The Use of High Content Imaging and Viability Readouts for Screening with 3D Spheroids Alison Gordon1 & Gary Allenby1 1 Aurelia Bioscience Ltd., Biocity, Pennyfoot Street, Nottingham, NG1 1GF

Introduction CellTiter Glo® 3D Viability

Cell based assays have widely been used in discovery, U87 Cells HCT116 Cells Spheroids were allowed to form in ULA however, it is becoming increasingly clear that the traditional 6000000 8×106 plates for 4-days, this was an optimum Untreated Cells Untreated Cells 5000000 ) 2D model of culture is not truly representative of how ) U 6 Nocodazole

b) U

time for cell compaction prior to drug L 6×10 L

R 4000000

( R

Paclitaxel

cells grow and respond in vivo. To address this disparity, 3D (

e e

treatment. To assess cellular viability, c c

n 3000000 6 Mitomycin C n

cell culture technology has made vast improvements in recent e 4×10

e c

WYE 687 c

spheroids were cultured and then s WYE 687 s

e 2000000

e n

years and is fast emerging as a new approach to drug i n Staurosporine i 6 Staurosporine

treated with an increasing m 2×10 m

u 1000000 u discovery allowing for a greater understanding of drug efficacy L Camptothecin concentration of compound. Each cell L Camptothecin and toxicity before move into the phase. Untreated Cells 0 0 line showed a dose-dependent -10 -9 -8 -7 -6 -5 -4 -10 -9 -8 -7 -6 -5 -4 decrease in viability with an increasing Log [Compound] (M) Log [Compound] (M) Aurelia has implemented 3D cell culture techniques which can concentration of compound. Figure 1: Cell viability as determined by the presence of ATP in U87 or HCT116 spheroids. easily be adapted into an HTS format. Ultra-low attachment Data is expressed as a percentage of the untreated control. (ULA) plates allow spheroids to form naturally. This technology relies upon a cells innate ability to grow and Nocodazole Paclitaxel Mitomycin C WYE 687 Staurosporine Camptothecin interact in a 3D format to form a spheroid. This provides a U87 416nM 535nM 935nM 1165nM 9nM 553nM better representation of in vivo cell physiology which is lost in HCT116 78nM 84nM 229nM 1622nM 10nM 376nM the tradition monolayer format. Using this technology we Table 1: Estimated IC values determined in the CellTiter Glo® 3D Viability assay in spheroids derived from U87 or HCT116 cells. were able to monitor drug potency and efficacy in either an 50 imaging-based assay (on the CellInsightTM CX5 High Content Screening (HCS) Platform) or by using commercially available Growth Inhibition assays such as the CellTiter Glo® 3D viability assay (Promega). Spheroid size was quantified on the CellInsightTM CX5 High Content Screening (HCS) Platform using an overlay on the brightfield imaging channel. Here we test two spheroid cultures derived from either a HCT116 (human colorectal carcinoma) or U87 (human Figure 2: HCT116 derived spheroid imaged in brightfield. The blue mask glioblastoma) cell line. Each cell line produces a compact which overlays the image details the size quantification. spheroid. We have used a number of readouts to generate a Prior to treatment, spheroid sizes showed an initial decrease due to cell compaction. After this time, in the absence of pharmacological profile of each spheroid type. compound, spheroids grew in a linear manner up to day 8 of the experiment. Compounds induced a dose-dependent inhibition of growth as characterized by a reduction in the spheroid size. This was consistently seen in the U87 cell line. In comparison, compounds paclitaxel and nocodazole induced a looser cytoskeletal organization in HCT116 spheroids at higher concentrations.

Methodology Mitomycin C Mitomycin C HCT116 and U87 cells were cultured in DMEM media 10mM 5mM 2.5mM 1.25mM 62.5nM 31.3nM 15.6nM DMSO supplemented with 10% FBS in TC-treated flasks at 37°C, 5% 10mM 5mM 2.5mM 1.25mM 62.5nM 31.3nM 15.6nM DMSO WYE687 WYE687 CO2. For spheroid formation, each cell line was seeded at 400 cells/well (40ml/well) in cell culture medium in 384-well ultra 10mM 5mM 2.5mM 1.25mM 62.5nM 31.3nM 15.6nM DMSO 10mM 5mM 2.5mM 1.25mM 62.5nM 31.3nM 15.6nM DMSO a) b) low attachment (ULA) microplates (Corning). The plates were Staurosporine Staurosporine incubated at 37°C, 5% CO2 for a period of 4 days to allow the cells to cluster and form a spheroid. Compounds were added 100nM 50nM 25nM 12.5nM 6.25nM 3.13nM 1.56nM DMSO 100nM 50nM 25nM 12.5nM 6.25nM 3.13nM 1.56nM DMSO on day 4 at a 1:2 dilution (40ml/well) and incubated with cells Figure 3: Brightfield images of spheroids consisting on U87 cells (left panel) or HCT116 cells (right panel). for a further 4 days before measurements were taken. A U87 Size B HCT116 Size C Staurosporine 400000 400000 400000 Nocodazole Nocodazole Day 4 The effect of each compound was determined in three Paclitaxel Paclitaxel Day 6

300000 a

300000 a 300000

e e r

different assays. Viability was measured using the CellTiter mitomycin C r mitomycin C Day 7

A A

a

d d

WYE 687 WYE 687 e Day 8

i i

r 200000 o

Glo® 3D Viability Assay (Promega). An equal volume of o

A r

200000 r 200000 e

Staurosporin e Staurosporin Day 4-Control

h h p

reagent was added to the culture medium. Plates were mixed p S Camptothecin S Camptothecin 100000 Day 6-Control for 5 minutes on an orbital shaker to ensure good cell lysis. 100000 Control 100000 Control Day 7-Control 0 Day 8-Control The plates were allowed to equilibrate at room temperature -10 -9 -8 -7 -6 -5 -4 -10 -9 -8 -7 -6 -5 -4 -9 -8 -7 for 20-25 minutes after which time luminescence was Log [Compound] (M) Log [Compound] (M) Log [Staurosporine] (M) determined using the Enspire® Plate Reader (PerkinElmer). Figure 4: Spheroid area (in pixels) plotted against compound concentration in the presence of U87 (A) or HCT116 (B) cells, data is taken from an 8-day The luminescent signal is directly proportional to the level of time point. The cytotoxic effects of staurosporine can be seen over time (C). ATP within the cells and hence gives a measurement of Nocodazole Paclitaxel Mitomycin C WYE 687 Staurosporine Camptothecin metabolic activity and cell viability. U87 154nM 26nM 177nM 315nM 21nM 26nM HCT116 - - 31nM 1407nM 23nM 4nM Over time, the size of the spheroid was monitored using Table 2: Estimated IC values derived from the spheroid size at day 8 derived from spheroids comprising of U87 or HCT116 cells. brightfield microscopy on the CellInsightTM CX5 High Content 50 Screening (HCS) Platform (Thermo Scientific). An algorithm was used to define the area of the spheroid using a 10x Spheroid Viability objective. Spheroid is characterized by a mixed population of cells whereby a hypoxic core is surrounded by a viable and replicating TM In addition to monitoring spheroid size, we also stained the collection of cells on the periphery. In the absence of compound, DRAQ7 staining increased in untreated spheroids over time. TM spheroids with DRAQ7 , a far-red DNA stain that will only Figure 5: HCT116 derived spheroid. The hypoxic core is stained with stain the nuclei of dead or permeabilised cells. Spheroids were DRAQ7TM (red), a far-red dye that will only stain dead cells. Viable cells are stained with DRAQ7TM at a 1:2 dilution to give a final stained with Calcein AM (green). concentration of 1.5mM. Spheroids were treated with compounds early in their growth phase prior to the formation of distinct necrotic core. Compounds induced a gradual increase in DRAQ7TM staining over time in a dose-dependent manner. Treatment with nocodazole or paclitaxel Summary and Conclusion significantly increased DRAQ7TM staining, this may be due to the loose nature of the spheroids in the presence of compound. • All compounds tested affected cell viability in a dose dependent manner (CellTiter Glo® 3D Viability Assay). While Paclitaxel Paclitaxel 10mM 5mM 2.5mM 1.25mM 62.5nM 31.3nM 15.6nM DMSO 10mM 5mM 2.5mM 1.25mM 62.5nM 31.3nM 15.6nM DMSO there was some variation in potency, not all compounds were able to restore viability to untreated levels within the Mitomycin C Mitomycin C concentration range tested. TM • Further investigation with a DRAQ7TM dead cell stain Figure 5: DRAQ7 staining of U87 spheroids (left panel) or HCT116 spheroids (right panel) at day-8 in the presence of increasing concentrations of compound. revealed that those compounds which showed reduced U87 Cells HCT116 Cells 8 4×10 8 Nocodazole 3.5×10

y Nocodazole

viability also appeared to have a necrotic core. This effect t y

i 8

t 3×10 i s 8 Paclitaxel s Paclitaxel

D n 3×10 n was also dose dependent. e 8 t 2.5×10

mitomycin C e t

n mitomycin C

I

n

I

l 8 2×10 8 WYE 687 l • Growth inhibition was determined by calculating spheroid a 2×10

t WYE 687

a t

o 8 o

T Staurosporin 1.5×10

T Staurosporin

area. We were able to monitor growth inhibition in a dose- 7

8 7 8

q 1×10 Camptothecin 1×10 q

a Camptothecin

r

a r and time- dependent manner in both spheroid types tested. D Untreated U87 Cells 5×107 D Untreated HCT116 cells Only nocodazole and paclitaxel increased spheroid size. They 0 0 -10 -8 -6 -4 -10 -8 -6 -4 appeared to cause shedding of peripheral cells and a loss of Log [Compound] (M) Log [Compound] (M) the tight spheroid morphology. Figure 4: DRAQ7TM intensity plotted against compound concentration using either U87 or HCT116 cells. • Taken together, we were able to build a pharmacological Nocodazole Paclitaxel Mitomycin C WYE 687 Staurosporine Camptothecin profile for each compound tested. U87 476nM 233nM 2107nM - - 219nM www.aureliabio.com HCT116 54nM 125nM 1857nM - - 495nM TM [email protected] Table 3: Estimated IC50 values derived from DRAQ7 intensity in spheroids - comprising of U87 or HCT116 cells.