Sage Journals: Discover world-class research

Abstract

Phenotypic cell-based high-throughput screenings play a central role in drug discovery and toxicology. The main tendency in cell screenings is the increase of the throughput and decrease of reaction volume in order to accelerate the experiments, reduce the costs, and enable screenings of rare cells. Conventionally, cell-based assays are performed in microtiter plates, which exist in 96- to 1536-wells formats and cannot be further miniaturized. In addition, performing screenings of suspension cells is associated with risk of losing cell content during the staining procedures and incompatibility with high-content microscopy. Here, we evaluate the Droplet-Microarray screening platform for culturing, screening, and imaging of suspension cells. We demonstrate pipetting-free cell seeding and proliferation of cells in individual droplets of 3–80 nL in volume. We developed a methodology to perform parallel treatment, staining, and fixation of suspension cells in individual droplets. Automated imaging of live suspension cells directly in the droplets combined with algorithms for pattern recognition for image analysis is demonstrated. We evaluated the developed methodology by performing a dose–response study with antineoplastic drugs. We believe that the DMA screening platform carries great potential to be adopted for broad spectrum of screenings of suspension cells.

Keywords

Droplet-Microarray high-throughput screenings suspension cells hydrophobicity

Introduction

High-throughput phenotypic screenings of live cells play a pivotal role in the fields of drug discovery and toxicology. Across all therapeutic areas, 32% of drugs approved between 1999 and 2008 were discovered using phenotype-based approaches.¹ Phenotypic screenings are also used as an in vitro model for toxicity tests and for studies of drug-induced liver injury (DILI).^2,3 The constantly growing need for new drugs and bioactive compounds, the creation of large chemical libraries, and the development of sophisticated biological assays have resulted in a constant growth of market for cell-based assays, which was already estimated to be about US$11 billion in 2015.⁴

Conventionally, cell-based assays are performed in microtiter plates, ranging from the 96-to 1536-well format. The tendency for increasing the throughput and decreasing the reaction volumes arises from the purpose of reducing the costs of experiments and enabling the screening of rare cells. The majority of spending, up to 75%, on high-throughput screenings (HTSs) goes toward reagents, microtiter plates, and assay kits.⁵ Decreasing the volume per experiment combined with the increase in throughput is a way to reduce the total costs of the HTS experiments. On the other hand, more researchers are switching to human primary and stem cells for their cell-based screening. Such cells are often available in limited amounts and cannot be easily expanded, which makes it difficult to perform large screens. Miniaturization can solve this problem by decreasing the amount of cells per experiment, allowing for more compounds to be screened on rare cell types. Miniaturization of microtiter plates has reached its physical limit as further reduction of volumes of wells is difficult due to capillary effects.⁶ At the same time, there is a clear need for further miniaturization in big pharmaceutical companies.⁷

Miniaturization of assay format by means of alternative technologies has been a trend for some years now. Such technologies are represented by cell microarrays introduced by Sabatini and colleagues and microfluidic platforms of different kinds adopted for cell-based applications.^8–12 Cell microarray technology, which is based on preprinting transfection mixtures onto a glass substrate prior to seeding a layer of adherent cells onto it, was widely used by multiple academic groups to perform phenotypic screens.^13–16 Cell-based assays in a microfluidic format, where liquids are confined in small channels and cavities, can benefit from low sample consumption, high-throughput, precise and rapid addition of defined solutions, and reduced time of experiment and analysis.¹⁷ However, along with mentioned advantages, these technologies have limitations. Cell microarrays are (1) limited to adherent cell cultures and not compatible with cells of a suspension nature, (2) designed for transfection assays and barely compatible with compound screenings, and (3) suffering from the risk of cross-contamination between the spots. Cell assays in microfluidic formats are proved to be advantageous compared with conventional method in such areas as single-cell sequencing, metabolism studies, ion channel research, and chemotaxis.¹⁷ At the same time, microfluidics does not offer optimal solutions in multiple areas, including toxicity testing or high-content screenings. Therefore, such technologies are still restricted in their applications and cannot substitute microtiter plates for high-throughput cell screens in big pharmaceutical companies.

The screening of suspension cells as a rule is more challenging than the screening of cells of an adherent nature. This is first due to potential loss of cells during washing, staining, and fixation procedures. Second, suspension cells are small and tend to clump while growing, which makes it difficult or impossible to perform microscopy and image analysis.¹⁸ In addition, suspension cells are not growing in one focal plane, being rather distributed throughout the whole volume. In order to perform microscopy of suspension cells, they are usually fixed first, making them not compatible with live imaging. Alternative miniaturized screening platforms are not always compatible with the screening of suspension cells due to the same reasons. For example, it is challenging to create cell microarrays with suspension cells, because cells would have to be immobilized on the surface using anchor molecules, which might affect the cell behavior and transfection efficacy.¹⁹

Recently, we developed a Droplet-Microarray (DMA) platform for cell screenings and have been continuously optimizing different aspects of this technology.^20–25 The DMA technology is based on superhydrophilic (SL)-superhydrophobic (SH) patterning. The effect of discontinuous dewetting enables formation of arrays of droplets trapped in SL areas. Biocompatibility of the polymer surface allowed us to adopt such arrays for cell culturing and create arrays of droplets containing cells. The unique feature of the DMA platform is the ability to form homogeneous droplet arrays spontaneously without a need for pipetting of each droplet. We demonstrated the compatibility of the DMA technology with culturing of different cell types in volumes ranging from 3 to 80 nL.²¹ We established different methods to deliver compounds and reagents into droplets containing cells and demonstrated the applicability of such methods for drug treatment and transfection of adherent cells. In the current study, we further exploited the potential of the DMA platform and developed protocols for the culturing and screening of suspension cells.

In this study, we developed a simple methodology to perform staining and fixation of suspension cells in individual miniaturized droplets in a parallel pipetting-free manner without losing the content of the droplets. We designed and manufactured a handheld device, which enables easy manipulations with cells. We also developed a semiautomated workflow for high-throughput compound screening of cells on the DMA platform based on microscopy of live cells and demonstrated its applicability by testing anticancer compounds. We believe that due to its unique features, the DMA screening platform in combination with methodology developed in this work, carries the potential to be widely adapted by other researchers for phenotypic screenings of suspension cells.

Materials and Methods

Preparation of DMA Slides

Preparation of SL-SH patterned surface of DMA slides was described in detail elsewhere.^21,26 Briefly, patterns were prepared as follows. Glass slides (Schott Nexterion, Jena, Germany) were first activated with 1 M NaOH (Carl Roth GmbH + Co. KG, Karsruhe, Germany) for 1 h, followed by neutralization with 1 M HCl (Merck KGaA, Darmstadt, Germany) for 30 min. Afterward, activated glass slides were modified with 20% v/v solution of 3-(trimethoxysilyl)propyl methacrylate (Sigma-Aldrich Chemie, Munich, Germany) in ethanol for 30 min at room temperature. A polymer layer was formed by applying 25 μL of polymerization mixture (24 wt% 2-hydroxyethyl methacrylate [HEMA] [Sigma-Aldrich Chemie], 16 wt% ethylene dimethacrylate [EDMA] [Sigma-Aldrich Chemie], 12 wt% 1-decanol, 48 wt% cyclohexanol [Sigma-Aldrich Chemie], and 0.4 wt% 2.2-dimethoxy-2-phenylacetophenone [Sigma-Aldrich]) onto a fluorinated glass slide, covering it with a modified glass slide and cross-linking the polymer by ultraviolet (UV) irradiation with 12 mW/cm² intensity and 260 nm wavelength for 15 min. Glass slides were fluorinated in trichloro(1H,1H,2H,2H-perfluorooctyl)silane (Sigma-Aldrich Chemie) vapor in desiccators under 50 mbar vacuum overnight. The polymer surface was modified with alkyne groups by incubating the slides in solution containing 45 mL of dichloromethane (Merck KGaA), 56 mg of 4-(dimethylamino)pyridine (Novabiochem, Merck KGaA), 111.6 mg of pentynoic acid (Sigma-Aldrich Chemie), and 180 µL of N,N′-diisopropylcarbodiimine (Alfa Aesar GmbH & Co KG, Karlsruhe, Germany) for 4 h under stirring at room temperature. The SH background was created by applying 5% v/v solution of 1H,1H,2H,2H-perfluorodecanethiol (Sigma-Aldrich Chemie) in acetone onto the polymer surface and irradiating the slide through a photomask (Rose fotomasken, Bergisch Gladbach, Germany) with 260 nm of UV light at 12 mW/cm² for 1 min. Afterward, SL spots were created by applying 10% v/v β-mercaptoethanol (Alfa Aesar) solution in 1:1 water–ethanol onto the patterned surface and irradiating the slide with 260 nm of UV light at 12 mW/cm² for 1 min.

Cell Culture

Jurkat human T-cell lymphocyte cells were cultured in RPMI-1640 medium supplemented with 10% heat-inactivated fetal bovine serum (FBS, Sigma-Aldrich) and 1% penicillin/streptomycin (Gibco, Life Technologies GmbH, Darmstadt, Germany). Cells were cultured in T25 flasks and diluted every 2–3 days until a cell density of 2 × 10⁵ cells/mL was achieved.

DMA slides for culturing were sterilized in ethanol and preconditioned by immersing in RPMI-1640 medium supplemented with 1% heat-inactivated FBS and 1% penicillin/streptomycin for 45 min in a cell culture incubator. Afterward, DMA slides were dried for 45 min under a clean bench. Cells were seeded onto a preconditioned DMA slide using a well-established protocol described in a previous work.²¹ Briefly, 1.7 mL of cell suspension was applied per field surrounded by a SH border ( Fig. 1a , Suppl. Video S1) for 90 s, followed by slight tilting of the DMA slide to allow spontaneous droplet formation²¹ (Suppl. Video S1). The seeding concentrations for Jurkat cells used to obtain different final concentrations in droplets on patterns with spots of three different sizes are summarized in Table 1 . In order to avoid evaporation, certain precautions were taken, as described elsewhere.²¹ Briefly, tilting of the slide and droplet formation was done in a closed Petri dish, which was immediately placed in a bigger Petri dish containing 10 mL of phosphate-buffered saline (PBS) and paper tissues placed along the perimeter of the dish.

Figure 1.

Culturing and microscopic analysis of suspension cells on the DMA platform. (a) Schematic representation of the DMA platform. (b) Microscope images of DMA slides with spots measured at 1000, 500, and 350 μm and containing Jurkat cells stained with CalceinAM. (c) Graphs showing the distribution of average fluorescent intensities of spots containing calcein-stained Jurkat cells. Distribution is shown for DMA slides with three different spot sizes.

Table 1.

Seeding Cell Concentrations and Corresponding Final Cell Concentrations of Jurkat Cells on DMA Platform.

1000 μm		500 μm		350 μm
Seeding Cell Concentration	Final Cell Concentration	Seeding Cell Concentration	Final Cell Concentration	Seeding Cell Concentration	Final Cell Concentration
1 × 10⁵ cell/mL	0.7 × 10⁶ cell/mL	5 × 10⁵ cell/mL	5 × 10⁶ cell/mL	5 × 10⁵ cell/mL	5 × 10⁶ cell/mL
2.5 × 10⁵ cell/mL	2.3 × 10⁶ cell/mL	10 × 10⁵ cell/mL	16 × 10⁶ cell/mL	10 × 10⁵ cell/mL	20 × 10⁶ cell/mL
3.5 × 10⁵ cell/mL	2.7 × 10⁶ cell/mL	15 × 10⁵ cell/mL	19 × 10⁶ cell/mL	15 × 10⁵ cell/mL	27 × 10⁶ cell/mL

Cell Growth and Viability Study

Jurkat cells were seeded onto DMA slides with spots of different sizes and the amount of cells was estimated by monitoring 10–20 spots per slide at 0, 24, 48, 72, and 96 hours after seeding. Images of each spot were taken at 10× resolution using a Keyence BZ-9000 microscope (KEYENCE, Osaka, Japan). The amount of cells was estimated by manual counting. The viability of cells was estimated at a final time point at 96 h by live/dead staining (see Live/Dead Staining of Cells section). The data from three independent experiments were included in the analysis.

Aligning Frame for Sandwiching

An aligning frame was designed and manufactured for the aligning and sandwiching of DMA slides. The following parameters were taken into consideration: (1) secure holding of the slides; (2) precise (nonadjustable) alignment in the XY plane of both slides, with respect to each other; (3) a precisely adjustable distance of 0–2 mm between the slides in the z axis (sandwiching); and (4) the distance between the bottom surface of the frame and the lower DMA slide does not exceed 1 mm for direct microscopy in the frame. The holder consists of two separate frame pieces, an upper and lower holder. The slides are securely held in place via spring-loaded brackets. Both frames are precisely aligned in the XY plane, with respect to each other. This is achieved by four stainless steel pins, located in the four corners of the lower holder. These pins are aligned with four mirroring holes on the upper frame. Four holes contain polytetrafluoroethylene (PTFE) bushings (GGB Heilbronn GmbH, Heilbronn, Germany). The precise and adjustable distance between the slides is achieved by four micrometer screws (Lehmann Präzision GmbH, Hardt, Germany). These are placed symmetrically in the upper holder while being supported by the lower holder. The distance between DMA slides is adjusted by rotation of the screws. The holders and brackets were machined out of aluminum (Maschinenbau Kaltenbach GmbH, Crailsheim, Germany) and black anodized (Grau & Wagenblast, Abtsgmünd, Germany).

Sandwiching Method

The sandwiching method was used for all manipulations with individual droplets on DMA, including staining and fixation of cells and compound addition. The DMA slide containing cells was fixed in the lower holder of the aligner. For staining and fixation, an empty DMA slide was fixed in the upper holder of the aligner and fixation or staining solution was spread onto a DMA slide using an automatic pipette (Suppl. Video S2). To initiate a screen, a slide containing drugs was fixed in the upper holder of the aligner. Afterward, the aligner was closed with four screws controlling the distance between the slides (see Fig. 3b , red arrows). In order to bring slides together, four screws were turned slowly counterclockwise until the droplets came into contact, which could be visually monitored. The DMA slide was incubated in the sandwiched position for 15 min in a cell culture incubator or on a bench. To avoid evaporation during the sandwiching step, the frame was covered with sterile wetted paper tissues. To open the frame, four screws controlling the distance between the slides (see Fig. 3b , red arrows) were turned clockwise until droplets were not in contact anymore. Afterward, the frame was open and the DMA slide containing cells was used according to requirements of a particular protocol.

Fixation of Cells

The fixation solution consisted of 7.4% formaldehyde and 50 µg/mL DAPI in Dulbecco’s PBS. Fixation of cells was performed using the sandwiching approach as described above for 15 min on a bench. Afterward, the aligning frame was opened and the DMA slide containing cells was dried on air and then mounted with Immu-Mount (Thermo Fisher Scientific Inc., Darmstadt, Germany) and dried at +4 °C. Images of each individual droplet were taken using a Keyence BZ-9000 microscope (KEYENCE).

Live/Dead Staining of Cells

The staining solution consisted of 50 µg/mL propidium iodide (PI) (Invitrogene, Merelbeke, Belgium) and 50 µg/mL CalceinAM (Thermo Scientific) in PBS containing CaCl₂ and MgCl₂. Live/dead staining was performed using the sandwiching approach as described above for 15 min in a cell culture incubator. Afterward, the aligner was opened and the DMA slide containing stained cells was placed into a four-well dish (Thermo Scientific Nunc) and immediately subjected to microscopic analysis. Depending on the procedure, images were taken using either a Keyence BZ-9000 microscope (KEYENCE) or automated screening Olympus IX81 microscope (Olympus, Tokyo, Japan) (for details, see Automated Imaging and Image Analysis section).

Screening Initiation

Printing of chemical compounds onto glass slides was performed as follows. Clean microscope glass slides (Schott Nexterion) were fluorinated by trichloro(1H,1H,2H,2H-perfluorooctyl)silane (Sigma-Aldrich) under 50 mbar vacuum in a sealed desiccator containing an open vial of the silane overnight. Afterward, different amounts of antineoplastic drugs (in DMSO or diluted in pure RPMI-1640 medium if necessary) were printed on fluorinated slides in a defined pattern geometry corresponding to the geometry of DMA slides using the sciFLEXARRAYER S11 liquid dispenser (Scienion, Berlin, Germany). The following antineoplastic drugs were used: doxorubicin (European Pharmacopolia Reference Standards, European Directorate for the Quality of Medicines & Healthcare, Strasbourg, France), daunorubicin, cytarabine, dexamethasone, vincristine, and etoposide (Sigma-Aldrich). Prior to use, slides containing drugs were dried in a sealed desiccator containing silica gel (Sigma-Aldrich) for 24 h in darkness.

Jurkat cells were seeded onto a DMA slide as described in the Cell Culture section 3 h prior to screening initiation. Screening was initiated by sandwiching the DMA slide containing cells with the slide containing preprinted drugs as described in the Sandwiching Method section for 15 min in cell culture incubator. Afterward, the frame was open and the DMA slide containing cells was returned to the cell culture incubator.

Automated Imaging and Image Analysis

The DMA slide containing cells was placed into a four-well dish (Thermo Scientific Nunc), which was fixed in a standard holder for microtiter plates for microscopy. Imaging was performed using the automated screening microscope Olympus IX81 (Olympus). The grid of the DMA pattern was defined in the microscope software. The autofocus function “interpolate AF” was used to minimize defects that can arise from uneven surface thickness and unparalleled slide position. Z stacks of 10 slices were made to count for suspension cells being on slightly different focal planes. The images of one central field of view per each SL spot were taken at 10× magnification, in three different channels (brighfield, mCherry [for PI staining], and green fluorescent protein [GFP, for CalceinAM staining]). Imaging of 100 SL spots took about 20 min and resulted in 7.67 GB of data in total.

For analysis, images were processed using the “Extended Depth” algorithm in ImageJ software. To estimate the number of calcein- and PI-stained cells, a special algorithm for pattern recognition was developed. Obtained images were 12-bit greyscale images with a dynamic range of 4096 brightness values. In each image, the minimum brightness was calculated and brightness exceeding a fourth of the dynamic range was defined as the signal threshold. Each GFP and PI image was processed by applying normalization, shading correction, Laplacian of Gaussian filtering, noise cancellation, and application of a watershed transform, resulting in images showing detected objects. All found objects were counted if the maximum brightness value overcame the above-mentioned threshold. Each counted object was assigned the features area and mean intensity. GFP-positive cells tend to form clusters where individual cells cannot be identified automatically. To account for this, we assumed that a regular cell consists of 200 pixels and divided the area of a cluster by this number, getting an estimation of the number of cells within the cluster. To account for dead cells that exhibit a low GFP signal, every object found to be positive in the GFP channel was double-checked in the PI channel, and excluded from the count if found to be PI positive. All found objects were counted in both channels, delivering an estimation of objects in both GFP and PI channels.

Statistics and Data Analysis

Every experiment was repeated at least three times, and the data from three independent experiments were included in the analysis. Compounds in each concentration and appropriate controls were spotted in 10 repeats. The dose–response curves of tested drugs were plotted in GraphPad Prism 7. The sigmoidal curves were created based on a nonlinear regression technique. IC50 values of the compounds were calculated in the program.

The Z′ factor was calculated as follows: $Z' = 1 - (3 σ p + 3 σ n) / (µ p - µ n)$ , where σp is the standard deviation of the positive control, σn is the standard deviation of the negative control, μp is the mean of the positive control, and μn is the mean of the negative control. Here, positive control is a drug in a 1000 µM concentration and negative control corresponds to DMSO control.

Results

Suspension Cells on Droplet-Microarray: Culturing and Manipulation

We chose a Jurkat T-cell leukemia suspension cell line as a model for validating all the procedures that can be performed using the DMA platform. The DMA platform was introduced in a previous work.^21,22 It consists of an array of SL spots on an SH background ( Fig. 1a ). Due to the extreme difference in wettability between SL and SH areas, arrays of droplets containing cells can be created by spreading the cell suspension on the surface of a DMA slide without the need to pipette each individual spot ( Fig. 1a , b , Suppl. Video S1). In this case, droplets are formed spontaneously due to the effect of discontinuous dewetting (Suppl. Video S1). In the current study, we used DMA slides with dimensions of a standard microscope glass slide, containing 588, 2187, and 4563 square SL spots measuring 1 mm, 500 µm, and 350 µm, respectively, which corresponds to approximately 6000, 25,000, and 50,000 droplets, respectively, for arrays with the size of a standard microtiter plate. In order to see if we can achieve homogeneous cell distribution throughout a DMA slide using the effect of discontinuous dewetting, we seeded Jurkat cells stained with CalceinAM for visualization on DMA slides with SL spots of three different sizes (1000 µm, 500 µm, and 350 µm). The graphs depicted in Figure 1c show the distribution of the mean fluorescence intensity of the droplets, which is proportional to the number of cells per droplet.

As a next step, we monitored viability and proliferation of Jurkat cells cultured on a DMA slide for 5 days without medium exchange. As demonstrated in Figure 2b and Supplemental Figure S1a,b, the growth rate of cells on DMA slides strongly depended on an initial cell seeding concentration. The optimal initial cell concentration that gives the best growth rate was used for further experiments. Jurkat cells were multiplied by factors 1.7, 1.4, and 1.2 over a culturing period of 96 h on DMA slides with 1000 µm and 500 or 350 µm spot sizes, respectively ( Fig. 2c ). We compared the proliferation rate of Jurkat cells on the DMA slides with that of cells in a 24-well plate, where the cells were seeded with the same initial cell density as on DMA slides with a 1000 µm spot size. The cells in the 24-well plate proliferated only until 48 h, whereas in the 1000 μm spots, they continuously proliferated up to 5 days of culturing ( Fig. 2c ). The slow growth of cells in the 24-well plate can be explained by very high cell densities, 3 × 10⁶ cells/mL after 48 h of culturing. It is known that Jurkat cells have very slow proliferation rate at this high density. Interestingly, we observed continuous cell proliferation in miniaturized droplets with the same cell densities ( Fig. 2c ). The viability of Jurkat cells cultured in the droplets of all three sizes was comparable to that of cells cultured by conventional methods ( Fig. 2a ). Even after 5 days of culturing without medium exchange, the viability of cells in droplets was 90% and 75% on the spots of 1000 μm and 500 or 350 μm, respectively. Based on these results, we can conclude that the DMA platform is suitable for screening applications, such as, for example, anticancer compound screens. Acute effects of antineoplastic drugs are usually estimated within 2–3 days in cell-based assays, and since most antineoplastic drugs influence rapidly proliferating cells, continuous proliferation of cells is critical in this case.

Figure 2.

Viability and growth rate of Jurkat cells cultured on DMA slides with different spot sizes. (a) Microscope images of Jurkat cells stained with calcein (live cells shown in green) and PI (dead cells shown in red) cultured on DMA slides with three different spot sizes for 24 h. (b) Graphs showing growth rates of Jurkat on DMA slides with 1000 µm spot size seeded in different initial cell densities. (c) Graph showing comparison of growth rates of Jurkat cells cultured on DMA slides with different spot sizes and in a 24-well plate.

Screening of suspension cells is challenging due to the risk of losing the cell content during washing and staining procedures. Therefore, in order to be able to use the DMA platform for screening applications, it is necessary to develop protocols enabling the addressing of each individual spot without disturbing the content of the droplets. Therefore, we validated the sandwiching method, introduced in a previous work,²² for performing parallel staining and fixation of suspension cells. The principle of the sandwiching method for staining and fixation of cells on DMA is schematically presented in Figure 3a . The staining or fixation solution was spread onto a DMA slide with a pipette taking advantage of the effect of discontinuous dewetting (Suppl. Video S2). Afterward, the dye solution was added to cells by aligning and sandwiching the DMA slide containing the dye with the DMA slide containing cells, which allows for precise addition of the solution of interest into each individual droplet containing cells preserving the content of the droplets and excluding any cross-interaction between the spots. After 15 min, the sandwich was opened and cells were subjected to microscopic analysis. To enable easy manipulations with cells, we designed and manufactured a handheld device ( Fig. 3b ). It consists of a lower holder, where a DMA slide containing cells is fixed, and an upper holder, where a DMA slide with staining or fixation solution is fixed. The design of the frame ensures perfect alignment of two slides and four screws on the top ( Fig. 3b , red arrows) control the distance between the slides, avoiding merging of the droplets during the sandwiching procedure. In order to enable microscopic analysis of cells on a DMA slide directly in the frame, we designed it with outer dimensions of standard microtiter plates and with a distance from a lower DMA slide being less than 1 mm. These features make the frame compatible with any kind of manual and automated microscopes. Using the sandwiching approach and designed frame, we were able to perform live/dead staining, as well as fixation of suspension cells on a DMA slide ( Fig. 3c , d ). There were no or very few cells observed on the slide containing the solution after opening the sandwich, and quantification of cells before and after the procedure across random droplets showed that 95% and 97% of cells in the cases of fixation and live staining, respectively, were still present in the spots ( Fig. 3c , d ). Thus, we demonstrated that the sandwiching approach and designed aligning frame can be used for staining and fixation of suspension cells directly on DMA slides without losing the content of the droplets.

Figure 3.

HTS of suspension cells using the DMA aligning frame. (a) Scheme showing the principle of staining and fixation procedures using the sandwich approach. (b) Photographs of the DMA aligner frame. (c,d) Microscope images of Jurkat cells on 1000 µm spots before and after live/dead staining (c) and fixation (d).

High-Throughput Screening Using the DMA Platform

To be able to use the DMA platform for HTS applications, it is necessary to be able to perform experiments in a parallel manner. For this purpose, we developed a workflow for high-throughput drug screening of suspension cells on the DMA platform depicted in Figure 4 . For a proof-of-concept study, we used DMA slides containing 588 spots of 1000 µm size ( Figs. 1a and 4(2) ). In this workflow, drugs in different concentrations, including appropriate controls, were printed onto a hydrophobic glass slide in geometry corresponding to the geometry of a DMA slide and dried ( Fig. 4(1) ). Jurkat cells were seeded onto a DMA slide using a protocol established in previous work²¹ (see Materials and Methods). This protocol allows for parallel pipetting-free seeding of cells on the full DMA slide using the effect of discontinuous dewetting ( Fig. 4(2) , Suppl. Video S1). The screening was initiated in parallel by sandwiching the DMA slide containing cells with the slide containing the preprinted compounds using the aligning frame ( Fig. 4(3) ). The sandwiched frame was placed into a cell culture incubator for 15 min to allow for the drugs to dissolve in the droplets. Afterward, the frame was opened ( Fig. 4(4) ) and the DMA slide containing cells was returned to the cell culture incubator ( Fig. 4(5) ). The effect of the compound was estimated 72 h after exposure by performing live/dead staining using the sandwiching approach ( Fig. 4(6) ). Afterward, the slide containing cells was placed in a four-well plate to avoid evaporation and imaged using an automated screening microscope Olympus IX81 ( Fig. 4(7) ). The grid of DMA was defined in the microscope software, and images of each spot at 10× magnification, in three different channels (brightfield, mCherry [for PI staining], and GFP [for CalceinAM staining]), were taken automatically. To account for slightly different focal positions of cells, a Z stack of 10 slices was taken and images were processed using the “Extended Depth” algorithm in ImageJ software. To estimate the viability of cells in the spots, we calculated the number of live (CalceinAM-stained) and dead (PI-stained) cells in the droplets. For this purpose, we developed a pattern recognition algorithm, which enabled automated image processing (see Materials and Methods and Fig.4(8) ).

Figure 4.

The workflow of HTS of suspension cells using the DMA platform.

Drug Treatment: Proof-of-Concept Study

To validate the developed protocol, we performed a proof-of-concept study where we tested Jurkat cells against antineoplastic drugs commonly used for the treatment of leukemia patients in the clinic and known to be effective against Jurkat cells ( Fig. 5 , Fig. S2). Using the workflow described above, we introduced Jurkat cells to different concentrations of the drugs. Cells were seeded onto DMA slides as described in Materials and Methods and placed in a cell culture incubator for 3 h before the screening initiation. Afterward, cells were introduced to 10 different concentrations of drugs and appropriate controls with 10 repeats for each concentration. Cells were incubated for 72 h, and the effect of drugs was estimated based on the viability of cells in treated spots compared with the control of untreated spots. Simultaneous live/dead staining of the cells, automated imaging, and image analysis were performed as described in the screening workflow. The obtained dose–response curves are demonstrated in Figure 5 and Supplemental Figure S2. A full response of Jurkat cells to doxorubicin and daunorubicin, with cell viability going down to zero by treatment with high concentrations of the drugs, was observed. The IC50 values of doxorubicin and daunorubicin obtained on the DMA platform were 24.4 and 136.8 μM, respectively. To evaluate the quality of the developed assay, we calculated the Z′ factor, which was 0.886917 and 0.899118 in the cases of doxorubicin and daunorubicin treatments, respectively. If the value of Z′ factor is between 0.5 and 1, the separation between the values of the positive and negative control is large and the assay is considered to be suitable for analysis.²⁷ Thus, it demonstrates the applicability of the DMA platform in combination with the developed workflow for screening assays.

Figure 5.

Drug treatment: Proof-of-concept study. Dose–response curves of doxorubicin (a) and daunorubicin (b).

Discussion

In the current study, we presented a system for HTS of suspension cells based on the DMA technology. DMA slides used in this work enable miniaturized screenings of live cells in droplets ranging from 3 to 80 nL in densities from 588 to 4563 spots per standard microscope glass slide, which corresponds to approximately 6000 and 50,000 spots per area of a standard microtiter plate, respectively. We utilized a previously established seeding method based on pipetting-free spreading of the cells on the whole DMA slide, taking advantage of the effect of discontinuous dewetting.^21,22 Homogeneous distribution of suspension cells between the droplets of different sizes was demonstrated. The single-step pipetting-free seeding results in savings in the pipetting steps, pipetting tips, robotics, and time of the experiment.

We demonstrated that Jurkat cells continuously proliferate in individual droplets of 80 nL volume for at least 5 days without medium exchange. This makes the DMA platform compatible with assays performed within several days, which are the majority of primary screens, including such screenings as small-compound screenings and genome-wide siRNA screens.^28–31 Cell proliferation itself is a critical parameter in many assays, for example, in testing of antineoplastic drugs that affect rapidly proliferating cells and might not be effective on slowly or nondividing cells.^32,33

We developed a simple methodology based on a sandwiching approach for parallel addressability of individual droplets containing cells. The steps of compound addition, staining, and fixation of cells are performed in a one-step simultaneous procedure on the whole DMA slide. Such parallelization could be beneficial in the case of very short incubation times or rapid cell responses, where seconds or minutes are critical for the accurate readout. Since the sandwiching approach does not require multiple pipetting steps, the processing time for hundreds of “microwells” reduces to a few minutes of hands-on time. Finally, the DMA platform, in combination with the sandwiching method, can potentially enable affordable HTSs without a need for any additional equipment, including liquid handling robotics. The DMA technology can also be used in combination with direct dispensing of compounds, reagents, and even cells into individual droplets using an appropriate liquid dispenser. In this case, the volume of medium and number of cells in the droplets could be better controlled, which is beneficial for some applications. However, in order to be able to perform such dispensing in a fast and parallel manner compatible with high throughput, we would need new technical solutions.

We established a workflow for automated imaging of live cells directly in the droplets. Due to small volumes, suspension cells tend to locate close to the surface rather than being distributed throughout the whole droplet. We accounted for the small differences in cell localization by acquiring a Z stack of 10 slices and processing images using the “Extended Depth” algorithm. This gave us good-quality images for pattern recognition analysis. We also established the procedure for fixation of suspension cells in droplets, which can be followed by microscopic analysis directly on the DMA platform. Microscopy of fixed or live suspension cells is challenging, because cells are not fixed to the surface and constantly moving, making it difficult to acquire high-quality images. For this purpose, cells of a suspension nature have to be attached to the surface by using, for example, poly-l-lysine-coated glass. This technique does not work for all cell types, and culturing suspension cells attached to the surface might alter their behavior and properties.^34,35 Therefore, most phenotypic HTSs based on microscopic readouts, especially high-content screenings, are restricted to the cells of an adherent nature. However, a big portion of important cell types, including immune cells, are of the suspension nature, and imaging is the only method that can be used for investigation of certain processes, such as, for example, cell–cell interaction during immune response. In general, high-content screenings offer high throughput and multiparameter complex readouts where even a small amount of cells is enough to collect sufficient information about the phenotype.³⁶ Therefore, a system that allows for culturing, treatment, staining, fixation, and microscopic analysis of suspension cells in one spot directly on the platform is highly beneficial for the screening community.

We developed an algorithm for automated counting of stained cells allowing for estimation of cell viability in a particular droplet. Estimation of viability using the microscopic approach might be more accurate than using bulk colorimetric or luminescence assays. Colorimetric methods such as MTT and MTS-based assays have been documented to exhibit cytotoxicity and give wrong viability estimation due to various nonspecific reactions.^37,38 Luminescence ATP assays are more accurate but give information only about the amount of live cells without distinguishing between different cell death types. In contrast, microscopic analysis offers the possibility to extract multiple morphological features, giving more information about the drug effect on the cells.

To test the established workflow, we performed a dose-dependent study of cytotoxicity of antineoplastic drugs on Jurkat cells. The IC50 values of doxorubicin and daunorubicin known from the literature range from 0.004444 to 29.9 and from 0.02 to 1,^39–42 respectively. Differences in IC50 values obtained on the DMA platform and those known from the literature are common and arise from differences in the experimental setup. The IC50 values obtained from different laboratories are known to vary and strongly depend on such parameters as platform, cell number, time of incubation, and readout assays.⁴³ We observed relatively high variability in cell response to drug concentrations from the middle of the effective dose range. It is accepted that the lower variance in the case of high or low doses of drug compared with middle doses can be observed.⁴⁴ Nevertheless, the Z′ factor calculated based on positive (a compound in concentration of 1000 μM) and negative (DMSO control) controls confirmed the suitability of the developed methodology for screening applications. We also analyzed the effect of vincristine, cytarabine, and etoposide on the viability of Jurkat cells (Suppl. Fig. S2). The dose-dependent curves of all three drugs plateaued at about 50%–60% of viability. A partial effect of etoposide and vincristine on cell viability could possibly be explained by the low solubility of these drugs in water: up to 1 mg/mL for etoposide and up to 0.00227 mg/mL for vincristine, which equals the final concentrations of 140 and 2.5 μM, respectively (PubChem). The solubility of drugs is a common problem in drug discovery and development, as well as in clinical response.⁴⁵ About 40% of compounds that are under development or approved for clinical applications are poorly water soluble.⁴⁵ There are various strategies to increase the solubility of compounds by means of cosolvents, surfactants, or lipid-based systems, which can be applied in the future to promote drug dissolution into the droplets.⁴⁵ In the case of cytarabine, which is highly soluble in water, the low effect might be explained by the lower rate of cell proliferation in nanoliter-sized droplets than in microtiter plates. Even though we demonstrated constant cell proliferation throughout the experiment in droplets, cells tend to proliferate faster in bigger volumes. Thus, there is a possibility that there is a subpopulation of slow or nonproliferating cells that are resistant to this drug, reducing its effect on the viability of the whole cell population. More optimization is required and more compounds need to be tested in order to be able to apply the DMA platform for screening of compounds of a wider range of properties.

Conclusions

Taken together, we have demonstrated the applicability of DMA system for miniaturized HTS of suspension cells. We utilized the DMA platform for culturing of suspension Jurkat cells and demonstrated homogeneous pipetting-free cell seeding and constant proliferation of cells in individual droplets of 3–80 nL volume from 2 to 5 days, depending on droplet size. We developed a simple methodology to perform parallel treatments, staining, and fixation of suspension cells in individual nanodroplets without losing the content of the droplets. A workflow for automated imaging of suspension cells in the droplets and developed algorithms for pattern recognition for image analysis were demonstrated. Finally, we tested the system by introducing Jurkat cells to antineoplastic drugs. We believe that the DMA screening platform carries great potential to be adopted for various miniaturized screening applications on cells of a suspension nature due to its unique properties allowing for single-step parallel manipulations with suspension cells, as well as automated fluorescent microscopy directly on the platform.

Footnotes

Acknowledgements

The authors are grateful to Dr. Christine Blattner (Institute of Toxicology and Genetics, KIT) for providing the Jurkat cell line.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The research was supported by the ERC starting grant (ID 337077-DropCellArray), ERC PoC grant (ID DLV-680913-CellScreenChip), and Helmholtz Association’s Initiative and Networking Fund (grant VH-NG-621).

Supplementary material for this article is available online.

References

Moffat

J. G.

Rudolph

Bailey

Phenotypic Screening in Cancer Drug Discovery—Past, Present and Future. Nat. Rev. Drug Discov. 2014, 13 (8), 588–602.

Soldatow

V. Y.

LeCluyse

E. L.

Griffith

L. G.

. In Vitro Models for Liver Toxicity Testing. Toxicol. Res. 2013, 2 (1), 23–39.

Kia

Sison

R. L. C.

Heslop

. Stem Cell-Derived Hepatocytes as a Predictive Model for Drug-Induced Liver Injury: Are We There Yet? Br. J. Clin. Pharmacol. 2013, 75 (4), 885–896.

Cell-Based Assays Market by Product (Cell Lines, Reagents, Kits, Instruments, Software, Services), Applications (Research, Drug Discovery, ADME Studies, Toxicology), End User (Pharmaceutical Companies, Academics, CRO), & Region—Global Forecast to

2020. Markets and Markets, November 2015.

Comley

New Tools Increase Accessibility to HCS Assays. Drug Discovery World, Summer 2012.

Mayr

L. M.

Fuerst

The Future of High-Throughput Screening. J. Biomol. Screen. 2008, 13 (6), 443–448.

Kumar

C. S. S. R

. Nanotechnology Tools in Pharmaceutical R&D. Mater. Today 2010, 12, (Suppl. 1), 24–30.

Ziauddin

Sabatini

D. M.

Microarrays of Cells Expressing Defined cDNAs. Nature 2001, 411 (6833), 107–110.

Wang

Kim

M.-C.

Marquez

. High-Density Microfluidic Arrays for Cell Cytotoxicity Analysis. Lab Chip 2007, 7 (6), 740–745.

10.

Clausell-Tormos

Lieber

Baret

J. C.

. Droplet-Based Microfluidic Platforms for the Encapsulation and Screening of Mammalian Cells and Multicellular Organisms. Chem. Biol. 2008, 15 (5), 427–437.

11.

Lecault

VanInsberghe

Sekulovic

. High-Throughput Analysis of Single Hematopoietic Stem Cell Proliferation in Microfluidic Cell Culture Arrays. Nat. Methods 2011, 8 (7), 581–586.

12.

Macosko

E. Z.

Basu

Satija

. Highly Parallel Genome-Wide Expression Profiling of Individual Cells Using Nanoliter Droplets. Cell 2015, 161 (5), 1202–1214.

13.

Papp

Szittner

Prechl

Life on a Microarray: Assessing Live Cell Functions in a Microarray Format. Cell. Mol. Life Sci. 2012, 69 (16), 2717–2725.

14.

Neumann

Walter

Heriche

J. K.

. Phenotypic Profiling of the Human Genome by Time-Lapse Microscopy Reveals Cell Division Genes. Nature 2010, 464 (7289), 721–727.

15.

Mannherz

Mertens

Hahn

. Functional Screening for Proapoptotic Genes by Reverse Transfection Cell Array Technology. Genomics 2006, 87 (5), 665–672.

16.

Palmer

E. L.

Miller

A. D.

Freeman

T. C.

Identification and Characterisation of Human Apoptosis Inducing Proteins Using Cell-Based Transfection Microarrays and Expression Analysis. BMC Genomics 2006, 7, 145.

17.

Pihl

Karlsson

Chiu

D. T.

Microfluidic Technologies in Drug Discovery. Drug Discov. Today 2005, 10 (20), 1377–1383.

18.

Götte

Gabriela

Image-Based High-Content Screening. In Drug Discovery, Drug Discovery and Development—Present and Future; Kapetanovic, I. M., Ed.; InTech: Rijeka, Croatia, 2011; chap. 14.

19.

Kato

Umezawa

Miyake

. Transfection Microarray of Nonadherent Cells on an Oleyl Poly(Ethylene Glycol) Ether-Modified Glass Slide. Biotechniques 2004, 37 (3), 444–448.

20.

Geyer

F. L.

Ueda

Liebel

. Superhydrophobic–Superhydrophilic Micropatterning: Towards Genome-on-a-Chip Cell Microarrays. Angew. Chem. Int. Ed. 2011, 50 (36), 8424–8427.

21.

Popova

A. A.

Demir

Hartanto

T. G.

. Droplet-Microarray on Superhydrophobic-Superhydrophilic Patterns for High-Throughput Live Cell Screenings. RSC Adv. 2016, 6 (44), 38263–38276.

22.

Popova

A. A.

Schillo

S. M.

Demir

. Droplet-Array (DA) Sandwich Chip: A Versatile Platform for High-Throughput Cell Screening Based on Superhydrophobic–Superhydrophilic Micropatterning. Adv. Mater. 2015, 27 (35), 5217–5222.

23.

Ueda

Geyer

F. L.

Nedashkivska

. DropletMicroarray: Facile Formation of Arrays of Microdroplets and Hydrogel Micropads for Cell Screening Applications. Lab Chip 2012, 12 (24), 5218–5224.

24.

Ueda

Levkin

P. A.

Emerging Applications of Superhydrophilic-Superhydrophobic Micropatterns. Adv. Mater. 2013, 25 (9), 1234–1247.

25.

Neto

A. I.

Demir

Popova

A. A.

. Fabrication of Hydrogel Particles of Defined Shapes Using Superhydrophobic-Hydrophilic Micropatterns. Adv. Mater. 2016, 28 (35), 7613–7619.

26.

Feng

Ueda

. Surface Patterning via Thiol-Yne Click Chemistry: An Extremely Fast and Versatile Approach to Superhydrophilic-Superhydrophobic Micropatterns. Adv. Mater. Interfaces 2014, 1 (7).

27.

Zhang

J.-H.

Chung

T. D. Y.

Oldenburg

K. R.

A Simple Statistical Parameter for Use in Evaluation and Validation of High Throughput Screening Assays. J. Biomol. Screen. 1999, 4 (2), 67–73.

28.

Paulsen

R. D.

Soni

D. V.

Wollman

. A Genome-Wide siRNA Screen Reveals Diverse Cellular Processes and Pathways That Mediate Genome Stability. Mol. Cell 2009, 35 (2), 228–239.

29.

Papageorgiou

Rapley

Mesirov

J. P.

. A Genome-Wide siRNA Screen in Mammalian Cells for Regulators of S6 Phosphorylation. PLoS One 2015, 10 (3), e0116096.

30.

Chen

Hsieh

J.-H.

Huang

. Cell-Based High-Throughput Screening for Aromatase Inhibitors in the Tox21 10K Library. Toxicol. Sci. 2015, 147 (2), 446–457.

31.

Geng

Whiteley

Ribbens

. Novel Patient Cell-Based HTS Assay for Identification of Small Molecules for a Lysosomal Storage Disease. PLoS One 2011, 6 (12), e29504.

32.

Naumov

G. N.

Townson

J. L.

MacDonald

I. C.

. Ineffectiveness of Doxorubicin Treatment on Solitary Dormant Mammary Carcinoma Cells or Late-Developing Metastases. Breast Cancer Res Treat. 2003, 82 (3), 199–206.

33.

Lundqvist

E. Å.

Fujiwara

Seoud

Principles of Chemotherapy. Int. J. Gynecol. Obstet. 2015, 131 (Suppl. 2), S146–S149.

34.

Kato

Umezawa

Funeriu

D. P.

. Immobilized Culture of Nonadherent Cells on an Oleyl Poly(Ethylene Glycol) Ether-Modified Surface. Biotechniques 2003, 35 (5), 1014–1018.

35.

Tokuyama

Fujii

S.-I.

Sato

. Microbioassay System for Antiallergic Drug Screening Using Suspension Cells Retaining in a Poly(Dimethylsiloxane) Microfluidic Device. Anal. Chem. 2005, 77 (10), 3309–3314.

36.

Feng

Mitchison

T. J.

Bender

. Multi-Parameter Phenotypic Profiling: Using Cellular Effects to Characterize Small-Molecule Compounds. Nat. Rev. Drug Discov. 2009, 8 (7), 567–578.

37.

Huang

K. T.

Chen

Walker

A. M.

Inaccuracies in MTS Assays: Major Distorting Effects of Medium, Serum Albumin, and Fatty Acids. Biotechniques 2004, 37 (3), 410–412.

38.

Riss

Is Your MTT Assay Really the Best Choice?

2014. http://www.promega.com/~/pdf/resources/pubhub/is-your-mtt-assay-really-the-best-choice/.

39.

Gervasoni

J. E.

Hindenburg

A. A.

Vezeridis

M. P.

. An Effective In Vitro Antitumor Response against Human Pancreatic Carcinoma with Paclitaxel and Daunorubicin by Induction of Both Necrosis and Apoptosis. Anticancer Res. 2004, 24 (5A), 2617–2626.

40.

Arican

G. Ö.

Soy

N. N.

Effects of Epirubicin and Daunorubicin on Cell Proliferation and Cell Death in HeLa Cells. J. Cell Mol. Biol. 2005, 4, 47–52.

41.

Rao

D.-R.

Zheng

F.-M.

. Curcumin Reduces Expression of Bcl-2, Leading to Apoptosis in Daunorubicin-Insensitive CD34+ Acute Myeloid Leukemia Cell Lines and Primary Sorted CD34+ Acute Myeloid Leukemia Cells. J. Transl. Med. 2011, 9 (1), 1–15.

42.

Svensson

S. P. S.

Lindgren

Powell

. Melanin Inhibits Cytotoxic Effects of Doxorubicin and Daunorubicin in MOLT 4 Cells. Pigment Cell Res. 2003, 16 (4), 351–354.

43.

Haibe-Kains

El-Hachem

Birkbak

N. J.

. Inconsistency in Large Pharmacogenomic Studies. Nature 2013, 504 (7480), 389–393.

44.

Lyles

R. H.

Poindexter

Evans

. Nonlinear Model-Based Estimates of IC(50) for Studies Involving Continuous Therapeutic Dose-Response Data. Contemp. Clin. Trials 2008, 29 (6), 878–886.

45.

Williams

H. D.

Trevaskis

N. L.

Charman

S. A.

. Strategies to Address Low Drug Solubility in Discovery and Development. Pharmacol. Rev. 2013, 65 (1), 315–499.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.07 MB

Evaluation of the Droplet-Microarray Platform for High-Throughput Screening of Suspension Cells

Abstract

Keywords

Introduction

Materials and Methods

Preparation of DMA Slides

Cell Culture

Cell Growth and Viability Study

Aligning Frame for Sandwiching

Sandwiching Method

Fixation of Cells

Live/Dead Staining of Cells

Screening Initiation

Automated Imaging and Image Analysis

Statistics and Data Analysis

Results

Suspension Cells on Droplet-Microarray: Culturing and Manipulation

High-Throughput Screening Using the DMA Platform

Drug Treatment: Proof-of-Concept Study

Discussion

Conclusions

Footnotes

Acknowledgements

Declaration of Conflicting Interests

Funding

References

Supplementary Material