Seconds-resolved in vivo measurements reveal rapid, serotonin transporter–independent plasma serotonin clearance

Key Points

• •Seconds-resolved serotonin measurements enable direct observation of uptake in blood ex vivo and clearance from the circulation in vivo.
• •Plasma serotonin is cleared even after SERT inhibition, indicating compensatory, SERT-independent pathway(s) can maintain serotonin balance.

Visual Abstract

Abstract

Serotonin potently regulates vascular tone, permeability, and platelet activation. Because of this, circulating plasma-free serotonin is normally maintained at very low concentrations through rapid, serotonin transporter (SERT)–mediated uptake into platelets. The well-established tolerability of SERT-blocking selective serotonin reuptake inhibitors (SSRIs), however, suggests that additional systemic mechanisms preserve serotonin homeostasis when SERT is inhibited. To test the hypothesized existence of such non-SERT plasma serotonin clearance, here we used an electrochemical biosensor to measure, with 8-second resolution, the dynamics of plasma serotonin in the presence and absence of pharmacologic SERT inhibition both in vivo in the rat jugular vein and ex vivo in fresh, whole blood. In vivo, we found that, following IV serotonin challenges in animals with intact platelet function, plasma serotonin concentrations rapidly (t_1/2 ∼2 minutes) fell below the sensor’s detection limit, ∼20nM in vivo. In contrast, after using the SSRI escitalopram to inhibit SERT, we observed a large, transient (ie, >8-14 minutes) elevation of baseline plasma serotonin, and subsequent challenges with exogenous serotonin exhibited a notably larger peak and a slower (t_1/2 ∼10 minutes) return to baseline. Ex vivo experiments performed using fresh, whole blood indicated that the level of SSRI employed in our studies is sufficient to effectively abolish platelet SERT activity, suggesting that the postescitalopram clearance of supraphysiological serotonin concentrations in intact animals occurs via yet-to-be-identified, SERT-independent clearance pathways occurring in the solid tissues. By revealing a SERT-independent compensatory clearance capacity that reduces plasma serotonin levels when platelet uptake is impaired, these findings refine current models of the platelet-mediated regulation of plasma serotonin.

Introduction

Free serotonin circulating in the plasma is a potent vasoconstrictor and exerts profound effects on a wide range of organs and pathways,1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12 including the liver,² the cardiovascular system,^3,5,7,13, 14, 15, 16 glucose and lipid metabolism,¹² insulin secretion,¹⁷ brown adipose tissue,¹⁸ and the immune system.^9,19 Most of the serotonin in blood is supplied by the enterochromaffin cells in the gut but is rapidly cleared by sequestration into platelets and metabolic removal via hepatic and pulmonary pathways,20, 21, 22, 23 ensuring that, under normal physiological conditions, plasma-free serotonin remains in the low-to-subnanomolar range.²³ Nevertheless, transient elevations of plasma serotonin are crucial for acute physiological responses to stress or injury. For example, platelets, which store 99% of the circulating serotonin, release it upon activation to promote aggregation and vasoconstriction as part of the hemostatic response.⁷ Because of this, aberrant increases in plasma serotonin are implicated in vascular complications of various pathological conditions, including atherosclerosis, myocardial infarction, pulmonary arterial hypertension, lupus, and sepsis.24, 25, 26

The regulation of plasma serotonin is thought to be mediated largely by the serotonin transporter (SERT), which is expressed in a number of cell and tissue types, including neurons, platelets, pulmonary endothelial cells, enterocytes (ie, the intestinal lining), and hepatocytes.¹⁶ Because the same protein is expressed in both the brain and platelets,²⁷ the selective serotonin reuptake inhibitors (SSRIs), typically prescribed to inhibit SERT in the central nervous system (for example, in treating depression), also inhibit serotonin reuptake in the peripheral tissues, including blood.²⁸ However, despite serotonin’s critical role in cardiovascular physiology and the widespread therapeutic use of SSRIs, our understanding of how these drugs impact serotonin homeostasis in the blood remains limited, in no small part due to the significant technical challenges associated with monitoring plasma serotonin. Specifically, in addition to the relatively poor (typically tens of minutes) time resolution of blood draws, collection and sample preparation almost invariably activate platelets, causing artifactual release of the serotonin.²⁹ In contrast, microdialysis is less likely to cause the artificial release of the serotonin. We have, however, identified only 3 examples in the literature where microdialysis was used to monitor plasma serotonin dynamics. One of these reported elevated plasma serotonin in human patients with cancer receiving cisplatin³⁰; however, in another, plasma serotonin was observed to increase in animals during induced cerebrovascular thrombosis.³¹ Finally, a third study used repeated blood sampling followed by ex vivo microdialysis to examine the impact of selective and nonselective pharmacological inhibition of SERT on plasma serotonin concentrations.³² The temporal resolution of all 3 of these prior studies was 10 to 30 minutes,30, 31, 32 limiting the insights they might provide regarding the dynamics of serotonin release and clearance, which, as we show here, occur on a few-minutes time scale. In contrast, we investigated the kinetics of serotonin release and clearance after acute inhibition of SERT by the SSRI escitalopram using an electrochemical aptamer–based (EAB) sensor that can measure serotonin dynamics in situ in the veins of live animals with 8-second temporal resolution.

Methods

We designed serotonin sensors following protocols previously developed for both in vivo and ex vivo applications.^33,34 For ex vivo testing, we employed 0.2 mm diameter gold wire cut to a length of 4.5 cm, soldered to an electrochemical connector, and insulated with heat-shrink polyolefin tubing, leaving 3 mm of the wire exposed for sensing. We first cleaned the wires electrochemically by sweeping the potential from −1.0 V to −2.0 V (vs Ag/AgCl) at 1 V/s for 1000 cycles in 0.5M NaOH in a 3 electrode configuration. To enhance surface roughness and increase the active area, gold wires were subsequently roughened electrochemically in 0.5M H₂SO₄, using 0.02-second pulses between 0.0 V and 2.2 V for 100 cycles.

For in vivo experiments, we employed a modified electrode assembly consistent with previously reported configurations.³³ Specifically, we bundled the 3 mm gold working electrode (0.2 mm diameter) together with a 10 mm Ag/AgCl (0.13 mm diameter) reference electrode and a 6 mm platinum (0.13 mm diameter) wire counter electrode. These components were enclosed within a common protective tubing to enable stable, safe insertion into the jugular vein. This arrangement allowed real-time, 8-second resolution measurement of serotonin in anesthetized animals while minimizing motion artifacts and maintaining electrochemical control throughout the recording.

We functionalized the gold working electrodes with a serotonin-specific DNA aptamer (CGA CTG GTA GGC AGA TAG GGG AAG CTG ATT CGA TGC GTG GGT CG) that had been modified with a methylene blue redox reporter on its 5′ terminus and a thiol group on its 3′ terminus. Prior to immobilization, aptamers were reduced using 10mM tris (2-carboxyethyl) phosphin for 1 hour, diluted to 500nM in phosphate-buffered saline containing 2mM MgCl₂, and incubated with the cleaned gold electrodes for 1 hour to allow self-assembly via thiol-gold coupling. After immobilization, we passivated the electrodes overnight in 10mM 6-mercapto-1-hexanol. All fabrication steps were performed at room temperature and completed sensors were rinsed and stored in phosphate-buffered saline until use the following day.

We derived our calibration curves at 37°C in HEPES (N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid)–modified Tyrode solution, which is composed of 137mM NaCl, 5.4mM KCl, 1.8mM CaCl₂, 0.5mM MgCl₂, 0.16mM NaH₂PO₄, 3.0mM NaHCO₃, 5.5mM glucose, and 5.0mM HEPES (pH 7.4).

We performed our in vivo experiments using adult Sprague-Dawley rats (aged 2-6 months, weighing 300-600 g; Charles River Laboratories, Wilmington, MA). Animals were housed in pairs under standard vivarium conditions with regulated temperature and humidity, maintained on a 12-hour light/dark cycle, and given unrestricted access to food and water. All experimental protocols adhered to the National Institutes of Health Guide for the Care and Use of Laboratory Animals³⁵ and were approved by the Institutional Animal Care and Use Committee at the University of California, Santa Barbara, CA. All surgical procedures and in vivo deployments followed institutional guidelines for animal research.

Anesthesia was initiated with 4% isoflurane delivered in an induction chamber and maintained throughout the procedure at 2% to 3% via a nose cone. Vital signs, including oxygen saturation and heart rate, were continuously monitored using a pulse oximeter (Nonin Medical, Plymouth, MN). The animal’s chest area was shaved and prepped using alternating applications of 70% ethanol and povidone-iodine. Both left and right jugular veins were carefully exposed via small incisions. We then inserted a custom catheter consisting of a stainless-steel cannula connected to silastic tubing into the left jugular vein for drug and serotonin administration. The EAB sensor was inserted into the right jugular vein and secured in place using sterile 6-0 silk sutures (Fine Science Tools, Foster City, CA) to ensure sensor stability during the recording. To maintain patency of the infusion line, 30 units of heparin were administered immediately following catheter insertion.

We delivered serotonin hydrochloride (Alfa Aesar, Haverhill, MA; catalog no. B21261) IV at doses ranging from 0.01 to 0.5 mg/kg, depending on the experimental conditions. For all in vivo experiments employing escitalopram, we dosed IV at 1 mg/kg over 4 minutes using 0.005M stock solution of escitalopram oxalate (Sigma-Aldrich, St. Louis, MO; catalog no. 46234001). The dosing of both molecules was performed via the indwelling jugular catheter using a 3 mL syringe mounted on a motorized syringe pump (KDS 200, KD Scientific, Holliston, MA). The infusion duration was adjusted to deliver each dose consistently and precisely. Prior to each administration, the baseline was monitored until a stable signal was achieved (defined as a fluctuation of no more than ±5% over 15 minutes), presumably reflecting the clearance of any serotonin released from platelets in response to sensor insertion.

We interrogated our sensors using square wave voltammetry performed at square-wave frequencies of 30 Hz and 200 Hz on a CH1040C multipotentiostat (CH Instruments), sweeping from −0.1 V to −0.45 V vs Ag/AgCl. Specifically, we used alternating measurements at these 2 frequencies to perform kinetic differential measurements, a commonly employed in vivo drift correction mechanism.^36,37 When deployed in the rat jugular vein, the sensor exhibited significantly higher noise than seen in vitro, leading to an estimated limit of detection of 25nM in the former placement. We defined this limit as the serotonin concentration in Tyrode buffer that corresponds to a kinetic differential measurement value 3 times the standard deviation of the baseline noise seen in vivo.

We performed our ex vivo measurements in whole blood as follows. We performed ex vivo measurements in whole blood collected from rats previously used in unrelated experiments, in accordance with animal use reduction principles. Blood was drawn via an indwelling catheter placed in the jugular vein, using the same surgical and sampling setup employed for in vivo studies. To prevent coagulation, the withdrawn blood was immediately transferred into sterile tubes preloaded with a small heparin solution at a volume ratio of ∼9:1. These animals had also received systemic heparin prior to blood collection. Samples were promptly placed in a 37°C water bath, and all subsequent measurements were conducted at this temperature to mimic physiological conditions.

Results

Our approach to the seconds-resolved in situ measurement of plasma serotonin employs a previously developed, in vivo EAB (Olena Lukoyanova, Ruben W. Kolkman, Nicole A. Emmons, Tod E. Kippin, and Kevin W. Plaxco, unpublished data, January 2026) sensor (Figure 1) that employs a 42-base aptamer of 20 nanomolar sensitivity and excellent selectivity against such potentially important interferents as 5-hydroxytryptophan and 5-hydroxyindoleacetic acid.³⁸ The resulting sensor achieves both excellent sensitivity over clinically relevant plasma serotonin concentrations (Figure 1B) and, as deployed here, physiologically relevant, 8-second time resolution (the latter is defined by the time required to collect the requisite pair of voltammograms). Critically for our pharmacological SERT inhibition studies, the sensor does not respond significantly to the SSRI escitalopram at 80μM, a concentration well above the highest possible plasma concentration expected in our experiments (Figure 1B).

Figure 1.

Sensor characterization. (A) Here, we have used a previously described EAB sensor targeting serotonin. Here, a serotonin-binding aptamer is tethered to a monolayer-coated gold electrode, and an MB redox reporter is attached to its distal terminus. Upon serotonin binding, the aptamer is thought to fold, altering electron transfer kinetics between the redox reporter and the electrode surface and producing an easily measured signal using square wave voltammetry. (B) The calibration curve we obtained in modified Tyrode buffer is shown and was used to convert the sensor output (calculated via KDM)^36,37 to concentrations (blue). Also the sensor’s response to Esc (pink) is shown. Even at the highest conceivable concentration the drug might reach in our experiments (ie, assuming no distribution into the tissues, no protein binding, and no elimination, such that the entire dose remains solely in the 35 mL/kg plasma volume of the rat),³⁹ the error induced by potential cross reactivity would amount to only 20nM serotonin, a level far lower than the concentrations we observe after Esc dosing. Esc, escitalopram; KDM, kinetic differential measurements; MB, methylene blue.

As our animal model, we employed anesthetized Sprague–Dawley rats, in which we placed a sensor in the right jugular vein and a drug delivery catheter in the left jugular vein. A concern is that, because platelets participate in the hemostatic response, vascular injury near the insertion site could cause initial serotonin concentrations to elevate. To minimize this, we waited 1.5 to 2 hours after venous insertion of the sensor before initiating measurements. Of note, under normal physiological conditions the free serotonin concentration in animal plasma is subnanomolar,^23,29 a value well below the sensor’s ∼25nM limit of detection in the jugular vein. We therefore assigned 0nM concentration to our stabilized baseline measurements and used this value as the single-point calibration for performing kinetic differential measurement drift correction.³⁷

The serotonin EAB sensor easily measures the rapid rise and subsequent clearance of exogenously introduced serotonin in systemic circulation. To confirm this, we challenged 2 animals with IV serotonin and recorded the resulting rise and recovery of its plasma concentration (Figure 2A-B). Of note, by injecting serotonin directly into the jugular vein, we bypass gut-derived serotonin release and hepatic first-pass metabolism, allowing us to isolate and examine the clearance mechanisms acting specifically on circulating serotonin in the bloodstream. In animal 1 (Figure 2A), doses <0.5 mg/kg were cleared rapidly, consistent with SERT-mediated uptake by platelets, vascular endothelium, liver, and lungs. At 0.5 mg/kg, however, we observed a substantial plasma serotonin peak of ∼170nM serotonin, presumably because under this challenge, the rate of delivery into the plasma exceeded its rate of transport into platelets. Note, a simple redistribution model based on estimated extracellular fluid volumes predicts that, in the absence of uptake by platelets and tissues, peak serotonin concentrations would be 70-fold (ie, assuming distribution is limited to the 0.035 L/kg of plasma volume^39,40) to 390-fold (ie, 12μM assuming complete distribution into the 0.194 L/kg of plasma plus interstitial fluid) higher than the value we observed. Consistent with this, the transient serotonin peak was rapidly cleared, exhibiting a half-life of ∼2.6 minutes. In animal 2, sequential 0.5 mg/kg injections produced similar transient peaks, supporting the hypothesis that these concentrations approach uptake saturation. Of note, the second peak, which exhibited a 1.8-minute elimination half-life (supplemental Figure 1), was substantially higher than the first, again suggestive of a cumulative saturation of SERT-mediated clearance pathways.

Figure 2.

Serotonin is rapidly cleared from circulation, preventing sustained accumulation even after rather large challenges with the molecule. (A-B) Serotonin concentration-time courses in 2 animals following IV serotonin challenges of the indicated doses (0.01-0.5 mg/kg) are shown. The blue-shaded regions denote the duration of the infusions. Note that sensor noise progressively increased with measurement duration, presumably due to the time-dependent instability of the immobilized DNA aptamer on the probe surface in blood.^41,42

Acute IV administration of the SSRI escitalopram produces a large, transient increase in plasma-free serotonin. To confirm this, we first dosed 4 animals with the drug at 1 mg/kg, which, given its 12 L/kg distribution volume,⁴³ corresponds to a concentration of ∼260μM, a value well above its 6.6nM inhibition.⁴⁴ In response to this challenge, plasma-free serotonin concentrations rose to between ∼30 and ∼140nM (Figure 3) then fell back to baseline with half-lives of 8 to 14 minutes (supplemental Figure 2). In contrast to the first dose of escitalopram, however, a second, equal dose delivered 30 to 68 minutes after the first failed to produce any detectable change in plasma-free serotonin in any of our 4 animals, consistent with SERT remaining saturated with the drug, which has a >1-hour half-life in rats.45, 46, 47

Figure 3.

Acute inhibition of SERT by Esc leads to a transient increase in serotonin. The figure shows plasma serotonin concentration–time courses in 4 animals after IV Esc administration. The time of the first 1 mg/kg Esc injection is designated as time zero. A subsequent 1 mg/kg Esc injection was administered at different time points for individual animals, marked by color-coded shaded regions corresponding to their respective response curves. No similar transient is seen upon this second challenge for any animal, suggesting that SERT inhibition remained effectively complete due to the first drug challenge.

Acute escitalopram treatment also slows plasma serotonin clearance. To check this, we first challenged 6 animals with IV serotonin at 0.05 mg/kg. In all 6 animals, plasma serotonin was cleared so rapidly that our sensors detected little to no increase, with levels remaining below ∼50nM (Figure 4). We then administered IV escitalopram at 2 mg/kg to 3 animals and vehicle control to the remaining 3. As expected, escitalopram produced a rapid, transient rise in plasma-free serotonin to >100nM, whereas vehicle administration produced no detectable change. We then performed a second serotonin challenge in all 6 animals. In the escitalopram-treated group this challenge produced a marked, transient increase in plasma-free serotonin, with levels exceeding 150nM (Figure 4A). In contrast, the second serotonin challenge produced little to no detectable increase in plasma-free serotonin in the vehicle-treated control group (Figure 4B).

Figure 4.

Acute inhibition of SERT by Esc impairs serotonin clearance. (A) To test this, we challenged 3 animals with 50 μg/kg of exogenous serotonin, followed by 2 mg/kg Esc and then a second 50 μg/kg exogenous serotonin. Although no significant peak was seen after the first serotonin challenge, after Esc treatment the same dose challenge caused a large serotonin peak in all 3 animals. (B) In contrast, in 3 animals treated with a vehicle injection rather than Esc, no such peak was observed.

Although the effects of SERT inhibition on the dynamics of plasma serotonin are significant, they do not entirely abolish plasma serotonin removal. To quantify this, we next compared the response to serotonin challenges performed before and after escitalopram or vehicle administration to derive a post/pre-response ratio (Figure 5). Specifically, we defined the injection-evoked response as the maximum serotonin concentration reached within 15 minutes after injection minus the basal serotonin concentration measured 20 seconds before injection. The mean post/pre-response ratio was 17-fold higher in escitalopram-treated animals than in vehicle-treated controls (Figure 5). Together with our earlier findings, this indicates that acute SERT inhibition with escitalopram transiently increases baseline plasma-free serotonin and slows the clearance of exogenous plasma serotonin challenges. Nevertheless, even after such SERT blockade, plasma serotonin is cleared, albeit more slowly than in the absence of such blockade. This residual clearance suggests that non-SERT mechanisms provide additional clearance capacity and prevent sustained serotonin accumulation in plasma.

Figure 5.

Esc increases the serotonin injection–evoked plasma-free serotonin response. Quantification and statistical analysis of the serotonin responses are shown in Figure 4. The serotonin injection–evoked response was defined as ΔC = C_max − C_baseline, where C_max was the maximum plasma free serotonin concentration observed at any time within 15 minutes after serotonin injection, and C_baseline was the concentration measured ∼20 seconds before the start of injection. The post/pre-response ratio was calculated as ΔCpost/ΔCpre, where “post” and “pre” refer to serotonin challenges after and before Esc or vehicle administration, respectively. Rats received serotonin at 0.05 mg/kg IV before and after treatment with Esc at 2 mg/kg IV or vehicle control. Red circles indicate individual animals; horizontal bars and error bars indicate mean and 95% confidence interval (n = 3 animals per group).

In contrast to the situation seen in vivo, escitalopram blocks detectable serotonin uptake in vitro in fresh, whole rat blood at 37°C, suggesting that the recovery seen in escitalopram-treated animals is due to nonblood pathways. To check this, we used a serotonin challenge–escitalopram treatment–repeat serotonin challenge design analogous to the in vivo experiments described in Figure 4A. Fresh rat blood was diluted at a 1:1 volume ratio with Tyrode-HEPES buffer to stabilize pH, reduce viscosity, and facilitate rapid mixing. Before escitalopram exposure, exogenous serotonin additions were rapidly cleared, such that free serotonin concentrations returned to near-baseline levels within 2 minutes (Figures 6A-C). Vehicle-only control challenges performed either before or after the initial addition of serotonin produced no measurable increase in free serotonin, indicating that the mixing alone did not artifactually elevate the sensor signal. Escitalopram treatment at 500nM increased baseline free serotonin concentrations by 50 to 100nM. The second identical serotonin challenge caused free serotonin concentrations to increase further and, unlike the response observed in vivo, remain elevated (Figure 6A-B). It thus appears that inhibition of SERT abolishes the ability of whole blood to clear serotonin from plasma. We repeated this experiment a third time using 50nM escitalopram and found that, despite the 10-fold lower inhibitor concentration, serotonin uptake remained substantially slowed (Figure 6C). Finally, we performed a similar experiment using undiluted whole blood (supplemental Figure 3A). In this, we challenged vehicle- and escitalopram-treated aliquots from the same blood sample with serotonin, confirming that the reduced serotonin clearance observed after escitalopram treatment is not due to prolonged ex vivo handling or differences in experimental timing.

Figure 6.

SERT inhibition in ex vivo whole blood samples impairs serotonin uptake by platelets. (A-C) Representative ex vivo experiments performed using blood samples from 3 independent animals. Administration of 50nM Esc (C), administration of 500nM Esc (A,B). (A) Serotonin concentration–time courses collected in fresh, whole rat blood diluted at a 1:1 volume ratio with Tyrode-HEPES buffer and maintained at 37°C. (i) Serotonin was added and mixed by pipetting, producing a transient increase in free serotonin followed by rapid clearance to near-baseline concentrations within 2 minutes. (ii) A vehicle mixing control, consisting of buffer matching the volume of the serotonin delivery buffer, also mixed by pipetting, did not increase free serotonin. (iii) Esc was added at 500nM, after which free serotonin increased gradually. (iv) Repeat serotonin challenge after Esc treatment produced a sustained elevation in free serotonin, consistent with inhibition of platelet SERT–mediated serotonin uptake. Filled and open circles represent simultaneous recordings from sensor 1 and sensor 2, respectively. (B) Serotonin concentration–time courses collected in fresh, whole rat blood diluted at a 1:1 volume ratio with Tyrode-HEPES buffer and maintained at 37°C. (i) A vehicle mixing control, consisting of buffer matching the volume of the serotonin delivery buffer, was added and mixed by pipetting. (ii) Serotonin was added and mixed by pipetting, producing a transient increase in free serotonin followed by rapid clearance to near-baseline concentrations within 5 minutes. (iii) Esc was added at 500nM, after which free serotonin increased gradually. (iv) Repeat serotonin challenge after Esc treatment produced a sustained elevation in free serotonin, consistent with inhibition of platelet SERT–mediated serotonin uptake. Recordings were done using 1 sensor. (C) (i) A vehicle mixing control, consisting of buffer matching the volume of the serotonin delivery buffer, was added and mixed by pipetting. (ii) Serotonin was added and mixed by pipetting, producing a transient increase in free serotonin followed by rapid clearance to near-baseline concentrations within 5 minutes. (iii) Esc was added at 50nM, after which free serotonin increased gradually. (iv) Repeat serotonin challenge after Esc treatment produced a very slow decrease in free serotonin, consistent with significant but not complete inhibition of platelet SERT–mediated serotonin uptake. Values indicate the mean and standard deviations of measurements collected using 4 sensors.

Discussion

Despite extensive study, key questions regarding the dynamics and mechanisms of serotonin homeostasis in blood have remained unanswered and, for the most part, even unaddressed. By leveraging the unprecedented temporal resolution of in vivo EAB sensors, we have answered these questions using studies performed both in vivo in the circulatory systems of live, intact animals and ex vivo in fresh, whole blood. In doing so, we found that (1) IV serotonin challenges are cleared within a few minutes and (2) pretreatment with an SSRI inhibitor causes both a rapid, transient rise in plasma-free serotonin and a reduction in the rate of clearance of exogenously dosed serotonin. Even in the latter case, however, plasma serotonin eventually returns to near basal levels. In contrast, the ex vivo treatment of fresh blood with the same SSRI entirely abolishes the clearance of plasma serotonin. Together these observations indicate that, although SERT-driven clearance dominates the dynamics of plasma serotonin both in vivo (ie, in blood in situ in the jugular vein of a live animal) and ex vivo (in whole blood at 37°C), SERT-independent, nonblood pathways continue to remove plasma serotonin in the body even when SERT is effectively and completely inhibited.

Exogenous serotonin challenges are rapidly eliminated in animals with fully functional (uninhibited) SERT, presumably as a means of preventing the molecule’s supraphysiological accumulation. At lower challenge doses (0.01-0.05 mg/kg), for example, serotonin clearance is so rapid that we detect little, if any, increase in the plasma concentration (Figure 2A). At higher challenge doses (eg, 0.5 mg/kg), we detect a transient rise in plasma serotonin concentration, presumably because SERT becomes saturated or platelet storage capacity is exceeded, allowing for temporary accumulation of serotonin in plasma (Figure 2A-B). We speculate that at these higher concentrations, although SERT becomes saturated, serotonin elimination remains relatively rapid, with clearance occurring in just 15 to 30 minutes. Consistent with this, the maximum concentrations we measured were ∼ 450- to 900-fold lower than those theoretically expected in the absence of clearance by platelet uptake or other mechanisms (ie, the dose divided by the expected plasma volume). Together, these observations confirm that the circulatory system maintains exceptionally tight regulation of plasma free serotonin, presumably through platelet uptake and hepatic and pulmonary clearance.⁴⁸ Given serotonin’s potent vasoactive properties and the widespread expression of high-affinity serotonin receptors (eg, 5-HT₁, 5-HT₂, and 5-HT₇) throughout the circulatory system,^13,49, 50, 51 this stringent control is perhaps not surprising.

The IV administration of the SSRI escitalopram induces a rapid, transient increase in endogenous plasma serotonin (Figure 3), indicating that SERT inhibition causes a short-term disruption of plasma serotonin homeostasis. The magnitude and duration of this effect varied slightly between animals (Figure 3), likely reflecting individual differences in serotonin clearance capacity. A second identical escitalopram dose, however, does not produce any measurable change in plasma serotonin, likely because the initial, 1 mg/kg dose is sufficient to fully inhibit SERT. This argument is supported by both SERT’s high affinity for escitalopram (K_i = 6.6nM)⁴⁴ and the drug’s >1-hour half-life in rats.45, 46, 47 We observed that the transient rise in plasma serotonin is consistent with findings of previous studies using the SSRI fluoxetine, which also observed a transient elevation in plasma serotonin.³² However, the orders-of-magnitude poorer time resolution of the prior study, which employed microdialysis with a 15-minute sampling period, renders direct comparisons with our results little more than qualitative.

As expected, SERT-mediated uptake by platelets, and potentially other SERT-dependent pathways, is essentially abolished following SSRI treatment. Specifically, serotonin clearance is much slower in escitalopram-treated animals than in animals with uninhibited SERT function. For example, although in the absence of SSRI a 0.05 mg/kg serotonin challenge is cleared so rapidly that it is not detectable, the same challenge elicits a pronounced plasma serotonin peak in escitalopram-treated animals (Figure 4). Similarly, the 0.05 mg/kg challenge produces serotonin peaks in SSRI-treated animals that are 2 to 4 times higher than those observed after 10-fold larger challenges (0.5 mg/kg) in SSRI-naïve animals (Figure 5), which then fell to baseline with a half-life of ∼10 minutes.

In contrast to the situation seen in vivo, where it slows but does not abolish serotonin clearance, escitalopram entirely abolishes serotonin clearance in fresh, whole blood in vitro (Figure 6). Together these observations suggest that, in a live, intact animal, a secondary, SERT-independent (or at least escitalopram-insensitive), nonblood pathway is present that can clear serotonin from the plasma only a few-fold more slowly than the escitalopram-sensitive pathways. Given that this effect arises immediately after escitalopram treatment, it appears that, rather than emerging as long-term compensatory adaptations, the pathway underlying it appears to be operating in real time. Having said that, although our data clearly point to the existence of a novel, SERT-independent clearance pathway (or pathways), the specific molecular mechanisms for this clearance remain unknown. Possible mediators of this clearance, however, include low-affinity, high-capacity organic cation transporters (OCTs) and plasma-membrane monoamine transporter (PMAT), which transport monoamines and are broadly expressed in peripheral organs, including the liver and kidney.52, 53 Notably, the absence of OCT and PMAT expression in platelets^54,55 is consistent with our observation that, when SERT is blocked by SSRI treatment ex vivo, whole blood fails to clear serotonin from plasma. Just as OCTs and PMAT operate alongside SERT in brain and gut,^27,56, 57, 58, 59 our results raise the possibility that they serve a complementary clearance role in the circulatory system. Future in vivo studies employing targeted genetic or pharmacological inhibition of candidate OCT isoforms and PMAT (eg, liver- or kidney-specific knockouts) will be needed to confirm their role and quantify their relative contributions.

The existence of the secondary serotonin clearance pathways may help explain why chronic SERT inhibition, whether pharmacological (via SSRIs) or genetic in origin, leads to depletion of platelet serotonin without resulting in pathologically elevated plasma-free serotonin.60, 61, 62, 63 Such compensatory capacity underscores the resilience of the circulatory system in maintaining plasma serotonin homeostasis and, in turn, vascular and hematologic homeostasis.

Conflict-of-interest disclosure: K.W.P. owns equity in and consults for a company that is commercializing in vivo electrochemical aptamer–based sensors. Following the completion of this work, N.A.E. became an employee at a company that is commercializing aptamer-based sensors. The remaining authors declare no competing financial interests.