Curcumin disrupts meiotic and mitotic divisions via spindle impairment and inhibition of CDK1 activity

Abstract

Objectives:  Curcumin, a natural compound, is a potent anti‐cancer agent, which inhibits cell division and/or induces cell death. It is believed that normal cells are less sensitive to curcumin than malignant cells; however, the mechanism(s) responsible for curcumin’s effect on normal cells are poorly understood. The aim of this study was to verify the hypothesis that curcumin affects normal cell division by influencing microtubule stability, using mouse oocyte and early embryo model systems.

Materials and methods:  Maturating mouse oocytes and two‐cell embryos were treated with different concentrations of curcumin (10–50 μm), and meiotic resumption and mitotic cleavage were analysed. Spindle and chromatin structure were visualized using confocal microscopy. In addition, acetylation and in vitro polymerization of tubulin, in the presence of curcumin, were investigated and the damage to double‐stranded DNA was studied using γH2A.X. CDK1 activity was measured.

Results and conclusions:  We have shown for the first time, that curcumin, in a dose‐dependent manner, delays and partially inhibits meiotic resumption of oocytes and inhibits meiotic and mitotic divisions by causing disruption of spindle structure and does not induce DNA damage. Our analysis indicated that curcumin affects CDK1 kinase activity but does not directly affect microtubule polymerization and tubulin acetylation. As our study showed that curcumin impairs generative and somatic cell division, its future clinical use or of its derivatives with improved bioavailability after oral administration, should take into consideration the possibility of extensive side‐effects on normal cells.

Introduction

Curcumin, the phytochemical derived from the rhizome of Curcuma longa and present in the spice named turmeric, has been used for millennia as a wound‐healing agent and for treating a variety of diseases, in traditional Indian and Chinese medicine. Recently, curcumin has attracted the attention of researchers and clinicians as an anti‐inflammatory and antioxidant agent with potential use in therapy for many diseases including cancer (1). Curcumin has been shown to inhibit proliferation and induce cell death in various cancer cells in vivo and in vitro (2). Our group has shown that curcumin is able to induce cell death in many different cancer cell lines, even those resistant to other types of treatment (3, 4, 5, 6, 7, 8, 9). Curcumin has a myriad molecular targets, but it is believed that the most far‐reaching physiological consequences of its effects stem from its ability to inhibit the transcription factor NF‐κB [reviewed recently in Aggarwal and Sung (10)]. As many cancer cells are characterized by activation of NF‐κB, it is believed that they are more sensitive to curcumin than normal cells. Several studies show that induction of cell death in cancer cells requires much lower concentration of curcumin than normal cells (11, 12, 13). We have shown that curcumin induced cell death in both cancer and normal cells, but human normal proliferating T cells (with active NF‐κB) (14) are more sensitive to curcumin than non‐proliferating T cells (with inactive NF‐κB), thus indicating that NF‐κB can be a one of, but not the only, curcumin target on the apoptotic pathway (15, 16).

Curcumin induces different types of cell death including apoptosis, mitotic catastrophe and autophagy [reviewed in Salvioli (17)]. Induction of apoptotic death of cancer cells by curcumin involves activation of the intrinsic and/or extrinsic signal transduction pathways. The extrinsic pathway is induced by ligation of death receptors, and the intrinsic pathway, comprised of mitochondrial and endoplasmic reticulum pathways, is induced by DNA damage or other kinds of stress. Although there is crosstalk between these two apoptotic pathways, they lead to activation of different initiator caspases, which in turn activate common effector caspases 3, 6 and 7 (18).

Mitotic catastrophe can often lead to cell death. The final step of mitotic catastrophe occurs either through necrosis or apoptosis. Mitotic catastrophe, usually triggered by DNA damage or stress, can be also triggered by treatment with agents that influence stability of microtubules (19). In cancer cells, this is a good indicator of mitotic spindle checkpoint proficiency (20). To the best of our knowledge, there are but few studies (including those from our laboratory), showing that curcumin can affect cell division by influencing structure of the mitotic spindle (4, 9, 21, 22). However, because these studies were performed on cancer cells, inhibition of NF‐κB could not be excluded as a factor involved in cell death. We wanted to verify the hypothesis that curcumin can inhibit proliferation of normal cells with no active NF‐κB present. This would indicate that at least in in vitro studies, curcumin acts also on normal cells. We performed studies on mouse oocytes and two‐cell embryos undergoing meiotic and mitotic divisions, respectively. These cells were chosen as their spindle (mitotic or meiotic) formation is easily observable, and although they undergo divisions, they lack active NF‐κB (23). To elucidate the mechanism responsible of faulty spindle formation in cells treated with curcumin, we studied tubulin acetylation and microtubule assembly in these cells. As CDK1 kinase has been shown to be involved in spindle formation, we measured activity of this enzyme in oocytes, which matured in its presence. We also investigated potential influence of curcumin on formation of double‐strand DNA breaks.

Materials and methods

This study was approved by The Local Ethic Committee No.1 in Warsaw (Poland). Oocytes and embryos used in all experiments were obtained from F1 (C57Bl/6xCBA/ H) 3‐ to 4‐month‐old female mice.

Reagents and media

All chemicals were obtained from Sigma‐Aldrich (Poznań, Poland), unless stated otherwise. Oocytes and embryos were cultured in M2 medium (M16 medium buffered with HEPES) (24). Curcumin was obtained from Cayman (Michigan, USA) and 10 mm stock solution was prepared in DMSO.

Collection and culture of oocytes and cleaving embryos

Oocytes.  Fully grown GV (germinal vesicle) oocytes (arrested in prophase of first meiosis) were obtained from ovaries of 3‐ to 4‐month‐old female mice injected with 10 IU of pregnant mares’ serum gonadotrophin (PMSG; Intervet, Boxmeer, The Netherlands) which stimulates growth of follicles. Forty‐eight hours after PMSG injection, the ovaries were dissected and placed in M2 supplemented with dibutyryl cyclic AMP (dbcAMP; 150 μg/ml) to prevent spontaneous resumption of meiosis (25) and oocytes were released from the largest follicles by puncturing them with a hypodermic needle. Isolated oocytes were cultured with or without curcumin, in droplets of M2 medium under mineral oil, at 37.5 °C in an atmosphere of 5% CO₂ in air. They were then cultured either for 6 h after GVBD (germinal vesicle breakdown) or during time allowed for extrusion of the first polar body. Oocytes were also pre‐incubated 6 h with curcumin. Pre‐incubation medium for control and curcumin‐treated cells was supplemented with dbcAMP to arrest oocytes in the GV stage. After pre‐incubation, oocytes were washed 3× in M2 and cultured until extrusion of first polar body. Oocytes were fixed for immunostaining or protein extracts were prepared for Western blotting (WB). Twenty‐five oocytes were collected for each sample.

Two‐cell embryos.  Two‐cell embryos were obtained from females induced to superovulate by injection of 10 IU of PMSG followed 48 h later by injection of 10 IU human chorionic gonadotrophin (hCG; Intervet); after injection of hCG, females were immediately mated with F1 males. Embryos were collected 42–44 h after hCG injection from oviducts of females with vaginal plugs and they were placed in M2 medium and cut into pieces. Embryos were transferred to M2 medium or M2 with curcumin (at concentrations of 10, 20, 30, 40 and 50 μm) and cultured for 6–8 h or until the control, untreated embryos reached four‐ or eight‐cell stage, and then embryos were fixed for immunostaining.

Immunostaining.  Zonae pellucidae were removed from oocytes or embryos with 0.5% pronase in Ringer’s solution, or with acidic Tyrode’s solution, pH 2.5 (26). Oocytes and embryos were fixed and permeabilized as described previously (27). Subsequently, they were incubated with primary antibodies, anti‐γH2A.X mouse monoclonal antibody (diluted 1:500; Upstate, Billerica, MA, USA) or mouse monoclonal anti‐acetylated α‐tubulin antibody (1:50), both in PBS containing 3% BSA and 0.05% Tween 20, overnight at 4 °C or with anti‐α‐tubulin antibody conjugated with FITC (1:50) for 1 h at RT. After washing (1× in PBS and 2× in PBS containing 3% BSA and 0.05% Tween 20), oocytes and embryos incubated previously with anti‐H2A.X antibody or anti‐acetylated α‐tubulin antibody were incubated with secondary antibodies (FITC‐conjugated donkey anti‐mouse IgG or TRITC‐conjugated goat anti‐mouse IgG both from Jackson Immunoresearch Laboratories, Bar Harbor, ME, USA) for 1 h at RT, and then were washed three times in PBS. Oocytes and embryos incubated with anti‐α‐tubulin conjugated with FITC were washed 3× in PBS. For chromatin visualization, cells were incubated in DRAQ5 (10 μm in PBS without Mg²⁺ and Ca²⁺; Biostatus, Leicestershire, UK) for 10 min at 37 °C. Samples were analysed using a confocal microscope (LSM 510; Carl Zeiss, Jena, Germany).

Western blotting.  Lysates from oocytes and embryos were prepared in Laemmli buffer (28) containing 5% mercaptoethanol, and WB was performed as described previously (3). α‐tubulin and acetylated α‐tubulin were immunodetected by overnight incubation at 4 °C with primary mouse monoclonal anti‐acetylated α‐tubulin or anti‐α‐tubulin (1:200) antibodies. Subsequently a peroxidase‐conjugated secondary antibody (goat anti‐mouse, 1:2 000, Pierce, Rockford, IL, USA) was used. Detection was performed using the SuperSignal West Dura Extended Duration Substrate (Pierce). Each experiment was performed twice. Detection of α‐tubulin was used as protein loading control.

Tubulin polymerization in vitro.  Native tubulin was purified from fresh porcine brain according to the modified method of Mandelkow (29), described in detail in Nieznanski (30), and was stored at −70 °C. Tubulin preparations were thawed, and cleared by centrifugation at 22 000 g for 20 min at 4 °C. Tubulin present in the supernatant was used in further studies. Protein concentration was determined by the Bradford method (31) using bovine serum albumin as standard. Electron microscopy was performed on tubulin samples polymerized in absence or in presence of 100 μm curcumin. Tubulin at 2 mg/ml was incubated in polymerization buffer containing 10% (w/v) glycerol, 1 mm GTP, 16 mm MgCl₂ and 10 mm sodium phosphate buffer, pH 7.0, for 30 min at 37 °C. The incubation was carried out in absence or in presence of 20 μm taxol (to stabilize microtubules). Samples were subsequently diluted 10‐fold in pre‐warmed polymerization buffer or in buffer supplemented with 1.1% glutaraldehyde, and immediately applied to electron microscopy grids covered with collodion (SPI Supplies, West Chester, PA, USA) and carbon. Ten microlitres of diluted sample was applied to grids for 40 s. Negative staining was performed with 2% (w/v) aqueous uranyl acetate (SPI Supplies) for 25 s. Grids were examined using a JEOL‐1200EX transmission electron microscope operated at 80 kV with a 50 μm objective aperture.

Measurement of CDK1 activity – histone H1 kinase assay.  Activity of CDK1 was calculated as its ability to phosphorylate histone H1, according to the method described by Verlhac et al. (32). GV oocytes were incubated without or with curcumin (10, 30 and 50 μm concentrations) during maturation in vitro and collected 6 h after GVBD. GV oocytes were used as a negative control. Oocytes were washed in PBS and pooled in groups of five in 1 μl of PBS. Samples were frozen and stored at −80 °C. Three microlitres of lysis buffer [containing 0.16 m glycerophosphate, 40 mm EGTA (pH 7.3), 30 mm MgCl₂, 2 mm DTT, protease inhibitor (diluted 1:20, Complete Protease Inhibitor Cocktail; Roche, Mannheim, Germany), and BSA (11.3 mg/ml); final concentrations] was added to each sample. Oocytes were lysed by freezing and thawing, and subsequently 1.5 μl of reaction buffer (containing 0.5 mg/ml histone H1, 5 mm ATP and 1.67 μCi/μl [32P]‐ATP) was added. Samples were incubated for 30 min at 30 °C. The reaction was stopped by addition of Laemmli buffer (28). Samples were boiled for 10 min and run on 12% SDS–PAGE. Gels were exposed to autoradiography films at −80 °C for 18–48 h. Intensity of bands on the films was measured with GelDoc using software Quantity One 4.2.2 (Bio‐Rad, Mississauga, Canada). Intensity of histone H1 phosphorylation provided the means for quantification of CDK1 activity. Experiments were repeated three times.

Results

Curcumin impaired I meiotic cleavage of oocytes and II mitotic cleavage of embryos

Mammalian oocytes arrest in prophase of the first meiotic division, which is characterized by chromatin decondensation and presence of nuclear membrane (germinal vesicle – GV integrity). After hormonal stimulation in vivo or in vitro, oocytes resume meiosis (meiotic maturation). The first visible marker of meiosis resumption is GVBD, which in mouse oocytes usually occurs 1.5–2 h after stimulation. GVBD is followed by chromatin condensation and appearance of meiotic chromosomes. Subsequently, the meiotic spindle forms and migrates to the oocyte’s cortex where the first meiotic division takes place. Nine hours after GVBD, the first polar body (I pb) is extruded, and the oocyte immediately enters into the second meiotic division. After entering metaphase II, oocytes arrest again (MII oocytes). Upon activation by spermatozoon or by parthenogenetic stimuli, they resume the cycle and, within 1–1.5 h, complete the second meiotic division (II polar body extrusion). We were interested to see whether curcumin treatment affected ability of the oocyte to resume meiosis, that is, ability to undergo GVBD. Fully grown GV oocytes were placed in medium (M2) without or in the presence of different concentrations of curcumin (10, 20, 30, 40 and 50 μm). During 1.5 h of culture, almost 85% of untreated GV oocytes underwent GVBD and 1 h later, all oocytes were in GVBD stage (Fig. 1A). Treatment with curcumin delayed GVBD: after 1.5 h treatment with 30 and 50 μm curcumin, only 44% and 28% of oocytes, respectively, resumed meiosis. Additional 1 h incubation in the presence of curcumin increased the percentage of GVBD oocytes to 85% and 64% respectively. These results indicate that curcumin treatment led to a delay and partial inhibition of GVBD.

Figure 1.

 Influence of different concentrations of curcumin on GVBD, first meiotic cleavage of oocytes and second and subsequent mitotic divisions of embryos. (A) Percentage of control and curcumin‐treated GV oocytes, which underwent GVBD, measured after 1.5 and 2.5 h of culture; each group contained ∼65 oocytes. (B) First meiosis completion of control and curcumin‐treated GV oocytes. GVBD – oocytes started meiotic resumption but did not extrude I polar body (Ipb). Ipb – oocytes emitted first polar body; control n = 193, curcumin: 10, 20 μm n = about 60, 30 μm n = 171, 40, 50 μm n = about 100 (n means number of oocytes). For statistical analysis, control and curcumin‐treated oocytes that extruded I pb were compared. (C) Completion of second and subsequent mitotic divisions in control and curcumin‐treated two‐cell embryos (9 and 24 h of culture with different concentrations of curcumin). 2‐cell, 4‐cell, 8‐cell describe two‐, four‐ and eight‐cell embryos. Each group contained ∼50 embryos. For statistical analysis, control and curcumin‐treated embryos that achieved four‐cell stage (9 h) or eight‐cell stage (24 h) were compared; Student’s t‐test; *P < 0.05, **P < 0.01, ***P < 0.001. Data were collected from six independent experiments. Curcumin delayed and partially inhibited GVBD in a dose‐dependent manner (A). Curcumin in a dose‐dependent manner decreased the ability to finish the first meiotic division (B) and second and subsequent mitotic cleavages (C). 50 μm curcumin inhibited meiotic and mitotic cleavages and after 24 h caused fragmentation all of the two‐cell embryos.

Subsequently, we wanted to see whether curcumin treatment affected completion of I meiotic division, that is, extrusion of I pb. GVBD oocytes were cultured for 9 h and during this time 90% control oocytes extruded I pb. Incubation of GVBD oocytes in the presence of 20 μm curcumin resulted in decrease in the number of oocytes extruding I pb, and in the presence of 50 μm curcumin, more than 90% of oocytes were arrested in meiosis I and did not extrude I pb (Fig. 1B). These results clearly indicate that curcumin inhibited I meiotic division of mouse oocytes. Incubation of GVBD oocytes in the presence of 10 μm curcumin slightly increased the number of oocytes, which completed meiosis I, but this result was not statistically significant (P > 0.05).

Next, we asked the question whether curcumin would also influence mitotic cleavage of two‐cell embryos. Two‐cell embryos were incubated for 24 h without or with curcumin. After 9 h, the vast majority (88%) of untreated control embryos underwent second mitosis (resulting in formation of four‐cell embryos) and after 24 h, virtually all control embryos contained eight cells, which indicated that the subsequent mitotic divisions were successfully completed (Fig. 1C). After 24 h treatment with 10 μm curcumin, almost all embryos achieved the eight‐cell stage. However, at 20 or 30 μm curcumin, mixture of two‐ and four‐cell embryos was observed. In the presence of 40 μm curcumin, only few embryos (5%) reached the eight‐cell stage, but the majority (80%) reached the four‐cell stage. In the presence of 50 μm curcumin, no two‐cell embryos divided and within 24 h of incubation, all fragmented, indicating that they underwent cell death (not shown). These results showed that curcumin inhibited both meiotic and mitotic divisions in a dose‐dependent manner, within the concentration range of 20–50 μm, and 50 μm curcumin was lethal for two‐cell embryos.

Curcumin did not induce DNA double‐strand breaks in oocytes or embryos

Comet assay studies of Blasiak et al. on human GM (gastric mucosa) cells and on peripheral blood lymphocytes showed that curcumin induced single‐ and double‐strand DNA breaks (SSB and DSB respectively) (33, 34). However, our studies showed that curcumin did not induce DNA DSB in malignant cells (4, 5). To investigate further the potential influence of curcumin on DNA, in the present study we checked whether curcumin induced DSB in mouse oocytes and embryos by performing immunostaining of γH2A.X, which recognizes DSB and is a commonly used marker of DNA damage (35). We were interested whether any double‐strand DNA breaks would arise in curcumin‐treated oocytes and embryos. As positive control, we used 300 nm doxorubicin (dox, inducer of DSB; 36).

Oocytes were cultured from GV stage in the presence of curcumin, for the period necessary for untreated oocytes to finish I meiotic division and achieve MII stage. Figure 2A c, c′ shows not recognizable chromosomes and strong γH2A.X signal in dox‐treated oocytes (DSB positive control). In contrast, chromatin of untreated control oocytes (Fig. 2A a, a′) and of those treated with 40 μm curcumin (Fig. 2A b, b′) did not show presence of γH2A.X signal. The strong signal of γH2A.X was only present in extruded I pb, which is known to undergo rapid degradation (Fig. 2A a, a′). γH2A.X foci visible in control oocytes (Fig. 2A a, a′) originated from the process of recombination and are unrelated to cell death.

Figure 2.

 Visualization of DNA damage by detection of γH2A.X in oocytes and embryos treated with curcumin. Chromatin shown in red, γH2A.X shown in green. a, b and c – merge of chromatin and γH2A.X staining; a′, b′ and c′γH2A.X only. (A) Oocytes treated with curcumin at GV stage and cultured during time necessary for finishing I meiosis. a. CTR – control, b. 40 μm curcumin, c. 300 nm doxorubicin. (B) Embryos treated with curcumin at two‐cell stage and cultured during time necessary to undergo second and subsequent mitosis. a. CTR – control eight‐cell embryo (nuclei of five cells visible, rest of cells are in different focal planes), b. four‐cell embryo treated with 40 μm curcumin, c. two‐cell embryo F treated with 300 nm doxorubicin (non dividing); arrow points to polar body. Approximately 50 oocytes or embryos in each group. Curcumin did not induce DNA damage. In oocytes, γH2A.X was detected only in the polar body, designated for degradation, but not in chromatin area of the egg (A). In embryos, no significant difference was observed between control and curcumin‐treated embryos (B). Scale bar: 20 μm.

Two‐cell embryos were cultured in the presence of curcumin for the period necessary to achieve eight‐cell stage, by untreated oocytes. Dox completely inhibited cleavage of embryos and induced a very strong γH2A.X signal, indicating DSB (Fig. 2B c, c′). In contrast, γH2A.X signal detected in curcumin‐treated embryos (40 μm curcumin –Fig. 2A b, b′ and 50 μm curcumin – not shown) was similar to that observed in control untreated oocytes (Fig. 2B a, a′).

Curcumin impaired meiotic and mitotic spindle structure

We have previously shown that curcumin acts as a mitotic spindle toxin and induces mitotic catastrophe in cancer cells (4). Moreover, Holy (21) showed that, in malignant cells, curcumin disrupted the mitotic spindle. Here, we studied the effect of curcumin treatment on meiotic and mitotic spindles of mouse eggs and embryos. Figure 3 shows the structure of meiotic (Fig. 3A) and mitotic (Fig. 3B) spindles after treatment with different doses of curcumin. The severity of curcumin‐induced spindle abnormalities was dose‐dependent. Lowest (10 μm) concentrations of curcumin had no effect on meiotic or mitotic spindle structure or meiotic and mitotic divisions (Fig. 3A b and 3B b). In contrast, treatment with 30 or 40 μm curcumin caused lengthening and irregularity of distribution of the spindle microtubules in GV oocytes (Fig. 3A c, d). In these, chromosomes were scattered and misaligned. Treatment of two‐cell embryos with 30 or 40 μm curcumin caused dramatic changes in structure of spindle microtubules; they were long and curly, particularly at the spindle’s poles (Fig. 3B c, d). Highest (50 μm) concentration of curcumin almost completely inhibited formation of both meiotic and mitotic spindles (Fig. 3A e and 3B e), although in oocytes and embryos, accumulation of tubulin was visible in the vicinity of condensing chromatin.

Figure 3.

 Spindle morphology of oocytes during first meiotic cleavages and two‐cell embryos during second and subsequent mitosis after treatment with different concentrations of curcumin. Chromatin shown in red, α‐tubulin in green. (A) Oocyte treated with curcumin at the GV stage and cultured during time necessary for completion of I meiosis. (B) Embryos treated with curcumin at two‐cell stage and cultured for time necessary to undergo second and subsequent cleavages. a. control, b. 20 μm, c. 30 μm, d. 40 μm and e. 50 μm curcumin. Scale bar: 20 μm. Curcumin changed spindle morphology in both oocytes and embryos in a dose‐dependent manner. Microtubules of spindles elongated and were irregular. In oocytes treated with 30 and 40 μm curcumin, chromosomes were dispersed and not properly aligned. Highest concentration of curcumin totally inhibited spindle formation; arrow points to the polar body.

Pre‐incubation of GV‐arrested ooocytes with 50 μm curcumin caused changes in spindle structure

To check whether pre‐incubation of oocytes with curcumin would also affect structure of the newly formed spindle, oocytes at GV stage, physiologically arrested in prophase of the first meiosis (such oocytes are equivalent of non‐dividing cells), were pre‐incubated for 6 h with 30 or 50 μm curcumin. To prevent meiotic resumption in culture during pre‐incubation, medium was supplemented with dbcAMP.

Light microscopic analysis of cell morphology showed that 80% of control oocytes, in comparison with 70% and 60% of oocytes pre‐treated with 30 and 50 μm, respectively, emitted I polar body (Fig. 4A). This indicates a weak negative effect of pre‐incubation on cleavage ability of the oocytes. Immunostaining showed not completed cytokinesis in the majority of oocytes pre‐incubated with 50 μm curcumin, despite presence of I polar body, which was still connected through the spindle’s microtubules (Fig. 4B). In many oocytes, the spindle was deformed and chromosomes were not properly aligned in the metaphase plate. Pre‐incubation with 50 μm concentration of curcumin impaired successful completion of I meiotic division by affecting morphology of the oocyte spindle. Concentration of 30 μm did not induced visible effects on the oocytes.

Figure 4.

 GV‐arrested oocytes pre‐incubated with curcumin. Ability to finish I meiotic division of GV‐arrested control oocytes and GV‐arrested oocytes pre‐incubated 6 h with 30 and 50 μm curcumin after replacement with fresh medium. (A) Percentage of oocytes, which extruded I polar body (Ipb). Each group contained ∼150 oocytes. GVBD – oocytes started meiotic resumption but did not extrude Ipb. (B) Chromatin shown in red, α‐tubulin shown in green. Control oocytes and oocytes pre‐incubated with 50 μm curcumin – after replacement with fresh medium and cultured for time necessary for completion of I meiosis. Scale bar: 20 μm. Pre‐incubation with 50 μm curcumin impaired cytokinesis and resulted in decrease in ability to finish first meiotic division. Polar body was still connected to oocyte (arrow).

Curcumin did not directly affect microtubule assembly and stability

As we observed that curcumin significantly disrupted formation and morphology of the spindle in dividing oocytes and embryos, we performed an in vitro assay with purified tubulin to determine whether curcumin would directly affect tubulin polymerization. Subsequently, samples of tubulin polymerized in presence and absence of curcumin were viewed using electron microscopy. Polymerization was performed in presence (Fig. 5A,B) or absence (Fig. 5C,D) of microtubule‐stabilizing agent–taxol, or assembled structures were chemically fixed in glutaraldehyde (not shown). It allowed to observe effects on microtubule morphology and yield of tubulin polymerization. We found that curcumin had no detectable effect on tubulin polymerization (Fig. 5C,D) or microtubule assembly and stability. Number and length of microtubules assembled in presence or in absence of curcumin were very similar (compare Fig. 5A,B). Furthermore, curcumin had no effect on oligomerization of tubulin (compare Fig. 5C,D).

Figure 5.

 Electron microscopy analysis of tubulin polymerization. Electron microscopy analysis of tubulin polymerized in the absence (A, C) and in presence of curcumin (B, D). Tubulin polymerization performed in presence (A and B) and in absence of taxol (C and D). Microtubules assembled in presence of taxol were stable and did not depolymerize upon dilution (A). Curcumin did not affect number of formed microtubules nor globular tubulin oligomers (B). Microtubules formed in absence of taxol disassemble into oligomers upon dilution (C). Curcumin neither stabilized microtubules nor affected tubulin oligomerization (D). Microtubules and tubulin oligomers in presence and absence of curcumin looked the same.

Curcumin did not affect tubulin acetylation

Tubulin modifications, including acetylation, play an important role in regulating microtubule properties, such as stability and structure, as well as microtubule‐based functions, such as ciliary beating, intracellular trafficking and cell division (37). As it was shown previously, curcumin can act as an inhibitor of histone deacetylase (38), substrates of which include histones and tubulin (39, 40, 41). Accordingly, we hypothesized that curcumin may change acetylation status of tubulin. To verify this hypothesis, we conducted experiments to detect tubulin acetylation by immunostaining and WB. However, neither immunostaining (not shown) nor WB analysis of acetylated tubulin revealed any significant differences between control and curcumin‐treated oocytes (Fig. 6) and embryos (not shown).

Figure 6.

 Western blot analysis of tubulin acetylation status in oocytes treated with curcumin. Western blot analysis of acetylated α‐tubulin and α‐tubulin. Oocytes were treated with 30 and 50 μm concentration of curcumin. (A) Oocytes treated at GV stage and cultured for time necessary for formation of meiotic spindle (about 6 h). Positive control, oocytes treated with TSA – commonly used inhibitor of deacetylases. (B) Oocytes treated at GV stage and cultured for time necessary to finish I meiotic division (about 11 h). 30, 50–30 and 50 μm concentration of curcumin. AcT, α‐acetylated tubulin; T, α‐tubulin. Each lane contained lysate from 25 oocytes. Representative picture from one of three independent experiments.

Curcumin influenced CDK1 activity

It has been shown previously that curcumin inhibited activity of CDK1 (42, 43, 44), the kinase necessary for G2 mitosis transition, also involved in spindle formation (45, 46, 47). Accordingly, we measured activity of CDK1 in curcumin‐treated oocytes. GV oocytes, which do not exhibit CDK1 activity, were used as negative control. As a positive control, oocytes which resumed meiosis (6 h after GVBD) and have natural high level of CDK1 activity were used. In oocytes treated with 10 μm curcumin, elevation of CDK1 activity was observed (around 135% of control). In oocytes treated with higher doses of curcumin, decrease in CDK1 activity was observed (30 μm– 80% of control: 50 μm– 74% of control) (Fig. 7A,B). Thus, we concluded that curcumin caused an increase or decrease (depending on concentration) in activity of CDK1 kinase.

Figure 7.

 Measurement of CDK1 activity – histone H1 kinase assay. Activity of CDK1 was analysed in GV oocytes (GV), in maturing oocytes 6 h after GVBD (control) and in oocytes treated with different concentrations of curcumin (10, 30, 50 μm). (A) Mean % of CDK1 activity from three experiments after densitometric analysis. (B) Representative image from one of three independent experiments. 10 μm curcumin increased CDK1 activity and 30 and 50 μm curcumin decreased CDK1 activity.

Discussion

In this study, we demonstrated for the first time, that curcumin influenced meiotic divisions of mouse oocytes and mitotic divisions of early mouse embryos by affecting spindle formation.

We showed that curcumin delayed and partially inhibited meiotic resumption, and around 40% of oocytes treated for 2.5 h with 50 μm curcumin did not undergo GVBD. Pavlok et al. (48, 49) have shown that 20 μm curcumin prevented GVBD and resumption of meiosis of bovine oocytes and completion of nuclear and cytoplasmic maturation. They suggested that these effects of curcumin could be beneficial for prolonging time of in vitro manipulation of oocytes leading to improvement in quality of in vitro‐produced embryos. A further study showed that antioxidants reversibly inhibited meiotic resumption in rodent oocytes (50), and it is also known that curcumin acts as an antioxidant.

We also have shown that curcumin is a potent inhibitor of meiotic division (in maturing mouse oocytes) and mitotic cleavage (in early mouse embryos). We showed that 50 μm curcumin inhibited extrusion of I pb in 90% of oocytes and completely blocked two‐cell embryo division. To our knowledge, this is the first report of any harmful effect of curcumin on maturation of mouse oocytes and development of pre‐implanted embryos. Other laboratories have reported very serious developmental defects caused by curcumin in zebrafish embryos (51).

We also performed experiments on the role of curcumin in induction of DNA damage. It has recently been reported that curcumin induced formation of stabile complexes of topoisomerase (topo) with DNA (either topo I or topo II) thus leading to permanent DNA strand breaks and cell death (52). Our study indicates that curcumin does not induce DSB in mouse oocytes and embryos. Even in cells treated with the highest (50 μm) concentration of curcumin, there were no visible γH2A.X foci. This provides indirect evidence that processes induced by curcumin in mouse oocytes and embryos differ from apoptotic death pathways induced by curcumin in malignant HL‐60 cells, in which the high level of γH2A.X signal indicated oligonucleosomal DNA degradation (5). We cannot exclude that strong signals observed in dox‐treated oocytes and embryos could be the result of apoptotic DNA damage. We also focused our studies on curcumin’s action on meiotic and mitotic divisions, with particular interest in spindle formation. This still remains a relatively unexplored field, although several studies have shown that curcumin was able to arrest cells in G₂/M phase of the cell cycle [reviewed in Karunagaran (53)]. Previously, we have shown that curcumin acts as a mitotic poison inducing mitotic catastrophe (4). In addition, Holy (21) showed that curcumin arrested MCF‐7 malignant cells in mitosis. These cells exhibited monopolar spindles, and no chromosome segregation was observed. Although, in our study we did not observe monopolar spindles, we observed abnormal meiotic and mitotic spindles leading to lack of chromosome segregation. In our study, 50 μm curcumin almost completely inhibited formation of spindles. Although morphological observation did not allow us to conclude whether curcumin affected polymerization or depolymerization of spindle microtubules, Gupta et al. (22) described that curcumin induced significant depolymerization of interphase and mitotic spindle microtubules in HeLa and MCF‐7 cells.

Previously, a role of curcumin in cytokinesis inhibition in dying somatic cells undergoing mitotic catastrophe has been demonstrated (9). One of our questions was whether curcumin would induce abnormalities only when present in the medium during cell division, or would it exert its effects also when oocytes were pre‐incubated in its presence prior to maturation, in GV stage, and subsequently matured in curcumin‐free medium. We observed that most oocytes, which were pre‐incubated in curcumin‐containing medium, were able to resume meiotic division. This indicates that only high concentration of curcumin (50 μm) could cause irreversible changes and impair cell division by affecting completion of cytokinesis. However, it has to be noted that because of poor bioavailability of curcumin after oral administration, 50 μm concentration has never been achieved in prophylactic or therapeutic applications in vivo.

To elucidate further any possible mechanism of curcumin‐induced changes to the spindle, we asked whether curcumin would directly influence tubulin polymerization. In the in vitro tubulin polymerization assay, we did not observe any effect of curcumin on tubulin polymerization, as well as not on microtubule formation and stability. Similarly, Thomas et al. (54) have recently demonstrated that curcumin had no effect on tubulin polymerization assayed turbidimetrically in vitro. This contradicts observations of Gupta et al. (22) who concluded that curcumin induces in vitro tubulin aggregation and inhibits microtubule assembly. These discrepancies may result from different experimental conditions, different purification protocols or, as suggested recently (54), from different curcuminoids used in their studies. One possible explanation is that Gupta et al. (22) performed electron microscopy of tubulin incubated in the presence of 1 m glutamate, known to induce the formation of sheets of parallel tubulin protofilaments rather than microtubules (55).

To determine the mechanism of curcumin’s effect on structure of microtubules, we studied acetylation status of tubulin, which is known to be important for spindle formation (56). However, similar to results from other laboratories, performed on human breast and prostate cancer cell lines (54), we did not observe any changes in acetylation level of tubulin in mouse oocytes and embryos.

Changes in spindle morphology, similar to those observed in our experiments, have also been reported for mouse and rat oocytes (57, 58). These authors suggested that spindle assembly during mitosis is tightly controlled by phosphorylation. Curcumin is a known inhibitor of many protein kinases [reviewed in Aggarwal (59)] and it is known that polymerization of tubulin depends on activity of many regulatory proteins, for example, microtubule‐associated proteins (MAPs) regulated by phosphorylation (60). Thus, it cannot be excluded that curcumin affects phosphorylation status of tubulin.

One of the targets of curcumin could be CDK1 kinase. This kinase is a key factor in controlling resumption of meiosis and we believe that delay and partial inhibition of GVBD could be a result of inhibitory effects of curcumin on activity of this kinase. In addition, it is known that CDK1 phosphorylates survivin (45, 46), and survivin is involved in regulation of spindle formation (47). We showed that curcumin in concentrations of above 30 μm decreases activity of CDK1. But curcumin at concentration of 10 μm increased histone H1 kinase activity of CDK1. This correlates with our observation that efficacy of meiotic resumption of oocytes treated with 10 μm curcumin was a little higher than that of controls. Inhibition of CDK1 by curcumin has been shown in several studies (42, 43, 44). Sahu et al. (42) observed decreasing CDK1 levels in cancer cells but not in normal immortalized cells. Others have shown decrease in the CDK1 levels in both normal and cancer cells (43, 44) but kinase activity of CDK1 after curcumin treatment has not been fully investigated.

Anti‐cancer activity of curcumin has been documented not only in vitro, but also in vivo, on laboratory animals [reviewed in Strimpakos and Sharma (61)]. This beneficial effect of curcumin, together with its very low bioavailability after oral administration and lack of toxicity in human volunteers, encouraged researchers to place it in the category of hormetins, which have damaging effects at higher doses but can induce an adaptive beneficial effect on cells or organisms at low doses (62). However, administration of the highest acceptable dose (8 g/day) by patients resulted in serum concentration of 1.75 ± 0.80 μm (54). In vitro treatment of cells required at least 10 times higher concentration of curcumin (61). This can be partially explained by curcumin’s pleiotropic activity, that is, its ability to affect many targets and many processes (which possibly modulate interaction between host and malignancy), and despite low bioavailability after oral administration, to counteract cancer development (anti‐inflammatory and anti‐angiogenic activities of curcumin). Moreover, although curcumin’s low systemic bioavailability may limit access of sufficient concentrations for pharmacological effect in certain tissues, chemical analogues and novel delivery methods are currently under development. We have demonstrated that in vitro application of curcumin at concentrations three times higher than that used for cancer cell lines can affect spindle microtubules and damage normal generative and somatic cells. This indicates that normal cells are also sensitive to curcumin. Therefore, the possibility of harmful side‐effects on human fertility and reproducibility should be considered when curcumin or curcumin derivatives with improved bioavailability, are used in clinical applications.