INTRODUCTION

Extracellular serotonin (5-hydroxytryptamine (5-HT)) in brain is regulated in part by the presynaptic serotonin reuptake transporter (5-HTT, SLC6A4). Mice with a partial (5-HTT+/−) or complete (5-HTT−/−) 5-HTT deletion, compared with wild-type (5-HTT+/+) mice, differ with regard to brain anatomy; brain concentrations, reuptake, synthesis, and release of 5-HT; 5-HT and GABA receptor densities; programmed cell death; and brain glucose metabolism (Bengel et al, 1998; Esaki et al, 2005; Fox et al, 2007a, 2008a; Mathews et al, 2004; Murphy et al, 2008; Murphy and Lesch, 2008). They also show increased anxiety- and depression-like behaviors and reduced aggressiveness on various tests (Fox et al, 2007a; Murphy et al, 2008; Murphy and Lesch, 2008).

Reduced serotonergic function in 5-HTT+/− and 5-HTT−/− mice is thought to be comparable to reduced serotonergic function in humans who carry one or two short (S) compared with long (L) alleles of the promoter-region polymorphism of 5-HTT (5-HTTLPR), or who express rs25531 or rs25532 variants of the 5-HTT allele (Murphy et al, 2008; Murphy and Lesch, 2008). Thus, studying 5-HTT-deficient mice could elucidate dysfunctional serotonergic neurotransmission in humans with these polymorphisms, and suggest new methods for identifying and quantifying this dysfunction.

For example, humans carrying one ‘S’ 5-HTTLPR allele have 50% reductions in 5-HTT expression and function in lymphocytes, platelets, and brain, compared with those with the LL genotype (Hu et al, 2006; Murphy et al, 2008; Murphy and Lesch, 2008; Praschak-Rieder et al, 2007; Wendland et al, 2008). They also have comparatively elevated anxiety, depression, and aggression-related personality traits, and increased susceptibility to depression associated with major negative life events (Caspi et al, 2003; Uher and McGuffin, 2008). They respond poorly to selective serotonin reuptake inhibitors (SSRIs; Hu et al, 2006; Murphy et al, 2004; Serretti et al, 2005), and are at increased risk for bipolar disorder, comorbid disorders accompanying alcoholism, and suicide (Baca-Garcia et al, 2007; Li and He, 2007; Marques et al, 2006; Masoliver et al, 2006). Extracellular striatal 5-HT concentrations are three- and sixfold higher, respectively, in 5-HTT+/− and 5-HTT−/− than 5-HTT+/+ mice (Mathews et al, 2004). 5-HT2A receptor density is reduced in the striatum but increased in the hypothalamus and septum of 5-HTT−/− compared with 5-HTT+/+ mice, whereas 5-HT2C receptor density is elevated in the amygdala and choroid plexus (Li et al, 2003).

Elevated extracellular 5-HT concentrations would be expected to increase 5-HT occupancy of the postsynaptic 5-HT2A/2C receptors that are coupled to cytosolic phospholipase A2 (cPLA2), and thereby tonically activate cPLA2 (Berg et al, 1998a; Clark et al, 1995; Felder et al, 1990). cPLA2 when activated selectively releases arachidonic acid (AA, 20:4n-6) from membrane phospholipid to initiate the AA signaling cascade (Fitzpatrick and Soberman, 2001; Shimizu and Wolfe, 1990). AA and its metabolites (eg, prostaglandins and endocannabinoids) can modify sleep, neural firing, neurotransmitter release, nociception, cerebral blood flow, and gene transcription (Bosetti, 2007). We, therefore, thought it of interest in this study to see whether this cascade is upregulated in 5-HTT-deficient mice.

Brain AA signaling involving cPLA2-coupled neuroreceptors can be imaged in unanesthetized rodents by infusing radiolabeled AA i.v. and measuring tracer uptake into brain with quantitative autoradiography (Rapoport, 2001; Robinson et al, 1992). k* for AA at baseline or following drug is independent of changes in cerebral blood flow, thus only reflecting brain AA metabolism (Chang et al, 1997). The flux Jin of AA, which represents the rate of regional brain AA consumption, as AA cannot be synthesized de novo in vertebrate tissue or converted from its circulating precursor, linoleic acid (18:2n-6) in brain (DeMar et al, 2006; Holman, 1986; Rapoport et al, 2001) can be calculated as the product of k* and the unesterified plasma AA concentration.

In this study, we used in vivo brain AA imaging to test whether the reported high levels of brain extracellular 5-HT in 5-HTT-deficient mice would tonically stimulate 5-HT2A/2C receptors to augment cPLA2 activity, and thereby elevate baseline values of k* and Jin, and of AA-derived eicosanoid concentrations. We also examined whether these changes would be accompanied by elevated cyclooxygenase (COX) activity and concentrations of COX-derived eicosanoids, as COX-1 and -2 have been reported to be functionally coupled to cPLA2 in brain (Bosetti and Weerasinghe, 2003; Fitzpatrick and Soberman, 2001; Kaufmann et al, 1996; Ong et al, 1999; Pardue et al, 2003; Sapirstein et al, 2005; Xu et al, 2008). Additionally, we checked whether k* responses to (+/−)-2,5-dimethoxy-4-iodophenyl-2-aminopropane (DOI), a partial 5-HT2A/2C agonist (Marek and Aghajanian, 1996), would be reduced in 5-HTT-deficient mice by displacing already bound 5-HT. The 1.5 mg/kg s.c. DOI dose that we chose has been reported to increase k* for AA significantly in mouse brain regions rich in 5-HT2A/2C receptors (Qu et al, 2005). Finally, we quantified head-twitch responses (HTRs) to DOI as a behavioral test of 5-HT2A receptor function (Willins and Meltzer, 1997), before and after pharmacological alteration of extracellular 5-HT (Cesana et al, 1993; Fox et al, 2007b). Parts of this study have been presented in abstract form (Basselin et al, 2007a).

MATERIALS AND METHODS

Animals

Experiments were conducted following the ‘Guide for the Care and Use of Laboratory Animals’ (National Institute of Health Publication No. 86-23) and were approved by the Animal Care and Use Committee of the Eunice Kennedy Shriver National Institute of Child Health and Human Development. Five- to nine-month-old male 5-HTT+/− and 5-HTT−/− mice and their littermate 5-HTT+/+ controls, derived from a C57BL/6J genetic background (Bengel et al, 1998), were maintained in an animal facility in which temperature, humidity, and light cycle were regulated with free access to water and a fixed diet (Rodent NIH-31 auto 18-4, Zeigler Bros, Gardners, PA). The diet contained (as percent of total fatty acids) 20.1% saturated, 22.5% monounsaturated, 47.9% linoleic, 5.1% α-linolenic, 0.02% AA, 2.0% eicosapentaenoic, and 2.3% docosahexaenoic acid.

Drugs

Unanesthetized mice received 0.9%. NaCl (saline) or 1.5 mg/kg s.c. DOI (Sigma-Aldrich, St Louis MO). [1-14C]AA in ethanol (53 mCi/mmol; 99.4% pure, Moravek Biomedicals, Brea, CA) was evaporated and resuspended in 5 mM HEPES buffer, pH 7.4, which contained 50 mg/ml of bovine serum albumin, essentially fatty acid free (Sigma-Aldrich). Tracer purity was ascertained to exceed 99% by gas chromatography, after converting AA into its methyl ester with 1% sulfuric acid in anhydrous methanol. The 5-HT synthesis inhibitor para-chlorophenylalanine (PCPA, 30 mg/ml prepared in distilled deionized water) and the 5-HT precursor 5-hydroxy-L-tryptophan (5-HTP, 5 mg/ml prepared in 5% Tween 80 in distilled water) were obtained from Sigma-Aldrich.

Surgical Procedures and Tracer Infusion

A mouse was anesthetized with 2–3% halothane in O2 and PE 10 polyethylene catheters were inserted into its right femoral artery and vein as reported (Basselin et al, 2006b; Qu et al, 2005). The wound site was closed with 454 Instant Adhesive (Loctite Corp. Hartford, CT), and the mouse was wrapped loosely, with its upper body remaining free, in a fast-setting plaster cast taped to a wooden block. It was allowed to recover from anesthesia for 3–4 h in a warming environment maintained at 25°C. Starting 20 min after s.c. DOI or saline injection, 45 μl [1-14C]AA (300 μCi/kg) was infused for 3 min through the femoral vein at a rate of 15 μl/min, using a Hamilton syringe and an infusion pump (Harvard Apparatus Model 22, Holliston, MA). Ten 15-20 μl arterial blood samples were collected at 0, 0.25, 1.0, 1.5, 2.0, 2.8, 3.2, 5.0, 10, and 19 min to determine radioactivity of unesterified AA in the plasma. At 20 min, the mouse was killed by an overdose of Nembutal® (50 mg/kg, i.v.). The brain was removed quickly within <30 s, frozen in 2-methylbutane in dry ice at −40°C, and stored at −80°C until sectioned.

Chemical Analysis

The blood samples collected before, during, and after [1-14C]AA infusions were centrifuged immediately (30 S at 18 000 g) to obtain plasma, which was stored at −80°C. Total lipids were extracted from 5 μl of thawed plasma with 1 ml chloroform:methanol (2:1, by vol) and 0.5 ml 0.1 M KCl, using a modified method of Folch (Folch et al, 1957). Radioactivity was determined in 100 μl of the lower organic phase by liquid scintillation counting. As reported earlier, more than 95–98% of total plasma and brain radioactivity at 5 min was radiolabeled AA (Lee et al, 2007).

Concentrations of unlabeled unesterified fatty acids were determined in 100–150 μl of frozen arterial plasma collected by heart puncture. Total lipids were extracted by the modified Folch method, and were separated by thin layer chromatography on silica gel 60 plates using the solvent system: heptane/diethyl ether/acetic acid (60:40:3, by vol). Unesterified fatty acids were scraped from the plate and methylated with 1% sulfuric acid (by vol) in anhydrous methanol for 3 h at 70°C, then separated and quantified by gas chromatography using heptadecanoic acid (17:0) as an internal standard.

Quantitative Autoradiography and Calculations

Frozen brains were cut in serial 20-μm thick coronal sections on a cryostat at −20°C, then placed for 4 weeks together with calibrated [14C]methylmethacrylate standards (Amersham, Arlington Heights, IL) on Ektascan C/RA film (Eastman Kodak Company, Rochester, NY). Radioactivity (nCi/g of brain) in 92 anatomically identified regions (Franklin and Paxinos, 1997) was measured bilaterally six times by quantitative densitometry, using the public domain NIH Image program 1.62 (http://rsb.info.nih.gov/nih-image/). Regional AA incorporation coefficients k* (ml/s/g brain) of AA were calculated as (Robinson et al, 1992),

where cbrain* (nCi/g brain) is brain radioactivity at 20 min after the onset of infusion as determined by densitometry, cbrain* (nCi/ml plasma) is the arterial plasma concentration of labeled unesterified AA as determined by scintillation counting, and t (min) is time after the onset of [1-14C]AA infusion. Integrals of plasma radioactivity (input function in denominator) were determined in each experiment by trapezoidal integration, and divided into cbrain* to calculate k* for each experiment.

Regional rate of incorporation of unesterified AA from plasma into brain phospholipids, Jin (fmol/s/g), was calculated as,

where cplasma (nmol/ml) is the plasma concentration of unlabeled unesterified AA.

Brain cPLA2 Activity

In separate experiments, mice were anesthetized with Nembutal (50 mg/kg, i.p.) and decapitated. The brain was rapidly excised, frozen in 2-methylbutane maintained at −40°C with dry ice, and stored at −80°C. Brain hemispheres were homogenized using a Teflon-glass homogenizer in 2 vol of ice-cold buffer containing 10 mM HEPES, pH 7.5, 1 mM EDTA, 0.34 M sucrose and protease inhibitor cocktail tablet (Complete, Roche, Mannheim, Germany). Homogenates were centrifuged at 14 000 g for 20 min, then at 100 000 g for 1 h at 4°C. Supernatants corresponding to the cytosolic fraction were assayed for cPLA2 activity, using a cPLA2 assay kit and secretory PLA2 and Ca2+-independent PLA2 inhibitors (Cayman, Ann Arbor MI).

Brain COX Activity

Brain hemispheres (see above) were homogenized using a Teflon-glass homogenizer in 1 ml of ice-cold lysate buffer containing 10 mM Tris-HCl, pH 7.8, 1% Igepal CA-630, 0.15 M NaCl, and 1 mM EDTA. Homogenates were centrifuged at 14 000 g for 20 min at 4°C. Brain COX activity was measured as the rate of PGE2 formation (pg PGE2/min/mg cytosolic protein) in homogenate cytosolic fractions diluted 1:5 with lysate buffer in the presence of 10 mM phenol, 18.2 mM (−)-epinephrine, 4.6 mM L-glutathione reduced, and 10 μM porcine hematin. The reaction was started by adding AA (Oxford Biochemical Research, Oxford, MI) to a final concentration of 0.1 mM, and the mixture was incubated at 37°C for 15 min. The reaction was terminated by adding 250 μl of 1 M HCl. PGE2 was extracted with ethyl acetate and quantified using a PGE2 immunoassay kit (Oxford Biochemical Research, Oxford, MI). A sample not containing AA was assayed and used for the blank determination.

In the same study, test drugs, Celebrex® (400 mg; Pfizer Inc., New York, NY; obtained from NIH Division of Veterinary Medicine, Bethesda, MD), a specific COX-2 inhibitor, and DOI were dissolved in dimethylsulfoxide at a concentration of 0.1% and in saline, respectively. These drugs were added to the mixture 10 min before adding AA (see above).

Brain PGE2, PGF2α, and TXB2 Concentrations

In separate experiments, mice were anesthetized with Nembutal (50 mg/kg, i.p.) and subjected to head-focused microwave irradiation (5.5 kW, 0.9 s; Cober Electronics, Stamford, CT) to stop postmortem changes (Anton et al, 1983). Frozen half brains were weighed, homogenized with 18 vol of hexane:isopropanol (3:2 by vol) using a glass Tenbroeck homogenizer, and the homogenate was centrifuged for 5 min at 800 g. Tissue residues then were rinsed with 3 × 2 vol of the same solvent. The resultant lipid extract was concentrated to dryness under nitrogen and resuspended in enzyme immunoassay buffer provided with the polyclonal PGE2, PGF2α and TXB2 kits (Oxford Biochemical Research, Oxford, MI).

Head-Twitch Responses

Mice were administered PCPA (300 mg/kg i.p.) or vehicle twice daily for 3 days (Cesana et al, 1993). On the fourth day, 18 h after the final dose of PCPA, mice were placed in a Plexiglas container. Following 15 min of habituation, DOI (2.5 mg/kg i.p.) was administered. HTR were counted for five 1-min periods starting 5 min after drug administration, and were summed over these five periods. In a separate experiment, HTR following administration of 5-HTP (80 mg/kg i.p.) were assessed in a similar manner (Fox et al, 2007b).

Statistical Analysis

A one-way analysis of variance (ANOVA) with a Bonferroni's post-test was used to compare mean body weights, cPLA2 and COX activities, and eicosanoid concentrations using GraphPad Prism version 4.0b for Macintosh (GraphPad Software, San Diego CA, www.graphpad.com). A two-way ANOVA was employed to examine the effects of two factors, genotype (5-HTT−/− or 5-HTT+/− vs 5-HTT+/+), and drug (DOI vs saline) using SPSS 16.0 (SPSS Inc., Chicago, IL, http://www.spss.com) on the arterial input function, plasma unesterified fatty-acid concentrations, k* and Jin. We report all main effect statistics (p- and F-values), although main effects in the context of a significant interaction may be difficult to interpret (Motulsky 2003; Tabachnick and Fidell, 2001). A one-way ANOVA with Bonferroni's post-test with correction for five comparisons (5-HTT+/+ plus DOI vs 5-HTT+/+ saline; 5-HTT+/− saline vs 5-HTT+/+ saline, 5-HTT+/− plus DOI vs 5-HTT+/− saline, 5-HTT−/− saline vs 5-HTT+/+ saline, and 5-HTT−/− plus DOI vs 5-HTT−/− saline) was performed. For k* and Jin, corrections for multiple comparisons across regions were not made because the purpose of this exploratory study was to identify regions that were involved in individual drug effect.

One-way (genotype) or two-way (genotype × drug condition) ANOVAs followed by Bonferroni's post-tests were used to assess differences in 5-HTP- and DOI-induced HTR, respectively. Data are reported as mean±SD, with statistical significance taken as p0.05.

RESULTS

Body Weight and Arterial Plasma Input Function

Mean body weight was significantly higher in 5-HTT−/− (p<0. 01) than 5-HTT+/+ mice (38.9±5.7 g (n=10) vs 29.5±5.4 g (n=11)). Body weight equaled 34.2±2.8 g (n=10) in 5-HTT+/− mice.

A two-way ANOVA revealed a significant effect of DOI (p=0.008) on integrated arterial plasma radioactivity (denominator of Eq. 1; plasma input function). Input functions [(nCi/ml × s)±SD, n=4–6] are 5-HTT+/+ plus saline, 135 476±21 938; 5-HTT+/+ plus DOI, 116 030±19 123; 5-HTT+/− plus saline, 130 028±10 446; 5-HTT+/− plus DOI, 106 044±11 708; 5-HTT−/− plus saline, 113 179±18 727, and 5-HTT−/− plus DOI, 105 875±11 947.

Plasma Concentrations of Unlabeled Unesterified Fatty Acids

A two-way ANOVA showed significant interactions between 5-HTT genotype and DOI for plasma concentrations of unesterified palmitoleic, stearic, oleic, and arachidonic acids (Table 1). Subsequent one-way ANOVAs with Bonferroni's post-tests showed that palmitoleic and oleic acid concentrations were higher in 5-HTT−/− and 5-HTT+/− mice than in 5-HTT+/+ mice by 57 and 34%, respectively, and that DOI compared with saline decreased palmitoleic acid in 5-HTT+/− and 5-HTT−/− mice by 52 and 38%, respectively. The mean unesterified plasma AA concentration did not differ significantly between groups. Where 5-HTT × DOI interactions were insignificant, the 5-HTT genotype had a main effect for palmitic, linoleic and α-linolenic acid, and DOI had a main effect for linoleic and α-linolenic acids.

Table 1 Unesterified plasma fatty-acid concentrations in 5-HTT+/+, 5-HTT+/−, and 5-HTT−/− mice after s.c. saline or DOI administration

Regional Brain AA Incorporation Coefficients k*

Figure 1 illustrates color-coded coronal autoradiographs of k* for AA from brains of 5-HTT+/+, 5-HTT+/−, and 5-HTT−/− mice injected with either saline (baseline) or DOI. 5-HTT+/− mice, and to a greater extent 5-HTT−/− mice, had higher baseline values of k* (Eq. 1) than the 5-HTT+/+ mice. Values of k* were elevated in 5-HTT+/+ mice injected with DOI compared with saline-injected mice, but reduced in 5-HTT+/− and 5-HTT−/− mice.

Figure 1
figure 1

Coronal autoradiographs of brain showing effects of DOI and 5-HTT genotype on regional AA incorporation coefficients k* in mice. Values of k* (ml/s/g brain × 10−4) are given on a color scale from 5 (blue) to 30 (red). Acg, anterior cingulate cortex; Aud, auditory cortex; CPu, caudate-putamen; Hipp, hippocampus; IPC, interpeduncular nucleus; Mot, motor cortex; SN, substantia nigra; Thal, thalamus; Vis, visual cortex.

Mean AA incorporation coefficients k* in each of 92 brain regions were compared among the different experimental groups and conditions using a two-way ANOVA. As illustrated in Table 2, 90 brain regions (but not the bed nucleus of the stria terminalis and the dorsal raphe nucleus, highlighted) had statistically significant genotype × drug interactions.

Table 2 Arachidonic acid incorporation coefficients (k*) in 5-HTT+/+, 5-HTT+/−, and 5-HTT−/− mice at baseline and in response to DOI

Effect of 5-HTT genotype on baseline values of k*

A one-way ANOVA with a Bonferroni's post-test showed that partial and total 5-HTT deletion significantly increased mean baseline values of k* for AA in 45 (by 20–67%) and 72 (by 21–71%) regions, respectively. Cerebral cortex, olfactory tubercle, hippocampus, nucleus accumbens, caudate-putamen, geniculate nucleus, thalamus, mammillary nucleus, mesencephalon, and rhombencephalon were affected in both genotypes. In the two regions with statistically insignificant genotype × drug interactions, 5-HTT genotype did not have any main effect.

Effect of DOI in 5-HTT+/+ mice

DOI compared with saline significantly increased k* for AA (by 17–65%) in 42 of 92 regions of the 5-HTT+/+ mice (Table 2). Positively affected regions included cerebral cortex (21 of 25 regions, average 33%), suprachiasmatic nucleus (39%), hippocampus CA1 (17%), caudate-putamen ventral (23%), geniculate nucleus (29%), subthalamic nucleus (26%), mesencephalon (6 of 9 regions, average 38%), rhombencephalon (6 of 10 regions, average 51%), white matter (1 of 4 regions, 19%), and nonblood–barrier regions (2 of 3, average 42%). In the two regions with insignificant genotype × drug interactions, DOI did not have any significant main effect.

Effect of DOI in 5-HTT+/− and 5-HTT−/− mice

DOI compared with saline did not significantly increase k* in any of the 42 regions where 5-HTT genotype × drug interactions were statistically significant (Table 2), but significantly reduced k* in 31 (−12 to −31%) and 83 (−16 to −63%) regions in 5-HTT+/− and 5-HTT−/− mice, respectively. In the two regions with statistically insignificant genotype × drug interactions, DOI did not have any main effect.

Patterns of significant differences in k*

Figure 2 presents difference patterns of k* responses to DOI in sagittal representations of the mouse brain. The 5-HTT+/−+DOI image compared with the 5-HTT+/++saline image illustrates the positive regional effects of DOI in the wild-type mice, whereas the 5-HTT+/−+saline image and the 5-HTT−/−+saline image compared with the 5-HTT+/++saline image illustrates the positive effects of a partial and complete deletion of 5-HTT, respectively, on baseline values of k*. The 5-HTT+/−+DOI and the 5-HTT−/−+DOI images compared with the 5-HTT+/−+saline and the 5-HTT−/−+DOI saline images show the negative effects of acute DOI on k* for AA, in mice with a partial and complete deletion of 5-HTT.

Figure 2
figure 2

Difference patterns of k* responses to DOI in sagittal representation of 5-HTT mouse brain. Regions in which k* was increased significantly (p<0.05) are solid gray, regions in which k* was decreased significantly are hatched. The 5-HTT+/−+DOI image is compared with the 5-HTT+/++saline image. The 5-HTT+/−+saline and the 5-HTT−/−+saline images are compared with the 5-HTT+/++saline image. The 5-HTT+/−+ DOI and the 5-HTT−/−+DOI images are compared with the 5-HTT+/−+saline and the 5-HTT−/−+saline, respectively. List of regions: A, amygdala; Acb, nucleus accumbens; AUD, auditory cortex; am, anteromedial thalamic nucleus; av, anteroventral thalamic nucleus; CbG, cerebellar gray matter; CbW, cerebellar white matter; Co, cochlear nucleus; CPu, caudate putamen; DLG, dorsal lateral geniculate nucleus; DR, dorsal raphe; Fr, frontal cortex; GP, globus pallidus; Hb, habenular complex; HIP, hippocampus; HYP, hypothalamus; IC, inferior colliculus; IPC, interpeduncular nucleus; LC, locus coeruleus; MI, mammillary nucleus; mG, medial geniculate nucleus; MolCbG, molecular layer of cerebellar gray matter; MOT, motor cortex; MR, median raphe; MVe, medial vestibular nucleus; OT, olfactory tubercle; PF, prefrontal cortex; PPTg, pedunculopontine tegmental nucleus; SN, substantia nigra; S, septum; SO, superior olive; Sp5, spinal trigeminal nucleus; SS, somatosensory cortex; SC, superior colliculus; SCgl, gray layer of superior colliculus; STH, subthalamic nucleus; THa, thalamus; Vis, visual cortex.

Regional Incorporation Rates of Unlabeled Unesterified AA from Plasma into Brain

Baseline and DOI-induced rates of incorporation of unlabeled unesterified AA from plasma into brain phospholipids, Jin, were calculated by multiplying individual regional values of k* by the plasma concentration of unlabeled unesterified AA (Eq. 2; data not shown). Each of the 92 regions showed a statistically significant 5-HTT genotype × drug interaction with regard to Jin. In 5-HTT+/+ mice, baseline Jin ranged from 4.19 fmol/s/g in the internal capsule (white matter) to 23.4 fmol/s/g in the choroid plexus. The partial and total 5-HTT deletions significantly increased Jin in 15 and 68 brain regions, respectively. DOI elevated Jin significantly in 71 out of 92 regions in the 5-HTT+/+ mice, whereas the drug significantly decreased Jin in 68 and 83 out of 92 regions in 5-HTT+/− and 5-HTT−/− mice, respectively.

Brain cPLA2 Activity

An in vitro assay with calcium chelators showed that brain cPLA2 activity was increased by 29 (p<0.001) and 34.5% (p<0.001) in 5-HTT+/− and 5-HTT−/− mice, respectively, compared with 5-HTT+/+ mice (Table 3). Activity did not differ significantly between the 5-HTT+/− and 5-HTT−/− mice. We did not analyze brains following DOI, because we could not reproduce the intracellular Ca2+ concentrations associated with DOI in vivo.

Table 3 Effect of 5-HTT genotype on global brain cPLA2 and COX activities, and eicosanoid concentrations

Brain COX Activity

Brain COX activity was decreased by 49.2 (p<0.001) and 74.2% (p<0.001) in 5-HTT+/− and 5-HTT−/− mice, respectively, compared to 5-HTT+/+ mice (Table 3). COX activity was 49% less in 5-HTT−/− than in 5-HTT+/− mice. Preincubation of 5-HTT+/+ homogenate with 100 μM DOI did not significantly affect COX activity (102.9±9.8 vs 112.4±13.9 pg/min/mg protein), indicating that DOI did not inhibit COX enzymes. On the other hand, 100 μM Celebrex, a selective COX-2 inhibitor used as a positive control, inhibited COX activity by 68% (35.5±4.2 vs 112.4±13.9 pg/min/mg protein, n=5, p<0.001).

Brain PGE2, PGF2α, and TXB2 Concentrations

As illustrated in Table 3, the basal brain PGE2 concentration was decreased significantly by 74 and 90% in 5-HTT+/− and 5-HTT−/− mice, respectively, compared with 5-HTT+/+ mice. Brain PGF2α was decreased significantly by 23 and 35% in 5-HTT+/− and 5-HTT−/− mice, respectively, and brain TXB2 was decreased significantly by 34 and 72% in 5-HTT+/− and 5-HTT−/− mice, respectively.

Head-Twitch Responses

After vehicle pretreatment, 5-HTT−/− mice had 86% fewer DOI-induced HTR than 5-HTT+/+ mice (p=0.006), similar to an earlier report (Qu et al, 2005); the number of responses did not differ significantly between 5-HTT+/− and 5-HTT+/+ mice, although 5-HTT+/− mice had an intermediate response (decreased 22% compared to 5-HTT+/+ mice; Figure 3). PCPA pretreatment increased the number of HTR in 5-HTT−/− mice by 386% (p<0.0001), with a trend toward an increase in 5-HTT+/− mice (p=0.08), but had no significant effect in 5-HTT+/+ mice. After PCPA pretreatment, DOI-induced HTR did not differ significantly among the three genotypes. These findings suggest that 5-HT depletion ‘normalized’ DOI-induced HTR in 5-HTT−/− mice (main effect of genotype (F(2,57)=6.33, p=0.003), main effect of pretreatment drug (F(1,57)=3.22, p=0.078), and genotype × pretreatment drug interaction (F(2, 57)=3.13, p=0.05).

Figure 3
figure 3

Effects of brain 5-HT depletion by PCPA pretreatment on DOI-induced head twitches in 5-HTT mice. At baseline (vehicle pretreatment), DOI induced fewer head twitches in 5-HTT−/− mice compared to 5-HTT+/+ mice, whereas DOI-induced head twitches were similar between 5-HTT+/+ and 5-HTT+/− mice. Pretreatment with PCPA, which depletes 5-HT levels, increased DOI-induced head twitches in 5-HTT−/− mice, with a trend toward a significant increase in 5-HTT+/− mice (p=0.08), compared to their vehicle-pretreated counterparts. Data represent the mean±SEM, n=7–13 per group. **p<0.01 compared to 5-HTT+/+ mice in the same pretreatment condition; ++++p<0.0001 compared to vehicle-pretreated mice of the same 5-HTT genotype.

Administration of the 5-HT precursor, 5-HTP, is reported to increase 5-HT syndrome behaviors (Fox et al, 2008a, 2007b) and brain 5-HT concentrations 2-5-fold in 5-HTT+/+ and 5-HTT+/− mice and 4.5–12-fold in 5-HTT−/− mice (Fox et al, 2008a). 5-HTT−/− mice given 5-HTP displayed 48% fewer HTR than did 5-HTT+/+ mice (p=0.037), whereas there was no difference between 5-HTT+/+ and 5-HTT+/− mice given 5-HTP (Figure 4; main effect of genotype (F(2,36)=4.24, p=0.022).

Figure 4
figure 4

5-HTP-induced head twitches in 5-HTT mice. The 5-HT precursor 5-HTP, which increases 5-HT levels, induced fewer head twitches in 5-HTT−/− mice compared to 5-HTT+/+ mice, with no difference between 5-HTT+/− and 5-HTT+/+ mice. Data represent the mean±SEM; n=9–16 per group. *p<0.05 compared to 5-HTT+/+ mice.

DISCUSSION

Baseline AA incorporation coefficients k* were increased significantly in 5-HTT−/− mice in 72 of 92 regions by 21–71%, and in 5-HTT+/− mice in 45 regions by 20–67%, compared to 5-HTT+/+ mice. Comparable increases were found for Jin as well. The increases were accompanied by elevated brain cPLA2 activity (29 and 35% in 5-HTT+/− and 5-HTT−/− mice, respectively), decreased COX activity (−49 and −74%, respectively) and decreased concentrations of the COX-derived AA metabolites, PGE2 (−74 and -90%, respectively), PGF2α (−24 and −35%, respectively), and TXB2 (−34 and −72%, respectively). The partial 5-HT2A/2C agonist DOI increased k* in 42 regions in wild-type mice, but decreased k* in 31 and 83 regions, respectively, in the 5-HTT+/− and 5-HTT−/− mice. DOI-induced HTR were reduced in 5-HTT−/− mice, but this decreased response was ‘normalized to 5-HTT+/+ levels after 5-HT depletion by pretreatment with PCPA.

Together, these studies suggest that elevated extracellular 5-HT levels in 5-HTT deficient mice, by increasing 5-HT occupancy of PLA2-coupled postsynaptic 5-HT2A/2C receptors, likely tonically activate cPLA2 and increase AA release from membrane phospholipid, thereby increasing baseline values of k* and Jin for AA. Tonic activation of cPLA2-coupled neuroreceptors also has been reported to increase cPLA2 expression (mRNA and activity) in rats treated chronically with a subconvulsive dose of N-methyl-D-aspartic acid (NMDA) to stimulate NMDA receptors (Lee et al, 2008; Rao et al, 2007), or with fluoxetine, a 5-HTT inhibitor, to stimulate 5-HT2A/2C receptors through elevated extracellular 5-HT (Lee et al, 2007; Qu et al, 2006; Stenfors and Ross, 2002). In these and the present case, excess neuroreceptor-induced AA release may have activated protein kinase C and nuclear transcription factor-κB to transcriptionally upregulate cPLA2 expression by a feedback mechanism (Toborek et al, 1999; Xu et al, 2002).

Although the increased Jin for AA in the deficient mice represented increased AA loss by brain metabolism (Demar et al, 2005; Holman, 1986; Rapoport et al, 2001), the reduced brain COX activity and PGE2, PGF2α, and TXB2 concentrations indicate that AA loss was not through COX-mediated pathways, but by other pathways, such as β-oxidation, formation of endocannabinoids, or oxidation by cytochrome P450 epoxygenase or lipoxygenase (Fitzpatrick and Soberman, 2001; Shimizu and Wolfe, 1990). Although there is a limitation that we do not know the exact pathways of increased loss, these might be determined in the future by measuring brain COX-2, COX-1, 5-lipoxygenase and cytochrome P450 epoxygenase activities and their metabolic products in the 5-HTT-deficient mice, in relation to altered behavior (Fox et al, 2007a; Murphy and Lesch, 2008). In this regard, endocannabinoids derivatives, such as anandamide, can induce anxiety-like behaviors in rodents (Rubino et al, 2008; Rutkowska et al, 2006). We do not have a ready explanation for the reductions in COX activity and in eicosanoid concentrations, in the face of increased cPLA2 activity, but such ‘uncoupling’ of the enzymes also was noted in COX-2 knockout mice (Bosetti et al, 2004; Zhang et al, 2002). cPLA2 and COX-2 normally are functionally coupled and colocalized on postsynaptic membranes in rodent brain (Bosetti and Weerasinghe, 2003; Fitzpatrick and Soberman, 2001; Kaufmann et al, 1996; Ong et al, 1999; Pardue et al, 2003; Sapirstein et al, 2005; Xu et al, 2008).

The decreased k* and Jin responses to DOI in the 5-HTT-deficient mice are not due to reductions in 5-HT2A/2C receptor density or their availability due to internalization, as binding studies indicate that 5-HT2A receptor density is reduced only in the striatum but is increased in the hypothalamus and septum of the mice, whereas 5-HT2C receptor density is elevated in the amygdala and choroid plexus (Li et al, 2003). The hypothalamus, septum, and amygdala belong to the limbic system, which is involved in emotional regulation. Altered 5-HT2A/2C receptors and their signaling may contribute to some of the behavioral changes observed in these mice, such as increased anxiety-like behaviors and reduced aggressiveness on various tests (Fox et al, 2007a; Murphy et al, 2008; Murphy and Lesch, 2008).

One possibility for the decreased k* and Jin responses to DOI is that, as a partial agonist, DOI displaced bound 5-HT from cPLA2-coupled 5-HT2A/2C receptors, and produced less activation compared with 5-HT (Marek and Aghajanian, 1996). Such displacement also can explain the decreased DOI-induced HTR in 5-HTT−/− mice, replicating an earlier report (Qu et al, 2005), as PCPA pretreatment, sufficient to deplete extracellular 5-HT by 67–94% in wild-type or 5-HTT−/− mice (Cesana et al, 1993; Fox et al, 2008b), returned the DOI-induced HTR in 5-HTT−/− mice to wild-type levels. Consistent with this interpretation, 5-HTT overexpressing mice have lower levels of extracellular 5-HT and higher DOI-induced HTR than do wild-type mice (Jennings et al, 2008). Postnatal PCPA administration is reported to prevent some aspects of the adult 5-HTT−/− behavioral phenotype (Alexandre et al, 2006), and it would be worthwhile to see whether it also prevented some of the differences in the AA signal (Fox et al, 2007a).

The serotonin precursor 5-HTP, which increases 5-HT levels in 5-HTT-deficient mice (Fox et al, 2008a), induces HTR in mice. In this study, 5-HTP induced fewer HTR in 5-HTT−/− vs 5-HTT+/+ mice, indicating that excessive baseline levels of synaptic 5-HT increases HTR. Brain extracellular concentrations of dopamine, glutamate, and acetylcholine are unchanged in 5-HTT-deficient mice and rats (Homberg et al, 2007; Mathews et al, 2004), and likely did not contribute to the elevations in k* and Jin for AA in the 5-HTT-deficient mice.

In the 5-HTT+/+ mice, statistically significant elevations in k* for AA in response to DOI occurred in the neocortex, mesencephalon, and rhombencephalon, which have high 5-HT2A receptor densities. The olfactory tubercle, hypothalamus, amygdala, hippocampus, and choroid plexus, which contain mainly 5-HT2C receptors (Li et al, 2003) were not activated significantly, suggesting that the k* responses to DOI were mediated mainly by 5-HT2A receptors. However, stimulation of 5-HT2A or 5-HT2C receptors by different agonists can activate cPLA2 to release AA (Berg et al, 1998b). Selective 5-HT2A and 5-HT2C antagonists could be used to distinguish the roles of the two receptor subtypes.

Dorsal raphe neurons of 5-HTT−/− and 5-HTT+/− mice exhibit decreased firing rates (Murphy and Lesch, 2008), and 5-HTT−/− mice have widespread reductions in brain glucose metabolism, a measure of energy consumption (Esaki et al, 2005; Sokoloff, 1999), despite their elevated values of k* and Jin. The obesity of the 5-HTT−/− mice is consistent with published data and is associated with increased plasma levels of insulin, leptin, cholesterol, and triglycerides (Murphy and Lesch, 2008). Their high plasma unesterified fatty-acid concentrations may be related to adrenocorticotropic hormone and corticosterone elevations and to anxiety-related behaviors (Gottschalk et al, 1969; John et al, 1987; Murphy and Lesch, 2008).

The baseline values of k* for AA in the 5-HTT+/+ mice agree with published values (Basselin et al, 2006b; Qu et al, 2005). An earlier study reported that baseline values did not differ significantly between 5-HTT−/− and 5-HTT+/+ mice, unlike the current findings. However, the earlier study is likely erroneous. Frozen brain sections in the earlier study were exposed to X-ray film for 6–10 weeks, rather than for 4 weeks as was carried out in this study, which resulted in saturation at high optical densities, with loss of linearity and discrimination (Basselin M, unpublished observations).

The 5-HTT+/− mouse is considered a model for humans who carry the ‘S’ compared with ‘L’ allele of the 5-HTTLPR, or who express rs25531 or rs25532 variants of the 5-HTT allele with regard to the levels of 5-HTT expression and function (Murphy et al, 2008; Murphy and Lesch, 2008). Individuals with these lesser-expressing 5-HTT polymorphisms are at risk for multiple psychiatric disorders, including bipolar disorder (Masoliver et al, 2006; Murphy et al, 2008; Murphy and Lesch, 2008). In this regard, elevated brain AA metabolism has been suggested to contribute to and be a risk factor for bipolar disorder (Basselin et al, 2006a, 2007b; Rao et al, 2008; Rapoport and Bosetti, 2002).

The new data in this study suggest that baseline k* and Jin for AA would be elevated in individuals with the ‘S’ 5-HTTLPR allele. This prediction might be tested with the use of positron emission tomography (PET) in human subjects in the resting state, following the i.v. injection of [1-11C]AA. In such PET studies, the coefficient of variation of k* for AA ranges from 12 to 16% (Esposito et al, 2008; Giovacchini et al, 2004). A power analysis (http://statpages.org/#Power) shows that statistically significant (α=5, statistical power=0.8) resting-state differences of 20% (the lower range of significant elevations in the 5-HTT-deficient mice) thus could be shown in eight subjects each belonging to long and short 5-HTTLPR groups. To date, PET studies assessing acute drug responses of the AA signal have not been performed in humans, but in any case, if DOI were given, it would be expected to produce untoward hallucinogenic effects (Marek and Aghajanian, 1996).

The present findings extend our knowledge of altered neurotransmission involving 5-HT, which is thought to contribute to depression and anxiety disorders. The data showed that reduced or absent 5-HTT function in mice results in the upregulation of baseline AA signaling involving 5-HT. Given the parallels between phenotypic abnormalities in 5-HTT-deficient mice and in human mood and anxiety disorders, these data provide a model for humans with 5-HTT polymorphisms and mutations that affect 5-HTT expression and function.