TubulinTracker, a Novel In Vitro Reporter Assay to Study Intracellular Microtubule Dynamics, Cell Cycle Progression, and Aneugenicity

Abstract

Aneuploidy is characterized by the presence of an abnormal number of chromosomes and is a common hallmark of cancer. However, exposure to aneugenic compounds does not necessarily lead to cancer. Aneugenic compounds are mainly identified using the in vitro micronucleus assay but this assay cannot standardly discriminate between aneugens and clastogens and cannot be used to identify the exact mode-of-action (MOA) of aneugens; tubulin stabilization, tubulin destabilization, or inhibition of mitotic kinases. To improve the classification of aneugenic substances and determine their MOA, we developed and validated the TubulinTracker assay that uses a green fluorescent protein-tagged tubulin reporter cell line to study microtubule stability using flow cytometry. Combining the assay with a DNA stain also enables cell cycle analysis. Substances whose exposure resulted in an accumulation of cells in G2/M phase, combined with increased or decreased tubulin levels, were classified as tubulin poisons. All known tubulin poisons included were classified correctly. Moreover, we correctly classified compounds, including aneugens that did not affect microtubule levels. However, the MOA of aneugens not affecting tubulin stability, such as Aurora kinase inhibitors, could not be identified. Here, we show that the TubulinTracker assay can be used to classify microtubule stabilizing and destabilizing compounds in living cells. This insight into the MOA of aneugenic agents is important, eg, to support a weight-of-evidence approach for risk assessment, and the classification as an aneugen as opposed to a clastogen or mutagen, has a big impact on the assessment.

Aneuploidy is defined as the presence of an abnormal number of chromosomes and is often caused by improper chromosomal segregation during mitosis (Lynch et al., 2019). Aneuploidy is considered a hallmark of cancer. Although this aberration is common and might repress cell proliferation and induce genome instability (Lynch et al., 2019), aneuploidy is not considered to be a universal initiator or promotor of cancer (Ben-David and Amon, 2020; Tweats et al., 2019). Other hereditary conditions with an aneugenic phenotype are, eg, Patau syndrome, Edwards’ syndrome, Down syndrome, and Turner syndrome.

Substances that can induce aneuploidy are called aneugens. Exposure to these substances often accumulates cells in the G2/M phase of the cell cycle (Brandsma et al., 2020; Manchado et al., 2012) and increases the presence of aneuploid cells. The main difference with clastogenic agents, substances that cause DNA breaks, is that those generally do not arrest cells in G2/M phase after short exposure or cause significant aneuploidy. Most aneugens, such as taxol and vinblastine (Kirkland et al., 2016), are not carcinogenic and aneugens that are carcinogenic, often have different genotoxic and nongenotoxic mechanisms that explain the carcinogenic phenotype (Tweats et al., 2019). Moreover, many aneugens are actually used to treat cancers.

The chemical mechanisms causing aneuploidy can be related to tubulin binding or inhibition of mitotic kinases (Lynch et al., 2019). Microtubules consist of α- and β-tubulin heterodimers and are important for maintaining cell morphology and migration (Lynch et al., 2019). During mitosis, microtubules are responsible for the formation and stability of the mitotic spindle and chromosomal segregation (Meunier and Vernos, 2012). Microtubule poisons impair tubulin dynamics and prevent correct formation of the bipolar spindle (Manchado et al., 2012). Some tubulin poisons, such as vinblastine, bind the interface of the α- and β-tubulin and destabilize microtubules (Lynch et al., 2019). Destabilization inhibits microtubule polymerization and can lead to complications during spindle formation and function (Chen and Horwitz, 2002; Manchado et al., 2012). Stabilizing microtubule poisons enhance polymerization (Manchado et al., 2012). Mitosis is then delayed by the spindle assembly checkpoint (SAC) to allow time to remove errors. This delay makes cells more susceptible to cell death or leads to mitotic slippage (Manchado et al., 2012). During the latter, cells can escape prolonged mitotic arrest instead of going into apoptosis, which leads to incorrect chromosome segregation and aneuploidy (Ohashi, 2016; Sinha et al., 2019). Mitotic kinases whose inhibition can lead to aneuploidy include the family of Aurora kinases (Lynch et al., 2019). These are important for proper alignment and segregation of chromosomes during mitosis and meiosis but do not affect microtubules (Goldenson and Crispino, 2015; Lynch et al., 2019).

Currently, there are several approaches to identify aneugenic and clastogenic compounds such as the chromosomal aberration (CA) test and in vitro micronucleus test (MN). Clastogenic agents disrupt DNA integrity and can cause chromosomal breakage, leading to CAs and micronuclei. More recently, the ToxTracker assay (Hendriks et al., 2012, 2016) was introduced to provide mechanistic insight into (geno)toxic substances, followed by the introduction of the Multiflow assay (Bryce et al., 2016) and ToxTracker Aneugen Clastogen Evaluation (ACE) (Brandsma et al., 2020) that can discriminate between aneugens and clastogens based on the effect these substances have on reporter activation and the cell cycle. Although classifying substances as an aneugen or clastogen adds value to risk assessment, it is important to provide information on the specific mode-of-action (MOA) as different key events (KEs) can lead to the same adverse outcome (Lynch et al., 2019; Sasaki et al., 2020). Insight in the MOA of aneugens provides more information on the KEs that precede aneuploidy and enables the classification of substances as tubulin poisons or Aurora kinase inhibitors. This helps to build weight-of-evidence (WOE) for an adverse outcome pathway (AOP) leading to aneuploidy and thereby contributes to the improvement of risk assessment of substances. Current follow-up assays that can study the MOA of aneugens assess tubulin polymerization (Mirigian et al., 2013; Stock et al., 2018) or tubulin levels using fluorophore-labeled taxol (Bernacki et al., 2019).

Here, we describe TubulinTracker, a novel assay to assess aneugenicity and microtubule stability after exposure to aneugenic compounds to study their MOA. TubulinTracker uses a green fluorescent protein (GFP)-tubulin reporter cell line to study microtubule levels using flow cytometry and perform cell cycle analysis. The assay improves risk assessment of aneugenic substances by classifying tubulin-affecting substances.

MATERIALS AND METHODS

Generation of GFP-tubulin reporter cells

GFP-tubulin cells were generated by BAC recombineering in C57/Bl6 B4418 mouse embryonic stem (mES) cells as described previously (Hendriks et al., 2012; Poser et al., 2008). Briefly, the Mitocheck BAC finder (https://www.mitocheck.org, last accessed February 2, 2022) was used to select bacterial strains containing a mouse Tuba1c-BAC and these were ordered from Thermo-Scientific (clone number RP23-43E11). The tubulin gene was modified at the N-terminus with a GFP using the Quick and easy BAC modification kit (Gene bridges). The RecE and RecT recombination enzymes-containing pRed/ET plasmid was transformed into electro-competent BAC strains. These strains were grown on L-arabinose for 30 min at 37°C to express the enzymes. Next, they were electroporated with a GFP-ires-neomycin/kanamycin selection cassette, grown at 37°C for 2 h to allow for recombination and plated on kanamycin selection plates. Modified BACs were isolated using the Nucleobond Xtra Midi DNA isolation kit (Macherey Nagel) and individual clones were analyzed for integration of the GFP cassette by PCR. Next, mES cells were transfected with modified BACs using Lipofectamine 2000 (Invitrogen). To do so, mES cells were seeded on gelatin-coated culture dishes 24 h before transfection. Finally, monoclonal cell lines were selected based on the level of GFP-tubulin using flow cytometry and the localization using an iRiS Digital microscope (Model: I10999, Logos Biosystems).

Cell culture and compound treatment

C57/Bl6 B4418 mES GFP-Tubulin cells were cultured in mES knock-out medium (Gibco) supplemented with 10% fetal calf serum (FCS, Thermo Scientific), 1 mM sodium pyruvate, 2 mM GlutaMax, 1× minimal essential medium nonessential amino acids, 100 mM β-mercaptoethanol, and leukemia inhibitory factor. Cells were propagated on irradiated primary mouse embryonic fibroblasts as feeders, as described previously (Hendriks et al., 2012), and maintained at 37°C and 5% CO₂. For exposure, mES cells were seeded on gelatin-coated 96-well plates at 50 000 cells/well 24 h prior to treatment and maintained at 37°C and 5% CO₂. For 24 h treatment, 30 000 cells/well were seeded. Medium was aspirated 24 h after seeding, and fresh complete cell culture medium was added containing the test chemicals (diluted 1:100 in the medium). mES cells were exposed to specified chemicals for the indicated exposure times at 37°C and 5% CO₂.

The following compounds were tested: taxol (CAS no.: 33069-62-4), epothilone A (CAS no.: 152044-53-6), ixabepilone (CAS no.: 219989-84-1), nocodazole (CAS no.: 31430-18-9), vinblastine sulfate (CAS no.: 143-67-9), vindesine (CAS no.: 59917-39-4), colcemid (Roche, 10295892001, CAS no.: 477-30-5), carbendazim (CAS no.: 10605-21-7), griseofulvin (CAS no.: 126-07-8), vinorelbine (CAS no: 125317-39-7), rigosertib (CAS no.: 592542-60-4), AMG900 (Selleckchem cat no.: S2719), hesperadin (Selleckchem, cat no.: S1529), VX680 (Selleckchem, CAS no.: 639089-54-6), CHR-6494 (CAS no.: 1333377-65-3), methyl methanesulfonate (MMS, CAS no.: 66-27-3), cisplatin (CAS no.: 15663-27-1), tunicamycin (CAS no.: 11089-65-9), 4-nitrophenol (CAS no.: 100-02-7), econazole (CAS no.: 27220-47-9), lovastatin (CAS no.: 75330-75-5), volasertib (Selleckchem, cat no.: S2235), sucrose (CAS no.: 57-50-1), and cytochalasin B from Drechslera dematioidea (CAS no.: 14930-96-2). All compounds were purchased from Merck/Sigma-Aldrich unless indicated otherwise. Cisplatin was dissolved in Phosphate-buffered saline (PBS, Gibco), sucrose was dissolved in milliQ water, and all other compounds were dissolved in dimethyl sulfoxide (DMSO, Sigma-Aldrich).

Cell cycle analysis, aneuploidy, and microtubule levels

Cells were prepared for flow cytometry analysis as described previously (Brandsma et al., 2020). Briefly, a DNA staining was applied 30 min before the end of chemical exposure to determine the cell cycle profile. Therefore, Hoechst-33342 (Sigma-Aldrich) was added in medium to a final concentration of 10 µg/ml for 30 min at 37°C, and cells were protected against light from this point onward. Next, cells were washed twice with PBS, dissociated using trypsin, and resuspended in 2% FCS in PBS. Cells were collected by centrifugation, incubated for 10 min in extraction buffer, and fixed for 10 min in 0.5% glutaraldehyde (CAS no.: 111-30-8, Sigma-Aldrich). After quenching the fixative with sodium borohydride (CAS no.: 16940-66-2, Sigma-Aldrich), cells were collected by centrifugation and collected in PBS for analysis on the MACSQuant X flow cytometer (Miltenyi Biotec) to determine the cell cycle profile (DNA content), percentage of aneuploidy (DNA content), and GFP-tubulin intensity. Additionally, cell numbers were analyzed to exclude measurements with high cytotoxicity, generally only observed after 24 h of exposure (data not shown).

Data analysis

Cell cycle profiles, aneuploidy, and microtubule levels (GFP signal) were determined using the MACSQuant X and analyzed using FlowLogic software (Inivai Technologies). GFP-tubulin signal remaining after extraction was determined by analyzing the average GFP in all cells or in the indicated cell cycle phase. For the fold change GFP-tubulin signal in G2/M phase, GFP-tubulin signal in G2/M cells of 3 independent experiments was used to calculate the average signal and levels of treated samples were normalized to the vehicle control to assess fold change and shown with the SEM. The percentage of aneuploid cells was determined by quantifying the percentage of cells that had a DNA content of more than 4n using FlowLogic. The average of 3 independent experiments was shown with the SEM and a substance was considered an aneugen when the percentage exceeds 4%. The 4% threshold for aneuploidy has been established previously (Brandsma et al., 2020) and was based on the average percentage of aneuploid cells (2.44%, n = 80) plus 2 SDs (0.79%). In the current dataset, the average value was 2.95% for DMSO with an SD of 0.45%. Cell cycle profiles were extracted from FlowLogic. For the cell cycle analysis, the percentage of cells in the G2/M phase was quantified and the average of 3 independent experiments was shown with the SEM. To estimate the area under the curve (AUC), the exposure doses and the fold change GFP-signal were normalized and used to calculate the AUC. An increased AUC above 2.0 indicates microtubule stabilization, and a reduced AUC below 0.65 indicates microtubule destabilization (see Results section for definition of threshold).

To determine whether significant changes were observed between treated samples and the vehicle control, a one-way ANOVA was used (GraphPad PRISM 9).

Microscopy images

For imaging, the GFP-tubulin mES cells were seeded on fibronectin-coated glass coverslips and cells were exposed as described above 24 h after seeding. After 3.5 h, the DNA stain Hoechst-33342 was added and cells were incubated for 30 min at 37°C. Cells were incubated in extraction buffer for 10 min, after which cells were fixed for 10 min in 0.5% glutaraldehyde. Fixation was quenched using sodium borohydride and cells were stored in PBS until imaging. Cells were imaged using an iRiS Digital microscope (Model: I10999, Logos Biosystems).

RESULTS

To be able to measure tubulin stability in a high-throughput manner, we created a GFP-tubulin reporter cell line using BAC recombination in mES cells. These cells were used to study physiologically relevant GFP-tubulin levels in response to different types of tubulin stabilizing and destabilizing compounds, Aurora kinase inhibitors, and compounds that did not affect tubulin as outlined in Figure 1.

To visualize the subcellular localization of GFP-tubulin, we assessed the GFP-tubulin signal using a fluorescent microscope (Figure 2). Vehicle-treated cells were stained with Hoechst-33342 to visualize the nuclei. In DMSO-treated mES cells, GFP-tubulin was localized in the cytoplasm where it forms the microtubules. Next, we exposed the cells to the tubulin stabilizing substance taxol or tubulin destabilizing substance vinblastine (Kirkland et al., 2016), stained the DNA and assessed the localization of GFP-tubulin (Figure 2B). Treatment with both tubulin-affecting compounds resulted in an increase in the number of cells that are in mitosis, as indicated by the arrows. Exposure of the cells to taxol resulted in increased GFP-tubulin levels, indicating that microtubules were stabilized. Exposure of the cells to vinblastine resulted in decreased GFP-tubulin levels, indicative of microtubule destabilization.

Next, we studied the GFP-tubulin signal in the different phases of the cell cycle, distinguished based on the DNA stain intensity, to determine if the cell cycle phase influences the tubulin levels (Figure 2C) in DMSO-treated cells. The GFP-tubulin intensity after extraction was lowest in G1 phase and increased during S phase to a maximum intensity in G2/M phase (Figure 2C), possibly due to an increase in the abundance of microtubules during the different cell cycle phases. Mitotic arrest leads to an increase in the number of cells in G2/M phase (Manchado et al., 2012) and can lead to an increase in the GFP-tubulin signal, due to either microtubule stabilization or inhibition of mitotic kinases. To be able to distinguish the direct effect on tubulin levels from the effect of mitotic arrest, only the GFP-tubulin intensity from cells in G2/M phase was quantified in experiments described below.

To determine the effect of aneugens and a clastogen on mitotic progression and microtubule levels after different exposure times, we studied the effect of taxol, vinblastine, AMG900, and cisplatin using flow cytometry (Figure 3). The Aurora A/B/C kinase inhibitor AMG900 is an aneugen but does not affect microtubules (Payton et al., 2010). The clastogenic DNA cross-linker cisplatin does not affect tubulin levels and was included as negative control (Kirkland et al., 2016). We visualized the cell cycle using Hoechst-33342 and the GFP-tubulin levels in microtubules from G2/M phase cells after 30 min, 1, 2, 4, and 6 h of exposure (Figure 3).

Exposure to the tubulin stabilizing aneugen taxol induced a time-dependent increase in cells that accumulate in the G2/M phase. The accumulation could be detected after 1 h of exposure and was more apparent after 4 h of exposure (Figs. 3A and 3B). Exposure to taxol resulted in an increase in GFP-tubulin levels over time, with the strongest effect observed for the highest treatment concentration (Figure 3C). The microtubule levels were further visualized using the AUC to determine the change in GFP-tubulin over time (Figure 3D). Exposure to taxol already stabilized the microtubules after 30 min and microtubules were further stabilized over time with a maximum stabilization after 4 h (Figs. 3C and 3D). Exposure to the tubulin destabilizing aneugen vinblastine also induced a time-dependent accumulation of cells in the G2/M phase of the cell cycle that was detectable after exposure for 2 h (Figure 3B). The GFP-tubulin signal was already strongly decreased after 30 min of exposure, indicating that microtubules were destabilized (Figure 3D). A stronger destabilizing effect was observed with increasing concentrations, although no further destabilization was observed above 62.5 nM (Figure 3C). Exposure to the Aurora kinase inhibitor AMG900 for 4 or 6 h induced an accumulation of cells in the G2/M phase but did not result in changes in the GFP-tubulin levels, indicating that microtubule stability is unaffected (Figs. 3C and 3D). Exposure to cisplatin did not affect the cell cycle. At the highest treatment concentrations and after 4 and 6 h, an increase in GFP-tubulin signal was observed, but the AUC for the highest treatment was 1.43 after 4 h and 1.44 after 6 h. Together, this suggests that changes in microtubule stability can already be detected after short exposure times but that the arrest in mitosis can best be studied after exposure for 4 h.

To further validate the TubulinTracker assay, a panel of 24 well-established reference substances was tested. The compound library contained 3 tubulin stabilizing substances, 8 destabilizing substances, and 13 substances that have no known effect on microtubules. The anticipated MOA and concentration ranges tested are summarized in Table 1. The compounds that do not affect microtubule levels have varying MOA. Three Aurora kinase inhibitors AMG900, hesperadin, and VX680 were included in the validation. Other compounds are known to induce protein stress, induce DNA or alkylating damage, inhibit actin microfilaments, or inhibit processes independent of tubulin, but are not known to be aneugens, except volasertib. The GFP-tubulin intensity was measured after exposure for 4 h to 10 concentrations in serial dilution (Figure 4A). The percentage of cells in G2/M phase for each compound is shown in Supplementary Figure 1. The 3 tubulin stabilizing compounds showed a strong increased GFP-tubulin signal after treatment representing microtubule stabilization. As an example, the result is shown for epothilone A (Figure 4B). The 8 tubulin destabilizing substances, such as rigosertib (Figure 4C), clearly showed a decreased GFP-tubulin signal and destabilized microtubules (Figure 4). Finally, the Aurora kinase inhibitors and other compounds, including cytochalasin B (Figure 4D) showed a GFP-tubulin signal of approximately 1 and did not influence microtubule levels (Figure 4A). Of these compounds, only exposure to the tubulin stabilizers, tubulin destabilizers, Aurora kinase inhibitors, CHR-6494, and volasertib resulted in an accumulation of cells in the G2/M phase as indicated by the yellow names and shown in Supplementary Figure 1. All other compounds did not affect the cell cycle progression of treated cells after 4 h.

Based on the results of these 24 substances, thresholds for the AUC for tubulin stabilizing and destabilizing substances were defined. The average AUC for the substances not affecting tubulin stability was 1.02 (SD: 0.18, n = 13). A tubulin destabilizing effect of more than 2 SD equals an AUC of < 0.66. The average AUC for tubulin destabilizing substances was 0.45 (SD: 0.09, n = 8). Two SD above the average AUC for tubulin destabilizers equals 0.62. The threshold for tubulin destabilization was therefore defined as 0.65. Following a similar argumentation for tubulin stabilizers, a tubulin stabilizing effect of more than 2 SD equals an AUC of more than 1.38. The average AUC for stabilizers was 7.07 (SD: 1.68, n = 3). The threshold for tubulin stabilization was defined as an AUC of more than 2.

To confirm whether the tested compounds can induce aneuploidy, we performed a cell cycle analysis after exposure to several known aneugens, a clastogen, and substances with a different MOA. Cell cycle analysis was performed by staining the DNA with Hoechst-33342 and using flow cytometry after exposure for 24 h to the aneugens taxol (Kirkland et al., 2016), ixabepilone (Lee et al., 2001), vinblastine (Kirkland et al., 2016), vinorelbine (Toso and Lindley, 1995), and AMG900 (Payton et al., 2010), the clastogen cisplatin (Kirkland et al., 2016), and to volasertib (Rudolph et al., 2009), lovastatin (Krukemyer and Talbert, 1987), and sucrose (Figure 5 and Table 1). Exposure for 24 h to the aneugens induced aneuploidy (Figs. 5A and 5B). This effect could not be detected after exposure to the clastogen cisplatin or other compounds, except after exposure to volasertib. This Plk1 inhibitor is a tubulin-independent mitotic inhibitor and therefore can induce aneuploidy (Rudolph et al., 2009). The induction of aneuploidy after exposure for 24 h correlates with an impaired progression through mitosis after shorter exposure periods as indicated by an increase in G2/M phase cell after 4 h of exposure (Supplementary Figure 1).

DISCUSSION

For successful safety assessment, it is important to correctly classify aneugenic compounds and to distinguish them from mutagenic and clastogenic substances. Identifying the specific MOA of aneugenic substances helps to build a WOE approach and support a specific AOP, eg, the induction of aneuploidy via the binding of chemicals to tubulin or to the catalytic domain of Aurora kinases (Sasaki et al., 2020). The TubulinTracker assay identifies substances that affect microtubule stability and thereby provides evidence for the first KE in the AOP for aneuploidy induced via the binding of chemicals to tubulin.

Tubulin poisons and Aurora kinase inhibitors are substances that have been shown to induce genotoxicity via an aneugenic MOA (Elhajouji et al., 1998; Gollapudi et al., 2014). The current approaches to identify aneugenic compounds consist of the CA test, the in vitro MN, the Multiflow assay (Lynch et al., 2019), and ToxTracker ACE (Brandsma et al., 2020). The CA test can only be used when the aberrations are assessed after the second round of mitosis to allow for aneuploidy formation. The MN test is one of the standard assays to detect aneuploidy. To discriminate between aneugenic and clastogenic compounds, the MN test can be combined with a kinetochore staining or fluorescent in situ hybridization staining to stain the centromeres and determine if the MN is the result of whole chromosome loss or fragment loss (Eastmond and Tucker, 1989a,b; Kirsch-Volders et al., 2003, 2011). The MN assay can also be combined with yH2Ax staining in high-content imaging to discriminate between clastogens and aneugens (Takeiri et al., 2019). The Multiflow assay uses several biomarkers to assess the aneugenicity of substances (Bryce et al., 2016). Finally, the ToxTracker ACE assay uses GFP-reporter activation in combination with cell cycle analysis to discriminate between aneugenic and clastogenic compounds (Brandsma et al., 2020).

To identify the exact MOA of aneugens and build WOE, follow-up assays are necessary to classify compounds as tubulin poisons or Aurora kinase inhibitors. For example, tubulin polymerization can be assessed using purified proteins (Mirigian et al., 2013) but this does not properly represent the effects of substances within the cellular context. Alternatively, tubulin polymerization can be studied using an antibody staining (Stock et al., 2018). This assay using an antibody needs fixation and staining of tubulin, making it less suitable for studying tubulin using live-cell imaging. Finally, tubulin levels can be studied by using fluorescently labeled taxol (Bernacki et al., 2019). Taxol has the downside that it arrests cells in mitosis and prevents analysis of the cell cycle.

Although the TubulinTracker assay is specifically developed to identify microtubule poisons, understanding the MOA of other kinase inhibitors such as Aurora kinase inhibitors is important for understanding the phenotype of other aneugens in the assay. There are 3 different Aurora kinases; Aurora kinase A, B, and C. Aurora kinase A is primarily involved in coordination of centrosome maturation, formation of the bipolar spindle, and chromosome separation (Goldenson and Crispino, 2015; Willems et al., 2018). Inhibition of Aurora kinase A leads to monopolar spindles and a G2/M arrest (Carpinelli and Moll, 2008; Girdler et al., 2006). Aurora kinase B is involved in chromosome condensation, microtubule attachment to chromosomes, SAC regulation, chromosome separation, and cytokinesis (Goldenson and Crispino, 2015; Willems et al., 2018). Inhibition of Aurora kinase B results in misaligned chromosomes and bi-nucleated cells (Vader et al., 2006). Finally, Aurora kinase C is important during meiosis (Goldenson and Crispino, 2015). In general, Aurora kinase inhibitors do not induce mitotic arrest but rather increase the number of cells in prophase compared with metaphase and anaphase (Ditchfield et al., 2003).

TubulinTracker is a novel assay to identify the MOA of tubulin poisons in the context of a living cell by assessing the effect of substances on microtubule polymerization and the cell cycle. TubulinTracker has the advantage of containing endogenous GFP-tubulin and therefore does not require antibody staining. It only requires extraction of the free tubulin pool to study the effect of substances on microtubules. The endogenous GFP-tubulin can be analyzed by flow cytometry and, including a DNA stain, enables cell cycle analysis. Furthermore, although our cells are fixed in our protocol, they do not necessarily need fixation and this allows direct visualization of microtubules in treated cells using microscopy. This can provide insight in the dynamics of tubulin poisons at a single-cell level. Additionally, the assay can be adapted to determine LC₅₀ values or determine the point of departure for tubulin stabilization of destabilization, depending on the application.

Tubulin destabilizing substances generally show a maximum effect after 3–4 h of exposure, whereas for tubulin destabilizing substances, the effect can be observed after 15 min and the maximum effect can be reached after 1 h (Morrison and Hergenrother, 2012). Here, we observed the effect of tubulin stabilizers and destabilizers already after 30 min (Figure 3C). The maximum effect for vinblastine was observed after 2 h and for taxol after 4 h. Exposure to cisplatin leads to an increase in GFP-tubulin at the higher treatment concentrations after 4 and 6 h in the time course experiment (Figure 3). However, there was no effect on the cell cycle observed up to 6 h, which was the case for all other tubulin poisons.

We have shown that only tubulin poisons specifically affect microtubule stability and that compounds that do not influence microtubule levels were correctly classified as non-tubulin poisons. For example, CHR-6494 and volasertib arrested cells in mitosis but did not change microtubule levels. CHR-6494 is a Haspin kinase inhibitor that is known to induce a G2/M arrest via abnormal spindle formation and defects in chromosome alignment and centrosomes (Amoussou et al., 2018). Volasertib targets Plk1 and prevents bipolar spindle formation leading to an arrest in early M phase (Gjertsen and Schöffski, 2015). Moreover, treatment with cytochalasin B, an inhibitor of actin microfilaments (MacLean-Fletcher and Pollard, 1980), did not change the microtubule levels. The TubulinTracker assay correctly identified all the tested, known tubulin stabilizing and destabilizing compounds and therefore has a high sensitivity.

As a follow-up to the TubulinTracker assay, to identify the MOA of aneugenic substances that arrest cells in G2/M phase but do not affect tubulin stability, additional analysis investigating, eg, the phosphorylation of histone H3 can be performed to identify the cause of aneuploidy (Bernacki et al., 2019).

Here, we have shown that TubulinTracker is a fast, easy, and highly reliable assay that combines cell cycle analysis with studying microtubule stability after compound treatment. This allows detailed MOA assessment of substances that previously tested positive in the in vitro micronucleus assay or were identified as an aneugen in the ToxTracker ACE assay (Brandsma et al., 2020). Classification of aneugens as tubulin poisons provides evidence to support a specific AOP and improves safety assessment of substances.

SUPPLEMENTARY DATA

Supplementary data are available at Toxicological Sciences online.

DECLARATION OF CONFLICTING INTERESTS

M.E.G, N.M., G.Z., R.D., T.O., G.H., and I.B. are employed at Toxys, a Dutch company that offers TubulinTracker as a commercial service.

ACKNOWLEDGMENTS

We would like to thank Hortense Billaud and Carolien Hoff for their contributions to developing the TubulinTracker assay.

REFERENCES