Address correspondence and reprint requests to Rachel Hill, PhD, Behavioural Neuroscience Laboratory, Florey Institute of Neuroscience and Mental Health, Kenneth Myer Building, University of Melbourne, Parkville, VIC 3010, Australia. E-mail: email@example.com
Neurodevelopmental psychiatric disorders such as schizophrenia may be caused by a combination of gene × environment, gene × gene, and/or gene × sex interactions. Reduced expression of both Reelin and Brain-Derived Neurotrophic factor (BDNF) has been associated with schizophrenia in human post-mortem studies. However, it remains unclear how Reelin and BDNF interact (gene × gene) and whether this is sex-specific (gene × sex). This study investigated BDNF-TrkB signaling in the hippocampus of male and female Reelin heterozygous (Rln+/−) mice. We found significantly increased levels of BDNF in the ventral hippocampus (VHP) of female, but not male Rln+/− compared to wild-type (WT) controls. While levels of TrkB were not significantly altered, phosphorylated TrkB (pTrkB) levels were significantly lower, again only in female Rln+/− compared to WT. This translated to downstream effects with a significant decrease in phosphorylated ERK1 (pERK1). No changes in BDNF, TrkB, pTrkB or pERK1/2 were observed in the dorsal hippocampus of Rln+/− mice. Ovariectomy (OVX) had no effect in WT controls, but caused a significant decrease in BDNF expression in the VHP of Rln+/− mice to the levels of intact WT controls. The high expression of BDNF was restored in OVX Rln+/− mice by 17β-estradiol treatment, suggesting that Rln+/− mice respond differently to an altered estradiol state than WT controls. In addition, while OVX had no significant effect on TrkB or ERK expression/phosphorylation, OVX + estradiol treatment markedly increased TrkB and ERK1 phosphorylation in Rln+/− and, to a lesser extent in WT controls, compared to intact genotype-matched controls. These data may provide a better understanding of the interaction of Reelin and BDNF in the hippocampus, which may be involved in schizophrenia.
Schizophrenia and gene × gene, gene × sex interactions
Several lines of evidence suggest that the hippocampus is central to the pathophysiology of schizophrenia. For example, subjects with schizophrenia show deficits in episodic memory, which is linked to hippocampal function (Weinberger 1999). Neuroimaging studies have shown reductions in the size and volume of the hippocampus in schizophrenia patients (Bogerts et al. 1990; Lawrie and Abukmeil 1998) and correlations between psychotic symptoms and hippocampal metabolic activity have been shown (Tamminga et al. 1992; Medoff et al. 2001). Animal models of disrupted hippocampal function have shown alterations in schizophrenia-like behaviors, such as locomotor hyperactivity, cognitive impairments, and disruptions to latent inhibition and prepulse inhibition [see review, (Lipska and Weinberger 2000)].
In the hippocampus of schizophrenia patients, markedly decreased protein and mRNA expression levels of both reelin (Impagnatiello et al. 1998; Fatemi et al. 2000) and of brain-derived neurotrophic factor (BDNF) and its receptor, tyrosine kinase receptor (TrkB) (Thompson Ray et al. 2011) have been found. Sex differences have been observed in the association between serum BDNF levels and schizophrenia; male patients with schizophrenia showed lower serum levels of BDNF compared to female patients, while no sex differences were observed in healthy controls (Xiu et al. 2009).
Reelin is an extracellular matrix glycoprotein involved during early embryogenesis in the migration of cerebral cortex neurons to their target layers (Badea et al. 2007). At post-natal stages, reelin is expressed at high levels in subsets of GABAergic neurons throughout the neocortex and hippocampus (Alcantara et al. 1998). In reeler homozygous knockout mice, the pyramidal cell layer of the hippocampus proper is split into a double layered structure, while granular cells of the dentate gyrus appear scattered (Stanfield and Cowan 1979). Reelin heterozygous (Rln+/−) mice show an approximate 50% reduction in protein expression of reelin in brain extracts including the corpus striatum and anterior hippocampus (Tueting et al. 1999), levels similar to those previously reported in schizophrenia patients (Impagnatiello et al. 1998). Rln+/− mice show alterations in hippocampal synaptic function (Weeber et al. 2002; Qiu et al. 2006) and both homozygous and heterozygous reeler mice show significant reductions in hippocampal spine density (Liu et al. 2001; Pappas et al. 2001; Niu et al. 2008).
BDNF also plays a crucial role in neuronal survival and migration, axonal and dendritic growth, and synaptic plasticity (Buckley et al. 2007). The mature form of BDNF binds to its receptor TrkB (Buckley et al. 2007), and causes phosphorylation of the receptor and consequent activation of a number of downstream signaling pathways involved in synaptic plasticity including mitogen-activated protein kinase/extracellular signal-regulated kinase (MAPK/ERK) (Yoshii and Constantine-Paton 2010).
Previous studies have suggested an interaction of reelin and BDNF. For example, protein levels of BDNF were increased, but TrkB phosphorylation was decreased in the frontal cortex of reelin heterozygous compared to wild-type (WT) control mice (Pillai and Mahadik 2008). However, it remains unclear if a similar interaction is observed in the hippocampus, and furthermore this previous study did not mention the sex of the mice studied. Given the integral roles that both proteins play during hippocampal development, the objectives of this study were to analyze BDNF-TrkB and downstream ERK1/2 signaling in the hippocampus of Rln+/− mice. Because several studies suggest that BDNF may be sex-dependently regulated and that sex steroid hormones, such as estrogen, play a role in the regulation of BDNF-TrkB signaling and in the pathogenesis of schizophrenia (Blurton-Jones et al. 2004; Sohrabji and Lewis 2006; Kaur et al. 2007; Ottem et al. 2007; Hill and van den Buuse 2011), we compared male and female mice and assessed the effects of ovariectomy and 17β-estradiol replacement.
The hippocampus can be anatomically divided into dorsal and ventral subregions, which have distinct afferent and efferent connections. The dorsal hippocampus (DHP) primarily receives inputs from the neocortex, while the ventral hippocampus (VHP) has direct reciprocal connections with the amygdala and hypothalamus (Fanselow and Dong 2010), nucleus accumbens and medial prefrontal cortex (Swerdlow and Geyer 1998; Wu et al. 2012). The DHP and VHP consequently take on differential functions, with the dorsal segment predominantly involved in cognition, including spatial memory, and the ventral segment said to regulate stress and anxiety-related behaviors (aan het Rot et al. 2009; Wu et al. 2012). Hence, we decided in this study to separately analyze the dorsal and ventral hippocampus.
Male and female reelin heterozygous (Rln+/−) mice and wild-type controls (WT) were derived from a breeding colony at the Florey Neuroscience Institutes, Parkville, Australia (van den Buuse et al. 2012). The animals were genotyped by PCR (Hammond et al. 2006) and housed under standard conditions, with ad libitum access to water and mouse chow. The first cohort of animals was 14–16 weeks of age with five to six animals per group. A second cohort of animals consisting of n =7–10 female WT and Rln+/− mice were either sham-operated or ovariectomized (OVX) at 10 weeks of age and implanted with 1 cm of silastic tubing which contained either 17β-estradiol or was left empty. This dose of estradiol has previously been shown in our laboratory to produce physiological levels of hormone replacement (Hill et al. 2012). During surgery, these mice were anesthetized using an isoflurane/oxygen gas mixture (I.S.O.®; Veterinary Companies of Australia, Artarmon, NSW, Australia). All mice which underwent surgery received a single 5 mg/kg injection of the non-steroidal anti-inflammatory analgesic, Carprofen (Rimadyl®; Pfizer, Sandwich, Kent, UK) to prevent post-operative pain and discomfort. Animals were killed at 15 weeks of age by cervical dislocation, their brains were removed, and the hippocampus dissected. The hippocampus was then further dissected in half into dorsal and ventral regions and snap frozen on dry ice for protein extraction. Uterine weight was recorded in this cohort to assess the effectiveness of OVX and estradiol replacement treatment. All experimental procedures were approved by the Animals Experimentation Ethics Committee of the Florey Neuroscience Institutes, University Of Melbourne, Australia.
Tissue samples were weighed and the appropriate amount of lysis buffer (50 mM Tris pH 8.0, 0.1% sodium dodecyl sulfate (SDS), 1% Triton X-100, 150 mM sodium chloride, dH2O, protease inhibitor cocktail set III (1 : 200), and phosphatase inhibitor cocktail set IV (1 : 50) (Merck, Kilsyth, Vic, Australia) was added according to the weight (1000 μL per 100 μg). Samples were homogenized and left on ice for 10 min, followed by rotation for 1 h. Samples were then centrifuged for 15 min at 14 000 g, and 3 μL of the resulting supernatant was used for protein assay using the BCA protein assay kit (Thermo Scientific, Rockford, IL, USA).
Western blot analysis
Western blot analysis was performed as previously described (Hill et al. 2011). Briefly, sample volume required for 50 μg of protein was added to an equal volume of loading buffer (0.4M Tris pH 6.8, 37.5% glycerol, 10% SDS, 1% 2–mercaptoethanol, 0.5% bromphenol blue, dH2O). Samples were then denatured for 10 min at 95°C before SDS-PAGE (15% or 10% acrylamide gel, 120 V, 1.5 h) and transferred to a nitrocellulose membrane which was incubated with primary antibody overnight at 4°C. Primary antibodies were rabbit anti-mature BDNF (H-117, 1 : 500, Santa Cruz Biotechnology, Santa Cruz, CA, USA) which produced a band of approximately 13.5 kDa, rabbit anti-TrkB (H-181, 1 : 1000, Santa Cruz Biotechnology) which produced two bands at approximately 140 kDa (full length TrkB) and 95 kDa (truncated TrkB), rabbit anti-phosphorylated TrkB (phospho-Tyr705, 1 : 1000, Signalway Antibody, Pearland, TX, USA) which produced a band at approximately 140 kDa, rabbit anti–p44/42 MAPK (ERK1/2) (1 : 2000, Cell Signaling, Danvers, MA, USA) which produced two bands at 44 (ERK1) and 42 kDa (ERK2), mouse anti-phospho p44/42 MAPK (ERK1/2) (1 : 2000, Cell Signaling) which produced two bands at 44 kDa (pERK1) and 42 kDa (pERK2) or mouse anti-β-actin (1 : 5000, Sigma-Aldrich, Castle Hill, NSW, Australia). The next day, the membrane was incubated with either anti-rabbit or anti-mouse IgG horseradish peroxidase-linked secondary antibodies (Cell Signaling). Images were captured using a Luminescence Image Analyzer (Fuji film LAS-4000; FujiFilm Life Science, Stamford, CT, USA) and analyzed using Multi Gauge software (FujiFilm Life Science). BDNF and TrkB levels were normalized against levels of the housekeeping gene, β-actin. In addition, to give an overall indication of TrkB activity, pTrkB levels were expressed as a ratio of full length TrkB levels. This was also the case for ERK1/2, expressed as a ratio of pERK1/2/ERK1/2. Each Western blot assay was repeated two-three times.
All data are expressed as mean ± standard error of the mean (SEM). Statistical analysis was done by two-way analysis of variance (anova) with genotype and sex of the animals or genotype and surgery condition as main factors (Systat 9.0; SPSS Inc, Chicago, IL, USA). Pairwise comparisons were performed by Mann–Whitney U-test between groups (GraphPad PRISM version 3.0 for Windows; GraphPad Software, San Diego, CA, USA). If p-values were less than 0.05, differences were considered statistically significant.
BDNF-TrkB and ERK1/2 signaling in male compared to female WT and Rln+/− mice
In the ventral hippocampus, there was a significant main effect of genotype (F(1,22) = 9.2, p =0.006) and sex (F(1,22) = 5.3, p =0.03) on BDNF expression (Fig. 1a). While no significant sex × genotype interaction was found, pairwise comparison showed a significant increase in BDNF expression in female, but not male Rln+/− compared to WT controls (Fig. 1a, p = 0.01). No significant effects of sex or genotype on full length TrkB expression were found in the VHP (Fig. 1b). However, analysis of TrkB phosphorylation (Fig 1c) revealed a main effect of genotype (F(1,18) = 33.0, p <0.001) and a significant sex × genotype interaction (F(1,18) = 12.0, p =0.003). Pairwise comparison showed a significant decrease in TrkB phosphorylation in female Rln+/− mice compared to female WT controls (Fig. 1c, p = 0.002), while phosphorylation levels were unchanged in male Rln+/− mice. Analysis of ERK1 phosphorylation (pERK1) revealed a significant effect of genotype (F(1,48) = 8.1, p =0.007) and pairwise comparison showed a significant decrease in pERK1 in female Rln+/− when compared to WT controls (Fig. 1d, p = 0.04). In contrast, analysis of ERK2 phosphorylation (pERK2) showed no significant effect of either sex or genotype in the ventral hippocampus (Fig. 1e).
In the dorsal hippocampus, we found no significant effect of either genotype or sex on BDNF (Fig. 2a), TrkB (Fig. 2b), phosphorylated TrkB (Fig 2c), or phosphorylated ERK1/2 (Fig. 2d/e) expression.
The effects of ovariectomy and 17β-estradiol replacement on BDNF-TrkB and ERK1/2 signaling in female WT and Rln+/− mice
There was a borderline significant statistical interaction of surgery condition and genotype on BDNF levels (F(2,58) = 3.2, p = 0.048), suggesting differential effects of OVX and OVX+E2 treatment between WT and Rln+/− mice. Indeed, there was a significant effect of surgery condition in Rln+/− mice (F(2,28) = 4.3, p = 0.024) but not in WT mice. Pairwise comparison confirmed the significantly higher levels of BDNF in intact Rln+/− compared to WT controls (Fig. 3a, p = 0.01) also found in the first cohort. In addition, OVX caused a significant decrease in BDNF expression in Rln+/− mice (Fig. 3a, p = 0.003) with levels reduced to those found in WT controls and less than half those in intact Rln+/− controls. Estradiol replacement was effective in restoring BDNF levels to those of intact Rln+/− mice (Fig 3a, p = 0.04). Neither OVX nor E2 replacement induced any significant effects on BDNF expression in WT mice (Fig. 3a).
No significant effects of genotype or surgery condition were found on full length (Fig. 3b) or truncated TrkB (data not shown) expression in the ventral hippocampus. In this cohort of mice, there was also no statistically significant main effect or interaction of genotype or surgery condition on TrkB phosphorylation, although pairwise comparison showed a strong trend for a similar reduction of pTrkB/TrkB in Rln+/− mice compared to WT controls (Fig. 3c, p = 0.08) as seen in the first cohort. There was also a significant increase in pTrkB/TrkB in OVX+E2 Rln+/− mice compared to intact Rln+/− mice (Fig. 3c, p = 0.04).
Analysis of ERK1 phosphorylation revealed a highly significant effect of surgery condition (main effect: F(2,51) = 16.2, p <0.001) which was different between the genotypes (F(2,51) = 4.4, p =0.017). Although E2 treatment appeared to increase ERK1 phosphorylation in both genotypes (Fig. 3d), this was only significant in Rln+/− mice (F(2,25) = 17.3, p <0.001). Pairwise comparison showed that pERK1/total ERK1 levels were significantly lower in sham-operated Rln+/− mice compared to intact WT controls (p =0.044), similar to the first cohort. OVX by itself had no significant effect on ERK1 phosphorylation; however, OVX + E2 replacement induced significantly increased ERK1 phosphorylation (pERK1/total ERK1) compared to placebo-treated Rln+/− mice (Fig. 3d, p = 0.001) and compared to Rln+/− OVX mice (Fig. 3d, p = 0.003). No significant effects of genotype, OVX or E2 replacement were found in ERK2 phosphorylation (pERK2/tERK2) in the VHP (Fig. 3e).
In the dorsal hippocampus no significant effects of genotype, OVX or E2 replacement were found on BDNF (Fig. 4a), TrkB (Fig. 4b), pTrkB (Fig. 4c) or pERK1 (Fig. 4d) expression. However, for pERK2, there was a main effect of surgery condition (F(2,38) = 4.4, p =0.02) reflecting significantly lower pERK2 levels in OVX+E2 groups of both genotypes compared to sham-operated controls. However, pairwise comparison only showed significant reduction of pERK2 in WT controls (Fig. 4e).
anova revealed that in both WT and Rln+/− mice, OVX caused the expected decrease in uterine weight, whereas treatment with estradiol significantly increased uterine weight when compared to the OVX and sham-operated condition (main effect of surgery condition: F(2,25) = 151.2, p <0.001, for pairwise comparison, see Fig. 5).
This study in Rln+/− mice found significant disruptions to BDNF-TrkB signaling that were specific to the female, but not male VHP. These changes in BDNF-TrkB signaling translated to downstream effects, with a significant decrease found in TrkB and ERK1 phosphorylation, again specific to the VHP. Ovariectomy had no effect in WT controls, but caused a significant decrease in BDNF expression in the VHP of Rln+/− mice to levels seen in sham-operated WT control mice. The high expression of BDNF was restored in OVX Rln+/− mice by 17β-estradiol replacement treatment, indicating a role for this hormone in mediating sex-specific alterations in BDNF expression in Rln+/− mice.
Functional consequences of altered BDNF-TrkB and ERK1/2 signaling in the VHP of Rln+/− mice
The finding of altered BDNF-TrkB and ERK1 signaling in the VHP, but not DHP, in Rln+/− mice is in line with previous observations in reelin homozygous and heterozygous mice that the characteristic shape of the VHP was particularly altered, more so than the DHP (Badea et al. 2007). BDNF plays a crucial role in neuronal survival and migration, axonal and dendritic growth, and synaptic plasticity (Buckley et al. 2007); hence, changes in BDNF-TrkB signaling may contribute to morphological changes of the VHP. Because the DHP and VHP subserve differential functions, with the dorsal segment predominantly involved in cognition, including spatial memory, and the ventral segment modulating stress and anxiety-related behaviors (aan het Rot et al. 2009), changes in BDNF-TrkB signaling may be involved in some of the behavioral changes previously observed in Rln+/− mice. For example, one study has suggested that Rln+/− mice show an age-dependent disruption in prepulse inhibition (PPI), an endophenotype of schizophrenia (Tueting et al. 1999). The hippocampus plays an important role in regulating PPI and this is thought to be mediated by connections from the VHP to the nucleus accumbens and medial prefrontal cortex (Swerdlow and Geyer 1998). Moreover, disruptions in PPI have been shown in VHP-lesioned rats (Swerdlow et al. 2008; Miller et al. 2010). VHP-lesioned rats also show spatial memory deficits, thought to be mediated via disruptions to hippocampal-prefrontal pathways (Brady et al. 2010). Previous studies in Rln+/− mice have furthermore shown decreased contextual fear conditioning and impaired hippocampal long term potentiation (LTP) and long term depression (LTD) (Qiu et al. 2006). Hippocampus-specific deletion of BDNF in adult mice induced impaired spatial learning in the Morris water maze (Heldt et al. 2007). ERK1 is suggested to control mainly cytoplasmic targets involved in fast-acting functions such as LTP (Marchi et al. 2008) and ERK1 knockout mice show subtle changes to synaptic plasticity and learning and memory (Pages et al. 1999; Mazzucchelli et al. 2002; Satoh et al. 2007). Rln+/− mice have also previously been shown to display increased anxiety, which is also associated with VHP function (Tueting et al. 1999). Altered BDNF-TrkB signaling by deletion of TrkB in adult progenitors of the hippocampus has been shown to cause increased anxiety-like behaviors in mice (Bergami et al. 2008). This would be in line with our findings of impairments in BDNF-TrkB signaling in the VHP. Taken together, disruptions to BDNF-TrkB and ERK1 phosphorylation found in this study may contribute to changes in PPI, learning and memory, LTP and anxiety previously found in Rln+/− mice.
A role for sex steroid hormones in regulating BDNF-TrkB signaling in Rln+/− mice
In addition to striking male–female differences, we found that the effects of reelin deficiency on BDNF-TrkB signaling were modulated by ovariectomy in a region-specific manner. In response to decreased TrkB phosphorylation and downstream ERK1 signaling, in female Rln+/− mice, sex steroid hormones, such as estradiol, may induce a compensatory increase in BDNF expression in the VHP. Consequently, when sex steroid hormones are removed by ovariectomy, this compensatory increase in BDNF expression is lost and, instead, BDNF expression is reduced to the level of intact WT controls. This inverse relationship between BDNF and TrkB phosphorylation has previously been shown in the frontal cortex of Rln+/− mice, where BDNF expression was increased, but TrkB phosphorylation was decreased (Pillai and Mahadik 2008). Similarly, OVX increased BDNF expression in the frontal cortex and striatum, but decreased TrkB phosphorylation of adolescent WT C57Bl/6 mice (Hill et al. 2012), once again demonstrating an inverse relationship between BDNF and TrkB. In this study, we show that this compensatory relationship may be altered in the VHP of Rln+/− mice, suggesting a major role for these sex steroid hormones in modulating reelin-BDNF interactions. In addition, this compensatory event would seem to be a direct effect of estradiol on BDNF expression [Fig. 6 (1)] as OVX significantly decreased BDNF expression, but had no effect on TrkB or ERK phosphorylation.
Estradiol replacement was effective in restoring BDNF levels to those of intact Rln+/− mice confirming that this compensatory event is indeed controlled by estradiol. Interestingly, while OVX showed no effect, OVX + E2 significantly increased ERK1 phosphorylation in Rln+/− mice (with a similar trend for TrkB phosphorylation). This suggests that constant administration of estradiol via an implant alters the BDNF-TrkB pathway differently than normally circulating and fluctuating estradiol, with implant E2 administration modulating particularly ERK1 phosphorylation [Fig. 5 (2, 3)] independent of BDNF. Other studies have also shown that pharmacological levels of estradiol may produce fast-acting, ER-independent effects, which could explain the indirect effects on TrkB and ERK, while lower, normally circulating physiological levels induce slower, ER-dependent actions, which may explain the direct actions upon BDNF gene expression in this study [for review, see (Hill and Boon 2009)]. Indeed, even in the normal estrus cycle, direct actions of high levels of estradiol in the proestrus stage have been shown to increase TrkB phosphorylation although ERK1 was not analyzed (Spencer-Segal et al. 2011). The molecular mechanism involved in these effects is likely to be independent of classical nuclear estrogen receptors and could be an estrogen-mediated activation of the MAPK/ERK pathway via activation of upstream kinases, such as B-Raf (Singh et al. 1999) and c-Src (Nethrapalli et al. 2001) or via activation of a putative plasma membrane-associated estrogen receptor (Toran-Allerand et al. 2002). A more recent study, however, has demonstrated that upon stimulation with estradiol, the estrogen receptor α forms a macromolecular signaling complex with the IGF-I receptor in the plasma membrane and this in turn activates downstream MAPK/ERK signaling (Jover-Mengual et al. 2007).
Whichever the molecular mechanisms involved, our results demonstrate that female Rln+/− mice respond differently to estradiol treatment than WT controls in terms of BDNF-TrkB and ERK signaling, and to this end, future studies should analyze the estrogen receptor expression profile (ERα, ERβ, and membrane bound ER) as well as aromastase and estradiol levels in female Rln+/− hippocampus. In a recent study in rats, high expression levels of ERα were found in reelin expressing Cajal-Retzius (CR) cells of the hippocampus, and exogenous application of estrogen in hippocampus slice cultures increased reelin expression, while aromatase inhibition reduced reelin expression in these CR cells (Bender et al. 2010).
Male Rln+/− mice show increased levels of testosterone and 17β-estradiol in the Purkinje cells of the cerebellum (Biamonte et al. 2009) and treatment with 17β-estradiol increased reelin mRNA, again particularly in males (Biamonte et al. 2009). This altered response to estradiol treatment in the cerebellum is similar to our findings in the ventral hippocampus, albeit ours were in females rather than in males. Overall, the available evidence suggests that reelin is highly regulated by estradiol and that BDNF may be a mediator of this interaction.
In contrast to the VHP, the DHP showed no significant changes in BDNF, TrkB, or phosphorylated TrkB expression. However, ERK2 phosphorylation was significantly altered by OVX+E2 in WT animals with a similar trend not reaching significance in Rln+/− mice. One possibility to explain these results may be that progesterone is involved in mediating ERK2 phosphorylation [Fig. 5 (4)]. Indeed, previous reports have shown that bilateral infusion of progesterone into the DHP of OVX mice caused significant increases in phosphor-ERK (Orr et al. 2012).
Functional and clinical implications
Our finding that female, but not male Rln+/− showed altered BDNF-TrkB signaling in the VHP is in line with human studies showing a female-specific association between genetic Reelin abnormalities and schizophrenia risk (Shifman et al. 2008) and with bipolar disorder (Goes et al. 2010). These studies, combined with our data, may suggest that women with a Reelin deficiency may be more susceptible to psychiatric illnesses, such as schizophrenia or bipolar disorder, than men and that ovarian hormones may be involved in this sex difference. Several rodent studies have suggested a role for estradiol in regulating BDNF expression (Blurton-Jones et al. 2004; Sohrabji and Lewis 2006), while in addition testosterone (Ottem et al. 2007) and progesterone (Kaur et al. 2007) have also been shown to alter BDNF-TrkB signaling. Our laboratory has previously shown that estrogen can modulate disruptions of PPI in mice, rats, and humans (Gogos et al. 2006a; Gogos et al. 2006b; Gogos et al. 2010). In addition, estrogens are also involved in the regulation of hippocampus-dependent learning and memory (Spencer et al. 2008). However, further studies are needed to investigate the effects of estrogen treatment on PPI and learning and memory in Rln+/− mice and, by extension, in humans.
We report here a significant disruption to TrkB and particularly downstream ERK1 phosphorylation in the VHP of female, but not male Rln+/− mice, most likely causing a compensatory increase in BDNF expression. The removal of sex steroid hormones by OVX disrupts this compensatory response to reelin deficiency in the VHP, resulting in a dramatic reduction in BDNF expression levels. Replacement treatment with 17β-estradiol effectively restored levels to intact Rln+/− controls. We suggest that ovarian hormones, such as estradiol, may be mediating sex-specific involvement of Reelin as a risk factor in the development of psychiatric illnesses, such as schizophrenia particularly in women (Shifman et al. 2008) via modulation of the BDNF-TrkB signaling pathway. In post-mortem tissue from subjects with schizophrenia, levels of both Reelin and BDNF are significantly reduced. The interaction of these two trophic factors, shown in this study, may be involved in some of the sex differences observed in the pathogenesis of schizophrenia.
The authors acknowledge the technical assistance of Perrin Kwek and Jac Kee Low. This study was supported by a postdoctoral training fellowship (RH), a senior research fellowship (MvdB) and a project grant from the National Health and Medical Research Council (NHMRC) of Australia, as well as operational infrastructure received from the State Government of Victoria. The authors have no conflict of interest to report. RH designed the study, carried out all molecular analysis and wrote the first draft of the manuscript. AG and YWCW contributed to experiments, data analysis, and manuscript editing. MvdB contributed to study design, data analysis, and manuscript editing.