Treatment of breast and colon cancer cell lines with anti-helmintic benzimidazoles mebendazole or albendazole results in selective apoptotic cell death

Anti-helmintic drugs mebendazole and albendazole are commonly used to treat a variety of parasitic infections. They have recently shown some promising results in pre-clinical in vitro and in vivo anti-cancer studies. We compare their efficacy in breast and colon cancer cell lines as well as in non-cancerous cells and elucidate their mechanism of action. The drugs were screened for cytotoxicity in MDA-MB-231, MCF-7 (breast cancer), HT-29 (colorectal cancer), and mesenchymal stromal cells, using the MTT assay. Their effects on the cell cycle, tubulin levels, and cell death mechanisms were analysed using flow cytometry and fluorescent microscopy. Mebendazole and albendazole were found to selectively kill cancer cells, being most potent in the colorectal cancer cell line HT-29, with both drugs having IC50 values of less than 1 µM at 48 h. Both mebendazole and albendazole induced classical apoptosis characterised by caspase-3 activation, phosphatidylserine exposure, DNA fragmentation, mitochondrial membrane permeability, and reactive oxygen species production. Cell cycle arrest in the G2/M phase was found, and tubulin polymerisation was disrupted. Mebendazole and albendazole were shown to cause selective cancer cell death via a mechanism of classical apoptosis and cell cycle arrest, involving the destabilisation of microtubules.


Introduction
Mebendazole (MBZ) and albendazole (ABZ) are broadspectrum benzimidazole anti-helmintic drugs. They are commonly prescribed to treat a range of parasitic worm infections in humans and domestic animals. These drugs have few adverse effects and low cost (Nygren and Larsson 2013). The anti-helmintics bind to the colchicine-binding domain of tubulin and inhibit microtubule polymerisation, leading to reduced glucose uptake in the gastrointestinal cells of helminths (Russell and Lacey 1995).
MBZ and ABZ have recently been shown to have promising anti-cancer effects in a range of in vitro studies Sasaki et al. 2002;Castro et al. 2016;Pinto et al. 2019;Lai et al. 2017;Zhang et al.2017Zhang et al. , 2019bLiu et al. 2020;Chu et al. 2009), as well as in a small number of xenograft models (Liu et al. 2020;Williamson et al. 2016). MBZ and ABZ have also been shown to act in synergy with ionising radiation in a number of cancer types (Markowitz et al. 2017;Patel et al. 2011;Skibinski et al. 2018;Zhang et al. 2019a), as has MBZ with non-steroidal anti-inflammatory sulindac in colorectal cancer (Williamson et al. 2016).
Here, we compare this information with the cytotoxicity of MBZ and ABZ in breast and colorectal cancer cell lines and in non-cancerous mesenchymal stromal cells, and describe the mechanism of action by measuring cell cycle progression, microtubule changes and markers of apoptosis, necrosis, and oxidative stress.

Cell culture
The three human cancer cell lines were MCF-7 (breast adenocarcinoma), MDA-MB-231 (triple negative breast cancer, kind gifts from Prof. Rhonda Rosengren), and HT-29 (colorectal adenocarcinoma, ATCC). The non-cancerous, human bone marrow-derived mesenchymal stromal-cell (MSC) line RCB2157 (Riken Biosciences) was also used. All cell lines were cultured in Dulbecco's Modified Eagle Medium (DMEM) with 5% heat inactivated FBS for cancer cell lines and 10% for MSCs, 100 units/mL penicillin, and 100 µg/mL streptomycin. Cells were incubated at 37 °C in humidified conditions with 5% CO 2 .

MTT assay
Cells were seeded and allowed to adhere overnight. They were then treated with MBZ, ALB or DMSO as a vehicle control at 0.1% v/v. The quantity of live cells remaining following 24 and 48 h treatment exposure was measured using the MTT assay, which assesses mitochondrial and endoplasmic reticulum dehydrogenase activity (Mosmann 1983). MTT was added to cells at a final concentration of 0.5 mg/mL, and incubated at 37 °C for 3 h. Formazan crystals were dissolved with DMSO, and the resulting absorbance read at 560 nm using the spectrophotometer.

Flow cytometry
Flow cytometry was carried out using a BD FACSCantoII ® flow cytometer and data analysed using FlowJo ® VX software. Cells were treated and after either 24 or 48 h were prepared and stained as outlined below, and then analysed using the appropriate filters of FITC and/or PE.

Phosphatidylserine exposure
Following incubation with treatments for 24 or 48 h, cells and media were harvested, centrifuged at 4 °C, 2000 rpm for 3 min, the pellet washed twice in 4 °C PBS, and the cells resuspended in Annexin V-binding buffer (HEPES; 10.0 mM (pH 7.4), NaCl; 140.0 mM, and CaCl 2 ; 2.5 mM) with propidium iodide (PI) and Annexin V conjugated to FITC. Cells were incubated for 15 min in the dark at room temperature.

Mitochondrial membrane permeability
Cells were treated and harvested, the pellet washed twice in 4 °C PBS, and the cells resuspended in PBS. Tetramethylrhodamine ethyl ester (TMRE) was added to a final concentration of 40 nM, and then incubated at 37 °C for 10 min.

Cell cycle
Following incubation with treatments for 24 or 48 h, media and cells were harvested, centrifuged at 4 °C, 2000 rpm for 3 min, washed twice in 4 °C PBS, and fixed with 70% ethanol at 4 °C for 30 min. Cell were washed twice in phosphatecitrate buffer (Na 2 HPO 4 ; 192 mM, citric acid; 4 mM (pH 7.8)), treated with 100 µg/mL RNAse A and stained with 50 µg/mL PI.

Caspase activation
Following treatment of cells for 24 or 48 h, media and cells were harvested, washed in cold PBS, and frozen at −80 °C overnight. Cells were then resuspended in buffer (100 mM HEPES, 10% sucrose, 0.1% CHAPS, and 10-4% Tergitol NP-40 at pH 7.25), DTT added to 5 mM and Ac-DEVD-AMC caspase substrate added to 50 µM. The plate was incubated in the dark for 2 h and the AMC fluorescence read at excitation 360 nm and emission 460 nm.

Oxidative stress
Oxidative stress levels were analysed by measuring reduced thiol content using the DTNB (5, 5″-dithio-bis-(2-nitrobenzoic acid)) assay. DTNB quantifies free sulfhydryl groups by producing a measurable yellow product (Ellman 1959). Using the molar extinction coefficient of 14,150 M −1 cm −1 , the free sulfhydryl content can be calculated. Cells were treated, incubated at 37 °C for 48 h, harvested and centrifuged at 4 °C, 2000 rpm for 3 min, washed with PBS twice, and sonicated for 2 min. Ten percent of SDS and DTNB at final concentration 30 µM were added and incubated at room temperature for 30 min. Absorbance was measured at 412 nm.

Fluorescent microscopy
For staining with anti-tubulin antibodies, cells were treated with for 8 h. Following fixation, cells were blocked for 60 min in blocking buffer (1X PBS, 5% BSA, 0.3% Triton X-100) and stained with primary antibody, anti--tubulin at 1:16,000, or anti-acetyl--tubulin at 1:800, overnight at 4 °C. After washing, cells were incubated with secondary antibodies and Hoechst 33,342 at 10 µg/mL for 1 h at room temperature in the dark. Prolong Gold was placed on the slides along with coverslips. The cells were analysed using the Nikon AI-R inverted fluorescent microscope with Nikon Intensilight C-HGFI mercury lamp (Olympus, Japan) at 20 × magnification.

Statistical analyses
Results were analysed using GraphPad Prism ® 6.0 software. Cell viability, flow cytometry, and DTNB experiments (N = 3) were analysed using one-way analysis of variance (ANOVA) and Dunnett's post hoc tests with statistical significance at P < 0.05.

MBZ and ABZ lower cancer cell viability, especially in colorectal cancer cell line HT29
The IC50 values of MBZ and ABZ over 48 h were determined in three cancer cell lines MCF-7, MDA-MB-231 (both breast cancer), and HT-29 (colon cancer), as well as in the non-transformed mesenchymal stromal-cell line RCB2157, using MTT cell viability assays ( Fig. 1a-d). The IC 50 values in the mesenchymal stromal-cell line RCB2157 were the highest by several fold (Table 1). The two breast cancer lines were less sensitive to the two drugs than the colorectal cancer cell line HT29.

MBZ and ABZ block the cell cycle at G2/M
Further analyses were carried out in the HT29 cells to determine the mechanism of the loss of cell viability after 24 or 48 h treatment with MBZ or ABZ. Progression through the cell cycle was analysed using fixation and staining with propidium iodide (PI). Paclitaxel, known to cause a block in the cell cycle at G2/M through binding to microtubules, was used as a positive control.
Treatment with 2 µM MBZ or ABZ or 50 nM paclitaxel resulted in significantly large increases in the G2/M phase compared to control (Fig. 1e, f). At 24 h, MBZ was at 60.18%, ABZ at 58.68%, and paclitaxel at 62.42% compared to vehicle-only control at 21.37% (P > 0.05). There were smaller, but still significant increases of cells in the > G2/M phase, and concomitant decreases in the proportion of cells in the G1 and S phases.

MBZ and ABZ induce apoptotic nuclear morphology and depolymerise microtubules
To investigate whether MBZ and ABZ interact with microtubules in a similar way to paclitaxel, HT-29 cells were stained for nucleus and microtubules following treatment with MBZ (2 µM), ABZ (2 µM), paclitaxel (50 nM), or DMSO vehicle control (0.1% v/v) for 8 h. Hoechst 33,342 staining was used for the nucleus, and antibodies against either -tubulin, a main component of the tubulin dimer that polymerises to form microtubules or against acetyl--tubulin, an indicator of stabilised microtubules, were used to visualise microtubule structure (Al-Bassam and Corbett 2012).
MBZ, ABZ, and paclitaxel treatments all produced the condensed nuclei characteristic of apoptosis. -tubulin and acetylated--tubulin both displayed a decreased fluorescent intensity following MBZ and ABZ treatment compared to control. By contrast, paclitaxel increased intensity of bothand acetylated--tubulin staining compared to control. This suggests that MBZ and ABZ promote depolymerisation of microtubules, whereas paclitaxel promotes polymerisation and stabilises microtubules, as expected.

MBZ and ABZ induce caspase-3 activity
Since it appeared that cell death was occurring concurrently with a block in the cell cycle, characteristics associated with various forms of apoptosis were investigated. Caspase-3 activity was measured in HT-29 cells following treatment with MBZ, ABZ, paclitaxel, or vehicle-only control as before (Fig. 2a). Both untreated control and DMSO vehicle control showed low levels of activity. ABZ and MBZ treatment both increased caspase activation around 550-fold at 24 h (P < 0.0001 compared to untreated control), which dropped back to 300-400-fold at 48 h, suggesting that the cell death process had peaked within the first 24 h. Paclitaxel treatment caused the greatest increase caspase activity, up to 750-fold of control at 24 h (P < 0.0001 compared to MBZ, ABZ, and controls), which is not surprising as paclitaxel is a classical apoptosis inducer and activates caspase-3 (Oyaizu et al 1999).

MBZ and ABZ cause early phosphatidylserine exposure
Following both 24 and 48 h treatments, cells were doublestained with fluorescently labelled Annexin V to identify early externalisation of phosphatidylserine residues, and PI to separate out the internal phosphatidylserine exposed to Annexin V by cell membrane permeability, occurring as a result of primary or secondary necrosis (Fig. 2b,c). Following 24 h treatment, MBZ and ABZ caused approximate 25% decreases in the percentage of viable cells (negative for both Annexin V and PI) compared to control (P < 0.0001), similar to that of the positive control paclitaxel (Fig. 2b). MBZ and ABZ caused a significant increase in the percentage of cells with phosphatidylserine exposure, with 22.83% of MBZ treated cells and 24.17% of ABZ treated cells positive for Annexin V only, similar to paclitaxel at 20.72%, whereas only 3.37% of control cells were positive (P < 0.0001). Numbers of double-stained cells undergoing secondary necrosis after treatment with paclitaxel, MBZ, and ABZ all significantly increased to MBZ 5.78%, ABZ 5.99%, and paclitaxel 7.94% versus control 2.12% (P < 0.0001). Necrotic cells, positive for PI only, were increased only in MBZ and ABZ treatments compared to control, 3.58% and 3.48% versus 0.54%, respectively. Forty-eight hour treatments produced similar results to that of 24 h treatments.

MBZ and ABZ treatment results in DNA fragmentation
Another indicator of apoptosis is the fragmentation of DNA content, known as sub-G1 DNA, measured using fixation of cells and PI staining. Here, following both 24 and 48 h treatments with MBZ, ABZ and paclitaxel, the percentage of cells with sub-G1 DNA content was significantly increased compared to control (Fig. 2d). After 24 h, MBZ treatment resulted in 13.84% (P = 0.001) of cells with sub-G1 DNA content, ABZ 11.53% (P = 0.026), paclitaxel 19.33% (P < 0.0001) and control 6.45%. DMSO caused no statistically significant change. Following 48 h treatment, all amounts increased, except in the controls.

MBZ and ABZ increase mitochondrial membrane permeability
HT-29 cells were treated with MBZ, ABZ, paclitaxel, DMSO, or left untreated for 24 and 48 h and the percentage of cells with higher mitochondrial membrane permeability, indicative of apoptosis, determined using the mitochondrialtargeted stain TMRE.
In the 24 h treatments, there were no significant differences between DMSO vehicle control (3.49 ± 0.38%) and control (3.36 ± 0.29%, P > 0.05). However, in MBZ, ABZ and paclitaxel treatments there were approximate sixfold increases in the percentage of cells with permeabilised mitochondrial membranes (Fig. 2e). Cells exposed to the treatments for 48 h exhibited similar results, suggesting that the majority of cells had passed through the earlier phases of apoptosis by this time.

MBZ and ABZ induce oxidative stress, resulting in reduced thiol oxidation
Indirect measurement of reactive oxygen species damage was undertaken using the DTNB assay, which quantifies reduced thiol content, from both protein and glutathione sources. This decreases as a consequence of oxidative stress. Following 48 h treatments, the reduced thiol concentration  c) and RCB2157 (mesenchymal stromal cell, d) lines were treated with MBZ, ABZ (1 nM-500 µM) or DMSO vehicle control (0.1% v/v) for 48 h, and relative cell number measured with MTT assays. IC 50 values and line-of-best-fit were calculated using non-linear regression. Each data point represents means ± SEM of three independent experiments (N = 3) carried out in quadruplicate. Following a 24 h (e) or 48 h (f) treatment of HT29 cells, DNA content was analysed using PI staining. Data represent mean ± SEM of the percentage of cells at each cell cycle stage (Sub-G1; G1; S; G2/M; > G2/M). Three independent experiments were conducted in triplicate with a minimum of 50,000 stopping events for each sample. Data were analysed using a two-way ANOVA with Dunnett's post hoc test comparing to control, where P < 0.05 was considered statistically significant. **P < 0.01, ****P < 0.0001. (g) HT-29 cells were treated with MBZ (2 µM), ABZ (2 µM), paclitaxel (50 nM), or DMSO vehicle control (0.1% v/v) for 8 h. Cells were stained with Hoechst 33,342, anti-alphatubulin, and anti-acetylated-alpha-tubulin, and imaged using the Nikon AI-R inverted fluorescent microscope with Nikon Intensilight C-HGFI mercury lamp (Olympus, Japan) at 20 × magnification, with consistent exposure ◂ Table 1 The IC50 values of MBZ and ABZ over 48 h a This value is that of the concentration that killed 50% of the cells treated. It is not a true IC 50 value (half maximal response), which could not be determined, as no complete dose response curve was obtained (Neubig et al. 2003 (Fig. 2f).

Discussion
The colon cancer cells, HT29, were the most sensitive to MBZ and ABZ. At 48 h, the HT29 cells were 10-(MCF-7) or 136-fold (MDA-MB-231) more sensitive to MBZ and 12-(MCF-7) or 77-fold (MDA-MB-231) more sensitive to ABZ than the breast cancer types tested. While direct comparisons between cancer types have not been made before, similar levels of potency to HT29s have been found in gastric cancer cells for MBZ (Pinto et al. 2019) and ABZ (Zhang et al. 2017) and in squamous cell carcinoma cell lines (Zhang et al. 2019b), whereas cell lines representing other cancer types were much less sensitive (Chu et al. 2009;Lai et al. 2017;Liu et al. 2020). IC 50 values for MBZ were also calculated in another study, for breast cancer cell lines MCF-7, MDA-MB-231, and SUM159PT (Zhang et al. 2019a). Notably, these values were substantially lower than those we report here, at 0.17 μM for MCF-7s and 0.34 μM for MDA-MB-231 cells. The two cell lines were cultured in different media, including double the level of fetal bovine serum, and the IC 50 values were reported after 72 h of treatment, rather than the 48 h we used, suggesting that breast cancer cell lines may take longer to respond to MBZ than the colorectal cancer cell line HT29.
Interestingly, this study also found that a 24 h dose of MBZ sensitised MDA-MB-231 and SUM159PT cells to treatment with ionising radiation, lowering the surviving fraction by 10-30% depending on radiation dose (Zhang et al. 2019a). Several other studies have also shown a sensitisation effect with radiation, with MBZ in meningioma cell lines and intracranial mouse models (Skibinski et al. 2018) and in glioma cell lines (Markowitz et al. 2017). MBZ was given before radiation treatment in both Zhang et al. (2019a) and Skibinski et al. (2018), including in food for the animal model, but both before and in the 9 h following radiation treatment by Markowitz et al. (2017), with both times of administration being found to be equally effective. Synergy was also found using ABZ in vitro with melanoma and small cell lung cancer cell lines when ABZ was given 12 h prior to irradiation (Patel et al. 2011).
We have shown that MBZ and ABZ have selectivity for cancer cells. Much higher concentrations were required to reduce the number of mesenchymal stromal cells. This is supported by the previous results showing similar effects in fibroblasts (WI-38), endothelial cells (HUVEC) , keratinocytes (Zhang et al. 2019b), human foreskin fibroblasts (Pinto et al. 2019), and primary canine fibroblasts (Lai et al. 2017).
In terms of the mechanism of action, we have shown that in HT29 cells, MBZ and ABZ both decrease proliferation and increase cell death levels. The drugs induce classical apoptosis, characterised by caspase-3 activation, mitochondrial membrane permeability and potential loss, phosphatidylserine exposure, DNA fragmentation, and some oxidative stress. Nuclear morphology is also typical of apoptosis, including condensation and blebbing.
The activation of caspases by MBZ and ABZ has been shown before, in a range of cancer cell types, including squamous cell carcinoma (Zhang et al. 2019b), gastric cancer (Pinto et al. 2019;Zhang et al. 2017), melanoma (Doudican et al. 2008), leukemia cells (Wang et al. 2019), and the HT29 cell line (Pourgholami et al. 2005). The generation of reactive oxygen species has been demonstrated once before, by ABZ in leukemia cells (Wang et al. 2019).
Other features of apoptosis have also been demonstrated previously, in separate studies. Phosphatidylserine exposure has been found with MBZ in melanoma (Doudican et al. 2008) and ABZ in gastric cancer cells (Zhang et al. 2017). Mitochondrial involvement in cell death has been shown by cytochrome c release in lung cancer cells treated with MBZ Sasaki et al. 2002). ABZ alters the mitochondrial membrane potential of MCF-7 cells, but only at 100 µM (Castro et al. 2016).
In combination with apoptosis, we have shown that these anti-helmintics also induce a block in the cell cycle at G2/M in HT29 cells, as well as a related depolarisation of microtubules, which is likely to lead to mitotic arrest. The G2/M block is similar to what is found with paclitaxel and has also been shown using ABZ in gastric cancer lines (Pinto et al. 2019;Zhang et al. 2017) and MBZ in lung cancer cell lines Sasaki et al. 2002). The concentration of free sulfhydryls (M) in HT-29 cells was determined following 48 h treatment. Data represents mean ± SEM of three independent experiments carried out in triplicate (N = 3). Data was analysed using a one-way ANOVA with Dunnett's post hoc test comparing to control, where P < 0.05 was considered statistically significant. *P < 0.05, **P < 0.01, ****P < 0.0001 ◂ Similar changes in microtubules to what was observed in HT29 cells treated with MBZ or ABZ have been found in other cancer cell types and described in a variety of ways. Depolymerisation of tubulin was found as a result of ABZ treatment in leukemia cells (Wang et al. 2019), ovarian cancer (Chu et al. 2009), canine glioma (Lai et al. 2017), and gastric cancer (Zhang et al. 2017) and with MBZ in lung cancer cells . In these last two studies, spindle formation was also shown to be inhibited. In melanoma cells, MBZ treatment caused diffuse staining with anti-alpha-tubulin antibody, showing microtubule disarray (Doudican et al. 2008). MBZ and ABZ treatments have also been shown to increase the ratio of soluble to insoluble tubulin in cancer cells, indicating tubulin depolymerisation. This was demonstrated for MBZ in lung cancer cells  and ABZ in gastric cancer cells (Zhang et al. 2017).
The findings presented here systematically characterise the classical apoptotic cell death caused by anti-helmintic drugs MBZ and ABZ in cancer cells. The apoptotic death is connected to a G2/M cell cycle arrest and the depolymerisation of microtubules. These findings strengthen and bring together the already available but fractured evidence in a wide range of other cancer cells types. We further show that MBZ and ABZ are selective, being many-fold more potent in cancer cells compared to non-cancerous cell types, and also demonstrate that anti-helmintics do not show equal efficacy against every cancer cell type, but do appear to be effective against the colorectal cancer cell line HT29. These conclusions make a strong case for a role for MBZ or ABZ as useful additions to current cancer therapies for at least some tumour types, improving their efficacy.