell number to volume ratio,14-17 such cultures also present new challenges. One such challenge is sampling and analyzing functional output in cell cultures with typical volumes of 1 to 5 µL. A number of approaches have been developed for sensitive detection of cell function in small volumes.18-20 Some of these approaches focused on immobilizing sensing elements inside microfluidic devices, in close proximity to cells. For example, the Heath group and others described immobilizing antibody arrays in microfluidic devices for multiplexed detection of cell-secreted proteins.21,22 Our group has previously developed electrodes functionalized with enzymes,23,24 peptides25 or aptamers26,27 for monitoring cellular secretions of reactive oxygen species, proteases or cytokines in microfluidic devices.

However, such “immobile” biosensors are often unsuitable for multi-day measurements under cell culture conditions due to saturation of antibody/aptamer binding sites or enzyme degradation. Furthermore, biorecognition elements such as enzymes or antibodies often become less active upon immobilization. An alternative strategy involves “mobile” biosensors where sensing elements are introduced into a microfluidic cell culture system during measurements, are subsequently washed out and replaced by the new batch of sensing elements. One of the early examples of such sensing approach was reported by Kennedy and co-workers who developed electrophoresis immunoassay in microfluidic devices for optical monitoring of hormone secretions from pancreatic islets with high temporal resolution.28,29 Of note here is recent report describing an electrochemical biosensor employing disposable antibodyfunctionalized beads for long-term monitoring of protein production in microfluidic cell cultures.30 Our lab has recently used bead-based immunoassay for optical detection of growth factor production in microfluidic hepatocyte cultures.31

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 1. Illustrations of the droplet microfluidic device and its operation. (A) Schematic illustration for concept of developed droplet-based biochemical assay. In the developed droplet device, three-different biochemical assays including glucose, LDH and bile acid assay could be accomplished at the same time. Since the water-in-oil droplet acts as an isolated biochemical reactor, each assay can be achieved without mutual-interferences of samples and reagents. (B) Photographs of hepatocyte-culturing microfluidic device and droplet device. For the cell injury analysis application, droplet device was connected with cell culturing device and the changes in biochemical metabolites in cell culture media was analyzed in the droplet assay module. (C) Photograph of droplet device. The microfluidic channels in flow layer and control layer were filled with red ink and blue ink, respectively. In most of the examples described above novel assays or sensing strategies were developed for the purpose of monitoring cell cultures in microfluidic devices. In our experience such novel assays take months to years to develop and validate. In contrast, there are a large number of robust and validated bulk-scale assays available commercially for monitoring indicators of cell function. While principles of action vary, a large proportion of such conventional assays employ redox enzymes and color- or fluorescence-emitting compounds to produce optical signal. One of our objectives was to adapt such conventional microtiter plate-based optical assays to analysis in small volumes of microfluidic devices. The bioanalysis platform described in this paper (see Figure 1) utilized two microfluidic technologies, microfluidic automation and droplet generation, for miniaturization and multiplexing of conventional assays. Microfluidic automation, the use of pneumatic microvalves to control liquid flow,32-35 was used to mix nanoliter volumes of sample and assay reagents in microfluidic devices. These binary sample/assay reagent mixtures were discretized into nanoliter water-in-oil droplets using pneumatic microvalves controlling flow through water-oil crossflow junction. Computer-controlled pneumatic valves enabled fluid flow actions such as flow and stop, reagent/sample loading, generation of water-in-oil droplets and incubation. To demonstrate the wide applicability of our platform, we implemented three enzymatic assays: a colorimetric glucose assay as well as fluorometric assays for LDH and total bile acids (Figure 1A). We then connected this bioanalysis platform to microfluidic cultures of hepatocyte spheroids and

monitored changes in hepatic viability/function daily during four days of injury (Figure 1B). In-droplet assays were sufficiently sensitive to detect changes in levels of LDH, glucose and bile acids in media conditioned by a small number of cells (144 spheroids). The concept of in-droplet microfluidic assays can be enhanced in the future by expanding the panel of analytes measured simultaneously as well as by adding immunoassays for on-chip detection of secreted proteins.

MATERIALS AND METHODS Fabrication of microfluidic devices. The dropletgenerating microfluidic device comprising a flow and a control layer was fabricated by multi-layer soft lithography 34,35A schematic detailing flow and control layers is presented in Figure S1. The microfluidic device for cultivation of hepatocytes consisted of two layers: a microwells layer and a flow layer (Figure S2A). There were 144 microwells, each with diameter of 350-µm and depth of 500-µm and the flow layer consisted of one chamber and two transport channels (See Supplementary Materials and Methods section S2 and S3 in Supporting Information for device fabrication details). Cultivation of hepatocytes in microfluidic devices. Primary rat hepatocytes were isolated from adult female Lewis rats weighing 110-200 g (Charles River Laboratories, Boston, MA, USA) using a two-step collagenase perfusion procedure.36 All animal experiments were performed under the National Institutes of Health (NIH) guidelines for the ethical care and use of laboratory animals, and the experimental protocol was approved by the Institutional Animal Care and

ACS Paragon Plus Environment

Page 2 of 9

Page 3 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Use Committee of the Mayo Clinic, Rochester, MN. The isolated primary hepatocytes were cultured in complete media with hormones (CH media). Cells (250 µL at 3×106 cells/mL) were seeded into devices as shown in Figure 1B. The spheroids were kept in culture for 3 days before starting metabolite measurements. Detailed descriptions for process of hepatocyte seeding and spheroid formation is presented in Supplementary Materials and Methods section S4 and Figure S2B. Operation of the automated microfluidic bioanalysis module. Fluid flow and valves actuation were controlled via a custom-made pneumatic system and LabVIEW interface. Experiments were performed on a fluorescence microscope (IX-83, Olympus, USA) outfitted with an environmental chamber (GM-8000, TOKAI HIT, Japan). The acquired microscopic images were analyzed for intensity of visible color or fluorescence using custom MATLAB script. Detailed information for device operation was described in Supplementary Materials and Methods section S5. Characterizing in-droplet biochemical assays. Glucose, LDH and bile acid assays were characterized in microfluidic devices and in standard microtiter plates. For glucose detection, we spiked glucose in DPBS into CH media to achieve eight different concentrations (0, 0.4, 0.8, 1.6, 3.1, 6.3, 12.5 and 25 mM). Similar dilutions were prepared for LDH, achieving concentrations of 0, 1.25, 2.5, 5, 10, 20, 40, and 80 U/L. For total bile acid assay, we followed manufacturer’s instructions in using sodium glycochenodeoxycholate as a reference bile acid. This compound was dissolved in CH media to achieve concentrations of 0, 8.1, 16.3, 32.5, 75, 150, 300 and 600 µM. The detailed procedures for preparation of assay reagent solutions and calibration study were presented in Supplementary Materials and Methods section S6 and S8. Monitoring hepatic injury using in-droplet microfluidic assays. Microfluidic devices containing hepatocyte spheroids were exposed to 400 µM palmitate (lipotoxic agent) in CH media for 24h to induce injury. In a control experiment, microfluidic cultures of hepatocytes were exposed to CH media without palmitate. After injuring cells for 24h, microfluidic cell culture device was connected to the bioanalysis module using Tygon tubing (Figure 1B). The inlets of the bioanalysis module were loaded with negative and positive control solutions as well as reagents for bile acid, LDH and glucose assays. The positive control solution contained 24.6 mM of glucose, 75 U/L of LDH, and 150 mM of bile acid. The negative control solution was DMEM media without glucose, LDH or bile acids. After both cell culture and bioanalysis devices were connected, they were placed in the environmental chamber mounted onto a fluorescence microscope for time-lapse imaging of in-droplet biochemical assays. The sequence of droplet formation was as follows: three droplets were formed by mixing glucose assay reagents with positive control solution, followed by three droplets of glucose assay reagents mixed with the negative control solutions, and three droplets formed by mixing glucose assay reagents with media from the cell culture device. The same sequence was repeated for LDH assay and then bile acid assay (27 droplets in total). Microfluidic cell cultures were analyzed every 24 h over the course of 96 h. Each analysis session lasted for 30 min.

RESULTS AND DISCUSSIONS

Design and operation of microfluidic in-droplet assays. The goal of this paper was to develop a microfluidic platform for miniaturizing and multiplexing conventional enzymatic assays. Key technological components used to accomplish this goal were pneumatic automated microvalves and droplet generation. Microvalves were ideally suited for on-chip mixing of solutions while droplets loaded with sample and assay components served as independent nanoliter wells where biochemical reactions occurred. The droplet-generating microfluidic device contained a control layer with ten pneumatic microvalves (Figure 1C and Figure S1B). The flow layer included six channels to introduce reagents and samples, a waste channel where the six channels converged, and a long serpentine channel for incubation of droplets (Figure 1C). Using a LabVIEW program, the microvalves were sequentially opened to allow for mixing of solutions and for automatic generation of water-in-oil droplets. The workflow of the droplet generation proceeded in steps described in Figure S3. First, mineral oil was injected into the main channel at 3 psi for 15 min. Next, valves were actuated to inject a small aliquot of media and reagents for one assay into the cross-flow junction where water-in-oil droplets were generated. Due to laminar nature of the flow minimal mixing between the two flow streams was observed prior to formation of droplets. After generation of droplets for biochemical assay 1 was completed, a new set of valves was opened to mix media with reagents for biochemical assay 2. Importantly, as shown in Figure S3, step 2, the valves were actuated to route media to waste in order to flush preceding solution out of the system. Subsequently, a new set of droplets containing a mixture of media and biochemical assay 2 was ejected into serpentine incubation channel (Figure S3, step 3). This sequence of steps could be repeated multiple times to populate incubation channel with droplets. The format of the device developed here, with six independently controlled inputs, allowed to analyze 5 biomarkers based on a small volume of input sample. We should note that once needed number of droplets was generated the flow in the device was stopped for 15 min to allow for biochemical reactions to take place inside the droplets and for optical signal to develop. The device can be reused multiple times by sweeping away the droplets from the serpentine channel. The Supplementary Video S1 highlights principle of operation of the device. Characterization of droplet generation. In the next set of experiments, we wanted to demonstrate that our automated microfluidic device allows controlling the volume of each droplet and the number of droplets formed. We initially evaluated the performance of our device to generate droplets of different volumes. As shown in Figure 2A, with valve V1 closed and valve V2 opened, solutions of yellow and blue dyes co-flow without mixing. Closing V2 and opening V1 pushes this mixture of yellow and blue dyes into the oil stream. Closing V1 for a brief time splits a section of the solution and creates a droplet. Immediately after the droplet is formed, yellow, blue and green color can be observed in the droplet, indicating that mixing has begun (Figure 2A, upper right panel). 10 s later, the droplet already shows a homogeneous green color indicating complete mixing (Figure 2A, lower right panel). We determined experimentally that the pneumatic pressure for aqueous and oil phases were 2 psi and 2.5 psi, respectively. Our platform allows to (i) generate droplets of different combinations of reagents and samples, (ii) tune the droplet volume as droplets are generated, (iii) precisely control

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

the amount of droplets, and (iii) create sequences of different combinations of droplets. These scenarios are shown in Figure 2B, in which droplets were generated with the desired sequences with varied volumes and color-combinations.

fluorescence signals in droplets therefore we chose assays with colorimetric and florigenic readouts. In all experiments involving in-droplet assays, our microfluidic devices were used to produce 0.84 nL volume droplets every 5 s.

Figure 2. Generation of droplets under varied parameter conditions. (A) Image of the droplet generation using two food dyes. Upper right panel shows a droplet right after the solutions enters in the oil-phase channel. The droplet still exhibits blue and yellow colors indicating incomplete mixing. Lower right panel shows a droplet which has been incubated for 10 s. Notice the homogeneous green color indicating complete mixing. (B) Photograph of droplets generated under varied sampling conditions. Droplets of different color combinations and different lengths were generated by actuating pneumatic valves. Scale bar = 1 mm. (C) Graph shows the correlation between the droplet volume and sampling time. Droplet volume can be tuned by controlling the time that valve V1 remains open. Adjusting this time from 0.05 to 0.25 s, produced droplets with lengths ranging from 220 to 720 µm (or 0.8 to 2.6 nL), as shown in Figure 2B and 2C). The number of droplets loaded into the same incubation channel may be controlled by the operating frequency of valve V1. Figure S4 shows droplets generated with valve operating frequency ranging from 15 to 250 mHz. As seen from these images, while droplet volume is kept the same (0.8 nL) by controlling duration of valve opening, the number of droplets loaded into a serpentine channel varies from 3 to 48. The results in Figures 2 and S4 demonstrate versatility of our microfluidic platform in controlling the size and number of droplets generated. Design of biochemical assays. To prove the concept of indroplet biochemical assays, we chose to monitor analytes associated with cell viability (LDH), energy metabolism (glucose) and liver-specific function (bile acid). We also wanted to show the ability to monitor absorbance and

Figure 3. Calibration curve for the colorimetric glucose assay. (A) Image of droplets containing different concentrations of glucose, ranging from 0-25 mM. (B) Enlarged images of representative droplets shown in A. (C) Calibration curve for the glucose assay. The glucose detection assay contained glucose oxidase (GOx) and horseradish peroxidase (HRP) as an enzyme pair to catalyze glucose, and 4-AAP/ADOS as a color-generating reagent.37,38 Enzymatic reaction underlying glucose detection assay is shown in Figure 1A. In this reaction, oxidation of glucose by GOx results in formation of H2O2 which in turn becomes oxidized to water in the presence of HRP. This HRPcatalyzed reaction produces magenta color due to reduction of color reagents (4-AAP/ADOS). The assay components were loaded into one of the inlets of the device while solutions with different concentrations of glucose were loaded into other inlets. Subsequently, contents of the assay were mixed with different concentrations of glucose by sequentially actuating pneumatic valves controlling flow from the inlets. Three droplets were generated per each concentration of glucose. These droplets were incubated under static conditions in a serpentine region of the microfluidic device for 5 min to allow for reaction and color development to take place. As shown in Figure 3A and 3B, the color intensity of droplets is proportional to glucose concentrations ranging from 0.39 to 25 mM. Intensity of magenta color was analyzed using MATLAB image analysis algorithm (Figure S5 and Figure 3C). The limit of detection (LOD), defined as three standard deviations above the blank, was determined to be 70 µM, while the mean coefficient of variation (COV) was calculated as 10.8%, indicating reproducibility. While commercial glucose assays report better sensitive (e.g. colorimetric glucose assay kit from

ACS Paragon Plus Environment

Page 4 of 9

Page 5 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Cell Biolabs, Inc., reports 7 µM of detection limit), the assay developed here was sufficiently sensitive to achieve objective of the present study – monitoring changes in metabolism of hepatocyte spheroids. The limit of detection for on-chip glucose assay may be improved in the future by further optimizing concentration of enzymes and chromogenic substrates.

Figure 4. Calibration curve for the fluorometric LDH assay. (A) Image of droplets containing different concentrations of LDH, ranging from 0-80 U/L. (B) Enlarged images of representative droplets shown in A. (C) Calibration curve for the fluorometric LDH assay. LDH is an intracellular enzyme in live cells but leaches out from dead/dying cells with disrupted cell membrane. The biochemical assay for detection of LDH is described in Figure 1A. This assay includes exogenous lactate that acts as substrate for LDH as well NAD+- a co-enzyme that becomes reduced to NADH by accepting electrons from LDH and is then oxidized by another enzyme, diaphorase (DPR).39,40 A signal is generated when non-fluorescent resazurin receives electrons from DPR and becomes a fluorescent compound resorufin (see Figure 1A). Mixing varying concentrations of LDH with assay components in the microfluidic device resulted in droplets with different levels of fluorescence signal as seen from Figure 4A to 4C. Quantification of fluorescence intensity revealed that this in-droplet microfluidic assay had a limit of detection of 0.5 U/L and COV of 4.1%. This limit of detection is comparable to commercial assays while volume requirements of our system are two orders of magnitude lower than standard bulk assay (1 µL vs 100 µL). In addition to LDH and glucose, which are generic indicators of cell function, we wanted to monitor a liver-specific marker. Bile acids are produced by hepatocytes and play an important role in digestion of food and elimination of waste products.41-43 We adapted a commercial assay for in-droplet detection of total bile acids. This enzymatic assay (Figure 1A) is based on the breakdown of a representative bile acid, 3α-

hydroxysteroid, by HSD and concomitant conversion of coenzyme NAD+ to NADH.

Figure 5. Calibration curve for the total bile acid assay. (A) Result image of droplet-based fluorometric bile acid assay for bile acid samples at concentrations from 0 to 150 µM. (B) Enlarged images of representative droplets shown in (A). (C) Calibration curve for droplet-based bile acid assay. Similar to LDH detection assay, fluorescence signal resulted from DRP oxidizing NADH then contributing to oxidation of non-fluorescent resazurin to fluorescent resorufin.44,45 Figure 5A-C show a representative calibration experiment where different concentrations of model bile acid (2 to 150 µM) were mixed with assay reagents and discretized into droplets. As seen from these images, in-droplet fluorescence intensity was proportional to the concentration of bile acid with linear range of 0 µM to 150 µM. Limit of detection (LOD) defined as three times over standard deviation (3σ) and COV were determined to be 2.1 µM and 4.7%, respectively. We carried out additional cross-reactivity experiments by creating matched and mis-matched mixtures of target analyte and assay reagents in a microtiter plate. The results of these experiments presented in Figure S6 demonstrate that only matched pairings produced absorbance or fluorescence signals. These experiments once again confirm specificity of the assays employed here. The performance of the in-droplet assay was comparable to commercial bulk scale assay for detection of total bile acids (e.g. a fluorometric total bile acid assay kit from BioAssay Systems LLC reports LOD of 1 µM with linear range extending to 150 µM). On-chip detection of liver injury using in-droplet microfluidic assays. Upon characterizing in-droplet assays individually, we wanted to highlight the possibility of running three different assays in the same device. Furthermore, we wanted to utilize microfluidic bioanalysis module in a setting where sampling and analysis in small volumes is of benefit.

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

While there are multiple applications for such a microfluidic device, from neonatology to small animal research, our focus in this paper was on analysis of microfluidic cell culture systems. Such systems, also known as organs-on-chip, are emerging as a valuable tool for elucidating interactions between different model organs or different cell types comprising the same organ.1-5 Organs-on-chip offer numerous benefits, they may contain human cells and may be used to model responses to injurious agents or drugs, however, small volumes of microfluidic cultures present challenges for analysis of cell function. While there are efforts to develop droplet-based microfluidic assays for cell analysis,46 to the best of our knowledge multiplexed analysis of cell function has not been demonstrated using such assays. In our study, indroplet microfluidic assays were utilized to monitor injury and

function of hepatocytes exposed to a toxic lipid, palmitate, in microfluidic devices. Palmitate-induced lipotoxicity causes cell damage and cell death via several pathways, including the activation of tumor necrosis factor-related apoptosis-inducing ligand receptor (TRAIL)-2 and/or the endoplasmic reticulum stress-mediated mitochondrial apoptotic pathway.47-50 To analyze response of hepatocytes to lipotoxic injury, we connected cell culture chamber to a bioanalysis module via Tygon tubing as shown in Figure 1B. This set-up offered considerable flexibility in being able to connect or disconnect bioanalysis module according to the needs of the experiment. However, connecting tubing increased the total volume of the system and possibly contributed to dilution of cell-conditioned media.

Figure 6. Cultivation and injury of hepatocyte spheroids in microfluidic devices. (A,B) Representative images of primary rat hepatocyte spheroids formed in the devices after 3 days of incubation with (A) C media and (B) palmitate. Scale bar: 350 µm. (C) Representative images of droplet-based metabolite assays for injured hepatocyte. Colorimetric glucose assay, fluorometric LDH assay and fluorometric bile acid assay were performed on samples supplied from the hepatocyte-culturing device. To validate the assay reagents and compensate for any spurious signal, positive and negative control solutions were analyzed in parallel. Left panel shows bright-filed images for the colorimetric glucose assay and right panel shows fluorescence images for LDH and bile acid assay. We carried out experiments to characterize sample dilution upon transfer of media from cell culture module to bioanalysis module. In these experiments, TRITC-dextran (MW 70 kDa) was loaded into cell culture chamber in lieu of cell-secreted product. Fresh CH media was slowly injected into the inlet port of a cell culture chamber (750 nL/min), displacing solution of fluorescent dextran and pushing it into a bioanalysis module (see Figure S7A). A small aliquot of this fluorescent solution was discretized into 300 droplets and analyzed using fluorescence microscopy. Analysis of indroplet fluorescence intensity revealed that first 20 droplets had high and relatively constant fluorescence whereas last 20 droplets had faint fluorescence. These experiments suggested that while mixing and dilution due to introduction of fresh media significantly decreased concentration of a model cell signal, the leading edge of cell-conditioned media entering bioanalysis module could be expected to retain high concentration of these signals. Moving forward, we used first 9 droplets for analysis of hepatic injury and function. Schematic of microfluidic devices used for spheroid formation

is shown in Figure S2. In a typical experiment, primary rat hepatocytes were seeded into a microfluidic device and formed 150 µm diameter spheroids within 24h. A cell culture chamber contained 144 such spheroids. Bright-field images of hepatocyte spheroids are shown in Figure 6A and 6B. The design of the device allowed us to carry out relatively sophisticated biochemical assays sequentially using droplet generation. In addition to analyzing the same small volume of hepatocyte-conditioned media for glucose, LDH and bile acids, each sensing session also included positive and negative control droplets for each analyte (see Figure 6C). The positive control droplets were spiked with known concentration of analyte in question and were used to assess quality of assay components while the negative control droplets were employed for background subtraction and normalization. As can be seen from Figure 6C, nine droplets where generated for each analyte being detected, three droplets each for positive control, cell-conditioned media and negative control. After generating 27 droplets, the flow in the device was stopped to provide incubation time and allow for optical signal to develop

ACS Paragon Plus Environment

Page 6 of 9

Page 7 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry inside droplets. Each sensing session lasted for 30 min and was carried out under physiological conditions. Results of these sensing sessions are summarized in Figure 7.

to day 1), decreased synthesis of bile acids (2.8 fold lower on day 4 compared to day 1). Leaking of LDH was consistent with cell injury and cytotoxicity of palmitate while decrease in bile acid synthesis could be attributed to a combination of hepatic cell death and loss of hepatic function during injury. Higher levels of glucose in hepatocyte cultures injured with palmitate could also be attributed to cell death (hepatocytes consuming glucose at day 4) as well as to lower hepatocyte function resulting in less glucose consumption. Overall, results presented in Figures 6 and 7 highlight utility of microfluidic in-droplet assays for sampling media effluent and analyzing functional output of microfluidic cell cultures.

CONCLUSIONS In this paper, we report the development of a fully automated droplet-based microfluidic platform for the multiplexing and temporal analysis of metabolites from nL to µL volume samples. The integration of droplet microfluidic technology and the automatic actuation of microvalves allowed flexible and precise flow control in the generation of droplets for multiplexing biochemical assays. We demonstrated a 3-plex analysis of glucose, LDH and bile acid in the same small volume of media. Each assay also included on-chip positive controls to validate integrity of assay reagents and negative controls for background subtraction. Importantly, we demonstrated the feasibility of in-droplet colorimetric and fluorescence-based assays. To highlight one application of this technology, we connected bioanalysis platform to a microfluidic cell culture system containing hepatocyte spheroids for periodic sampling of hepatic function and viability during 4-day injury experiment. Exposure of hepatocytes to lipotoxic agent, palmitate, caused an increase in LDH, marker of cytotoxicity, and a decrease in bile acid synthesis, marker of hepatic function. Thus, our bioanalysis platform enabled periodic sampling of small volume of conditioned media from an organ-on-chip device. In this future, this bioanalysis platform may be redesigned to increase multiplexing capacity and may be integrated with organ-onchip devices for automated sampling and monitoring of interorgan communications. Figure 7. The use bioanalysis microfluidic device for monitoring hepatocyte spheroid responses to injury. Measurements were made once per day over the course of 4 days. (A) Glucose, (B) LDH, and (C) total bile acid were measured every 24 h.

As can be seen from these data, in the absence of palmitate, levels of glucose and bile acid production in hepatocyte spheroids remained unchanged while LDH production increased moderately by 28%. In comparison, hepatocyte spheroids exposed to 400 µM palmitate in microfluidic devices exhibited increased production of LDH (2 fold higher at day 4 compared to day 1), decreased synthesis of bile acids (2.8 fold lower on day 4 compared to day 1). Leaking of LDH was consistent with cell injury and cytotoxicity of palmitate while decrease in bile acid synthesis could be attributed to a combination of hepatic cell death and loss of hepatic function during injury. Higher levels of glucose in hepatocyte cultures injured with palmitate could also be attributed to cell death (hepatocytes consuming glucose at day 4) as well as to lower hepatocyte function resulting in less glucose consumption. As can be seen from these data, in the absence of palmitate, levels of glucose and bile acid production in hepatocyte spheroids remained unchanged while LDH production increased moderately by 28%. In comparison, hepatocyte spheroids exposed to 400 µM palmitate in microfluidic devices exhibited increased production of LDH (2 fold higher at day 4 compared

ASSOCIATED CONTENT Supporting Information The Supporting information is available free of charge on the ACS publications website at DOI: Supplementary Martials and Methods describing details in materials, device fabrication, cultivation of primary rat hepatocyte spheroids, device operation and biochemical assay reagents (S1S8), Supplementary Figures (Figure S1–S7) and Supplementary Video (Video S1) (DOCX and AVI).

AUTHOR INFORMATION Corresponding Author Jose Luis Garcia Cordero, PhD; email: [email protected] Alexander Revzin, PhD email: [email protected]

Author Contributions ‡These authors contributed equally.

Notes The authors declare no competing financial interest

ACKNOWLEDGMENT

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The authors thank Dr. Harmeet Malhi for valuable comments. This work was supported in part by NIH (DK107255). JLGC and DFCA acknowledge support from Mexico's CONACyT under grant CB-256097 and CB-286368.

REFERENCES (1) Bhatia, S. N.; Ingber, D. E. Nat. Biotechnol. 2014, 32, 760772. (2) Benam, K. H.; Dauth, S.; Hassell, B.; Herland, A.; Jain, A.; Jang, K. J.; Karalis, K.; Kim, H. J.; MacQueen, L.; Mahmoodian, R.; Musah, S.; Torisawa, Y. S.; van der Meer, A. D.; Villenave, R.; Yadid, M.; Parker, K. K.; Ingber, D. E. Annu. Rev. Pathol. 2015, 10, 195-262. (3) Esch, E. W.; Bahinski, A.; Huh, D. Nat. Rev. Drug Discov. 2015, 14, 248-260. (4) Kimura, H.; Sakai, Y.; Fujii, T. Drug Metab. Pharmacokinet. 2018, 33, 43-48. (5) Ronaldson-Bouchard, K.; Vunjak-Novakovic, G. Cell Stem Cell 2018, 22, 310-324. (6) Gori, M.; Simonelli, M. C.; Giannitelli, S. M.; Businaro, L.; Trombetta, M.; Rainer, A. PLoS One 2016, 11, e0159729. (7) Yu, F.; Deng, R.; Hao Tong, W.; Huan, L.; Chan Way, N.; IslamBadhan, A.; Iliescu, C.; Yu, H. Sci. Rep. 2017, 7, 14528. (8) Huh, D.; Leslie, D. C.; Matthews, B. D.; Fraser, J. P.; Jurek, S.; Hamilton, G. A.; Thorneloe, K. S.; McAlexander, M. A.; Ingber, D. E. Sci. Transl. Med. 2012, 4, 159ra147. (9) Xu, Z.; Li, E.; Guo, Z.; Yu, R.; Hao, H.; Xu, Y.; Sun, Z.; Li, X.; Lyu, J.; Wang, Q. ACS. Appl. Mater. Interfaces 2016, 8, 2584025847. (10) Kim, H. J.; Huh, D.; Hamilton, G.; Ingber, D. E. Lab Chip 2012, 12, 2165-2174. (11) Trietsch, S. J.; Naumovska, E.; Kurek, D.; Setyawati, M. C.; Vormann, M. K.; Wilschut, K. J.; Lanz, H. L.; Nicolas, A.; Ng, C. P.; Joore, J.; Kustermann, S.; Roth, A.; Hankemeier, T.; Moisan, A.; Vulto, P. Nat. Commun. 2017, 8, 262. (12) Bauer, S.; Wennberg Huldt, C.; Kanebratt, K. P.; Durieux, I.; Gunne, D.; Andersson, S.; Ewart, L.; Haynes, W. G.; Maschmeyer, I.; Winter, A.; Ammala, C.; Marx, U.; Andersson, T. B. Sci. Rep. 2017, 7, 14620. (13) Lee, S. H.; Hong, S.; Song, J.; Cho, B.; Han, E. J.; Kondapavulur, S.; Kim, D.; Lee, L. P. Adv. Healthc. Mater. 2018, 7. (14) Yu, H.; Meyvantsson, I.; Shkel, I. A.; Beebe, D. J. Lab Chip 2005, 5, 1089-1095. (15) Paguirigan, A. L.; Beebe, D. J. Integr. Biol. (Camb.) 2009, 1, 182-195. (16) Mehling, M.; Tay, S. Curr. Opin. Biotechnol. 2014, 25, 95102. (17) Haque, A.; Gheibi, P.; Gao, Y.; Foster, E.; Son, K. J.; You, J.; Stybayeva, G.; Patel, D.; Revzin, A. Sci. Rep. 2016, 6, 33980. (18) Kaestli, A. J.; Junkin, M.; Tay, S. Lab Chip 2017, 17, 41244133. (19) Maoz, B. M.; Herland, A.; Henry, O. Y. F.; Leineweber, W. D.; Yadid, M.; Doyle, J.; Mannix, R.; Kujala, V. J.; FitzGerald, E. A.; Parker, K. K.; Ingber, D. E. Lab Chip 2017, 17, 2294-2302. (20) Cui, X.; Liu, Y.; Hu, D.; Qian, W.; Tin, C.; Sun, D.; Chen, W.; Lam, R. H. W. Lab Chip 2018, 18, 522-531. (21) Lu, Y.; Chen, J. J.; Mu, L.; Xue, Q.; Wu, Y.; Wu, P. H.; Li, J.; Vortmeyer, A. O.; Miller-Jensen, K.; Wirtz, D.; Fan, R. Anal. Chem. 2013, 85, 2548-2556. (22) Kravchenko-Balasha, N.; Wang, J.; Remacle, F.; Levine, R. D.; Heath, J. R. Proc. Natl. Acad. Sci. U. S. A. 2014, 111, 65216526. (23) Enomoto, J.; Matharu, Z.; Revzin, A. Methods Enzymol. 2013, 526, 107-121.

(24) Matharu, Z.; Enomoto, J.; Revzin, A. Anal. Chem. 2013, 85, 932-939. (25) Shin, D. S.; Liu, Y.; Gao, Y.; Kwa, T.; Matharu, Z.; Revzin, A. Anal. Chem. 2013, 85, 220-227. (26) Liu, Y.; Kwa, T.; Revzin, A. Biomaterials 2012, 33, 73477355. (27) Zhou, Q.; Patel, D.; Kwa, T.; Haque, A.; Matharu, Z.; Stybayeva, G.; Gao, Y.; Diehl, A. M.; Revzin, A. Lab Chip 2015, 15, 4467-4478. (28) Lu, S.; Kennedy, R. T. Methods Mol Biol 2017, 1547, 57-67. (29) Lu, S.; Dugan, C. E.; Kennedy, R. T. Anal. Chem. 2018, 90, 5171-5178. (30) Riahi, R.; Shaegh, S. A.; Ghaderi, M.; Zhang, Y. S.; Shin, S. R.; Aleman, J.; Massa, S.; Kim, D.; Dokmeci, M. R.; Khademhosseini, A. Sci. Rep. 2016, 6, 24598. (31) Son, K. J.; Gheibi, P.; Stybayeva, G.; Rahimian, A.; Revzin, A. Microsyst. Nanoeng. 2017, 3, 17025. (32) Marcy, Y.; Ishoey, T.; Lasken, R. S.; Stockwell, T. B.; Walenz, B. P.; Halpern, A. L.; Beeson, K. Y.; Goldberg, S. M.; Quake, S. R. PLoS Genet. 2007, 3, 1702-1708. (33) Chen, Y.; Zhang, B.; Feng, H.; Shu, W.; Chen, G. Y.; Zhong, J. F. Lab Chip 2012, 12, 3930-3935. (34) Jimenez-Valdes, R. J.; Rodriguez-Moncayo, R.; CedilloAlcantar, D. F.; Garcia-Cordero, J. L. Anal. Chem. 2017, 89, 5210-5220. (35) Gonzalez-Suarez, A. M.; Pena-Del Castillo, J. G.; Hernandez-Cruz, A.; Garcia-Cordero, J. L. Anal. Chem. 2018, 90, 8331-8336. (36) Dunn J. C.; Yarmush M. L.; Koebe H. G.; Tompkins R.G. FASEB Journal 1989, 3, 174-177. (37) Duk Han, Y.; Jin Chun, H.; Yoon, H. C. Biosens. Bioelectron. 2014, 59, 259-268. (38) Park C., H. Y. D., Kim H. V., Lee J., Yoon H. C., and Park S. Lab Chip 2018, 18, 1533-1538. (39) Jenie, S. N.; Prieto-Simon, B.; Voelcker, N. H. Biosens. Bioelectron. 2015, 74, 637-643. (40) Li, Z.; Cheng, H.; Shao, S.; Lu, X.; Mo, L.; Tsang, J.; Zeng, P.; Guo, Z.; Wang, S.; Nathanson, D. A.; Heath, J. R.; Wei, W.; Xue, M. Angew. Chem. Int. Ed. Engl. 2018, 57, 11554-11558. (41) Hofmann, A. F.; Hagey, L. R. Cell. Mol. Life Sci. 2008, 65, 2461-2483. (42) Chiang, J. Y. J. Lipid Res. 2009, 50, 1955-1966. (43) Dawson, P. A.; Lan, T.; Rao, A. J. Lipid Res. 2009, 50, 23402357. (44) Barnes, S.; Spenney, J. G. Clin. Chim. Acta 1980, 102, 241245. (45) Starkey, B. J.; Marks, V. Clin. Chim. Acta 1982, 119, 165177. (46) Li X.; Hu. J.; Easley C. J. Lab Chip 2018, 18, 2926-2935. (47) Wei, Y.; Wang, D.; Topczewski, F.; Pagliassotti, M. J. Am. J. Physiol. Endocrinol. Metab. 2006, 291, E275-281. (48) Nakamura, S.; Takamura, T.; Matsuzawa-Nagata, N.; Takayama, H.; Misu, H.; Noda, H.; Nabemoto, S.; Kurita, S.; Ota, T.; Ando, H.; Miyamoto, K.; Kaneko, S. J. Biol. Chem. 2009, 284, 14809-14818. (49) Pfaffenbach, K. T.; Gentile, C. L.; Nivala, A. M.; Wang, D.; Wei, Y.; Pagliassotti, M. J. Am. J. Physiol. Endocrinol. Metab. 2010, 298, E1027-1035. (50) Malhi, H.; Kaufman, R. J. J. Hepatol. 2011, 54, 795-809.

ACS Paragon Plus Environment

Page 8 of 9

Page 9 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

For TOC only

ACS Paragon Plus Environment