In this study, we investigated the mechanism of action of mirtazapine with respect to its ability to increase BDNF levels and, in particular, we aimed to determine whether this effect is due to an increase in BDNF mRNA translatability. Specifically, we tested in SH-SY5Y human neuroblastoma cells the alternative hypotheses that mirtazapine could increase BDNF translation either by direct activation of serotonergic/noradrenergic receptors or, alternatively, by increased release of 5HT/NE neurotransmitters. These two hypotheses were formulated based on the current literature and on our bioinformatic analysis which confirmed that serotonergic and noradrenergic receptors are among the numerous targets of mirtazapine. Experimentally, we found that 1h mirtazapine treatment reduced or had no effect on the serotonergic receptors 5HT1A, 5HT2A, 5HT2B and 5HT2C and the two adrenoceptors ADRA1D and ADRB1 protein expression levels, whereas it selectively increased the ADRA2A and ADRB2 receptors. In addition, we investigated the translation of the individual BDNF transcripts I, IIc, IV and VI, in response to 5HT and NE and we found an increase only for exon IIc after 3h treatment with NE, but not with 5HT. Notably, both endogenous BDNF and exogenous exon IIc protein levels were increased by mirtazapine at 3h. The involvement of noradrenergic signaling in the regulation of BDNF translation was further supported by the finding that inhibition of either ADRA2A or ADRB2 receptors prevented the effects of NE on exon IIc translation, whereas only the inhibition of ADRB2 could block the effect of mirtazapine on translation of the endogenous BDNF and BDNF exon IIc transcript. Finally, we demonstrated that mirtazapine treatment promotes increased release of NE in the neuroblastoma cell supernatant. We propose that the increase in BDNF protein levels induced by mirtazapine could be largely accounted by exon IIc transcript translation. Furthermore, we propose a mechanism of action of mirtazapine on BDNF translation through increased NE release and activation of the ADRA2A and ADRB2 receptors, which are also upregulated by mirtazapine treatment.
The first biological question addressed in our study was whether mirtazapine could modulate BDNF expression levels, specifically at the level of translational regulation rather than at the transcriptional level. The current literature provides many examples of the transcriptional regulation of BDNF transcripts in response to antidepressants and mood stabilizers [34, 35]. In particular, two of the major mood stabilizers, lithium and valproate, have been shown to increase the translational activation of the BDNF exon IV promoter [36]. In the rat prefrontal cortex, prolonged, but not short-term fluoxetine treatment up-regulates BDNF mRNA expression, indicating that this change takes time to develop, possibly explaining the slow onset of antidepressant activity observed with fluoxetine [37]. In 2017, Rogóz et al. investigated the effect of repeated administration of three different antidepressants (escitalopram, fluoxetine, and mirtazapine) alone or in combination with low doses of the antipsychotic drug risperidone in Wistar rats. They found that escitalopram increased hippocampal BDNF mRNA expression only at the highest concentration tested (10 mg/kg). In addition, while BDNF mRNA levels were increased in the cortex by fluoxetine alone, to have a similar effect mirtazapine and escitalopram required the combination with risperidone [24]. In a previous study, the same group also found increased BDNF mRNA levels in rat brain after treatment with mirtazapine only [38]. Another study showed that hippocampal BDNF mRNA expression was increased after two days treatment with the SNRI reboxetine combined with physical exercise, whereas the SSRI citalopram required two weeks of treatment to produce an increase in BDNF mRNA [39]. There are also studies in the literature reporting effects of antidepressants on BDNF protein levels. Repeated administration of escitalopram increased BDNF protein levels in the hippocampus, while mirtazapine and fluoxetine affected BDNF protein both in the cortex and hippocampus only in combination with risperidone [24]. In addition, a study showed that the antidepressants desipramine, fluoxetine and phenelzine increased BDNF protein levels in the frontal cortex after chronic, but not acute treatment. In the same study, were also tested two antipsychotics, haloperidol, and clozapine, and it was found that they can both positively regulate BDNF protein levels in the frontal cortex after chronic treatment [40]. In our study, we saw an increase in total endogenous BDNF protein in SH-SY5Y human neuroblastoma cells after 3h treatment with 10µM mirtazapine, but not with the lower dose (1µM). While there is a general lack of dose-dependent studies on antidepressants effects on BDNF protein production, it has been reported that BDNF mRNA levels were increased in both rat hippocampus and cortex only at high (10 mg/kg) but not at low mirtazapine doses (5 mg/kg) [38].
From a methodological point of view, it must be considered that the overall amount of BDNF protein level is the final result of translation of multiple transcripts, giving the same final protein product. Thus, the only way to investigate translation of individual BDNF mRNAs is to study artificial constructs for the different BDNF exons. Indeed, two previous studies investigated the translatability of the different BDNF transcripts under basal conditions and in response to various stimuli using synthetic constructs [20, 41]. Here, we investigated if BDNF I, IIc, IV and VI transcripts are translated in response to mirtazapine, serotonin and norepinephrine treatments using plasmids containing BDNF 5’UTR-luciferase-3’UTR constructs ([20]; patent no. PCT/EP2010/067081). In agreement with Vaghi et al., we found that MTZ10 increased the translatability of BDNF exon IIc and that 5HT has no effect on the translatability of the different exons, even at a higher concentration of 5HT (100µM) with respect to that previously used (50µM) by Vaghi et al. In contrast, we observed a significant increase in the translatability of exon IIc after treatment with NE that was not found by Vaghi et al. [20], who instead showed increased exon VI translation. The main difference between the two studies is the concentration of NE used to treat SH-SY5Y cells. While Vaghi et al. used 50µM of NE, in the present study treatments were performed at a lower dose (10µM) of the neurotransmitter. It was already shown that each 5’UTR-luciferase BDNF constructs have a different degree of translatability [20, 41]. In particular, exon VI showed the lowest rate of basal translatability which may explain why low doses of NE had no effect on translatability of this BDNF variant compared to 50µM. Instead, the selective increase of BDNF exon IIc translation at low NE doses, could be due to the specific pharmacological dynamics of the beta-adrenergic receptors [42].
A second key biological question addressed by this study is whether the increase in BDNF translation is accounted by direct mirtazapine-induced activation of serotonergic/noradrenergic receptors or is mediated by the release of neurotransmitters. We found that mirtazapine treatment leads to increased norepinephrine release with the same time course as the mirtazapine-induced increase in total endogenous BDNF, i.e. after 3h treatment. It is known that mirtazapine, unlike other antidepressants such as SSRI or SNRI, does not inhibit the reuptake of serotonin, norepinephrine, or dopamine [22]. Indeed, the blockade of alpha2-adrenergic autoreceptors and heteroreceptors leads to increased norepinephrine release [21]. In addition, we found that the effect of NE and mirtazapine on BDNF translation, is prevented by the inhibition of the noradrenergic ADRB2 and ADR2A, strongly implying these receptors in the regulation of BDNF levels. It was previously shown that treating primary cultures of hippocampal neurons with norepinephrine increases BDNF levels [26]. Furthermore, serotonin and norepinephrine differentially increase BDNF expression [43]. This is in line with our results in which, norepinephrine, but not serotonin, induced an increase in both exogenous and endogenous BDNF.
Mirtazapine is a multi-target drug, capable of interacting with a number of serotonergic and adrenergic receptors. Remarkably, our western blot analysis showed that 1h mirtazapine treatment is sufficient to change the relative levels of 5HT1A (downregulated) and 2 adrenergic receptors (upregulated). While the binding profile of this antidepressant to these receptors is well established [44], no other studies had previously described an altering effect of mirtazapine on the levels of serotonergic and adrenergic receptors. Of note, a recent study showed that early life exposure to the SSRI citalopram, in rats increased mRNA expression of 5HT1A in the medial preoptic area and 5HT1B, 5HT2A and 5HT2C in the prefrontal cortex [45]. More in general, studies on the effects of mirtazapine on specific protein levels are scant and we only found a study showing that mirtazapine can increase the hepatic CXCL10 chemokine and its cognate receptor CXCR3 [46]. Thus, our results provide a completely new perspective in which mirtazapine, in addition to promoting NE release, also acts by regulating the protein levels of serotonergic and adrenergic receptors, favoring noradrenergic responses through an increased expression of ADRA2A and ADRB2, and a concomitant reduction in the expression of serotonergic receptors.
This study also has some limitations. The first limitation is posed by the cell line that we have used, the SH-SY5Y cells which are derived from a bone marrow biopsy containing metastatic neuroblastoma cells [47]. These cells exhibit a neuron-like phenotype upon induced differentiation but have a major disadvantage as they can reach an undefined differentiation state [48]. On the other hand, these cells are a popular in vitro model used to study neurodevelopmental disorders and express both serotonergic [49, 50] and adrenergic receptors [29, 51]. In particular, SH-SY5Y cells are used in studies focusing on the adrenergic system [52].
A second limitation concerns the number and type of receptors investigated. Since mirtazapine is a multi-target drug, and the SH-SY5Y cells can also express dopaminergic [53], opioid, muscarinic and NGF receptors [54] other receptors may be involved in the regulation of BDNF production by mirtazapine. However, this study was focused on serotonergic and noradrenergic receptors and further studies are required to investigate additional receptors.
One further limitation of the present study relates to the fact that BDNF is probably not the only neurotrophic factor that is increased by mirtazapine. Indeed, a recent study has shown that mirtazapine enhances the production of glial cell line-derived neurotrophic factor (GDNF) in astrocytes by increasing the phosphorylation of ERK1/2 through the LPA1 receptor [55].
In conclusion, our results shed light on one of the possible mechanisms of action of mirtazapine in regulating the increase in total BDNF levels. In particular, we have shown that mirtazapine acts by increasing the translatability of exon II of BDNF, supporting the hypothesis proposed by Vaghi et al. that BDNF mRNA variants represent a quantitative code for the regulation of BDNF protein [20]. Furthermore, our study shows that the effect of mirtazapine on BDNF translation is strictly related to the noradrenergic system, at least in neuroblastoma cells. Confirming the systematic review of Gillman in which the author suggested that mirtazapine has not any significant serotonergic effects [56]. Indeed, this antidepressant can increase norepinephrine release and it requires the activation of the ADRB2 receptor. Such a mechanism of translational regulation of BDNF, and specifically of the BDNF exon IIc transcript, has not been described before and is of interest for understanding the neurotrophic effect of antidepressants.