Analysis of Drug Resistant Properties of A2780 Ovarian Cancer Cell Lines

Using Label-Free Automated Microscopy (IncuCyte Zoom)

Authors Andrew Bashford and Jim Cooper

Scientific Development Group

Culture Collections

National Infection Service

Public Health England

Abstract

The A2780 ovarian cancer cell line and its reported cisplatin and Adriamycin (generic name:

doxorubicin) drug resistant derivatives (A2780cis and A2780ADR) have been available for

over 20 years and offer researchers the potential to study drug resistance in the field of

cancer therapeutic research. The cell lines were quantitatively characterised for their

resistant properties soon after derivation by various technically demanding and inconsistent

methods, where relative IC50 values were reported as a measure of resistance. We

hypothesised that despite several decades of cell banking and storage, the A2780cis and

A2780ADR cell lines had retained their drug resistant characteristics as compared to the

parental cell line and by utilising novel automated microscopy and image analysis software,

we could quantify these properties using a label free kinetic assay and end point nuclear

fluorescence staining. Our results qualitatively and quantitatively demonstrated drug

resistance in the A2780cis and A2780ADR cells and the IC50 values derived using both the

kinetic assay and nuclear staining were consistent with previously published data derived

using alternative methodologies. This study has independently verified that the cell lines

retain the drug resistance properties as reported decades previously. We believe this novel

approach using an automated, high throughput format allows the rapid and robust

quantification of the drug resistant properties of cell lines and will be of use to researchers in

the field working with these and other cancer derived cells.

Introduction

The ovarian cancer cell line A2780 (ECACC 93112519) and its drug resistant derivatives

(A2780cis, ECACC 93112517 and A2780ADR, ECACC 93112520) are widely used tools for

cancer research and have been in general use for over 20 years. The resistant cell lines

were generated by monolayer culture of A2780 cells in increasing concentrations of

Adriamycin (generic name: doxorubicin) (Behrens et al. 1984) or cis-diammineplatinum(II)

dichloride (cisplatin) (Behrens et al. 1987). Together these cell lines represent an excellent

means to study cancer biology, resistance to chemotherapeutic agents and enable the

search for new therapeutic agents.

Drug resistance in cell lines is best compared using IC50 values; the dose of a drug which

causes a 50% reduction in cell viability. A number of different methodologies have been

employed to determine this value. These include the use of agar-based clonogenic assays

(Puck and Marcus 1956), a modified version: the human tumour stem cell assay (HTSCA)

(Hamburger and Salmon 1977), the resazurin assay for cell viability (Tardito et al. 2009),

ATP assays for cell viability (Fan et al. 2004), HTSCA combined with thymidine incorporation

(Friedman and Glaubiger 1982), HTSCA combined with the ATP assay (Sevin et al. 1988),

MTT cell proliferation assay (Yamaue et al. 1996) and dye exclusion methods (Weisenthal et

al. 1983). These approaches have different durations of growth, are labour intensive and can

often yield different experimental results.

The European Collection of Authenticated Cell Cultures (ECACC) and its official distributors

are the exclusive suppliers of A2780, A2780cis and A2780ADR cell cultures. Here we report

a modern approach using the automated IncuCyte ZOOM system and a label-free

assessment of IC50 values.

Methods

Authenticated (Short Tandem Repeat (STR) profiled) cell lines were provided from ECACC

and grown in accordance with the recommended procedures. In short, cells were grown in

RPMI-1640 media with 2mM Glutamine (R8758, Sigma Aldrich) and 10% FBS (Sigma

Aldrich). Initial cell expansion was performed in T175 flasks (431466, Corning) seeding at

densities of 5x104 cell/cm2 (A2780) or 3x104 cell/cm2

(A2780cis, A2780ADR). At the first

passage cells were washed with PBS, treated with 0.05% Trypsin-EDTA (Gibco®) and 100μL

of cells seeded into black walled cell bind 96 well plates (3340, Corning) at densities of

3x104 cell/cm2 (A2780) and 1x104 cell/cm2 (A2780cis, A2780ADR). For each cell type thirty-

two wells were seeded per ninety-six well plate and two plates were seeded in total.

Drug dilutions were added to cells immediately after they were plated. Cisplatin

(P4394, Sigma) was dissolved in PBS to make a 5mM working stock. This was serially

diluted using culture media and 100μL added to cells to make final concentrations of:

Fig 1) A2780 cells and its drug resistant clones; A2780cis, A2780ADR. STR profiles demonstrating cell line identity for a) A2780, b) A2780cis and c) A2780ADR. Representative images showing 0 hr time point for untreated cells d) A2780, e) A2780cis and f) A2780ADR. 72 hr time point for untreated cells g) A2780, h) A2780cis and i) A2780ADR. Scale bar: 70μm

500μM, 100μM, 10μM, 1μM, 0.1μM, 0.01μM, 0.001μM. For each cell type a total of 8 wells

(4 per plate) were treated with each concentration. Adriamycin (44583, Sigma) was

dissolved in DMSO to make a 10mg/ml stock. This was serially diluted using culture media

and 100μL added to cells to make final concentrations of: 150μM, 100μM, 10μM, 1μM,

0.1μM, 0.01μM, 0.001μM. For each cell type a total of 6 wells (3 per plate) were treated with

each concentration of Adriamycin. DMSO was also serially diluted using culture media to act

as a vehicle control for Adriamycin treatments; 100μL was added to cells to make final

concentrations of: 0.85%, 0.57%, 0.057%, 0.0057%, 0.00057%, 0.000057%, 0.0000057%.

For each cell type a total of 2 wells (1 per plate) were treated with each concentration of

DMSO.

Plates were incubated (37°C, 5% CO2) in the IncuCyte ZOOM (Essen BioScience, Ann

Arbor, Michigan, US) and phase images taken every two hours using the 20x objective lens.

Nuclear stain Syto16 (S7578, Life Technologies) was diluted in PBS to make 5μM stock.

After 120 hours wells were washed with PBS and Syto16 added. Plates were returned to the

IncuCyte ZOOM, incubated for 20 minutes and re-imaged using both phase and green

fluorescence channels. Data were analysed using the IncuCyte ZOOM software and

processing definitions modified to quantify confluence and cell number. Processing definition

were set as follows:

- Confluence: Segmentation Adjustment of 0.1, hole fill 0.00μm2, adjust size -2 pixels.

Filters: Area min 100μm2.

- Cell count (main): Parameters: Top-Hat threshold adjustment with radius of 10μm

and threshold = 2.000, edge split on with sensitivity = 0. Filters: area min 60.00μm2,

mean intensity min 25.00.

- Cell count (low, high debris): Parameters: Adaptive threshold adjustment with

threshold = 5.000. Edge split on with sensitivity = 11, hole fill 0.00μm2 and adjust size

= -1 pixels. Filters: area min 90.00μm2, area max 900.00μm2, eccentricity max 0.850.

Accuracy of processing definitions was qualitatively assessed to determine whether main or

low cell count processing definitions where appropriate.

GraphPad Prism 7 was used to plot drug concentration against confluence or cell number

and fit non-linear regression to determine the IC50 values. To test the effect of DMSO, values

obtained from Adriamycin treated wells were normalised using values obtained from

corresponding DMSO concentrations. No difference was seen in IC50 values calculated

using normalized or raw values.

Fig 2) The IncuCyte ZOOM is able to accurately quantify confluence and nuclear count. Representative images from A2780 cells after 120 hr showing a) phase data, b) fluorescence of nuclear stain Syto16. The IncuCyte ZOOM trained processing definition to quantify c) percentage confluence and d) nuclear count as number per mm

2.

Scale bar: 70μm

Results

All human cells supplied by ECACC undergo Short Tandem Repeat (STR) profiling to

ensure their correct identity (Fig 1a-c). To independently verify the drug sensitivities of

A2780, A2780cis and A2780ADR, cells were cultured using the IncuCyte ZOOM, an

automated live-cell analysis system which accurately quantifies confluence and cell number

which prompted us to test its use as a novel, label-free method for determining IC50 values.

Cells were grown in serial dilutions of cisplatin or Adriamycin and imaged every two hours

(Fig 1d-f). Untreated cells reached mid log phase after approximately 72 hours (Fig 1g-i) and

they showed similar morphology between the different cell types. Nuclear dye Syto16 was

added to the cells at the end of the experiment (120 hours) and cells imaged (Fig 2a,b). The

IncuCyte ZOOM created processing definitions which were modified to quantify cell

confluence, based on phase images, or cell number, based on the fluorescent nuclear stain

(Fig 2c,d). The two drugs caused different cellular effects at high concentrations, with

Adriamycin producing a greater amount of cellular debris. This resulted in the needs for two

different processing definitions in order to obtain accurate count cells; one to accommodate

the low cell number with high debris (low, high debris) and one to accommodate high cell

number with low debris (main).

IC50 values were calculated using cell confluence (72 hours, Fig 3) and cell count (120

hours, Fig 4) representing the mid-log phase and end-point of the experiment respectively

(summarised in Table1). Both methods of IC50 calculation produced very similar values for

each cell type, demonstrating equivalence between the two techniques. When treated with

cisplatin, the A2780 cells had a range of IC50 values between 0.4-0.7μM. The A2780cis cells

showed a ten-fold resistance to cisplatin with an IC50 range between 7-8μM and the

A2780ADR showed a slight resistance with IC50 values of 1μM. When treated with

Adriamycin, the A2780 cells had an IC50 range between 0.004-0.007μM. In comparison the

A2780ADR cells had an IC50 range between 0.06-0.9μM, again showing over ten-fold

resistance and the A2780cis cells had an IC50 range between 0.01-0.02μM. These results

are strikingly similar with those published over 20 years ago and generated using different

experimental techniques (Table 1).

Fig 3) Confluence-based IC50 values show A2780cis and A2780ADR have higher drug resistance than parental A2780 cells. Graphs used to calculate IC50 values from cell confluence data at 72 hours. Cells treated with cisplatin (a-c) or Adriamycin (d-f). a, d) A2780; b, e) A2780cis and c, f) A2780ADR. GraphPad Prism was used to fit non-linear regression to the data and calculate IC50 values. Error bars are SEM generated from 3 (Adriamycin) or 4 (cisplatin) technical replicates at each drug concentration.

Fig 4) Cell number-based IC50 values show A2780cis and A2780ADR have higher drug resistance than parental A2780 cells. Graphs used to calculate IC50 values from cell number data at 120 hours. Cells treated with cisplatin (a-c) or Adriamycin (d-f). a, d) A2780; b, e) A2780cis and c, f) A2780ADR. GraphPad Prism was used to fit non-linear regression to the data and calculate IC50 values. Error bars are SEM generated from 6 (Adriamycin) or 8 (cisplatin) technical replicates at each drug concentration.

Discussion

In this report we have demonstrated that A2780cis and A2780ADR are ten-fold resistant to

their respective therapeutic agents. These results are consistent with the originally published

data describing these cell lines (Behrens et al. 1984; Behrens et al. 1987).

The use of the IncuCyte ZOOM enabled the automation of imaging and data processing.

The label-free approach relied on the assumption that the reduced cell division and growth

caused by a drug resulted in fewer cells at a given time point, either through reduced

proliferation or increased cell death, ultimately leading to a lower confluence. In these

analyses an important consideration is that changes in cell morphology can also affect

confluence measurements. Importantly this did not appear to be a significant factor in this

study as demonstrated by the similarity in data distribution generated using the two methods

(compare Fig3 & Fig4). The cell count is not without its limitations, however, due to the

addition of a nuclear label it was only completed at a single time point and the debris created

by cell death required two processing definitions to accurately map cell number. Qualitative

inspection of the fluorescence-based cell counting map, when overlaid on the phase images,

suggested a small number of miscounted cells. This was consistent throughout the dataset

and could not be improved through modification of the processing definitions. Despite the

caveats of confluence and cell count methodologies both produced strikingly similar IC50

values. This agreement between the two approaches strengthens the validity of the

experimental methods.

The initial publications describing the development of A2780cis and A2780ADR used agar

based clonogenic assays that took between 14 and 21 days (Behrens et al. 1987; Behrens

et al. 1984; Louie et al. 1985; Ozols et al. 1980). Similar studies looking at A2780 have

previously used agar based clonogenic assays grown for 7-10 days (Masuda et al. 1988) or

even fixed and stained cells after 96 hours (Sasaki et al. 1991). Despite the difference in

experimental duration, the results generated in this study are highly consistent with the

originally published literature (Table1).

This study has demonstrated that A2780 and its drug resistant counterparts, A2780cis and

A2780ADR, have the same character reported over 20 years ago. It shows the banking

procedures used in ECACC have maintained the cellular phenotype ensuring the cells are fit

for purpose. Furthermore, this study has demonstrated a label-free, fast and convenient way

to determine IC50 values using the IncuCyte ZOOM.

Table 1. IC50 values for A2780, A2780cis and A2780ADR from experimental data and published

studies

Drug Method Time (hr)

IC50 values

A2780 (µM) A2780cis (µM) A2780adr (µM)

Experimental IC50

Cisplatin

Cell number 120 0.736 7.044 1.009

Confluence 72 0.438 7.812 0.997

Range : 0.4 - 0.7 7 - 8 1

Published IC50

Cisplatin

(Behrens et al., 1987) 1.1 8 -

(Sasaki et al., 1990) 0.303 - 1.364

(Masuda et al., 1988) 1.02 38.5 -

Range: 0.3-1.1 8-39 1.4

Experimental IC50

Adriamycin

Cell number 120 0.007 0.016 0.070

Cell number (DMSO normalised) 120 0.005 0.012 0.060

Confluence 72 0.004 0.023 0.071

Confluence (DMSO normalised) 72 0.004 0.010 0.086

Range : 0.004 - 0.007 0.01 - 0.02 0.06 – 0.9

Published IC50

Adriamycin

(Louie et al., 1985) 0.01 - 1.6

(Sasaki et al., 1990) 0.11 - 710

(Hamilton et al., 1984) 0.015 - 2.25

Range: 0.01-0.1 - 1.6-710

References

Behrens, B. C., T. C. Hamilton, H. Masuda, K. R. Grotzinger, J. Whang-Peng, K. G. Louie, T.

Knutsen, W. M. McKoy, R. C. Young, and R. F. Ozols. 1987. 'Characterization of a

cis-diamminedichloroplatinum(II)-resistant human ovarian cancer cell line and its use

in evaluation of platinum analogues', Cancer Res, 47: 414-8.

Behrens, B. C., K. G. Louie, T. C. Hamilton, G. Curt, T. Kinsetla, R. C. Young, and R. F.

Ozols. 1984. 'Resistance and cross-resistance of human ovarian cancer cell lines to

Adriamycin, melphalan, and irradiation', Proc. Am. Assoc. Cancer Res, 25: 336.

Fan, C. W., H. A. Fan, S. H. Hsu, C. C. Chan, S. Y. Chen, Y. H. Hsu, and E. C. Chan. 2004.

'An in vitro short time-high dose drug exposure assay for predicting 5FU-resistance

of colorectal cancer', Cancer Lett, 214: 181-8.

Friedman, H. M., and D. L. Glaubiger. 1982. 'Assessment of in vitro drug sensitivity of

human tumor cells using [3H]thymidine incorporation in a modified human tumor

stem cell assay', Cancer Res, 42: 4683-9.

Hamburger, A., and S. E. Salmon. 1977. 'Primary bioassay of human myeloma stem cells', J

Clin Invest, 60: 846-54.

Louie, K. G., B. C. Behrens, T. J. Kinsella, T. C. Hamilton, K. R. Grotzinger, W. M. McKoy,

M. A. Winker, and R. F. Ozols. 1985. 'Radiation survival parameters of antineoplastic

drug-sensitive and -resistant human ovarian cancer cell lines and their modification

by buthionine sulfoximine', Cancer Res, 45: 2110-5.

Masuda, H., R. F. Ozols, G. M. Lai, A. Fojo, M. Rothenberg, and T. C. Hamilton. 1988.

'Increased DNA repair as a mechanism of acquired resistance to cis-

diamminedichloroplatinum (II) in human ovarian cancer cell lines', Cancer Res, 48:

5713-6.

Ozols, R. F., J. K. Willson, K. R. Grotzinger, and R. C. Young. 1980. 'Cloning of human

ovarian cancer cells in soft agar from malignant and peritoneal washings', Cancer

Res, 40: 2743-7.

Puck, T. T., and P. I. Marcus. 1956. 'Action of x-rays on mammalian cells', J Exp Med, 103:

653-66.

Sasaki, H., K. Takada, Y. Terashima, H. Ekimoto, K. Takahashi, T. Tsuruo, and M.

Fukushima. 1991. 'Human ovarian cancer cell lines resistant to cisplatin, doxorubicin,

and L-phenylalanine mustard are sensitive to delta 7-prostaglandin A1 and delta 12-

prostaglandin J2', Gynecol Oncol, 41: 36-40.

Sevin, B. U., Z. L. Peng, J. P. Perras, P. Ganjei, M. Penalver, and H. E. Averette. 1988.

'Application of an ATP-bioluminescence assay in human tumor chemosensitivity

testing', Gynecol Oncol, 31: 191-204.

Tardito, S., C. Isella, E. Medico, L. Marchio, E. Bevilacqua, M. Hatzoglou, O. Bussolati, and

R. Franchi-Gazzola. 2009. 'The thioxotriazole copper(II) complex A0 induces

endoplasmic reticulum stress and paraptotic death in human cancer cells', J Biol

Chem, 284: 24306-19.

Weisenthal, L. M., J. A. Marsden, P. L. Dill, and C. K. Macaluso. 1983. 'A novel dye

exclusion method for testing in vitro chemosensitivity of human tumors', Cancer Res,

43: 749-57.

Yamaue, H., H. Tanimura, M. Nakamori, K. Noguchi, M. Iwahashi, M. Tani, T. Hotta, K.

Murakami, and K. Ishimoto. 1996. 'Clinical evaluation of chemosensitivity testing for

patients with colorectal cancer using MTT assay', Dis Colon Rectum, 39: 416-22.