Modifying 5-HT1A Receptor Gene Expression as a New Target for Antidepressant Therapy.

Major depression is the most common form of mental illness, and is treated with antidepressant compounds that increase serotonin (5-HT) neurotransmission. Increased 5-HT1A autoreceptor levels in the raphe nuclei act as a "brake" to inhibit the 5-HT system, leading to depression and resistance to antidepressants. Several 5-HT1A receptor agonists (buspirone, flesinoxan, ipsapirone) that preferentially desensitize 5-HT1A autoreceptors have been tested for augmentation of antidepressant drugs with mixed results. One explanation could be the presence of the C(-1019)G 5-HT1A promoter polymorphism that prevents gene repression of the 5-HT1A autoreceptor. Furthermore, down-regulation of 5-HT1A autoreceptor expression, not simply desensitization of receptor signaling, appears to be required to enhance and accelerate antidepressant action. The current review focuses on the transcriptional regulators of 5-HT1A autoreceptor expression, their roles in permitting response to 5-HT1A-targeted treatments and their potential as targets for new antidepressant compounds for treatment-resistant depression.

Introduction

Depression and serotonin

Major depression is the most prevalent form of mental illness, twice as frequent in women as in men (Doris et al., 1999; Fava and Kendler, 2000), and is predicted to increase from fourth to second (first in high income countries) highest global burden of disease by 2030 (Ustun et al., 2004; Mathers and Loncar, 2006). There is a need for improved antidepressant treatments since only 30% of patients remit with current strategies (Trivedi et al., 2006a,b, 2008), while 15% attempt suicide (Mann et al., 2001; Mann, 2005). Although other neurotransmitters (e.g., noradrenaline, dopamine, glutamate, neurotrophins) are indirectly involved in depression (Blier, 2003; Duman, 2004; Nestler and Carlezon Jr., 2006; Skolnick et al., 2009), multiple lines of evidence implicate reduced 5-HT neurotransmission as a primary defect in depression (Millan, 2004; Wong et al., 2005; Tremblay and Blier, 2006; Jans et al., 2007; aan het Rot et al., 2009). Virtually all antidepressant treatments, including serotonin reuptake inhibitors (SSRIs), increase 5-HT neurotransmission, either directly or indirectly (Blier, 2003; Berton and Nestler, 2006; Savitz et al., 2009).

5-HT1A receptors and the 5-HT system

The brain serotonin (5-HT) system originates from neurons of the raphe nuclei that express tryptophan hydroxylase 2 (TPH2), the rate-limiting enzyme for 5-HT synthesis (Walther et al., 2003). These neurons project widely throughout the brain (Hornung, 2003) to regulate many physiological functions, including sleep, mood and stress reactivity and are implicated in mental illnesses, including MDD and anxiety (Barnes and Sharp, 1999; Young and Leyton, 2002; Gordon and Hen, 2004; Lanfumey et al., 2008; Savitz et al., 2009). Among the 17 human 5-HT receptor genes, we have focused on the 5-HT1A receptor since it is abundant in corticolimbic regions that are implicated in mood and emotion (Albert et al., 1996; Albert and Lemonde, 2004). The 5-HT1A receptor also functions presynaptically as the major somatodendritic autoreceptor on 5-HT neurons (Sotelo et al., 1990; Riad et al., 2000) where it acts as a “brake” to inhibit the activity of the entire 5-HT system (Hjorth et al., 1996; Richer et al., 2002; Bortolozzi et al., 2004) (Figure 1). Hence, the mechanisms that regulate 5-HT1A autoreceptor levels set the tone of the 5-HT system.

Figure 1

Regulation of 5-HT1A receptors in depression. Shown figuratively are 5-HT1A-expressing presynaptic raphe neuron and post-synaptic prefrontal cortical neuron. The 5-HT1A gene and its regulators (Deaf1, Freud-1, Freud-2, REST, Hes) are shown; 5-HT1A receptors (circle/triangle) and serotonin (triangles) are shown to indicate expected changes in their levels in depression. In depression, 5-HT1A receptor RNA and protein levels are increased presynaptically, leading to inhibition of 5-HT neurons (red-) and reduced firing. Paradoxically, the presynaptic levels of repressors REST and Deaf1 are increased, and there is also a trend for increase in Freud-1, perhaps to compensate for 5-HT1A overexpression due to the G(−1019) allele, for example. Post-synaptically, 5-HT1A receptors appear reduced, and levels of Deaf1 are also lower, although Freud-1 and Freud-2 (trend) increase. See text for details and references.

5-HT1A Receptors in Depression

5-HT1A receptors in animal models of depression

Despite their limited validity to model human depression, animal behavioral studies have provided valuable insights into the biological function of 5-HT1A receptors in depression and anxiety. Importantly, 5-HT1A-null mice display increased anxiety behavior (Heisler et al., 1998; Parks et al., 1998; Ramboz et al., 1998) and are unresponsive to SSRI treatment (Santarelli et al., 2003), while early developmental overexpression of the 5-HT1A receptor decreases anxiety (Kusserow et al., 2004). 5-HT1A receptor expression in early postnatal forebrain of 5-HT1A−/− mice rescues the anxiety phenotype, suggesting that post-synaptic 5-HT1A receptors are critical in the development of the anxiety phenotype (Gross et al., 2002). Conversely, mice with a 30% increase in 5-HT1A autoreceptors display reduced 5-HT neuron firing, reduced 5-HT release, increased depression behavior, but no change in anxiety behavior (Richardson-Jones et al., 2010). Hence, while reduction in post-synaptic 5-HT1A receptors is implicated in anxiety, increased levels of presynaptic 5-HT1A autoreceptors inhibit 5-HT neurotransmission and leads to depression (Figure 1).

5-HT1A receptors in clinical depression

In human depression, there remain conflicting findings concerning changes in 5-HT1A binding potential (Savitz et al., 2009). Most studies find reductions in post-synaptic 5-HT1A receptors in subsets of prefrontal and temporal cortical regions in depression (Sargent et al., 2000; Bhagwagar et al., 2004; Shively et al., 2006; Moses-Kolko et al., 2008) and anxiety (Tauscher et al., 2001; Neumeister et al., 2004; Sullivan et al., 2005; Lanzenberger et al., 2007; Nash et al., 2008; Akimova et al., 2009). In contrast, 5-HT1A autoreceptors are increased in depressed patients (Parsey et al., 2006a,b) and in post-mortem raphe tissue from depressed suicide victims (Stockmeier et al., 1998; Boldrini et al., 2008). However, decreases in raphe 5-HT1A binding have also been reported, but may reflect a reduction in 5-HT neurons (Drevets et al., 2007). These region-specific alterations in 5-HT1A receptor levels in depression are consistent with a global reduction in 5-HT neurotransmission, since increased 5-HT1A autoreceptors reduce 5-HT neuronal activity, while reduced post-synaptic 5-HT1A receptors reduce behavioral response to 5-HT (Albert and Lemonde, 2004).

5-HT1A receptors in antidepressant response

Although SSRIs rapidly enter the brain and block 5-HT reuptake, chronic 3-week treatment is required for clinical response, due to recurrent inhibition of raphe activity by 5-HT1A autoreceptors (Figure 1). During this period, 5-HT1A autoreceptors desensitize (Le Poul et al., 1995; Mochizuki et al., 2002; Shen et al., 2002; Elena Castro et al., 2003), leading to reduced 5-HT1A RNA and autoreceptor levels, particularly in animal models of depression and depressed subjects (Welner et al., 1989; Fanelli and McMonagle-Strucko, 1992; Sibug et al., 1998; Le Poul et al., 2000; Meltzer et al., 2004; Rabiner et al., 2004; Berney et al., 2008), resulting in restoration of raphe firing activity and 5-HT release. Chronic antidepressant treatment reverses changes in 5-HT1A receptor RNA expression in chronic stress models (Burnet et al., 1995; Greenwood et al., 2003, 2005; Morley-Fletcher et al., 2004), as well as in depression (Lopez et al., 1998; Lopez-Figueroa et al., 2004). However, several studies in “normal” animals have found that chronic antidepressant treatments desensitize 5-HT1A receptors without a concomitant reduction in receptor levels (Le Poul et al., 1995; Elena Castro et al., 2003; Rossi et al., 2008). Because acute desensitization (uncoupling and internalization) occurs rapidly (second–minute) and is reversible (Riad et al., 2001, 2004), we postulated that acute desensitization is not sufficient for antidepressant response, but that a reduction in the synthesis of new 5-HT1A receptors may be required, and could account for the longer time course of the treatment (Albert et al., 1996; Albert and Lemonde, 2004). Interestingly, a 30% reduction in 5-HT1A autoreceptors is sufficient to permit a rapid and robust response to chronic SSRI treatment in mice that do not otherwise respond (Richardson-Jones et al., 2010), suggesting that transcriptional down-regulation of the 5-HT1A mRNA could be a key mechanism in determining antidepressant response.

Transcriptional Regulation of the 5-HT1A Receptor Gene in Mental Illness

Transcriptional regulation of the 5-HT1A receptor gene

Based on the hypothesis that impaired repression of the 5-HT1A receptor gene may predispose individuals to depression, we have identified key transcription regulators of the 5-HT1A receptor (Albert et al., 1996; Albert and Lemonde, 2004). The 5-HT1A receptor gene is intronless with a non-selective CG-rich housekeeping promoter containing multiple strong Sp1/MAZ enhancers that drive expression in all cells (Albert et al., 1996; Parks and Shenk, 1996; Storring et al., 1999) and an upstream repressor region that silences its expression in non-neuronal cells, but also represses transcription in neuronal cells that express 5-HT1A receptors (Ou et al., 2000; Lemonde et al., 2004b). The promoter also contains an NFkB-response element, that may mediate 5-HT1A induction in immune cells, although its function in neurons remains unaddressed. Repressors include Freud-1/CC2D1A and Freud-2/CC2D1B (Five-prime Repressor Element Under Dual repression binding protein), calcium-sensitive repressors of pre- (Freud-1) and post-synaptic 5-HT1A receptor expression in neurons (Ou et al., 2003; Lemonde et al., 2004b; Hadjighassem et al., 2009). We have also identified Deaf1 and Hes1/Hes5 as regulators at the C(−1019)G 5-HT1A polymorphism that repress 5-HT1A autoreceptor transcription at the C-allele but not the G-allele (Lemonde et al., 2003). Consistent with this, the G/G genotype is associated increased 5-HT1A autoreceptor binding potential in depressed subjects (Parsey et al., 2006b). Deaf1 is colocalized with 5-HT1A receptors in raphe and post-synaptic neurons, and expressed from development to adulthood; while Hes proteins are restricted to neural precursors and immature neurons. In Hes1−/− mice, 5-HT1A receptor RNA expression is greatly up-regulated in embryonic midbrain (Jacobsen et al., 2008), suggesting that the G(−1019) allele may result in an early embryonic alteration in 5-HT1A receptor levels. Interestingly, Deaf1 displays tissue-specific activity, repressing 5-HT1A autoreceptor RNA and protein expression in raphe cells, but enhancing 5-HT1A promoter activity in non-serotonergic neurons (Lemonde et al., 2003; Czesak et al., 2006). Based on this, the G-allele is expected to increase 5-HT1A autoreceptor levels to reduce 5-HT neuronal activity, and decrease post-synaptic 5-HT1A receptors, synergistically reducing 5-HT neurotransmission.

Transcriptional dysregulation of the 5-HT1A receptor gene in mental illness

Changes in 5-HT1A receptor RNA and protein levels are observed in brain regions of MDD subjects as discussed above, and suggest that alterations in 5-HT1A receptor gene expression could confer susceptibility. We initially associated the C(−1019)G polymorphism with MDD and suicide (Lemonde et al., 2003). This association has been replicated in most studies (Parsey et al., 2006b; Anttila et al., 2007; Kraus et al., 2007; Le François et al., 2008; Kishi et al., 2009; Neff et al., 2009). In our original study (Lemonde et al., 2003), both G/G genotype (P = 0.0134*) and G-allele frequencies (P = 0.0043**) were associated with depression in females (81 normal; 74 depressed). Males (53 control; 55 depressed) showed the same trends for genotype and allele frequency (P = 0.0846 and 0.0574), but they were not significant, suggesting a stronger association in females. However the allele and genotype ratios in depressed/control were similar (G/G genotype frequency 2.36× and 2.39× for females and males; G-allele frequency 1.45× and 1.33×). Hence there does not appear to be a clear gender effect of the polymorphism for association with depression.

Recently, the homozygous 5-HT1A G/G(−1019) risk genotype has been associated with increased levels of 5-HT1A autoreceptors in the raphe of depressed and bipolar depressed subjects (Parsey et al., 2006b; Sullivan et al., 2009), which is consistent with the increase in 5-HT1A autoreceptors observed in post-mortem studies of depressed suicides (Stockmeier et al., 1998; Drevets et al., 2007; Boldrini et al., 2008). This association supports the hypothesis that the 5-HT1A G(−1019) allele is a risk factor for depression by increasing 5-HT1A autoreceptor levels to reduce 5-HT neurotransmission (Albert and Lemonde, 2004). Interestingly, in serotonin neurons of depressed females, RNA levels of both Deaf1 and REST [a repressor of 5-HT1A receptor transcription (Lemonde et al., 2004b)] are increased, suggesting a compensatory change to re-establish normal 5-HT1A autoreceptor levels (Goswami et al., 2010). Despite these compensatory changes, there was a trend toward increased 5-HT1A RNA levels in depressed female raphe tissue, consistent with previous studies showing increased 5-HT1A autoreceptors. The recent finding that mice with selective increase in presynaptic 5-HT1A autoreceptors have depression-like behavior and are resistant to antidepressants (Richardson-Jones et al., 2010), provides validation of the hypothesis that dysregulation of the 5-HT1A autoreceptor is implicated in major depression (Albert et al., 1996; Albert and Lemonde, 2004).

In addition to its association with depression, we and others have found that the G(−1019) allele associates with reduced response to chronic SSRI or SSRI/pindolol treatment (Lemonde et al., 2004a; Le François et al., 2008; Villafuerte et al., 2009; Yevtushenko et al., 2010). Interestingly, antidepressant response to flibanserin, a 5-HT1A agonist which directly targets 5-HT1A receptor desensitization, was the most attenuated in subjects with the G/G genotype. This suggests that the Deaf1 site is critical for antidepressant action to down-regulate 5-HT1A autoreceptor expression. However, particularly in Asian populations where the G-allele is more prevalent, there is evidence that SSRI response is greater in the G/G genotype (Kato et al., 2009), suggesting that alternate regulatory factors (e.g., Freud-1, REST) could augment 5-HT1A autoreceptor down-regulation. Therefore, mechanisms preventing 5-HT1A autoreceptor de-repression could provide effective therapy for treatment-resistant depressed patients.

Thus, alterations in transcription factors that regulate 5-HT1A expression could either contribute to the dysregulation of the receptor or compensate for changes associated with mental illness.

Targeting 5-HT1A Autoreceptors for Antidepressant Response

Targeting 5-HT1A autoreceptors to augment to antidepressant efficacy

Several compounds that target 5-HT1A receptors including pindolol and buspirone, have been used in augmentation with SSRIs to accelerate or enhance their antidepressant action. Pindolol is a β-adrenergic receptor blocker also shown to act as a weak partial agonist of the 5-HT1A receptor (Newman-Tancredi et al., 1998). Pindolol preferentially competes with endogenous 5-HT at presynaptic 5-HT1A receptors (Rabiner et al., 2004; Serrats et al., 2004) to increase serotonergic neurotransmission. Clinical studies have demonstrated that pindolol can accelerate response to SSRIs, even in drug-resistant depression (Artigas et al., 1994; Blier and Bergeron, 1995; Portella et al., 2009). However, the ability of pindolol to improve SSRI treatment has not been consistent, in part due to low occupancy of 5-HT1A receptor sites at the dose of pindolol used in most studies (Martinez et al., 2000). In contrast to pindolol, buspirone functions as strong 5-HT1A partial agonist that can specifically desensitize 5-HT1A autoreceptors (Sim-Selley et al., 2000) to relieve serotonergic autoinhibition. Similarly to pindolol, buspirone was able to improve SSRI treatment in some studies (Harvey and Balon, 1995; Trivedi et al., 2006a) but not consistently (Blier and Ward, 2003). Differences between pindolol and buspirone appear due to their differential antagonist or agonist activity at 5-HT1A receptors. For example, unlike buspirone, acute pindolol treatment potentiates citalopram-mediated increase in 5-HT in rats (Hjorth, 1996). However, buspirone has unique agonist activities that can differ from full agonists like 5-HT. Buspirone and its analog BMY7378 inhibited Gi2-dependent constitutive activity of the 5-HT1A receptor, while full agonists (5-HT, 8OH-DPAT) or antagonists WAY100635, pindolol and NAN-190 had little or no effect (Albert et al., 1999). In the raphe nuclei, buspirone recruits Gi2, Gi3, and Go to the 5-HT1A receptor and inhibits adenylyl cyclase, while 8OH-DPAT only recruited Gi3 and did not inhibit forskolin-stimulated adenylyl cyclase (Valdizan et al., 2009). The preferential activity of buspirone in raphe cells may explain the efficacy of buspirone to decrease 5-HT synthesis in olfactory-bulbectomized rats (Watanabe et al., 2006). Other compounds that act at 5-HT1A receptors, such as atypical antipsychotics aripiprazole (Abilify) or SNRI milnacipran (Savella), can also ameliorate response to SSRIs. However, while these compounds desensitize the 5-HT1A autoreceptor (Mochizuki et al., 2002), they have other targets that can participate in their antidepressant actions. In addition it is important to note that drugs that indirectly target 5-HT neurons (Guiard et al., 2008) such as bupropion or α2-adrenergic receptor antagonists may also enhance response to SSRI response (Tremblay and Blier, 2006; Trivedi et al., 2006a; Blier et al., 2010). Thus, the concept of a multi-system therapeutic strategy appears to provide enhanced treatment response and remission rates compared to SSRI alone (Blier et al., 2010).

Targeting 5-HT1A gene transcription for the treatment of depression

The inconsistent results using 5-HT1A agonists to augment or accelerate response to SSRIs, may be due to their limited ability to decrease presynaptic 5-HT1A receptor levels. In fact, effective treatment appears to require long-term adaptive changes in addition to rapid desensitization of 5-HT1A autoreceptors, such as reduced transcription of 5-HT1A autoreceptors or induction of post-synaptic 5-HT1A receptors. For example, it was recently shown that aripiprazole increases hippocampal and cortical 5-HT1A receptors (Han et al., 2009). This suggests that long-term activation of presynaptic 5-HT1A receptors might trigger transcriptional changes that follow the rapid initial desensitization of the receptors and lead to long-term alterations in 5-HT1A receptor levels. The recent finding that a 30% transcription-mediated increase in 5-HT1A autoreceptors is sufficient to abolish response to chronic antidepressant treatment in adult mice (Richardson-Jones et al., 2010), suggests that an ability to even slightly down-regulate 5-HT1A autoreceptor expression could greatly improve SSRI response in treatment-resistant patients.

Additionally, it would be important to determine how 5-HT1A agonists affect the expression and activity of key transcriptional regulators of the 5-HT1A receptor gene. Recent findings suggested that alterations in Deaf1 occur both in raphe and prefrontal cortex of depressed females (Szewczyk et al., 2009; Goswami et al., 2010), suggesting that females may be more responsive to drugs that could upregulate Deaf1 expression. Up-regulation of Deaf-1 levels would be expected to repress presynaptic and induce post-synaptic 5-HT1A gene expression to increase serotonin neurotransmission in heterozygous or C/C homozygous subjects. However, patients with the 5-HT1A G/G(−1019) genotype would be expected to be resistant since the G-allele does not respond to Deaf1.

Since depression is twice as frequent in females compared to males, gender-specific alterations in 5-HT1A autoreceptor expression could mediate an increased susceptibility. In female rats and serotonin transporter-deficient female mice, ovariectomy results in increased raphe 5-HT1A receptor RNA and binding (Bouali et al., 2003; Le Saux and Di Paolo, 2005) and increased anxiety behavior (Imwalle et al., 2005), while estrogen treatment reverses changes in 5-HT1A levels and anxiety, increasing raphe responsiveness (Abizaid et al., 2005). Similarly, estrogen receptor-beta knockout female mice display increased anxiety and reduced raphe 5-HT levels (Imwalle et al., 2005). In female rhesus macaques, ovariectomy leads to an increase in nuclear NFkB in 5-HT raphe neurons, while estrogen/progesterone reverses this (Bethea et al., 2006) and reduces 5-HT1A autoreceptor protein level (Henderson and Bethea, 2008), which could be driven by NFkB-response element in the 5-HT1A promoter. Thus estrogen appear to be protective for anxiety and negatively regulates 5-HT1A autoreceptor levels to enhance raphe activity, but the mechanisms of this regulation remain to be clarified.

Antidepressant treatments could also target cortical 5-HT1A receptors, which are reduced in depression. Down-regulation of Freud-1 or Freud-2 in post-synaptic regions could lead to an increase in 5-HT1A receptor RNA and normalize post-synaptic 5-HT1A receptor expression and restore response of post-synaptic neuron to 5-HT. Interestingly, Freud-1 RNA and protein was down-regulated in the prefrontal cortex of chronically stressed rats (Iyo et al., 2009), while 5-HT1A RNA level was increased, although 5-HT1A protein was reduced. Similarly in the prefrontal cortex of younger (<58 years) depressed subjects (especially males), Freud-1 RNA and protein were decreased, while 5-HT1A receptor protein was also reduced (Szewczyk et al., 2010). Thus, despite increased 5-HT1A RNA other post-translational processes, such as receptor down-regulation, may reduce cortical 5-HT1A protein levels in depression. Interestingly, cortical Freud-1 levels were not altered by chronic treatment of macaques with SSRI, suggesting that other types of antidepressant treatments may be required to induce this effect. Alternately, Freud-1 activity could be reduced, for example by calcium mobilization, which inactivates Freud-1 (Ou et al., 2003).

Conclusion

Alterations in pre- and post-synaptic 5-HT1A receptor RNA and protein expression in depression suggest that transcriptional dysregulation of the receptor is involved. Currently used 5-HT1A selective ligands are well known to induce desensitization of 5-HT1A receptors, but should be tested for their ability to alter 5-HT1A gene transcription. Among the known transcriptional regulators of the 5-HT1A receptor, Deaf-1 is a very attractive target due to its ability to repress presynaptic 5-HT1A expression, while inducing post-synaptic 5-HT1A receptors. Thus, identification of ligands or strategies that activate or induce Deaf-1 could simultaneously target pre- and post-synaptic 5-HT1A receptor expression to increase serotonergic tone.

Conflict of Interest Statement

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.