The experiments were approved by the local ethical committee at Karolinska Institute (N351/08) and conducted in accordance with the European Communities Council Directive of 24 November 1986 (86/609/EEC). Adult male WT and TAAR1 KO mice on a C57Bl6 background were used (Di Cara et al., 2011). They were housed in temperature- and humidity-controlled rooms (20°C, 53% humidity) with a 12 h dark/light cycle. They had access to standard lab pellets and water ad libitum.

Mouse brains were rapidly removed and placed in ice-cold, oxygenated (95% O₂/5% CO₂) artificial cerebrospinal fluid (aCSF) containing (in mM): 126 NaCl, 2.5 KCl, 1.2 NaH₂PO₄, 1.3 MgCl₂, 2.4 CaCl₂, 10 glucose and 26 NaHCO₃, pH 7.4. Coronal slices (300 μm thick) were prepared using a Leica vibratome (Leica, Wetzlar, Germany). Dorsal striata were dissected from the slices in ice-cold aCSF buffer. Each slice was placed in a polypropylene incubation tube with 2 ml fresh aCSF buffer. The slices were preincubated at 30°C under constant oxygenation (95% O₂/5% CO₂) for 60 min with a change of buffer after 30 min. The buffer was then replaced with fresh aCSF and slices were treated with tyramine (1, 10, 100 μM; Sigma-Aldrich, St. Louis, MO, United States), T₁AM (1, 10, 100 μM; synthesized by Servier, kind gift from Mark J. Millan), β-PEA (100 μM; Sigma-Aldrich), SCH23390 (5 μM; Sigma-Aldrich), EPPTB (10 nM, synthesized by Servier, kind gift from Mark J. Millan), KN-92 (10 μM; Sigma-RBI), H-89 (10 μM; Calbiochem, Gibbstown, NJ, United States) and 8-CPT-2Me-cAMP (10 μM; Tocris Bioscience, Bristol, United Kingdom), alone or in combination. The higher doses of all compounds exceeds by far the IC₅₀ or Kd values for their respective target, but it is known that much higher concentrations are needed to exert actions in brain slices when compared to cell culture systems (Nishi et al., 1997). After the drug treatment, the buffer was removed, the slices rapidly frozen on dry ice and stored at -80°C until assayed.

Immunoblotting was performed as described earlier (Qi et al., 2009). Frozen tissue samples were sonicated in 1% SDS, transferred to Eppendorf tubes and boiled for additional 10 min. Small aliquots of the homogenate were retained for protein determination using the bicinchoninic acid protein assay method (Pierce, Rockford, IL, United States). Equal amounts of protein (20 μg) were loaded onto 12% acrylamide gels, and the proteins were separated by SDS-PAGE and transferred to Immobilon^®-P Polyvinylidene Difluoride membranes (Sigma). Immunoblotting was performed on the membranes using P-Ser¹⁹-TH (Merck Millipore, Billerica, MA, United States), P-Ser³¹-TH (Millipore), P-Ser⁴⁰-TH (Millipore), P-Ser⁸⁴⁵-GluA1 (UBI), and antibodies, which are not phosphorylation state-specific to estimate total levels of TH (Millipore) and GluA1 (UBI). The antibody binding was detected by incubation with goat anti-mouse or anti-rabbit horseradish peroxidase-linked IgG (1:6000–8000 dilution) and detected using ECL immunoblotting detection reagents (GE Healthcare, Little Chalfont, United Kingdom).

Dorsal striatal slices were incubated for 5 min with T₁AM (10 μM) or tyramine (100 μM), and then for 15 min with T₁AM or tyramine along with the L-amino acid decarboxylase inhibitor NSD-1015 (100 μM, Sigma-Aldrich). After the removal of the solutions, tissue slices were frozen and sonicated (10,000 g for 10 min) in 100 μL perchloric acid (0.1 mM). The pellets were resuspended in 100 μl 1% sodium dodecyl sulfate and the protein content was determined. The level of L-DOPA in the supernatant was determined using HPLC coupled to an electrochemical detection system with a refrigerated microsampling unit (model CMA/200; CMA Microdialysis, Kista, Sweden). The HPLC apparatus comprised an HPLC pump (model 2150; Pharmacia LKB Biotechnology AB, Uppsala, Sweden) that kept a constant flow of 0.2 mL/min of the mobile phase (0.12 m NaH₂PO₄H₂O; 0.09 m EDTA, 0.05 mm 1-octanesulfonic acid, and 15% methanol, pH 4.2) and a pressure of ∼0.5 bar on a reverse-phase ion pair C-18 column prepacked with Biophase ODS 5 μm particles (BAS, West Lafayette, IN, United States). L-DOPA was detected with an amperometric detector (model LC-4C; BAS) and a glassy carbon electrode set at 0.75 V. The limit of detection was ∼10 nM.

Sagittal striatal brain slices were prepared and maintained as above. Amperometric detection of DA release was performed as described earlier (Zhang et al., 2014a). Carbon fiber electrodes (10 μm in diameter, World Precision Instruments, Hertfordshire, England) had an active part (100 μm) that was positioned within the dorsal striatum in the brain slice. A constant voltage of +500 mV was applied to the carbon fiber via an Axopatch 200B amplifier (Axon Instruments) and currents were recorded with the same amplifier. A stimulating electrode (patch electrode filled with aCSF) was placed on the slice surface, in the vicinity of the carbon fiber electrode. Stimulation consisted of a single pulse (0.1 ms, 8–14 μA) applied every minute, which evoked a response corresponding to oxidation of DA at the surface of the electrode. When the carbon fiber electrode was held at 0 mV, stimulation of the slice did not produce any current.

Adult male WT and TAAR1 KO mice were injected with saline or T₁AM (20 mg/kg, i.p.) and killed by decapitation 30 or 60 min post-dose. All brains were immediately removed, snap frozen, and stored at -80°C until further analysis. The frozen brain tissues were cut using a cryostat-microtome (Leica CM3050S; Leica Microsystems, Welzlar, Germany) at a thickness of 14 μm, thaw-mounted onto conductive indium tin oxide (ITO) glass slides (Bruker Daltonics), and stored at -80°C. Sections were dried gently under a flow of nitrogen and desiccated at room temperature for 15 min, after which they were imaged optically using a photo scanner (Epson perfection V500). The samples were then coated with derivatization reagents, 2, 4-diphenylpyrylium tetrafluoroborate (DPP-TFB). Stock solution of DPP-TFB (8 mg in 1.2 ml MeOH) was prepared and diluted in 6 mL of 70% methanol containing 3.5 μL of trimethylamine. An automated pneumatic sprayer (TM-Sprayer, HTX Technologies, Carrboro, NC, United States) was used to spray DPP-TFB solution over the tissue sections. The nozzle temperature was set at 80°C and the reagent was sprayed for 30 passes over the tissue sections at a linear velocity of 110 cm/min with a flow rate of about 80 μL/min. Samples were then incubated for 15 min (dried by nitrogen flow every 5 min) in a chamber saturated with vapor from a 50% methanol solution. MALDI-MSI experiment was performed using a MALDI-TOF/TOF (Ultraflextreme, Bruker Daltonics, Bremen, Germany) mass spectrometer with a Smartbeam II 2 kHz laser in positive ion mode. The laser power was optimized at the start of each run and then held constant during the MALDI-MSI experiment.

Autoradiograms from western blotting experiments were digitized using a Dia-Scanner (Epson Perfection 4870 PHOTO). Optical density values were measured using NIH Scion Image for Windows (alfa 4.0.3.2; © 2000–2001 Scion Corporation). Biochemical data were analyzed using one-way ANOVAs followed by Newman–Keuls post hoc test. Data from Amperometry were acquired and analyzed with the pClamp 9 or pClamp 10 software (Axon Instruments). Data are expressed as % of the baseline response measured for each slice during the 5–10 min preceding start of perfusion with T₁AM. Statistical significance of the results was assessed by using Student’s t-test for paired observations (comparisons with baseline within single groups) or one-way ANOVA multiple comparison test followed by Newman–Keuls post hoc test since the samples from WT and KO mice were loaded in separated gels. The numbers of individual replicates are shown in the graphs, while p-values, degrees of freedom and F values are detailed in the results part.

We first studied the dose responses of the tyramine and T₁AM on the phosphorylation of TH and GluA1 in striatal slices from TAAR1 receptor WT and KO mice. To study effects of compounds in the both genotypes, their individual baseline was set at 100%. One way ANOVA analysis showed that tyramine caused a significant change of P-Ser⁴⁰-TH in both WT (F_[3,20]: 60.28; p < 0.0001) and KO mice (F_[3,20]: 9.928, p = 0.0003). Post hoc test showed that the lower (1 and 10 μM) concentrations of tyramine did not have any significant effect, whereas 100 μM tyramine significantly reduced phosphorylation of P-Ser⁴⁰-TH in both groups of mice, suggesting an effect independent of TAAR1 (Figure 1A). Meanwhile, there were no effects at the same concentration of tyramine on P-Ser¹⁹-TH (WT: F_[3,20]: 2.913; p = 0.06; KO: F_[3,20]: 2.718; p = 0.07) and P-Ser³¹-TH (WT: F_[3,20]: 0.9955; p = 0.42; KO: F_[3,20]: 0.9157; p = 0.45) in neither WT nor KO mice (Figure 1A). One way ANOVA analysis in T₁AM treated mice revealed that only in the WT but not in the KO mice, the drug affected significantly the levels of P-Ser¹⁹ (WT: F_[3,20]: 3.641; p = 0.0004; KO: F_[3,20]: 0.1953, p = 0.8983) and P-Ser⁴⁰-TH (WT: F_[3,16]: 4.393, p = 0.02; KO: F_[3,16]: 0.823, p = 0.5001). Post hoc test showed that 10 μM T₁AM enhanced P-Ser¹⁹ and P-Ser⁴⁰-TH in the striatum of WT but not TAAR1 KO mice (Figure 1B). The effects of T₁AM were biphasic with a further increase in drug concentrations resulting in less phosphorylation of TH. There was no effect of T₁AM on P-Ser³¹-TH both in WT (F_[3,20]: 0.7911, p = 0.5131) and KO mice (F_[3,20]: 0.446, p = 0.7228) (Figure 1B). Tyramine altered significantly the phosphorylation of P-Ser⁸⁴⁵-GluA1 in WT (F_[3,20]: 9.387; p = 0.0004) and KO mice (F_[3,20]: 22.89; p < 0.0001) (Figure 1C). In contrast, T₁AM had no effect on P-Ser⁸⁴⁵-GluA1 in striatal slices from either TAAR1 WT (F_[3,20]: 0.8744; p = 0.4709) or KO mice (F_[3,20]: 0.4367; p = 0.7295) (Figure 1D). These data suggest that tyramine and T₁AM act differently on pre- and post-synaptic striatal targets.

To further study the effect of a high dose of endogenous TAs, we incubated striatal slices from WT mice with the tyramine (100 μM), β-PEA (100 μM) alone or with the D₁ receptor antagonist SCH23390 (5 μM). As shown in Figure 2, we found that β-PEA, like tyramine, decreased P-Ser⁴⁰-TH (F_[5,18]: 68.88; p < 0.0001) while not significantly affecting P-Ser¹⁹-TH (F_[5,18]: 1.442; p = 0.2573) or P-Ser³¹-TH (F_[5,18]: 1.18; p = 0.3574). The effects of tyramine and β-PEA on P-Ser⁴⁰-TH were not affected by SCH23390. On the other hand, tyramine significantly enhanced P-Ser⁸⁴⁵-GluA1 (F_[5,18]: 2.455; p < 0.0001), an effect that was reversed to baseline by D₁ receptor blockade using SCH23390. Likewise, β-PEA tended to increase P-Ser⁸⁴⁵-GluA1, but this effect did not reach significance. The data of β-PEA was further to confirm that classical TAs and T₁AM act differently.

There was also a baseline increase of TH activity in TAAR1 KO mice as compared to WT mice (Di Cara et al., 2011). To study effects of compounds in the both genotypes, their individual baseline was set at 100%. One way ANOVA revealed significant difference among the groups in WT mice (F_[2,18]: 6.856; p = 0.0061) (Figure 3). Post hoc test showed that T₁AM (10 μM), but not tyramine (100 μM), induced L-DOPA accumulation in the presence of a DOPA decarboxylase inhibitor in WT mice. No effects were detected in TAAR1 KO mice (F_[2,18]: 0.08823; p = 0.9159), indicative of a TAAR1-mediated mechanism of action of T₁AM.

To further study intracellular signaling cascades underlying the T₁AM-induced phosphorylation of TH, we combined the T₁AM with KN-92 or H-89, inhibitors of CamKII and protein kinase A (PKA), respectively. The effects of T₁AM on both P-Ser¹⁹ (F_[7,75]: 3.651; p = 0.0019) and P-Ser⁴⁰-TH (F_[7,75]: 3.871; p = 0.0012) could be significantly inhibited by either KN-92 or H-89 (Figure 4). Since TAAR1 is a Gs-coupled receptor and generates cAMP, we also examined the effects of 8-CPT-2Me-cAMP, an EPAC (Exchange Protein directly Activated by cAMP) activator, on TH phosphorylation. 8-CPT-2Me-cAMP alone tended to increase TH phosphorylation, but did not interact with T₁AM (Figure 4).

We evaluated the effect of T₁AM on stimulation-evoked release of DA from DA-containing fibers present in sagittal striatal slices, as shown in Figure 5. We found that bath application of T₁AM (10 μM) significantly increased the amplitude of evoked DA release measured with carbon fiber electrodes coupled to amperometry in dorsal striatal brain slices from WT mice, and that this effect was significantly reduced in TAAR1 KO mice. In presence of the TAAR1 antagonist EPPTB (10 nM), the effect of T₁AM on evoked DA release in WT mice was significantly reduced, but not completely blocked (F_[2,20]: 7,252; p = 0.0043). EPPTB had no effect on evoked DA release by itself (data not shown). Similar to the results as above, T₁AM significantly increased P-Ser⁴⁰-TH (F_[3,30]: 3.384; p = 0.0309). This effect was blocked when T₁AM was combined with EPPTB, suggesting a mechanism of action mediated via TAAR1.

T₁AM was derivatized by 2, 4 diphenyl pyranylium and detected by MALDI-MSI. No clear endogenous signal of T₁AM was found in uninjected sections neither from WT nor TAAR1 KO mice. However, widespread signals corresponding to derivatized T₁AM was detected in mice intraperitoneally administered with T₁AM (20 mg/kg). The concentration of the drug appeared higher after 30 min compared to 60 min post-dose (Figure 6).

Alterations of DA synthesis are regulated via phosphorylation of TH, the rate-limiting enzyme in the synthesis of catecholamines (Daubner et al., 2011), and Ser¹⁹, Ser³¹, and Ser⁴⁰ have been identified as the functionally most important sites of TH phosphorylation (Haycock and Haycock, 1991). Phosphorylation of Ser¹⁹ is induced by enhanced intracellular Ca²⁺ concentrations and activation of CaM kinase II, whereas phosphorylation at Ser³¹ is induced by extracellular signal-regulated protein kinases, and phosphorylation of Ser⁴⁰ is catalyzed by PKA (Haycock, 1993). TH phosphorylation at Ser⁴⁰ and Ser³¹ leads to increased TH activity, whereas phosphorylation at Ser¹⁹ exerts a positive modulatory influence on Ser⁴⁰ phosphorylation (Bobrovskaya et al., 2004; Dunkley et al., 2004). Notably, TAAR1 KO animals exhibit a basal increase in TH activity and increased basal phosphorylation at Ser¹⁹, Ser³¹, and Ser⁴⁰ in striatal slices compared to WT animals, perhaps due to developmental compensations (Di Cara et al., 2011). An increased TH activity at Ser¹⁹, Ser³¹, and Ser⁴⁰ was observed in WT mice following administration of 1 to 10 μM T₁AM, but this effect was attenuated upon increasing or lowering the concentration, indicating a bell-shaped dose–response. The effects of T₁AM on TH activity were abolished in TAAR1 KO mice, identifying TAAR1 as a mediator of these actions. TAAR1 is coupled with stimulatory G_s proteins as well as G protein independent pathways and, upon activation, TAAR1 signals through the cAMP/PKA/CREB, β-arrestin2/Akt/GSK-3β and the protein kinase C (PKC)/Ca++/NFAT pathways (Borowsky et al., 2001; Bunzow et al., 2001; Panas et al., 2012; Harmeier et al., 2015). Here we found that the T₁AM-mediated increases on Ser¹⁹ and Ser⁴⁰ TH were inhibited by blockade of CamKII and PKA by either KN-92 or H-89, respectively. The protein responsible for the phosphorylation of TH at the site Ser³¹ is MAPK (Haycock et al., 1992). As a consequence we suppose that the TAAR1’s activation by T₁AM stimulates the activity of MAPK, either through PKA or through an alternative direct pathway like β-arrestin2 (Pierce and Lefkowitz, 2001). In order to confirm that the observed phosphorylation of TH leads to increased enzymatic activity, we made measurements of DOPA with HPLC in striatal slices. T₁AM induced a higher level of DOPA accumulation in the presence of a DOPA decarboxylase inhibitor in WT mice. This effect of T₁AM was abolished in KO counterparts. Figure 7 shows a schematic, and somewhat speculative, drawing of the proposed signaling pathway induced by T₁AM/TAAR1/PKA activation. TAAR1 is a G_s protein-coupled receptor and activation results in increased cAMP via activation of adenylyl cyclase and PKA signaling, which directly phosphorylates TH at Ser⁴⁰. Another pathway could involve activation of inhibitor 1 by PKA, which inhibits protein phosphatase 1 (PP-1). PP-1 inhibits the phosphorylation of CamKII, which activates phosphorylation of Ser¹⁹-TH. CamKII can also inhibit protein phosphatase 2A which activates phosphorylation of Ser⁴⁰-TH. The regulation of the suggested signal transduction pathways could explain how KN-92 and H-89, inhibitors of CamKII and PKA respectively, both can block the effect of T₁AM. 8-CPT-2Me-cAMP, the EPAC analog, had no influence on TH phosphorylation. In summary, this higher phosphorylation rate leads to the accumulation of DOPA.

In contrast to T₁AM, under the present conditions, the endogenous TAs tyramine and β-PEA appeared to elicit mainly non-TAAR1 dependent effects. Tyramine, at a high dose, reduced phosphorylation of Ser⁴⁰-TH, but increased phosphorylation of the post-synaptic AMPA receptor GluA1 subunit in striatal slices from both WT and TAAR1 KO mice. Similarly, another endogenous TA, β-PEA also reduced phosphorylation of Ser⁴⁰-TH and tended to increase phosphorylation of Ser⁸⁴⁵-GluA1 in slices from WT mice. The quantification of GluA1 phosphorylation on Ser⁸⁴⁵, help us to evaluate the status of GluA1 in the post-synaptic membrane of corticostriatal and thalamostriatal synapses, which are the main classes of glutamatergic synapse in the striatum (Smith et al., 2004). The corticostriatal and thalamocortical projection neurons innervate the principal population of MSNs, but also various subtypes of interneurons (Smith et al., 2004). By measuring the phosphorylation of GluA1 we cannot rule out which types of interneurons and MSNs (D₁ positive or D₂ positive) are activated. Nevertheless, it is shown that D₁ agonists and D₂ antagonists induce robust increases in GluA1 phosphorylation while D₂ agonists and D₁ antagonists have no effect (Xue et al., 2017). Considering this fact, we deduce that the effect of tyramine on GluA1 trafficking could be explained by either the post-synaptic regulation of D₁ or D₂ receptors. D₁ receptor blockade by SCH23390 blocked the tyramine-induced phosphorylation of GluA1 subunits, but had no effect on tyramine- or β-PEA-reduced Ser⁴⁰-TH phosphorylation, suggesting that the post-synaptic effects of tyramine are dependent on D₁-receptor activation. According, GluA1 upregulation may be a consequence of the dopamine’s net effect upon D₁ and D₂ receptors (Xue et al., 2017). In our study, we observed that tyramine attenuated the TH phosphorylation at Ser⁴⁰. However, diminished enzymatic activity of TH does not mean a reduction in dopamine release, whilst a negative feedback mechanism for dopamine control of TH activity has been documented (Lindgren et al., 2001; Daubner et al., 2011). It is conceivable that tyramine has amphetamine-like effects on the excitability of the post-synaptic membrane and possibly leads to the vesicular leak of dopamine by its interaction with VMAT2 (Zhu et al., 2007). Moreover, it has been suggested that TAs act like amphetamines and could increase extracellular DA levels by promoting DA release via inducing reversal of the dopamine transporter (DAT) and by displacing DA from vesicular stores (Sulzer et al., 1995; Jones et al., 1998; Mundorf et al., 1999). β-PEA, which is structurally related to amphetamine, has been proposed to act as an endogenous amphetamine (Janssen et al., 1999), and has previously been shown to increase extracellular levels of DA in striatum and NAc via a DAT-dependent mechanism (Sotnikova et al., 2004; Murata et al., 2009). Notably, Xie and Miller found that TAs, including tyramine and β-PEA, do not directly activate monoamine autoreceptors (Xie and Miller, 2008). However, they have been proposed to indirectly activate dopamine autoreceptors by enhancing the efflux of dopamine (Geracitano et al., 2004). One possible explanation for tyramine’s TAAR1 independent effect on GluA1 phosphorylation, may be that this TA acts through MSN-localized TAAR1 to affect the availability of GluN1 and through VMAT2 to alter the surface density of GluA1. Indeed, several studies have supposed that tyramine can affect glutamate receptor membrane availability through MSN-localized TAAR1 (Alvarsson et al., 2015; Espinoza et al., 2015a; Sukhanov et al., 2016). To conclude, our data support the notion that T₁AM can modulate DA synthesis via a mechanism of action that involves presynaptic TAAR1. Moreover, we suppose a direct effect of tyramine on dopamine release that could lead to the observed decline in TH phosphorylation due to secondary activation of indirect D₂ autoreceptors (Lindgren et al., 2001).

Our experiments suggest that striatal dopamine release can be enhanced by T₁AM-mediated TAAR1 activation. However, most previous slice experiments addressing the modulatory influence of TAAR1 on the dopaminergic system have been performed in the VTA (Lindemann et al., 2008; Revel et al., 2012), which may be a source of discrepancy between our study and previous findings. Previous slice experiments in the VTA of TAAR1 KO mice revealed enhanced spontaneous firing rates of dopaminergic neurons in TAAR1 KO mice compared to WT mice, suggesting that TAAR1 exerts an attenuating effect on dopaminergic neuron activity (Lindemann et al., 2008). This has been supported by slice experiments in mouse VTA using the specific ligands RO5166017 and EPPTB to stimulate and block TAAR1, respectively (Bradaia et al., 2009; Revel et al., 2011). However, the increased DA neuron firing rate observed in TAAR1 KO mice did not lead to enhanced basal levels of extracellular striatal DA compared to WT mice as detected by microdialysis (Lindemann et al., 2008). Indeed, mice overexpressing TAAR1, like TAAR1 KO mice, also exhibit an enhanced spontaneous firing activity of monoaminergic neurons of the VTA, DRN, and locus coeruleus (Revel et al., 2012). Moreover, it is likely that the functional outcome of TAAR1 activation differs between specific classes of ligand and distinct brain regions depending on the characteristics of the dopaminergic innervation and basal tone. Although midbrain DA neurons are considered to be relatively homogenous, emerging data support a high level of diversity among VTA and SNc neurons as regards electrophysiological properties, synaptic connectivity, protein expression profiles, and behavioral functions (Roeper, 2013). The expression levels of two TAAR1 related proteins, D₂ receptor and GIRK2, are implicated in the differences between the two subpopulations (Bradaia et al., 2009; Reyes et al., 2012; Bifsha et al., 2014; Brichta and Greengard, 2014). In agreement with this, it is reported that TAAR1 has a differential role in dopamine release between VTA and SNc projection sites in striatum (Leo et al., 2014). In contrast, Di Cara et al. (2011) showed that TAAR1 decreases the amplitude of Methylenedioxymethamphetamine (MDMA) induced dopamine release both in ventral and dorsal striatum. In the same study it was observed that the TAAR1 agonist, o-phenyl-3-iodotyramine (o-PIT) blunted the para-chloroamphetamine (PCA) induced dopamine release in both structures (Di Cara et al., 2011). Accordingly, TAAR1 may exert a complex pattern of effects on dopaminergic terminals in ventral as compared to dorsal compartments of the striatum. Furthermore, both the VTA and SNc can be further distinguished regarding the expression of calbindin D28k (CB), with the highest density of CB positive neurons located in VTA (Surmeier et al., 2017). CB positive neurons have the tendency to send projections in CB poor islands in striatum (striosomes), while CB negative cells mainly innervate CB rich regions of the striatum (striatal matrix) (Brimblecombe and Cragg, 2017; Sgobio et al., 2017). It has been reported that the evoked striatal DA release differs between these two compartments but also that the dopamine release ratio of striosome over matrix is higher in the ventral than dorsal striatum (Salinas et al., 2016). Consequently, TAAR1 could have diverse effects not only among VTA and SNc neurons but also between CB positive and negative subgroups.

Our findings with T₁AM, may also be explained by differences in the methodologies and protocols employed to evoke and measure dopamine release. For example, the use of strong stimulation intensities might evoke maximal release. Conversely, local, low intensity stimulation, as used in the present study, allows for observation of both inhibition and potentiation of dopamine release. In addition, recent studies have demonstrated that dopamine release in brain slices can be evoked by direct stimulation of dopaminergic axons and indirectly by stimulation of cholinergic interneurons in the striatum (Threlfell et al., 2012; Zhang et al., 2014b). It has not yet been established whether cholinergic neurons express TAAR1 but the contrasting effects of RO5166017 and T₁AM might result from differences in the involvement of cholinergic control of dopamine release between different experimental paradigms. Taken together, these studies raise the question of a possible differential control of dopamine release by TAAR1 receptors in cholinergic interneurons and in dopamine axon terminals.

In this study, we investigated the action of T₁AM at TAAR1 on dopaminergic terminals as compared to those of TAs. However, T₁AM is also known to be an agonist of TAAR5 (Dinter et al., 2015c). Moreover, the β-phenylethylamine-like structure affords T₁AM the ability to bind with various members of GPCR superfamily and ion channels (Chiellini et al., 2017; Khajavi et al., 2017). It is indeed claimed that T₁AM interacts with α2a adrenergic receptors, β2-adrenergic receptors and muscarinic receptors (Kleinau et al., 2011; Dinter et al., 2015a,b; Laurino et al., 2016, 2017). Notably, outside the CNS, T₁AM has been found to differentially regulate insulin secretion through actions at TAAR1 and α2a adrenergic receptor (Chiellini et al., 2017; Lehmphul et al., 2017). Hence, despite blockade of the actions of T₁AM in KO mice and by pharmacological antagonist, the possibility that it exerts actions via other mechanisms should not be excluded.

In this study we incubated the slices in a T₁AM containing buffer. It is important to access the roles of T₁AM in the intact brain. We show here that T₁AM can be detected by MALDI-MSI in mouse brain slices 30 and 60 min after systemic administration. Since T₁AM was detected in many brain areas, we can conclude that T₁AM can penetrate the blood brain barrier. This finding is in accordance with previous studies showing effects on glucose metabolism by intraperitoneally administered T₁AM (Klieverik et al., 2009). Using MALDI-MSI, no clear endogenous levels of T₁AM could be detected. However, it will be interesting to study T₁AM levels in pathological states, particularly in hyperthyroid conditions. In addition to a circulating source, direct enzymatic transformation of T₄ to T₁AM may occur in neurons. The responsible enzyme for this reaction is ornithine decarboxylase (Hoefig et al., 2016), which is expressed by neuronal and astroglial cell types of the CNS (Bernstein and Müller, 1999). Apart from T₁AM itself, its metabolite 3-iodothyroacetic acid (TA1) is implicated in the modulation of histaminergic neurotransmission and might likewise interact with dopaminergic pathways: this remains to be clarified (Laurino et al., 2015).