3D hydrogel cultures for high-throughput drug discovery

Abstract

Our increased understanding of how a cell’s microenvironment influences its behavior has fueled an interest in three-dimensional (3D) cell cultures for drug discovery. Particularly, scaffold-based 3D cultures are expected to recapitulate in vivo tissue stiffness and extracellular matrix composition more accurately than standard two-dimensional (2D) monolayer cultures. Here we present a 3D hydrogel cell culture setup suitable for automated screening with standard high-throughput screening (HTS) liquid handling equipment commonly found in a drug discovery laboratory. Further, we describe the steps required to validate the assay system for compound screening.

1. Introduction

Attrition rates in drug development remain high and significantly contribute to the immense cost of bringing new drugs to the clinic. Undoubtedly, there is an urgent need for new technologies that better accommodate precision in the preclinical stages of drug discovery. The rapidly expanding knowledge of how the microenvironment of a cell attenuates its phenotype, behavior and response to drugs and treatments, asks for new cell culture techniques in cell-based high-throughput/content screening that more closely recapitulate in vivo biology than the currently prevalent two-dimensional (2D) monolayers [1,2]. This is particularly important when the targets for drug discovery are part of the intricate signaling systems that sense changes in a cell’s microenvironment and transmit these changes to guide cellular behavior. In recent years, three-dimensional (3D) cell cultures have moved to the forefront as more in vivo-like environments. Current 3D culture systems are roughly categorized into non-scaffold, anchorage-independent (e.g. spheroids) and scaffold-based (e.g. hydrogel) systems with more complex systems such as organoids and microfluidic devices becoming more common [1,3–5]. However, remaining technological challenges must be overcome in order to apply 3D cultures to truly automated high-throughput drug discovery.

Scaffold-based 3D cultures are especially well-suited to mimic intricate cell-cell and cell-extracellular matrix (ECM) interactions usually found in a tissue. We previously showed that the self-assembling and hydrogelating MAX8 β-hairpin peptide is suitable as a 3D cell culture technology for automated high-throughput drug discovery [6,7]. We have shown that MAX8 combines biocompatibility and tunability in function and stiffness [6–14] with unique mechanical properties (e.g., shear-thinning, injectable solid with immediate rehealing) [8–10,15–23] that allow automatic handling with standard high-throughput screening (HTS) liquid handling equipment commonly found in a drug discovery laboratory [6,7]. Here we describe the basic steps of generating MAX8–3D cell cultures in a 384-well setup using liquid handling robotics, assay validation using cell viability as a measure, and performing a pilot screen.

2. Materials

Prepare all solutions using ultrapure water and analytical grade reagents. For the culture of human cells, institutional biosafety procedures apply.

2.1. Cell culture

1. Cell line: ONS-76 medulloblastoma cells, a human pediatric brain cancer cell line, as the example (see Note 1).

2. Culture media: Dulbecco’s Minimal Essential Media (DMEM) supplemented with 10% fetal bovine serum (FBS), penicillin-streptomycin-glutamine (1X).

3. Sterile Phosphate-Buffered Saline (1X PBS), pH 7.4: NaCl (9g/L), Na₂HPO₄ (0.795g/L), KH₂PO₄ (0.144g/L).

4. Trypsin/EDTA (0.05%), tissue culture grade.

5. 10 cm tissue culture plates.

6. 384well, sterile, white, flat bottom plates for culture of cell/gel mix (see Note 2).

7. Sterile centrifuge tubes.

8. Benchtop swinging bucket centrifuge.

9. Hemacytometer.

10. Humidified CO₂ incubator set at 37 °C, 5% CO₂ for cell culture.

11. Biological safety cabinet for manipulation of cells in sterile environment.

2.2. Peptide hydrogel

1. β-hairpin peptide: MAX8-RGDS [RGDS-VKVKVKVK-(V^DPPT)-KVEVKVKV-NH2] (see Note 3).

2. HEPES pH7.4 (50 mM): dilute 1ml sterile 1M HEPES into 19ml sterile ultrapure water. Confirm pH is near 7.4 with pH paper.

3. DMEM cell culture medium without any additives.

2.3. Compound screening

1. Commerically available, single step, luminescence-based cell viability assay that allows for continuous monitoring of cell growth. Prepare as 5x assay mix.

2. SAHA (suberanilohydroxamic acid, vorinostat; 50 mM): Resuspend 10mg SAHA in 757ul DMSO.

3. Compound screening library: MicroSource Spectrum Collection or any other screening library in a 384-well format (see Note 4).

4. DMSO (99.9%) for drug dilution, stored desiccated.

5. Isopropanol and ultrapure water for washing tips.

6. 384-well polypropylene plates for drug dilution into 5X assay mix.

7. 384-well small volume polypropylene plates for dispensing cell/gel mix with automated liquid handling system.

8. Automated reagent dispenser.

9. Automated workstation for liquid handling, including dispensing tips.

10. Robotic pin tool for initial drug dilutions from the library.

11. Lint-free blotting paper for tip cleaning/drying.

12. Plate reader to measure luminescence signal in cps (counts per second)

2.4. Statistical analysis

1. Statistical analysis software.

2. QC (quality control) plates to determine Z’ factor [24] (see Note 5).

3. Methods

Prior to setting up 3D cell cultures for compound screening, the following parameters should be established for each cell line and each assay to be used: (1) cell growth curves to determine optimal cell numbers and incubation times in the desired hydrogel concentration, (2) DMSO sensitivity of cells encapsulated into the hydrogel, (3) assay parameters such as signal-to-noise ratio, (4) substrate concentrations or incubation times.

While QC plates are routinely included during compound screening, it is advisable to run some QC plates prior to the screen to validate and possibly optimize the assay setup. Fig. 1 shows a chart of the work flow. An example for the assay described here is in [6,7] (see Note 6).

Figure 1. Flowchart for HTS in 3D hydrogel cultures.

CPS, counts per second.

3.1. Cell culture

1. Cells are grown in appropriate growth media on tissue culture plastic in culture media in a humidified incubator at 37° C, 5% CO₂, to sub-confluent density. Manipulation of cells is done in a biosafety cabinet to maintain sterility, except when using the liquid handling system (see Note 7).

2. To harvest cells from the tissue culture plasticware, remove culture media, then rinse cells twice with PBS. Add 0.05% trypsin/EDTA solution to cells (1ml/100mm plate) and incubate at room temperature for several minutes. Cells will appear rounded under microscopic inspection and start floating off plate during incubation. At this point, add culture media to cell/trypsin suspension and triturate mixture to single cell status. Use a 10uL aliquot on a hemacytometer to determine cell number.

3. Recover cells from suspension by pelleting at 300x g for 5min. Remove the supernatant, and resuspend the cell pellet in serum-free DMEM to a density of 1×10⁶/mL.

3.2. Dispensing hydrogel-cell constructs in 384-well plates

1. Peptide is dissolved in 50 mM HEPES buffer (pH 7.4) to make a 0.5 wt% stock (50 mg peptide per 10 ml of buffer) and stored at 4°C. This forms a gel-like substance that can be prepared the day before use.

2. Mix cell suspension (see 3.1.3) with an equal volume of the hydrogel solution (0.5 wt%) to yield a final concentration of 2000 cells/4uL of 0.25 wt% hydrogel (see Note 8). This is the cell/gel mix used for drug screening.

3. To each assay plate, dispense culture media at 36ul/well into 384-well sterile, white, flat bottom plates using the automated reagent dispenser (see Note 9).

4. Add 4uL/well of cell/gel mix using the automated liquid handling workstation into columns 1 and 3–23 of each plate. Columns 2 and 24 receive media only or hydrogel only, respectively, as low (background, no cells) controls (see Note 10).

5. Culture plates in 37 °C, 5% CO₂ humidified incubator for 24 h before addition of compounds for assay (see Note 11).

3.3. Compound screening

This section describes the distribution of a compound screening library into the assay plates and determination of cell viability upon treatment. The library used is distributed in a 384-well format, with compounds suspended in DMSO.

1. Prepare the cell viability assay mix as 5x assay mix and transfer the reagent (25uL/well) into 384-well plates using the automated liquid dispenser.

2. The library compounds are then diluted into the 5x assay mix using the pin tool head to deliver 50nL DMSO drug stock into the 25uL assay mix (1:500 dilution). Only wells in columns 3 through 22 receive drug, as columns 1, 2, 23, and 24 have only DMSO in the compound plate so that columns 1 and 23 have no drug for the positive control, and columns 2 and 24 have no cells for low (background) control. A schematic setup of the plates is shown in Fig 2.

3. This intermediate dilution of compounds is then dispensed as a 10uL aliquot into the cell/gel assay plates using the liquid handling workstation and appropriate tips, according to manufacturer’s instructions (see Note 12). The final volume of 50uL/well in each assay plate contains 1x cell viability assay mix and 1:2500 dilution of drug stock and a final DMSO concentration of 0.04%. The control wells (columns 1 and 23 for high control, and 2 and 24 for low control) also have final concentration of 1x cell viability assay mix and 0.04% DMSO.

4. Assay plates are then incubated in a humidified 37 °C, 5% CO₂ incubator for 48 h to allow for cell growth.

5. Plates are brought to room temperature for 30min (see Note 13) before reading luminescence on the plate reader. Results are expressed as counts per second (cps).

6. Quality control (QC) plates are included with each screen to validate the integrity of the assay each time it is run. With this assay, the compound SAHA (suberanilohydroxamic acid, vorinostat) is used as the control compound to show inhibition of cell viability. Depending on cell type a different control compound may be chosen. QC plates are set up with the same process as the compound screening plates, using the intermediate dilution with the pin tool, followed by addition of an aliquot of this drug dilution to the cell/gel assay plate, but the QC plates have the same concentration (close to IC₅₀) of SAHA in all wells of columns 3 through 22. These plates are then incubated for 48 h, followed by luminescence reading after reaching room temperature, just as is done with the compound screening plates.

Figure 2. Single dose assay plate format.

Wells in columns 1 and 23 (green) have cells but no drug (high control) and wells in columns 2 and 24 (red) have culture medium but no cells for background control. Assay wells in columns 3 through 22 (blue) receive cells and drug compounds.

3.4. Statistical analysis

1. For each plate run in the assay, the Z’ factor and Coefficient of Variation (CV) are calculated to verify the robustness and reproducibility of the assay [6,7,24,25]. A Z’ factor >0.5, and CV < 20% represent a respectable assay.

2. The Z’ factor is calculated as follows:

   $\mathrm{Z}\mathrm{’}=1-\frac{3\mathrm{*}\left({\sigma }_{High}+{\sigma }_{Low}\right)}{\left|{\mu }_{High}-{\mu }_{Low}\right|}$, with $\mathrm{\sigma }$ representing the standard deviation, $\mathrm{\mu }$ the mean, ‘High’ the wells without drug, and ‘Low’ the wells without cells.

3. A Z’ factor in the range between 0.5 and 1 indicates that the assay is suitable to HTS. For Z’ factors between 0 and 0.5, the assay is still acceptable, however, re-optimization is strongly recommended before large-scale application.

4. The CV is calculated as follows: $\mathrm{C}\mathrm{V}=100\mathrm{*}(\mathrm{\sigma }/\mathrm{\mu })$, with $\mathrm{\sigma }$ representing the standard deviation, $\mathrm{\mu }$ the mean.

5. The lower the CV, the less variability in data points and a CV < 20% is desirable.

6. The effect of each compound on the viability of the cell/gel mix in each well is calculated as percent inhibition.

   $\mathrm{P}\mathrm{e}\mathrm{r}\mathrm{c}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{I}\mathrm{n}\mathrm{h}\mathrm{i}\mathrm{b}\mathrm{i}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}=100-100\mathrm{*}\left(CP{S}_{Treated}-CP{S}_{Low}/CP{S}_{High}-CP{S}_{Low}\right)$, where $CP{S}_{Treated}$ is the signal from the compound well, $CP{S}_{Low}$ is the average signal from the background wells on the plate, and $CP{S}_{High}$ is the average signal from the untreated control wells on the plate.

7. The classification of a compound as a ‘hit’ is made by comparing its percent inhibition to the standard deviation of the signal of the background (Low) wells of the plate. Inhibition greater than 3x this standard deviation is worth pursuing as a ‘hit’ (see Note 14).

3.5. Dose response curves of compounds of interest

1. Sixteen-point dose response curves are generated for each compound of interest, by placing the highest dose of compound in the top row of the drug dilution plate, then diluting 2x into DMSO as an equal amount of compound is added to the DMSO in the row below. Dilutions are done sequentially from top to bottom of the plate using the automated workstation (see Note 15 and Fig. 3).

2. The control compound SAHA is included in each dose response plate to monitor plate-to-plate variation. Columns 1, 2, 23, and 24 contain DMSO only, for the ‘High’ and ‘Low’ wells of each plate.

3. Dose response assays should also be performed using intermediate dilutions for each compound. Proceed with compound screening (see section 3.3), adjusting the volumes of compound and cell viability assay mix accordingly.

4. Data is analyzed with appropriate statistical analysis software using a nonlinear regression of dose-response.

5. Compounds that are validated by dose response are then resynthesized and dose response assays are repeated to verify the identity of the compound producing the results.

Figure 3. Dose response plate format.

As for the single dose assay plates, wells in columns 1 and 23 (green) have cells but no drug (high control) and wells in columns 2 and 24 (red) have medium but no cells for background control. Wells in column 22 (orange) receive decreasing concentrations of control compound (SAHA as an example). Columns 3 through 21 (blue) receive 16-point dilutions of test compounds.

4. Notes

1. MAX8 hydrogel is biocompatible with a wide range of cells, including various cancer cell lines, mesenchymal stem cells and primary neuronal cultures such as cerebellar granule cells [6–12].

2. We recommend using high-quality 384-well plates that prevent cross-contamination, are molded to exacting standards and meet ANSI/SLAS dimensions.

3. Although MAX8 is biocompatible, some cells exhibit higher growth rates in MAX8-RGDS [6,7]. The peptide may also be synthesized with other adhesive ligands. MAX8-RGDS peptide can be obtained from commercial sources or synthesized with an automated peptide synthesizer, using standard Fmoc-based solid phase peptide synthesis. MAX8 peptide synthesis has been described in detail [8,10]. Ensure that the peptide has high purity (>95%) and the lyophilized powder should be free of any residual contaminants that may interfere with cell viability.

4. The MicroSource Spectrum library contains bioactive compounds and natural products at a 10mM concentration. The library is stored desiccated in sealed plates at −30°C to maintain integrity of compounds.

5. Before any high throughput screen is undertaken, a screening assay must be developed and validated to ascertain its quality. Common measures are the Z’ factor and coefficient of variation (CV) values calculated as described in 3.4.

6. For the cell viability assay used for the compound screen described here, we first validated that the luminescence obtained in the readout correlates with the number of cells encapsulated in the hydrogel (Fig. 4a) and that the cells proliferate without a plateau in growth throughout the 72 h timeframe of the experiment (Fig. 4b). The background luminescence from dead cells was determined to be low compared to that of viable cells (Fig. 4c). The Signal-to-Noise Ratio was calculated as follows: S/N = CPS_Treated/CPS_Low

7. The viability of the encapsulated cells was determined to be unaffected up to a 1% final concentration of DMSO. DMSO is the solvent for the compounds in the screening library and is introduced into the assay with the addition of the screening compounds. When a known inhibitor (vorinostat, SAHA) is tested in the assay on a 384-well plate, the variability in response to the drug (grey triangles) across the plate can be determined, as well as the Z’ factor calculated from the response to high control (blue diamonds, no drug) and low control (orange squares, no cells) (Fig. 4D). Using several QC plates, the calculated Z’ factor was 0.58, within the desired 0.5 to 1 robust assay range.

8. For workstations lacking an environmental enclosure, care should be taken to avoid contamination of cultures by keeping the room environment clean.

9. 0.25 wt% MAX8 is a relatively soft gel. If a stronger gel is desired, the peptide stock solution can be adjusted accordingly.

10. Plates that are not cell culture treated should be used for this step to avoid cells adhering to the plastic of the plate. Culture media must be added to plates before cell/gel mix to maintain viability of cells.

11. Assay plates should be set up in duplicate or triplicate and the volumes of the peptide stock solution should be adjusted accordingly.

12. We prefer to let the cells equilibrate in the hydrogel for 24 h before proceeding to a library screen. Depending on the cells used for the screen and their growth curves, longer or shorter equilibration times may be needed.

13. To reduce cost and plastic waste, tips for the automated liquid handling station can be washed between transfer uses. This requires blotting tips on lint-free blotting paper between solutions and using large volumes (400–500ml) of PBS to remove the previous solution. The PBS should be triturated to remove/dilute any liquid on the internal surface of the tips. The tips are again blotted onto blotting paper, followed by dipping and triturating in a container of isopropanol. The blotting and isopropanol rinsing is repeated, then tips are blotted on blotting paper, followed by trituration just in the air to dry the tips. The first solution used for rinsing should be similar in composition to the solution being removed. In this example, PBS is used since the compounds are in media, and PBS is an aqueous, pH-balanced solution like the media, but without additives. If the compounds were in a DMSO solution, then the first solution in this process should be DMSO/water (1:1). The final two rinses are done in isopropanol to further dilute/remove any residuals and to allow fast drying. Any similar washing process used in the laboratory should be validated by testing the removal of a measurable substance, such as fluorescein. Transfer fluorescein solution to a plate, wash tips used with desired method, then use washed tips to transfer non-fluorescent solution to new plate. If there is any measurable fluorescence in this new plate, your washing process is not thorough enough to prevent transfer from plate to plate. You will need to use larger volumes in your washes, or a larger number of washes in the process, or change tips between plates.

14. Temperature affects luminescence, so read all plates at the same temperature to avoid variation due to this parameter. We choose to bring plates to room temperature to ensure that there is no plate-to-plate variation due to plate temperature.

15. As an example, in one of our previous screens the cutoff level was 57% inhibition (Fig. 5). Compounds with greater than 57% inhibition were assayed with dose response curves to verify activity.

16. Example for the 16-point dose response curve. 10uL of compounds in wells of row A are added to 10uL DMSO in adjacent wells of row B for the first 2x dilution. After thorough mixing, 10uL of diluted compounds in wells of row B are added to 10uL DMSO in adjacent wells in row C for next 2x dilution in series. This 2x dilution series continues through each row of the plate to result in 16-point dose series for each compound on the plate (Fig. 3). Fig. 6 shows examples of dose response curves obtained from four different compounds.

Figure 4. Basic characterization of 3D cultures for HTS.

(A) Cell viability signal from an increasing number of ONS76 cells encapsulated in MAX8 hydrogel. (B) Longitudinal growth measurements of ONS-76 cells encapsulated in MAX8. (C) Cell viability test comparing untreated control cells, cell treated with ethanol to induce cell death (dead cells), and wells with medium but no cells as background control. (D) Quality control plate to determine the statistical robustness of the assay. Blue, untreated cells (maximum signal); orange, background control; grey, cells treated with LD50 of SAHA. (A-D) CPS; counts per second.

Figure 5. Graphical summary of a 3D pilot screen.

ONS-76 cells were encapsulated in MAX8-RGDS as described and a pilot screen was performed as described in the methods section. The line indicates the threshold of three standard deviations used to identify hits.

Figure 6. Examples of dose response curves of selected compounds.

16-fold dose response curves of ONS-76 cells encapsulated in MAX8-RGDS using a 384-well setup done in triplicates. 24 h after plating, the cells were treated with individual compounds for 48 h and Cell viability was determined using a luminescence-based assay. CPS, counts per second.