All the brain's a stage for serotonin: the forgotten story of serotonin diffusion across cell membranes

1. Introduction

In the conventional model of serotonin neurotransmission in the central nervous system, serotonin is released into synapses¹ in forebrain regions from neurons that originate primarily in the dorsal raphe nucleus (DRN). Serotonin then affects post-synaptic² neurons through receptor-mediated signalling pathways, presumably without ever entering the post-synaptic neuron. The serotonin transporter (SERT), which is expressed adjacent to release sites on serotonergic neurons, mediates reuptake of serotonin into the releasing cell, limiting the magnitude and duration of target cell responses. SERT has high affinity for serotonin, but a low transport capacity, and SERT-mediated transport is an energy-dependent process [1–3].

According to the conventional model, serotonin does not cross the membrane of the post-synaptic neuron because its positive charge at physiological pH prevents diffusion across the post-synaptic cell's plasma membrane, and because SERT is expressed only on serotonergic neurons. However, research since the 1970s has shown that serotonin can cross cell membranes through a diffusion-like process, despite having a positive charge [3–9]. Subsequent research solved this apparent paradox through the discovery of multiple transporters in addition to SERT that are widely expressed in the brain, and that can mediate diffusion of serotonin and other monoamines across cell membranes. These transporters, which include the organic cation transporters (OCT1, OCT2 and OCT3) and the plasma membrane monoamine transporter (PMAT), are broadly specific and are able to move catecholamines and histamine in addition to serotonin [10,11]. In comparison to SERT, they have relatively low affinity for serotonin, but they have substantially higher capacity [10] and, unlike SERT, transport mediated by these carriers is energy-independent [12–14]. The rate of serotonin transport via these higher capacity carriers can exceed the rate of serotonin reuptake via SERT at concentrations of serotonin commonly found in serotonergic synapses, suggesting that carrier-mediated diffusion of serotonin is physiologically consequential [15]. Importantly, these transporters are expressed in neuronal and glial cells in the brain and have been shown to play roles in the regulation of extracellular monoamine concentrations [16–18]. Through these transporters, serotonin may enter non-serotonergic neurons, a possibility not considered in the conventional model of serotonergic neurotransmission. Because most neuroscientific textbooks have neglected the roles of these transporters in the clearance of extracellular serotonin, most neuroscientists are unaware of their existence and potential role in regulating serotonergic neurotransmission.

The failure of the conventional model to incorporate mechanisms that allow serotonin to cross cell membranes and enter non-serotonergic cells perpetuates a limited view of the mechanisms by which serotonin may regulate cellular function and, consequently, influence brain function and behaviour. Transporter-mediated clearance is a key determinant of the magnitude, duration, and physical spread of released serotonin. Ignoring the contribution of non-SERT transporters to serotonin clearance limits our understanding of a key regulator of serotonergic neurotransmission. It also limits research aimed at understanding the potential intracellular actions of serotonin in post-synaptic neurons and glial cells. The limited view perpetuated by the conventional model also has implications for the pharmacological strategies widely used to treat depression and other neuropsychiatric conditions. In the USA, nearly 13% of people aged 12 or older have taken an antidepressant in the last month [19], and the most commonly prescribed antidepressants are the selective serotonin reuptake inhibitors (SSRIs), which elevate extracellular serotonin concentrations by inhibiting SERT-mediated serotonin transport. While these are among the most widely prescribed drugs, and while many patients benefit from them, the mechanisms by which SSRIs exert their therapeutic effects, and by which serotonin regulates mood and anxiety, are still not understood [20–23]. The ability of non-serotonergic neurons and glia to take up serotonin indicates that current models of serotonin's role in depression, and of the actions of widely used antidepressants, must be re-evaluated to incorporate additional mechanisms.

In this paper, we first provide some relevant background on serotonin before reviewing the molecular mechanisms that allow serotonin to cross cell membranes and enter forebrain neurons. We then provide a brief review of some of the known intracellular functions of serotonin. We suggest that intracellular actions of serotonin may contribute to the well-described effects of serotonin on critical neuronal functions, including plasticity, and we discuss some implications of the ability of serotonin to cross cell membranes and enter forebrain neurons.

2. Background

Serotonin, a monoamine derived from the amino acid tryptophan, exerts powerful and pervasive influences on behaviour by regulating the activity of neurons throughout the brain. It is synthesized by a small population of serotonergic neurons in the brainstem raphe nuclei. These neurons project widely throughout the brain such that there is essentially no region of the brain that does not receive serotonergic innervation [1,24]. Serotonin synthesis is a two-step process [25]. The first step, which is rate-limiting, is the oxidation of tryptophan to 5-hydroxytryptophan (5-HTP), catalysed by the enzyme tryptophan hydroxylase (Tph). There are two variants of the tryptophan hydroxylase enzyme: Tph1 is expressed in peripheral tissues including the pineal gland, while Tph2 is expressed in the brain [26]. The second step is the conversion of 5-HTP to 5-hydroxytryptamine (5-HT, serotonin) catalysed by aromatic l-amino acid decarboxylase (AADC). Serotonin is unstable at physiological pH and, after synthesis, it is usually packaged into vesicles via vesicular monoamine transporter 2 (VMAT2). Uptake of serotonin into these vesicles is an active process; VMAT2 uses a proton gradient to transport serotonin into vesicles [27]. Serotonin is also synthesized by enterochromaffin cells in intestinal epithelium. This serotonin acts locally on intestinal neurons to regulate gastrointestinal smooth muscle tone and contractions; however, much of it is taken up into blood platelets in the portal vein, packaged in vesicles, and circulated via these platelets throughout the periphery [20]. Peripherally synthesized serotonin does not enter the brain, as it cannot cross the blood-brain-barrier, but platelets were key model systems in early studies of serotonin transport and release [1,20].

After it is released, serotonin exerts powerful and diverse effects on neuronal and other cells via a large family of receptors, the 5-HT receptors, consisting of seven distinct classes (5-HT₁ to 5-HT₇) [28]. With the exception of the 5-HT₃ receptor, which is an ion channel, all of the 5-HT receptors are G-protein-coupled receptors. This diversity of receptors is thought to underlie the diverse cell type- and brain region-specific responses to serotonin.

During a single neuronal impulse, local extracellular concentrations of serotonin can reach 100 nM, but they can reach micromolar levels with repeated firing [29]. Most serotonergic synapses in the prefrontal cortex are non-junctional, which allows serotonin to diffuse out of the synapse before it is removed by SERT [30]. While SERT is transiently expressed in non-serotonergic neurons in early development, this is not usually the case in the brains of adult mammals [31] such that synaptic serotonin cleared by SERT is usually taken back into the serotonergic neuron. Serotonin can diffuse more than 20 µm while maintaining a micromolar to nanomolar concentration, while a typical synapse size is on the order of magnitude of tens of nanometres [6].

The duration, peak concentration, and physical spread of transmitted serotonin, and therefore the magnitude and duration of serotonin actions on target cells, are largely determined by transport mechanisms that clear serotonin from the extracellular space. Serotonin clearance has long been attributed exclusively to SERT, which is expressed in tissues throughout the body, including the heart, blood vessels, platelets, liver, gallbladder, adrenal gland, kidney, immune system and lungs [32–35]. This widespread expression allows many organs to actively take up serotonin, following the exocytosis of serotonin from platelets [36]. It is perhaps not surprising, then, that serotonin is involved in just about every major process in the body: serotonin regulates the functioning of many organs (e.g. heart, blood vessels, lungs, muscles, pancreas, kidney) and processes (e.g. clotting, development, thermoregulation, hunger, reproduction, immune function), and it functions as a potent neuromodulator in the central nervous system, across the spectrum of vertebrate and invertebrate eukaryotic organisms [1,20,37]. Intracellular serotonin can be metabolized into 5-hydroxy-indoleacetic acid (5-HIAA) in a reaction catalysed by the enzyme monoamine oxidase (MAO), which occurs as two isoforms, MAO-A and MAO-B, which display different affinities for serotonin and different cellular expression patterns [1]. Both MAO isoforms are localized to the outer mitochondrial membrane, but their catalytic surfaces are oriented differently, with MAO-A oriented on the cytosolic face, and MAO-B oriented facing the intermembrane space [38].

3. Molecular mechanisms that allow serotonin to enter forebrain neurons

Starting in the late 1960s, platelets—which store serotonin in granules—were used as models to elucidate mechanisms of serotonin neurotransmission, under the assumption that they would serve as a good model of neuronal transport [39,40]. This research found two transport mechanisms [8]: (1) an active transport mechanism (via SERT) that predominated at low extracellular serotonin concentrations (less than 100 nM) and (2) a passive diffusion-like mechanism that predominated at high extracellular concentrations (greater than 100 nM). Serotonin has been found to cross membranes of many cell types through a diffusion-like process, including kidney cells [4,9], heart cells [5], endothelial cells [41], neuronal synaptosomes [6] and neurons [7]. We now know that this diffusion-like mechanism is mediated by non-specific carriers. In this context, the micromolar concentrations of serotonin elicited with repeated neuronal firing [28] lie well within the range required for carrier-mediated diffusion to predominate.

One possible reason for the lack of attention to non-SERT mechanisms of serotonin uptake is that platelet cells were models of serotonergic neurons, which synthesize and transmit serotonin [8,39,40,42]. They were not considered models of non-serotonergic, forebrain neurons. Nevertheless, we can think of no rational reason why early researchers would have assumed that passive transport of serotonin across cell membranes was limited to serotonergic neurons, particularly since this mechanism had been demonstrated in multiple cell types [42]. Another possible reason why the evidence of a diffusion-like mechanism was ignored is because researchers did not understand the mechanism, due to serotonin's positive charge. This is also not a satisfying answer, because it should have stimulated research that attempted to resolve the paradox.

In any event, it is surprising that the evidence serotonin crosses cell membranes through a diffusion-like process was not even incorporated into the textbook descriptions of serotonergic neurons [1–3]. Some researchers in the early 1970s thought that the transport of serotonin into serotonergic neurons by diffusion was not as physiologically important as SERT uptake [42]. Of course, that does not mean that the diffusion-like uptake of serotonin was generally unimportant in the brain.

The paradox regarding serotonin's transport through a diffusion-like process despite having a positive charge was resolved by the discovery of several molecular mechanisms of passive transport [10,14]. Passive, diffusion-like transport of serotonin has now been attributed to a family of broadly specific monoamine transporters. These transporters are all considered members of the ‘uptake₂’ family of monoamine transporters, to contrast them with SERT, which is a member of the ‘uptake₁’ family. Uptake₁ and uptake₂ activities were originally observed in studies of catecholamine uptake in heart tissue. Uptake₁ was a high-affinity (K_d = 0.27 µM), low-capacity (V_max = 1.22 nmol min⁻¹ g⁻¹ tissue) transport process, inhibited by cocaine and desipramine, while uptake₂ had lower affinity (K_d = 252 µM) and higher capacity (V_max = 100 nmol min⁻¹ g⁻¹) for catecholamines and was insensitive to cocaine and desipramine [43,44]. Because of its relatively low affinity, uptake₂ was originally believed to mediate catecholamine uptake only when its substrates were at very high concentrations. However, later studies demonstrated uptake₂ contributions to catecholamine clearance under both high and low substrate concentrations [45].

Uptake₂-like activity has been attributed to at least four distinct transporters, all of which mediate high-capacity transport not only of catecholamines, but also of serotonin and, in some cases, histamine. These transporters, organic cation transporters (OCT1, OCT2 and OCT3), and the PMAT, all mediate high-capacity, bidirectional, sodium-independent transport of serotonin and other monoamines, with each transporter displaying distinct transport efficiencies for the various monoamines [10,46] (table 1). The PMAT is thought to have some selectivity for serotonin transport in the brain. The K_m and IC₅₀ values for serotonin transport by OCT transporters, multidrug and toxin extrusion protein 1 (MATE1), and PMAT as determined following expression in cell lines (table 1) may underestimate their functional role in neural tissues, as the IC₅₀ value for serotonin inhibition of [³H]-histamine uptake—a known substrate of OCT2 and OCT3 [47]—in rat hypothalamic tissue was estimated to be 43 µM [13]. Functionally, Xenopus laevis oocytes, which are made to express a variant of PMAT found throughout the human brain, alongside the human SERT protein, increase the uptake of serotonin by 2.5- to 3-fold, at 1 µM concentration, relative to when SERT alone was expressed [48]. The contribution of monoamine transporters to overall serotonin uptake in the brain, therefore, may be considerable.

Furthermore, low-affinity, high-capacity serotonin transporters, such as OCT3, which is widely expressed throughout the brain in both neurons and astrocytes [16], are functionally important and may play a role in the modulation of serotonergic signalling following exposure to aversive stimuli [12,13,49–56]. Such data suggest that a sustained elevation in serotonin transmission into the forebrain could boost uptake into post-synaptic neurons through these transporters (table 1), as has been demonstrated for OCT3 [50–52].

Although all uptake₂ transporters except for PMAT were originally described in peripheral tissues, all have since been detected in brain tissue [17,18] and specifically in forebrain neurons [57–60] and, for OCT3, in astrocytes and ependymal cells [16,59].

We have conducted studies using immuno-electron microscopy to examine the subcellular localization of OCT3 in the amygdala and demonstrated that, in addition to axonal and dendritic plasma membranes, OCT3 is also localized to Golgi, mitochondrial, and outer nuclear membranes [59], suggesting a potential mechanism through which serotonin may enter or exit these organelles (figure 1).

• Download figure
• Open in new tab
• Download PowerPoint

In that same study, we observed similar localization patterns in cortical neurons (figure 2), although we did not publish those results at that time. Of particular interest is the fact that OCT3-expressing mitochondria were often observed adjacent to dendritic plasma membranes that also expressed OCT3, suggesting that, in these cortical neurons, serotonin or other OCT3 substrates may be transported across the post-synaptic membrane from the extracellular space, directly into the mitochondrial intermembrane space (figure 2a,b).

• Download figure
• Open in new tab
• Download PowerPoint

In short, decades of research have demonstrated that serotonin can cross neuronal and other cell membranes under physiological conditions through a passive non-SERT, carrier-mediated process. The relative inattention to this mechanism, in neurobiology textbooks [1–3] and elsewhere, means that most models of serotonergic neurotransmission are incomplete.

4. The distribution of the monoamine oxidase enzymes

During development, both variants of MAO are often co-expressed [61]. However, postnatally, MAO-A, which has the higher affinity for serotonin, tends to be more abundant in non-serotonergic forebrain neurons such as pyramidal neurons in the cortex and dentate granule neurons of the hippocampus [62,63]. Conversely, the low-affinity MAO-B is more abundant in serotonergic neurons, postnatally. Interestingly, MAO-A and MAO-B are co-expressed in astrocytes of the adult brain [64].

The regional differences in the expression of MAO-A and MAO-B have puzzled researchers [62,63]. The greater expression of the low affinity MAO-B enzyme in serotonergic neurons may be an adaptation to minimize serotonin metabolism in those neurons, maintaining the availability for neurotransmission. The presence of MAO-A in forebrain neurons poses a conundrum for the conventional model of serotonin transmission, and it should have been another clue that the model was wrong or incomplete. While the preferential expression of the high-affinity variant of MAO in forebrain neurons is itself no guarantee of serotonin being present in these neurons, it does indicate that MAO-expressing neurons have the capacity to metabolize monoamines, including serotonin. Furthermore, it suggests the hypothesis that, in these neurons, MAO-A acts to metabolize serotonin taken up from the extracellular space.

There are alternative hypotheses that try to explain the expression of MAO-A under the conventional model of serotonin transmission. One such hypothesis proposes that MAO-A acts as a scavenger for errant monoamines that have inadvertently lost their way and entered neurons where they have no appropriate physiological role [64]. Another hypothesis suggests that MAO-A might eliminate foreign amines that are false neurotransmitters [60]. However, neither of these hypotheses offers a natural explanation for the fact that MAO-A and MAO-B have higher affinities for deprotonated serotonin [38]. Below, we provide an explanation for this fact.

5. Post-synaptic and intracellular functions of serotonin

Surprisingly, the ability of serotonin to cross cell membranes through a diffusion-like process has even been neglected in the textbooks of researchers [3,65] who were instrumental in demonstrating it [8]. Again, the reasons for this are unclear. By the 1970s, serotonin was known to have both excitatory and inhibitory effects on post-synaptic neurons [66,67]. Generally, researchers thought that the excitatory and inhibitory effects required the existence of multiple serotonin receptor subtypes expressed on post-synaptic neuron membranes [66]. Thus, it is possible that most researchers thought that serotonin affected post-synaptic neurons primarily through receptor-mediated signalling pathways, negating other possible mechanisms. This might have implied that even if serotonin entered post-synaptic neurons through diffusion-like processes, it would have no important intracellular effects. Nevertheless, the intraneural metabolism of serotonin led some researchers to question whether serotonin had intraneural functions [68].

The strong anatomical and physiological evidence that cells other than serotonergic neurons can take up serotonin via a variety of transporters raises important questions about the fate of serotonin in these cells. Indeed, we now know that serotonin has important intracellular effects, and these effects may complement serotonin's plasma membrane receptor-mediated effects on neuronal firing. Below, we describe evidence for three potential fates of intracellular serotonin: (1) metabolism by MAO at the mitochondrion; (2) covalent attachment to proteins at glutamine residues by transglutaminase enzymes (i.e. serotonylation); and (3) activation of serotonin receptors at intracellular membranes.

6. Discussion

The conventional model of serotonin neurotransmission does not include mechanisms mediating serotonin transport across the post-synaptic membrane. In the absence of a transport mechanism, the positive charge at physiological pH should prevent the passive diffusion of serotonin across cell membranes, and yet several decades of research have demonstrated that serotonin passes across cell membranes through a diffusion-like process [4–9]. The discovery of high-capacity, low-affinity, sodium-independent transporters, including OCT3 and PMAT, resolves this paradox. These transporters, which had not yet been discovered in the 1970s when the traditional model of serotonergic signalling was developed, offer a gateway for serotonin to passively diffuse across cell membranes despite its positive charge.

Carrier-mediated diffusion puts the intracellular environment of post-synaptic forebrain neurons within reach of serotonin. Similarly, the widespread expression of SERT and non-specific carriers in the periphery arguably gives every organ the capacity to take up serotonin from the bloodstream [20]. Thus, the cytoplasm of every cell in the body is potentially a stage for serotonin's actions.

7. Conclusion

There are multiple low-affinity, high-capacity, sodium-independent transporters—such as OCT3—that are widely expressed in the brain and allow the carrier-mediated diffusion of serotonin into post-synaptic neurons and glial cells. The evidence that serotonin can cross cell membranes through a diffusion-like mechanism despite having a positive charge—now several decades old—is no longer a paradox and should no longer be ignored. There is also considerable evidence that serotonin enters forebrain neurons and interacts with organelles, such as the Golgi apparatus, mitochondria and nucleus. Even small concentrations of serotonin could have important intracellular effects, similar to how small concentrations of melatonin have important intracellular effects [112,113]. The story of serotonin neurotransmission is far from complete, but research on this is hindered by the fact that the mechanisms by which serotonin can enter forebrain neurons are not widely known. We encourage the textbooks to modify the conventional model to include carrier-mediated diffusion mechanisms, and we encourage new research that focuses on the intracellular functions of serotonin, particularly those in post-synaptic neurons and glia.

Data accessibility

This article has no additional data.

Authors' contributions

P.W.A.: conceptualization, project administration, supervision, writing—original draft, writing—review and editing; C.B.: conceptualization, writing—review and editing; L.B.: conceptualization, writing—review and editing; L.G.: conceptualization, writing—review and editing; P.J.G.: conceptualization, data curation, writing—review and editing; C.A.L.: conceptualization, writing—review and editing; V.M.P.: data curation, writing—review and editing.

All authors gave final approval for publication and agreed to be held accountable for the work performed therein.

Conflict of interest declaration

We declare we have no competing interests.

Funding

C.A.L. is currently supported by the National Center for Complementary and Integrative Health (grant nos. R01AT010005 and R41AT011390), the Colorado Office of Economic Development and International Trade (OEDIT) Advanced Industries Accelerator Program (grant no. CTGG1-2020-3064), the Department of the Navy, Office of Naval Research Multidisciplinary University Research Initiative (MURI) Award (grant no. N00014-15-1-2809), the Institute for Cannabis Research (grant no. ICR-FY23-Lowry), and an anonymous donor through Benefunder. P.J.G. is supported by the National Institute on Drug Abuse (grant no. DA032895), the Charles E. Kubly Mental Health Research Center, and the Just Live Foundation. V.M.P. is supported by the National Institute on Drug Abuse (grant no. DA042943).