|Home | About | Journals | Submit | Contact Us | Français|
Dysfunctions of the brain serotonin (5-HT) system are often associated with affective disorders, such as depression. The raphe nuclei target the limbic system and most forebrain areas and constitute the main source of 5-HT in the brain. All 5-HT neurons express tryptophan hydroxylase-2 (TPH2), the brain specific, rate-limiting enzyme for 5-HT synthesis. ERbeta agonists have been shown to attenuate anxiety-and despair-like behaviors in rodent models. Therefore, we tested the hypothesis that ERbeta may contribute to the regulation of gene expression in 5-HT neurons of the dorsal raphe nuclei (DRN) by examining the effects of systemic and local application of the selective ERbeta agonist diarylpropionitrile (DPN) on tph2 mRNA expression.
Ovariectomized (OVX) female rats were injected subcutaneously (s.c.) with DPN or vehicle once daily for 8 days. In situ hybridization revealed that systemic DPN-treatment elevated basal tph2 mRNA expression in the caudal and mid-dorsal DRN. Behavioral testing of all animals in the open field (OF) and on the elevated plus maze (EPM) on days 6 and 7 of treatment confirmed the anxiolytic nature of ERbeta activation.
Another cohort of female OVX rats was stereotaxically implanted bilaterally with hormone-containing wax pellets flanking the DRN. Pellets contained either 17-beta-estradiol (E), DPN, or no hormone. Both DPN and E significantly enhanced tph2 mRNA expression in the mid-dorsal DRN. DPN also increased tph2 mRNA in the caudal DRN. DPN- and E-treated rats displayed a more active stress-coping behavior in the forced-swim test (FST). No behavioral differences were found in the OF or on the EPM.
These data indicate that ERbeta acts at the level of the rat DRN to modulate tph2 mRNA expression and thereby influence 5-HT synthesis in DRN subregions. Our results also suggest that local activation of ERbeta neurons in the DRN may be sufficient to decrease despair-like behavior, but not anxiolytic behaviors.
Major depressive disorder (MDD) affects about 17% of Americans (Kessler et al., 1994; Williams et al., 2007), and is a complex, heterogeneous disease (Winokur, 1997; Ellard, 2001; Weissman, 2002). A deficiency in 5-HT neurotransmission is a leading hypothesis regarding the development and pathophysiology of this disease (Owens and Nemeroff, 1994; Arango et al., 2002; Perlis et al., 2002; Lesch, 2004). An influence of gonadal hormones on the etiology of depression is indicated based on the incidence, duration, severity and rate of reoccurrence of depressive disorders, which are over twice as high in women compared to men (Earls, 1987; Angold and Worthman, 1993; Weissman et al., 1993; Kornstein et al., 1995). Moreover, women also tend to respond differently than men to common antidepressant treatments, such as selective serotonin-reuptake inhibitors (Kornstein, 1997; Gorman, 2006). These observations, together with animal studies reporting sex differences in the regulation of emotion (Steenbergen et al., 1990; Caldarone et al., 2003; Shors and Leuner, 2003; Toufexis, 2007) and interactions between estrogen and the 5-HT system (for review see (Amin et al., 2005)) suggest an involvement of estrogen receptors (ERs) in the etiology of MDD.
In animal models, estrogen can exert both anxiolytic, and anxiogenic actions depending on the behavioral context (Koss et al., 2004; Hiroi and Neumaier, 2006). This ambiguity may be explained by the existence of two different estrogen receptor systems, ERalpha and ERbeta. The endogenous ligand estradiol binds to and activates both receptor types with similar affinity (Kuiper et al., 1997). However, ERalpha-selective agonists are anxiogenic, while ERbeta-selective compounds are anxiolytic and anti-depressive (Walf et al., 2004; Lund et al., 2005; Weiser et al., 2009). Moreover, in flinders-sensitive rats, a strain selectively bred for depressive-like behaviors, ERbeta agonists reduce the animals’ passive floating and immobility behavior (Overstreet et al., 2006) during the forced swim test (FST), a test established to assess despair-like behavior in rodents (Porsolt et al., 1977). This rat strain also displays abnormal levels of 5-HT(2A) receptor mRNA in the perirhinal cortex, piriform cortex, medial anterodorsal amygdala and in the hippocampus, a phenotype that is reversed by estradiol-treatment (Osterlund et al., 1999).
The brainstem dorsal raphe nuclei (DRN) constitute the primary 5-HT system of the brain. Distinct DRN subdivisions give rise to axons that innervate most forebrain areas, including areas crucial for the regulation of emotion and stress-coping behavior, such as the amygdala and the paraventricular nucleus of the hypothalamus (Imai et al., 1986; Petrov et al., 1992). Other subregions of the DRN send projections to motivational areas like the prefrontal cortex (Lowry, 2002; Abrams et al., 2004), while axons from the caudal DRN target limbic structures, such as the hippocampus, the entorhinal cortex and the septum (Kohler and Steinbusch, 1982).
Tryptophan-hydroxylase 2 (TPH2), the brain-specific version of TPH (Walther et al., 2003; Zhang et al., 2004), catalyses the rate-limiting step of 5-HT synthesis. Disruption or dysfunction of the tph2 gene is strongly correlated with affective disorders (Zill et al., 2004; Zhang et al., 2005; Haghighi et al., 2008), and abnormal tph2 mRNA expression may be responsible for many of those pathologies.
Previous studies have demonstrated that ERbeta is robustly expressed within the DRN of rodents (Shughrue et al., 1997a; Lu et al., 2001; Mitra et al., 2003; Nomura et al., 2005; Vanderhorst et al., 2005), primates (Gundlah et al., 2000; Gundlah et al., 2001) and guinea pigs (Lu et al., 1999), whereas ERalpha is only expressed to a small extent in the DRN of these species. Given the importance of the 5-HT system in anxiety- and depressive disorders together with the anxiolytic nature of ERbeta and its presence in the raphe complex, we speculate that ERbeta activation may regulate gene expression in 5-HT neurons.
Results from earlier studies suggest that ERbeta activation can regulate transcription and neurotransmission in the brainstem. Alves et al. (2000) reported estradiol-induced progestin receptor expression in the DRN of ERalpha null mice, suggesting a role for ERbeta. Also, phytoestrogens that selectively bind ERbeta have been shown to improve mood and 5-HT neurotransmission in the cynomolgus monkey (Shively et al., 2003). In turn, ERbeta null mice are characterized by increased anxiety in conjunction with lower 5-HT content and decreased tph mRNA in the DRN (Imwalle et al., 2005; Nomura et al., 2005). In macaques, estradiol and a combination of the ovarian steroids progesterone and estradiol each caused an elevation of tph2 mRNA in the DRN (Sanchez et al., 2005). In rats, a recent study showed that estradiol itself increases tph2 mRNA expression specifically in those DRN subregions that are associated with attenuated anxiety (Hiroi et al., 2006).
These findings support the hypothesis that ERbeta activation in the DRN may be sufficient to alter behavioral parameters and 5-HT-neuronal gene expression in the DRN. Therefore, we examined the effects of chronic systemic versus local, intracerebral delivery of the ERbeta agonist, diarylpropionitrile (DPN), in OVX female rats on anxiety-and despair-like behaviors and on tph2 mRNA expression in the DRN.
All animal surgeries, behavioral tests and experimental protocols followed NIH guidelines and were approved by the Animal Care and Use Committee (ACUC) at Colorado State University. Young adult female Sprague-Dawley rats (200–250 g body weight, Charles River Laboratories, Wilmington, MA) were fed a phytoestrogen-free diet (modified AIN-93G with corn oil substituted for soy oil; Dyets, Philadelphia, PA), double housed, and kept under standard laboratory conditions (12:12 h light-dark cycle, lights on at 0600 h, 22 °C, 60 % humidity, and ad libitum access to water and food). The animals were handled and their weight monitored every other day for the duration of both experiments. Surgical procedures were performed under isoflurane- (for OVX) or ketamine-anaesthesia (100mg/kg of 93% ketamine / 5% xylazine / 2% acepromazine for stereotaxic surgeries).
One week after arrival, all rats underwent bilateral OVX to remove circulating gonadal steroids. This procedure also ensured a constant level of ERbeta expression within the brain because receptor levels are regulated by hormones (Suzuki and Handa, 2005). Chronic 8-day systemic (s.c.) or local (intracerebral) treatment with ER ligands began one week after OVX. All animals were double housed with a partner of the same treatment group.
Rats were injected s.c. with the ERbeta agonist, diarylpropionitrile (DPN, 2 mg/kg; n=8) or vehicle (27% hydroxypropyl-beta-cyclodextrin from CTD Inc., High Springs, FL; n=8) in phosphate buffered saline once per day at 0600 h. DPN was synthesized de novo following an established protocol (Lund et al., 2005). In the morning of day 6, rats were tested for anxiety-like behavior in the OF, and on day 7 on the EPM. All animals were killed by decapitation on day 8 between 1000 and 1200 h under basal, non-stress conditions. This occurred 4 h after the last DPN injection to avoid potential acute effects of agonist treatment. For all animals, brains were removed from the skull, fresh-frozen in pre-cooled 2-methylbutane (−40° C) and stored at −80° C until sectioning.
Three groups of rats were stereotaxically implanted bilaterally with wax pellets (2.00 mm long, 0.25 mm in diameter) flanking the DRN at coordinates: 8.0 mm posterior to bregma, ± 1.5 mm lateral, 5.5 mm deep, at a 7° angle. Each pellet contained either 0.5 μM DPN (n=10), 0.5 μM 17-beta-estradiol (Sigma, St. Louis, MO; E, n=10) or beeswax only (VWR International, Bristol, CT; vehicle control, n=10). The dosage of DPN and E and the procedure for stereotaxic wax pellet implantation was based on previous studies by Lund et al. (2006). A second control group of OVX animals remained unoperated (C, n=7). All animals were tested in the OF, EPM and FST on days 5, 6 and 7 of treatment. Animals were returned to their home cage after each test. All animals were killed on day 8 between 1000 and 1200 h, their brains removed, fresh-frozen and stored at −80° C until sectioning.
To measure anxiety-like behaviors, all rats were tested in the OF and on the EPM on two consecutive days between 1000 and 1200 h for 5 min each. In the OF, the following parameters were scored: locomotor activity (total of square line crossings), number of rears at walls, time spent in center squares, time spent in outer squares, time spent grooming, and number of fecal boli. On the EPM, the latency until first open arm entry, the time spent in the open and closed arms, the number of closed and open arm entries (locomotor activity), the time spent grooming, and the number of fecal boli were recorded as described in Lund et al. (2005)
To measure despair-like behavior, rats were tested in the FST (Porsolt et al., 2001) for 5 min on day 7, using 25 °C tap water. In the afternoon of the previous day, all rats were trained for the FST for 3 min each, without recording their behavior. The time spent paddling (normal stress-coping behavior: slow-pace front and hind leg movements), the time spent struggling (active behavior: high-pace front leg paddling and strong hind leg strokes), the time spent floating (passive, despair-like behavior: minimal leg movements; stiff, floating body posture) and the number of dives (active behavior) was recorded.
For the production of a riboprobe specific for tph2 mRNA, total RNA was isolated from microdissected DRN tissue samples (Palkovits et al., 1975), following the protocol established by Chomczynski and Sacchi (1987). 1 μg RNA was reverse transcribed into 10 total cDNA with the reverse transcriptase MMLV-RT (Invitrogen, Carlsbad, CA) at 37° C for 50 min, using 1 μl oligo dT primers, dNTPs (100 mM each), 1st strand buffer (100 mM Tris–Cl–900 mM KCl–1 mM MgCl) and 2.5 mM DTT. Subsequently, a 583 bp fragment of tph2 cDNA was amplified by RT-PCR (forward primer: 5′-GGG GTG TTG TGT TTC GGG-3′, reverse primer: 5′-GTG GTG ATT AGG CAT TCC-3′). PCR conditions were: 45 s denaturation at 95° C, 45 s annealing at 55° C, 45 s elongation at 72° C, and a final 7-min elongation step at 72° C after 35 cycles. The 50 μl PCR reaction volume contained 1.5 mM Mg2+, 0.2 mM dNTPs, 0.2 μM forward and reverse primer, 50 ng template cDNA, and 1.0 unit Taq DNA polymerase (Eppendorf, Westbury, NY). After gel-purification (Qiagen, Valencia, CA), the PCR product was subcloned into the linearized 4.0 kb TOPO-vector pCR®II (Invitrogen, Carlsbad, CA), and amplified in chemically competent TOP10 bacterial cells (Invitrogen). Successful clones were verified via sequencing (Retrogen, San Diego, CA). Antisense and sense (control) tph2 cRNAs were transcribed from the plasmid in the presence of [35S]-UTP, following linearization with restriction enzymes, BlpI or XbaI respectively. As confirmed in two hybridization test runs, the antisense probe specifically detected tph2 mRNA in the dorsal and median raphe nuclei, but did not hybridize with tph1 mRNA in the pineal gland (Patel et al., 2004; Malek et al., 2005). There was no hybridization above background seen when using the sense-directed control probe.
A series of coronal 16-μm brainstem sections between bregma −6.5 mm to −9.5 mm (Paxinos and Watson, 1998) was cut at −20° C using a cryostat (Leica, Wetzlar, Germany) and thaw-mounted onto positively charged slides (Superfrost Plus, VWR Scientific, West Chester, PA). All sections were stored at −80° C until assayed. For in situ hybridization, tissue sections were thawed at room temperature, fixed within 10% paraformaldehyde, acetylated with 0.25% acetic anhydride, dehydrated in a graded series of alcohols, and air-dried. Next, sections were incubated with hybridization solution (50% formamide, 0.60 M NaCl, 0.02 M Tris, 0.01 M EDTA, 10% dextran sulfate, 2 M Denhart’s solution, 50 mM dithiothreitol, 0.2% SDS, 100 mg/ml salmon testis DNA, 500 mg/ml total yeast RNA, and 50 mg/ml yeast transfer RNA), containing radiolabeled cRNA at a concentration of 2 × 107 cpm/ml, in humidified chambers at 60° C overnight. After hybridization, slides were rinsed in 2 × SSC. Non-hybridized RNA was digested in a 30 mg/ml RNase A solution for 30 min at 37° C. A final high stringency wash (0.1 × SSC, room temperature) preceded dehydration in graded alcohols. Hybridization was first examined by opposing slides to a 35S-sensitive Biomax MR film (Kodak, Rochester, NY) for 14 hours. Subsequently, hybridization was detected using photographic emulsion-coated slide autoradiography (NTB-3; Kodak). After a 2 day-incubation at 4° C in the dark, all slides were developed (Kodak D-19) and counterstained with cresyl violet.
We assumed that the compounds used in the present study successfully diffused into all rostro-caudal and medial-to-lateral subregions of the DRN if both pellets were placed within a maximal radius of 0.5 mm from the center of a coronal DRN section. This criterion was defined empirically by Lund et al. (2006) who found the diffusion of [3H]-labeled E to be confined within a 0.5 mm area surrounding the wax pellet. Based on this anatomical criterion, one individual of the vehicle-group had to be excluded from data analysis. Each section containing the DRN (12 sections per animal) was evaluated using bright-field microscopy and the center location of each pellet was estimated and mapped using a rat brain atlas (Paxinos and Watson, 1998).
Cytoplasmic detection of tph2 mRNA in individual 5-HT neurons was verified via photomicrographs. Counterstaining with cresyl violet allowed for distinction between silvergrain-labeled tph2 mRNA localized around purple-labeled nuclei. Dark-field images were captured by a Zeiss AxioCam HR camera on an Axioplan 2 microscope controlled by Axiovision, version 3.1, software. Three rostral, mid and caudal sections per animal were atlas matched (Paxinos and Watson, 1998) and used for analysis via ImageJ software (version 1.31). Matched dark-field images were inverted in order to cause silvergrains to appear as dark pixels in the inverted picture. Matrices in approximate shape of the DRN subregions of interest were then utilized to assess tph2 mRNA expression within the DRN. The density of black pixels was measured and expressed as arbitrary density units [AdU] for each subregion. After subtraction of background activity (determined in an adjacent area devoid of labeling), six (for the lateral DRN) or three values each (for the dorso-rostral, ventro-rostral, dorso-mid, ventro-mid, dorso-caudal, ventro-caudal DRN) were averaged per animal to obtain an individual value for statistical analysis.
All data are expressed as the mean ± standard error of the mean (SEM). For studies with only two treatment groups, Student’s t- test was used for pair-wise data comparison. Results of all studies with more than two treatment groups were analyzed by one-way ANOVA (factor treatment) followed by Tukey’s post hoc test where appropriate, using SPSS 12.0 for Windows software. Results were considered significantly different when p < 0.05.
Vehicle animals gained 25.16 ± 2.79 g, DPN-treated animals 31.17 ± 3.90 g from day 1 to day 8 of treatment. There were no significant group differences in weight gain.
Animals were tested for anxiety-like behaviors in the open field (OF) and on the elevated plus maze (EPM). In the OF (Fig. 1A), DPN-treated animals displayed more rears at the walls than their vehicle-treated counterparts (p < 0.01). The total number of square line crossings (a measure of overall activity) did not differ between the two groups, suggesting that there was no overall effect on activity. The total time spent in inner or outer squares, the time spent grooming and the number of fecal boli were not significantly different between the two treatment groups.
On the EPM, DPN treatment caused the rats to enter the open arms sooner (p < 0.01), more often (p < 0.05), and stay on the open arms longer (p < 0.05) than vehicle controls (Fig. 1B). The number of entries into the closed arms did not differ between the two treatment groups, again indicating that DPN did not increase the rats’ overall activity or locomotor behavior. None of the other parameters recorded (time grooming, time in closed arms, fecal boli) revealed significant group differences.
To test the hypothesis that systemic ERbeta activation may upregulate tph2 mRNA expression, in situ hybridization was performed, and the density of tph2 mRNA expression was measured in all subregions of the DRN. Fig. 2 displays representative dark-field photomicrographs of tph2 mRNA hybridization in the DRN of systemically treated rats. Compared to vehicle-treated animals, daily s.c. administration of DPN significantly enhanced tph2 mRNA levels in the dorso-mid (p < 0.05), the dorso- caudal (p < 0.05) and the ventro-caudal (p < 0.05) DRN (Fig. 3A). Accordingly, total tph2 mRNA levels in the entire caudal DRN were almost doubled in the DPN group (p < 0.01; Fig. 3B), compared to vehicle controls. Tph2 mRNA expression in the rostral DRN was not elevated by DPN. Fig. 3C illustrates in principle the matrix-based, digital analysis of inverted dark-field images that was used for all photomicrographs.
The average weight gain during experiment 2 was 28.57 ± 1.87 g for the unoperated control group, 25.00 ± 2.21 g for the vehicle control animals, 25.80 ± 3.30 g for E-treated animals, and 30.50 ± 2.25 g in the DPN group. No significant differences in weight gain were found between any of the treatment groups.
Fig. 4A shows a representative bright-field photomicrograph of the bilaterally implanted wax pellets placed to flank the DRN. To confirm the cellular silvergrain-labeling of neurons containing tph2 mRNA, brain tissue was counterstained with cresyl violet at an intensity sufficient to label the nucleus, but light enough to not alter or interfere with the identification of silvergrains (Fig. 4B). A schematic of actual bilateral wax pellet localization in the brainstem of all animals in experiment 2 is depicted in Fig. 4C. If the center of any of the two pellets was found to be more than 0.5 mm away from the estimated center of the DRN, the respective animal was excluded from all data analysis. This was true for one vehicle-treated rat (black dots in Fig. 4C), which reduced the size of this group from 10 to 9 animals.
All animals were tested for anxiety-related behavior in the OF and on the EPM. Local, DRN-targeted delivery of neither DPN nor E altered any of the parameters measured in the OF and the EPM, compared to unoperated and vehicle controls (Fig. 5, A and B). In the FST, rats that were implanted with bilateral, DRN-flanking E- or DPN-pellets displayed a more active stress-coping strategy by spending more time actively struggling in the water than did vehicle controls (Fig. 5C; F3,32=4.628; p < 0.05). Animals of all treatment groups spent about the same time paddling (neutral stress-coping behavior). Neither E- nor DPN-treatment significantly altered the time spent floating, compared to controls. However, DPN-treated animals spent more time passively floating than their E-treated counterparts (Fig. 5C; F3,32=6.040; p<0.05). The number of dives did not differ between any of the groups.
Animals from the unoperated control group did not differ from the regular vehicle control group in any of the behavioral paradigms, indicating that there was no effect of brain surgery on anxiety-like behaviors.
To investigate whether site-specific activation of ERs through local delivery of DPN or E itself is sufficient to enhance the expression of tph2 mRNA, in situ hybridization was performed. Fig. 6 displays representative dark-field photomicrographs of tph2 mRNA hybridization in the brainstem of locally DPN-, E-, or vehicle-treated rats. In the dorso-mid DRN, local DPN-administration elevated tph2 mRNA expression to approximately 2.5-fold of the level seen in controls (Fig 7A; F3,32=18.197; both p < 0.01), resulting in a more than 1.5-fold overall increase of tph2 mRNA expression for the entire mid-DRN, compared to vehicle and unoperated controls (Fig. 7A and B; F3,32=9.096; both p < 0.01). Regarding its effect on tph2 mRNA in the mid DRN, local DPN treatment also resulted in a stronger response than that of E (Fig. 7A and B; p < 0.01). In both the dorso- and ventro-caudal DRN, a 2- to 3-fold increase in tph2 mRNA expression (dorso-caudal: F3,32=13.720; p < 0.01; ventro-caudal: F3,32=14.962; p < 0.05) was seen when DPN-treated animals were compared to unoperated and vehicle-treated controls. In summary, local DPN-treatment more than doubled the expression of tph2 mRNA in the entire caudal DRN, compared to vehicle (Fig. 7A and B; F3,32=6.040; p < 0.01) and unoperated controls (p < 0.05). Tph2 mRNA expression in the rostral DRN was not elevated by DPN.
Treatment with estradiol caused similar, but less intense, effects on tph2 mRNA expression compared to DPN. Local E-treatment caused tph2 mRNA levels in the dorso-mid DRN to be about 1.8-fold higher than in either control group (Fig. 7A; F3,32=10.007; p < 0.05), resulting in an overall 1.5-fold increase of tph2 mRNA in the entire mid DRN (Fig. 7A and B; F3,32=7.902; p < 0.05), compared to both control groups. In the dorso-caudal subregion of the DRN, local delivery of E caused a significant increase in tph2 mRNA compared to the vehicle control group (Fig. 7A; F3,32=14.665; p < 0.05), in the ventro-caudal part compared to the unoperated control group (Fig. 7A; F3,32=9.192; p < 0.05). Overall expression of tph2 mRNA in the caudal DRN was almost doubled by local E-treatment compared to vehicle controls (Fig. 7A and B; F3,32=8.360; p < 0.05).
Tph2 mRNA levels did not differ for any subregion between the vehicle- and the unoperated control group, ensuring that the implantation of wax pellets adjacent to the DRN did not alter tph2 gene expression.
The results of our studies show that both systemic and local activation of ERbeta in or around the DRN increased the expression of tph2 mRNA in a subregion-dependent manner. Estrogen treatment caused similar overall effects, but to a lesser extent than DPN. Furthermore, animals treated locally with DPN or E showed decreased despair-like behavior. However, only systemic delivery of DPN decreased anxiety-like behavior, while local administration of DPN failed to have the same effect.
The systemic delivery of DPN to OVX rats reduced several anxiety-related behaviors, confirming the anxiolytic nature of ERbeta (Krezel et al., 2001; Lund et al., 2005; Rocha et al., 2005; Weiser et al., 2009). In contrast, local DPN- or estradiol-treatment of the raphe nuclei failed to decrease anxiety-like behavior, indicating that estradiol’s action in regulating anxiety-like behaviors may primarily involve other brain areas, such as the hypothalamic PVN (Herman et al., 2002; Donner et al., 2007; Blume et al., 2008; Neumann, 2008), the lateral septum (Henry et al., 2006), the amygdala (Bosch et al., 2007), or the bed nucleus of the stria terminalis (Davis et al., 1997; Walker et al., 2003). Interestingly, our studies demonstrate that local delivery of DPN or estradiol is sufficient to decrease despair-like behaviors in the FST. It remains unclear why DPN-treated rats spent less time with passive floating behavior than estradiol-treated animals. Yet, in accordance with studies from other research groups, both estradiol and DPN increased the time rats spent actively struggling. While recent studies already revealed an antidepressant function of ERbeta (Walf et al., 2004; Rocha et al., 2005; Hughes et al., 2008), our experiments indicate a site of action for the observed effect. Thus, high anxiety and despair-like behavior may be closely related to the phenotype of depression (Chaby et al., 1993; Leibbrand et al., 1999; Farabaugh et al., 2005; Godart et al., 2006; Mittal et al., 2006), but the involvement of ERbeta may be through two distinct neuroanatomical regions. Further studies will be required to determine how ERbeta-mediated elevations of tph2 mRNA expression in the mid-dorsal and caudal DRN are coupled to the attenuation of despair-like behavior.
Not only the route and site of drug administration (systemic versus local), but also the different time course of delivery (once daily s.c. versus constant intracerebral) may have contributed to the discrepancy regarding anxiety-related behaviors between our two experiments. The concentration of DPN in the plasma of male rats that received a 1mg/kg s.c. injection of DPN has been shown to peak after 1 hour and decrease rapidly to undetectable levels within 3 hours post injection (Patisaul et al., 2009). However, this doesn’t mean that the biological activity of DPN decreases within the same timeframe. Lund et al. (2005) used an in vitro binding approach and differential centrifugation for separation of bound and unbound receptors to estimate that DPN occupies neural ERs with a half life of about 8 hours. These and other studies indicate that DPN, phytoestrogens, and estradiol itself are likely to be sequestered by ERs in reproductive tissues and in the brain, and are able to exert physiological effects for many hours after incorporation (Blaustein et al., 1979; Dehennin et al., 1982; Morton et al., 1997; Walf et al., 2004; Walf and Frye, 2005). Although animals in our experiment weren’t sacrificed until 4 hours after the last s.c. injection of DPN, it is therefore not unlikely that the behavioral discrepancies observed were due to delayed acute effects of the last DPN injection, compared to a constant chronic exposure to the compound diffusing from locally implanted wax pellets. Chronic wax-mediated, application of DPN for 7 days around the paraventricular nucleus has previously proven sufficient to inhibit the neuronal and corticosterone response of the hypothalamo-pituitary adrenal axis to an acute stressor (Lund et al., 2006). The effects of local DPN-treatment on cellular parameters (tph2 mRNA) in our study further strengthen the assumption that DPN is stable and biologically active for a long time when dissolved and administered in beeswax. However, it would be of further interest to identify the biological stability and metabolic fate of DPN following administration.
The observed DPN- or estrogen-induced increase in tph2 mRNA expression was mainly restricted to the dorso-mid and the caudal DRN. These findings are consistent with observations by Hiroi et al. (2006), who reported elevated tph2 mRNA expression specifically in the dorso- and ventro-caudal DRN following systemic estrogen treatment in female OVX rats. Furthermore, recent studies in rats and mice indicate that depression-related behavioral paradigms (Keeney and Hogg, 1999; Becker et al., 2007) like social defeat (Gardner et al., 2005) or inescapable stress (Grahn et al., 1999; Amat et al., 2005) selectively activate the dorsal and caudal parts of the DRN. These subregions correspondingly give rise to projections targeting forebrain areas involved in the control of emotional behavior (Lowry et al., 2005; Lowry et al., 2008). The mid-dorsal DRN, for instance, sends out collateral projections to emotion- and stress-related brain areas, that could, for instance, simultaneously modulate the hypothalamic PVN and the basolateral or central nucleus of the amygdala (Lowry, 2002; Hale et al., 2008b; Hale et al., 2008a). 5-HT axons from the caudal DRN target limbic structures like the hippocampus, the entorhinal cortex and the septum (Kohler et al., 1982; Kohler and Steinbusch, 1982), indicating that an alteration in tph2 expression and 5-HT neurotransmission by estrogens may improve memory and learning deficits that are associated with depression (Shors et al., 1998; Burriss et al., 2008; Liu et al., 2008). Within the entire caudal DRN it is the dorso-caudal subdivision that has been suggested to play a crucial role in changes associated with affective disorders (Commons et al., 2003). In clinical studies of drug-free, depressed suicide victims, the dorso-caudal DRN subregions also exhibited elevated TPH2 protein and tph2 mRNA expression (Bonkale et al., 2006; Bach-Mizrachi et al., 2008). However, a pathological increase in tph2 expression may explain this apparent paradox. Pathologically elevated tph2 mRNA and protein may reflect a compensatory feedback response to low overall 5-HT concentrations in the brain of depressed patients (Mann et al., 1989; Owens and Nemeroff, 1994; Placidi et al., 2001). Also, most of the brains assessed in these clinical studies were derived from male individuals, not females. Consequently, more detailed studies quantifying TPH2 protein and local 5-HT release and turnover within the DRN itself (autoregulation) and in target areas of the DRN in both male and female animal models of anxiety and depression are required to answer this question.
Overall, local estradiol treatment had a similar, but less intense effect on tph2 mRNA expression and on despair-like behavior than the selective ERbeta agonist. This difference between E and DPN could be explained by the non-selective action of estradiol on both ERalpha and ERbeta. Our previous data and those of others suggest that ERalpha and ERbeta have opposing actions on stress related behaviors (Liu et al., 2002; Lund et al., 2005; Toufexis et al., 2007; Weiser et al., 2009). While increased ERalpha mRNA and single nucleotide polymorphisms (SNPs) in the gene coding for ERalpha are associated with mental illness, specifically with depression (Perlman et al., 2005; Mill et al., 2008), ERbeta-mediated actions have been found to exert anxiolytic and antidepressant effects in various animal models (Imwalle et al., 2005; Lund et al., 2005; Rocha et al., 2005). Since estradiol can bind to both receptor types with equal affinity (Kuiper et al., 1997; Lund et al., 2005), the possibility exists that it could activate two functionally opposing mechanisms, both ultimately balancing tph2 mRNA expression and 5-HT-dependent behaviors.
At present, the exact patterns for ERbeta expression in DRN 5-HT neurons are controversial. Lu et al. (2001) demonstrated that 5-HT neurons of the DRN of rats contain ERbeta, whereas Sheng et al. (2004) were unable to identify ERbeta in 5-HT neurons. In mice, Nomura et al. (2005) revealed that ERbeta, but not ERalpha is located within 5-HT neurons. ERbeta has also been shown in 5-HT neurons of the guinea pig (Lu et al., 1999) and rhesus monkey (Gundlah et al., 2001). Furthermore, ERbeta2 (Chung et al., 2007), a novel splice variant carrying an 18-amino acid insert between the fifth and the sixth exon in the ligand-binding domain of ERbeta, has been shown in the DRN of female rats. In contrast, ERalpha may only be expressed in non-5-HT-, but possibly GABAergic interneurons (Hart et al., 2001; Su et al., 2001), placing ERalpha in a position to interfere with the negative feedback regulation of 5-HT-neuronal function (Haddjeri et al., 2000; Liu et al., 2000). Differences in the expression of ERalpha versus ERbeta in the midbrain of rats (Shughrue et al., 1997a; Shughrue et al., 1997b; Lu et al., 2001), mice (Nomura et al., 2005; Vanderhorst et al., 2005), guinea pigs (Lu et al., 1999; Warembourg and Leroy, 2004), and cats (VanderHorst et al., 1998) suggest that species differences may exist in the modulation of the 5-HT system by gonadal steroids. In non-human primates, ERbeta, but not ERalpha, seems to be the predominant ER expressed in raphe 5-HT neurons (Gundlah et al., 2000; Gundlah et al., 2001; Vanderhorst et al., 2009). While ER expression in the human DRN remains to be fully described, estrogens would be expected to exert mainly anxiolytic and anti-depressive actions in humans if our expression profile of ERbeta in the brain resembles the pattern found in other primates.
The molecular mechanisms by which ERbeta and ERalpha may directly or indirectly modulate tph2 gene expression are still unknown. Although most ER-induced changes in gene transcription are due to classic effects of the steroid receptors acting as nuclear transcription factors, it is possible that other mechanisms may be present, particularly given the recent studies showing rapid, membrane mediated mechanisms of ER action (Cato et al., 2002; Mhyre and Dorsa, 2006; Levin, 2008).
In conclusion, our results show that chronic, local activation of ERbeta alters tph2 mRNA expression in the DRN in a subregion-dependent manner, and, at the same time, facilitates active stress-coping behavior. Interactions between ERbeta and 5-HT neurons of the DRN may be key regulators of anti-depressive behavior, whereas other brain circuits seem to be necessary for ERbeta to exert its anxiolytic actions. Our observations also raise the question whether physiological changes in circulating estradiol can differentially influence behaviors in women across the menstrual cycle. One possibility is that an altered ratio of ERalpha versus ERbeta expression or a disruption of normal ER-regulation of tph2 expression in the midbrain might contribute to mood disorders like premenstrual syndrome (Rubinow, 1992; Arpels, 1996; Schmidt et al., 1998) or premenstrual dysphoric disorder (Gorman, 2006). These data demonstrate the potential of ERbeta as a pharmaceutical target for treating affective disorders. The future development of an ERbeta agonist for clinical use that facilitates 5-HT function and emotional stability in menopausal women without the risk of breast- or gynecological cancer associated with ERalpha-mediated actions (Chen et al., 2008) would be highly beneficial.
The authors are grateful to Dr. Michael Weiser for de novo synthesis of diarylpropionitrile (DPN) and for help with stereotaxic surgeries.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.