Cell lines

Human fibrosarcoma HT1080 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) (Invitrogen) supplemented with 10% (v/v) tet system-approved fetal bovine serum (Clontech Laboratories, Inc.) at 37°C in 5% CO₂. Hypoxanthine phosphoribosyltransferase (HPRT)-deficient Chinese hamster ovary (CHO) cells (JCRB0218) carrying the alphoid^tetO-HAC were maintained in Ham's F-12 nutrient mixture (Invitrogen) plus 10% FBS with 8 μg/ml of BS (Funakoshi). After loading of the EGFP transgene cassette into the alphoid^tetO-HAC, the CHO cells were cultured in 1× HAT supplemented medium.

Loading of the EGFP transgene cassette into the loxP site of alphoid^tetO-HAC in CHO cells

A total of 3 to 5 μg of a EGFP transgene plasmid (or X3.1-I-EGFP-I described previously [13]) and 1 to 2 μg of the Cre expression pCpG-iCre vector DNA were co-transformed into HPRT-deficient CHO cells containing the alphoid^tetO-HAC by lipofection with FuGENERHD transfection reagent (Roche) or Lipofectamine 2000 (Invitrogen). HPRT-positive colonies were selected after 2 to 3 weeks growth in HAT medium. For each experiment, from 5 to 7 clones were usually selected. Correct loading of the EGFP transgene cassette into the HAC was confirmed by genomic PCR with a specific pair of primers that diagnose reconstitution of the HPRT gene [9].

Microcell-mediated chromosome transfer

The alphoid^tetO-HAC containing the EGFP transgene cassette (EGFP-HAC) was transferred from CHO cells to HT1080 cells using a standard microcell-mediated chromosome transfer (MMCT) protocol [13, 14]. Blasticidin (BS) was used to select resistant colonies containing the HAC. Typically, three to ten BS^R colonies were obtained in one MMCT experiment. BS^R colonies were analyzed by FISH for the presence of the autonomous form of the HAC. The co-transfer of CHO chromosomes was examined using a sensitive PCR test for rodent-specific SINE elements [9].

Flow cytometry

Analysis of EGFP expression was performed on a FACS Calibur instrument (BD Biosciences) using CellQuest acquisition software and analyzed statistically with FlowJo software [15, 16]. The cells were harvested by trypsin-treatment. Intensities of fluorescence were determined by flow cytometry. A minimum of 4 x 10⁴ cells was analyzed for each cell sample.

Drug treatment

Calculation of the rate of HAC loss after drug treatment

The formula, P _n = P ₀(1 − R) ⁿ [17], routinely used to calculate the rate (R) of spontaneous HAC (or chromosome) loss, cannot be applied when cells are treated by a single dose of drug. So in our study, we first determined the normal rate of spontaneous HAC miss-segregation (R_Normal) in the host cell line HT1080 using the formula, $P_{\mathit{normal}}=P_0{\left(\frac{2-R_{\mathit{Normal}}}{2}\right)}^{n_1}$ (Figure 1); where P₀ is the percentage of EGFP(+) cells at the start of the experiment as determined by FACS. These cells were cultured under HAC selection conditions using blaticidin. P_Normal is the percentage EGFP(+) cells after culturing without HAC selection (no blasticidin) for a duration of t_1. In this study t₁ was 14 days. n₁ is the number of cell doublings that occurs during culturing without blasticidin selection. The doubling time of HT1080 under normal growth conditions is approximately 18 hours. The number of cell divisions (n) is calculated by (t / host cell doubling time).

Once (R_Normal) was obtained, the rate of HAC loss induced by drug treatment (R_Drug) is then determined using the formula, $P_{\mathit{Treated}}=P_0{\left(\frac{2-R_{\mathit{Drug}}}{2}\right)}^{n_2}{\left(\frac{2-R_{\mathit{Normal}}}{2}\right)}^{n_3}$. Justification of this algorithm is presented in Figure 1. As before, P₀ represents the percentage of EGFP(+) cells at the start of the experiment, cultured under HAC selection condition. P_Treated is the percentage of EGFP(+) cells at the end of a drug treatment experiment with a duration of (t₂ + t₃), where t₂ is the duration of drug treatment and t₃ is the duration of culturing after the drug is removed. (t₂ + t₃) was 14 days in this study. n₂ is the number of cell doublings that occurs during drug treatment, while n₃ is the number of cell doublings that occurs during the culturing without selection after drug treatment.

In the present study, the duration of most drug treatments were less than the duration of a single cell cycle of HT1080 (t₂ <18 hr). We made the assumption that any significant increase in HAC loss occurs only during the first mitotic division after washing off a drug (n₂ = 1). Thus n₃ = (14 d / 18 hr - 1). The exceptions, reversine and ZM-447439 both severely inhibits cell growth, thus despite the 3 day drug treatment only one cell division is assumed to have occurred. We also assumed that HAC stability in subsequent cell divisions is no different from that of untreated cells.

The algorithm we used is valid between the ranges of 0 to 1. R values large than 1 indicate that the assumptions made in this model are incorrect. The assumption of synchronous growth in the model means that the estimated mis-segregation rate is lower than real values. As the spontaneous rate of HAC mis-segregation (R_Normal) was found to be low, this algorithm is relatively insensitive to the number of cell divisions that occurs post drug treatment. It is worth noting that the frequency of HAC loss in the clones carrying two copies of the HAC is indistinguishable from those containing a single copy of the HAC (data not shown). Therefore, the model is applicable when a cell inherits two copies of the HAC due to non-disjunction.

Cell viability test

MTS tetrazolium cell viability assays were done according to the manufacturer’s instructions (CellTiter 96 AQueous Assay reagent; Promega). Briefly, the CellTiter 96 AQueous One Solution Reagent was added to each well and incubated at 37°C for 3 h. Cell proliferation was determined by measuring the absorbance at 490 nm using a microtiter plate reader (Molecular Devices, Sunnyvale, CA). The half-maximal inhibitory concentration (IC50) was obtained from the MTS viability curves using GraphPad Prism 5. Experiments were carried out in triplicate.

FISH analysis with PNA probes

The presence of the HAC in an autonomous form was confirmed by FISH analysis as previously described [8, 9]. HT1080 cells containing the HAC were grown in DMEM medium to 70-80% confluence. Metaphase cells were obtained by adding colcemid (Gibco) to a final concentration of 0.05 μg/ml and incubating overnight. Media was aspirated, and the plate washed with 1x PBS. Cells were removed from the plate by 0.25% Typsin, washed off with DMEM, pelleted and resuspended in 10 ml of 50 mM KCl hypotonic solution for 30 min at 37°C. Cells were fixed by three washes of fixative solution (75% acetic acid, 25% methanol). Between each wash, cells were pelleted by centrifugation at 900 rpm for 4 min. Metaphase cells were evenly spread on a microscope slide and the fixative solution evaporated over boiling water. Dry slides were rehydrated with 1× PBS for 15 min, and fixed in 4% formaldehyde-1× PBS for 2 min, followed by three 5 min 1× PBS washes and ethanol series dehydration. PNA (peptide nucleic acid) labeled probes used were telomere (CCCTAA)₃-Cy3) (PerSeptive Biosystems, Inc.) and tetO-alphoid array (FITC-OO-ACCACTCCCTATCAG) (Panagene, South Korea). Ten nanomol of each PNA probe was mixed with hybridization buffer and applied to the slide, followed by denaturation at 80°C for 3 min. Slides were hybridized for 2 hours at room temperature in the dark. Slides were washed twice in 70% formamide, 10 mM Tris pH 7.2, 0.1% BSA and followed by three washes with 1xTBS, 0.08% Tween-20. Slides were dehydrated gradually with a series of 70%, 90% and 100% ethanol washes and mounted (Vectorshield with DAPI). Images were captured using a Zeiss Microscope (Axiophot) equipped with a cooled-charge-coupled device (CCD) camera (Cool SNAP HQ, Photometric) and analyzed by IP lab software (Signal Analytics). The PNA-DNA hybrid probes demonstrated a high hybridization efficiency, staining intensity and adopt a stable duplex form with complementary nucleic acid.

FISH analysis with the BAC probe

HT1080 cells were processed for fluorescence in situ hybridization (FISH) after drug treatment followed by the 14 day washout. The probe used for FISH was BAC32-2-mer(tetO) DNA containing 40 kb of alphoid-tetO array cloned into a BAC vector as described previously [8]. Specifically, a BAC32-2-mer(tetO) clone contains an amplified synthetic alphoid DNA dimer. One monomer of this dimer is an alphoid DNA consensus sequence carrying the tetO sequence; another monomer is alphoid DNA from chromosome 17. This probe is specific to the HAC but also gives a low signal with centromeric regions of several endogenous chromosomes. BAC DNA was digoxigenin-labeled using a nick-translation kit with digoxigenin-11dUTP or biotin16-dUTP (Roche Diagnostics). Images were captured as before. The probe was denatured for 5 min at 95°C and added to the slides, which were incubated at 72°C for 2 min before overnight incubation at 39°C. After washes with 0.1 × SSC at 65°C followed by a wash with 4 × SSC + 0.1% Tween 20 at room temperature, standard procedures were used to detect biotinylated probes. Slides were mounted with VectaShield and screened for the presence or absence of the HAC. Between 70–250 metaphases were analyzed for each experiment (one drug treatment).

Genomic DNA preparation and PCR analysis

Genomic DNA for PCR analysis was prepared using a QIAmp DNA Mini Kit (QIAGEN Inc., Valencia, CA, USA). Reconstitution of the HPRT gene after Cre/lox-mediated recombination was determined by specific primers, Lox137-R 5′-agccttctgtacacatttcttctc-3′ and Rev #65′-gctctactaagcagatggccacagaactag-3′. Cross contamination by hamster chromosomes was determined by specific primers detecting hamster SINEs: Cons B2 5′-ccatctgtaatgagatctgatgc-3′, Ham B2-F 5′-gctcagaggttaagagcactgac-3′ and Ham B2-R 5′-tgcttccatgtatatctgcacac-3′. PCR products for sequencing were separated by agarose gel electrophoresis and gel extracted.

Micronucleus formation assay (MNi)

Experimental design for identification of drugs that elevate CIN in cancer cells

Figure 2 shows a general scheme of the new assay developed for measuring chromosome instability (CIN) based on the use of a human artificial chromosome (HAC) carrying the EGFP (enhanced green fluorescence protein) transgene. Due to the presence of a functional kinetochore, the EGFP-HAC is maintained as a non-essential 47th chromosome that replicates and segregates like a normal chromosome in human cells. Thus, the cells that inherit the HAC display green fluorescence, while cells that lack it do not. Normally, after growing in non-selective medium (i.e. in the absence of selection for the HAC), the majority of cells contain one copy of the HAC. After drug treatment, there are two possibilities: either no change in the EGFP level (no response to the drug) or an increased percentage of cells without HAC due to HAC segregation or replication errors (response to the drug treatment). It is expected that the control untreated cells should show uniform green fluorescence, while those that have lost HAC after drug treatment will exhibit reduced fluorescence that can be detected by flow cytometry.

Construction of the HAC carrying a single copy of the EGFPtransgene and its transfer to human cells

The EGFP transgene cassette (X3.1-I-EGFP-I) was described previously [13]. To adapt the alphoid^tetO-HAC [8] for CIN studies, the EGFP transgene cassette was inserted into a single loxP loading site of the HAC [13] in hamster HPRT-deficient Chinese Hamster Ovary (CHO) cells (Figure 3A, B, 2B). In this cassette, EGFP is flanked by the cHS4 insulator sequences to protect the transgene from epigenetic silencing. Targeting of the EGFP-cHS4 cassette into the loxP site was accompanied by reconstitution of the HPRT gene, allowing cell selection on HAT medium. Therefore, recombinant clones were selected by growth in HAT medium after two-three weeks. PCR analysis with specific primers (see Methods) confirmed that the HPRT gene was indeed reconstituted in all five drug-resistant clones analyzed (data not shown). The efficiency of cassette targeting into the HAC was <1-3 × 10⁻⁴. All of the HPRT⁺ transfectants expressed the EGFP transgene, which was detected by fluorescence microscopy. FISH analysis of CHO metaphase spreads revealed the HAC in an autonomous form in four clones (data not shown). One clone containing the HAC was chosen for further experiments.

Effect of aneugens and clastogens on the rate of HAC mis-segregation during mitotic divisions

We next investigated whether the EGFP/HAC-based assay could be used to detect compounds that cause chromosome loss and mis-segregation. HT1080 cells with an autonomously propagated EGFP-HAC were treated with eight known aneugens: taxol (pacilitaxel), docetaxel, peloruside A, ixabepilone, nocodazole, ZM447439, reversine, and SAHA (Table 1). Taxol, docetaxel, peloruside A, ixabepilone, and nocodazole are microtubule-targeting drugs [19]; ZM447439 and reversine are inhibitors of Aurora B and MPS1, respectively, and function in the spindle assembly checkpoint [20, 21]. SAHA is an inhibitor of histone deacetylases (HDAC) [22]. Cells were also treated with the well-known clastogenic DNA topoisomerase II inhibitor VP16 (etoposide) [23].

For each compound, a cell cytotoxicity assay was carried out to determine IC50 values, i.e., the conditions under which the viability of cells would be around 50%. We chose this parameter in order to normalize the results at the same percentage of viable cells. The evaluation of chromosome instability at the IC50 values has been used in many studies involving micronucleus scoring [24, 25]. Time treatments and drug concentrations corresponding to IC50 are summarized in Table 1. For all compounds, the IC50 values were in the range clinically relevant concentrations, where applicable. After treatment, the cells were grown for two weeks in the absence of selection. Samples were analyzed every few days, and the proportion of non-fluorescent cells was determined. Non-fluorescent cells could be detected after a few days, and treated and untreated cell populations were clearly distinguishable after 5–7 days. The delay between HAC loss and the appearance of non-fluorescent cells is due to the persistence of the EGFP protein, which has a quite high half-life. Based on our analysis, sampling time has a broad interval (5–14 days) without a significant effect on the calculated rate of HAC loss (see Additional file 1). In most experiments, the cells were analyzed 14 days after drug washout. Figure 4 shows representative flow cytometry histograms illustrating loss of fluorescence for the cells treated by a single doze of taxol. It is important that the measuring is highly reproducible: The raw FACS data of three independent populations for drug treatments have standard deviations less than 1%.

Figure 5 summarizes the estimated rates of HAC loss in response to different drugs. The rate of HAC loss (R_Drug) induced by drug treatment was calculated from the proportion of non-fluorescent cells in the population (Table 3) using the formula $P_0{\left(\frac{2-R_{\mathit{Drug}}}{2}\right)}^{n_2}{\left(\frac{2-R_{\mathit{Normal}}}{2}\right)}^{n_3}$ (see more details in Methods and Figure 1 legend).

As seen, a single dose of taxol greatly increases (~50 times) the rate of HAC loss. Taxol (pacilitaxel) was isolated ~ 40 years ago and is currently administered in a large variety of indications, including solid tumors and haematological malignancies [19] and references therein]. While its mechanism of action remains controversial, accumulating data suggest that at clinically relevant concentrations, taxol-mediated cell death involves elevation of chromosome mis-segregation that is incompatible with cancer cell survival [26] and references therein]. HAC loss was also significantly increased by peloruside A, ixabepilone, which also binds to the microtubule as taxol, and by the Aurora B inhibitor ZM-447439. For the other analyzed compounds except SAHA, the rates of HAC loss were also increased compared to controls, but much lower compared to those obtained after taxol or peloruside A treatment. The lowest increase of HAC loss was obtained after treatment by the inhibitor of histone deacetylases, SAHA.

In separate experiments, the relationship of HAC loss to chemical dose was investigated. Treatment of cells with higher doses of the compounds kills more cells but does not increase the rate of HAC loss. This may be explained if cells arrest, then apoptosis in response to damage of the spindle or if the daughter cells are not viable. We also performed the experiments with lower doses of drugs, i.e. with concentrations ½, ¼, ¹/₈ and ¹/₁₆ of the concentration corresponding to IC50. For some drugs, the frequency of HAC loss was dose independent while for others displayed a linear dose-dependence. For example, HAC loss was decreased in an almost linear fashion after treatment by VP16 (which is inhibitor of TOP2) with lower doses. In contrast, our analysis revealed a non-linear dose-dependent decrease in HAC loss frequencies for nocodazole and taxol, which were previously reported as compounds that interact with the mitotic spindle (thresholded concentration-effect response) [27].

FISH analysis was used to confirm that the appearance of non-fluorescent cells detected by flow cytometry corresponds to HAC loss events and to exclude possible alternative explanations such as mutations in the EGFP gene or its epigenetic silencing. Quantitative analysis of metaphase chromosome spreads using a HAC-specific probe (see Methods) correlated with the data on HAC loss determined by FACS (Table 3). Therefore, we conclude that the appearance of non-fluorescent cells is caused by HAC loss.

An unexpected result was the different effects of microtubule-binding drugs on HAC stability. Ixabepilone, docetaxel, and peloruside A are microtubule-stabilizing agents similar to taxol [19]. However, each drug exhibited a different effect on HAC mitotic stability, suggesting a different cell response to these compounds. Interestingly, under these conditions, all of the analyzed drugs induced micronuclei formation in the HAC-containing HT1080 cells (Table 2) similar to that reported for other cell lines [27]. However, there was no detectable correlation between frequencies of MNi formation (that varied between 37% and 63%) and rates of HAC loss determined by FISH and FACS (Table 3). Thus, aneugens with a similar mechanism of action and cytotoxicity may differ from each other by their effect on the mitotic stability of the non-essential human artificial chromosome. It is unlikely that karyotyping or another variant of MNi tests can detect such differences between the compounds. Based on these results, we conclude that a newly developed EGFP/HAC-based assay for measuring CIN is a more sensitized system than other previously described methods to detect chromosome mis-segregation.