Allopregnanolone Prevents Dieldrin-Induced NMDA Receptor Internalization and Neurotoxicity By

NEUROENDOCRINOLOGY

Allopregnanolone Prevents Dieldrin-Induced NMDA Receptor Internalization and Neurotoxicity by

Preserving GABAA Receptor Function

Víctor Briz, Jyoti Parkash, Sara Sa´ nchez-Redondo, Vincent Prevot, and Cristina Sun˜ol

Department of Neurochemistry and Neuropharmacology (V.B., S.S.-R., C.S.), Instituto de Investigaciones Biome´ dicas de Barcelona, Consejo Superior de Investigaciones Científicas-Institut d’Investigacions Biomèdiques August Pi i Sunyer (IIBB-CSIC-IDIBAPS), Centro de Investigación Biomédica en Red Epidemiology and Public Health (CIBERESP), E-08036, Barcelona, Spain; and Institut National de la Sante´ et de la Recherche Me´ dicale (J.P., V.P.), Jean-Pierre Aubert Research Centre, Development and Plasticity of the Postnatal Brain, Unite´ 837, F-59045, and Université Droit e Santé Lille (J.P., V.P.), Universite´ Lille Nord de France, School of Medicine, F59000, Lille, France

Dieldrin is an endocrine disruptor that accumulates in mammalian adipose tissue and brain. It ␥ induces convulsions due to its antagonism of the -aminobutyric acid A receptor (GABAAR). We have previously reported that long-term exposure to dieldrin causes the internalization of the

N-methyl-D-aspartate receptor (NMDAR) as a result of persistent GABAAR inhibition. Because the neurosteroids 17␤-estradiol (E2) and allopregnanolone are known to modulate the function and

trafficking of GABAAR and NMDAR, we examined the effects of E2 and allopregnanolone on

dieldrin-induced GABAAR inhibition, NMDAR internalization, and neuronal death in cortical neurons. We found that 1 nM E2 increased the membrane expression of NR1/NR2B receptors and postsynaptic density 95 but did not induce their physical association. In contrast, 10 nM E2 had no effect on these proteins but reduced NR2A membrane expression. We also found that exposure to 60 nM dieldrin for 6 d in vitro caused the internalization of NR1 and NR2B but not NR2A. Treatment with either 1 nM E2 or 10 ␮M allopregnanolone prevented the dieldrin-induced reduction in membrane levels of the NR1/NR2B receptors. Furthermore, prolonged exposure to 200 nM dieldrin down-regulated the expression of NR2A; this was inhibited only by allopregnanolone. Although both hormones restored NMDAR function, as measured by the NMDA-induced rise in intracellular

calcium, allopregnanolone (but not E2) reversed the inhibition of GABAAR and neuronal death caused by prolonged exposure to dieldrin. Our results indicate that allopregnanolone protects cortical neurons against the neurotoxicity caused by long-term exposure to dieldrin by maintaining

GABAAR and NMDAR functionality. (Endocrinology 153: 847–860, 2012)

he persistent organochlorine pesticide dieldrin is an ease (4, 5) and disrupts neuronal development by altering Tendocrine disruptor due to its interaction with estro- the expression of GABAAR subunits (6). There is increas- gen receptors (ER) (1), including those found in neurons ing evidence that long-term changes in glutamatergic and (2). It produces convulsions after acute intoxication be- GABAergic neurotransmission eventually lead to regula- cause it blocks the ␥-aminobutyric acid A receptor tion of N-methyl-D-aspartate receptor (NMDAR) expres-

(GABAAR) (3). Dieldrin is also considered an environ- sion and trafficking, which restores the homeostatic bal- mental risk factor in the development of Parkinson’s dis- ance between excitatory and inhibitory synaptic activity

2ϩ ISSN Print 0013-7227 ISSN Online 1945-7170 Abbreviations: [Ca ]i, Intracellular calcium concentration; DIV, days in vitro; DMSO, di- Printed in U.S.A. methylsulfoxide; E2, 17␤-estradiol; ER, estrogen receptor; GABA, ␥-amino-n-butyric acid; ␥ Copyright © 2012 by The Endocrine Society GABAAR, -aminobutyric acid A receptor; HBSS, Hank’s balanced salt solution; MTT, doi: 10.1210/en.2011-1333 Received June 2, 2011. Accepted November 16, 2011. 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl-tetrazolium bromide; NMDA, N-methyl-D-aspar- First Published Online December 13, 2011 tate; NMDAR, N-methyl-D-aspartate receptor; PSD95, postsynaptic density 95; SAP102, synapse-associated protein 102.

Endocrinology, February 2012, 153(2):847–860 endo.endojournals.org 847

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 11 February 2014. at 02:57 For personal use only. No other uses without permission. . All rights reserved. 848 Briz et al. Allopregnanolone Prevents Dieldrin Toxicity Endocrinology, February 2012, 153(2):847–860

(7–9). We have previously reported that long-term expo- Materials and Methods sure to dieldrin induces the internalization of the NMDAR in cerebellar granule cells and in cortical neurons as a Materials consequence of permanent blockade of the GABA R (10, Pregnant NMRI mice (16th gestational day) were obtained A from Charles River, Iffa Credo (Saint Germain-sur-l’Arbreste, 11). These alterations may explain the cognitive and be- France). Dimethylsulfoxide (DMSO), dieldrin, E2, ␥-amino-n- havioral deficits observed in animals that are chronically butyric acid (GABA), L-glycine, N-methyl-D-aspartate (NMDA), exposed to dieldrin (12, 13). However, the precise mech- and 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl-tetrazolium bro- anisms underlying such impairments remain elusive. mide (MTT) were from Sigma Chemical Co. (St. Louis, MO). On the other hand, the neurosteroids 17␤-estradiol 3␣-Hydroxy-5␣-pregnan-20-one or allopregnanolone was from (E2) and allopregnanolone are important for the develop- the Institute of Organic Chemistry and Biochemistry (Prague, ment, maturation, and function of the central nervous sys- Czech Republic) (21). Fluo-3 acetoxymethyl ester was from Mo- lecular Probes (Invitrogen, Carlsbad, CA). [3H]Flunitrazepam tem. These hormones are involved in brain processes such was from PerkinElmer (Boston, MA). as reproductive behavior, memory, and neuronal survival. Moreover, these effects have been often related to their Neuronal cultures modulation of GABA and glutamate neurotransmission Primary cultures of cortical neurons were prepared from the (14–17). For instance, E2 regulates GABA synthesis (18) cerebral cortices of embryonic d 16 NMRI mouse fetuses as and induces synaptogenesis and synaptic plasticity by in- described previously (11). Animals were handled in compliance creasing NMDAR membrane expression (19, 20). Allo- with protocol DMA1852 of the University of Barcelona, ap- pregnanolone is a metabolite of progesterone that acts as proved by the Generalitat de Catalun˜ a, Spain, following Euro- pean Union guidelines. a positive allosteric modulator of the GABAAR (21). Fur- thermore, both hormones are neuroprotective against dif- Chemical treatments ferent models of neurodegeneration, both in vivo and in Stock solutions of each compound were prepared in DMSO vitro (14, 22–28). and frozen in aliquots of 100 ␮l. The final concentration of In the present work, we examined the ability of E2 and DMSO in the culture medium was less than 0.5%. Cultured allopregnanolone to prevent the reduction in GABA and neurons were treated with dieldrin at 2, 4, or6dinvitro (DIV) glutamate neurotransmission observed after long-term ex- by adding the stock solution to the culture medium. The medium posure to dieldrin. Primary cultures of cortical neurons remained unchanged until the experiments were performed at 8 DIV (exposure for 6, 4, and 2 DIV, respectively). In some ex- (mainly composed of GABAergic and glutamatergic neu- periments, neurons were treated with dieldrin at 1–2 DIV, and rons) were used to evaluate the effects of the neurosteroids the experiments were performed at 4–5 DIV. Long-term or pro- on GABAAR and NMDAR functionality after dieldrin ex- longed exposure to dieldrin is generally referred as exposure for posure. The NMDAR is a heterotetramer composed of 6 DIV, unless otherwise stated. The concentrations of dieldrin two NR1 and two NR2 subunits. Although the NR1 sub- used for long-term experiments (60 and 200 nM, 6 DIV) were unit is necessary for the cell membrane expression and devoid of cytotoxicity (11). With the exception of the cell viability assay, experiments with mature neurons in which a higher functionality of the NMDAR, the NR2 (NR2A–D) sub- concentration of dieldrin (10 ␮M) was used, were performed 2 unit composition determines the calcium-permeable chan- DIV before neuronal damage was detected. Treatments with the nel gating kinetics and the pharmacological properties of neurosteroids E2 and allopregnanolone were normally applied the receptor (29). Different roles in synaptic plasticity, from 6–8 DIV, unless otherwise stated. memory, neuronal survival, and excitotoxicity have also been attributed to NR2A- and NR2B-containing Membrane extraction and coimmunoprecipitation NMDAR (30–32). In addition, the interaction of NR2A Membrane extraction was performed as previously described and NR2B with the scaffold proteins postsynaptic density (34), with few modifications. In brief, cells grown on six-well plates were washed three times with ice-cold PBS (137 mM NaCl, 95 (PSD95) and synapse-associated protein 102 (SAP102) 2.7 mM KCl, 10 mM Na2HPO4,and2mM KH2PO4). Protein determines the synaptic localization and trafficking of the extracts from three wells were used for each sample. The mem- NMDAR (33). Thus, we evaluated the membrane expres- brane samples were mixed with the same volume of 2ϫ sample sion of the NMDAR subunits NR1, NR2A, and NR2B buffer (Invitrogen) and boiled for 5 min. Supernatants from ul- and the scaffold proteins PSD95 and SAP102 after E2, tracentrifugation (cytosolic fraction) were diluted in 4ϫ sample allopregnanolone, and/or dieldrin treatment. Finally, we buffer (Invitrogen) and boiled for 5 min. tested the neuroprotective action of these hormones Coimmunoprecipitation was performed as described previously (34–36), with few modifications. Briefly, similar amounts against the neuronal death caused by dieldrin. We found of protein (ϳ0.25 mg) from membrane extracts were prepared in that allopregnanolone, but not E2, prevented dieldrin- a total volume of 750 ␮l lysis buffer X-1 and then incubated induced neurotoxicity by maintaining GABAAR and overnight with gentle rocking at 4 C with 2 ␮g of both anti- NMDAR functionality. PSD95 antibodies [MA1-045 (lot no. 172-121) and MA1-046

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 11 February 2014. at 02:57 For personal use only. No other uses without permission. . All rights reserved. Endocrinology, February 2012, 153(2):847–860 endo.endojournals.org 849

(lot no. 173-138); Affinity BioReagents, Golden, CO] or 2 ␮g were washed three times in a modified Hank’s balanced salt anti-NR2B (71-8600, lot no. 1407642; Invitrogen). solution (HBSS) buffer (136 mM NaCl, 5.4 mM KCl, 1.2 mM

CaCl2, 1.4 mM MgCl2, 1.0 mM NaH2P04,10mM HEPES, and 9 Western blot mM glucose, adjusted to pH 7.3) and then exposed to 1.3–2.0 nM 3 ␮ Before loading, the samples were boiled, and 50 ␮g of protein [ H]flunitrazepam in HBSS in the absence or presence of 30 M from membrane fraction samples or 25 ␮g from cytosolic frac- GABA for 30 min at 25 C. tion samples were subjected to electrophoresis as previously described (34–36). The antibodies used for immunodetection [3H]MK801 binding were mouse monoclonal anti-NMDAR1 (05-432, lot no. NMDAR binding assay was performed as previously de- DAM1557769, 1:3000; Millipore, Billerica, MA), anti-PSD95 scribed (10), with modifications. Cultures grown in 24-well (MA1-046, 1:500), and anti-␣-spectrin (MAB1622, lot no. plates for at least 8 DIV were washed with PBS and then exposed 3 24050176, 1:1000; Millipore) and rabbit polyclonal anti-NR2A for 15 min at 37 C to 4–5 nM [ H]MK801 in PBS solution con- (AB1555P, lot no. JC1650245, 1:500; Millipore), anti-NR2B taining 100 ␮M glutamate and 100 ␮M glycine. After washing (1:500), and anti-SAP102 (3733S, lot no. 1, 1:1000; Cell Sig- with cold PBS, cells were collected in 0.2 M NaOH, and then their naling, Beverly, MA). A goat polyclonal anti-actin (sc-1616, lot radioactivity and protein content were measured. Nonspecific no. H1010, 1:1000; Santa Cruz Biotechnology, Santa Cruz, CA) binding was determined in the presence of 100 ␮M nonlabeled was always used as a control of the amount of protein loaded. MK801. Secondary antibodies used for Western blot detection were an- timouse (1:8000), antirabbit (1:10,000), and antigoat/sheep (1: Cell viability 10,000), all horseradish peroxidase conjugated (Sigma). Cell viability was assessed by measuring the reduction of the colored formazan salt MTT by mitochondrial activity as previ- Reverse transcription-polymerase chain reaction ously described (11). Total cell RNA was purified following the method of Chom- czynski and Sacchi (37). RNA integrity and concentration were Data analysis assessed by spectral analysis using a NanoDrop Spectrophotom- Unless otherwise stated, data are presented as mean Ϯ SE of at eter (Thermo Scientific, Wilmington, DE). Equal amounts of least three experiments from independent culture batches, each RNA (ϳ1 ␮g) were retro-transcribed with Omniscript reverse one performed in triplicate. Statistical comparisons were made transcriptase (QIAGEN, Valencia, CA) following the manufac- by one-way ANOVA followed by Dunnett’s post hoc compari- turer’s guidelines. To ensure that cDNA samples were not con- taminated with genomic DNA, all samples were also incubated son test, and by two-way ANOVA followed by Bonferroni post with a reverse transcription reaction mix lacking Omniscript hoc test when comparing two factors. The level of significance Ͻ reverse transcriptase. Amplification of each transcript was per- was set at P 0.05. Graphics were created and statistical analysis formed in 20 ␮l of a PCR cocktail containing cDNA derived from performed using the software package GraphPad Prism version 50 ng mRNA, 8 pmol of each primer, and 0.6 U KAPA Taq DNA 4.0, 2003 (GraphPad Software Inc., San Diego, CA). polymerase (Kapa Biosystems, Woburn, MA). After incubation at 95 C for 5 min, 32 cycles of amplification (15 sec at 95 C, 30 sec at 62 C, and 40 sec at 72 C) or 30 cycles of amplification (15 Results sec at 95 C, 30 sec at 56 C, and 15 sec at 72 C) were performed for NR2A and GAPDH, respectively. The number of amplifi- Allopregnanolone and E2 prevented cation cycles for each primer pair and cDNA preparation was set dieldrin-induced reduction of NR1 and NR2B to ensure that no saturation occurred. Sense and antisense PCR primers were as follows: 5Ј-CCCACCTACTCAGGCCACTT-3Ј membrane expression and NMDAR functionality and 5Ј-CCGACTGTCCCTGGAGCAAT-3Ј (nucleotides 3664– We have previously reported that exposure to dieldrin 3683 and 4251–4232 in databank accession NM_008170.2) for for 6 DIV reduces the number of NMDAR in the cell mem- NR2A and 5Ј-AACGACCCCTTCATTGAC-3Ј and 5Ј-TCCA brane as a result of persistent GABA R inhibition (10, 11). Ј A CGACATACTCAGCAC-3 (nucleotides 144–161 and 334– Because NR2A and NR2B are the most expressed NR2 316 in databank accession NM_008084.2) for GAPDH. subunits in mature cortical neurons (33, 29, 39), we aimed Determination of intracellular calcium to study the effects of prolonged exposure to dieldrin concentration and/or to the neurosteroids allopregnanolone and E2 on 2ϩ NR1, NR2A, and NR2B membrane protein levels. Expo- The intracellular calcium concentration ([Ca ]i) was determined by measuring Fluo-3 acetoxymethyl ester (Molecular sure to 60 nM dieldrin for 6 DIV reduced the membrane Probes) fluorescence, as previously described (11). This fluores- expression of the NR1 (Fig. 1A) and NR2B (Fig. 2A) sub- cent probe is specific for the evaluation of glutamate receptor units of the NMDAR. A simultaneous increase of NR1 function (10, 11). and NR2B levels was found in the cytosolic fraction (Figs. 1B and 2B). No differences were observed in the total [3H]Flunitrazepam binding The binding assay was performed on cells grown in 24-well protein levels of NR2B from whole-cell homogenates plates as previously described (21), with a few modifications. (Supplemental Fig. 1, published on The Endocrine Soci- After treatment with allopregnanolone and/or dieldrin, cultures ety’s Journals Online web site at http://endo.endojournals.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 11 February 2014. at 02:57 For personal use only. No other uses without permission. . All rights reserved. 850 Briz et al. Allopregnanolone Prevents Dieldrin Toxicity Endocrinology, February 2012, 153(2):847–860

of dieldrin on NMDAR are time and concentration dependent. Because changes in the phosphorylation state of NR2B have been associated with alterations in its cellular location (40, 41), we wondered whether tyrosine phosphorylation of NR2B was modified by dieldrin before this NMDAR subunit was internalized. Exposure to 60 nM dieldrin for both 2 and 4 DIV significantly reduced NR2B phosphorylation at Tyr1472 with respect to control cells (Fig. 2C), suggesting that this posttranslational modification precedes the internalization of NR2B-containing NMDAR. To confirm that the observed changes in NMDAR localization have functional relevance, we evaluated NMDAR function by measuring NMDA-induced in- 2ϩ crease in [Ca ]i. We have previously re- FIG. 1. Effects of neurosteroids on dieldrin-induced internalization of NR1 subunit of ported that exposure to 60 nM dieldrin 2ϩ NMDAR. Cortical neurons were exposed to DMSO [control (C)] or 60 nM dieldrin (D60) for 6 for 6 DIV reduced the increase in [Ca ]i DIV and to 10 ␮M allopregnanolone (AP), 1 nM E2 (E1) or 10 nM E2 (E10) for 2 DIV. The induced by NMDA (11). Figure 3 shows protein levels of NR1 and actin were analyzed on membrane (panel A) and cytosolic (panel B) that this reduction was prevented by fractions from the cultures by Western blot. Representative immunoblots are shown on the top (panel A) or on the right (panel B), and densitometric quantification of the immunoblots is treatment with either E2 (at 1 nM but shown on the bottom (panel A) or on the left (panel B) of each panel. Data are mean Ϯ SE of not at 10 nM)or10␮M allopregnano- three to seven independent experiments. Statistical comparisons were made by two-way lone (Fig. 3). In addition, 1 nM E2 en- ANOVA: **, P Ͻ 0.01; ***, P Ͻ 0.001 vs. control; #, P Ͻ 0.05; ###, P Ͻ 0.001 vs. dieldrin- treated cells; $$, P Ͻ 0.01 vs. E1-treated cells. hanced the NMDA-induced increase in 2ϩ [Ca ]i with respect to control cells, un- org), as we previously reported for NR1 (11). These ob- like 10 nM E2 (Fig. 3A) and allopregnanolone (Fig. 3B). servations confirm that dieldrin induces changes in the Consistent with this, [3H]MK801 binding to cell surface localization but not in the overall expression of these NMDA receptors tested under the above conditions re- NMDAR subunits, indicating that these subunits are in- produced the effects of dieldrin and the neurosteroids on ternalized after pesticide treatment. Treatment with 1 nM NMDAR function and localization (Supplemental Fig. 3). E2 but not with 10 nM E2 increased the membrane expression of both NR1 and NR2B. Moreover, 1 nM E2 Allopregnanolone but not E2 inhibited the prevented the reduction in membrane expression of both down-regulation of NR2A caused by dieldrin subunits induced by dieldrin (Figs. 1A and 2A). However, Exposure to dieldrin for 6 DIV diminished NR2A mem- E2 treatment did not prevent the increase in cytosolic NR1 brane expression in cortical neurons in a concentration- (Fig. 1B) and NR2B (data not shown) levels induced by dependent manner, achieving statistical significance at dieldrin. In contrast, treatment with 10 ␮M allopregnano- 200 nM (Fig. 4A). Addition of allopregnanolone but not E2 lone had no effect on the membrane expression of any of restored the basal levels of NR2A at the cell membrane the NMDAR subunits studied, but it inhibited the dieldrin-induced reduction in membrane NR1 and NR2B lev- after dieldrin exposure (Fig. 4B). Unlike NR1 and NR2B, els (Figs. 1A and 2A). Furthermore, allopregnanolone pre- the membrane expression of NR2A was significantly re- vented the increase in the cytosolic levels of NR1 (Fig. 1B) duced after treatment with 10 nM (but not 1 nM) E2 (Fig. and NR2B (Fig. 2B). Exposure to 60 nM dieldrin for 2 DIV 4A). However, no concomitant increase of NR2A in the did not affect the membrane expression of any of NMDAR cytosolic fraction was observed after dieldrin or E2 expo- subunits studied, and a higher pesticide concentration sure (Fig. 4C). This prompted us to examine the effect of (200 nM) was needed to achieve a significant effect (Sup- dieldrin on NR2A mRNA levels. Analysis by RT-PCR plemental Fig. 2). These results indicate that the effects showed that long-term exposure to 200 nM dieldrin caused

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 11 February 2014. at 02:57 For personal use only. No other uses without permission. . All rights reserved. Endocrinology, February 2012, 153(2):847–860 endo.endojournals.org 851

nM) E2 strongly increased the membrane expression of PSD95 (Fig. 5A). We then studied the association of PSD95 with the NMDAR subunits to determine whether the effects of E2 on PSD95 were accompanied by increased interaction with the NMDAR subunits. We also tested the possibility that dieldrin could disrupt the interaction between these proteins. Thus, we immunoprecipitated membrane extracts from our cultures with an antibody raised against PSD95 and subsequently analyzed the presence of NR2A and NR2B in the immunoprecipitates by immunoblotting. PSD95 was highly coupled to NR2A-containing (Fig. 5C) but not to NR2B-containing (Supple- mental Fig. 4A) NMDAR in our cultures. Moreover, long-term exposure to 200 nM dieldrin totally disrupted the physical association between PSD95 and NR2A (Fig. 5C). These observations confirm that the effects of dieldrin are specific to NMDAR and are not a consequence of the general reduction of the PSD, i.e. neurodegeneration. How- ever, although 1 nM E2 promoted the membrane expression of PSD95 (Fig. FIG. 2. Effects of neurosteroids on dieldrin-induced internalization of NR2B subunit of NMDAR. Cortical neurons were exposed to DMSO [control (C)] or 60 nM dieldrin (D60) for 6 5A), this was not accompanied by in- DIV and to 10 ␮M allopregnanolone or 1 nM E2 (E1) for 2 DIV (and combinations), and creased interaction with NR2A (Fig. subsequently, the protein levels of NR2B and actin were analyzed on membrane (panel A) and 5C). Likewise, E2 did not induce any cytosolic (panel B) fractions from the cultures by Western blot. Panel C, Cultures were physical association between PSD95 exposed to DMSO or 60 nM dieldrin for 2–4 DIV, and the protein levels of phospho-NR2B (Tyr1472) and actin were analyzed on whole-cell extracts by Western blot. Representative and NR2B (Supplemental Fig. 4A). immunoblots are shown on the top and densitometric quantification of the immunoblots is Consistent with this, immunoblot anal- Ϯ shown on the bottom of each panel. Data are mean SE of three to six independent ysis of the supernatants after immuno- experiments. Statistical comparisons were made by two-way ANOVA: *, P Ͻ 0.05; **, P Ͻ 0.01; ***, P Ͻ 0.001 vs. control; #, P Ͻ 0.05; ##, P Ͻ 0.01; ###, P Ͻ 0.001 vs. dieldrin- precipitation with PSD95 antibody treated cells. showed increased expression of the obligatory NR1 subunit in E2-treated the down-regulation of NR2A, this effect being reversed cultures, confirming that E2 stimulated by allopregnanolone (Fig. 4D). cell membrane insertion of NMDAR not associated with PSD95 (Supplemental Fig. 4B). Because SAP102 is the E2 promoted PSD95 expression but not its scaffold protein that usually holds NR1/NR2B receptors association with NMDAR at the cell membrane in the adult cortex (33), we tested Because PSD95 and SAP102 are the two main scaffold whether E2 could induce their association in our cultures. proteins that hold the NMDAR on the cell membrane (33), Thus, we performed immunoprecipitation of the NR2B we wondered whether the effects of dieldrin and E2 on subunit of NMDAR in membrane extracts followed by NMDAR subunits could be related to changes in the immunoblotting for SAP102. Figure 5D shows the phys- expression of these proteins. Figure 5, A and B, shows ical association between NR2B and SAP102 in control that prolonged exposure to dieldrin (60–200 nM) did neurons. E2 greatly enhanced the interaction between not alter the membrane protein levels of PSD95 or these proteins (Fig. 5D, top), without altering the overall SAP102. In contrast, treatment with 1 nM (but not 10 expression of SAP102 on the cell membrane (Fig. 5D, bot-

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 11 February 2014. at 02:57 For personal use only. No other uses without permission. . All rights reserved. 852 Briz et al. Allopregnanolone Prevents Dieldrin Toxicity Endocrinology, February 2012, 153(2):847–860

To determine whether the neuroprotective effect of allopregnanolone against dieldrin was due to its positive

modulation of the GABAAR, we evaluated GABAAR function by measuring the ability of GABA to increase [3H]flunitrazepam binding (21). Cultured mature neurons were exposed to 10 ␮M dieldrin for 2 DIV, a condition that did not induce cell death (data not shown). Dieldrin did not modify basal [3H]flunitrazepam binding; however, it com- pletely abolished the increase in [3H]flunitrazepam binding induced by 30 ␮M GABA. Treatment with 10 ␮M allopregnanolone significantly reversed the effect of dieldrin

on the GABAAR. Allopregnanolone alone slightly reduced the effect of GABA on [3H]flunitrazepam binding (Fig. 7A). In contrast, 1 nM E2 did not affect the inhibition of

GABAAR function caused by dieldrin (the GABA-induced increase in [3H]flunitrazepam binding amounted to 254 Ϯ 3, 130 Ϯ 7, and 132 Ϯ 3% in control, dieldrin-treated and dieldrin- plus E2-treated cultures, respectively, n ϭ 3). These results suggest that the protective effects of allopregnanolone against dieldrin neurotoxicity are most

likely mediated through the modulation of the GABAAR.

FIG. 3. Effects of the neurosteroids E2 (panel A) and allopregnanolone Allopregnanolone prevented dieldrin-induced (panel B) on dieldrin-induced reduction of NMDAR functionality. Fluo-3 NR2B cleavage and loss of association with fluorescence was measured immediately after treatment with 100 ␮M NMDA in neurons exposed to DMSO or dieldrin for 6 DIV and to 10 SAP102 in mature neurons ␮M allopregnanolone (AP), 1 nM E2 (E1), or 10 nM E2 (E10) for 2 DIV Finally, we tested whether the effects of dieldrin on (and combinations). [Ca2ϩ] (calculated from fluorescence values) is i NMDAR subunits were also observed in mature neurons expressed as percentage with respect to cells incubated with HBSS without NMDA and are mean Ϯ SE of three to five independent before the onset of the neuronal damage produced by this experiments, each performed in triplicate. Statistical comparisons were pollutant. Thus, we exposed the cells to 10 ␮M dieldrin for 2 made by two-way ANOVA: *, P Ͻ 0.05; **, P Ͻ 0.01 vs. DMSO- DIV both in the absence and presence of allopregnanolone. treated cells; ##, P Ͻ 0.01 vs. dieldrin-treated cells; $$, P Ͻ 0.01 vs. E1-treated cells. As previously observed at lower concentrations, dieldrin reduced both NR2A and NR2B membrane protein levels (Fig. tom). Taken together, the results indicate that E2 stimu- 7B) and NR2A mRNA expression (data not shown), these lates the cell membrane insertion of NMDAR not associ- effects being prevented by allopregnanolone (Fig. 7B). How- ated with PSD95, but with SAP102. ever, on this occasion, there was no increase in cytosolic NR2B expression (data not shown). Conversely, a break- ϳ Allopregnanolone reversed GABAAR inhibition and down product of NR2B ( 120 kDa) appeared in cultures neuronal death induced by dieldrin exposed to dieldrin. Again, this effect was reversed by treat- We also tested the possible neuroprotective effects of ment with allopregnanolone. In contrast, no truncated allopregnanolone and E2 against dieldrin-induced neu- NR2A fragments were observed after dieldrin exposure (Fig. ronal death under different exposure paradigms, based 7B). Cleavage of NMDAR subunits is commonly triggered on preliminary data from our laboratory. Cultured cor- by calpains in neurons and heterologous cells (42, 43). Both tical neurons were exposed to 3–10 ␮M dieldrin for 4 calpains and caspases degrade ␣-spectrin, a protein that links DIV beginning on DIV 2 (immature neurons; Fig. 6A) or NMDAR to actin cytoskeleton (44). Therefore, we measured DIV 5 (mature neurons; Fig. 6B). Treatment with 1 nM calpain activity by determining the proteolysis of ␣-spectrin E2 was unable to prevent the neuronal damage caused under the above conditions. Dieldrin exposure induced the by the pesticide, whereas 10 ␮M allopregnanolone sig- calpain- but not caspase-dependent cleavage of ␣-spectrin nificantly attenuated dieldrin neurotoxicity, regardless (fragments of 150 and 120 kDa, respectively), this effect be- of the maturity of the neurons in culture. In contrast, ing prevented by allopregnanolone (Fig. 7C). lower concentrations of allopregnanolone did not res- We next wondered whether the interaction between cue cells from dieldrin-induced cell death, not even in NR2B and SAP102 was affected by dieldrin-induced NR2B the presence of GABA (Supplemental Fig. 5). proteolysis. Exposure to 10 ␮M dieldrin for 2 DIV reduced

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 11 February 2014. at 02:57 For personal use only. No other uses without permission. . All rights reserved. Endocrinology, February 2012, 153(2):847–860 endo.endojournals.org 853

FIG. 4. Effects of neurosteroids on dieldrin-induced down-regulation of NR2A subunit of NMDAR. Cortical neurons were exposed to DMSO [control (C)], 60 nM dieldrin (D60), or 200 nM dieldrin (D200) for 6 DIV and to 10 ␮M allopregnanolone, 1 nM E2 (E1), or 10 nM E2 (E10) for 2 DIV (and combinations), and subsequently, the protein levels of NR2A and actin were analyzed on membrane (panels A and B) and cytosolic (panel C) fractions from the cultures by Western blot. Representative immunoblots are shown on the top and densitometric quantification of the immunoblots is shown on the bottom of each panel (A and B). Data are mean Ϯ SE of three to five independent experiments. Statistical comparisons were made by one-way ANOVA (panel A) or two-way ANOVA (panel B): *, P Ͻ 0.05; **, P Ͻ 0.01 vs. control; #, P Ͻ 0.05 vs. dieldrin-treated cells. Panel D, Cultures were exposed to DMSO [control (C)] or 200 nM dieldrin (D200) for 6 DIV and to 10 ␮M allopregnanolone for 2 DIV, and subsequently, the mRNA levels of NR2A and GAPDH were analyzed by RT-PCR. the interaction of SAP102 with NR2B compared with con- roprotective against the toxicity caused by dieldrin through trol, this effect being prevented by allopregnanolone treat- its positive modulation of the GABAAR. ment (Fig. 7D, top). In contrast, the membrane expression of SAP102 before immunoprecipitation remained unaffected in Homeostatic scaling down of NMDAR induced by all conditions tested (Fig. 7D, bottom). long-term exposure to dieldrin The concentrations of dieldrin used were similar to those found in the human brain (4), illustrating the toxi- Discussion cological relevance of the data presented. We have previously reported that prolonged exposure to low concen- In the present study, we show that the neurosteroids allo- trations of this pollutant produces partial but permanent pregnanolone and E2 prevented the reduction in NMDAR blockade of the GABA R (11). Disinhibition of synaptic functionality produced by long-term exposure to the endo- A activity induced by GABA R antagonists initially causes crine disruptor dieldrin in primary cultures of cortical neu- A rons. Moreover, the results provide insights into the mech- hyperexcitability, which, if it persists, produces compen- anisms of neurotoxicity of dieldrin because this pesticide satory down-regulation of the postsynaptic response to regulated NMDAR subunits in different ways, leading to the glutamate over a 48-h period (45). Furthermore, the ho- internalization of NR1/NR2B receptors and to down-regu- meostatic balance between GABA and glutamate neu- lation of NR2A. We also found that allopregnanolone re- rotransmission, known as synaptic scaling, has been verses the effects of dieldrin on all the NMDAR subunits shown to involve NMDAR trafficking (7). Here, we found studied, whereas E2 had a compensatory effect only on NR1/ that long-term exposure to dieldrin reduced the membrane NR2B receptors and did not prevent their internalization expression of all the NMDAR subunits studied. These re- (Fig. 8). In addition, allopregnanolone but not E2 was neu- sults confirm our previous observations in which NR1

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 11 February 2014. at 02:57 For personal use only. No other uses without permission. . All rights reserved. 854 Briz et al. Allopregnanolone Prevents Dieldrin Toxicity Endocrinology, February 2012, 153(2):847–860

ized (47) and cleaved by calpains (42). In agreement with previous studies (33), we found that PSD95 specifically interacts with NR2A- but not with NR2B-containing NMDAR in cortical neurons. In contrast, NR2B was phys- ically associated with SAP102, which lacks palmitoylation motifs and so is unable to cluster and to prevent the calpain-dependent proteolysis of NR2 subunits (42). The subunit-specific association with these scaffold proteins most likely explains why NR2B but not NR2A was internalized together with NR1 after dieldrin exposure. Consis- tent with this, NR2B undergoes more robust constitutive internalization (es- pecially when located extrasynapti- cally) than NR2A (47). Our results suggest that the reduced tyrosine phosphorylation of NR2B in dieldrin- treated cells drives its translocation to extrasynaptic sites, as has been de- FIG. 5. E2 increases PSD95 membrane expression and the association of NR2B with SAP102. scribed to occur during epileptogen- Cortical neurons were exposed to DMSO [control (C)], 60 nM dieldrin (D60), or 200 nM dieldrin (D200) for 6 DIV and to 10 ␮M allopregnanolone, 1 nM E2 (E1), or 10 nM E2 (E10) for esis (40). As a result of this, NR2B- 2 DIV, and subsequently, the protein levels of PSD95 (panel A) and SAP102 (panel B) were containing NMDAR would be more analyzed on membrane fractions from the cultures by Western blot. Representative susceptible of being internalized by immunoblots are shown on the top and densitometric quantification of the immunoblots is shown on the bottom of each panel (A and B). Data are mean Ϯ SE of two to six independent clathrins (41, 47). experiments. Statistical comparisons were made by one-way ANOVA: ***, P Ͻ 0.001 vs. The respective interactions of NR2 Ͻ control; ##, P 0.01 vs. E1-treated cells. Panels C and D, After the treatments, membrane subunits with SAP102 and PSD95 may extracts from the cultures were immunoprecipitated with an antibody raised against PSD95 (panel C) or NR2B (panel D, top), and then the proteins levels of NR2A and PSD95 or NR2B also explain why NR2B but not NR2A and SAP102 were analyzed by Western blot, respectively. Molecular weights are indicated on was degraded by calpains after expo- the right of the immunoblots. Straight Western blot analysis for SAP102 and actin was sure to dieldrin in mature neurons. Fur- performed in membrane extracts from control and E1-treated cells (Panel D, bottom). thermore, the loss of association between SAP102 and NR2B observed immunostaining was lower in the plasma membrane and after dieldrin treatment supports the idea that truncated higher in the cytosol after long-term exposure to dieldrin NR2B subunits are no longer able to interact with scaf-

(11). Likewise, chronic treatment with other GABAAR fold proteins despite remaining on the cell membrane. antagonists has been shown to diminish NR1, NR2A, and The proteolysis of the NR2B but not the NR2A subunit NR2B membrane levels (9). In our hands, the NR1 and by calpains in mature cortical neurons (43) and in vivo NR2B subunits were those that were first affected by diel- after ischemia or status epilepticus (48, 49) has been drin exposure. In addition, unlike NR2A, which was reg- reported previously. However, the physiological or ulated at the mRNA level, these NMDAR subunits were pathological role of truncated NR2B subunits present in internalized after dieldrin exposure. Overall, distinct the membrane remains unknown. mechanisms appear to be involved in the regulation of Although NR1 and NR2B are first expressed in cul- NMDAR subunits by this endocrine disruptor. tured cortical neurons clustering at both synaptic and at NR2 trafficking is regulated by dual palmitoylation extrasynaptic sites as early as 3 DIV, NR2A is almost un- that can either cause retention in the Golgi apparatus or detectable up to 7 DIV but becomes predominant at the tyrosine phosphorylation-mediated membrane stabiliza- synapsis in mature neurons (39). Furthermore, NR2A but tion (46). Activity-dependent phosphorylation of NR2B not NR2B mRNA expression is dependent on synaptic at Tyr1472 targets NMDAR to synapses (40, 41). More- activity mediated through NMDAR (50). Thus, dieldrin- over, association with PSD95 stops NR2A being internal- induced internalization of NR2B-containing NMDAR

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 11 February 2014. at 02:57 For personal use only. No other uses without permission. . All rights reserved. Endocrinology, February 2012, 153(2):847–860 endo.endojournals.org 855

reported to decrease, increase, or remain unaffected after E2 treatment (27, 51, 52). In the present study, we found that treatment for 2 DIV with 1 nM but not 10 nM E2 increased the membrane expression of NR1, NR2B, and PSD95. In contrast, 10 nM E2 reduced NR2A expression as previously described (27), whereas 1 nM E2 did not. Thus, the conflicting results obtained by different studies may be due to the different doses of E2 used and to the different NMDAR subunits analyzed. In support of this, it has been reported that the effects of E2 on neuronal outgrowth in cultured cortical neurons show an inverted U- shaped dose-response curve and moreover that they de- pend on the cortical region (14). The long-term effects of E2 on NMDAR expression involve de novo protein synthesis as a result of nuclear ER transactivation (51, 52). Furthermore, E2 treatment did not prevent the internalization of NR1 and NR2B induced by dieldrin. All together, these findings indicate that the mechanisms by which these compounds modulate NR2B- containing NMDAR differ. In addition, we found that the increase in membrane PSD95 levels induced by 1 nM E2 FIG. 6. Effects of neurosteroids on dieldrin-induced reduction of cell did not result in an increased association with NMDAR. viability in immature (panel A) and mature (panel B) cortical neurons Similarly, we have recently reported that the association (CN). Panel A, Immature cortical neurons were exposed from 2–5 DIV between PSD95 and NR2B in the preoptic area is not mod- to DMSO or 3 ␮M dieldrin in the absence [control (C)] and presence of 10 ␮M allopregnanolone (AP) or 1 nM E2 (E1). Panel B, Mature cortical ified in proestrus vs. diestrus or after E2 treatment of neurons were exposed from 5–9 DIV to DMSO or 10 ␮M dieldrin in the ovariectomized rats (35, 36). Our results suggest different absence [control (C)] and presence of 10 ␮M allopregnanolone (AP) or mechanisms for the regulation of PSD95 and NMDAR 1nM E2 (E1). MTT values are expressed as percentage with respect to expression by E2. Consistent with this, previous studies control (DMSO-treated cells) and are mean Ϯ SE of four to five independent experiments, each in triplicate. Statistical comparisons have shown that E2 promotes PSD95 expression via the were made by two-way ANOVA: ***, P Ͻ 0.001 vs. control; #, P Ͻ phosphatidylinositol 3-kinase/Akt pathway and ER␤ ac- Ͻ 0.05; ###, P 0.001 vs. dieldrin-treated cells. tivation (15, 53). On the other hand, activation of ER␣ is responsible for the E2-induced up-regulation of NR1/ precedes and maybe causes the disruption in the develop- NR2B receptors in hippocampal neurons, whereas both mental up-regulation of NR2A observed at a higher con- ER appear to contribute in the cortex (19, 52). The ex- centration of pesticide. However, more studies are re- pression of ER␣ and ER␤ in cortical neurons (2) would quired to confirm this hypothesis. We recently reported support the contribution of both isoforms in these estro- that dieldrin mimics some of the effects of E2 in cortical genic actions. neurons (2). Therefore, another possibility is that diel- Although E2 treatment counteracted the reduction of drin could modulate NR2A expression through an ER- NR1 and NR2B membrane expression and NMDAR dependent mechanism. However, this is unlikely be- functionality induced by long-term exposure to dieldrin, it cause the affinity of dieldrin for ER in cortical neurons did not protect cortical neurons from its neurotoxic ef- is at least two orders of magnitude higher (ϳ30 ␮M) (2) fects. The inability of E2 to prevent both GABAAR block- than the concentrations used here for long-term exper- ade and subsequent NR2A down-regulation induced by iments (60–200 nM). prolonged exposure to this pesticide could reflect its lack of neuroprotective effects against dieldrin-induced neuro- Effects of E2 and allopregnanolone on GABA and nal damage. In addition, it has been reported that E2 has glutamate neurotransmission negative effects on cortical GABA levels (18), which may E2 positively modulates synaptic plasticity and en- also account for its lack of neuroprotection. In contrast, hances memory performance, these effects being associ- allopregnanolone reversed the inhibitory effects of diel- ated with an increase in synaptic NR1, NR2B, and PSD95 drin at the GABAAR and restored the functionality and protein levels in the hippocampus (15, 19, 20). In contrast, membrane expression of the NMDAR. These positive NMDAR expression and binding in the cortex has been actions on both GABAergic and glutamatergic neu-

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 11 February 2014. at 02:57 For personal use only. No other uses without permission. . All rights reserved. 856 Briz et al. Allopregnanolone Prevents Dieldrin Toxicity Endocrinology, February 2012, 153(2):847–860

FIG. 7. Effects of allopregnanolone and dieldrin on GABAA functionality, calpain activity, and calpain-mediated cleavage of NR2 subunits of NMDAR in mature cortical neurons. Cells were exposed to DMSO [control (C)] or 10 ␮M dieldrin (D) for 2 DIV both in the absence and presence of 10 ␮M allopregnanolone (AP). Panel A, After the treatments, the cells were exposed to 1.3–2.0 nM [3H]flunitrazepam in HBSS in the absence (basal) or presence of 30 ␮M GABA. Values are expressed as percentage of basal binding and are mean Ϯ SE of three independent experiments, each in triplicate. Statistical comparisons were made by two-way ANOVA: ***, P Ͻ 0.001 vs. basal; #, P Ͻ 0.05; ###, P Ͻ 0.001 vs. DMSO-treated cells; $$, P Ͻ 0.01 vs. dieldrin-treated cells. Panel B, After the treatments, the protein levels of NR2A (bottom), NR2B (top), and actin were analyzed on membrane fractions from the cultures by Western blot. Panel C, Calpain and caspase activities were determined by Western blot analysis of ␣-spectrin in whole-cell lysates from cortical neurons previously exposed to 10 ␮M dieldrin and/or 10 ␮M allopregnanolone for 2 DIV. Panel D, After the treatments, membrane extracts from the cultures were straightly subjected to Western blot analysis of SAP102 and actin (bottom) or immunoprecipitated with an antibody raised against NR2B and subsequently analyzed by Western blot for NR2B and SAP102 (top). Molecular weights are indicated on the right (panels B and D) or on the left (panel C) of the immunoblots. rotransmission may underlie the neuroprotective effect tylbicyclophosphorothionate binding (21, 54). Thus, al- of allopregnanolone against dieldrin-induced cell death. lopregnanolone most likely interferes with dieldrin bind-

Consistent with this, neuroprotection by allopregnano- ing to GABAAR, and as a result, this neurosteroid can lone against both in vitro and in vivo models of ischemia prevent the negative regulation of NMDAR and the neu- has been related to GABAAR stabilization on the cell mem- ronal damage induced by long-term exposure to dieldrin. brane (25). Accordingly, allopregnanolone was found to be anticon- Both acute and prolonged exposure to dieldrin block vulsant and neuroprotective against the noncompetitive the GABAAR through its interaction with the t-butylbicy- GABAAR antagonist picrotoxinin (22, 55), which shares clophosphorothionate binding site (3, 11). The neuroste- the same binding site with dieldrin (3). Nevertheless, the roid allopregnanolone has been shown to acutely increase allosteric properties of allopregnanolone at the benzodi- [3H]flunitrazepam binding and also to reduce [35S]t-bu- azepine sites are lost after prolonged exposure (56), ap-

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 11 February 2014. at 02:57 For personal use only. No other uses without permission. . All rights reserved. Endocrinology, February 2012, 153(2):847–860 endo.endojournals.org 857

FIG. 8. Proposed model for the modulation of GABA and glutamate neurotransmission by dieldrin, E2, and allopregnanolone. A, Under normal conditions, cortical neurons express NR2A-containing NMDAR that are associated with PSD95 (and targeted to synaptic sites) and NR2B-containing NMDAR linked to SAP102, which are mainly located at extrasynaptic sites and so with a high rate of internalization. Some NR2B subunits are phosphorylated at Tyr1472, modification that targets the receptor to the PSD and impedes its internalization. B, The sustained inhibition of the

GABAAR produced by dieldrin (through binding at the channel pore) causes a compensatory scaling down of NMDAR, which starts with a reduction on NR2B tyrosine phosphorylation and translocation of NR1/NR2B receptors to extrasynaptic sites, followed by an increased 2ϩ internalization rate of these receptors and a reduced NR2A expression. As a result of this, the increase in [Ca ]i in response to NMDAR activation 2ϩ is lower than in control cells. C, E2 increases the membrane expression of NR1/NR2B receptors (enhancing [Ca ]i rise in response to NMDAR activation with respect to control cells) and of PSD95, but it does not induce their interaction. D, E2 counteracts the reduction in membrane NR2B- containing NMDAR induced by dieldrin, but it does not prevent their internalization nor NR2A down-regulation. Although NMDAR functionality is similar to control cells, GABAAR function is still impaired. E, Allopregnanolone binds to the GABAAR, and it does not affect NMDAR functionality or subunit expression/localization. F, As an allosteric modulator, allopregnanolone produces conformational changes at the GABAAR that most likely impedes dieldrin binding to the receptor. Therefore, allopregnanolone restores GABAAR function in the presence of dieldrin and reverses the effects of the endocrine disruptor on NMDAR functionality and subunit expression/localization. Figures were produced using Servier Medical Art (http://www.servier.com/servier-medical-art). parently due to GABAAR desensitization (57). This may number of benzodiazepine binding sites (basal binding) explain the reduced response to GABA observed in allo- remains unchanged. Furthermore, although the EC50 of pregnanolone-treated neurons, despite the fact that the allopregnanolone on [3H]flunitrazepam binding is around

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 11 February 2014. at 02:57 For personal use only. No other uses without permission. . All rights reserved. 858 Briz et al. Allopregnanolone Prevents Dieldrin Toxicity Endocrinology, February 2012, 153(2):847–860

␮ ␮ 1 M, a full potentiation of GABAARby10 M allopreg- Concluding remarks nanolone (21) was necessary to prevent the neuronal dam- In summary, the present study shows that prolonged

age caused by this pesticide, as it was also the case for GABAAR blockade by dieldrin triggers the internalization ␣-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid- of NR2B-containing NMDAR and also disrupts NR2A and NMDA-mediated excitotoxicity (24, 28). The results subunit developmental up-regulation. Both NR2A- and indicate that the allosteric properties of allopregnanolone NR2B-containing NMDAR have a crucial role in cortical (present at nanomolar concentrations) are not sufficient to synaptic plasticity and memory (30, 32). Therefore, the achieve neuroprotection against dieldrin toxicity, but the results presented here may contribute to our understand- GABA-mimetic action is required. ing of the molecular mechanisms underlying the cognitive and behavioral deficits found in animals after chronic ex- Mechanisms of dieldrin neurotoxicity in cortical posure to this endocrine disruptor (12, 13). In addition, neurons the observed neuroprotective actions of allopregnanolone The neurotoxicity of dieldrin in dopaminergic neurons support its use as a potential therapeutic hormone against involves the production of reactive oxygen species and insults or neurodegenerative disorders in which GABAe- caspase-dependent apoptosis (5). We have recently ob- rgic and/or glutamatergic systems are compromised. served that the dieldrin-induced apoptosis of cerebellar granule cells is also mediated by caspases (manuscript in preparation). However, these proteases do not appear to Acknowledgments be activated in cortical neurons after dieldrin exposure, as We thank Dr. Alexander Kasal Institute of Organic Chemistry indicated by the absence of caspase-dependent spectrin and Biochemistry, Prague, Czech Republic) for supplying degradation. In contrast, the 150-kDa fragment in the allopregnanolone. ␣-spectrin proteolytic pattern suggests that calpains were active before neuronal death was observed in dieldrin- Address all correspondence and requests for reprints to: treated cortical neurons. Nevertheless, the results do not Cristina Sun˜ ol, Department of Neurochemistry and Neurophar- rule out the possible involvement of caspases in late stages macology, Institut d’Investigacions Biome`diques de Barcelona, of the neuronal damage caused by dieldrin, because the Consejo Superior de Investigaciones Científicas, Rossello´ 161, caspase-specific fragments of ␣-spectrin have been shown E-08036, Barcelona, Spain. E-mail: [email protected]. to appear later than calpain-dependent fragments after This work was supported by the Spanish Ministry of Science and Innovation (FIS 10/0453) (to C.S.) and Generalitat de Cata- prolonged seizures (58). Furthermore, excessive NMDA- lunya (2009/SGR/214) (to C.S.); the French Agence Nationale mediated calpain activation can trigger both proapoptotic pour la Recherche Grants ANR-07-NEURO-026-03 and ANR- pathways through the Bcl-2 family of proteins and necro- 09-BLAN-0267 (to V.P.); the European Society for Pediatric En- sis, as has been reported after status epilepticus and oxy- docrinology Research Unit Grant 2009-2011 coordinated by Dr. gen-glucose deprivation (59, 60). In this regard, allopreg- Sabine Heger (Children’s Hospital Bult, Hannover, Germany) nanolone protects against different toxic insults through (to V.P.). the modulation of Bcl-2 proteins (23, 28). Overall, the Disclosure Summary: The authors have nothing to disclose. involvement of cysteine proteases and the Bcl-2 family of proteins in both dieldrin neurotoxicity and neuroprotection by allopregnanolone deserves further investigation. References Suppression of NMDAR activity (as observed in 1. Soto AM, Chung KL, Sonnenschein C 1994 The pesticides endo- long-term dieldrin-treated cells) can also trigger apo- sulfan, toxaphene, and dieldrin have estrogenic effects on human ptosis in cortical neurons, which can be reversed by estrogen-sensitive cells. Environ Health Perspect 102:380–383 restoring synaptic activity (38). In addition, NR2A-con- 2. Briz V, Molina-Molina JM, Sańchez-Redondo S, Fernańdez MF, Grimalt JO, Olea N, Rodríguez-Farre´ E, SunõlC2011 Differential taining NMDAR appear to mediate the prosurvival effects estrogenic effects of the persistent organochlorine pesticides diel- of NMDA, whereas activation of NR2B-containing drin, endosulfan, and lindane in primary neuronal cultures. Toxicol NMDAR results in excitotoxicity (31). This might explain Sci 120:413–427 why prolonged exposure to dieldrin initially prevents glu- 3. Pome´s A, Rodríguez-Farre´ E, SunõlC 1994 Disruption of GABA- dependent chloride flux by cyclodienes and hexachlorocyclo- tamate excitotoxicity (11) but eventually leads to neuro- hexanes in primary cultures of cortical neurons. J Pharmacol Exp degeneration at higher concentrations. All together, the Ther 271:1616–1623 results suggest that the regulation of NR2A expression 4. Corrigan FM, Wienburg CL, Shore RF, Daniel SE, Mann D 2000 Organochlorine insecticides in substantia nigra in Parkinson’s dis- plays an important role in the neurotoxic effects of ease. J Toxicol Environ Health A 59:229–234 dieldrin. 5. Kitazawa M, Anantharam V, Kanthasamy AG 2003 Dieldrin in-

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 11 February 2014. at 02:57 For personal use only. No other uses without permission. . All rights reserved. Endocrinology, February 2012, 153(2):847–860 endo.endojournals.org 859

duces apoptosis by promoting caspase-3-dependent proteolytic and allopregnanolone protect sympathoadrenal medulla cells cleavage of protein kinase C␦ in dopaminergic cells: relevance to against apoptosis via antiapoptotic Bcl-2 proteins. Proc Natl Acad oxidative stress and dopaminergic degeneration. Neuroscience 119: Sci USA 101:8209–8214 945–964 24. Ciriza I, Azcoitia I, Garcia-Segura LM 2004 Reduced progesterone 6. Liu J, Morrow AL, Devaud LL, Grayson DR, Lauder JM 1997 Reg- metabolites protect rat hippocampal neurones from kainic acid ex-

ulation of GABAA receptor subunit mRNA expression by the pes- citotoxicity in vivo. J Neuroendocrinol 16:58–63 ticide dieldrin in embryonic brainstem cultures: a quantitative, com- 25. Kelley MH, Taguchi N, Ardeshiri A, Kuroiwa M, Hurn PD, Tray- petitive reverse transcription-polymerase chain reaction study. stman RJ, Herson PS 2008 Ischemic insult to cerebellar Purkinje cells J Neurosci Res 49:645–653 causes diminished GABAA receptor function and allopregnanolone 7. Pe´rez-Otan˜ o I, Ehlers MD 2005 Homeostatic plasticity and NMDA neuroprotection is associated with GABAA receptor stabilization. receptor trafficking. Trends Neurosci 28:229–238 J Neurochem 107:668–678 8. Rao A, Craig AM 1997 Activity regulates the synaptic localization 26. Mendez P, Azcoitia I, Garcia-Segura LM 2005 Interdependence of of the NMDA receptor in hippocampal neurons. Neuron 19:801– oestrogen and insulin-like growth factor-I in the brain: potential for 812 analysing neuroprotective mechanisms. J Endocrinol 185:11–17 9. Swann JW, Le JT, Lam TT, Owens J, Mayer AT 2007 The impact 27. Numakawa Y, Matsumoto T, Yokomaku D, Taguchi T, Niki E, of chronic network hyperexcitability on developing glutamatergic Hatanaka H, Kunugi H, Numakawa T 2007 17␤-Estradiol protects synapses. Eur J Neurosci 26:975–991 cortical neurons against oxidative stress-induced cell death 10. Babot Z, Vilaro´ MT, SunõlC2007 Long-term exposure to dieldrin through reduction in the activity of mitogen-activated protein reduces gamma-aminobutyric acid type A and N-methyl-D-aspar- kinase and in the accumulation of intracellular calcium. Endo- tate receptor function in primary cultures of mouse cerebellar gran- crinology 148:627–637 ule cells. J Neurosci Res 85:3687–3695 28. Xilouri M, Papazafiri P 2006 Anti-apoptotic effects of allopreg- 11. Briz V, Galofre´ M, SunõlC2010 Reduction of glutamatergic neu- nanolone on P19 neurons. Eur J Neurosci 23:43–54 rotransmission by prolonged exposure to dieldrin involves NMDA 29. Monyer H, Burnashev N, Laurie DJ, Sakmann B, Seeburg PH 1994 receptor internalization and metabotropic glutamate receptor 5 Developmental and regional expression in the rat brain and func- downregulation. Toxicol Sci 113:138–149 tional properties of four NMDA receptors. Neuron 12:529–540 12. Carlson JN, Rosellini RA 1987 Exposure to low doses of the envi- 30. Barker GR, Warburton EC, Koder T, Dolman NP, More JC, Aggle- ronmental chemical dieldrin causes behavioral deficits in animals ton JP, Bashir ZI, Auberson YP, Jane DE, Brown MW 2006 The prevented from coping with stress. Psychopharmacology (Berl) 91: different effects on recognition memory of perirhinal kainate and 122–126 NMDA glutamate receptor antagonism: implications for underlying 13. Topinka MA, Eisenstein EM, Siegel D, Trofeli R, Barraco D 1984 plasticity mechanisms. J Neurosci 26:3561–3566 The effects of dieldrin and chlordimeform on learning and memory 31. Liu Y, Wong TP, Aarts M, Rooyakkers A, Liu L, Lai TW, Wu DC, in the cockroach, Periplaneta americana: a study in behavioral tox- Lu J, Tymianski M, Craig AM, Wang YT 2007 NMDA receptor icology. J Toxicol Environ Health 13:705–719 subunits have differential roles in mediating excitotoxic neuronal 14. Brinton RD, Tran J, Proffitt P, Montoya M 1997 17␤-Estradiol death both in vitro and in vivo. J Neurosci 27:2846–2857 enhances the outgrowth and survival of neocortical neurons in cul- 32. Massey PV, Johnson BE, Moult PR, Auberson YP, Brown MW, ture. Neurochem Res 22:1339–1351 Molnar E, Collingridge GL, Bashir ZI 2004 Differential roles of 15. Liu F, Day M, Mun˜ iz LC, Bitran D, Arias R, Revilla-Sanchez R, NR2A and NR2B-containing NMDA receptors in cortical long- Grauer S, Zhang G, Kelley C, Pulito V, Sung A, Mervis RF, Navarra term potentiation and long-term depression. J Neurosci 24:7821– R, Hirst WD, Reinhart PH, Marquis KL, Moss SJ, Pangalos MN, 7828 Brandon NJ 2008 Activation of estrogen receptor-␤ regulates hip- 33. van Zundert B, Yoshii A, Constantine-Paton M 2004 Receptor com- pocampal synaptic plasticity and improves memory. Nat Neurosci partmentalization and trafficking at glutamate synapses: a develop- 11:334–343 mental proposal. Trends Neurosci 27:428–437 16. Mellon SH 2007 Neurosteroid regulation of central nervous system 34. Parkash J, d’Anglemont de Tassigny X, Bellefontaine N, Campagne development. Pharmacol Ther 116:107–124 C, Mazure D, Bueé-Scherrer V, Prevot V 2010 Phosphorylation of 17. Schwarz JM, McCarthy MM 2008 Steroid-induced sexual differ- N-methyl-D-aspartic acid receptor-associated neuronal nitric oxide entiation of the developing brain: multiple pathways, one goal. synthase depends on estrogens and modulates hypothalamic nitric J Neurochem 105:1561–1572 oxide production during the ovarian cycle. Endocrinology 151: 18. Blurton-Jones M, Tuszynski MH 2006 Estradiol-induced modula- 2723–2735 tion of estrogen receptor-␤ and GABA within the adult neocortex: 35. d’Anglemont de Tassigny X, Campagne C, Dehouck B, Leroy D, a potential transsynaptic mechanism for estrogen modulation of Holstein GR, Beauvillain JC, Bueé-Scherrer V, Prevot V 2007 Cou- BDNF. J Comp Neurol 499:603–612 pling of neuronal nitric oxide synthase to NMDA receptors via post- 19. Jelks KB, Wylie R, Floyd CL, McAllister AK, Wise P 2007 Estradiol synaptic density-95 depends on estrogen and contributes to the cen- targets synaptic proteins to induce glutamatergic synapse formation tral control of adult female reproduction. J Neurosci 27:6103–6114 in cultured hippocampal neurons: critical role of estrogen recep- 36. d’Anglemont de Tassigny X, Campagne C, Steculorum S, Prevot V tor-␣. J Neurosci 27:6903–6913 2009 Estradiol induces physical association of neuronal nitric oxide 20. Smith CC, McMahon LL 2006 Estradiol-induced increase in the synthase with NMDA receptor and promotes nitric oxide formation magnitude of long-term potentiation is prevented by blocking via estrogen receptor activation in primary neuronal cultures. J Neu- NR2B-containing receptors. J Neurosci 26:8517–8522 rochem 109:214–224 21. Sun˜ ol C, García DA, Bujons J, Kristofíkova´ Z, Matya´s L, Babot Z, 37. Chomczynski P, Sacchi N 1987 Single-step method of RNA isolation Kasal A 2006 Activity of B-nor analogues of neurosteroids on the by acid guanidinium thiocyanate-phenol-chloroform extraction.

GABAA receptor in primary neuronal cultures. J Med Chem 49: Anal Biochem 162:156–159 3225–3234 38. Le´veille´ F, Papadia S, Fricker M, Bell KF, Soriano FX, Martel MA, 22. Brinton RD 1994 The neurosteroid 3␣-hydroxy-5␣-pregnan-20- Puddifoot C, Habel M, Wyllie DJ, Ikonomidou C, Tolkovsky AM, one induces cytoarchitectural regression in cultured fetal hippocam- Hardingham GE 2010 Suppression of the intrinsic apoptosis path- pal neurons. J Neurosci 14:2763–2774 way by synaptic activity. J Neurosci 30:2623–2635 23. Charalampopoulos I, Tsatsanis C, Dermitzaki E, Alexaki VI, Cas- 39. Zhong J, Russell SL, Pritchett DB, Molinoff PB, Williams K 1994 tanas E, Margioris AN, Gravanis A 2004 Dehydroepiandrosterone Expression of mRNAs encoding subunits of the N-methyl-D-aspar-

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 11 February 2014. at 02:57 For personal use only. No other uses without permission. . All rights reserved. 860 Briz et al. Allopregnanolone Prevents Dieldrin Toxicity Endocrinology, February 2012, 153(2):847–860

tate receptor in cultured cortical neurons. Mol Pharmacol 45:846– tivity-dependent developmental regulation of NMDA receptor sub- 853 unit expression in cultured neocortical neurons. J Neurochem 75: 40. Frasca A, Aalbers M, Frigerio F, Fiordaliso F, Salio M, Gobbi M, 1590–1599 Cagnotto A, Gardoni F, Battaglia GS, Hoogland G, Di Luca M, 51. Brann DW, Zamorano PL, Chorich LP, Mahesh VB 1993 Steroid Vezzani A 2011 Misplaced NMDA receptors in epileptogenesis con- hormone effects on NMDA receptor binding and NMDA receptor tribute to excitotoxicity. Neurobiol Dis 43:507–515 mRNA levels in the hypothalamus and cerebral cortex of the adult 41. Prybylowski K, Chang K, Sans N, Kan L, Vicini S, Wenthold RJ rat. Neuroendocrinology 58:666–672 2005 The synaptic localization of NR2B-containing NMDA recep- 52. Morissette M, Le Saux M, Di Paolo T 2008 Effect of oestrogen tors is controlled by interactions with PDZ proteins and AP-2. Neu- receptor ␣- and ␤-agonists on brain N-methyl-D-aspartate receptors. ron 47:845–857 J Neuroendocrinol 20:1006–1014 42. Dong YN, Waxman EA, Lynch DR 2004 Interactions of postsyn- 53. Akama KT, McEwen BS 2003 Estrogen stimulates postsynaptic aptic density-95 and the NMDA receptor 2 subunit control calpain- density-95 rapid protein synthesis via the Akt/protein kinase B path- mediated cleavage of the NMDA receptor. J Neurosci 24:11035– way. J Neurosci 23:2333–2339 11045 54. Hawkinson JE, Kimbrough CL, Belelli D, Lambert JJ, Purdy RH, Lan NC 43. Dong YN, Wu HY, Hsu FC, Coulter DA, Lynch DR 2006 Devel- 1994 Correlation of neuroactive steroid modulation of [35S]t-butylbicy- opmental and cell-selective variations in N-methyl-D-aspartate re- clophosphorothionate and [3H]flunitrazepam binding and ␥-aminobu-

ceptor degradation by calpain. J Neurochem 99:206–217 tyric acidA receptor function. Mol Pharmacol 46:977–985 44. Rajgopal Y, Vemuri MC 2002 Calpain activation and ␣-spectrin 55. Belelli D, Bolger MB, Gee KW 1989 Anticonvulsant profile of the cleavage in rat brain by ethanol. Neurosci Lett 321:187–191 progesterone metabolite 5␣-pregnan-3␣-ol-20-one. Eur J Pharma- 45. Turrigiano GG, Leslie KR, Desai NS, Rutherford LC, Nelson SB col 166:325–329 ␥ 1998 Activity-dependent scaling of quantal amplitude in neocortical 56. Friedman L, Gibbs TT, Farb DH 1993 -Aminobutyric acidA re- neurons. Nature 391:892–896 ceptor regulation: chronic treatment with pregnanolone uncouples 46. Hayashi T, Thomas GM, Huganir RL 2009 Dual palmitoylation of allosteric interactions between steroid and benzodiazepine recogni- NR2 subunits regulates NMDA receptor trafficking. Neuron 64: tion sites. Mol Pharmacol 44:191–197 213–226 57. Gyenes M, Wang Q, Gibbs TT, Farb DH 1994 Phosphorylation 47. Lavezzari G, McCallum J, Dewey CM, Roche KW 2004 Subunit- factors control neurotransmitter and neuromodulator actions at the specific regulation of NMDA receptor endocytosis. J Neurosci 24: ␥-aminobutyric acid type A receptor. Mol Pharmacol 46:542–549 6383–6391 58. Wang S, Wang S, Shan P, Song Z, Dai T, Wang R, Chi Z 2008 48. Arau´ jo IM, Xapelli S, Gil JM, Mohapel P, Peterse´n A, Pinheiro PS, Mu-calpain mediates hippocampal neuron death in rats after lith- Malva JO, Bahr BA, Brundin P, Carvalho CM 2005 Proteolysis of ium-pilocarpine-induced status epilepticus. Brain Res Bull 76: NR2B by calpain in the hippocampus of epileptic rats. Neuroreport 90–96 16:393–396 59. Fujikawa DG 2005 Prolonged seizures and cellular injury: under- 49. Simpkins KL, Guttmann RP, Dong Y, Chen Z, Sokol S, Neumar standing the connection. Epilepsy Behav 7(Suppl 3):S3–S11 RW, Lynch DR 2003 Selective activation induced cleavage of the 60. Yamashima T 2000 Implication of cysteine proteases calpain, ca- NR2B subunit by calpain. J Neurosci 23:11322–11331 thepsin and caspase in ischemic neuronal death of primates. Prog 50. Hoffmann H, Gremme T, Hatt H, Gottmann K 2000 Synaptic ac- Neurobiol 62:273–295