Mutation Research 759 (2014) 9–20

Contents lists available at ScienceDirect

Mutation Research/Genetic Toxicology and Environmental Mutagenesis journal homepage: www.elsevier.com/locate/gentox Community address: www.elsevier.com/locate/mutres

Mercury specifically induces LINE-1 activity in a human neuroblastoma cell line Laleh Habibi a , Mohammad Ali Shokrgozar b,∗∗ , Mina Tabrizi a , Mohammad Hossein Modarressi a , Seyed Mohammad Akrami a,∗ a b

Department of Medical Genetics, School of Medicine, Tehran University of Medical Sciences, Tehran, Iran Iranian National Cell Bank, Pasteur Institute of Iran, Iran

a r t i c l e

i n f o

Article history: Received 11 April 2013 Received in revised form 25 June 2013 Accepted 5 July 2013 Available online 13 November 2013 Keywords: L1 retrotransposition Mercury Cobalt Neuron

a b s t r a c t L1 retro-elements comprise 17% of the human genome. Approximately 100 copies of these autonomous mobile elements are active in our DNA and can cause mutations, gene disruptions, and genomic instability. Therefore, human cells control the activities of L1 elements, in order to prevent their deleterious effects through different mechanisms. However, some toxic agents increase the retrotransposition activity of L1 elements in somatic cells. In order to identify specific effects of neurotoxic metals on L1 activity in neuronal cells, we studied the effects of mercury and cobalt on L1-retroelement activity by measuring levels of cellular transcription, protein expression, and genomic retrotransposition in a neuroblastoma cell line compared with the effects in three non-neuronal cell lines. Our results show that mercury increased the expression of L1 RNA, the activity of the L1 5 UTR, and L1 retrotransposition exclusively in the neuroblastoma cell line but not in non-neuronal cell lines. However, cobalt increased the expression of L1 RNA in neuroblastoma cells, HeLa cells, and wild-type human fibroblasts, and also increased the activity of the L1 5 UTR as well as the SV40 promoter in HeLa cells but not in neuroblastoma cells. Exposure to cobalt did not result in increased retrotransposition activity in HeLa cells or neuroblastoma cells. We conclude that non-toxic levels of the neurotoxic agent mercury could influence DNA by increasing L1 activities, specifically in neuronal cells, and may make these cells susceptible to neurodegeneration over time. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Long interspersed elements-1 (LINE-1, L1) retrotransposons are members of ancient mobile DNA families that still have the ability to retrotranspose autonomously through the human genome [1]. L1s occupy 17–20% of our genome, whereas coding genes occupy only 1% of our genome [2]. It is believed that these elements are responsible for shaping the genome [3], for creating retro-genes [4], and for the movement of other retro-elements, including Alus and SVAs [5,6]. Typical human full-length L1s possess the following elements: a 5 UTR promoter region, open reading frames ORF1 (coding for RNA chaperone) [7] and ORF2 (encoding a 150-kDa protein with three domains: reverse transcriptase (RT), endonuclease (EN) [8] and Cysteine-rich (unknown function)), a 3 UTR, and

∗ Corresponding author at: Department of Medical Genetics, School of Medicine, Tehran University of Medical Sciences, Keshavarz BLV, Poursina St. Tehran, Iran. Tel.: +98 21 88953005; fax: +98 21 88953005. ∗∗ Corresponding author at: Iranian National Cell Bank, Pasteur Institute of Iran, Tehran, Iran. Tel.: +98 21 66492595; fax: +98 21 66492595. E-mail addresses: [email protected] (M.A. Shokrgozar), [email protected] (S.M. Akrami). 1383-5718/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.mrgentox.2013.07.015

a polyA tail [9]. Autonomous activity of L1s proceeds via a copyand-paste mechanism called target primed reverse transcription (TPRT) [10]. This mechanism is not usually successful and creates many truncated copies (i.e. without 5 UTR, ORF1 or ORF2 regions) of L1 retrotransposons in each activity [11]. Therefore, it is estimated that just 80–100 copies of L1 retrotransposons have retained the ability to retrotranspose autonomously in human DNA [9]. Although L1 retrotransposition activities have played a role in shaping the genome in the course of evolution, it is well known that these activities may also produce genetic disorders, primarily as a result of gene interruption. Known genetic disorders that have been shown to correlate directly with L1 retrotransposition are Duchenne muscular dystrophy [12], haemophilia [13], and neurofibromatosis [14]. Recently, L1 elements have been found inserted in tumour-suppressor genes, which led scientists to propose the involvement of L1 in cancer pathogenesis [15,16]. For this reason, L1 retrotransposition is blocked in most human somatic cells and multiple cellular defense mechanisms have come into existence through evolution to avoid the abovementioned genomic problems and instabilities caused by L1 activity [17,18]. Despite mechanisms developed by our cells to restrict L1 retrotransposition [18] it is believed that external factors could stimulate this process, besides basic cellular activities. Effects of heavy metals

10

L. Habibi et al. / Mutation Research 759 (2014) 9–20

including mercury, cobalt, and cadmium on L1 retrotransposition have been described in HeLa and NIH3T3 cells in transient and stable assays [19]. A 2.5-fold increase in L1 retrotransposition was reported after addition of nickel for 48 h to HeLa cells [20]. Another group showed the 1-h effect of oxidative stress on stimulating L1 retrotransposition in a neuroblastoma cell line [21]. The effect of carcinogenic agents and magnetic fields on L1 retrotransposition has been described recently [22,23]. Harmful heavy metals such as mercury (Hg) and cobalt (Co) are ubiquitous in the environment and could affect our health by threatening specific organs [24,25]. Hg is among the most widespread neurotoxic heavy metals [26] in our environment, entering our body mainly through seafood, dental filling plaques, and other sources. This metal can pass through the blood–brain barrier, accumulate in neurons, and be converted to Hg2+ ions. These ions mainly interact with neuronal metabolism, causing oxidative stress and DNA damage [27,28]. It has been reported that mercury ions can persist and influence neuronal health for many years after their first entry into cells [27]. Cobalt is an essential metal for haematopoiesis [29], but it is carcinogenic [30] and high concentrations of this metal result in oxidative stress. It can also induce apoptosis in different cell types, including neurons [31]. In this study, we tried to measure the specific effect of two previously studied toxic heavy metals, Hg and Co, in inducing expression of endogenous L1s and L1 retrotransposition in a neuronal cell line. We assumed that the induction of L1 activity by these neurotoxic metals may threaten the integrity of neuronal DNA and thus be one of the mechanisms behind Hg- and Co-induced neurotoxicity [32]. To test our hypothesis, we measured endogenous L1 expression and exogenous retrotransposition as well as L1 5 UTR activity in a human neuroblastoma cell line [33] and three different nonneuronal cell lines after treatment with different concentrations of the metals. 2. Materials and methods 2.1. Cells The BE (2)-M17 human neuroblastoma cell line (Sigma, 95011816) was grown in DMEM: F12 (1:1) with 15% FBS, 1% non-essential amino acids, 1% l-glutamine and 1% penicillin–streptomycin (GIBCO); HeLa cells were grown in DMEM with 10% FBS (GIBCO); fibroblasts from ATM patients and from controls were grown in IMEM with 20% FBS and 1% penicillin–streptomycin (GIBCO). All cells were incubated at 37 ◦ C in an atmosphere containing 5% CO2 .

with expression of the house-keeping gene GAPDH, which would not change after induction in different cells (internal control). Real-time PCR was performed in triplicate for each sample and its related nonRT sample (RNA samples in cDNA synthesis process that lack reverse transcriptase in the reaction) as a control for any DNA contamination. The reaction was run in an ABI (Applied Biosystem) machine as follows: 500 ng of DNA, 10 ␮l of Sybr green (Gotaq, Promega), 300 nM final concentration for forward and reverse primers, up to 20 ␮l double-distilled water, and a PCR programme as follows: 95 ◦ C for 2 min, 95 ◦ C 15 s, 57 ◦ C 30 s and 72 ◦ C 15 s, for 40 cycles. Data were analyzed with the CT equation. GAPDH was used as an internal control. The fold-change for N51 PCR products was calculated in treated vs non-treated samples in each group of cells.

2.4. Protein extraction and Western blotting NB, HeLa, HF WT and HF D cells were grown as duplicates for 7 days in media containing Hg and Co. The cells were trypsinized and centrifuged at 1500 × g, at 4 ◦ C. A mixture of the following reagents was added to the cell pellets: RIPA lysis buffer (Radio Immuno Precipitation Assay buffer) (92.8%) (Sigma), cocktail proteinase inhibitor 2 and 3 (each reagent 0.92%) (Sigma), PMSF (phenylmethylsulfonyl fluoride) as protease inhibitor 100 mM (0.46%)(Sigma), lithium acetylide/ethylenediamine complex 25× (4.6%) (Sigma), and 2-mercaptoethanol (0.23%) (Sigma). Dissolved pellets were incubated for 15 min on ice and then centrifuged at 18,000 × g at 4 ◦ C for 10 min. The protein in the supernatant solution was quantified by use of the BSA protein-quantification kit (Life technology). Western blot was accomplished by use of a 10% polyacrylamide gel. Two hundred ␮g/ml of whole lysate protein from NB, HF WT, HF D and 80 ␮g/ml from HeLa cells were loaded onto the gel. The proteins were run together with dual-colour protein ladder (Invitrogen) for 90 min at 100 V. Transfer was done at 250 mA for 70 min. The nitrocellulose paper was blocked for 2 h with 10% milk, then incubated overnight with L1-ORF1 antibody (from Garcia-Perez JL lab) (5:10,000) at 4 ◦ C. After washing the membrane with TBS-T, secondary antibody, FITC-tagged IgG from rabbit (1:10,000) was added and incubated at room temperature (RT) for 1 h. After washing the membrane, ORF1 protein was checked with ODYSSEY (Life technology). The membrane was then incubated with actin antibody (1:20,000) and Cy5-tagged IgG from mouse as secondary antibody, simultaneously for 1 h at RT. The membrane was subsequently checked with ODYSSEY.

2.5. Luciferase assay NB, HeLa, HF WT and HF D cells were first grown in media containing heavy metals for 5 days. On day 6, the cells were seeded in 12-well plates (20,000/well). Transfection was done the next day under two different conditions: (1) transfection in media containing Hg and Co (the media was replaced again with media containing Hg and Co for this group of cells before transfection) and (2) transfection in normal media (changing the media containing Hg and Co with normal media in this group of cells before transfection) (Fig. 3a). The transfection included two co-transfection processes: (1) L1 5 UTR firefly-tagged vector plus Renilla vector (0.5 ␮g of L1 5 UTR and 0.2 ␮g of Renilla vector), and (2) SV40 promoter firefly-tagged vector plus Renilla vector (0.5 ␮g of L1 5 UTR and 0.2 ␮g of Renilla vector), 100 ␮l of OPTIMEM (GIBCO) and 30 ␮l of Fugene 6 (Promega). The transfections were done according to the Fugene-6 protocol. Twenty hours later the cells were lysed and the luciferase assay was performed with the Promega luciferase assay kit, following the protocol.

2.2. Metal solutions Four stock solutions of mercury (Hg) and cobalt (Co) [34] were prepared, as follows: One mg HgCl2 (Sigma–Aldrich) was dissolved in 10 ml DMEM: F12 serum-free media. Stock solutions of 1 ␮g/L, 5 ␮g/L, 10 ␮g/L, and 15 ␮g/L were prepared in specific complete media for each cell type. One mg CoCl2 (Sigma–Aldrich) was dissolved in 10 ml DMEM: F12 serum-free media. Stock solutions of 0.1 ␮g/L, 5 ␮g/L, 10 ␮g/L, and 20 ␮g/L were prepared in specific complete media for each cell type. 2.3. Total RNA extraction, cDNA synthesis and qRT-PCR BE(2)-M17 (NB) cells, HeLa cells, normal human fibroblasts (HF WT), and human fibroblasts derived from ATM patients (HF D) were seeded and grown for 7 days in the presence of different concentrations of Hg and Co (Fig. 1a). Total RNA was extracted by use of the TRIZOL (Sigma) protocol. To eliminate DNA from RNA samples, 1 ␮g of RNA was treated with DNAse I (Invitrogen). Total cDNA was synthesized from 1 ␮g of DNAse I-treated RNA with the ABI cDNA synthesis kit with random hexamer primers. Five hundred ng of cDNA was used for qRT-PCR. Real-time PCR was conducted for the first segment of the LINE-1 element, by using N51 primer set F: GAATGATTTTGACGAGCTGAGAGAA, R: GTCCTCCCGTAGCTCAGAGTAATT (Sigma) (Fig. 1a) and GAPDH primers as an internal control, F: TGCACCACCAACTGCTTAGC, R: GGCATGGACTGTGGTCATGAG (Sigma). The N51 primer set recognizes the border region between L1 5 UTR and ORF1 in most of the active L1 elements in human genome. The extent of L1-RNA expression induced by heavy metals was measured by use of the N51 primer set and compared

2.6. Retrotransposition assay and vectors NB and HeLa cells (2 × 104 ) were seeded in 6-well plates. Sixteen hours later, the cells were transfected with the following vectors (Fig. 4a): Ks101/L1.3-sv(A)+: a vector containing the full-length human L1.3 element that is driven by L1 5 UTR promoter and tagged with the neomycin-resistance gene that is disrupted by an intron [35]. Kubc101/L1.3-sv(A)+: a vector containing the human L1.3 element that is driven with both UBC and L1 5 UTR promoters and tagged with the neomycin-resistance gene that is disrupted by an intron. Ks101/L1.3-sv(A)+/RT (−): a vector containing the full-length human L1.3 element that is driven by just the L1 5 UTR promoter, with a mutation in the reverse-transcriptase domain of the ORF2 gene in the L1 element and tagged with the neomycin-resistance gene that is disrupted by an intron. pBSKS-NEO: a vector that contains the neomycin-resistance gene. The cells were transfected with 1 ␮g of each vector, 30 ␮l of Fugene 6 (Promega) and 100 ␮l of OPTIMEM (GIBCO) according to the Fugene-6 protocol. After 24 h the transfection media was removed and replaced with media containing heavy metals (Fig. 4a). Forty-eight hours later the cells were grown in the presence of metals in neomycin-containing media for 14 days (the amount of neomycin was 500 ␮g/ml for NB cells and 400 ␮g/ml for HeLa cells). The selection media was changed every 24 h. After day 14, the media was removed and the cells washed with PBS (to remove the dead cells). Two ml of fixative was added to each well. After 30 min the fixative was removed and replaced with crystal violet dye. Five minutes later, the dye was washed out and the stained colonies were counted.

L. Habibi et al. / Mutation Research 759 (2014) 9–20

11

Fig. 1. Endogenous LINE-1-RNA expression analysis after treatment of different cells with mercury and cobalt. (a) Schematic view of the LINE-1 retro-element, position of the N51 primer and transcription analysis. (b) Graphs showing the effect of mercury at toxic and non-toxic concentrations on endogenous LINE-1 expression. Mercury up-regulated the expression of L1 in NB cells (p-value < 0.05), but not in HeLa cells, wild-type human fibroblast and Human fibroblast from ATM patients (p-value > 0.53). (c) Graphs showing the effect of Co at toxic (p-value < 0.05) and non-toxic concentrations (p-value > 0.12) on endogenous LINE-1 expression. Data show that Co up-regulates L1 expression in both HeLa and NB cells (the concentrations 1 and 5 ␮g/L of Hg, and 10 and 20 ␮g/L of Co for analysis of the effect of metals on human fibroblasts were selected after analysis of the NB and HeLa cells, and after assessment of the toxicity of the metals to these cells). *p-Value < 0.05. NB: neuroblastoma, BE (2)-M17; HF WT: wild-type human fibroblast; HF D: human fibroblast from ATM patients.

12

L. Habibi et al. / Mutation Research 759 (2014) 9–20

Fig. 2. Analysis of endogenous expression of ORF1 protein after treatment of different cell types with mercury and cobalt. (A) Western-blot analysis for ORF1 protein and Actin after treating cells with different concentrations of Hg and Co. (B) Quantification of the intensity of ORF1 bands relative to that of Actin by use of ODYSSEY software. No clear up-regulation of ORF1 protein compared with Actin is seen in treated compared with untreated NB cells, HeLa cells, wild-type human fibroblasts, and fibroblasts derived from ATM patients (the concentrations of 1 and 5 ␮g/L Hg, and 10 and 20 ␮g/L Co for analysis of the effect of metals on human fibroblasts were selected after analysis of the NB and HeLa cells and after assessment of the toxicity of metals for these cells). NB: neuroblastoma, BE(2)-M17; HF WT: wild type human fibroblast; HF D: human fibroblast from ATM patients.

L. Habibi et al. / Mutation Research 759 (2014) 9–20

13

Fig. 3. The activation of the L1 5 UTR compared with the SV40 promoter after treatment of cells with mercury and cobalt. (a) Schematic view of study design and vectors. (b) Luciferase analysis for activation of L1 promoter with mercury. Results show that mercury at non-toxic concentration (1 ␮g/L) activates L1 5 UTR promoter (p-value < 0.001) in NB cells, but not in other cell types. Similar activation could not be seen for SV40 promoter in NB cells (p-value = 0.311) or for the 5 UTR in HeLa cells and fibroblasts (p-value > 0.421). (c) Cobalt treatment did not increase the activity of the L1 promoter in NB cells (p-value > 0.340); a non-significant increase in L1 5 UTR activity could be seen in HeLa cells (p-value > 0.511) and in wild-type human fibroblast (p-value = 0.169). This metal increased the activity of the SV40 promoter in HeLa cells (p-value < 0.05) and fibroblast from ATM patients (p-value < 0.05) (the concentrations of 1 and 5 ␮g/L for mercury, and 10 and 20 ␮g/L for cobalt for this assay were selected dependent on data from RT-qPCR and Western-blotting, and on the viability of the cells under stress conditions). *p-value < 0.05 NB: neuroblastoma, BE(2)-M17.

14

L. Habibi et al. / Mutation Research 759 (2014) 9–20

Fig. 3. (Continued).

2.7. MTT assay NB and HeLa cells (2 × 10 ) were seeded in 96-well plates (6 wells for each concentration) and incubated overnight. Media (200 ␮l) containing Hg and Co was then replaced with complete media without heavy metals. After 24 h of incubation at 37 ◦ C in a 5% CO2 atmosphere, 100 ␮l of 0.005 ␮g/ml MTT dye (Sigma–Aldrich) were added to the cells. The plate was incubated for 4 h at 37 ◦ C. The plate was read with an ELISA reader (BIOHIT.PIc) at 545 nm vs 630 nm (excitation/emission). Control cells were grown in normal media. 4

2.8. Propidium iodide (PI) cell-cycle assay NB cells (5 × 105 ) were seeded and incubated in the presence of Hg at concentrations of 1 and 5 ␮g/L. Seventy-two hours later, the cells were collected and fixed in 4.5 ml cold 70% ethanol. After 2 h, cells were pelleted by centrifugation at 600 × g for 5 min at room temperature (RT). Cell pellets were washed once with PBS and centrifuged at 600 × g for 5 min at RT. Pellets were re-suspended in a solution containing 200 ␮l PI dye (1 mg/ml, Sigma–Aldrich, USA), 10 mL Triton X-100 and 2 mg RNase (Bioneer, South Korea). Cells were incubated for 30 min in 37 ◦ C. PI emission was then read at 617 nm wavelength by means of flow cytometry (Partek-Germany). 2.9. Statistical analysis Real time PCRs for each cell type and for the two metals were done in three independent assays and each PCR in triplicate; the Western blotting and the retrotransposition assay were done in two independent experiments and each assay in duplicate; promoter activities were tested once as quadruple; the MTT assay in three independent experiments each time as hexad; and the PI test in two independent assays, each time in duplicate. All data were analyzed after checking their symmetry coefficient by use of ANOVA (Analysis of variance) (post hoc, Dunnett (2-sided), and LSD tests) in SPSS 18.0 software.

3. Results 3.1. Effect of mercury and cobalt on endogenous L1-expression in different cell lines Human neuroblastoma (BE(2)-M17) (NB) and HeLa cells were grown in the presence of Hg (1, 5, 10, 15 ␮g/L) or Co (0.1, 5, 10,

20 ␮g/L); wild-type human fibroblasts (HF WT), and fibroblasts derived from an ATM patient (HF D) were grown in the presence of Hg (1 and 5 ␮g/L) or Co (10 and 20 ␮g/L). RNA from treated cells was extracted after 7 days and qRT-PCR was conducted for the first part of the L1 element (last part of L1 5 UTR and first part of the ORF1 region) (Fig. 1a). NB cells under Hg-stress had significantly increased L1-RNA expression, up to 1.46- and 1.48-fold for the concentrations 1 ␮g/L and 5 ␮g/L (p-value = 0.021), respectively. An almost two-fold increase in L1-RNA expression was seen with Hg at 10 ␮g/L (p-value = 0.050). Any significant up-regulation of L1 mRNA could not be observed in non-neuronal cell lines including HeLa, HF WT, and HF D (Fig. 1b). Treatment of the cells with Co increased the expression of L1 RNA in NB cells up to 1.58 and 1.68-fold (not significant) for nontoxic concentrations (0.1 and 5 ␮g/L, respectively) and more than 2-fold at toxic concentrations (10 and 20 ␮g/L) (p-value < 0.05). In HeLa cells the effect of Co on L1-RNA expression was more prominent, with an increase of more than 2-fold at lower concentrations (0.1 and 5 ␮g/L, not significant), three-fold at 10 ␮g/L (p-value = 0.04) and more than 5-fold at the most toxic concentration (20 ␮g/L, p-value < 0.001) (Fig. 1c). The fibroblasts were less responsive to Co effects; HF WT showed non-significant increases up to 1.62-fold at a concentration of 20 ␮g/L for L1 expression, however HF D showed a non-significant down-regulation of L1-mRNA after treatment with Co (Fig. 1c). 3.2. Effects of mercury and cobalt on endogenous L1-protein expression (ORF1) To analyze the effects of Hg and Co on endogenous L1-protein expression (ORF1) in NB cells compared with non-neuronal cell lines, we grew NB and HeLa cells in the presence of four concentrations and HF WT and HF D with two concentrations of Hg and Co for 7 days (Fig. 2). Overall, the amount of ORF1 protein in NB cells was

L. Habibi et al. / Mutation Research 759 (2014) 9–20

15

Fig. 4. Retrotransposition assay for NB and HeLa cells treated with mercury and cobalt. (a) Schematic view of the mechanism of the retrotransposition assay, and study design. (b) Without treatment the rate of retrotransposition in NB cells is less than in HeLa cells. Mercury (1 ␮g/L) caused an increase in the number of neomycin-resistant colonies of NB cells, but it did not have any significant effect on HeLa cells. (c) Treatment with Co does not increase the number of neomycin-resistant colonies in NB and HeLa cells. The rate of retrotransposition in human fibroblasts is basically very low, so we could not check the effect of metals on exogenous L1 retrotransposition in these cells. HeLa cells were used as control since they support retrotransposition. Ks101/L1.3-sv(A)+ (this vector contains 5 UTR as a promoter for L1); Kubc101/L1.3-sv(A)+ (this vector contains UBC + 5 UTR as a promoter for L1); Ks101/L1.3-sv(A)+/RT (−) (this vector carries a mutation in the reverse-transcriptase domain of ORF2 and is used as a negative control); p-BSKS-NEO (this vector is used to check the efficiency of retrotransposition). NB: neuroblastoma, BE(2)-M17.

less than that in HeLa cells in control conditions (concordant with RNA levels observed, data not shown). HF WT and HF D cells did not show any ORF1-protein expression, as previously reported [36,37]. We also did not find any significant difference in the amount of ORF1 protein between treatment and control conditions for both Hg and Co in all groups of cells (Fig. 2). 3.3. Effects of mercury and cobalt on L1 5 UTR-activity compared with non-specific SV40 promoter-activity in NB and non-neuronal cell lines In order to search for specific effects of Hg and Co on the activation of the L1 5 UTR compared with a general promoter (SV40) in NB and non-neuronal cell lines, we grew selected cells for 5 days at two concentrations of Hg (1 and 5 ␮g/L) and Co (10 and 20 ␮g/L). These conditions were chosen for this part based on data obtained from L1 expression studies and depending on the toxicity of metals for the cells. After seeding the cells at day 6 in media containing Hg or Co, co-transfection (5 UTR + Renilla and SV40 + Renilla) was carried out in two different groups of cells. One group was maintained in media containing Hg or Co; these cells were grown for 7 days in the presence of the metal and transfection was performed in

media containing Hg or Co. The second group was grown in media containing Hg or Co, then switched to normal media, and transfection was performed subsequently. In this second group, cells were grown for 20 h under non-stressed conditions (Fig. 3a) and subsequently transfected. The following two conditions were tested: (1) to assess the effects of removing stress conditions on the activity of the L1 5 UTR, (2) to see if Hg and Co ions in cell-culture media have an effect on transfection efficiency of the vectors (Fig. 3a). The results show that Hg at a concentration of 1 ␮g/L (pvalue < 0.001) but not 5 ␮g/L (p-value = 0.071) gives rise to a significant 3.1-fold increase in activity of L1 5 UTR in NB cells. This effect could not be seen in any other non-neuronal cell line (p-value > 0.425). The SV40 promoter did not show a significant increase (p-value > 0.311) in activity in NB cells after treatment with Hg (Fig. 3b), possibly showing the specific effect of Hg on L1 promoter activity in NB cells. Removing the media containing Hg for 20 h resulted in decreased activity of the L1 5 UTR down to 1.42fold at a concentration of 1 ␮g/L in NB cells (p-value = 0.001) and a non-significant increase in activity of the SV40 promoter in NB (p-value = 0.411) (Suppl. Fig. 1a). Cobalt did not have any effect on L1 5 UTR activity in NB cells (p-value > 0.340). A non-significant increase (p-value > 0.511) in

16

L. Habibi et al. / Mutation Research 759 (2014) 9–20

Fig. 4. (Continued).

activity of the L1 promoter was seen in HeLa cells (1.5-fold, at concentration 10 ␮g/L) and in HF WT cells (p-value = 0.169) (1.42fold, at 20 ␮g/L). Cobalt also increased the activity of the SV40 promoter in HeLa cells: 1.33-fold at a concentration of 10 ␮g/L (p-value > 0.849) and more than 3-fold at 20 ␮g/L (p-value < 0.05) (Fig. 3c). This finding ruled out any Co specificity for activating the L1 5 UTR in HeLa cells. Removing Co from the media did not show any remarkable increase in 5 UTR activity in NB cells (pvalue > 0.170), but it caused a significant increase (p-value < 0.05) in the amount of L1 promoter activity in HeLa cells and a decrease in HF WT cells (p-value < 0.05). This effect was observed for the SV40 promoter-activity in HeLa cells but not in the HF WT cell line (Suppl. Fig. 1b). 3.4. Effect of mercury and cobalt on L1 retrotransposition in NB cells compared with that in HeLa cells The L1 retrotransposition assay was done after transfection of NB and HeLa cells with L1.3 neomycin-tagged vectors [38]. In this study, human fibroblasts (wild-type and ATM-derived) were omitted because previous reports have shown that these cell types do not support L1 retrotransposition [36,37]. HeLa cells were used as a positive control for NB because, intrinsically and without induction, these cells can support L1 retrotransposition [38]. The cells were transfected with Ks101/L1.3-sv(A)+ (full-length L1.3 NeomycinR -tagged vector), Kubc101/L1.3-sv(A)+ (full length

L1.3 NeomycinR -tagged vector – the L1 element in this vector is driven by both 5 UTR and the UBC promoter), Ks101/L1.3/RT(−) (full length L1.3 NeomycinR -tagged vector – carries a mutation in the reverse-transcriptase (RT) domain of the ORF2 protein in the L1 element). The Ks101/L1.3/RT(−) vector cannot support L1 retrotransposition because of the defective RT domain and was used as a negative control. pBSKS-NEO vector (containing the NeomycinR gene) was used as a control of transfection efficiency. Both transfected cell types were grown in media containing Hg and Co together with neomycin selection for 14 days. Stained neomycin-resistant colonies were counted after this period of time. Resistant colonies showed the number of cells in which L1 retrotransposition had taken place; therefore, these cells could express the neomycin-resistance gene and grow in neomycin-containing selection media (Fig. 4a). Counting neomycin-resistant colonies showed that Hg increased L1 retrotransposition up to 2.3-fold in NB cells at the non-toxic concentration of 1 ␮g/L, and 1.64-fold at a concentration of 5 ␮g/L. These conditions did not seem to have any effect on the number of L1 retrotransposition events in HeLa cells (Fig. 4b). Media containing Hg increased the number of neomycinresistant colonies in NB cells transfected with pBSKS-NEO from a basal level up to 1.3-fold (Fig. 4a). This phenomenon may suggest a role for mercury in NB cell proliferation and may affect our retrotransposition data. In order to check for this effect, we performed a PI test for NB cells treated with mercury at concentrations of 1 and 5 ␮g/L. The result shows that no significant change (p-value > 0.109)

L. Habibi et al. / Mutation Research 759 (2014) 9–20

17

Fig. 4. (Continued).

could be seen in G1, S and G2 phases of the cell cycle between treated and untreated cells (Fig. 5). Cobalt did not increase the L1 retrotransposition process above the basal level (untreated cells) in both NB and HeLa cells (Fig. 4c). Control vectors (L1.3 RT (−) and pBSKS-NEO) in HeLa cells grown in media containing cobalt did not demonstrate any change. Although Co increased the expression of the NeomycinR -gene from the pBSKS-NEO vector above basal level in NB cells (untreated transfected cells with pBSKS-NEO vector), it seems that this event did

not have any effect on the number of NB cells that support retrotransposition in media containing Co (Fig. 4c). 4. Discussion Our genome has been a host to the LINE-1 elements for millions of years and it is believed that LINE-1 retrotranspositions have played an important role in shaping and restructuring the

Fig. 5. PI test to check the effect of mercury on proliferation of neuroblastoma cells. The data do not show any significant change (p-value > 0.109) in cell-cycle phase between untreated and mercury-treated cells at concentrations of 1 and 5 ␮g/L.

18

L. Habibi et al. / Mutation Research 759 (2014) 9–20

Fig. 6. Schematic overview of the results obtained in this study. (a) Effect of mercury (Hg2+ ions) on transcription, protein and retrotransposition level of the LINE-1 retroelement in different cell lines. Mercury increased the expression of L1 RNA, L1 5 UTR activity and retrotransposition specifically in NB cells, but not in HeLa cells, wild-type human fibroblasts or human fibroblasts from ATM patients. (b) Effect of cobalt (Co2+ ions) at the transcription, protein and retrotransposition level of the LINE-1 retro-element in different cell lines. Cobalt increased the amount of L1 RNA in NB, HeLa and wild-type human fibroblasts. It did not have any significant effect on L1-ORF1 protein level and 5 UTR activity in any of the cell types studied. Cobalt did not affect retrotransposition in NB and HeLa cells either. NB: Neuroblastoma, BE(2)-M17; HF: wild-type human fibroblasts; HF D: human fibroblasts from ATM patients.

eukaryotic genome [39]. However, L1 retrotransposition, including uncontrolled copying and pasting, pose a threat to our health [40]. L1 retrotransposition and expression is confined to the embryonic phase, the cancerous state, and to some somatic cell types including neurons [16,41,42]. In addition, it has been shown that some environmental stress factors, including heavy metals and H2 O2 , increased the frequency of L1 retrotransposition in NIH3T3 (stable assays in Kale et al.) and NB cells [19–21]. In this study, we tried to clarify the specific effect of Hg compared with Co on L1 activity in a neuronal cell line in comparison

with several non-neuronal cell lines. Our data show that Hg, at each concentration tested, significantly up-regulated L1-RNA expression only in NB cells but not in non-neuronal cells (Fig. 6a). The effect of Co on L1 expression was not cell-specific, as HeLa cells and HF WT cells, in addition to the neuronal cell line, showed significant upregulation of L1 RNA at toxic concentrations (Fig. 6b). Unlike our observations at the RNA level, no detectable change could be found at the L1-protein level (ORF1) after treatment with Hg or Co (Fig. 6a and b). The different results obtained in this part of our study could be related to the complexity of L1 biology. There is evidence that

L. Habibi et al. / Mutation Research 759 (2014) 9–20

19

Fig. 7. MTT assay after treatments with mercury and cobalt; toxic effects were assessed of mercury and cobalt at different concentrations on NB and HeLa cells. Mercury has a significant (p-value < 0.05) toxic effect on NB cells at concentrations of 5, 10 and 15 ␮g/L, but not at 1 ␮g/L. This metal did not show any significant (p-value > 0.134) toxic effects at selected concentrations on HeLa cells. Cobalt did not show any significant toxicity in both neuroblastoma and HeLa cells (p-value > 0.357) NB: Neuroblastoma, BE(2)-M17. *p-value < 0.05.

L1 RNA will undergo various processing events such as premature poly-adenylation [43] and multiple splicing processes to inhibit its activity [44]. Therefore, an increase in L1 RNA may not ultimately result in protein production. In addition, agents such as Co may affect RNA stability or induce hypomethylation of the L1 5 UTR in, e.g., HeLa cells without affecting L1-protein production. We should note also that the RT-qPCR technique cannot discriminate between full-length L1 RNA or RNA expressed from other non-mobile fulllength L1s [45], which may interfere also with the interpretation of data from real-time PCR and Western blot. The activity of the 5 UTR was stimulated by Hg, but only in NB cells. The specificity of this event was demonstrated when Hg treatment did not stimulate the SV40 promoter in these cells. In addition, removal of Hg stress decreased L1-promoter activity in NB cells. Co ions increased both L1 and SV40 promoter activity in HeLa cells, which demonstrates the nonspecific effect of this metal in this cell line. Interestingly, Co increased L1 but not SV40 promoter activity in HF WT cells. Removing Co from the media did not produce a significant change in 5 UTR activity in HeLa cells but decreased the activity of this promoter in HF WT cells. This specific effect of Hg on 5 UTR activity in NB cells may reflect other pathways besides known toxic routes [46] that Hg could activate to stimulate gene expression in neurons [47]. However Co may not have specific targets in neuronal cells as in other cell lines. Retrotransposition results were correlated with the data obtained for endogenous L1 expression and L1 5 UTR activation by Hg, specifically in NB cells (Fig. 6a). Our retrotransposition data for

HeLa cells confirmed the transient results from a previous study [19]. Indeed, this report about L1 activation in NIH3T3 cells after treatment with Hg is consistent with removing L1-silencing rather than its retrotransposition. In another study, we also observed removal of L1 silencing in non-dividing NB cells by Hg, but not other heavy metals [48]. Co did not increase retrotransposition in NB nor in HeLa cells (Fig. 6b). Our search for toxic effects of Hg and Co in NB and HeLa cell lines has shown that Hg but not Co has a significant toxic effect in NB cells after 24 h. Such toxicity was not seen in HeLa cells (Fig. 7). Importantly, data obtained for L1 activity in NB cells after treatment with Hg came from experiments with non-toxic concentrations. To the best of our knowledge, this is the first report to demonstrate a specific role of mercury, a widespread neurotoxic heavy metal, in L1 activity in NB cells. It has been accepted that the effect of non-toxic concentrations of Hg can be tolerated by neurons. Here we show that even these concentrations may result in structural changes in DNA by increasing the activities of mobile DNA elements. Such activities could affect the expression of genes in neurons and make cells susceptible to degeneration over time. Further experiments, including studies on the effect of these factors on the methylation status of the whole genome and more specifically the L1 5 UTR could help unravel mechanisms involved in enhancing L1 retrotransposition and endogenous expression in NB cells compared with non-neuronal cell lines.

20

L. Habibi et al. / Mutation Research 759 (2014) 9–20

Conflict of interest statement No competing financial interests exist. Acknowledgements We would like to thank Dr Garcia Perez JL and Dr Munoz Lopez M for their nice scientific discussion and providing vectors, Tehran University of Medical Sciences for financial support of this project (Grant number 9765) and Iran Health Ministry. 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.2013. 07.015. References [1] H.H. Kazazian Jr., J.V. Moran, The impact of L1 retrotransposons on the human genome, Nat. Genet. 19 (1998) 19–24. [2] E.S. Lander, L.M. Linton, B. Birren, C. Nusbaum, M.C. Zody, et al., Initial sequencing and analysis of the human genome, Nature 409 (2001) 860–921. [3] A.F. Smit, The origin of interspersed repeats in the human genome, Curr. Opin. Genet. Dev. 6 (1996) 743–748. [4] L. Lin, B. Chen, J. Dai, L1 elements, processed pseudogenes and retrogenes in mammalian genomes, IUBMB Life 58 (2006) 677–685. [5] M. Dewannieux, C. Esnault, T. Heidmann, LINE-mediated retrotransposition of marked Alu sequences, Nat. Genet. 35 (2003) 41–48. [6] E.M. Ostertag, J.L. Goodier, Y. Zhang, H.H. Kazazian Jr., SVA elements are nonautonomous retrotransposons that cause disease in humans, Am. J. Hum. Genet. 73 (2003) 1444–1451. [7] S. Martin, Nucleic acid chaperone properties of ORF1p from the non-LTR retrotransposon, LINE-1, RNA Biol. 7 (2010) 706–711. [8] Q. Feng, J.V. Moran, H.H. Kazazian Jr., J.D. Boeke, Human L1 retrotransposon encodes a conserved endonuclease required for retrotransposition, Cell 87 (1996) 905–916. [9] J.L. Goodier, H.H. Kazazian Jr., Retrotransposons revisited: the restraint and rehabilitation of parasites, Cell 135 (2008) 23–35. [10] G.J. Cost, Q. Feng, A. Jacquier, J.D. Boeke, Human, L1 element target-primed reverse transcription in vitro, EMBO J. 21 (2002) 5899–5910. [11] S. Boissinot, A. Entezam, A.V. Furano, Selection against deleterious line-1containing loci in the human lineage, Mol. Biol. Evol. 18 (2001) 926–935. [12] N. Narita, H. Nishio, Y. Kitoh, Y. Ishikawa, Y. Ishikawa, et al., Insertion of a 5 truncated L1 element into the 3 end of exon 44 of the dystrophin gene resulted in skipping of the exon during splicing in a case of Duchenne muscular dystrophy, J. Clin. Invest. 91 (1993) 1862–1867. [13] N. Van de Water, R. Williams, P. Ockelford, P. Browett, A 20.7 kb deletion within the factor VIII gene associated with LINE-1 element insertion, Thromb. Haemost. 79 (1998) 938–942. [14] K. Wimmer, T. Callens, A. Wernstedt, L. Messiaen, The NF1 gene contains hotspots for L1 endonuclease-dependent de novo insertion, PLoS Genet. 7 (2011) e1002371. [15] S. Solyom, A.D. Ewing, E.P. Rahrmann, T. Doucet, H.H. Nelson, et al., Extensive somatic L1 retrotransposition in colorectal tumors, Genome Res. 22 (2012) 2328–2338. ˜ [16] R. Shukla, K.R. Upton, M. Munoz-Lopez, D.J. Gerhardt, M.E. Fisher, et al., Endogenous retrotransposition activates oncogenic pathways in hepatocellular carcinoma, Cell 153 (2013) 101–111. [17] H.L. Levin, J.V. Moran, Dynamic interactions between transposable elements and their hosts, Nat. Rev. Genet. 12 (2011) 615–627. [18] J.M. Rosser, W. An, L1 expression and regulation in humans and rodents, Front. Biosci. (Elite Ed.) 4 (2012) 2203–2225. [19] S.P. Kale, L. Moore, P.L. Deininger, A.M. Roy-Engel, Heavy metals stimulate human LINE-1 retrotransposition, Int. J. Environ. Res. Public Health 2 (2005) 14–23. [20] M. El-Sawy, S.P. Kale, C. Dugan, T.Q. Nguyen, V. Belancio, et al., Nickel stimulates L1 retrotransposition by a post-transcriptional mechanism, J. Mol. Biol. 354 (2005) 246–257.

[21] G. Giorgi, P. Marcantonio, B. Del Re, LINE-1 retrotransposition in human neuroblastoma cells is affected by oxidative stress, Cell Tissue Res. 346 (2011) 383–391. [22] N. Okudaira, M. Goto, R. Yanobu-Takanashi, M. Tamura, A. An, et al., Involvement of retrotransposition of long interspersed nucleotide element1 in skin tumorigenesis induced by 7,12-dimethylbenz[a]anthracene and 12-O-tetradecanoylphorbol-13-acetate, Cancer Sci. 102 (2011) 2000–2006. [23] B. Del Re, P. Marcantonio, E. Gavoc¸i, F. Bersani, G. Giorgi, Assessing LINE-1 retrotransposition activity in neuroblastoma cells exposed to extremely lowfrequency pulsed magnetic fields, Mut. Res. 749 (2012) 76–81. [24] M. Aschner, S. Ceccatelli, Are neuropathological conditions relevant to ethylmercury exposure? Neurotox. Res. 18 (2010) 59–68. ´ ´ [25] A. Cebulska-Wasilewska, A. Panek, Z. Zabinski, P. Moszczynski, W.W. Au, Occupational exposure to mercury vapour on genotoxicity and DNA repair, Mutat. Res. 586 (2005) 102–114. [26] J. Segura Aguilar, R.M. Kostrzewa, Neurotoxins and neurotoxic species implicated in neurodegeneration, Neurotox. Res. 6 (2004) 615–630. [27] J. Mutter, A. Curth, J. Naumann, R. Deth, H. Walach, Does inorganic mercury play a role in Alzheimer’s disease? A systematic review and an integrated molecular mechanism, J. Alzheimers Dis. 22 (2012) 357–374. [28] E.Y. Ben-Ozer, A.J. Rosenspire, M.J. McCabe Jr., R.G. Worth, A.L. Kindzelskii, et al., Mercuric chloride damages cellular DNA by a non-apoptotic mechanism, Mutat. Res. 470 (2000) 19–27. [29] S. Okamoto, L.D. Eltis, The biological occurrence and trafficking of cobalt, Metallomics 3 (2011) 963–970. [30] M. De Boeck, M. Kirsch-Volders, D. Lison, Cobalt and antimony: genotoxicity and carcinogenicity, Mutat. Res. 533 (2003) 135–152. [31] J.M. Matés, J.A. Segura, F.J. Alonso, Márquez F J., Roles of dioxins and heavy metals in cancer and neurological diseases using ROS-mediated mechanisms, Free Radic. Biol. Med. 49 (2010) 1328–1341. [32] L. Habibi, S.M. Akrami, M.A. Shokrgozar, Does L1 retrotransposition cause neuronal loss in neurodegenerative disorders? Iran J. Med. Hypotheses Ideas 4 (2010) 4. [33] P. Shastry, A. Basu, M.S. Rajadhyaksha, Neuroblastoma cell lines – a versatile in vitro model in neurobiology, Int. J. Neurosci. 108 (2001) 109–126. [34] Goldfrank, Goldfrank’s Toxicologic Emergencies, 9/e, McGraw-Hill, United state, New York, 2011, pp. 688–767. [35] J.L. Garcia-Perez, M.C.N. Marchetto, A.R. Muotri, N.G. Coufal, F.H. Gage, et al., LINE-1 retrotransposition in human embryonic stem cells, Hum. Mol. Genet. 16 (2007) 1569–1577. ˜ [36] S. Wissing, M. Munoz-Lopez, A. Macia, Z. Yang, M. Montano, et al., Reprogramming somatic cells into iPS cells activates LINE-1 retroelement mobility, Hum. Mol. Genet. 21 (2012) 208–218. [37] X. Shi, V. Seluanov, V. Gorbunova, Cell divisions are required for L1 retrotransposition, Mol. Cell. Biol. 27 (2007) 1264–1270. [38] J.V. Moran, S.E. Holmes, T.P. Naas, R.J. DeBerardinis, J.D. Boeke, et al., High frequency retrotransposition in cultured mammalian cells, Cell 87 (1996) 917–927. [39] R. Cordaux, M.A. Batzer, The impact of retrotransposons on human genome evolution, Nat. Rev. Genet. 10 (2009) 691–703. [40] D.C. Hancks, H.H. Kazazian Jr., Active human retrotransposons: variation and disease, Curr. Opin. Genet. Dev. 22 (2012) 191–203. [41] J.A. Van den Hurk, I.C. Meij, M.C. Seleme, H. Kano, K. Nikopoulos, et al., L1 retrotransposition can occur early in human embryonic development, Hum. Mol. Genet. 16 (2007) 1587–1592. [42] J.K. Baillie, M.W. Barnett, K.R. Upton, D.J. Gerhardt, T.A. Richmond, et al., Somatic retrotransposition alters the genetic landscape of the human brain, Nature 47 (2011) 534–537. [43] V. Perepelitsa-Belancio, P. Deininger, RNA truncation by premature polyadenylation attenuates human mobile element activity, Nat. Genet. 35 (2003) 363–366. [44] V.P. Belancio, A.M. Roy-Engel, P. Deininger, The impact of multiple splice sites in human L1 elements, Gene 411 (2008) 38–45. [45] V.P. Belancio, D.J. Hedges, P. Deininger, LINE-1 RNA splicing and influences on mammalian gene expression, Nucl. Acids Res. 34 (2006) 1512–1521. [46] M. Valko, H. Morris, M.T. Cronin, Metals, toxicity and oxidative stress, J. Exp. Bot. 53 (2002) 1351–1365. [47] C. Bland, M.D. Rand, Methylmercury induces activation of Notch signaling, Neurotoxicology 27 (2006) 982–991. [48] L. Habibi, M.A. Shokrgozar, M. Motamedi, S.M. Akrami, Effect of heavy metals on silencing of engineered long interspersed element-1 retrotransposon in non-dividing neuroblastoma cell line, Iran. Biomed. J. 17 (2013) 171–178.