Phenethylamine is a substrate of monoamine oxidase B in the paraventricular thalamic nucleus

Monoamine oxidase (MAO) is a key enzyme responsible for the degradation of neurotransmitters and trace amines. MAO has two subtypes (MAO-A and MAO-B) that are encoded by different genes. In the brain, MAO-B is highly expressed in the paraventricular thalamic nucleus (PVT); however, its substrate in PVT remains unclear. To identify the MAO-B substrate in PVT, we generated Maob knockout (KO) mice and measured five candidate substrates (i.e., noradrenaline, dopamine, 3-methoxytyramine, serotonin, and phenethylamine [PEA]) by liquid chromatography tandem mass spectrometry. We showed that only PEA levels were markedly elevated in the PVT of Maob KO mice. To exclude the influence of peripheral MAO-B deficiency, we developed brain-specific Maob KO mice, finding that PEA in the PVT was increased in brain-specific Maob KO mice, whereas the extent of PEA increase was less than that in global Maob KO mice. Given that plasma PEA levels were elevated in global KO mice, but not in brain–specific KO mice, and that PEA passes across the blood–brain barrier, the substantial accumulation of PEA in the PVT of Maob KO mice was likely due to the increase in plasma PEA. These data suggest that PEA is a substrate of MAO-B in the PVT as well as other tissues.

www.nature.com/scientificreports/ isoenzyme, and the inhibitors potentially act on mechanisms other than MAO inhibition. A gene knockout (KO) technique using the CRISPR/Cas9 system can be a more rigorous approach to identifying MAO-B substrates. Previous studies reported elevated PEA levels in the brain of Maob KO mice 27,28 , although they used wholebrain samples or bulk tissues. PEA is synthesized in DA neurons 29 and so-called "D-cells" or "D-neurons" that are immunoreactive for aromatic l-amino acid decarboxylase but contain neither tyrosine hydroxylase nor serotonin 30,31 . PEA, as well as tyramine, tryptamine, and octopamine, is called a "trace amine", and its concentration in the brain is several 100-fold lower than those of classical MA neurotransmitters 29,[32][33][34] . It is technically challenging to measure PEA levels in a small nucleus and, to the best of our knowledge, there are no reports on measurement of PEA concentration in the PVT. Furthermore, PEA is reportedly present in various food items and produced by bacteria, including gut microbiota [35][36][37][38] . These peripheral PEAs are absorbed into the bloodstream and readily pass across the blood-brain barrier (BBB) [39][40][41][42] .
In this study, we focused on the PVT while also targeting the seven other brain regions and successfully measured levels of PEA and other MAs in small areas, including the PVT, using advanced high-performance liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS) capable of highly sensitive measurements with small sample amounts. The concentrations of PEA and other MAs in the PVT were compared between wild-type (WT) mice and Maob KO mice in order to identify the MAO-B substrate in the PVT. Additionally, we developed brain-specific Maob KO mice, as well as global Maob KO mice, based on the ability of peripheral PEA to pass across the BBB.

Results
Generation of Maob KO mice and Maob flox mice. To develop Maob KO and Maob flox alleles, we cut the 5ʹ-and 3ʹ-flanking regions of the Maob gene in mouse fertilized eggs using the CRISPR/Cas9 system (Fig. 2a). The injected CRISPR/Cas9 cocktail contained two kinds of single-stranded oligonucleotides (ssODNs) comprising the lox2272 sequence and homology arms, which allowed for homology-directed repair to induce insertion of the lox2272 sequence at the two cut positions or non-homologous end-joining of the two cut ends to induce deletion of the Maob gene ( Supplementary Fig. S1). According to the results of genotyping PCR, we obtained four individuals harboring the Maob KO allele and four individuals harboring the Maob flox allele.  www.nature.com/scientificreports/ Sanger sequencing confirmed that the Maob gene was deleted or floxed with lox2272 ( Supplementary Fig. S1). We then performed whole-exome sequencing using one of the third final generation (F3) offspring from Maob KO or Maob flox lines and verified that they harbored no damaging mutations that might have occurred because of the off-target activity of Cas9. Reverse transcription-quantitative PCR analysis showed no Maob expression in any tissues collected from the global Maob KO mice (Fig. 2b). We then determined whether the 130-kb deletion affected the expression of adjacent genes, Maoa and Ndp. These mRNA levels normalized against either Actb (Fig. 2b), Gapdh, or calbindin 2 (Calb2; a PVT neuron marker) ( Supplementary Fig. S1) did not change significantly in any of the tissues.

PEA levels in the PVT and other brain regions are elevated in global Maob KO mice. Global
Maob KO mice and their WT littermates were anesthetized, and the brains were fixed with focused microwave irradiation to obtain a snapshot of steady-state substance (MAs and metabolites) levels 43 . We sampled the PVT region (φ1 mm × 2 mm thickness; Supplementary Fig. S2) and measured PEA, NA, DA, 3-methoxytyramine (3-MT), 5-HT, and 5-hydroxyindoleacetic acid (5-HIAA) using a highly sensitive and specific LC-MS/MS method. Representative chromatograms are presented in Supplementary Fig. S3. Two-way analysis of variance (ANOVA) with the main effect of genotype and substance showed a significant difference depending on the genotype × substance interaction (genotype: df = 1, F = 24.15, P = 3 × 10 −6 ; substance: df = 6, F = 39.11, P < 2 × 10 −16 ; and genotype × substance interaction, df = 6, F = 22.88, P < 2 × 10 −16 ). When comparing the genotypes for each substance, only PEA was markedly increased in Maob KO mice (P = 0.0014, Student's t-test with Bonferroni correction) (Fig. 3a). Seven other brain regions (caudate-putamen, NAc, Ins, paraventricular hypothalamic nucleus, occipital cortex, VTA, and DR) were also sampled and assayed ( Supplementary Fig. S2). PEA levels were elevated in the other brain regions, as well as in the PVT (Fig. 3b). No other MAs or metabolites in any brain region differed significantly between Maob KO and WT mice ( Supplementary Fig. S4).
The other MAs and metabolites were not significantly altered in brain-specific KO mice ( Supplementary Fig. S5).
Plasma PEA level is elevated in global Maob KO mice but not in brain-specific Maob KO mice. PEA passes across the BBB, and plasma PEA is distributed in the brain parenchyma 39,41 . The observed larger increase in PEA level in the PVT of global Maob KO mice could be attributed to increased plasma PEA levels induced by peripheral MAO-B deficiency. To test this possibility, we measured plasma PEA levels, as well as those of other MAs and metabolites, in global KO mice and brain-specific KO mice. Plasma PEA level was higher in global KO mice than in WT mice (P = 0.0325) (Fig. 5a), whereas plasma PEA level in brain-specific KO mice was not significantly increased relative to that in control Nestin-Cre mice (Fig. 5b). Additionally, the 5-HT concentration and 5-HT/5-HIAA ratio tended to increase in the plasma of global KO mice. These results suggest that PEA is metabolized by MAO-B in peripheral tissues, and that elevated levels of plasma PEA induced increases in PEA levels in brains of global Maob KO mice.

Discussion
MA neurotransmission profoundly affects a wide range of brain areas and nuclei, the functions of which are highly differentiated and interdependent. However, it is challenging to examine the MA metabolism of small nuclei, such as the PVT, whose volume is at most 1 mm 3 in mice. A recent study using imaging MS detected classical MAs at relatively high abundance in the PVT 24 , although PEA was not investigated. In the present study, we successfully measured trace amine PEA levels in small brain regions, including the PVT, using a microwave fixation method and high-sensitivity  www.nature.com/scientificreports/ from the periphery. However, the influence of the former is likely insignificant, because we did not detect any increase in PEA levels in brain tissues, except for the PVT, from brain-specific Maob KO mice (Fig. 4). This study illustrates the need for careful interpretation of PEA levels in the brains of global Maob KO mice or following systemic Maob inhibitor treatment.
A recent study reported that noradrenergic afferent terminals of the locus coeruleus to PVT projection neurons mediate DA release in the PVT 17 , which leaves open the possibility that NA and DA could also be MAO-B substrates in the PVT. Indeed, MAO-B is co-expressed in astrocytes along with vesicular monoamine transporter 2 (VMAT2) and organic cation transporter 3 and could metabolize DA under conditions of VMAT2 dysfunction 9 . Additionally, DA is also converted to 3-MT ( Fig. 1), which could be another candidate MAO-B substrate. Moreover, 5-HT levels in the brains of Maoa/Maob double KO mice were higher than those of single Maoa KO mice, suggesting that MAO-B partly contributes to 5-HT metabolism 44 . However, the present study did not identify the metabolisms of these MAs by MAO-B in the PVT ( Fig. 3a and Supplementary Fig. S5). Furthermore, the lack of DA alteration in Maob KO mice is consistent with the findings that DA is metabolized predominantly by MAO-A rather than MAO-B in rodents 1,[45][46][47][48] . Analysis of metabolites is important for investigating NA and DA metabolism, however, these concentrations in the PVT, which is a small region, were not determined in this study.
This study has several limitations. First, the genetic KO system, either global or brain-specific, deleted the Maob gene from the immediate-early stage of ontogeny. There is a possibility that unidentified compensatory mechanisms for metabolic homeostasis might occur and influence MA levels. Second, our measurements of the MAs and metabolites were performed at steady state. We could not evaluate the effect of MAO-B deficit on MA dynamics after stimulation, task, or stress, each of which can accelerate MA release. Third, this investigation was a target analysis of the selected MAs and metabolites. We cannot exclude the possibility that MAO-B can metabolize unidentified substrates. In particular, it should be noted that there are many trace amines in the brain other than PEA, including tyramine, tryptamine, and octopamine, which are reportedly common substrates of both MAO types according to biochemical studies 3 . Additionally, we have to consider also the fact that MAO-B plays a role in the putrescine-degradation pathway, resulting in the production of gamma-aminobutyric acids in glial cells 8,10 . Another limitation of the study is that the metabolism of DA, NA, and PEA was evaluated solely by measuring substrate concentrations. The PVT is so small that it was necessary to measure many substrates simultaneously in one measurement. However, metabolites produced from DA (i.e., 3,4-dihydroxyphenylacetic acid [DOPAC], and homovanillic acid), NA (i.e., 3-methoxy-4-hydroxyphenylglycol, and vanillylmandelic acid), and PEA (i.e., phenylacetic acid) could not be quantified because of the low ionization efficiency and the use of microwave irradiation for brain-sample preparation, which avoided postmortem degradation and maintained www.nature.com/scientificreports/ high MA concentrations and low metabolite concentrations 43 . A study using bulk tissues showed that the ratio of DOPAC to DA prepared by microwave irradiation was ~ sevenfold lower than that prepared by conventional method 43 . Because the PVT is innervated predominantly by serotonergic neurons rather than dopaminergic neurons 12,21 and DA is metabolized primarily by MAO-A in rodents 1,45-48 , we focused on PEA and 5-HT metabolism in this study. In summary, these findings revealed a potential physiological role of MAO-B in the PVT. The PVT is anatomically in contact with the third ventricle (3 V) and positioned close to the choroid plexus at the 3 V, making it a potential entry point for PEA from the peripheral blood. PEA is known as "the body's own amphetamine" since it has pharmacological effects similar to those of amphetamine, especially its effect on altering MA transporter function 29,34,[49][50][51] . This action potentially disturbs the fine-tuning of MA neurotransmission in the PVT, considering that many MA receptors are localized in this region 11,17,22,23 . If MAO-B plays a role in protecting neurons receiving MA neurotransmission from perturbations by PEA as a "false transmitter" 4,7 , it is reasonable that MAO-B is highly expressed in the PVT. However, PEA primarily demonstrates a high affinity for trace amine-associated receptor 1 (TAAR1), which is a G-protein coupled receptor identified in 2001 [52][53][54] . Taar1 mRNA is expressed in monoaminergic nuclei such as the VTA and DR 52,55 , and TAAR1 activation can modulate monoaminergic neural activity and animal behavior [54][55][56][57] . The mechanism of action of TAAR1 is intriguing from a clinical standpoint, because a recent study reported that a TAAR1 agonist showed significant therapeutic effects in patients with schizophrenia 58 . In this case, PEA could be a valuable neuromodulator that reflects external and internal environments, such as diet and gut microbiota, and MAO-B could be a regulator of this PEA signaling. Here, we described the potential role of MAO-B in the PVT, with a focus on PEA metabolism; however, there are many other MAs in the brain that could be metabolized by MAO-B. In future studies, further investigation will be needed for a more comprehensive metabolome analysis of PVT-specific and cell type-specific MAO-B inactivation. This will enable elucidation of the physiological importance of MAO-B in the PVT, as a promising target region for understanding the mechanism of mood regulation and emotional behaviors underlying mental disorders.

Materials and methods
All animal care and experimental procedures were approved by the RIKEN Wako Animal Experiment Committee (approval number: W2019-2-040 [4]). All animal experiments were performed in accordance with ARRIVE guidelines and the guidelines for the proper conduct of animal experiments published by the Science Council of Japan.

Animals.
The experiments were performed using adult male mice, with all animals housed with ad libitum access to food and water. The animal care unit was maintained on a 12-:12-h light-dark cycle (08:00-20:00) with controlled temperature (23 ± 2 °C) and humidity (50 ± 15%). Maob KO mice and Maob flox mice were developed using the CRISPR Generation of Maob KO mice and Maob flox mice using the CRISPR/Cas9 system. We cut the two target positions using the CRISPR/Cas9 system and generated a large deletion at the Maob locus by nonhomologous end-joining or inserted the lox2272 sequences into the two cut positions by homology-directed repair. For this purpose, we designed two crRNAs targeting the upstream and downstream regions of the Maob gene and two ssODNs containing the lox2272 sequences with 80-nt homology arms ( Supplementary Fig. S1). We prepared a CRISPR/Cas9 cocktail containing 100 ng/μL Cas9 protein (Alt-R S.p. Cas9 Nuclease V3), 10 μM tracrRNA (Alt-R CRISPR-Cas9 tracrRNA), 5 μM each crRNA (Alt-R CRISPR-Cas9 crRNA), 10 ng/μL of each ssODN (PAGE Ultramer DNA Oligo) in nuclease-free duplex buffer with 100 mM potassium acetate and 30 mM 4-(2-hydroxyethyl)-1-piperazine ethanesulfonic acid (HEPES). These constructs for the CRISPR/Cas9 system were purchased from Integrated DNA Technologies (Coralville, IA, USA). The cocktail was injected into the pronuclei of 500 fertilized mouse eggs. A total of 286 embryos were transplanted to surrogate mothers, and 53 pups were obtained (F0 generation). We performed genotyping by PCR using genomic DNA extracted from tail biopsies using the following primer sequences: to detect the Maob KO allele, AGA GCA AGG CCA CTC TCA AAG and AGA GGA AGC TGA GAC TCT GACTG; to detect lox2272 insertion in the upstream region, AGA GCA AGG CCA CTC TCA AAG and CTA ACA GTG GCC TCA AAG CAG; and to detect lox2272 insertion in the downstream region, CAT CCA TGT GCT AAA CAA TGC and AGA GGA AGC TGA GAC TCT GACTG. F0 mice carrying the Maob KO allele (four individuals) or Maob flox allele (four individuals) were crossbred with C57BL/6J WT mice. Sanger sequencing was performed to confirm the large deletion at the Maob locus or correct insertion of the lox2272 sequences in the F1 offspring.
Assessment of off-target mutations by whole-exome sequencing. We performed whole-exome sequencing using DNA samples from the F3 offspring of Maob KO and Maob flox lines for comparison with a C57BL/6J WT mouse. Genomic DNA was extracted from the kidneys using the GenElute mammalian genomic DNA miniprep kit (Sigma-Aldrich, St. Louis, MO, USA). DNA capture, library preparation, DNA sequencing, and variant calling were performed by Riken Genesis (Tokyo, Japan). Briefly, a DNA library was prepared following the SureSelect Target Enrichment System workflow using a SureSelect Mouse All Exon kit (Agilent Tech- www.nature.com/scientificreports/ nologies, Santa Clara, CA, USA). Sequencing was performed on a NovaSeq 6000 platform (Illumina, San Diego, CA, USA) to produce 150-bp reads. Sequencing reads were aligned to the mouse reference (mm10) the Burrows-Wheeler Aligner (v.0.7.10; http:// bio-bwa. sourc eforge. net/). Annotation for all variants was made using dbSNP151, CCDS (NCBI, release 21), and RefSeq (UCSC Genome Browser, dumped 20181125). We applied a further filtering process (Supplementary Table S1) to identify the private variants carried by each F3 mouse. The Maob KO mouse (female, Maob(−/+)) harbored a frameshift mutation (Chr 16: 23114138, ins A) in exon 10 of Rfc4 gene. Because the probability of a loss-of-function intolerant (pLi) score of the Rfc4 gene was 0 according to gnomAD (https:// gnomad. broad insti tute. org/), we judged that this frameshift mutation in a heterozygous manner could not affect the results of the following experiments. No private variants were detected in the Maob flox mouse (male, Maob(flox/Y)). We performed the following experiments using F4 and F5 mice.

Validation of loss of Maob expression by reverse transcription-quantitative PCR.
We assayed the mRNA expression of Maoa, Maob, Ndp, and Calb2 using male Maob(−/Y) mice and their littermate Maob(+/Y) (12-13 weeks old, n = 4 for each group). Actb and Gapdh were also measured as internal normalizers. We collected the PVT, DR, cerebellum, liver, and intestine from decapitated individuals. Total RNA was isolated from the tissues using TRIzol reagent (Thermo Fisher Scientific, Waltham, MA, USA) and extracted using the RNeasy mini kit (Qiagen, Venlo, Netherlands) according to the manufacturer instructions. The RNA integrity number was measured using an Agilent 2100 bioanalyzer (Agilent Technologies) and confirmed to be in the range of 7.3 to 9.6. RNA was then reverse transcribed using SuperScript III First-Strand Synthesis SuperMix (Thermo Fisher Scientific). Target mRNA levels were analyzed using specific primers (Supplementary Table S2) and TB Green Premix Ex Taq II (Takara Bio, Shiga, Japan). Quantitative PCR was performed using a QuantStudio 12 K Flex (Thermo Fisher Scientific). All PCRs were performed in quadruplicate. When the threshold cycle (Ct) value was > 30 in more than two PCR reactions, mRNA expression was evaluated as being 'not detected' . Brain sample preparation for LC-MS/MS. Animals were anesthetized with isoflurane and subjected to focused microwave irradiation (5 kW, 1.05 s) using an MMW-05 apparatus (Muromachi Kikai, Tokyo, Japan). This was performed between 13:00 and 15:00 h. After microwave irradiation, brains were collected and sliced using a brain matrix (2-mm thickness for the PVT and paraventricular hypothalamic nucleus; 1-mm thickness for bilateral sampling of the other regions). The seven regions were set as controls; the NAc, Ins, and DR were anatomically connected to the PVT and the other regions (the caudate-putamen, paraventricular hypothalamic nucleus, occipital cortex, and VTA) were not connected to the PVT. Each target region was collected by punch biopsy (φ1.0 mm) on ice ( Supplementary Fig. S2). Punch biopsies were purchased from Kai Industries (Gifu, Japan), and the punched specimens were stored at − 80 °C and used for subsequent extraction treatment within one week. Extraction of MAs and metabolites was performed using a modified method based on a previous report 59 . We added 90 μL of 1.89% FA/1 mM AA in ice-cold water to each sample, which was then homogenized using an ultrasonic probe (VCX 130; Sonics and Materials, Newtown, CT, USA) at 20% power for 20 s. We then added 10 μL of 1 μM DHBA/1 mM AA in ice-cold water as an internal control, and 10 μL of 1 mM AA in ice-cold water was added as a replacement for addition of the standard substance for calibration curve preparation. After centrifugation at 14,000×g for 15 min at 4 °C, 85 μL of the supernatant was collected, and 340 μL of ACN with 1% FA/1 mM AA was added for deproteinization. After centrifugation at 14,000×g for 5 min at 4 °C, 400 μL of the supernatant was collected, evaporated using a concentrator (CVE-3100 and UT-2000, Tokyo Rikakikai, Tokyo, Japan), and stored at − 80 °C until the LC-MS/MS assay.

Analysis of MAs and metabolites by
For quantitative analysis, we generated calibration curves for the target MAs and metabolites in each batch experiment. In advance, we applied microwave irradiation to a C57BL/6J WT mouse, harvested the whole brain, prepared the homogenate within 90 μL of 1.89% FA/1 mM AA in ice-cold water per mg tissue weight, and stored the samples at − 80 °C until use. Five aliquots (90 μL) of the homogenate were taken, and we added with 10 μL of 1 μM DHBA/1 mM AA in ice-cold water as an internal control and 10 μL of standard substance solution with 1 mM AA to each aliquot. The standard substance solutions included mixtures of PEA (0-50 μM [0, 0.78125, 3.125, 12.5, 50 μM]), NA (0-5000 μM), DA (0-5000 μM), 3-MT (0-50 μM), 5-HT (0-5000 μM), and 5-HIAA (0-5000 μM). We then proceeded with the same extraction treatment used for the brain samples.
Preparation of plasma samples for LC-MS/MS analysis. We used animals distinct from the individuals used for the brain LC-MS/MS experiments. The animals were anesthetized with isoflurane, and blood was collected from the heart using needle aspiration. The sampling syringes were rinsed with 0.5 M ethylenediaminetetraacetic acid disodium salt in advance. The blood was centrifuged at 1000×g for 20 min at 4 °C, and plasma was collected. We www.nature.com/scientificreports/ (0-5000 nM). We then added 60 μL of ACN with 1% FA/1 mM AA, followed by centrifugation at 19,000×g for 10 min at 4 °C. Finally, 80 μL of the supernatant was collected, evaporated to dryness, and stored at − 80 °C. Quantification was performed using the standard addition method. When the calculated value of the substance levels based on the calibration curve was < 0, the concentration was set to 0 nM (see Supplementary Information).

Detection of MAs and metabolites by LC-MS/MS analysis.
All dried samples were reconstituted in 30 μL of 1% ACN with 0.1% FA/1 mM AA. After centrifugation, 5 μL of the supernatant was injected into the LC-MS/MS system consisting of a Vanquish UHPLC and TSQ Vantage EMR (Thermo Fisher Scientific). Separation of the analytes was carried out on a YMC Triart C18 column (2.0 × 100 mm). Mobile phases A (0.1% FA) and B (ACN) were used for linear gradient elution: 0-1 min, 0% B; 1-4 min, 0-60% B; 4-4.1 min, 60-95% B; 4.1-6 min, 95% B; 6-6.1 min, 95-0% B; and 6.1-8 min, 0% B. The flow rate was set at 0.3 mL/min. The electrospray conditions were as follows: spray voltage, 3000 V; vaporizer temperature, 450 °C; sheath gas pressure, 50; auxiliary gas pressure, 15; and collision gas pressure, 1.0 mTorr. Argon was used as the collision gas. The mass spectrometer was operated in positive/negative-ion multiple-reaction monitoring (MRM) mode for the detection of the molecular transition of each analyte (Supplementary Table S3). TraceFindersoftware (v.4.0; Thermo Fisher Scientific) was used for data processing. Ion peak detection and peak area integration were performed based on the analysis parameter settings shown in Supplementary Table S4. When the retention time was shifted and the ion peak was detected inadequately by the parameter settings, the ion peak detection and peak area integration were corrected manually.
Statistical analysis. Data were analyzed using Kyplot (KyensLab, Tokyo, Japan) or R (https:// www.r-proje ct. org/). For analysis of reverse transcription-quantitative PCR in the comparison of global Maob KO mice with WT mice, Student's t-test with Bonferroni correction was used. For analysis of LC-MS/MS experiments in the comparison of global Maob KO mice with WT mice or between brain-specific Maob KO mice and control Nestin-Cre mice, two-way ANOVA with the main effect of genotype and substance was applied in the PVT and each other respective brain region. When analyzing the PEA levels, a two-way ANOVA with the main effect of genotype and region was applied. Student's t-test was also used in the analyses of LC-MS/MS experiments.

Data availability
All data generated or analyzed during this study are included in this published article (and its Supplementary Information). www.nature.com/scientificreports/