Literature Review - Systematic Research 4 Pages

Mutation Research 793 (2015) 71–78

Contents lists available at ScienceDirect

Mutation Research/Genetic Toxicology and Environmental Mutagenesis

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / g e n t o x C o m m u n i t y a d d r e s s : w w w . e l s e v i e r . c o m / l o c a t e / m u t r e s

Low concentrations of caffeine induce asymmetric cell division as observed in vitro by means of the CBMN-assay and iFISH

Vasiliki I. Hatzi ∗, Maria Karakosta, Katarzyna Barszczewska, Ioanna Karachristou, Gabriel Pantelias, Georgia I. Terzoudi Laboratory of Health Physics, Radiobiology & Cytogenetics, Institute of Nuclear & Radiological Sciences & Technology, Energy & Safety, NCSR “Demokritos”, Athens, Greece

a r t i c l e i n f o

Article history: Received 31 July 2015 Accepted 3 August 2015 Available online 5 August 2015

Keywords: Caffeine Asymmetric cell division Chromatid breaks Aneuploidy Micronuclei

a b s t r a c t

The dual role of caffeine as a chromosomal damage inducer and G2/M-checkpoint abrogator is well known but it is observed mainly at relatively high concentrations. At low concentrations, caffeine enhances the cytogenetic effects of several carcinogens and its intake during pregnancy has been recently reported to cause adverse birth outcomes. Interestingly, a threshold below which this association is not apparent was not identified. Since chromosomal abnormalities and aneuploidy are the major genetic etiologies of spontaneous abortions and adverse birth outcomes, we re-evaluate here the effects of caffeine at the cytogenetic level and propose a model for the mechanisms involved. Our hypothesis is that low caf- feine concentrations affect DNA replication and cause chromosomal aberrations and asymmetric cell divisions not easily detected at metaphase since damaged cells are delayed during their G2/M-phase transition and the low caffeine concentrations cannot abrogate the G2-checkpoint. To test this hypoth- esis, caffeine-induced chromatid breaks and micronuclei in peripheral blood lymphocytes (PBLs) were evaluated in vitro after low caffeine concentration exposures, followed by a short treatment with 4 mM of caffeine to abrogate the G2-checkpoint. The results show a statistically significant increase in chromatid breaks at caffeine concentrations ≥1 mM. When caffeine was applied for G2/M-checkpoint abrogation, a statistically significant increase in chromatid breaks, compared to an active checkpoint, was only observed at 4 mM of caffeine. The potential of low concentrations to induce asymmetric cell divisions was tested by applying a methodology combining the cytochalasin-B mediated cytokinesis-block micronucleus assay (CBMN) with interphase FISH (iFISH), using selected centromeric probes. Interestingly, low caffeine con- centrations induce a dose dependent aneuploidy through asymmetric cell divisions, which are caused by misalignment of chromosomes through a mechanism unrelated to the formation of chromatid breaks. The cytogenetic approach used, combining CBMN with iFISH, is proposed as a valuable tool to test chemically induced asymmetric cell divisions.

© 2015 Elsevier B.V. All rights reserved.

1. Introduction

Caffeine is probably one of the most extensively studied nat- urally occurring dietary chemicals that have been linked both to beneficial and adverse health effects [1–11]. It occurs nat- urally in more than 60 plant species throughout the world and is one of the most frequently ingested neuroactive drugs to which people are exposed through food, beverages (e.g. coffee, tea, cola, energy drinks, chocolate) and medicines. The pharmacological effects of caffeine are known to be involved in cell cycle perturbation, programmed cell death or

∗ Corresponding author. Fax: +30 210 6534710. E-mail address: [email protected] (V.I. Hatzi).

apoptosis, drugs toxicity enhancement and clastogenicity [2,9–23]. At cellular level, caffeine has a dual role: it affects the DNA structure and DNA synthesis (S-phase) through the intercala- tion of DNA [24], possibly via a local unwinding and, in parallel, at high concentrations, it abrogates G2/M cell-cycle checkpoint [25], allowing thus the damaged cells to proceed to the next phases of the cell cycle. Even though these effects could cause adverse cellular cancer related effects, the majority of references in the literature support that high concentrations of caffeine (>1 mM) are needed to reveal these effects. In addition, there is limited evidence for the carcinogenicity of caffeine (IARC, Group 3) in experimental animals and humans [24]. Particularly, there is lack of information concerning the potentiating mech- anism of caffeine in cancer or cancer related cytogenetic effects [26,27].

http://dx.doi.org/10.1016/j.mrgentox.2015.08.002 1383-5718/© 2015 Elsevier B.V. All rights reserved.

72 V.I. Hatzi et al. / Mutation Research 793 (2015) 71–78

In contrast with the studies on the absence or even bene- ficial effects of caffeine [28] at low concentrations, numerous reports support that caffeine acts synergistically and modifies the mutagenic and carcinogenic potential of UV, ionizing radia- tion and chemical mutagens in various human and experimental models, increasing chemical and radiation risk [16,20–22,29–32]. Specifically, it has been shown that low concentrations of caf- feine (0.1–10 �M) act synergistically when combined with ionizing radiation and increase in a dose dependent manner the radiation induced chromosomal damage, revealing the potential mutagenic profile of caffeine [33,34]. Interestingly, a very recent systemic review links caffeine intake during pregnancy with adverse birth outcomes [1]. Greenwood’s et al. meta-analysis supports that consumption of caffeine is associated with increase in risk of spon- taneous abortion (14% increase in risk), stillbirth (19% increase in risk), preterm delivery (2% increase in risk), low birth weight (increase in risk 7%) and small for gestational age (SGA) (increase in risk 10%) [1]. Strikingly, it is concluded that there is no iden- tifiable threshold below which the associations are not apparent [1]. Given that chromosomal abnormalities and autosomal aneu- ploidies [35–37] are the major genetic etiologies of spontaneous abortions and adverse birth outcomes and also that caffeine is one of the most frequently ingested chemicals, the clastogenic and aneugenic effects of caffeine, particularly at low concentrations, need to be re-evaluated.

In the present study, we focus on the investigation of the cyto- genetic effects of low concentrations of caffeine in vitro (i.e. <1 mM) in human peripheral blood lymphocytes (PBLs). For this purpose, we hypothesize that caffeine at low concentrations can also affect cellular DNA replication and cause chromosomal aberrations and asymmetric cell division but this effect cannot be easily detected at mitosis, as low caffeine concentration cannot abrogate the G2- checkpoint and the damaged cells are delayed during their G2/M transition. If this hypothesis is correct, the cytogenetic effects of caffeine at low exposure concentrations could be revealed when the damaged cells are forced to proceed to mitosis through the abrogation of their G2-checkpoint. Specifically, we used a 1 hour and 20 min 4 mM caffeine treatment for G2-checkpoint abrogation. Moreover, we tested the potential of low concentrations of caffeine to induce asymmetric cell divisions. To investigate whether the asymmetric cell divisions arise from misalignment of chromosomes of a normal parental cell, we evaluated the yield of caffeine-induced asymmetric cell divisions through the combination of the CBMN assay (cytokinesis-block micronucleus assay) with iFISH (inter- phase FISH) using the centromeric probes for chromosomes 11 and 17 (CEP11 and CEP17). The exposure of cells to cytochalasin-B prior to iFISH analysis ensures that the abnormal cells arise from the same parental cell and enables as well the visualization of the centromeric regions in micronuclei.

2. Materials and methods

2.1. Cultures of blood lymphocytes and caffeine exposures

Peripheral blood samples were taken by venipuncture from three healthy individuals and collected in heparinized tubes. Blood cultures were initiated by adding 0.5 ml of whole blood to each culture tube that contained 5 ml of RPMI medium, supplemented with 10% fetal bovine serum (FBS), 1% L-glutamine, 1% antibiotics [penicillin (100 U/ml)- streptomycin (100 �g/ml)] and 2% phy- tohemagglutinin. All incubations were at 37 ◦C in a humidified incubator (5% CO2, 95% air). All culture media were provided from Biochrom-AG, unless stated otherwise. Anhydrous caffeine (purity ≥99%, MW: 194.2) was purchased from Fluka. Caffeine was diluted in PBS and added, in blood culture, to reach a final concentration

ranging from 0.01 �M to 10 mM. To test the effect of caffeine at lower concentrations, the following experimental procedure was applied: 48 h after whole blood culture initiation, the cells were exposed to a range of caffeine concentrations (0.01 �M-10 mM) for 24 h. Four hours after the addition of caffeine, a lower colcemid con- centration (final concentration 0.05 �g/ml instead of 0.1 �g/ml), was added for 20 h to collect all the potentially damaged cells that will reach M-phase during this time period at mitosis. For G2/M checkpoint abrogation, caffeine treated cycling PBLs were exposed to 4 mM of caffeine and incubated for 1 hour and 20 min at 37 ◦C before fixation, as described previously by Terzoudi et al. [38,39]. The cultured cells were harvested by centrifugation (1300 rpm), treated with hypotonic KCl (75 mM) (Sigma–Aldrich) and fixed with freshly prepared 3:1 methanol-acetic acid (v/v) (purchased from Fluka and Baker, respectively).

2.2. Chromosomal aberration analysis

Cell suspension (20 �l) was dropped on wet slides. The slides were air dried and stained in a 3% solution of Giemsa dye (Merck) for 10 min and rinsed with water. Air-dried slides were embed- ded with cover slips and coded for analysis. For each experimental point, 300 well spread metaphases were analyzed for chromatid type aberrations and the total number of chromosomes by micro- scopic examination. To study the effect of 0.01 �M–4 mM caffeine on chromosome number of PBLs at the subsequent metaphase, we counted the chromosome number in 300 well spread metaphases and the percentage of cells with chromosome number 46, >46 and <46 was calculated. 2000 nuclei were also examined to determine the mitotic index (MI, i.e., percentage of cells undergoing mitosis). For each experimental point, mean values and standard deviations were obtained from three independent experiments. Data were evaluated statistically by Student’s t-test. All P values were con- sidered statistically significant at P < 0.05.

2.3. Cytokinesis-block micronucleus (CBMN) assay

The cytokinesis-block micronucleus assay (CBMN test) was per- formed using the cytochalasin-B technique described by Fenech and Morley [40]. At 44 h after peripheral blood culture initiation, cells were exposed to selected doses of caffeine (0.01 �M–10 mM) and were blocked in cytokinesis with the addition of cytochalasin- B (Sigma, St. Louis, MO; final concentration, 5.56 �g/mL). The total incubation time for all cultures was 72 h. The cells were harvested, fixed according to the protocol of Fenech [41] and stained for 10 min with 5% Giemsa (Merck). For each experimental point, about 2000 binucleated cells (BNC) were scored blindly, following the scor- ing criteria outlined by the HUMN Project [42], and the number of micronuclei per 2000 binucleated cells was recorded for each experimental point. Data were evaluated statistically by Chi-square test. All p values were considered statistically significant at P < 0.05. The nuclear division index (NDI) was calculated using the for- mula NDI = (M1 + 2M2 + 3M3 + 4M4)/n, where M1 to M4 indicates the number of cells with one to four nuclei, respectively, and n is the number of cells scored. To calculate the NDI, a minimum of 500 cells was also scored to determine the percentage of cells with 1, 2, 3 and ≥4 nuclei.

2.4Interphase fluorescence in situ hybridization (iFISH) To evaluate the yield of caffeine-induced asymmetric cell

divisions, we combined the CBMN assay with iFISH using the com- mercial probes CEP11 (D11Z1) SpectrumGreen Probe and CEP17 (D17Z1) SpectrumOrange Probe. DNA probes were applied follow- ing the standard procedures outlined by the manufacturer (Vysis Inc., Downers Grove, IL). A normal hybridization pattern for each cell consisted of two green (CEP11) and two red signals (CEP17), otherwise the pattern was considered abnormal. At least a total

V.I. Hatzi et al. / Mutation Research 793 (2015) 71–78 73

of 400 cells (i.e. 200 cytochalasin-B induced binucleated cells) from each donor, were evaluated for each experimental point. Dual-color FISH images were digitally generated using the ISIS FISH-imaging software (MetaSystems, Germany). The statistical significance of differences in normal and abnormal hybridization patterns between cells treated with caffeine and controls were cal- culated using �2-statistics with continuity correction or Fisher’s exact test. P-values less than 0.05 were considered statistically sig- nificant. Odds ratio (OR) was calculated and reported with 95% confidence intervals (95% CI).

3. Results

3.1. Low concentrations of caffeine in vitro do not induce an increase in chromatid type aberrations in PBLs

To investigate the effects of low concentrations of caffeine, PBLs obtained from healthy individuals were exposed to a range of caf- feine concentrations from 0.01 �M to 10 mM for 24 h and the cells were examined at the subsequent metaphase. Considering that caf- feine affects the S-phase of the cell cycle, giving rise to chromatid breaks (Fig. 1), colcemid was added 4 h after the addition of caffeine and remained in the culture until fixation process, so that caffeine- damaged cells delayed in S-phase could be included in the analysis. Our results show that a 24 h caffeine exposure induces an increase in chromatid type aberrations in PBLs with active G2/M checkpoint, but this increase follows a statistically significant pattern only at caffeine concentrations ≥1mM (Table 1 ). To test the potential effect of low concentrations of caffeine, exposed cells from the same indi- viduals were subjected to G2/M-checkpoint abrogation [38,39], so that G2-blocked cells can proceed to mitosis. The results show that, when G2/M checkpoint is abrogated in vitro, a statistically signifi- cant increase (t = 2.8, P < 0.05) in chromatid type aberrations is only observed at the highest concentration tested (i.e. 4 mM) (Table 1). Moreover, no statistically significant differences are observed in mitotic indices between the G2/M-checkpoint abrogated and non- abrogated cells.

3.2. Caffeine increases the yield of micronuclei induction, as observed by means of CBMN assay

The effects of caffeine on micronuclei (MN) induction were also tested by means of the CBMN assay (Fig. 2). For this purpose, PBL cultures were exposed for 24 h to concentrations of caffeine: 10 �M, 250 �M, 500 �M, 1 mM, 4 mM and 10 mM. Micronuclei per 1,000 binucleated cells were scored and the percentages of cells with 1, 2, 3 and +4 micronuclei were also obtained. The results (Table 2 ) show that caffeine affects MN yield from concentrations ≥10 �M, with a statistically significant increase in the number of MN only after exposure to concentrations of caffeine ≥1 mM: i.e. at a concentra- tion that has been also shown to induce chromatid type aberrations. No effect was observed on the nuclear division index after exposure to the concentration range of caffeine being tested (10 �M–10 mM).

3.3. Low concentrations of caffeine in vitro affect the chromosomal distribution in PBLs daughter cells

The caffeine effect on misalignment of chromosomes in daugh- ter cells of exposed PBLs was also tested. For this purpose, PBLs were exposed for 24 h to a concentration range of caffeine (0.01 �M–4 mM) and the number of chromosomes was evaluated at metaphase level. The results (Table 3) show a statistically sig- nificant increase in the misalignment of chromosomes in daughter cells of PBLs at caffeine exposures ≥10 �M.

3.4. Low concentrations of caffeine in vitro induce asymmetric cell division in PBLs, as observed by combining CBMN-assay and iFISH

Caffeine-induced asymmetric cell division was tested on chro- mosomes 11 and 17 by means of CBMN-assay and iFISH analysis using the commercial probes CEP11 (D11Z1) SpectrumGreen Probe and CEP17 (D17Z1) SpectrumOrange Probe (Vysis Inc., Downers Grove, IL). In a normal cell division, it is expected that each daughter cell will contain a pair of each chromosome [i.e. two chromosomes 11 (green signal) and two chromosomes 17 (red signal)]. In each nuclear pair, the total number of each chromosome it is expected to be 4 (normal balanced profile). In our experiments, cells with a nor- mal balanced profile are visualised as containing two nuclei each with two red and two green signals (and in binucleated cells a total number of signals 4 red and 4 green) (Fig. 3). The asymmetric cell division is determined as a pair of nuclei with abnormal-balanced profile and is visualised as number of signals other than two red and two green in each nucleus, but with a total number of signals 4 red and 4 green in the pair. In pair of nuclei with an unbalanaced profile, the total number of signals obtained, that arises from the same parental cell and share the same cytoplasm, is other than 4 red and 4 green. Our results demonstrate that 6–11% of binucleated cells are asymmetrically divided with an unbalanced profile. Even in control sample, the abnormal hybridization pattern could be up to 8% (cut off value for each probe is 4%; according to E.C.A. guide- lines and quality assurance for acquired cytogenetics 2013). Our results show that caffeine induces a statistically significant increase in abnormal cells (i.e. other than 2 green and 2 red signals) when compared to control (unexposed) samples (i.e. 2 green and 2 red signals) after in vitro exposure of PBLs to concentrations of caffeine ≥10 �M (Table 4 ). Table 4 presents the total divisions analysed (200 divisions, i.e. 400 cells) from each donor.

4. Discussion

Based on the recent evidence that caffeine intake during preg- nancy has been reported to cause adverse birth outcomes [1], we re-evaluate in this manuscript the effects of low concentrations of caffeine in vitro on the induction of chromatid type aberrations, micronuclei formation and asymmetric cell division in human PBLs. Particularly, the clastogenic effects of caffeine in human PBLs in vitro are thoroughly studied. Our results demonstrate that a 24 h exposure to caffeine induces a statistically significant (t = 2.25, P < 0.05) increase in chromatid type aberrations in PBLs only at exposure concentrations ≥1 mM (Table 1). Interestingly, a lack of chromatid type exchanges is observed (Fig. 1) and a statistically significant increase (X2 test, P = 0.044) in micronuclei induction is also demonstrated for concentrations of caffeine ≥1 mM (Table 2). Regarding the clastogenicity of caffeine, these results are consis- tent with the early observations of Ostertag and Greif [43] for high caffeine concentrations. In the present work, we hypothesize that, following exposure to low caffeine concentration, a clasto- genic effect is induced which is not easily detected at metaphase, since the damaged cells are delayed during their G2/M-phase tran- sition and the low caffeine concentrations cannot abrogate the G2-checkpoint.

To test this hypothesis in PBLs in vitro, we used an approach based on G2-checkpoint abrogation with a short treatment with 4 mM of caffeine after the exposure to low caffeine concentra- tions. Our results show that the increase in chromatid breaks is not statistically significant at caffeine concentration less than 1 mM, a fact that does not support the tested hypothesis. This observation raises the question whether caffeine-affected cells in S-phase can be progressed in the cell cycle and blocked at G2-

74 V.I. Hatzi et al. / Mutation Research 793 (2015) 71–78

Fig. 1. Metaphases with chromatid type aberrations after in vitro 24 h exposure of cycling peripheral blood lymphocytes to 10 mM caffeine with approximately (a) 40, (b) 60 and (c) 80 chromosome breaks as well as a metaphase (d) with pulverized chromosomes.

Table 1 Mitotic index and yield of chromatid breaks after exposure of PBL for 24 h to various concentrations of caffeine. The exposed cells had either active or abrogated G2/M checkpoints.

N. of cells analyzed per experimental point

Mitotic index (%) Damaged cells (%) Yield of chromatid breaks per cella

Statistics (t-test, P) For caffeine induced chromatid breaks

Statistics (t-test, P) Between different checkpoint status

Control 300 10.71 3.30% 0.033 ± 0.02 t = 1.44 P > 0.1

Control (abrogated G2/M) 300 9.83 6.33% 0.063 ± 0.03 0,01 �M 300 9.40 3.30% 0.030 ± 0.03 t = 0.14,

P > 0.5 t = 1.22 P > 0.1

0,01 �M(abrogated G2/M) 300 9.20 6.00% 0.060 ± 0.03 0,1 �M 300 9.01 3.00% 0.030 ± 0.02 t = 0.18,

P > 0.5 t = 1.92 P > 0.1

0,1 �M (abrogated G2/M) 300 8.02 6.33% 0.070 ± 0.03 1 �M 300 9.90 3.67% 0.037 ± 0.02 t = 0.24,

P > 0.5 t = 2.63 P > 0.05

1 �M (abrogated G2/M) 300 8.89 8.00% 0.080 ± 0.02 10 �M 300 9.84 4.67% 0.047 ± 0.24 t = 0.10, P > 0.5 t = 0.18

P > 0.5 10 �M (abrogated G2/M) 300 9.34 7.33% 0.073 ± 0.02 100 �M 300 9.92 6.67% 0.070 ± 0.05 t = 1.19,

P > 0.5 t = 1.62 P > 0.1

100 �M (abrogated G2/M) 300 8.63 12.67% 0.130 ± 0.04 1 mM 300 7.89 23.33% 0.520 ± 0.12 t = 6.93, P < 0.01 t = 3.33

P < 0.02 1 mM (abrogated G2/M) 300 6.42 35.00% 0.820 ± 0.10 4 mM 300 5.36 56.67% 1.130 ± 0.14 t = 13.43, P < 0.001 t = 14.02

P < 0.001 4 mM (abrogated G2/M) 300 4.31 63.00% 3.040 ± 0.19 a Mean values and standard deviations were calculated from three independent experiments.

V.I. Hatzi et al. / Mutation Research 793 (2015) 71–78 75

Fig. 2. Micronuclei (MN) after 24 h exposure to high concentrations (4–10 mM) of caffeine (a) unexposed binucleated cells (BNUC), (b, c) BNUC with 2 and 12 MN, (d) a multi-micronucleated cell, (e) nucleoplasmic bridges (NPBs), (f) nuclear buds (NBUD).

Table 2 Caffeine induced chromosomal damage by means of MN induction: Distribution of MN formation after exposure of lymphocytes from 10 �M to 10 mM of caffeine for 24 h.

Caffeine BNCs(Total) NDI Total MNn (%) Micronuclei (MN) in binucleated cells (BNCs) n (%) P-value* (vs control)

0 1 2 3 4 5 6 Control 2000 1.60 ± 0.03 27 (1,35) 1973 (98.05) 21(1.05) 3(0,15) – – – – Caffeine 10 �M 2000 1.54 ± 0.05 35 (1.75) 1965 (98.25) 23(1.15) 3(0.15) 2(0.10) – – – 0.551 Caffeine250 �M 2000 1.63 ± 0.02 39 (1.95) 1961 (98.05) 25(1.25) 4(0.20) 2(0.10) – – – 0.472 Caffeine500 �M 2000 1.60 ± 0.01 47 (2.35) 1953 (98.15) 35(1.75) 6(0.30) – – – – 0.100 Caffeine 1 mM 2000 1.68 ± 0.03 74 (3.70) 1926 (96.3) 52 (2.6) 5(0.25) 4(0.20) – – – <0.0001 Caffeine 4 mM 2000 1.65 ± 0.02 >80 (>4)* 1920 (96) 51(2.55) 7(0.35) 5(0.25) – – – <0.0001 Caffeine 10 mM 2000 1.64 ± 0.03 >95 (>4.75)* 1905 (95.25) 67(3.35) 5(0.25) 6(0.30) – – – <0.0001

Chi square test. n: number. NDI: Nuclear division index.

* Multi-micronucleated, non-countable cells were also observed (see Fig. 2).

Table 3 Distribution of chromosome number in metaphases of PBL after exposure to a range of caffeine concentrations (0.01 �M–4 mM) for 24 h.

Caffeine Total cellsn Number of chromosomes per metaphase analyzed n (%) P-value* (vs control)

<46 46 >46

0 �M (control) 300 0 (0.00) 300 (100.00) 0 (0.00) 0.01 �M 300 3 (1.00) 297 (99.00) 0 (0.00) ns 0.1 �M 300 6 (2.00) 294 (98.00) 0 (0.00) ns 1 �M 300 4 (1.33) 295 (98.33) 1 (0.66) 0.080 10 �M 300 16 (5.33) 275 (91.66) 9 (3.00) <0.001 100 �M 300 18 (6.00) 273 (91.00) 9 (3.00) <0.001 1 mM 300 26 (8.66) 260 (86.66) 14 (4.66) <0.001 4 mM 300 56 (18.66) 210 (70.00) 34 (11.33) <0.0001

n: number. ns: non-statistically significant difference.

phase. The fact that the mitotic indices calculated after exposure of PBLs to various caffeine concentrations do not reveal any dif- ference between G2/M checkpoint abrogated and non-abrogated cells, implies that caffeine-induced damaged cells are not blocked at the G2/M checkpoint. In addition, the observation that caf- feine, despite its S-phase dependent mode of action, fails to induce chromatid type exchanges in PBLs (i.e.: the typical damage for S-phase dependent clastogens), may suggest that caffeine affects the replicative mechanism of the DNA and delays replication fork progression rather than inducing a direct clastogenic effect. This

conclusion is in agreement with earlier work reported by Johans- son et al. [44]. Such a mechanism of caffeine action is reinforced by the fact that, after the removal of caffeine (>1 mM) from the cell culture, the mitotic index approaches again numerically the mitotic index of the unexposed cells without showing any appar- ent chromatid type damage like chromatid breaks or exchanges. If there were a direct clastogenic action of caffeine, the activation of cellular repair machinery would give rise to the formation of chromatid type exchanges. The absence of such chromatid type exchanges, even at exposure of PBLs to high concentrations of

76 V.I. Hatzi et al. / Mutation Research 793 (2015) 71–78

Fig. 3. Caffeine at doses ≥10 �M induced asymmetric cell divisions and abnormal hybridization patterns of binucleated cells with or without MN as observed using the combination of CBMN-assay and iFISH analysis. In iFiSH analysis, the centromeric probes CEP17 (spectrum orange, red signal) and CEP11 (spectrum green, green signal) have been used. An example of normal hybridization pattern (a) as well as examples of abnormal balanced hybridization patterns (b–h) are presented.

Table 4 Percentage of asymmetric PBL divisions after exposure to low concentrations of caffeine (0.01 �M–500 �M) by means of combined CBMN and iFISH analysis using the commercial probes; CEP11 (D11Z1) spectrum green probe and CEP17 (D17Z1) spectrum orange probe (Vysis Inc., Downers Grove, IL). A statistically significant increase in abnormal cell divisions was observed at caffeine concentrations ≥10 �M.

Caffeine Total divisions n

Normal divisions n (%)

Abnormal divisions (balanced) n (%)

Abnormal divisions (unbalanced) n (%)

Abnormal divisions (total) n (%)

P-value OR (95% CI)

Control(0 �M) 222 207 (93.24) 1 (0.45) 14 (6.31) 15 (6.76) 0.01 �M 214 188 (87.85) 2 (0.93) 24 (9.96) 26 (12.15) 0.510*

0,1 �M 202 186 (92.08) 2 (0.99) 14 (6.93) 16 (7.92) 0.504*

1 �M 243 220 (90.53) 1 (0.41) 22 (9.05) 23 (9.47) 0.966*

10 �M 253 213 (84.19) 11 (4.35) 29 (11.46) 40 (15.81) 0.012** 10.69 (1.37–83.55) 250 �M 240 180 (75.00) 32 (13.33) 28 (11.67) 60 (25.00) <0.0001** 36.80(4.97–272.02) 500 �M 208 156 (75.00) 36 (17.31) 16 (7.69) 52 (25.00) <0.0001** 47.77(6.48–352.22)

n: number * Chi square test (P-value was estimated comparing normal divisions and abnormal balanced divisions among controls and cells treated with 0.01 �M–1 �M caffeine).

** Fisher’s exact test (P-value was estimated comparing each time normal and abnormal balanced divisions between controls and cells treated with 10 �M–500 �M caffeine).

caffeine, does not support, therefore, a direct clastogenic profile and opens up new questions on the mode of action of this chemi- cal. Based on these results, we may propose the following model: caffeine induces replication fork progression delay and, particu- larly at high concentrations, it allows entrance of cells at mitosis before completion of DNA replication. Due to the presence of stalled replication forks during G2/M transition and the induction of chro- matin condensation, a failure of chromatin organization occurs and gives rise to chromatin breakage, without apparent chromosome exchanges, as shown in Fig. 1. The fluctuations observed in the

extent of chromosomal damage between the different caffeine- exposed cells, at the same caffeine concentration, depends on the stage of the S-phase and the level of completion of DNA repli- cation at the time of exposure to caffeine. The above-proposed mechanism, underlying the effects of caffeine at the chromosomal level, is in agreement with our earlier work on the involvement of chromatin dynamics in the formation of chromosomal aberrations [45,46].

Considering that chromatin disruption could cause adverse cytogenetic effects, the question that arises is whether the observed

V.I. Hatzi et al. / Mutation Research 793 (2015) 71–78 77

caffeine-induced chromatin fragmentation might lead to phe- nomena of aneuploidy and asymmetric cell divisions. To test the potential of caffeine to induce asymmetric cell division we used cytochalasin-B to block cytokinesis prior to iFISH analysis using centromere 11 and centromere 17 probes (CEP11, CEP17). This approach ensures that the caffeine-induced asymmetrically divided cells arise from the same normal parental cell and also that the divided nuclei remain within the same cytoplasm. CEP11 and CEP17 probes were chosen since they host numerous critical genes related related to carcinogenesis, autism, cell cycle regulation and mitotic spindle formation: ATM (11q22-q23), autism-related gene neurexin 1 (chrom 11), bladder neoplasm genes (11p), BRCA1 (17q21), p53 (17p13.1) and Aurora-B Kinase gene (17p13), HER2 (17q12), RAD51C (17q22), RARA (17q21). This combined method- ology enables the visualization of centromeric regions in the micronuclei as well as the monitoring of the chemically-induced asymmetric cell divisions. The combination of micronucleus assay with iFISH has been applied previously in cytogenetic studies for low-dose radiation exposure evaluation in human PBLs [47], for detection of hyperdiploidy and chromosome breakage in inter- phase human PBLs following exposure to hydroquinone [48], as well as for detection of chromosome malsegregation to the daugh- ter nuclei in cytokinesis-blocked transgenic mouse splenocytes [49]. In the present study, the application of cytochalasin-B treated PBLs prior to iFISH analysis enables not only the detection of the chromosomal breakage in the micronuclei but it also ensures that the caffeine-induced asymmetrically divided cells arise from mis- alignment of chromosomes of the same parental cell, enabling thus the detection of the abnormally divided daughter cells with a bal- anced profile.

Interestingly, using this methodology, we present the first evi- dence that low concentrations of caffeine (≥10 �M) induce a statistically significant (OR = 5.60, 95% CI = 2.12–14.82, P = 0.0003) increase in asymmetric cell division in healthy human cells in vitro (Table 4). This result might be critical since asymmetric cell divi- sion is highly associated with aneuploidy, which is a common cause of poor reproductive outcomes in humans [50] and severe birth defects [1]. The in vitro effects of caffeine on asymmet- ric cell division have been reported previously but only for high caffeine concentrations (10 mM) in cancer cells (HeLa and U2OS cells) [51]. The association of caffeine with increased frequencies of aneuploidy, diploidy and the duplication phenotype of human male sperm has been also demonstrated [50]. The caffeine-induced asymmetric cell division that we observe at a low concentration (10 �M) (Table 4), without an apparent induction of chromatid damage (Table 1), suggests that the underlying mechanism of asymmetric cell division may not be directly associated with the induction of chromatid breaks. Interestingly, the biological effects of caffeine have been also linked with detachment and fragmenta- tion of kinetochores of mammalian chromosomes [52]. In fact, the effect of low concentrations of caffeine on asymmetric cell division could arise from the effect of the chemical on mitotic spindle check- point and more specifically on kinetochores function [50,53–55].

Although caffeine does not accumulate in the human body, even after high consumption of coffee, an inter-individual variation in caffeine half life time has been reported [56]. An increase in the plasma concentration has been also observed in the last half of gestation in rabbits [57]. A gradual prolongation of the half-time of caffeine was also shown during pregnancy in humans [58] and caffeine clearance from the human body had increased by more than three-fold at a time period of 2–12 weeks [59]. Therefore, the increased half-life of caffeine during gestation may have an impact on the data obtained by meta-analysis presented by Greenwood et al. [1] and opens up new questions on the fate of caffeine-induced damage in the human body.

5. Conclusions

The results of the present study imply that a 24 h in vitro exposure of PBLs to caffeine at concentrations ≥1 mM induces replication fork progression delay and, particularly at high concen- trations, it allows entrance of cells at mitosis before completion of DNA replication. The presence of stalled replication forks during G2/M transition and the induction of chromatin condensation give rise to a failure of chromatin organization and chromatid breaks without apparent formation of chromosome exchanges. According to the model we propose, caffeine-induced damaged cells are not blocked at the G2/M checkpoint, since abrogation of G2/M check- point following exposure of PBLs to caffeine does not reveal an effect of low concentrations of caffeine. However, a 24 hour expo- sure of cycling PBLs to low concentrations of caffeine (≥10 �M) induces asymmetric cell division in vitro, as observed by means of conventional metaphase analysis and by a new approach that combines the CBMN assay with iFISH. Interestingly, the results suggest that the observed asymmetric cell division is caused by misalignment of chromosomes through a mechanism unrelated to the formation of chromatid breaks. The application of a combined methodology using CBMN-assay with iFISH is also proposed as a valuable tool for the estimation of chemically-induced asymmetric cell division.

Conflict of interest

The authors declare no conflict of interest and no competing interest.

Acknowledgements

The authors acknowledge funding from the European Social Fund - ESF and Greek National funds, through the Operational Pro- gram THALIS – UOA, MIS: 377177, as well as the Greek General Secretariat for Research and Technology and the European Regional Development Fund, under the Action “Development Grants For Research Institutions–KRIPIS” of OPCE II. The authors would like to especially thank Vasiliki Karkoulia, MSc Nursing, NCSR “Demokri- tos”, for peripheral blood venipuncture from the healthy volunteers that participated in the present study.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.mrgentox.2015. 08.002

References

[1] D.C. Greenwood, N.J. Thatcher, J. Ye, L. Garrard, G. Keogh, L.G. King, J.E. Cade, Caffeine intake during pregnancy and adverse birth outcomes: a systematic review and dose-response meta-analysis, Eur. J. Epidemiol. 29 (2014) 725–734.

[2] F. Zeidan-Chulia, D.P. Gelain, E.A. Kolling, J.L. Rybarczyk-Filho, P. Ambrosi, S.R. Terra, A.S. Pires, J.B. da Rocha, G.A. Behr, J.C. Moreira, Major components of energy drinks (caffeine, taurine, and guarana) exert cytotoxic effects on human neuronal SH-SY5Y cells by decreasing reactive oxygen species production, Oxid. Med. Cell. Longev. 2013 (2013) 791–795.

[3] S. Uccella, A. Mariani, A.H. Wang, R.A. Vierkant, W.A. Cliby, K. Robien, K.E. Anderson, J.R. Cerhan, Intake of coffee, caffeine and other methylxanthines and risk of Type I vs Type II endometrial cancer, Br. J. Cancer 109 (2013) 1908–1913.

[4] S.Y. Suh, Y.S. Choi, S.C. Oh, Y.S. Kim, K. Cho, W.K. Bae, J.H. Lee, A.R. Seo, H.Y. Ahn, Caffeine as an adjuvant therapy to opioids in cancer pain: a randomized, double-blind, placebo-controlled trial, J. Pain. Symptom Manage 46 (2013) 474–482.

[5] J.V. Aranda, K. Beharry, G.B. Valencia, G. Natarajan, J. Davis, Caffeine impact on neonatal morbidities, J. Matern. Fetal. Neonatal. Med. 23 (Suppl 3) (2010) 20–23.

78 V.I. Hatzi et al. / Mutation Research 793 (2015) 71–78

[6] K.A. Hahn, L.A. Wise, K.J. Rothman, E.M. Mikkelsen, S.B. Brogly, H.T. Sorensen, A.H. Riis, E.E. Hatch, Caffeine and caffeinated beverage consumption and risk of spontaneous abortion, Hum. Reprod. 30 (5) (2015) 1246–1255.

[7] J.C. Verster, Caffeine consumption in children, adolescents and adults, Curr. Drug Abuse Rev. 7 (2015) 133–134.

[8] J.R. Whayne Tf, Coffee: a selected overview of beneficial or harmful effects on the cardiovascular system? Curr. Vasc. Pharmacol. (2014) 1, Epub ahead of print].

[9] B.A. Kihlman, H.C. Andersson, Effects of caffeine on chromosomes in cells of higher eukaryotic organisms, Rev. Environ. Health 7 (1987) 279–382.

[10] B.A. Kihlman, Effects of caffeine on the genetic material, Mutat. Res. 26 (1974) 53–71.

[11] L.M. Stoilov, L.H. Mullenders, A.T. Natarajan, Caffeine potentiates or protects against radiation-induced DNA and chromosomal damage in human lymphocytes depending on temperature and concentration, Mutat. Res. 311 (1994) 169–174.

[12] D.M. Graham, Caffeine-its identity, dietary sources, intake and biological effects, Nutr. Rev. 36 (1978) 97–102.

[13] P. Nawrot, S. Jordan, J. Eastwood, J. Rotstein, A. Hugenholtz, M. Feeley, Effects of caffeine on human health, Food Addit Contam 20 (2003) 1–30.

[14] Y. Shiraishi, K. Yamamoto, A.A. Sandberg, Effects of caffeine on chromosome aberrations and sister-chromatid exchanges induced by mitomycin C in BrdU-labeled human chromosomes, Mutat. Res. 62 (1979) 139–149.

[15] K. Hansson, A.T. Natarajan, B.A. Kihlman, Effect of caffeine in G2 on X-ray-induced chromosomal aberrations and mitotic inhibition in ataxia telangiectasia fibroblast and lymphoblastoid cells, Hum. Genet. 67 (1984) 329–335.

[16] A.T. Natarajan, G. Obe, F.N. Dulout, The effect of caffeine posttreatment on X-ray-induced chromosomal aberrations in human blood lymphocytes in vitro, Hum. Genet. 54 (1980) 183–189.

[17] T.G. Natarajan, N. Ganesan, P. Carter-Nolan, C.A. Tucker, P.G. Shields, L.L. Adams-Campbell, Gamma-Radiation-induced chromosomal mutagen sensitivity is associated with breast cancer risk in African–American women: caffeine modulates the outcome of mutagen sensitivity assay, Cancer Epidemiol. Biomarkers Prev. 15 (2006) 437–442.

[18] W. Kuhlmann, H.G. Fromme, E.M. Heege, W. Ostertag, The mutagenic action of caffeine in higher organisms, Cancer Res. 28 (1968) 2375–2389.

[19] B.A. Kihlman, Caffeine and Chromosomes, Elsevier, Amsterdam, 1977. [20] B.A. Kihlman, K. Hansson, H.C. Andersson, The effects of post-treatments with

caffeine during S and G2 on the frequencies of chromosomal aberrations induced by thiotepa in root tips of Vicia faba and in human lymphocytes in vitro, Mutat. Res. 104 (1982) 323–330.

[21] B.A. Kihlman, S. Sturelid, B. Hartley-Asp, K. Nilsson, The enhancement by caffeine of the frequencies of chromosomal aberrations induced in plant and animal cells by chemical and physical agents, Mutat. Res. 26 (1974) 105–122.

[22] F. Darroudi, A.T. Natarajan, Cytological characterization of chinese hamster ovary X-ray-sensitive mutant cells xrs 5 and xrs 6 I. Induction of chromosomal aberrations by X-irradiation and its modulation with 3-aminobenzamide and caffeine, Mutat. Res. 177 (1987) 133–148.

[23] P.C. Kesavan, A.T. Natarajan, Protection and potentiation of radiation clastogenesis by caffeine: nature of possible initial events, Mutat. Res. 143 (1985) 61–68.

[24] S. Tornaletti, P. Russo, S. Parodi, A.M. Pedrini, Studies on DNA binding of caffeine and derivatives: evidence of intercalation by DNA-unwinding experiments, Biochim. Biophys. Acta 1007 (1989) 112–115.

[25] H.R. Pelayo, P. Lastres, C. De la Torre, Replication and G2 checkpoints: their response to caffeine, Planta 212 (2001) 444–453.

[26] I. Skoog, Is it dangerous or beneficial to drink coffee? Reflections on a meta-analysis on risk at birth and a population study on risk in late life, Eur. J. Epidemiol. 29 (2014) 665–666.

[27] H.M. Phelps, C.E. Phelps, Caffeine ingestion and breast cancer. A negative correlation, Cancer 61 (1988) 1051–1054.

[28] I.A. Ludwig, M.N. Clifford, M.E. Lean, H. Ashihara, A. Crozier, Coffee: biochemistry and potential impact on health, Food Funct. 5 (2014) 1695–1717.

[29] K.J. Stelzer, W.J. Koh, H. Kurtz, B.E. Greer, T.W. Griffin, Caffeine consumption is associated with decreased severe late toxicity after radiation to the pelvis, Int. J. Radiat. Oncol. Biol. Phys. 30 (1994) 411–417.

[30] K.B. Anjaria, B.S. Rao, Effect of caffeine on the genotoxic effects of gamma radiation and 4-NQO in diploid yeast, J. Environ. Pathol. Toxicol. Oncol. 20 (2001) 39–45.

[31] S.R. Musk, C.S. Downes, R.T. Johnson, Caffeine induces uncoordinated expression of cell cycle functions after ultraviolet irradiation. Accelerated cycle transit, sister chromatid exchanges and premature chromosome condensation in a transformed Indian muntjac cell line, J. Cell. Sci. 90 (Pt 4) (1988) 591–599.

[32] A.A. van Zeeland, Post-treatment with caffeine and the induction of gene mutations by ultraviolet irradiation and ethyl methanesulphonate in V-79 Chinese hamster cells in culture, Mutat. Res. 50 (1978) 145–151.

[33] L.G. Airapetian, A.V. Sevan’kaev, Effect of caffeine on the frequency of radiation-induced chromosome aberrations at different stages of the cell cycle, Tsitol. Genet. 11 (1977) 468–472.

[34] S. Shukla, K.B. Anjaria, N.N. Bhat, K. Shirsath, B. Streedevi, Effect of caffeine on radiation induced micronuclei in human lymphocytes, Radiat. Prot. Environ. 33 (2010) 195–198.

[35] S. Munne, C. Magli, M. Bahce, J. Fung, M. Legator, L. Morrison, J. Cohert, L. Gianaroli, Preimplantation diagnosis of the aneuploidies most commonly found in spontaneous abortions and live births: XY, 13, 14, 15, 16 21, 22, Prenat. Diagn. 18 (1998) 1459–1466.

[36] K. Haoud, S. Mellali, L. Gouas, A. Tchirkov, P. Vago, S. Moulessehoul, Prevalence of aneuploidies in products of spontaneous abortion: interest of FISH and MLPA, Morphologie 98 (2014) 40–46.

[37] M.D. Stephenson, K.A. Awartani, W.P. Robinson, Cytogenetic analysis of miscarriages from couples with recurrent miscarriage: a case-control study, Hum. Reprod. 17 (2002) 446–451.

[38] G.I. Terzoudi, V.I. Hatzi, K. Barszczewska, K.N. Manola, C. Stavropoulou, P. Angelakis, G.E. Pantelias, G2-checkpoint abrogation in irradiated lymphocytes: a new cytogenetic approach to assess individual radiosensitivity and predisposition to cancer, Int. J. Oncol. 35 (2009) 1223–1230.

[39] G.E. Pantelias, G.I. Terzoudi, A standardized G2-assay for the prediction of individual radiosensitivity, Radiother. Oncol. 101 (2011) 28–34.

[40] M. Fenech, A.A. Morley, Cytokinesis-block micronucleus method in human lymphocytes: effect of in vivo ageing and low dose X-irradiation, Mutat. Res. 161 (1986) 193–198.

[41] M. Fenech, The in vitro micronucleus technique, Mutat. Res. 455 (2000) 81–95. [42] M. Fenech, W.P. Chang, M. Kirsch-Volders, N. Holland, S. Bonassi, E. Zeiger,

H.U.M. project, HUMN project: detailed description of the scoring criteria for the cytokinesis-block micronucleus assay using isolated human lymphocyte cultures, Mutat. Res. 534 (2003) 65–75.

[43] W. Ostertag, B.J. Greif, Die erzeugung von chromatidenbrüchen durch coffein in leukocytenkulturen des menschen, Humangenetic 3 (1967) 282–294.

[44] F. Johansson, A. Lagerqvist, S. Filippi, F. Palitti, K. Erixon, T. Helleday, D. Jenssen, Caffeine delays replication fork progression and enhances UV-induced homologous recombination in Chinese hamster cell lines, DNA Repair 5 (2006) 1449–1458 (Amst).

[45] G.E. Pantelias, G.I. Terzoudi, Functional cell-cycle chromatin conformation changes in the presence of DNA damage result into chromatid breaks: a new insight in the formation of radiation-induced chromosomal aberrations based on the direct observation of interphase chromatin, Mutat. Res. 701 (2010) 27–37.

[46] G.I. Terzoudi, V.I. Hatzi, C. Donta-Bakoyianni, G.E. Pantelias, Chromatin dynamics during cell cycle mediate conversion of DNA damage into chromatid breaks and affect formation of chromosomal aberrations: biological and clinical significance, Mutat. Res. 711 (2011) 174–186.

[47] A. Baeyens, R. Swanson, O. Herd, E. Ainsbury, T. Mabhengu, P. Willem, H. Thierens, J.P. Slabbert, A. Vral, A semi-automated micronucleus-centromere assay to assess low-dose radiation exposure in human lymphocytes, Int. J. Radiat. Biol. 87 (2011) 923–931.

[48] D.A. Eastmond, D.S. Rupa, L.S. Hasegawa, Detection of hyperdiploidy and chromosome breakage in interphase human lymphocytes following exposure to the benzene metabolite hydroquinone using multicolor fluorescence in situ hybridization with DNA probes, Mutat. Res. 322 (1994) 9–20.

[49] J.J. Boei, A.T. Natarajan, Detection of chromosome malsegregation to the daughter nuclei in cytokinesis-blocked transgenic mouse splenocytes, Chromosome Res. 3 (1995) 45–53.

[50] W.A. Robbins, M.F. Vine, K.Y. Truong, R.B. Everson, Use of fluorescence in situ hybridization (FISH) to assess effects of smoking, caffeine, and alcohol on aneuploidy load in sperm of healthy men, Environ. Mol. Mutagen. 30 (2) (1997) 175–183.

[51] Y. Katsuki, S. Nakada, T. Yokoyama, I. Imoto, J. Inazawa, M. Nagasawa, S. Mizutani, Caffeine yields aneuploidy through asymmetrical cell division caused by misalignment of chromosomes, Cancer Sci. 99 (2008) 1539–1545.

[52] R.P. Zinkowski, S.L. McCune, R.D. Balczon, P.N. Rao, B.R. Brinkley, The centromere and aneuploidy: I. Caffeine-induced detachment and fragmentation of kinetochores of mammalian chromosomes, Prog. Clin. Biol. Res. 318 (1989) 43–60.

[53] E. Kitamura, K. Tanaka, Y. Kitamura, T.U. Tanaka, Kinetochore microtubule interaction during S phase in Saccharomyces cerevisiae, Genes Dev. 21 (2007) 3319–3330.

[54] B. Alberts, A. Johnson, J. Lewis, M. Raff, K. Roberts, P. Walter, Molecular Biology of the Cell, 4th edition, Garland Science, New York, 2002.

[55] S. Westermann, A. Schleiffer, Family matters: structural and functional conservation of centromere-associated proteins from yeast to humans, Trends Cell. Biol. 23 (2013) 260–269.

[56] IARC, Caffeine, IARC Monogr. Eval. Carcinog Risks Hum. 51 (1991) 291–390. [57] S.H. Dorrbecker, P.A. Kramer, B.R. Dorrbecker, J.R. Raye, Caffeine disposition in

the pregnant rabbit. II. Fetal distribution of caffeine and paraxanthine during chronic maternal caffeine administration, Dev. Pharmacol. Ther. 11 (1988) 118–124.

[58] A. Aldridge, J.V. Aranda, A.H. Neims, Caffeine metabolism in the newborn, Clin. Pharmacol. Ther. 25 (1979) 447–453.

[59] W.D. Parsons, J.G. Pelletier, Delayed elimination of caffeine by women in the last 2 weeks of pregnancy, Can. Med. Assoc. J. 127 (1982) 377–380.