Gestational Age and Sex Influence the Susceptibility of Human Neural Progenitor Cells to Low Levels of MeHg

The developing nervous system is highly susceptible to methylmercury (MeHg), a widespread environmental neurotoxic contaminant. A wide range of morphological and functional outcomes have been described; however, there are still open questions regarding the mechanisms behind the developmental neurotoxic effects induced by low-level exposure. In the present study, we have examined the effects of nanomolar concentrations of MeHg on primary fetal human progenitor cells (hNPCs) with special focus on the role played by developmental stage and sex on the neurotoxic outcome. We found that neurospheres derived from earlier gestational time points exhibit higher susceptibility to MeHg, as they undergo apoptosis at a much lower dose (25 nM) as compared to neurospheres established from older fetuses (100 nM). At subapoptotic concentrations (10 nM), MeHg inhibited neuronal differentiation and maturation of hNPCs, as shown by a reduced number of Tuj1-positive cells and a visible reduction in neurite extension and cell migration, associated with a misregulation of Notch1 and BDNF signaling pathways. Interestingly, cells derived from male fetuses showed more severe alterations of neuronal morphology as compared to cells from females, indicating that the MeHg-induced impairment of neurite extension and cell migration is sex-dependent. Accordingly, the expression of the CDKL5 gene, a major factor regulating neurite outgrowth, was significantly more downregulated in male-derived cells. Altogether, gestational age and sex appear to be critical factors influencing in vitro hNPC sensitivity to low levels of MeHg. Electronic supplementary material The online version of this article (doi:10.1007/s12640-017-9786-x) contains supplementary material, which is available to authorized users.


Introduction
Methylmercury (MeHg) is a widespread environmental contaminant well known to be particularly harmful during nervous system development. The developmental neurotoxicity in humans has been recognized after poisoning catastrophes (Harada 1995), where children with neurological impairments were born by seemingly symptom-free women, highlighting that the developing nervous system is much more vulnerable than the adult one. Histopathological examinations during autopsies of MeHg exposed infants have shown alterations in specific brain areas, such as cerebellum and cerebral cortex (Roegge et al. 2006;Johansson et al. 2007), and signs of defects in neuronal organization and migration (Choi et al. 1978;Wilson et al. 2005;Fahrion et al. 2012).
Epidemiological data and behavioral studies on experimental animals exposed in utero have established that subcytotoxic doses that do not induce apoptosis or major histopathological signs still cause long-lasting impairments Johansson et al. 2007;Castoldi et al. 2008;Onishchenko et al. 2008). Moreover, behavioral studies in experimental animals, prenatally exposed to MeHg, showed in males but not in females reduced motor activity (Rossi et al. 1997; Giménez-Karin Edoff and Marilena Raciti are equally contributing authors. Electronic supplementary material The online version of this article (doi:10.1007/s12640-017-9786-x) contains supplementary material, which is available to authorized users. Llort et al. 2001) and depression-like behavior Onishchenko et al. 2008).
The molecular mechanisms behind low-level MeHg-induced developmental neurotoxicity have been studied in different experimental models both in vivo and in vitro, and neural stem and progenitor cells have been shown to be among the most sensitive targets (Tamm et al. 2006;Johansson et al. 2007;Tamm et al. 2008). We have found that rat embryonic NPCs exposed to very low concentrations of MeHg undergo alterations in proliferation capacity and increased susceptibility to oxidative stress and that these changes are heritable as they are present in daughter cells never directly exposed to the neurotoxicant . These cellular alterations are accompanied by changes in DNA methylation, suggesting the involvement of epigenetic mechanisms ). In addition, we found that male mice exposed to low levels of MeHg in utero exhibited reduced hippocampal neurogenesis even as adults and had fewer granule neurons in the dentate gyrus .
To further investigate the mechanisms involved in low doses of MeHg neurotoxicity, we examined putative sexrelated differences in the susceptibility to MeHg in primary human progenitor cells (hNPCs), with special focus on neuronal differentiation and maturation. Our data show that 10 nM MeHg inhibits neuronal differentiation and that the underlying mechanism probably targets the Notch signaling, a key regulator of neurogenesis (Louvi and Artavanis-Tsakonas 2006;Imayoshi et al. 2010).
There are several studies showing that MeHg disrupts neuronal migration (Heidemann et al. 2001;Moors et al. 2007;Moors et al. 2009;Guo et al. 2013), which may represent one of the main factors mediating MeHg developmental neurotoxicity. However, the mechanisms involved need to be further elucidated. Here we show that subapoptotic concentrations of MeHg induce an impaired migration associated to misexpression of cyclin-dependent kinase-like 5 (CDKL5), a key gene regulating neuronal morphogenesis and dendritic arborization by a mechanisms involving BDNF-Rac1 signaling (Chen et al. 2010). Moreover, MeHg interferes with neuronal maturation in a sex-dependent manner, as the observed alterations are more pronounced in cultures established from male fetuses.

Chemicals
All chemicals were and reagents were obtained from Life Technologies and Sigma-Aldrich unless otherwise stated. MeHg hydroxide was purchased from ALFA, Johnson Matthey, Karlsruhe, Germany.

Cell Culture
Postconception week (PCW) 8.5 hNPC cultures were established from human fetal central nervous system tissue. The Regional Ethics Committee, Stockholm, Sweden (nos. 2008/158-33/3, 2011/1101 approved the procedure. Briefly, cortical forebrain tissue was collected from clinical first trimester routine abortions, after obtaining informed consent by women undergoing termination of pregnancy. The human tissue was homogenized with a glass-Teflon homogenizer and cultured at 100,000-200,000 cells/ml in NS medium supplemented with 20 ng/ml epidermal growth factor (EGF), 20 ng/ml bFGF, and 10 ng/ml ciliary neurotrophic factor (CNTF) (all from R&D), as previously described (Åkesson et al. 2009). Neurospheres cultures were passaged every 7-14 days by using TrypLE Express (Life Technologies), and fresh medium was added twice a week. hNPC was expanded as free-floating neurospheres in Corning® non-treated culture dishes or Corning® ultra-low attachment culture dishes (100 mm × 20 mm, Sigma-Aldrich) and maintained in a humidified atmosphere at 37°C and 5% CO 2 . All the following experiments were performed in hNPCs between passages 5 and 10.
hNPCs from PCW 16 were obtained from Lonza (Verviers SPRL) and cultured as previously described (Moors et al. 2009;Moors et al. 2012) at 37°C and 5% CO 2 as a suspension culture in defined serum-free media composed of Dulbecco's modified Eagle medium (DMEM) and Hams F12 (3:1), supplemented with penicillin/streptomycin (50 U/ml), B27 1:50 (Invitrogen), 20 ng/ml EGF and 20 ng/ml recombinant human fibroblast growth factor (FGF; R&D Systems). Passaging was performed mechanically by cutting large spheres into smaller pieces using a McIlwain tissue chopper (Svendsen et al. 1998). Growing the human neural progenitor cells as neurospheres allows large numbers of cells to be expanded in small volumes of medium. However, to achieve homogenous levels of MeHg exposure for proliferation or differentiation assays, we dissociated the neurospheres to single cells before exposing them to MeHg containing medium. For migration assays, we used intact neurospheres.
All the experiments were performed using doses and times of exposure that do not induce apoptosis. For proliferation analyses, single cells were plated onto poly-D-lysine (PDL) and laminin coated glass coverslips (diameter 12 mm, placed in Nunclon® Δ Multidishes, 24 wells, flat bottom) and kept in DMEM/F12/ N2 (DFN) medium (DMEM/Hams F12 3:1, supplemented with N2, 1:100,Invitrogen), supplemented with FGF and EGF. The next day, cells were exposed to 10-100 nM MeHg in FGF/EGF-supplemented DFN medium for 24 h (see also supplementary material). The exposure was performed by replacing the culture medium with FGF/EGFsupplemented DFN medium containing MeHg. In the control cell cultures, no MeHg was added to the replacement medium.
For the proliferation studies, more than 5100 cells/nuclei per treatment (in total) were counted.
For differentiation analyses, dissociated cells were plated on PDL and laminin coated glass coverslips (placed in Nunclon® Δ Multidishes, 24 wells, flat bottom) at a density of 40.000 cells per 12 mm coverslip, in MeHg-containing (10 nM MeHg) or MeHg-free DFN medium. Gene expression analysis, immunostaining, neurite length quantification, and apoptosis assays were performed after 96 h of differentiation. For the apoptosis studies, more than 3.300 cells/nuclei per treatment (in total) were counted.

Sex Determination
For sex determination, genomic DNA was harvested from about 50 neurospheres using a genomic DNA extraction kit. The DNA was eluted in water and subjected to PCR with two primer pairs against the AMELX/AMELY gene (Nakahori et al. 1991). AMELX/AMELY is a single copy gene, located on the X and Y chromosomes. X-and Y-specific products with different sizes were simultaneously detected because of difference in the lengths of corresponding introns (Fig. 1b).
Images were collected from random fields by a Nikon inverted fluorescent microscope (Nikon Eclipse Ti-S) equipped with a Nikon Digital Sight DS-Qi1MC camera. For quantitative analysis, the images were batch processed (to avoid bias) using the Volocity image analysis software (Demo-version, PerkinElmer) or ImageJ (http://imagej.net/ ImageJ).

Migration Assay
For migration assay, neurospheres were plated in PDL and laminin-coated multi-well plates (24 or 48 wells, Nunc) and left to attach overnight in DFN medium. The following morning, cultures were switched to MeHg-containing DFN medium (10 nM MeHg) or fresh MeHg-free DFN medium and placed in a Cell-IQ incubator (Chip-Man Technologies) for live imaging over the next 26 h. Phase contrast images of each neurospheres were collected every 30 min.

RNA Extraction, cDNA Synthesis, and Quantitative RT-PCR
For mRNA extraction and quantification, total RNA was isolated using the peq Gold Microspin Total RNA Kit (peqLab GmbH, Erlangen, Germany). Complementary DNA (cDNA) was synthetized from at least 1 μg RNA by using Superscript II First-Strand cDNA Synthesis Kit according to the manufacturer's protocol. Amplification reactions were set up, and product accumulation was measured by quantitative realtime (qRT) PCR analyses based on SYBR Green detection via ABI Prism 7000. Sequence Detection System with SDS software (version 2.1; Applied Biosystems, Foster City, CA). The qRT-PCR cycle conditions were 50°C for 2 min, 95°C for 10 min, 95°C for 15 s, and 60°C for 1 min (40 cycles). Expression levels were normalized to the housekeeping genes β-actin and ribosomal protein-like 13 (ΔCt = Ct (target gene) − Ct (housekeeping gene), which showed no MeHginduced changes in gene expression (data not shown). R e l a t i v e e x p r e s s i o n l e v e l s w e r e c a l c u l a t e d a s ΔΔCt = ΔCtMeHg − ΔCtcontrol, and expression changes were calculated as 2−ΔΔCt. Primers were used at a final concentration of 4 μM. Primer sequences and annealing temperatures used for qRT-PCRs were as follows: Product specificity was determined via melting curve analyses (temperature ramp from 60 to 95°C) and agarose gel electrophoresis. All experiments were done on three replicate samples from two independent cell preparations from different donors.

Statistics
All experiments were performed on cells from at least three different fetuses in at least two replicate cultures (see Table 1). For statistical analysis, ANOVA followed by Tukey's post hoc test was used for comparisons between control and cultures exposed to different concentrations of MeHg. Factorial (twoway) ANOVA was used to relate MeHg exposure and sex differences. Student's t test was used for comparisons of two groups. The significance value was set at p < 0.05. Values are shown as mean ± SEM unless otherwise stated.

Fetus Developmental Age Influences Cytotoxicity of MeHg
The developing nervous system is particularly sensitive to insults during the first trimester of pregnancy (Miodovnik 2011). Therefore, as a first aim of the present study, we wanted to evaluate neurodevelopmental effects of low doses of MeHg in human neurospheres derived from PCW 8.5 fetuses (four male and four female), representative of an early stage of neurogenesis (Stiles and Jernigan 2010) (Fig. 1a).
Neurospheres generated from 8.5-week-old fetuses are characterized by the expression of well-established radial glia associated markers, namely Sox2, Pax6, and Nestin (Lendahl et al. 1990;Götz et al. 1998;Graham et al. 2003) (Fig. 1c). Clonal neurospheres are composed by heterogeneous cellular populations including neural stem cells and neuronal and glial progenitors in different stages of differentiation (Suslov et al. 2002;Jensen and Parmar 2006). Therefore, when dissociated neurospheres are differentiated, the progenitor population will rapidly initiate neuronal and glial differentiation (Fig. 1d), while the neural stem cell population will keep a radial glialike morphology ( Fig. 1c; supplementary Movie 1).
To assess whether MeHg toxicity is influenced by the developmental stage, we used hNPC cultures also from PCW 16 fetuses and assayed the effects of four MeHg concentrations (in the range 2.5-100 nM) on the apoptosis rate. To this purpose, we evaluated chromatin condensation and quantified nuclei with apoptotic morphology (Darzynkiewicz et al. 1992) (Fig. 2a, b). In hNPC cultures from PCW 8.5-week fetuses, we found significantly increased apoptosis after exposure to 25 nM MeHg (Fig. 2c), while cultures established from 16 weeks fetuses (PCW 16) showed a comparable amount of apoptotic cells only after exposure to 100 nM MeHg (Fig. 2d), indicating that the susceptibility of hNPCs to MeHg is related to the developmental stage.
Subapoptotic Doses of MeHg Affect Neuronal Differentiation of PCW 8.5 hNPCs Next, we wanted to evaluate the effect on neuronal differentiation of 10 nM MeHg, a subtoxic concentration that does not affect the proliferation of PCW 8.5-derived hNPCs (see Suppl. Fig. 1). After 4 days of spontaneous differentiation, MeHgtreated hNPCs showed a significant reduction in the number of newly formed neurons (Tuj1-positive) (Fig. 3a, b, c). It is known that MeHg is able to activate the Notch signaling pathway by regulating ADAM metalloproteases (Bland and Rand 2006;Tamm et al. 2008); therefore, we looked at the expression of the well-known Notch signaling target HES5. Consistently, MeHg exposure induced an upregulation of HES5, as shown in Fig. 3d.
We next looked at the glial compartment and found no changes in the number of GFAP expressing cells (Fig. 3a, b, e), suggesting that MeHg stalls neuronal differentiation rather than causing a shift toward the glial fate. To assess whether the observed MeHg-induced inhibition of neuronal differentiation was sex-related, we compared the expression level of the same neuronal and glial markers in hNPCs from male and female fetuses in the presence of MeHg but we could not find any significant change (data not shown). Due to its crucial role in regulating neuronal differentiation (Numakawa et al. 2010), we quantified BDNF expression on differentiating PCW 8.5 hNPCs and found it to be significantly lower in cultures exposed to MeHg (Fig. 4a). There was no sex-related change in BDNF expression when comparing MeHg-treated male cells to female cells (Fig. 4b).

MeHg Exposure Affects Neurite Extension and Neuronal Migration in a Sex-Related Manner
We next measured the neurite length of immature neurons in cultures treated with 10 nM MeHg and found a significant difference after 4 days of spontaneous differentiation, as compared to control (Fig. 5a). Importantly, when comparing male versus female cultures, neurite length was significantly reduced in cultures of developing neurons with male karyotype (80% of control in females and 59% of control in males; Fig. 5b).
Moreover, MeHg exposure reduced cell migration over a period of 18 h to about 77% of control (Fig. 5c, d), and the male-karyotype cultures were more affected as compared to cultures from female karyotype (86% of control in females and 63% of control in males; Fig. 5e). In light of the CDKL5 (cyclin-dependent kinase-like 5) role as a critical regulator of neuronal morphogenesis (Chen et al. 2010), we looked at its expression in MeHg-exposed and control cultures of differentiating PCW 8.5 hNPCs. CDKL5 expression was significantly decreased in MeHg-treated cultures, and the expression was even lower in cultures with male karyotype, as compared to female karyotype cells (Fig. 4c, d).

Discussion
In the present study, we show that the cytotoxicity of low doses of MeHg in human neurospheres is influenced by gestational age and that subapoptotic concentrations of MeHg impair neuronal maturation in a sex-dependent manner.
We found that hNPCs from an earlier fetal period are more susceptible to low doses MeHg as compared to cells from older fetuses, suggesting that MeHg toxicity is influenced by fetal developmental age and it is tempting to speculate that the molecular mechanisms involved may be related to the antioxidative enzymes expression level.
Previous studies from our group have shown that prenatal exposure to MeHg in the nanomolar range can inhibit neuronal differentiation (Tamm et al. 2008;Bose et al. 2012). In close accordance, the present data indicate that exposure to subapoptotic concentrations of MeHg interferes with neuronal differentiation of hNPCs, as shown by a decreased number of Tuj1-positive cells. It is well established that Notch signaling plays a key role in neurogenesis as its activation induces the expression of transcriptional repressor genes, including Hes1 and Hes5, leading to the inhibition of neuronal differentiation (Imayoshi et al. 2010). In light of its role as Notch effector, the increased expression of HES5 observed in our cells exposed to MeHg suggests that the molecular mechanism underlying the decreased neuronal differentiation involves an overactivation of Notch signaling that keeps progenitor cells in their undifferentiated state. This is further supported by our data showing that the inhibition of neuronal differentiation is not associated with an increase in the number of glial cells, indicating that MeHg does not induce a shift toward the glial fate and non-neuronal cells retain their progenitor identity as a consequence of Notch signaling activation.
In the attempt to identify additional factors mediating MeHg detrimental effects on hNPC, we analyzed BDNF expression, a key factor for neuronal differentiation and survival (Numakawa et al. 2010). According to previous studies, perinatal exposure to MeHg induces an increase in DNA methylation and a concomitant decrease of H3 acetylation in the BDNF promoter region, leading to the repression of its expression (Tsankova et al. 2006;Onishchenko et al. 2008). Consistently, we found a decreased BDNF expression level in MeHg-exposed hNPC, suggesting that alterations in this pathway may be directly involved in the disruption of hNPC differentiation following MeHg treatment.
After a deeper characterization of differentiated hNPC, we found a reduced neurite extension and cell migration following exposure to MeHg. As previously shown, exposure to nM concentration of MeHg disrupts neuronal migration and inhibits axonal morphogenesis of NPCs in vivo and in vitro (Heidemann et al. 2001;Moors et al. 2007;Moors et al. 2009;Guo et al. 2013); however, the underlying mechanisms are still not clearly identified. A study by Guo et al. showed that exposure to low levels of MeHg suppresses the expression of three key proteins involved in the regulation of neuronal migration, namely Rac1, Cdc42, and RhoA (Guo et al. 2013). Another reasonable molecular mechanism may involve the BDNF-activated pathways. Indeed, it is known that after BDNF binding to TrkB receptor, several signaling pathways are activated, including the MAPK/ERK1/2, a crucial pathway promoting cell migration (Huang et al. 2004;Moors et al. 2007). Thus, the BDNF downregulation observed in our samples after MeHg exposure may contribute to the migration impairment.
Neuronal process elongation and maintenance are regulated by nerve growth factor (NGF) in a microtubule-dependent manner (Drubin et al. 1985), and it is known that MeHg interferes with the axonal outgrowth process by poisoning microtubule assembly (Miura et al. 2000;Heidemann et al. 2001). A recent study by Fujimura et al. (2016) showed that prenatal exposure to low-dose MeHg was associated with a significant downregulation of eukaryote elongation factor 1A1 (eEF1A1), a key factor regulating neurite outgrowth, through NGF/TrkA activated pathway (Fujimura et al. 2016). According to other studies (Inamura et al. 2005  of neurite extension. Indeed, it has been shown that BDNF promotes eEF1A phosphorylation and that the consequent increased eEF1A activity leads to an enhanced protein synthesis, resulting in the promotion of neurite extension in cortical neurons (Inamura et al. 2005). Therefore, it is likely that the BDNF downregulation observed in MeHg-exposed cells results in a further decrease of EF1A activity, which may crucially contribute to the detrimental effects on neurites extension. However, more experiment is needed to further support this hyphotesis. It is worth noting that in our study, MeHg induced a reduction in the expression level of CDKL5, a critical gene regulating neuronal morphogenesis which is mutated in the Hanefeld variant of Rett syndrome (Chen et al. 2010). CDKL5 has been shown to exert different effects within neuronal cells depending on its subcellular location, i.e., cytoplasmic or nuclear (Rusconi et al. 2008). In the cytoplasm, CDKL5 regulates neuronal morphogenesis and dendritic arborization by a mechanism involving BDNF-Rac1 signaling (Chen et al. 2010). Indeed, it has been suggested that BDNF activates CDKL5 that, in turn, triggers Rac1 activity to regulate neuronal morphogenesis through the actin cytoskeleton remodeling (Chen et al. 2010). Thus, it is likely that the downregulation of CDKL5 observed in our MeHg-exposed cells prevents the BDNF-mediated activation of Rac1, leading to the defective axonal morphogenesis.  An interesting phenomenon observed in this study is the sex dependence of the MeHg-induced defects in neurite extension and cell migration, which were more pronounced in cultures established from male fetuses. This is in agreement with epidemiological and experimental studies showing that males are more susceptible to MeHg neurotoxicity as compared to females (McKeown-Eyssen et al. 1983;Rossi et al. 1997;Grandjean et al. 1998;Giménez-Llort et al. 2001;Björklund et al. 2007). As reported in previous studies, there are sex-related differences in the antioxidant defense system activity and in the peroxide production (Carrillo et al. 1992;Borrás et al. 2003). Indeed, it has been shown that mitochondria from female rats produce less peroxide than those from male rats of the same age (Borrás et al. 2003). Moreover, both gene expression level and enzymatic activity of Mnsuperoxide dismutase and glutathione peroxidase were found to be significantly higher in female as compared to male rats (Borrás et al. 2003). In this regard, it is important to underline that a major event mediating MeHginduced neurotoxicity is represented by its interaction with thiols from GSH (Sumi 2008;Farina et al. 2010). The subsequent decreased capacity of the entire antioxidant GSH system (Shanker et al. 2005;Stringari et al. 2008) may likely lead to the increased susceptibility to MeHg observed in cells from male fetuses.
In conclusion, the present study shows that the gestational age is a critical factor influencing hNPC sensitivity to low levels of MeHg. Subcytotoxic doses of MeHg impair neuronal differentiation and maturation of hNPC in a sex-dependent manner as shown by the more pronounced inhibition of neurites outgrowth and cell migration in MeHg-exposed cells from male fetuses. Our data point to Notch, CDNK5, and BDNF as critical players in the cascade of intracellular events leading to MeHg-induced in vitro neurotoxicity.