|Home | About | Journals | Submit | Contact Us | Français|
Uptake resolved by high-speed chronoamperometry on a second-by-second basis has revealed important differences in brain serotonin transporter function associated with genetic variability. Here, we use chronoamperometry to investigate variations in serotonin transport in primary lymphocytes associated with the rhesus serotonin transporter gene-linked polymorphism (rh5-HTTLPR), a promoter polymorphism whose orthologues occur only in higher order primates including humans. Serotonin clearance by lymphocytes is Na+-dependent and inhibited by the serotonin-selective reuptake inhibitor paroxetine (Paxil), indicative of active uptake by serotonin transporters. Moreover, reductions in serotonin uptake rates are evident in lymphocytes from monkeys with one or two copies of the short ‘s’ allele of the rh5-HTTLPR (s/s < s/l < l/l). These findings illustrate that rh5-HTTLPR-related alterations in serotonin uptake are present during adulthood in peripheral blood cells natively expressing serotonin transporters. Moreover, they suggest that lymphocytes can be used as peripheral biomarkers for investigating genetic or pharmacologic alterations in serotonin transporter function. Use of boron-doped diamond microelectrodes for measuring serotonin uptake, in contrast to carbon fiber microelectrodes used previously in the brain, enabled these high-sensitivity and high-resolution measurements. Boron-doped diamond microelectrodes show excellent signal-to-noise and signal-to-background ratios due mainly to low background currents and are highly resistant to fouling when exposed to lymphocytes or high concentrations of serotonin.
Microelectrode voltammetry, which comprises a group of related electrochemical techniques, is a rapidly growing means to explore chemical neurotransmission (1). The expanding use of voltammetry for neurochemical studies is largely due to advantages associated with the ability to perform in situ measurements with excellent spatial and temporal resolution. Moreover, strategies for parlaying electrochemical detection of nonelectroactive neurotransmitters have been devised (2,3). The benefits of making direct measurements in tissue are highly evident in recent work using high-speed chronoamperometry to investigate changes in serotonin uptake rates associated with decreases in serotonin transporter (SERT) gene expression. Reduced brain serotonin uptake associated with the loss of one functional copy of the SERT gene in mice is readily observed using chronoamperometry ex vivo(4) and in vivo(5). By contrast, these changes are not differentiable using standard radiochemical methods (6,7). Furthermore, maximal uptake rates appear to be underestimated, while the affinity of serotonin for its transporter is overestimated using traditional [3H]-serotonin uptake. We have attributed this to a loss of transported serotonin occurring during the filtration process employed in radiometric methods (6). The radiochemical assay also suffers from shortcomings associated with estimating uptake rates using a single time point, in contrast to chronoamperometry and other voltammetry methods, which provide information on neurotransmitter transport over the entire time course with second to millisecond resolution.
Carbon fiber microelectrodes (CFMs) have most often been used for voltammetry measurements of neurotransmitters (8). Advantages associated with this type of microelectrode include the commercial availability of carbon fibers with diameters ranging from 3 to 30 μm and the ease of fabrication of small cylindrical or disk-shaped microelectrodes providing spatially resolved measurements in the brain (9). Despite these advantages, CFMs have limitations associated with sp2-hybridized carbon and the presence of surface hydroxyl groups, both of which promote strong adsorption of polar analytes via dipole−dipole, ion−dipole, and hydrogen bonding interactions (10−13). Additional problems arise when using CFMs to detect serotonin, which forms an unstable oxidation byproduct that can dimerize (14,15). These products adhere to CFM surfaces causing electrode fouling (16). Thus, CFMs are not ideal for measurements of serotonin at concentrations greater than 1 μM (17) or for longer exposures to biological preparations.
Boron-doped diamond is a highly stable sp3-hybridized form of carbon that lacks an extended π-electron system and whose surface is dominated by hydrogen-termination (18−22). Together, these characteristics impart electrodes fabricated from boron-doped diamond with a nonpolar surface chemistry making them more resistant to fouling (23). Serotonin release by enterochromaffin cells in the intestinal mucosa, where tissue serotonin concentrations are estimated to be in the micromolar range (24), has been measured using boron-doped diamond microelectrodes (BDMs) and constant potential amperometry (23,25). Here differences in uptake associated with lower levels of SERT expression in neonatal intestinal mucosa were differentiated from those in adult gut tissue (26).
The serotonin transporter is a major research focus because of its primary role in regulating the neuromodulatory effects of serotonin in the central and peripheral nervous systems (27,28). The SERT is the primary molecular target for the serotonin-selective reuptake inhibitors (SSRIs), which are commonly prescribed drugs used in the treatment of depression and anxiety disorders. Widely abused drugs, including 3,4-methylenedioxymethamphetamine (MDMA; Ecstasy) and cocaine, also have primary mechanisms of action at the SERT (7,29,30).
Beyond its pharmacologic significance, a number of commonly occurring polymorphisms have been discovered in the noncoding regions of the human SERT gene (31−35). Additionally, rare coding region single nucleotide polymorphisms (SNPs) have been pinpointed (36), some of which are of psychiatric relevance (35,37,38). A 43-base pair insertion/deletion promoter polymorphism, termed the human serotonin transporter gene-linked polymorphic region (h5-HTTLPR), is the most extensively studied SERT gene variant (Figure (Figure1)1) (31). This polymorphism has been linked to differences in anxiety-related personality traits, vulnerability to developing neuropsychiatric disorders, and variability in drug responsiveness (28,31,39−41). The h5-HTTLPR is hypothesized to influence brain function and behavior via allele-specific SERT promoter activity (42,43). Decreases in SERT mRNA (31) and protein binding in postmortem human brain (44) have been associated with the lower expressing short ‘s’ allele. However, not all studies are in agreement. For example, SERT mRNA levels in human postmortem raphe tissue (45) and in vivo human SERT binding measured by positron emission tomography (PET) (46) have failed to show h5-HTTLPR-associated allele specific differences.
Rhesus monkeys (M. mulatta) express a serotonin transporter gene-linked polymorphic region, the rh5-HTTLPR, which is evolutionarily related to the h5-HTTLPR(47,48). The rh5-HTTLPR is a 21 base pair insertion/deletion promoter polymorphism that occurs at a locus slightly shifted from that in humans. Serotonin transmission, stress responsiveness, and social behavior are influenced by the rh5-HTTLPR49−52, similar to associations between the h5-HTTLPR and human behavior. At the molecular level, the rh5-HTTLPR is also hypothesized to affect transcription of SERT. However, its influence on SERT mRNA, protein levels, and serotonin uptake has not been determined, and only its effects in a reporter gene assay have been described (53).
The close parallels between the h5-HTTLPR and rh5-HTTLPR provide opportunities to use the latter to advantage for investigating this evolutionarily conserved polymorphism, which is absent in lower order mammals including rodents (47). For example, the h5-HTTLPR possesses additional functional promoter region SNPs (32−34), multiplying the number of genotypes and complicating their interpretation (54). By contrast, orthologous SNPs have not been reported in the rh5-HTTLPR. Furthermore, establishing the relationship between the effects of the rh5-HTTLPR in peripheral tissues vs the brain is possible. By contrast, with the exception of in vivo imaging techniques, which currently have low resolution, making these types of connections in living humans are difficult.
Here, we take a key step toward understanding the effects of the rh5-HTTLPR on SERT function. We employ high-speed chronoamperometry to explore serotonin uptake in nontransformed primary rhesus lymphocytes, which natively express low levels of SERT. We investigate whether BDMs have advantages over CFMs for measuring serotonin uptake in these cells. Furthermore, we test the hypothesis that rh5-HTTLPR genotype is associated with variations in serotonin uptake rates in adult animals. On the basis of our findings, we anticipate that readily obtainable peripheral blood lymphocytes, in combination with sensitive electrochemical methods, will provide opportunities for future studies aimed at understanding the effects of gene variants, neurological disease states, and drug treatment on SERT function.
Cyclic voltammetry was used to compare background and serotonin-related currents at BDMs vs CFMs. Figure Figure22 depicts representative cyclic voltammetric i-E curves obtained at 1 V/s in assay buffer (background) or 10 μM serotonin. Boron-doped diamond microelectrodes exhibited 6-fold lower background currents on average compared to those of CFMs, even though the surface areas of BDMs (~30,000 μm2) were larger than CFM surface areas (~11,000 μm2). Lower background currents at BDMs have been attributed to the absence of redox-active or ionizable surface groups (reduced pseudocapacitance) and slightly lower concentrations of internal charge carriers (reduced capacitance) (19). The mean oxidation to background current ratio produced by 10 μM serotonin at BDMs (0.58 ± 0.03) was significantly higher than that at CFMs (0.13 ± 0.03; P < 0.0001).
For serotonin, the half-wave oxidation potential (E1/2ox) at BDMs (0.54 ± 0.008 V) was significantly greater than that at CFMs (0.32 ± 0.005 V; P < 0.0001). Oxidation of monoamine neurotransmitters proceeds via an inner-sphere electron transfer pathway; thus, electron transport kinetics are highly sensitive to electrode surface characteristics (18,19). Adsorption of serotonin on electrode surfaces facilitates charge transfer leading to an increase in the heterogeneous electron transfer rate constant and faster kinetics. A positive shift in the oxidation potential for serotonin at BDMs vs CFMs has been hypothesized to be due to slower reaction kinetics resulting from lower adsorption of serotonin at BDM surfaces. Similar increases in E1/2ox at BDMs compared to CFMs have also been reported for norepinephrine (55).
These cyclic voltammetry data suggested that oxidation potentials higher than those typically used with CFMs would be needed to produce maximal serotonin currents when employing BDMs in combination with chronoamperometry. To investigate this, integrated currents produced by 1 μM serotonin at different oxidative step potentials were evaluated, while the reductive and resting potentials were held constant at 0 V. Varying Eox from 0.3 to 0.7 V in 0.1 V increments produced significant increases in integrated current that reached a plateau at 0.8 V (P < 0.01 for steps below 0.8 V vs 0.8 V), after which no significant increases in integrated current were observed (P > 0.05 for steps above 0.8 V vs 0.8 V) (Figure (Figure3A).3A). Increasing Eox to values higher than 0.8 V resulted in significant decreases in signal-to-noise ratios (P < 0.01 for 1 V vs 0.8 V) and signal-to-background ratios (P < 0.01 for 0.9−1 V vs 0.8 V) (Figure (Figure3A)3A) mainly attributable to increases in noise and background currents, respectively. Together, these results suggest that an Eox of 0.8 V produces the highest integrated oxidation current while also maximizing signal-to-noise and signal-to-background ratios for detecting serotonin at BDMs by chronoamperometry.
Next, we compared the responses of BDMs vs bare CFMs to serotonin concentrations ranging from 0.5 to 10 μM (Figure (Figure3B).3B). Boron-doped diamond microelectrodes showed highly linear responses over the entire concentration range investigated (R2 = 0.9908). By comparison, CFMs showed lower linearity (R2 = 0.9579) with a better fit to a nonlinear curve (one-site exponential, R2 = 0.9789). Analysis of the electrode responses at each serotonin concentration showed significantly lower responses of CFMs compared to BDMs (P < 0.01) at 2.5 μM and higher, probably due to fouling of CFM surfaces.
Carbon fiber microelectrodes have been used extensively for in vivo and ex vivo detection of neurotransmitters, and application of Nafion to CFMs has been shown to increase neurotransmitter current, signal-to-background ratios, and limits of detection (56−58). Boron-doped diamond, which is still in its early stages for use in monoamine neurotransmitter sensing, has been compared to glassy carbon (20,59) or bare CFMs (23). Here, we contrasted BDMs with both bare and Nafion-coated CFMs for measuring serotonin using chronoamperometry by assessing four key electrode properties: signal (nC/mm2), signal-to-noise, signal-to-background (%), and limits of detection (nM) (Table 1). With the exception of the signal itself, which was only modestly higher at BDMs vs bare CFMs (P < 0.05), BDMs showed highly improved figures of merit (P < 0.001 vs CFMs and Nafion-coated CFMs), further indicating that boron-doped diamond is a superior electrode material for serotonin sensing compared to carbon fiber primarily based on lower background currents and noise.
Carbon-fiber electrode surfaces show adsorption of polar analytes such as serotonin and its oxidation products (14,15). Furthermore, biological macromolecules adsorb to CFM surfaces, and both of these processes cause fouling, thus altering time-dependent amperometric responses to analytes (60,61). Commonly used techniques to reduce CFM fouling involve coating electrodes with anionic polymers, e.g., Nafion, or cleaning with organic solvents such as isopropanol (57,58,62,63). By contrast, BDMs exhibit minimal surface adsorption, primarily because of the absence of an extended π-electron system and a lack of polar carbon−oxygen surface groups.
Previously, Patel et al. investigated the ability of BDMs to resist fouling due to brief exposure (10 s) to high concentrations of serotonin (10 μM) comparing this to bare CFMs (23). Here, we extended those findings using conditions relevant to the present experiments, i.e., exposure to 10 μM serotonin or cell suspensions for 20 min, the time frame over which serotonin uptake occurs in the ex vivo experiments described below. Additionally, we compared the fouling properties of BDMs and CFMs to Nafion-coated CFMs. Boron-doped diamond microelectrodes showed significantly reduced fouling compared to both types of carbon fiber electrodes in the presence of high concentrations of serotonin or lymphocyte suspensions, as shown in Figures Figures4A4A and and4B,4B, respectively. Exposure to 10 μM serotonin for 20 min resulted in significant decreases (P < 0.001) in electrode sensitivity for bare (−55 ± 5%) or Nafion-coated (−40 ± 7%) CFMs, whereas BDMs showed nonsignificant decreases in sensitivity that were less than 10% (Figure (Figure4A).4A). Cleaning BDMs with isopropanol led to a nearly complete recovery of sensitivity (97 ± 4%) indicating that serotonin reaction products are weakly adsorbed on these electrode surfaces (Figure (Figure4A).4A). By contrast, cleaning CFMs or Nafion-coated CFMs in isopropanol for 5 min did not significantly regenerate sensitivity (P > 0.05 vs after fouling).
Similarly, exposing electrodes to lymphocytes suspended in assay buffer (2 million cells/mL) for 20 min resulted in significant decreases (P < 0.001) in sensitivity for bare CFMs (−60 ± 5%) and Nafion-coated CFMs (−35 ± 5%), compared to a nonsignificant 8.0 ± 5% decrease in sensitivity for BDMs (Figure (Figure4B).4B). Cleaning electrodes with isopropanol after biological fouling produced results similar to those obtained after exposure to serotonin, with only BDMs being completely regenerated (100 ± 2%). On the basis of these data and prior work in this developing field (23,25,26,64), it is evident that BDMs are significantly more resistant to fouling then CFMs, even when using common pre- and post-treatments for the latter. Thus, BDMs appear to be uniquely suited for measuring serotonin at higher concentrations and over longer time frames in biological preparations.
On the basis of our previous work in mouse brain synaptosomes (4,6), we anticipated that concentrations of serotonin >2 μM might be necessary to achieve maximal uptake rates in lymphocytes, hence our choice of higher concentrations of serotonin (10 μM) for the fouling study and upper limits of the calibration curves in the experiments described above. However, when we carried out initial uptake experiments in rhesus lymphocytes using BDMs and chronoamperometry, the data indicated that serotonin uptake in these cells is lower and that maximal uptake rates occurred at concentrations of serotonin <2 μM (see below). Although, CFMs show linear responses to serotonin at lower concentrations, the performance of BDMs was significantly better across the board with respect to responses to serotonin, reduced fouling, and reusability. Thus, we chose to use BDMs over CFMs or Nafion-coated CFMs for the experiments that follow.
Transport of serotonin by SERT depends on sodium, chloride, and potassium concentration gradients maintained across the plasma membrane directly or indirectly by the Na+/K+-ATPase. Initially, substrate binding sites at SERT are accessible to the extracellular fluid allowing SERT to bind serotonin, Na+, and Cl− in a 1:1:1 stoichiometry (65). This leads to a conformation change, which results in the translocation of serotonin (and Na+ and Cl−) across the plasma membrane followed by its release into the cytoplasm. An intracellular potassium ion then binds to the SERT causing confirmation reversal. On the basis of this model of transport, a Na+ gradient maintained across the plasma membrane ([Na+]out [Na+]in) is necessary to energetically drive the active transport of serotonin by SERT.
To investigate whether time-dependent decreases in amperometric current after the addition of serotonin to lymphocytes are due to an active transport process, we substituted equimolar LiCl for NaCl in the assay buffer and determined its effects on current with respect to time. Representative oxidative currents reflecting serotonin uptake by lymphocytes suspended in normal assay buffer containing 150 mM Na+ vs assay buffer devoid of Na+ are compared in Figure Figure5A5A and B, respectively. In the presence of sodium, suspensions of rhesus lymphocytes (2−4 million cells/mL) cleared 0.5 μM serotonin in ~15 min, as evidenced by a return of the oxidative current to baseline (Figure (Figure5A).5A). The mean uptake rate for pooled mixed-genotype lymphocytes was 3.3 ± 0.5 pmol serotonin/million cells-min (N = 6). By contrast, in the absence of extracellular sodium, the current remained constant with respect to time over the course of 15 min after the addition of 0.5 μM serotonin (Figure (Figure5B).5B). The latter was repeated using three independent lymphocyte samples and three different BDMs, and in each case, the experiment produced similar results. These results support the hypothesis that decreases in oxidative current after serotonin addition to lymphocytes are due to a Na+-dependent active uptake mechanism.
Other Na+- dependent plasma membrane transporters putatively expressed by lymphocytes, such as the dopamine transporter, have the capacity to transport serotonin, albeit with considerably lower affinity for serotonin than SERT (66,67). Paroxetine, a high affinity SSRI (Ki = 10 nM for SERT) (68), was used to investigate whether decreases in current after the addition of serotonin to lymphocytes are specifically due to SERT activity. Samples of rhesus peripheral blood lymphocytes were divided, and half of each sample was incubated with 100 nM paroxetine in assay buffer for 45 min, while the other half remained in assay buffer (4,6). Both samples were then resuspended in freshly oxygenated assay buffer. A representative serotonin uptake curve after paroxetine pretreatment is shown in Figure Figure5C.5C. Oxidative current was monitored for 20 min after the addition of 0.5 μM serotonin to lymphocyte suspensions. No serotonin clearance was observed in lymphocytes preincubated with paroxetine compared to that in lymphocytes preincubated in assay buffer alone. These data suggest that SERT is primarily responsible for taking up serotonin in lymphocytes, resulting in time-dependent decreases in extracellular serotonin detected by chronoamperometry. Paroxetine preincubation experiments were performed on three separate days using separate frozen stocks of pooled lymphocytes and different BDMs with similar results each day.
Rhesus macaques express a promoter polymorphism that is evolutionarily related to the human 5-HTTLPR; however, studies have not been carried out to investigate serotonin uptake with respect to the rhesus gene variant. Here, we assessed SERT function in native (nontransformed) lymphocytes and observed allele-specific decreases in serotonin uptake associated with the short form of the rh5-HTTLPR (Figure (Figure6).6). Serotonin uptake rates were significantly decreased by 30% in lymphocytes from monkeys with the s/s genotype (P < 0.001) and 25% in the s/longl genotype (P < 0.05) compared to those in animals with the l/l genotype. Thus, similar to the h5-HTTLPR, serotonin uptake is decreased in peripheral blood cells isolated from adult animals carrying the short rh5-HTTLPR allele, which is hypothesized to be associated with decreased SERT expression.
Serotonin uptake rates and SERT protein levels have each been evaluated in human transformed lymphoblasts and platelets with respect to the h5-HTTLPR. Lesch et al. first reported a decrease in serotonin uptake and SERT binding associated with the s allele in human transformed lymphoblastoid cell lines (31). Later studies on serotonin transport in human platelets corroborated these findings (69−71), with the exception of one study by Kaiser et al. (72). However, most studies on SERT binding as a measure of SERT protein levels in platelets have failed to find decreases associated with the h5-HTTLPRs allele (69,70,73) and only Stoltenberg et al. reported a reduction in SERT binding in s allele platelets (74). It is not clear what the underlying reasons are for the discrepancies between SERT binding and function nor what biological significance this may have.
Uptake studies in human platelets and lymphoblastoid cells have utilized radiochemical assay to assess serotonin uptake rates. However, we have shown that kinetic parameters obtained using this technique underestimate maximal uptake rates in brain synaptosomes because of a loss of transported serotonin occurring during high pressure vacuum filtration (6). Here, we observe uptake rates in rhesus primary lymphocytes by chronoamperometry comparable to [3H]-serotonin uptake rates reported in platelets. Future studies will be needed to determine whether intact cells retain transported serotonin during vacuum filtration unlike synaptosomes, which are derived from previously disrupted neurons. Gaining a better understanding of these technical aspects will allow more accurate comparisons of uptake rates in intact cells obtained by radiochemical assay vs voltammetry methods.
Similar to studies in peripheral cells, studies on the effects of the 5-HTTLPR in human brain have resulted in conflicting findings. Little et al. reported a significant decrease in SERT binding in postmortem midbrain associated with the s/l but not the s/s genotype compared to the l/l genotype; however, the number of s/s individuals investigated was small (44). By contrast, Naylor and co-workers observed no significant differences with respect to the h5-HTTLPR genotype in postmortem hippocampus (75). Studies measuring in vivo SERT binding potential by PET or single photon emission computed tomography (SPECT) have likewise yielded mixed results. Heinz et al. observed an s allele-dependent decrease in SERT binding in human midbrain (76). However, van Dyke et al. found increased binding in s/s individuals compared to that in s/l individuals in the brain stem (77). Others have reported no differences in SERT binding with respect to h5-HTTLPR genotype using PET or SPECT (46,78,79). In rhesus monkeys, SERT binding evaluated by SPECT was not significantly different with respect to rh5-HTTLPR genotype in the brain stem (80). The low resolution of PET or SPECT imaging might account for these mostly negative findings in vivo.
Mice and rats with reduced SERT expression have also been produced to investigate the effects of SERT deficiency on neurochemistry, neurophysiology, neuropharmacology, and behavior (4−6,28−30,39,54,81−84). Rodents with constitutive loss of SERT expression exhibit increases in anxiety-related behavior (28,82,83) that bears resemblance to increases in anxiety-related personality traits reported in humans carrying one or two copies of the h5-HTTLPRs allele. Moreover, postnatal disruption of SERT function by administration of serotonin reuptake inhibiting antidepressants in mice and rats leads to an increased anxiety-related phenotype during adulthood (85−87). Together, these data from genetic and pharmacologic rodent models, when viewed particularly in light of the negative data on the h5-HTTLPR and its association with differential SERT binding in adult human brain and platelets, have led to the hypothesis that the effects of the h5-HTTLPR on SERT expression and function might be limited to key postnatal periods (46,88). Reduced serotonin uptake in early development is thought to alter the formation of neuronal circuits responsible for modulating anxiety-related behavior during adulthood.
In contrast to this hypothesis wherein the effects of the 5-HTTLPR are limited to development, we observe significant differences in serotonin uptake rates associated with the rh5-HTTLPR in native rhesus lymphocytes obtained from adult animals using highly sensitive electrochemical methods. Our results strongly suggest that the rh5-HTTLPR influences SERT function not only during critical developmental periods but that these effects persist into adulthood. Previously, we and others have elucidated modest but biologically important differences in serotonin uptake rates in the brains of mice with partial constitutive reductions in SERT expression using chronoamperometry that are on the order of the differences in uptake rates observed here in rhesus lymphocytes. We hypothesize that the negative findings on the h5-HTTLPR and its association with SERT expression and/or function described above might largely be due to the use of methods with low resolution for determining uptake rates, e.g., radiochemical uptake methods or in vivo SERT binding, e.g., PET or SPECT imaging. Understanding whether the effects of the rh5-HTTLPR and more importantly, the h5-HTTLPR, persist into adulthood will be critical for advancing our fundamental understanding of the mechanisms by which this ubiquitous gene variant influences the architecture of human personality. Moreover, these mechanisms are also important for interpreting large individual variations in treatment response to SSRIs (41) and for using SERT activity in peripheral lymphocytes as a biomarker for personalizing first-line treatments for patients suffering from depression and anxiety disorders.
With recent advances in fabrication methods for thin-film boron-doped diamond (22) and reductions in the sizes of BDMs (64,89), these electrodes are becoming increasingly suited for making measurements of neurotransmitters in biological environments with the high temporal and spatial resolution needed to elucidate smaller magnitude changes that are nonetheless of principal biological significance. The unique surface chemistry of BDMs imparts excellent resistance to fouling. However, unlike coatings such as Nafion, they do not afford intrinsic advantages for detecting cationic monoamine neurotransmitters over anionic metabolites (90), the latter of which are typically encountered at high concentrations in vivo. However, a number of BDM surface modifications have been reported to provide selectivity along these lines. Anodized BDMs show separation of peak oxidative potentials for dopamine vs ascorbate, facilitating selective dopamine detection (89,91). Weng et al. electrodeposited gold clusters on boron-doped diamond electrodes followed by self-assembled monolayers of mercaptoacetic acid, and together, these modifications both enhanced the separation of oxidative peak potentials between dopamine and ascorbate and increased the limits of detection for dopamine (92). Moreover, permselective films on boron-doped diamond have been used to selectively detect dopamine or serotonin in the presence of ascorbate (93,94). In the present study, selectivity for cations was not an issue since we controlled the composition of the extracellular solution containing serotonin.
We conclude that reduced SERT function is associated with the short allele of the rh5-HTTLPR in peripheral blood lymphocytes isolated from adult rhesus monkeys. Our results illustrate further the sensitivity of chronoamperometry for differentiating serotonin uptake rates. They also demonstrate the effectiveness of boron-doped diamond microelectrodes for measuring serotonin uptake in peripheral cells natively expressing SERT. However, a number of key questions will need to be answered prior to determining whether peripheral lymphocytes can be used as potential biomarkers to reveal alterations in central SERT function associated with genetic variability and/or antidepressant responses in humans. First, are rh5-HTTLPR-related differences in SERT function associated with corresponding changes in SERT mRNA and/or protein levels in rhesus lymphocytes? Second, are rh5-HTTLPR-associated effects in the periphery paralleled in the brain, or are the ramifications of the 5-HTTLPR for anxiety-related personality traits and susceptibility to stress-related psychiatric disorders based on developmental central effects possibly interacting with adult peripheral alterations in serotonin transport? Third, how does the situation in rhesus monkeys compare to that in humans? The rh5-HTTLPR, in combination with electrochemical methods, provides many unique advantages for investigating some of these questions, including the possibility of carrying out in vivo neurochemical studies to determine associations between peripheral and central SERT function in monkeys (64,89).
To understand at the molecular level factors that influence susceptibility to anxiety disorders and depression, as well as drug mechanisms and efficacy, elucidating the effects of SERT polymorphisms, alone and in combination, on SERT expression and function will be necessary. Boron-doped diamond electrodes when used in combination with voltammetry can be used to detect modest change in serotonin uptake rates. Furthermore, the application of these methods to peripherally accessible tissues such as lymphocytes together comprise a powerful system to explore and to understand interneuronal chemical communication.
Venous blood was collected from Chinese rhesus macaques (Macaca mulatta) under anesthesia. Experimental protocols strictly adhered to National Institutes of Health guidelines and were approved by the University of Pittsburgh School of Medicine Institutional Animal Care and Use Committee. High molecular weight genomic DNA was isolated by standard methods, and genotyping of the rh5-HTTLPR was carried out as described by Lesch et al. (47). Monkeys for the present study were selected from a cohort of 42 animals based on their rh5-HTTLPR genotype. Study animals were 6.9 ± 0.1 years of age and weighed 7.8 ± 0.4 kg at the time of blood collection. Genotypes were N = 6 for l/l, N = 3 for s/l, and N=6 for s/s. Blood was collected from additional monkeys that had not been genotyped, and these blood samples were mixed together. Lymphocytes isolated from this pooled blood were used for protocol development and other experiments that were not dependent on genotype.
Peripheral blood lymphocytes were isolated from anticoagulated blood by Ficoll−Paque PLUS gradient centrifugation according to the manufacturer's instructions (GE Healthcare, Piscataway, NJ). Whole blood was diluted with RPMI media (Invitrogen Corporation, Carlsbad, CA) and added into Accuspin tubes (Sigma Aldrich, St. Louis, MO) containing Ficoll (GE Healthcare, Piscataway, NJ) and then centrifuged for 25−30 min at 900 g at room temperature. The entire top white layer was collected in 5% RPMI media followed by centrifugation at 900g for 10 min. The supernatant was then resuspended in 10−20 mL of 10% RPMI. If detected, red blood cells were lysed, and the samples were vortexed and rinsed with fresh 10% RPMI and centrifuged for an additional 10 min at 900g. Cell pellets were brought up in a freezing media (RPMI containing heat inactivated fetal bovine serum and 10% dimethylsulfoxide) and were placed at −20 °C for 1 h prior to freezing overnight in a Mr. Frosty (Thermo Fisher Scientific, Rochester, NY) at −80 °C. Long-term storage was in liquid nitrogen.
Representative frozen samples were thawed in assay buffer, and cell type and cell survival were analyzed by flow cytometry. Forward and side scatter plots indicated that cell populations were predominantly lymphocytes containing a few monocytes with no evidence of platelets or red blood cells. Trypan blue exclusion was also used to assess cell survival (see below).
Boron-doped diamond microelectrodes were fabricated using previously described procedures (22). Briefly, boron-doped diamond was deposited over chemically etched platinum wires using microwave-assisted chemical vapor deposition (1.5 kW, 2.54 GHz, ASTeX, Woburn, MA) to produce cylindrical electrode geometries. Boron doping was achieved during the vapor deposition process by adding 10 ppm B2H6 to 0.5% CH4/H2. After deposition, each diamond-coated platinum wire was glued to a copper wire using silver epoxy, and the joint was sealed in a polypropylene pipet tip using a heat gun.
Carbon fiber microelectrodes (CFMs) were prepared as described previously with minor modifications (4,6). Some electrodes were coated with Nafion, a perfluorinated ion-exchange polymer used to reduce fouling (57). Electrodes were cleaned with isopropanol, and Nafion was applied by dipping CFMs for 30 s followed by drying at 100 °C for 5 min between coats. The coating process was repeated 6 times, and coating effectiveness was evaluated by challenging electrodes with 100 μM ascorbate. Only Nafion-coated electrodes with selectivity for serotonin over ascorbate >1000:1 were used in subsequent experiments.
Reference electrodes were fabricated by electrochemically coating silver wires (Alfa Aesar, Ward Hill, MA) with AgCl. Silver wires were immersed in 3 M HCl saturated with NaCl, and a 9 V potential was applied for 5 min.
Cyclic voltammetry and high-speed chronoamperometry were performed using a Universal Electrochemistry Instrument and TarHeel CV software, version 1.0 (University of North Carolina, Chapel Hill, NC). Chronoamperometry raw data from the TarHeel CV software were further analyzed using a custom software module, ChronoAmp, written in-house. For chronoamperometry, a 1 Hz square wave step potential was applied to the working electrode, which consisted of an oxidative potential at 0.8 V for BDD electrodes or 0.55 V for CFMs, except where otherwise indicated. The oxidative potential was applied for 100 ms followed by a 0 V reductive potential for 100 ms (Figure (Figure7B).7B). The potential was held at 0 V for an additional 800 ms to reduce fouling due to serotonin and its oxidation products (delayed-pulse mode) (95). Potentials were applied with respect to Ag/AgCl reference electrodes. The last 80 ms of the oxidative and reductive phases representing faradaic current were integrated.
Integrated current values prior to the addition of serotonin were defined as background current, while integrated current after the addition of serotonin was defined as total current. The oxidation and reduction currents due to serotonin (i.e., signal) were calculated by subtracting background currents from total currents. Background-subtracted currents were plotted with respect to time (Figure (Figure7C).7C). Noise was calculated as the standard deviation of the background current measured over 10 scans. Limits of detection were defined as the concentrations of serotonin needed to generate signals equivalent to three times the noise values. In addition to background and total current, the current for 80 ms just prior to the application of the oxidative pulse was integrated to calculate the drift in background current and voltage (IR) drop due to resistance in the circuit. The drift current was subtracted from the oxidative and reductive currents to minimize the contribution of electrode drift and voltage drop to the determination of faradic current resulting from serotonin.
For evaluation of responses to 10 μM serotonin by cyclic voltammetry, BDMs or CFMs were held at −0.4 V with respect to Ag/AgCl reference electrodes and scanned from −0.4 to 1.0 V to −0.4 at 1 V/s every 10 s. Background and oxidative currents were determined at the potential needed to produce half maximal current in the presence of serotonin (E1/2ox) at the respective electrodes. All cyclic voltammograms used to measure signal and background currents were averaged over 10 scans.
Experiments were performed in 12-well polystyrene plates (BD Biosciences, San Jose, CA). Reference and working electrodes were carefully lowered into each well containing assay buffer (150 mM NaCl, 5 mM KCl, 1.2 mM MgCl2, 5 mM glucose, 10 mM HEPES, and 2 mM CaCl2, pH 7.4) or lymphocytes suspended in assay buffer, and the background current was monitored (Figure (Figure7A).7A). Data collection began when the background current stabilized, i.e., changes in current were <0.1 nA for BDMs or <0.2 nA for CFMs for a minimum of 100 s. The background current was then recorded for 50 s followed by injection of serotonin into the assay buffer. Solutions of lymphocytes or buffer were briefly stirred, and the current was recorded for an additional 120 s for calibration experiments or 1200 s for uptake experiments without further stirring. Electrodes were calibrated prior to each experiment against known concentrations of serotonin in assay buffer (0.1−1 μM unless otherwise noted), and the linearity of the responses were calculated. Electrodes were also calibrated after each experiment to determine changes in electrode sensitivity due to fouling.
Frozen lymphocytes (~10 million cells/mL) were thawed by adding 12−15 mL of 37 °C assay buffer. Small volumes (200 μL) of lymphocytes were used for counting live cells using Trypan blue exclusion. On average, 75% of the cells were alive after the thawing procedure. Lymphocytes were then centrifuged at 340g for 7 min. Pellets containing lymphocytes were resuspended by gently vortexing in assay buffer to produce final concentrations of 2−4 million cells/mL except where otherwise noted. Prior to each experiment, numbers of live cells were determined, and uptake rates were normalized per million live cells. Solutions of lymphocytes were kept at 4 °C for no more than 4 h before uptake experiments.
Immediately prior to measuring serotonin uptake, lymphocytes were centrifuged at 340g for 7 min. Freshly oxygenated assay buffer saturated with 95% O2/5% CO2 by bubbling the gas mixture for at least 30 min was utilized at room temperature for all experiments. Cells were resuspended in an appropriate amount of assay buffer by gently vortexing. In some experiments, frozen lymphocytes were thawed in oxygenated assay buffer containing 150 mM LiCl substituted for NaCl. Prior to measuring uptake, half of these lymphocytes were resuspended in LiCl-substituted oxygenated assay buffer and the other half in normal oxygenated assay buffer to investigate Na+-dependent uptake. In other experiments, divided solutions of lymphocytes were preincubated for 45 min with paroxetine (100 nM) in assay buffer or assay buffer alone to investigate the effects of SERT inhibition on uptake.
Serotonin, paroxetine, Nafion, and chemicals for lymphocyte isolation were purchased from Sigma-Aldrich (St. Louis, MO). All chemicals used for assay buffer preparation were purchased from VWR (West Chester, PA).
For chronoamperometry experiments, the first order rate constant, b (s−1), was determined by fitting serotonin clearance data to the following exponential decay function (96):
where y is the serotonin concentration (μM) at any given time (t), to is the time (s) at the start of uptake, and A is the concentration of serotonin (μM) at t0.
Initial serotonin uptake rates, υ (μM/s), were then calculated using the following formula:
Uptake rates were determined to be not detectable (negligible serotonin uptake) in cases where a <10% decrease in current occurred over 20 min after serotonin injection.
Data comparing two means were analyzed using two-tailed unpaired t-tests. One-way analysis of variance (ANOVA) was used in cases where more than two means were compared, followed by either Dunnett’s Multiple Comparison or Tukey’s post hoc tests. Two-way ANOVA with repeated measures was used to analyze calibration data using serotonin concentrations as the repeated measure and to analyze fouling data with treatment as the repeated measure. Calibration data were also analyzed by multiple regression analysis and nonlinear curve fitting. All statistical analyses were performed using GraphPad Prism v.4 for Mac (GraphPad Software, La Jolla, CA). All values are expressed as means ± standard errors of the mean (SEMs), with differences of P < 0.05 considered statistically significant. Significant differences are denoted in the figures as *P < 0.05, **P < 0.01, ***P < 0.001, and ††P < 0.01.
We would like to remember Professor Peter C. Eklund, who contributed greatly to our understanding of carbon in its many forms. We are grateful to Dr. Charles Bradberry for guidance regarding nonhuman primates and Drs. Andrew Ewing and Michael Heien for advice on electrochemical methods. We also acknowledge Ms. Stefanie Altieri and Mr. Brendan Beikmann for assistance with the experiments. This project was supported by funding from the National Institute of Mental Health (MH064756 to AMA).
The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Mental Health or the National Institutes of Health.