Monoamine oxidases (MAO; MAO-A and MAO-B in mammals) are enzymes catalyzing the degradation of biogenic amines, including monoamine neurotransmitters. In humans, coding mutations in MAOs are extremely rare and deleterious. Here, we assessed the structural and biochemical consequences of a point mutation (P106L) in the single mao gene of the blind cavefish, Astyanax mexicanus. This mutation decreased mao enzymatic activity by ∼3-fold and affected the enzyme kinetics parameters, in line with potential structure–function alterations. HPLC measurements in brains of four A. mexicanus genetic lines (mutant and non-mutant cavefish, and mutant and non-mutant surface fish) showed major disturbances in serotonin, dopamine, noradrenaline and metabolite levels in mutants and demonstrated that the P106L mao mutation is responsible for monoaminergic disequilibrium in the P106L mao mutant cavefish brain. The outcomes of the mutation were different in the posterior brain (containing the raphe nucleus) and the anterior brain (containing fish-specific hypothalamic serotonergic clusters), revealing contrasting properties in neurotransmitter homeostasis in these different neuronal groups. We also discovered that the effects of the mutation were partially compensated by a decrease in activity of TPH, the serotonin biosynthesis rate-limiting enzyme. Finally, the neurochemical outcomes of the mao P106L mutation differed in many respects from a treatment with deprenyl, an irreversible MAO inhibitor, showing that genetic and pharmacological interference with MAO function are not the same. Our results shed light on our understanding of cavefish evolution, on the specificities of fish monoaminergic systems, and on MAO-dependent homeostasis of brain neurochemistry in general.

Monoamine oxidase (MAO) is a flavoprotein that catalyzes the oxidative deamination of biogenic amines – monoamine neurotransmitters and dietary amines. It is located at the outer membrane of mitochondria in neuronal, glial and other cells, in platelets, liver and kidney for example (Berry et al., 1994; Levitt et al., 1982). In mammals, MAO exists under two isoforms that share 70% identity in primary sequence and differ in terms of substrate and inhibitor specificities (Shih, 1988). MAO-A prefers serotonin (5-hydroxytryptamine, 5-HT) and noradrenaline as substrates and is sensitive to the irreversible inhibitor clorgyline, whereas MAO-B prefers phenylethylamine and benzylamine as substrates and is more sensitive to the selective inhibitor deprenyl. Dopamine and tyramine are common substrates for both isoenzymes (Shih et al., 1999). The two MAO genes are closely linked on chromosome X and share the same exon–intron organization, suggesting that they have evolved from an ancestral gene by tandem duplication (Grimsby et al., 1991).

The enzyme's biomedical relevance is well supported by the proven efficacy of pharmacological inhibitors in treating depression (Youdim et al., 2006), for example, and by the major neurological defects caused by its malfunction or de-regulation. In humans, only two families have been reported with mutations in MAO-A, and they show cognitive and mental impairment, with severe behavioral disturbances including violent behavior and impulsive aggressiveness (Brunner et al., 1993; Piton et al., 2014). In mice, MAO-A and MAO-B knock-out lines with elevated brain monoamine levels also display complex behavioral alterations similar to the human ‘Brunner syndrome’, and modifications of the stress response and anxiety-like behavior, respectively (Cases et al., 1995; Grimsby et al., 1997). The double MAO-A/B knock-out mice show hallmarks of autism spectrum disorder and greater neuropathological alterations than single knock-out mice (Bortolato and Shih, 2011). Moreover, altered MAO gene promoter methylation and its interaction with environmental factors emerges as an important mechanism in the etiology of several mental disorders (Ziegler and Domschke, 2018). Finally, recent work suggests that besides its neurophysiological aspects, MAO plays important roles in tumorigenesis, diabetes, obesity and cardiovascular disease (Deshwal et al., 2017; Shih, 2018).

The first fish mao was cloned from trout and displays approximately 70% identity with MAO-A and MAO-B (Chen et al., 1994). In fact, teleost fish have only one mao in their genomes. Fish mao is homologous to the mammalian MAOs, is expressed in various tissues like in mammals, and degrades tyramine, serotonin, phenylethylamine (PEA) and dopamine (Aldeco et al., 2011; Anichtchik et al., 2006; Chen et al., 1994; Elipot et al., 2014; Setini et al., 2005). In contrast, some authors have reported that dopamine is a poor substrate for zebrafish MAO on the basis that its inhibition by deprenyl elevates serotonin but not dopamine levels (Sallinen et al., 2009). The teleost MAO seems to exhibit functional properties that are more similar to those of mammalian MAO-A (Aldeco et al., 2011; Sallinen et al., 2009), and the zebrafish mao knock-out line shows various neurological alterations (Baronio et al., 2022).

In the blind and depigmented, cave-dwelling form of the fish species Astyanax mexicanus (Elliott, 2018; Mitchell et al., 1977), a natural point mutation in the mao gene (P106L) has been reported (Elipot et al., 2014). The substituted proline 106 is homologous to the proline 114 and 105 in human MAO-A and -B, respectively. Therefore, through comparative studies with its river-dwelling conspecific, the cavefish natural mao mutant could help us better understand the properties of fish MAO and the impact of MAO deficiencies in the vertebrate brain. Moreover, as the zebrafish mao knock-out is non-viable (Baronio et al., 2022), the A. mexicanus mao P106L is an excellent tool to study the effect of MAO deficiency in fish.

A previous study measured MAO enzymatic activities and brain monoamine levels on A. mexicanus surface fish (SF) and cavefish (CF) originating from the Pachón cave (Elipot et al., 2014). However, we have recently discovered that the P106L mutation is not fixed in the Pachón CF population (Pierre et al., 2020). Therefore, the Pachón individuals used in the previous study were probably a mix of mutants, non-mutants and heterozygotes, prompting the need to re-assess biochemical parameters on samples with homogeneous genetic backgrounds. Moreover, it was suggested that the mutation in the gene Oca2 (Ocular and cutaneous albinism 2) that causes cavefish albinism (Protas et al., 2006) could provide a surplus of L-tyrosine, the common precursor of the catecholamine and melanin synthesis pathways, and could explain the higher levels of dopamine and noradrenaline in cavefish brains (Bilandžija et al., 2013). The goal of the present study, using four A. mexicanus genetic lines (mao mutant and non-mutant cavefish; mao mutant and non-mutant surface fish) was therefore to decipher the exact contribution of the mao P106L mutation in monoaminergic brain homeostasis.

The consequences of the P106L mao mutation on physiological, behavioral and neurodevelopmental traits in cavefish were reported in Pierre et al. (2020). Surprisingly, mao P106L has modest deleterious phenotypic effects and mainly alters anxiety-like behaviors in cavefish, reducing their basal cortisol levels but dramatically enhancing the amplitude of the stress response after a change in environment (Pierre et al., 2020). In line with mao P106L modulating the stress response, we also noticed that the mutation changes basal adrenaline body levels, and basal noradrenaline, serotonin and dopamine brain levels as well as their variations after a stressful stimulus (Pierre et al., 2020). To further characterize the biochemical effects of the P106L mao mutation, in the present study we focused on enzyme structure, enzymatic activity and brain neurochemistry. We report that P106L corresponds to a partial loss-of-function mutation causing the disequilibrium in serotonin, dopamine and noradrenaline neurotransmission in mutant cavefish brains, but that compensatory mechanisms may buffer the deleterious effects of the mutation.

Fish husbandry

Laboratory stocks of Astyanax mexicanus (De Filippi 1853) surface fish and cavefish (Pachón population) were obtained in 2004–2006 from the Jeffery laboratory at the University of Maryland, College Park, MD, USA, and were then bred in our local institute facility. Fish were maintained at 23–26°C on 12 h:12 h light:dark cycle and they were fed twice a day with dry food. The fry were raised in Petri dishes and fed with micro-worms after opening of the mouth [∼6 days post-fertilization (dpf)]. Animals were treated according to the French and European regulations for handling of animals in research. S.R.'s authorization for use of A. mexicanus in research is 91-116 and the Paris Centre-Sud Ethic Committee protocol authorization number related to this work is 2017-05#8604. The animal facility of the Paris-Saclay Institute of Neuroscience received authorization B 91 272 108 from the Veterinary Services of Essonne, France, in 2021.

Fish lines

To obtain Pachón cavefish (CF) without the P106L mutation, we crossed heterozygote fish identified among our laboratory Pachón breeding colony, taking advantage of the fact that the P106L mutation is not fixed in the Pachón population (see Pierre et al., 2020). To obtain a surface fish (SF) line carrying the mutation, a cross between a SF (wild-type mao) and a CF carrying the P106L (homozygote mutant) was followed by four backcrosses with SF (see Pierre et al., 2020). Then, to obtain SF homozygote mutants, we intercrossed the last generation together. Note that the generation time between the spawn of the n generation and the spawn of the n+1 generation was approximately 8 months.

mao P106L allele genotyping

To genotype adults, we took fin-clips and performed a lysis with proteinase K in lysis buffer (100 mmol l−1 Tris; 2 mmol l−1 EDTA; 0.2% Triton; 0.01 µg µl−1 PK), followed by a PCR (primer F-GGGAAATCATATCCATTCAAGGGG; primer R-CTCCATGCCCATCTTGTCCATAG), and a purification of DNA (NucleoSpin® Gel and PCR Clean-up). We used the genotyping service of Eurofins Genomics. Homozygotes and heterozygotes at position 106 were easily detected and identified on sequence chromatograms. We did not genotype larvae as they were obtained from genotyped groups of adult breeding fish.

Sampling for biochemistry

Larvae at 6 dpf were anesthetized in water at 1°C, and the head was cut from the body. A sample (n=1) was formed with 15 heads or bodies in 400 µl of HCl (10−3 mol l−1). It means that, for example in Fig. 1A, four tubes for SF and eight tubes for CF were measured, each containing 15 heads. Thus, a total of 60 SF heads and 120 CF heads were used for this experiment. The same protocol was used for 1-month-old larvae, but samples were made up of 10 heads or bodies.

Fig. 1.

Enzymatic properties of P106L mutant and non-mutant MAO in Astyanax mexicanus. (A) Measures of MAO enzymatic activity in brains of 6 dpf cavefish that are homozygous P106L mao mutants (m) or non-mutants (+). (B,C) Measures of Km and Vmax on serotonin (B; 5-HT) or dopamine (C; DA) as substrates, on brain extracts from 6 dpf cavefish with the indicated genotypes. In this and subsequent figures, the number of samples included is indicated under box plots. The outlines of the boxes indicate the morphotype (SF, blue and CF, red) and the colors inside bars indicate the genotype (+/wild-type in blue; mao P106L mutant in red). P-values for Mann–Whitney pairwise comparisons are shown (**P<0.01).

Fig. 1.

Enzymatic properties of P106L mutant and non-mutant MAO in Astyanax mexicanus. (A) Measures of MAO enzymatic activity in brains of 6 dpf cavefish that are homozygous P106L mao mutants (m) or non-mutants (+). (B,C) Measures of Km and Vmax on serotonin (B; 5-HT) or dopamine (C; DA) as substrates, on brain extracts from 6 dpf cavefish with the indicated genotypes. In this and subsequent figures, the number of samples included is indicated under box plots. The outlines of the boxes indicate the morphotype (SF, blue and CF, red) and the colors inside bars indicate the genotype (+/wild-type in blue; mao P106L mutant in red). P-values for Mann–Whitney pairwise comparisons are shown (**P<0.01).

Adults (5 months old) were quickly anesthetized in water at 1°C. Then, the head was cut with a scalpel, and the brain was dissected out. The fish brain serotonergic system is composed of two groups of clusters of neurons, in the hypothalamus and in the raphe, respectively. Hypothalamic and raphe clusters are distinct in terms of gene expression and transcriptional factors involved during their development (Lillesaar, 2011). To investigate whether the P106L mao mutation could have distinct consequences in the two groups of serotonergic clusters, in 5-month-old fish we analyzed separately the anterior (including the hypothalamic clusters) and the posterior (including the raphe nuclei) parts of the brain [a cut at the midbrain/hindbrain junction was possible on older, 5 months post-fertilization (mpf) brains]. The two parts of the brain were individually placed in 400 µl of HCl (10−3 mol l−1).

For circadian rhythm analyses, fish aged 5 mpf were entrained during one week on a 12 h:12 h light:dark regime, and brain monoamine levels were assayed every 2 h over a 26 h period.

HPLC and enzymatic activities

For HPLC, before analysis, tissues were crushed and centrifuged at 20,000 g for 1 h. The supernatant was analyzed by fluorimetry for serotonin and by coulometry for catecholamines.

For enzymatic activities, MAO (EC.1.4.3.4.) enzymatic activities were determined on brain homogenates as described in Elipot et al. (2014) by radioenzymology using either [14C]-5-hydroxytryptamine creatinine sulphate (serotonin, 5-HT, 1.96 GBq mmol−1, Amersham GE Healthcare, Saclay, France), 3,4-[ring 2,5,6-3H]-dihydroxyphenylethylamine (dopamine, DA, 9.25 MBq mmol−1, NEN PerkinElmer, PerkinElmer France SAS, Villebon sur Yvette, France) or [1,14C]-tyramine hydrochloride (1.85 GBq mmol−1, Amersham GE Healthcare, Saclay, France, final concentration 20 μmol l−1) as substrates. In brief, 20 μl aliquots of the homogenates were pre-incubated (possibly with an inhibitor) in a total volume of 220 μl at 37°C for 30 min. At the end of the pre-incubation period, the enzyme reaction was started by addition of 80 μl of one MAO substrate and the incubation continued for 15 min at 37°C. Then, the reaction was stopped by addition of 200 μl of HCl 2 mol l−1, and the deaminated metabolites were extracted by vigorous shaking for 10 min with 5 ml of diethylether (5-HT extraction), n-heptane (DA extraction) or anisole (tyramine extraction). After centrifugation (1000 g; 30 s) the water phase was frozen in dry ice and the organic layer poured into plastic vials containing 5 ml of scintillation cocktail. Finally, the radioactivity was determined in a scintillation spectrometer. Reaction mixtures as described above, but lacking the brain homogenate, served as blanks. The optimized incubation time (15 min) and MAO concentration (final concentration in the reaction mixture: 0.01 mg ml−1) were selected in the linear range from the time-dependent and MAO concentration-dependent studies. Saturation experiments were carried out with isotopic dilutions of 5-HT and DA at concentrations of 2, 5, 10, 25, 50, 100, 250 and 500 μmol l−1. GraphPad Prism 5 was applied to fit a Michaelis–Menten model to the data to obtain the Km (the concentration of substrate for which the speed of the enzymatic reaction is half the maximum speed Vmax, which gives an indication of the affinity of the enzyme for its substrate) and Vmax (the speed of the reaction when the substrate concentration is infinite) values in the saturation experiments.

Deprenyl treatments

Adults were placed individually in 600 ml water containing 10 µmol l−1 of the irreversible mao inhibitor deprenyl for 5 h for an acute treatment. The 10 µmol l−1 concentration was chosen based on previous experience showing that MAO inhibition is effective at this concentration after HPLC measurements of monoamines and their metabolites (e.g. Elipot et al., 2014).

Three-dimensional modeling

Three-dimensional modeling was performed with iTasser (https://zhanglab.ccmb.med.umich.edu/I-TASSER/), 3D alignment with TM-align, and secondary structure predictions and normalized B-score calculation with ResQ (Yang et al., 2015, 2016). ResQ (estimation of residue-specific quality and B-factor) is a method for estimating B-factor and residue-level quality in protein structure prediction, based on local variations of modeling simulations and the uncertainty of homologous alignments. Hence, ResQ estimates the global accuracy of the model.

Structural information reported in the Results include C-alpha trace, TM-score and B-factor. The C-alpha trace is the backbone of the protein, composed of the chain of all the alpha carbon atoms of the amino acids. The TM-score is calculated from the distances between corresponding amino acids in two proteins, after superposition in space. It is scored between 0 (=totally different structures) and 1 (=identical proteins). A score of >0.5 indicates proteins with similar folding which belong to the same structural class. The B-score (or B-factor, calculated by ResQ) measures the uncertainty/mobility of an atom in dynamic protein 3D structures, namely, the displacement of the atomic positions from its mean position. If the B-score is less than zero, the residue is defined as stable, and flexible otherwise.

The models used for A. mexicanus MAO (wild-type and mutant) were obtained with iTASSER.

The models used for MAO-A et MAO-B were obtained from PDB Protein databank (https://www.rcsb.org/).

Statistical analyses

Dataset 1 shows raw data and exhaustive statistical analyses using Mann–Whitney pairwise comparisons, and one-way, two-way and three-way ANOVAs. No statistical method was used to predetermine sample size. No data were excluded from analysis, and sample allocation was random after genotyping. Sex was not considered in the analyses as it was impossible to determine sex in A. mexicanus before the age of 6–7 months without dissection or euthanization. Analyses were not blind (note that for anatomical analyses, brains from SF or CF are easily recognizable by eye size and presence of pigmentation).

For Fig. 1, statistical analyses were performed using non-parametric Mann–Whitney tests with BiostaTGV (https://biostatgv.sentiweb.fr/; statistical tests performed in R; normal distribution was not tested).

For Fig. 3A,B, one-way ANOVA were performed to account for comparisons across the three groups compared, with Tukey post hoc tests for pairwise comparisons. For Figs 3C,D and 4, comparisons were carried out by two-way ANOVA to account for comparisons between genotypes (mao mutant or non-mutant) and morphotypes (SF versus CF) as well as their interactions, with Tukey post hoc tests for pairwise comparisons. For Fig. 5, three-way ANOVAs were performed to account for comparisons between genotypes (mao mutant or non-mutant), treatment type (deprenyl or control) and brain region considered (anterior versus posterior) as well as their interactions, with Tukey post hoc tests. ANOVA analyses were performed in R. Graphs show box plots representing the median, minimum and maximum, and first and third quartile of the datasets. The number of samples included is systematically indicated below the first series of data on each graph. P-values from Tukey post hoc comparisons are shown on graphs. Statistical significance was set at P<0.05. The raw data and full statistics summary are reported in Dataset 1.

The P106L mutation severely decreases the enzymatic activity of A. mexicanus MAO

First, we measured MAO enzymatic activities on brain extracts from 6 dpf Pachón cavefish larvae with homogeneous genetic backgrounds, obtained from crosses (see Materials and Methods): mao P106L homozygous mutants and non-mutants. The P106L mutation caused a ∼3-fold reduction of MAO enzymatic activity, using tyramine as a substrate (P=0.004, Mann–Whitney test; Fig. 1A). This reduction was stronger than the 2-fold reduction we previously reported on samples that were probably non-homogeneous genetically (Elipot et al., 2014). Of note, MAO activity in brains from heterozygotes was also slightly reduced as compared with non-mutant controls (15.1±0.39, n=3; P=0.057, Mann–Whitney), probably in line with the fact that MAO functions as a dimer.

Further, to obtain insights on MAO affinity and kinetics towards its substrates, we determined its Km and Vmax. Km and Vmax were first determined for 5-HT and compared in brain extracts from P106L mao mutant and non-mutant 6 dpf cavefish (Fig. 1B). In mutant extracts, the Km was 2- to 3-fold higher (P=0.002, Mann–Whitney), indicating a decreased affinity for the substrate, and the Vmax was 2- to 3-fold lower (P=0.002, Mann–Whitney), in concordance with the decreased MAO activity described above. We also determined the Km and Vmax of the enzyme for DA (Fig. 1C). The affinity of MAO for DA was much lower than for 5-HT (40 times). The effect of the P106L mutation was also milder on the affinity for DA (–13% in mutant brains, P=0.002, Mann–Whitney), and the mutation did not significantly affect the Vmax (P=0.126).

The P106L mutation probably limits entrance of the substrate to the enzyme active site

To obtain insights into how the mao P106L mutation affects the structure and function of the enzyme, we accessed 3D-modeling of the mutant and non-mutant proteins. First, a 3D model of A. mexicanus MAO was obtained with iTasser (Fig. 2A). Of note, the P106L mutation should theoretically remove a bend in the polypeptide chain. However, the protein modeled with either a proline or a leucine in position 106 had the same Cα trace (i.e. the backbone of the protein, composed of the chain of all the alpha carbon atoms of the amino acids) and perfectly overlapped (Fig. 2B), with a TM score between 0.995 and 1 depending on the models used. The TM score indicates 3D structural similarity and is scored between 0 and 1, with a value of 1 indicating a perfect match between two structures. Thus, the P106L mutation did not change the structure of the A. mexicanus MAO protein. In fact, A. mexicanus MAO showed a structure very similar to human MAOs (Son et al., 2008), in particular with the entrance channel for the substrate towards the deep active site being formed by three loops (yellow in Fig. 2A). The TM scores were 0.97415 and 0.94971 when compared with MAO-A and MAO-B, respectively. The proline106 in A. mexicanus MAO is homologous to the proline114 in human MAO-A, so that the P106L mutation found in cavefish is located at the edge of one of the three loops (Fig. 2A). In terms of secondary structure, the predictions obtained with ResQ showed two short additional strands (or beta-pleated sheets) formed next to position 106 in the mutant protein (Fig. 2C, green arrows). As alpha helixes and beta-pleated sheets are more rigid than unfolded coil, the P106L mutation may rigidify one of the loops forming the entrance of the active site and limit the access of substrates. It should be noted, however, that the normalized B-score whose calculation is based on local variations of modeling simulations, and so is an indicator of the flexibility, was not very different between the two proteins, including around position 106 (Fig. 2C, blue lines).

Fig. 2.

Structure of P106L mutant and non-mutant MAO proteins. (A) Three-dimensional model of the globular part of wild-type A. mexicanus MAO, with FAD cofactor in red. Blue amino acids belong to the active site and bind serotonin (Binda et al., 2002; Veselovsky et al., 2004); the three loops in yellow form the entrance to the active site (Son et al., 2008). The position 106 is in green (arrow). (B) P106L mutant (pink) and non-mutant MAO (blue) 3D structure superposition (C-alpha traces). (C) Normalized B-factor (blue line; indicates flexibility versus stability of a residue) and secondary structure prediction (red, alpha helix; green, beta strands) are reported along the protein sequences for the two MAO proteins. The normalized B-score is calculated with local variations of modeling simulations. If the normalized B-score is below zero, the residue is defined as stable and flexible otherwise. Green arrows point to two additional beta sheets predicted in the P106L mutant MAO.

Fig. 2.

Structure of P106L mutant and non-mutant MAO proteins. (A) Three-dimensional model of the globular part of wild-type A. mexicanus MAO, with FAD cofactor in red. Blue amino acids belong to the active site and bind serotonin (Binda et al., 2002; Veselovsky et al., 2004); the three loops in yellow form the entrance to the active site (Son et al., 2008). The position 106 is in green (arrow). (B) P106L mutant (pink) and non-mutant MAO (blue) 3D structure superposition (C-alpha traces). (C) Normalized B-factor (blue line; indicates flexibility versus stability of a residue) and secondary structure prediction (red, alpha helix; green, beta strands) are reported along the protein sequences for the two MAO proteins. The normalized B-score is calculated with local variations of modeling simulations. If the normalized B-score is below zero, the residue is defined as stable and flexible otherwise. Green arrows point to two additional beta sheets predicted in the P106L mutant MAO.

Finally, to obtain insights into the severity of the P106L mutation from another, independent perspective, we assessed its pathogenicity score in MutPred2, a machine learning-based predictor of the impact of missense variants in humans (Pejaver et al., 2020). The cavefish mao P106L mutation had a MutPred2 score, i.e. the probability that the amino acid change is pathogenic, of 0.456. This is close to the pathogenicity threshold (0.5). Of note, another non-synonymous mutation just next to position 106, M107I, found as heterozygote in a wild-caught surface fish individual (Pierre et al., 2020), had a MutPred2 score of 0.247. Thus, contrarily to P106L in cavefish, the M107I variant found in a surface fish is not deleterious for the enzyme function.

Altogether, these data provide structural, biochemical and computational evidence that the cavefish mao P106L mutation corresponds to a partial loss-of-function mutation. Below, we further analyze its consequences on cavefish brain neurochemistry.

The P106L mutation affects monoamine homeostasis

Region-specific effects on brain monoamine and metabolite levels

To assess the consequences of decreased MAO enzymatic activity on brain neurochemistry, we measured monoamine and metabolite levels in heads and bodies of 6 dpf larvae and brains of 1 and 5 mpf (=young adults) A. mexicanus. We reasoned that comparing surface fish and cavefish either with wild-type, non-mutant mao or with the P106L mao mutation would enable us to disentangle the effects of the morphotype (surface versus cave) and the effects of the genotype (mao wild-type versus P106L) as well as the interactions between the two factors on brain monoamines (the full ANOVA summary is reported in Dataset 1).

As expected, in 6 dpf larval heads, the levels of the three monoamines 5-HT, NA and DA were higher in P106L mao mutants than in non-mutant samples (Fig. 3A). The P106L mutation increased the levels of 5-HT, NA and DA by 1.26, 1.15 and 1.30 times, respectively (P=0.0003, P=0.006 and P=7.8e-06, respectively, ANOVA with Tukey post hoc comparisons). Conversely, the levels of 5-HT and DA metabolites (5HIAA, DOPAC and HVA) were lower in CF mao mutants (P=0.0004, P=1.1e-05 and P=0.0039, respectively, ANOVA with Tukey post hoc comparisons). The NA metabolite VMA was non-detectable (not shown). Importantly, the levels of 5-HT, NA and DA were identical in SF and non-mutant CF (P=0.68, P=0.86 and P=0.31, respectively, ANOVA with Tukey post hoc comparisons). These features were also true in 1-month-old juveniles (Fig. S1).

Fig. 3.

Monoamine and metabolite levels in mao P106L mutant or non-mutant A. mexicanus, surface fish and Pachón cavefish. On all graphs, the mutant (m) or non-mutant (+) state of the mao P106l mutation is indicated, as well as the morph (SF or CF), the type of sample (larval brain, body or anterior/posterior adult brain) and the molecule measured. The full statistics summaries for one-way ANOVA (A,B) and two-way ANOVA (C,D) testing the effects of genotype, morphotype and their interaction on the dependent variable are reported in Dataset 1. P-values for multiple comparisons using post hoc Tukey comparisons are shown (*P<0.05; **P<0.01; ***P<0.001; ****P<0.0001). (A,B) Measures of monoamines and metabolites in heads (A) and bodies (B) of 6 dpf larvae and calculated ratios 5-HIAA/5-HT and (DOPAC+HVA)/DA in SF and CF, non-mutant (+) or homozygote mao P106L mutants (m). Each sample (n=1) corresponds to 15 mixed heads of the same genotype. Adre, adrenaline. (C,D) Measures of monoamines and metabolites in the brain of 5-month-old fishes in SF and CF, non-mutant (+), and homozygote mao P106L mutants (m). For 5-HT and 5HIAA, the data are presented separately for the anterior and posterior brain (C). For DA, NA and metabolites, the values correspond to measures from whole brains (D).

Fig. 3.

Monoamine and metabolite levels in mao P106L mutant or non-mutant A. mexicanus, surface fish and Pachón cavefish. On all graphs, the mutant (m) or non-mutant (+) state of the mao P106l mutation is indicated, as well as the morph (SF or CF), the type of sample (larval brain, body or anterior/posterior adult brain) and the molecule measured. The full statistics summaries for one-way ANOVA (A,B) and two-way ANOVA (C,D) testing the effects of genotype, morphotype and their interaction on the dependent variable are reported in Dataset 1. P-values for multiple comparisons using post hoc Tukey comparisons are shown (*P<0.05; **P<0.01; ***P<0.001; ****P<0.0001). (A,B) Measures of monoamines and metabolites in heads (A) and bodies (B) of 6 dpf larvae and calculated ratios 5-HIAA/5-HT and (DOPAC+HVA)/DA in SF and CF, non-mutant (+) or homozygote mao P106L mutants (m). Each sample (n=1) corresponds to 15 mixed heads of the same genotype. Adre, adrenaline. (C,D) Measures of monoamines and metabolites in the brain of 5-month-old fishes in SF and CF, non-mutant (+), and homozygote mao P106L mutants (m). For 5-HT and 5HIAA, the data are presented separately for the anterior and posterior brain (C). For DA, NA and metabolites, the values correspond to measures from whole brains (D).

The index of neurotransmission can be estimated by the metabolite/neurotransmitter ratio (inversely proportional to the ‘strength’ of neurotransmission or signaling). In the heads of 6 dpf larvae, the 5HIAA/5-HT and (DOPAC+HVA)/DA ratios were lower in CF, both mutant and non-mutant, and were further decreased by the mutation within CF (Fig. 3A, right). These data suggested a strong 5-HT and DA signaling in P106L mao mutant CF larval heads.

We also measured monoamines and metabolites in the bodies of 6 dpf larvae. There, the P106L mutation also increased by 1.17, 1.23 and 1.17 times the levels of 5-HT, NA and adrenaline, respectively (P=0.0005, P=0.003 and P=0.017, respectively, ANOVA with Tukey post hoc comparisons; Fig. 3B). The metabolites 5HIAA, DOPAC and HVA were decreased (P=0.0015, P=0.0005 and P=0.025, respectively, ANOVA with Tukey post hoc comparisons). VMA and DA were non-detectable in bodies. The only morph-dependent difference between non-mutant CF and SF was a higher level of 5-HT in SF. This difference between non-mutant CF and SF disappeared in 1-month-old fish (Fig. S1).

In adult, 5-month-old fish, the effects of the P106L mao mutation were analyzed separately in the anterior and the posterior parts of the brain, containing the 5-HT hypothalamic clusters and hindbrain nuclei, respectively. Like for 6 dpf and 1 mpf brains, for the two brain parts, 5-HT levels were identical in non-mutant SF and non-mutant CF (P=0.99 and P=0.53, respectively, two-way ANOVA with Tukey post hoc comparisons; Fig. 3C), suggesting that the P106L mao mutation is fully responsible for 5-HT transmission disequilibrium in mutant CF. In cavefish carrying the P106L mutation, serotonin levels were increased by 1.58 times in the anterior brain (P=1.74e-05, two-way ANOVA with Tukey post hoc comparisons) but were unaffected in the posterior brain (P=0.54). However, in surface fish, the P106L mutation led to an increase in 5-HT levels in both the hypothalamus and the raphe (P=0.0006 and P=0.0014, two-way ANOVA with Tukey post hoc comparisons), thus uncovering a morph-specific difference in serotonin metabolism in the raphe. Finally and as expected, P106L decreased 5HIAA levels in both parts of the brain in both morphs (Fig. 3C). In both morphs, the heterozygotes showed intermediate levels between homozygote mutants and non-mutant individuals (not shown).

Regarding DA and NA, we examined 5 mpf whole brain contents. The two morphs showed similar variations according to mao genotypes (Fig. 3D). The P106L mutation increased NA levels significantly only in SF, although the same trend was observed in CF (P=0.0008 in SF and P=0.058 in CF; two-way ANOVA with Tukey post hoc comparisons), but increased DA levels only in CF (P=2.96e-08 in CF and P=0.66 in SF). The DA metabolites DOPAC and HVA were globally decreased. Like for 5-HT, heterozygotes showed intermediate levels between homozygote mutants and wild-type individuals (not shown). Importantly, NA and DA levels were identical in non-mutant SF and non-mutant CF (P=0.13 and P=0.66, respectively; two-way ANOVA with Tukey post hoc comparisons; Fig. 3D), suggesting that the P106L mao mutation could be fully responsible for monoamine transmission disequilibrium in mutant CF.

To summarize, an effect of the P106L mao genotype was systematically observed on all monoamine and metabolite levels (Fig. 3CD; see summary statistics for two-way ANOVA in Dataset 1). The cave versus surface morphotype had a strong influence as well on five out of the six molecules assessed (5-HT, NA, DA, 5HIAA and HVA but not DOPAC), and an interaction between the genotype and the morphotype could also be detected for three of them (5HIAA, DA and HVA).

Finally, vertebrates including teleosts show daily variations of monoamine levels (Fingerman, 1976; Khan and Joy, 1988). Yet A. mexicanus cavefish show a disruption of their endogenous biological clock and its entrainment by light (Beale et al., 2013; Moran et al., 2014) and mao P106L alters brain monoamine levels (above), prompting us to investigate whether circadian variation in monoamine levels would persist in mutants. After entrainment during 1 week on a 12 h:12 h light:dark regime, cyclic variations of monoamines (5-HT) and metabolites (5HIAA, DOPAC and HVA) seemed present in 5 mpf non-mutant A. mexicanus surface fish and cavefish, suggesting that daily variations are unaltered in the cave morph under these conditions of light:dark entrainment (Figs S2 and S3; note that statistics could not be tested because individual data points were lost for CF with non-mutant mao). However, rhythmicity appeared abolished in cavefish carrying the mao P106L mutation, suggesting that in cavefish, daily variations in monoamines can still be entrained by light despite the defective circadian clock, and that this process relies on MAO-dependent regulatory mechanisms.

P106L mutant CF show a reduced TPH activity

The enzymatic activity of MAO was reduced 2.64 times by the P106L mutation in cavefish (Fig. 1A). However, the levels of neurotransmitters were more modestly increased (1.58× for 5-HT in the anterior brain; 2.20× for DA) (Fig. 3). We therefore wondered whether a compensation could occur at the level of neurotransmitters synthesis. We measured and compared the activity of tryptophan hydroxylase (TPH, the 5-HT synthesis rate-limiting enzyme) in the four fish lines at 5 mpf. The two non-mutant morphs showed the same TPH activity, in the two parts of the brain (P=0.84 and P=0.78, respectively, two-way ANOVA with Tukey post hoc comparisons; Fig. 4). Yet again, morph-specific and region-specific differences were observed. TPH activity was strongly decreased in mutant CF (P=6.55e-08, two-way ANOVA with Tukey post hoc comparisons; Fig. 4). The reduction was proportionately more important in the anterior part (–53%) than in the posterior part (–28%) of the brain. In mutant SF, however, the reduction of TPH activity was not observed (P=0.12 and P=0.49 in the two brain parts, respectively, two-way ANOVA with Tukey post hoc comparisons). This result may explain why the posterior brain of CF are unaffected by the P106L mao mutation, conversely to SF. In sum, a significant compensation of the effects of the mutation exists at the 5-HT synthesis level in CF but not SF, and the anterior and posterior 5-HT nuclei may behave differently in this respect.

Fig. 4.

Tryptophan hydroxylase activity in mao P106L mutant or non-mutant cavefish and surface fish. TPH activity was measured in the anterior and posterior brain of 5 month-old individuals, SF and CF, mutant (m) and non-mutant (+). The full statistics summaries for two-way ANOVA testing the effects of genotype, morphotype and their interaction on the dependent variable are reported in Dataset 1. P-values for multiple comparisons using post hoc Tukey comparisons are shown (***P<0.001; ****P<0.0001).

Fig. 4.

Tryptophan hydroxylase activity in mao P106L mutant or non-mutant cavefish and surface fish. TPH activity was measured in the anterior and posterior brain of 5 month-old individuals, SF and CF, mutant (m) and non-mutant (+). The full statistics summaries for two-way ANOVA testing the effects of genotype, morphotype and their interaction on the dependent variable are reported in Dataset 1. P-values for multiple comparisons using post hoc Tukey comparisons are shown (***P<0.001; ****P<0.0001).

An acute treatment with deprenyl does not mimic the P106L genetic mutation

We next asked to what extent the effects of the P106L mutation could be compared with MAO inhibition by drug treatment. We thus analyzed the effects of deprenyl, a selective and irreversible MAO inhibitor, on the anterior and posterior brain neurochemistry of genotyped cavefish individuals, either mao mutants or non-mutants.

Acute treatment on 5-month-old adult cavefish increased 5-HT levels by 2.03× and 2.36× in non-mutant CF, and by 1.35× and 2.47× in mutant CF, in the anterior and the posterior part of the brain, respectively (Fig. 5; three-way ANOVA with Tukey post hoc tests, full statistics summary in Dataset 1). As expected, the treatment also decreased 5HIAA levels, except in the posterior brain of mutant CF. The modifications caused by acute MAO inhibition by deprenyl, at the concentration used (10 µmol l−1), on 5HIAA levels were stronger than those observed with the P106L mutation (compare Figs 3C and 5). Of note, this acute pharmacological treatment also increased 5-HT levels in the posterior part of the brain, in contrast to the mutation (Fig. 5). Hence, on the levels of 5-HT and its 5HIAA metabolite, the three-way ANOVA comparisons detected significant effects of the mao P106L genotype (P=7.8e-04 and P=6.8e-07, respectively) and of the deprenyl treatment (P=5.5e-11 and P=3.2e-12), as well as interactions between genotype, treatment and brain region considered.

Fig. 5.

Effects of acute deprenyl treatment on cavefish brain neurochemistry. Monoamine and metabolite levels in the anterior and posterior parts of the brain of 5-month-old fish, after acute 10 µmol l−1 deprenyl treatment (black squares) or in control fish (white squares). Fish used are CF, both mutant (m) and non-mutant (+). The full statistics summary for three-way ANOVA testing the effects of genotype, brain region, treatment and their interaction on the dependent variable is reported in Dataset 1. P-values for multiple comparisons using post hoc Tukey comparisons are shown (*P<0.05; **P<0.01; ***P<0.001; ****P<0.0001).

Fig. 5.

Effects of acute deprenyl treatment on cavefish brain neurochemistry. Monoamine and metabolite levels in the anterior and posterior parts of the brain of 5-month-old fish, after acute 10 µmol l−1 deprenyl treatment (black squares) or in control fish (white squares). Fish used are CF, both mutant (m) and non-mutant (+). The full statistics summary for three-way ANOVA testing the effects of genotype, brain region, treatment and their interaction on the dependent variable is reported in Dataset 1. P-values for multiple comparisons using post hoc Tukey comparisons are shown (*P<0.05; **P<0.01; ***P<0.001; ****P<0.0001).

In contrast, and surprisingly, deprenyl had no effect on NA and DA levels, but their metabolites were systematically and significantly decreased (with one exception; Fig. 5). Taken together, these results showed that an acute MAO inhibition had markedly different effects on brain neurochemistry than a chronic MAO deficiency owing to a partial loss-of-function genetic mutation.

Here, we report that mao P106L leads to a 3-fold reduction of MAO enzymatic activity, thus corresponding to a partial loss-of-function mutation, which can be explained by in silico structure–function analyses. In the ‘natural mutant’ cavefish brain, the P106L mao mutation causes major modifications in brain neurochemistry and a disequilibrium in 5-HT, DA and NA neurotransmission indices. Further, the differences observed between cavefish and surface fish carrying the P106L mao mutation suggest that the morph-specific genetic background has some influence on monoamine metabolism. We found that 5-HT homeostasis is partially compensated for at synthesis level, and that pharmacological treatments reducing MAO activity have markedly different outcomes than the P106L genetic mutation in cavefish, further supporting that plastic and compensatory mechanisms must occur and may buffer the deleterious effects of mutation. Finally, we discovered that the P106L mao mutation has distinct consequences in the two groups of serotonergic clusters in the hypothalamus and in the raphe, respectively, suggesting that 5-HT hypothalamic and raphe neurons ‘react’ differently to compromised MAO activity.

Together, our data provide insights into understanding: (1) A. mexicanus cavefish phenotype and evolution, (2) the specificities of fish monoaminergic and particularly serotonergic systems, and (3) MAO-dependent regulation of brain neurochemistry in general. These three aspects are discussed below (Fig. 6).

Fig. 6.

Summary of the results. The variations in brain neurochemistry observed between surface fish and cavefish, and non-mutants and mutants are summarized. The four-way comparisons allow us to infer genotype-specific, morph-specific and brain-region-specific differences.

Fig. 6.

Summary of the results. The variations in brain neurochemistry observed between surface fish and cavefish, and non-mutants and mutants are summarized. The four-way comparisons allow us to infer genotype-specific, morph-specific and brain-region-specific differences.

mao P106L: specific considerations related to cavefish biology

The first crystal structures of human MAO-B and rat MAO-A were obtained in 2002 and 2004, respectively (Binda et al., 2002; Ma et al., 2004), and subsequently, increasingly higher resolution structures (De Colibus et al., 2005; Son et al., 2008) helped the development of effective and selective inhibitors. The C-terminal region of the protein corresponds to transmembrane alpha helices that anchor the enzyme to the mitochondrial outer membrane. The active site is a flat hydrophobic cavity occupied by the covalently bound FAD coenzyme at the distal end. To access the active site, the substrate (or inhibitor) must negotiate a turn near the intersection of the enzyme with the surface of the membrane, and enter through three loops that form a narrow path. In the human MAO-A, the entrance is surrounded by residues V93-E95, Y109-P112 and F208-N2012, which lie in three different loops, and therefore conformational fluctuations and flexibility of the three loops are critical for opening the entry for substrates into the active site (Binda et al., 2002; Son et al., 2008). Mutations that cause a rigidification of the loop have been associated with a reduction of the enzymatic activity (Son et al., 2008). In the A. mexicanus MAO, we have found that the overall 3D structure of the enzyme is very well conserved. However, the P106L mutation is located at the entrance of the active site formed by these three loops, and may confer additional rigidity to the loops, as suggested by predicted extra beta-strands on the secondary structure, which are more rigid than unfolded coil. Therefore, we propose that the P106L mutation in cavefish MAO may impair the conformational fluctuations necessary for loop flexibility and substrate entrance into the enzyme active site. This hypothesis derived from structural in silico modeling is well supported by computational predictions on the pathogenicity of the mutation obtained from the machine-learning based predictor MutPred2 (Pejaver et al., 2020) and, most importantly, by biochemical evidence (present study) showing that A. mexicanus MAO enzymatic activity is reduced by two-thirds by the P106L mutation. Thus, structural and biochemical evidence strongly suggest that P106L corresponds to a partial loss-of-function mutation. As a consequence, the oxidative deamination of monoamine neurotransmitters and biogenic amines must be strongly impaired in cavefish metabolism.

The cave-dwelling morphs of A. mexicanus are albino. Their lack of melanin pigmentation is due to a mutation in the gene Oca2 (Ocular and cutaneous albinism 2; involved in the transport of melanin precursor into melanosomes) (Protas et al., 2006). It was suggested that a potential benefit of albinism in cavefish could be to provide a surplus of L-tyrosine, the common precursor of the catecholamine and melanin synthesis pathways, and that this could explain the higher levels of DA and NA in cavefish brains (Bilandžija et al., 2013). In support of this hypothesis, morpholino knock-down of Oca2 expression in surface fish embryos decreased their pigmentation as expected, and simultaneously increased the L-tyrosine and DA content in 3-day-old larvae (Bilandžija et al., 2013). In contrast to this hypothesis, here we bring strong, genetically based evidence at different stages, from 6 days to 1 month and up to 5 months of age, that the differences in brain 5-HT, NA and DA content between SF and CF are fully and solely due to the P106L mao mutation. Indeed, the three monoamine HPLC levels were the same in non-mutant SF and non-mutant CF, and they were altered in mutant CF, thus pointing to the causal role of mao P106L in monoamine homeostasis in cavefish.

Importantly, the fact that mutant genotypes of cavefish and surface fish do not show identical values of NA and DA brain levels (although they always vary in the same direction as compared with non-mutants) suggests that the P106L mutation has slightly different outcomes depending on the genetic background, either surface or cave. These backgrounds are associated with many polymorphisms and/or mutations, mostly unknown and including the Oca2 alleles, which must differentially interact with the P106L mao mutation in the two morphs.

In a previous paper on the neuro-behavioral effects of the mao P106L mutation, we have reported surprisingly little effect on the cavefish phenotype, besides a clear involvement of altered brain monoamine homeostasis in the ‘stressability’, or amplitude of response to a stress represented by a change of environment (Pierre et al., 2020). This is at odds with the major cognitive and behavioral deficits caused by the only two MAO-A mutations (Glu313Stop and Cys266Phe) discovered so far in two human families (Brunner et al., 1993; Piton et al., 2014), and with the two-third reduction of MAO enzymatic activity and the ∼33% increase in monoamine brain levels we have measured in mutant fish. This discrepancy raises two points of discussion. First, there must be multiple and complex compensatory plastic mechanisms occurring in the brain and body of cavefish, which explain how they can cope with their mao genetic mutation. It will be important to study these plastic changes, as they can provide valuable information on general mechanisms of brain homeostasis. Here, we have found that this includes, at least in part, a compensation at the level of neurotransmitter synthesis, as shown by reduction of TPH activity in P106L mao cavefish mutants. Second, it is well known that cavefish display metabolic alterations, accumulate fat, and suffer hyperglycemia and diabetes owing to a mutation in insulin receptor (Aspiras et al., 2015; Riddle et al., 2018; Salin et al., 2010). As a potential role of MAO in metabolic and cardiovascular diseases has recently been raised (Deshwal et al., 2017; Shih, 2018), it will be important as well to examine a possible contribution of mao P106L to non-neural, metabolic phenotypes of cavefish.

General considerations on MAO and the 5-HT system in fish

There are two major differences between mammals and fish concerning MAO and the 5-HT system (Lillesaar, 2011). Fish possess only one MAO enzyme as opposed to MAO-A and -B in mammals, and they possess clusters of 5-HT neurons both in the hypothalamus and in the raphe nucleus, as opposed to raphe only in mammals. Astyanax mexicanus does not escape the rule in these respects (Elipot et al., 2013, 2014; Pierre et al., 2020).

Fish MAO has been suggested to have properties closer to mammalian MAO-A, including a suspected poor affinity for DA as a substrate. This idea emerged because inhibition of MAO by deprenyl in larval zebrafish strongly elevates 5-HT levels, but not DA or NA levels (Sallinen et al., 2009). The same holds true in A. mexicanus after deprenyl treatment (Elipot et al., 2014; present study). In line with this notion, here we found that the affinity of MAO was 40 times lower for DA than for 5-HT. However, after acute deprenyl treatment, we found no change in DA and NA levels like in zebrafish, but a strong decrease in the levels of DA metabolites DOPAC and HVA, which is counterintuitive if DA is not a MAO substrate (NB: DOPAC can only be produced by MAO-dependent degradation; therefore, its presence suggests that MAO does degrade DA). This discrepancy between the variations of metabolites and neurotransmitters cannot be explained easily either if one considers the alternative catabolic pathway for DA, through catechol-O-methyltransferase (COMT). Even if the fish MAO was to degrade efficiently DA, a conundrum would remain. Why and how do DA and NA levels remain stable when their metabolites vary? An unknown, fish-specific mechanism must exist, which buffers the expected rise in DA and NA when MAO is pharmacologically inhibited – but not when it is genetically mutated. Indeed, in the case of the mao P106L cavefish mutation, neurotransmitters and their metabolites show logical, inversely correlated variations, which also tends to suggest that DA and NA are bona fide MAO substrates in fish. These observations suggest that a ‘reference value’ for the level of these catecholamines might exist, and would depend on the mao genotype.

Interestingly, in our experiments, the hypothalamic and the raphe serotonergic neurons did not respond in the same manner to MAO perturbations (Fig. 6). In fact, although hypothalamic 5-HT levels varied according to the mao genotype, in the posterior brain the 5-HT levels remained the same in mao P106L CF mutants. This confirmed and reinforced previous results (Elipot et al., 2014), and suggested a compensation of the effects of the mao P106L mutation, in the cavefish raphe, to tightly regulate the 5HT levels. The decreased TPH activity we have observed in the posterior brain of the mutant cavefish – but not the mutant surface fish – may participate in the process. Furthermore, in CF, an acute MAO inhibition by deprenyl treatment did modify 5-HT raphe levels, whereas a chronic inhibition by the P106L genetic mutation did not. This suggests that the compensation in the raphe, if it exists, takes time to set up.

During embryonic development in zebrafish, the specification of the 5-HT neurons of the hypothalamus and the raphe is controlled by different transcription factors – pet1 in the raphe (Lillesaar et al., 2007) like in mammals, versus Etv5b in the hypothalamus (Bosco et al., 2013) – suggesting that the serotonergic identity can be acquired by convergent mechanisms. Moreover, hypothalamic and raphe 5-HT neurons express distinct paralogs of the 5-HT pathway markers: tph1 and sertb in the hypothalamus, and tph2 and serta in the raphe (Lillesaar, 2011). The timing of their differentiation is also different. We propose that the different ways the anterior/hypothalamic and the posterior/raphe 5-HT neurons respond to the partial loss-of-function mao mutation in cavefish must be related to their distinct serotonergic identity conferred by their distinct embryonic origins. Hence, the cavefish used as a model for evolutionary biology can provide interesting insights, in a comparative perspective, into basic knowledge on neurotransmission homeostasis.

General considerations on MAO-dependent brain chemistry

In the A. mexicanus brain, as expected, the P106L mao mutation increased the levels of 5-HT, DA and NA (except for 5-HT in the raphe, discussed above and below) and decreased the levels of their metabolites. These results are consistent with those found in MAO-A simple mutant or MAO-A/B double mutant mice, where the NA, 5-HT and DA levels are increased, and the 5-HIAA and DOPAC levels are decreased (Cases et al., 1995; Popova et al., 2001; Chen et al., 2004).

Here, we discuss the sometimes-contrasting results observed in P106L mao mutants versus acute deprenyl treatment. The P106L mao mutation changed DA and NA levels whereas acute deprenyl treatment did not. The latter is consistent with Sallinen et al. (2009), who found an increase of 5-HT levels but not NA and DA levels after deprenyl treatment in larval zebrafish. Therefore, pharmacological and genetic interference on MAO function does not have the same outcomes on fish brain neurochemistry. In mammals too, such contrasting effects have been observed. Let us take the raphe as an example to discuss the possible underlying mechanisms.

The mammalian raphe receives noradrenergic projections from the locus coeruleus (Levitt and Moore, 1979; Baraban and Aghajanian, 1981) and is under tonic activation via α1-adrenoreceptors (Baraban and Aghajanian, 1980). In contrast, 5-HT1A autoreceptors and α2-adrenoreceptors mediate an inhibitory effect on raphe 5-HT neuronal activity (Andrade et al., 2015; Tao and Hjorth, 1992; Numazawa et al., 1995; Raiteri et al., 1990; Trendelenburg et al., 1994; Starke et al., 1989). The 5-HT2A/2B/2C, 4 and 6 receptors (Belmer et al., 2018; Lucas and Debonnel, 2002; Liu et al., 2000; Boothman et al., 2006; Brouard et al., 2015), the endocanabinoid system (Haj-Dahmane and Shen, 2011) and GABAergic neurons (Hernández-Vázquez et al., 2019) also modulate the 5-HT system. Little is known about serotonergic modulation in fish, but the teleost and mammalian monoaminergic systems share numerous similarities (Panula et al., 2006) and their raphe 5-HT neurons can be considered homologous. Among others, projections to the raphe from the locus coeruleus (Ma, 1994), expression of α2-adrenoreceptors and the 5-HT1A receptor in the raphe (Ampatzis et al., 2008; Norton et al., 2008) have been described in fish.

In mammals, acute MAO-A inhibition causes a decrease of firing activity of 5-HT dorsal raphe neurons (Blier and De Montigny, 1985). The underlying suggested mechanism is that acute MAO-A inhibition would increase levels of synaptic 5-HT, which normally activates inhibitory 5-HT1A autoreceptors (Haddjeri et al., 1998). Along this line, 5-HT would accumulate within the cells because of the reduction of both 5-HT degradation by MAO-A after re-uptake, and 5-HT release and re-uptake owing to the decrease in firing activity (Haddjeri et al., 1998).

Conversely, with a chronic MAO-A inhibition (chronic pharmacological treatment or MAO-A knockout), the increase in 5-HT levels is more moderate (Haddjeri et al., 1998). Also, a long-term treatment with befloxatone (MAO-A inhibitor) induces a desensitization of inhibitory α2-adrenoreceptors (Blier and Bouchard, 1994; Mongeau et al., 1994; Owesson et al., 2002), and decreases their density (Cohen et al., 1982). Moreover, several studies suggested that a chronic MAO inhibition disrupts the inhibition of firing rate mediated by 5-HT1A autoreceptors (Owesson et al., 2002; Palfreyman et al., 1986; Evrard et al., 2002). Together, these effects lead to a disinhibition of 5-HT neurons.

A previous study reported a complete recovery of the raphe firing activity with chronic treatment by clorgyline (MAO-A inhibitor) (Blier and De Montigny, 1985), whereas another reported that the firing activity of 5-HT neurons was still lowered by 40% in MAO-A KO mice (Evrard et al., 2002). In sum, different responses of the 5-HT raphe system to acute or chronic MAO inhibition are described in mammals: decrease of 5-HT raphe activity with acute inhibition, and complete or partial re-establishment of 5-HT raphe activity with chronic inhibition. Equivalent studies were not performed in fish, but they give possible explanations to the different effects of acute (deprenyl) and chronic (P106L mutation) MAO inhibition on 5-HT raphe levels in fish.

Finally, we discovered a compensation for the effects of the P106L mutation by a decrease of TPH activity in mutant CF, and in both parts of the brain (note that such decrease is not observed in mutant SF). This could attenuate the increase of 5-HT levels in mutant CF, and explain why MAO enzymatic activity is proportionately more reduced than monoamine levels are enhanced in mutant brains. Of note, we did not detect such a significant reduction of TPH activity in our previous study when we used cavefish samples that were probably not genetically homogeneous (Elipot et al., 2014). This highlights the importance of re-examining all parameters in genetic lines with known genotypes, and the potential hidden pitfalls of working on natural, polymorphic populations. Further studies are needed to determine whether the expression of tph is decreased and, more generally, whether and how the brain transcriptome as a whole is profoundly affected or dysregulated in the mao mutant context. The mechanism of the ‘TPH compensation’ is also unknown, but one possibility could be that excess synaptic 5-HT leads to a negative feedback on 5-HT synthesis. For example, increased TPH activity has been shown after destruction of 5-HT-containing nerve terminals in mice (Stachowiak et al., 1986). But conversely, Popova et al. (2001) showed that in MOA-A KO mice, the TPH activity in the frontal cortex, hippocampus and amygdala was increased, which did not attenuate but rather aggravated the effects of the MAO-A deficit.

Future experiments will have to test whether the synthesis of DA and NA is modulated as well by the P106L mutation, which could compensate for the increase in DA and NA levels in mao mutants. Indeed, in mammals, several studies have shown modifications of TH activity after acute or chronic inhibition of MAO-B (Vrana et al., 1992; Lamensdorf and Finberg, 1997).

Conclusion and perspectives

Using the cavefish as a natural mutant, we have uncovered several novel aspects of monoaminergic regulation in fish. Fish are doubly special as compared with mammals because they have only one mao gene but possess diversified and diffuse clusters of monoaminergic neurons in their brains. However, they can serve as valuable models because most characteristics and most functions are shared with mammals. Our findings open many avenues for future studies. This will include the understanding of the enigmatic differences in homeostatic properties between hypothalamic and raphe serotonergic neurons, and the deciphering of the (probably) complex plastic compensatory mechanisms that occur in cavefish brains and allow them to thrive in caves despite a deleterious mutation in a major enzyme, MAO. Altogether, our work using the two morphs of A. mexicanus as a model system illustrates how brain homeostasis can be shaped by evolution.

We thank Stéphane Père, Victor Simon and Krystel Saroul for taking care of our Astyanax mexicanus colony hosted in the animal facility of the Institute.

Author contributions

Conceptualization: C.P., S.R.; Methodology: C.P., J.C., J.-M.L., J.L.; Validation: C.P., S.R.; Formal analysis: C.P., J.C., J.-M.L., S.R.; Investigation: C.P., J.C., J.-M.L.; Resources: S.R.; Data curation: C.P., J.C., J.-M.L., S.R.; Writing - original draft: C.P., S.R.; Writing - review & editing: S.R.; Visualization: C.P., S.R.; Supervision: S.R.; Project administration: S.R.; Funding acquisition: S.R.

Funding

This research was supported by grants from Equipe Fondation pour la Recherche Médicale (FRM) (DEQ20150331745 and EQU202003010144), Centre National pour la Recherche Scientifique and Institut Diversité Evolution Ecologie du Vivant to S.R. C.P. received PhD fellowships from the French Ministry of Research and from FRM.

Data availability

All relevant data can be found within the article and its supplementary information.

Aldeco
,
M.
,
Arslan
,
B. K.
and
Edmondson
,
D. E.
(
2011
).
Catalytic and inhibitor binding properties of zebrafish monoamine oxidase (zMAO): comparisons with human MAO A and MAO B
.
Comp. Biochem. Physiol. B Biochem. Mol. Biol.
159
,
78
-
83
.
Ampatzis
,
K.
,
Kentouri
,
M.
and
Dermon
,
C. R.
(
2008
).
Neuronal and glial localization of α2A-adrenoceptors in the adult zebrafish (Danio rerio) brain
.
J. Comp. Neurol.
508
,
72
-
93
.
Andrade
,
R.
,
Huereca
,
D.
,
Lyons
,
J. G.
,
Andrade
,
E. M.
and
Mcgregor
,
K. M.
(
2015
).
5-HT1A receptor-mediated autoinhibition and the control of serotonergic cell firing
.
ACS Chem. Neurosci.
6
,
1110
-
1115
.
Anichtchik
,
O.
,
Sallinen
,
V.
,
Peitsaro
,
N.
and
Panula
,
P.
(
2006
).
Distinct structure and activity of monoamine oxidase in the brain of zebrafish (Danio rerio)
.
J. Comp. Neurol.
498
,
598
-
610
.
Aspiras
,
A. C.
,
Rohner
,
N.
,
Martineau
,
B.
,
Borowsky
,
R. L.
and
Tabin
,
C. J.
(
2015
).
Melanocortin 4 receptor mutations contribute to the adaptation of cavefish to nutrient-poor conditions
.
Proc. Natl. Acad. Sci. USA
112
,
9668
-
9673
.
Baraban
,
J. M.
and
Aghajanian
,
G. K.
(
1980
).
Suppression of firing activity of 5-HT neurons in the dorsal raphe by alpha-adrenoceptor antagonists
.
Neuropharmacology
19
,
355
-
363
.
Baraban
,
J. M.
and
Aghajanian
,
G. K.
(
1981
).
Noradrenergic innervation of serotonergic neurons in the dorsal raphe: demonstration by electron microscopic autoradiography
.
Brain Res.
204
,
1
-
11
.
Baronio
,
D.
,
Chen
,
Y. C.
and
Panula
,
P.
(
2022
).
Abnormal brain development of monoamine oxidase mutant zebrafish and impaired social interaction of heterozygous fish
.
Dis. Model. Mech.
15
,
dmm049133
.
Beale
,
A.
,
Guibal
,
C.
,
Tamai
,
T. K.
,
Klotz
,
L.
,
Cowen
,
S.
,
Peyric
,
E.
,
Reynoso
,
V. H.
,
Yamamoto
,
Y.
and
Whitmore
,
D.
(
2013
).
Circadian rhythms in Mexican blind cavefish Astyanax mexicanus in the lab and in the field
.
Nat. Commun.
4
,
2769
.
Belmer
,
A.
,
Quentin
,
E.
,
Diaz
,
S. L.
,
Guiard
,
B. P.
,
Fernandez
,
S. P.
,
Doly
,
S.
,
Banas
,
S. M.
,
Pitychoutis
,
P. M.
,
Moutkine
,
I.
,
Muzerelle
,
A.
et al. 
(
2018
).
Positive regulation of raphe serotonin neurons by serotonin 2B receptors
.
Neuropsychopharmacology
43
,
1623
-
1632
.
Berry
,
M. D.
,
Juorio
,
A. V.
,
Paterson
,
I. A.
(
1994
).
The functional role of monoamine oxidases A and B in the mammalian central nervous system
.
Prog. Neurobiol.
42
,
375
-
391
.
Bilandžija
,
H.
,
Ma
,
L.
,
Parkhurst
,
A.
and
Jeffery
,
W. R.
(
2013
).
A potential benefit of albinism in Astyanax cavefish: downregulation of the oca2 gene increases tyrosine and catecholamine levels as an alternative to melanin synthesis
.
PLoS One
8
,
e80823
.
Binda
,
C.
,
Newton-Vinson
,
P.
,
Hubálek
,
F.
,
Edmondson
,
D. E.
and
Mattevi
,
A.
(
2002
).
Structure of human monoamine oxidase B, a drug target for the treatment of neurological disorders
.
Nat. Struct. Biol.
9
,
22
-
26
.
Blier
,
P.
and
Bouchard
,
C.
(
1994
).
Modulation of 5-HT release in the guinea-pig brain following long-term administration of antidepressant drugs
.
Br. J. Pharmacol.
113
,
485
-
495
.
Blier
,
P.
and
De Montigny
,
C.
(
1985
).
Serotoninergic but not noradrenergic neurons long-term treatment with monoamine oxidase inhibitors
.
Neuroscience
16
,
949
-
955
.
Boothman
,
L.
,
Raley
,
J.
,
Denk
,
F.
,
Hirani
,
E.
and
Sharp
,
T.
(
2006
).
In vivo evidence that 5-HT 2C receptors inhibit 5-HT neuronal activity via a GABAergic mechanism
.
Br. J. Pharmacol.
149
,
861
-
869
.
Bortolato
,
M.
and
Shih
,
J. C.
(
2011
).
Behavioral Outcomes of Monoamine Oxidase Deficiency: Preclinical and Clinical Evidence
, 1st edn.
Elsevier
.
Bosco
,
A.
,
Bureau
,
C.
,
Affaticati
,
P.
,
Gaspar
,
P.
,
Bally-Cuif
,
L.
and
Lillesaar
,
C.
(
2013
).
Development of hypothalamic serotoninergic neurons requires Fgf signalling via the ETS-domain transcription factor Etv5b
.
Dev.
140
,
372
-
384
.
Brouard
,
J. T.
,
Schweimer
,
J. V.
,
Houlton
,
R.
,
Burnham
,
K. E.
,
Quérée
,
P.
and
Sharp
,
T.
(
2015
).
Pharmacological evidence for 5-HT6 receptor modulation of 5-HT neuron firing in vivo
.
ACS Chem. Neurosci.
6
,
1241
-
1247
.
Brunner
,
H. G.
,
Nelen
,
M.
,
Breakefield
,
X. O.
,
Ropers
,
H. H.
,
Oost
,
B. A.
and
Van
, (
1993
).
Abnormal behavior associated with a point mutation in the structural gene for monoamine oxidase A
.
Science
262
,
578
-
580
.
Cases
,
O.
,
Seif
,
I.
,
Grimsby
,
J.
,
Gaspar
,
P.
,
Chen
,
K.
,
Pournin
,
S.
,
Müller
,
U.
,
Aguet
,
M.
,
Babinet
,
C.
,
Shih
,
J. C.
et al. 
(
1995
).
Aggressive behavior and altered amounts of brain serotonin and norepinephrine in mice lacking MAOA
.
Science
268
,
1763
-
1766
.
Chen
,
K.
,
Wu
,
H. F.
,
Grimsby
,
J.
and
Shih
,
J. C.
(
1994
).
Cloning of a novel monoamine oxidase cDNA from trout liver
.
Mol. Pharmacol.
46
,
1226
-
1233
.
Chen
,
K.
,
Holschneider
,
D. P.
,
Wu
,
W.
,
Rebrini
,
I.
and
Shih
,
J. C.
(
2004
).
A spontaneous point mutation produces monoamine oxidase A/B knock-out mice with greatly elevated monoamines and anxiety-like behavior
.
J. Biol. Chem.
279
,
39645
-
39652
.
Cohen
,
R. M.
,
Campbell
,
I. C.
,
Dau'phin
,
M.
,
Tallman
,
J. F.
and
Murphy
,
D. L.
(
1982
).
Changes in α- and β-receptor densities in rat brain as a result of treatment with monoamine oxidase inhibiting antidepressants
.
Neuropharmacology
21
,
293
-
298
.
De Colibus
,
L.
,
Li
,
M.
,
Binda
,
C.
,
Lustig
,
A.
,
Edmondson
,
D. E.
and
Mattevi
,
A.
(
2005
).
Three-dimensional structure of human monoamine oxidase A (MAO A): relation to the structures of rat MAO A and human MAO B
.
Proc. Natl. Acad. Sci. USA
102
,
12684
-
12689
.
Deshwal
,
S.
,
Di Sante
,
M.
,
Di Lisa
,
F.
, and
Kaludercic
,
N.
(
2017
).
Emerging role of monoamine oxidase as a therapeutic target for cardiovascular disease
.
Curr. Opin. Pharmacol.
33
,
64
-
69
.
Elipot
,
Y.
,
Hinaux
,
H.
,
Callebert
,
J.
and
Rétaux
,
S.
(
2013
).
Evolutionary shift from fighting to foraging in blind cavefish through changes in the serotonin network
.
Curr. Biol.
23
,
1
-
10
.
Elipot
,
Y.
,
Hinaux
,
H.
,
Callebert
,
J.
,
Launay
,
J. M.
,
Blin
,
M.
and
Rétaux
,
S.
(
2014
).
A mutation in the enzyme monoamine oxidase explains part of the Astyanax cavefish behavioural syndrome
.
Nat. Commun.
5
,
3647
.
Elliott
,
W. R.
(
2018
).
The Astyanax caves of Mexico. Cavefishes of Tamaulipas, San Luis Potosi, and Guerrero
.
Assoc. Mex. Cave Stud. Bull.
2
.
Evrard
,
A.
,
Malagié
,
I.
,
Laporte
,
A. M.
,
Boni
,
C.
,
Hanoun
,
N.
,
Trillat
,
A. C.
,
Seif
,
I.
,
De Maeyer
,
E.
,
Gardier
,
A.
,
Hamon
,
M.
et al. 
(
2002
).
Altered regulation of the 5-HT system in the brain of MAO-A knock-out mice
.
Eur. J. Neurosci.
15
,
841
-
851
.
Fingerman
,
S. W.
(
1976
).
Circadian rhythms of brain 5-hydroxytryptamine and swimming activity in the teleost, Fundulus grandis
.
Comp. Biochem. Physiol. Part C, Comp
.
54
,
49
-
53
.
Grimsby
,
J.
,
Chen
,
K.
,
Wang
,
L. J.
,
Lan
,
N. C.
and
Shih
,
J. C.
(
1991
).
Human monoamine oxidase A and B genes exhibit identical exon-intron organization
.
Proc. Natl. Acad. Sci. USA
88
,
3637
-
3641
.
Grimsby
,
J.
,
Toth
,
M.
,
Chen
,
K.
,
Kumazawa
,
T.
,
Klaidman
,
L.
,
Adams
,
J. D.
,
Karoum
,
F.
,
Gal
,
J.
and
Shih
,
J. C.
(
1997
).
Increased stress response and beta-phenylethylamine in MAOB-deficient mice
.
Nat. Genet.
17
,
206
-
210
.
Haddjeri
,
N.
,
De Montigny
,
C.
,
Curet
,
O.
and
Blier
,
P.
(
1998
).
Effect of the reversible monoamine oxidase A-inhibitor befloxatone on the rat 5-hydroxytryptamine neurotransmission
.
Eur. J. Pharmacol.
343
,
179
-
192
.
Haj-Dahmane
,
S.
and
Shen
,
R.-Y.
(
2011
).
Modulation of the serotonin system by endocannabinoid signaling
.
Neuropharmacology
61
,
414
-
420
.
Hernández-Vázquez
,
F.
,
Garduño
,
J.
and
Hernández-López
,
S.
(
2019
).
GABAergic modulation of serotonergic neurons in the dorsal raphe nucleus
.
Rev. Neurosci.
30
,
289
-
303
.
Khan
,
I. A.
and
Joy
,
K. P.
(
1988
).
Seasonal and daily variations in hypothalamic monoamine levels and monoamine oxidase activity in the teleost Channa punctatus (Bloch)
.
Chronobiol. Int.
5
,
311
-
316
.
Lamensdorf
,
I.
and
Finberg
,
J. P. M.
(
1997
).
Reduced striatal tyrosine hydroxylase activity is not accompanied by change in responsiveness of dopaminergic receptors following chronic treatment with deprenyl
.
Neuropharmacology
36
,
1455
-
1461
.
Levitt
,
P.
and
Moore
,
R. Y.
(
1979
).
Origin and organization of brainstem catecholamine innervation in the rat
.
J. Comp. Neurol.
186
,
505
-
528
.
Levitt
,
P.
,
Pintar
,
J. E.
and
Breakefield
,
X. O.
(
1982
).
Immunocytochemical demonstration of monoamine oxidase B in brain astrocytes and serotonergic neurons
.
Proc. Natl. Acad. Sci. USA
79
,
6385
-
6389
.
Lillesaar
,
C.
(
2011
).
The serotonergic system in fish
.
J. Chem. Neuroanat.
41
,
294
-
308
.
Lillesaar
,
C.
,
Tannhäuser
,
B.
,
Stigloher
,
C.
,
Kremmer
,
E.
and
Bally-Cuif
,
L.
(
2007
).
The serotonergic phenotype is acquired by converging genetic mechanisms within the zebrafish central nervous system
.
Dev. Dyn.
236
,
1072
-
1084
.
Liu
,
R.
,
Jolas
,
T.
and
Aghajanian
,
G.
(
2000
).
Serotonin 5-HT2 receptors activate local GABA inhibitory inputs to serotonergic neurons of the dorsal raphe nucleus
.
Brain Res.
873
,
34
-
45
.
Lucas
,
G.
and
Debonnel
,
G.
(
2002
).
5-HT4 receptors exert a frequency-related facilitatory control on dorsal raphé nucleus 5-HT neuronal activity
.
Eur. J. Neurosci.
16
,
817
-
822
.
Ma
,
P. K. M.
(
1994
).
Catecholaminergic systems in the zebrafish. II. Projection pathways and pattern of termination of the locus coeruleus
.
J. Comp. Neurol.
344
,
256
-
269
.
Ma
,
J.
,
Yoshimura
,
M.
,
Yamashita
,
E.
,
Nakagawa
,
A.
,
Ito
,
A.
and
Tsukihara
,
T.
(
2004
).
Structure of rat monoamine oxidase A and its specific recognitions for substrates and inhibitors
.
J. Mol. Biol.
338
,
103
-
114
.
Mitchell
,
R.
,
Russell
,
W.
and
Elliott
,
W.
(
1977
).
Mexican eyeless characin fishes, genus Astyanax: environment, distribution, and evolution
.
Spec. Publ. Mus. Texas Tech. Univ
12
,
1
-
89
.
Mongeau
,
R.
,
De Montigny
,
C.
, and
Blier
,
P.
(
1994
).
Electrophysiologic evidence for desensitization of α2-adrenoceptors on serotonin terminals following long-term treatment with drugs increasing norepinephrine synaptic concentration
.
Neuropsychopharmacology
10
,
41
-
51
.
Moran
,
D.
,
Softley
,
R.
and
Warrant
,
E. J.
(
2014
).
Eyeless Mexican cavefish save energy by eliminating the circadian rhythm in metabolism
.
PLoS One
9
,
e107877
.
Norton
,
W. H. J.
,
Folchert
,
A.
and
Bally-Cuif
,
L.
(
2008
).
Comparative analysis of serotonin receptor (HTR1A/HTR1B families) and transporter (slc6a4a/b) gene expression in the zebrafish brain
.
J. Comp. Neurol
.
511
,
521
-
542
.
Numazawa
,
R.
,
Yoshioka
,
M.
,
Matsumoto
,
M.
,
Togashi
,
H.
,
Kemmotsu
,
O.
and
Saito
,
H.
(
1995
).
Pharmacological characterization of α2-adrenoceptor regulated serotonin release in the rat hippocampus
.
Neurosci. Lett.
192
,
161
-
164
.
Owesson
,
C. Å.
,
Hopwood
,
S. E.
,
Callado
,
L. F.
,
Seif
,
I.
,
Mclaughlin
,
D. P.
and
Stamford
,
J. A.
(
2002
).
Altered presynaptic function in monoaminergic neurons of monoamine oxidase-A knockout mice
.
Eur. J. Neurosci.
15
,
1516
-
1522
.
Palfreyman
,
M. G.
,
Mir
,
A. K.
,
Kubina
,
M.
,
Middlemiss
,
D. N.
,
Richards
,
M.
,
Tricklebank
,
M. D.
and
Fozard
,
J. R.
(
1986
).
Monoamine receptor sensitivity changes following chronic administration of MDL 72394, a site-directed inhibitor of monoamine oxidase
.
Eur. J. Pharmacol.
130
,
73
-
89
.
Panula
,
P.
,
Sallinen
,
V.
,
Sundvik
,
M.
,
Kolehmainen
,
J.
,
Torkko
,
V.
,
Tiittula
,
A.
,
Moshnyakov
,
M.
and
Podlasz
,
P.
(
2006
).
Modulatory neurotransmitter systems and behavior: towards zebrafish models of neurodegenerative diseases
.
Zebrafish
3
,
235
-
247
.
Pejaver
,
V.
,
Urresti
,
J.
,
Lugo-Martinez
,
J.
,
Pagel
,
K. A.
,
Lin
,
G. N.
,
Nam
,
H.-J.
,
Mort
,
M.
,
Cooper
,
D. N.
,
Sebat
,
J.
,
Iakoucheva
,
L. M.
et al. 
(
2020
).
MutPred2: inferring the molecular and phenotypic impact of amino acid variants
.
Nat. Commun.
11
,
5918
.
Piton
,
A.
,
Poquet
,
H.
,
Redin
,
C.
,
Masurel
,
A.
,
Lauer
,
J.
,
Muller
,
J.
,
Thevenon
,
J.
,
Herenger
,
Y.
,
Chancenotte
,
S.
,
Bonnet
,
M.
et al. 
(
2014
).
20 ans après: a second mutation in MAOA identified by targeted high-throughput sequencing in a family with altered behavior and cognition
.
Eur. J. Hum. Genet.
22
,
776
-
783
.
Pierre
,
C.
,
Pradere
,
N.
,
Froc
,
C.
,
Ornelas-Garcia
,
P.
,
Callebert
,
J.
and
Retaux
,
S.
(
2020
).
A mutation in monoamine oxidase (MAO) affects the evolution of stress behavior in the blind cavefish Astyanax mexicanus
.
J. Exp. Biol.
223
,
jeb226092
.
Popova
,
N. K.
,
Gilinsky
,
M. A.
,
Amstislavskaya
,
T. G.
,
Morosova
,
E. A.
,
Seif
,
I.
,
Maeyer
,
E.
and
De
, (
2001
).
Regional serotonin metabolism in the brain of transgenic mice lacking monoamine oxidase A
.
J. Neurosci. Res.
66
,
423
-
427
.
Protas
,
M. E.
,
Hersey
,
C.
,
Kochanek
,
D.
,
Zhou
,
Y.
,
Wilkens
,
H.
,
Jeffery
,
W. R.
,
Zon
,
L. I.
,
Borowsky
,
R.
and
Tabin
,
C. J.
(
2006
).
Genetic analysis of cavefish reveals molecular convergence in the evolution of albinism
.
Nat. Genet.
38
,
107
-
111
.
Raiteri
,
M.
,
Maura
,
G.
,
Folghera
,
S.
,
Cavazzani
,
P.
,
Andrioli
,
G. C.
,
Schlicker
,
E.
,
Schalnus
,
R.
and
Göthert
,
M.
(
1990
).
Modulation of 5-hydroxytryptamine release by presynaptic inhibitory α2-adrenoceptors in the human cerebral cortex
.
Naunyn. Schmiedebergs. Arch. Pharmacol.
342
,
508
-
512
.
Riddle
,
M. R.
,
Aspiras
,
A. C.
,
Gaudenz
,
K.
,
Peuss
,
R.
,
Sung
,
J. Y.
,
Martineau
,
B.
,
Peavey
,
M.
,
Box
,
A. C.
,
Tabin
,
J. A.
,
Mcgaugh
,
S.
et al. 
(
2018
).
Insulin resistance in cavefish as an adaptation to a nutrient-limited environment
.
Nature
555
,
647
-
651
.
Salin
,
K.
,
Voituron
,
Y.
,
Mourin
,
J.
and
Hervant
,
F.
(
2010
).
Cave colonization without fasting capacities: an example with the fish Astyanax fasciatus mexicanus
.
Comp. Biochem. Physiol. A Mol. Integr. Physiol.
156
,
451
-
457
.
Sallinen
,
V.
,
Sundvik
,
M.
,
Reenilä
,
I.
,
Peitsaro
,
N.
,
Khrustalyov
,
D.
,
Anichtchik
,
O.
,
Toleikyte
,
G.
,
Kaslin
,
J.
and
Panula
,
P.
(
2009
).
Hyperserotonergic phenotype after monoamine oxidase inhibition in larval zebrafish
.
J. Neurochem.
109
,
403
-
415
.
Setini
,
A.
,
Pierucci
,
F.
,
Senatori
,
O.
and
Nicotra
,
A.
(
2005
).
Molecular characterization of monoamine oxidase in zebrafish (Danio rerio)
.
Comp. Biochem. Physiol. B Biochem. Mol. Biol.
140
,
153
-
161
.
Shih
,
J. C.
(
2018
).
Monoamine oxidase isoenzymes: genes, functions and targets for behavior and cancer therapy
.
J. Neural Transm.
125
,
1553
-
1566
.
Shih
,
J. C.
,
Chen
,
K.
and
Ridd
,
M. J.
(
1999
).
Monoamine oxidase: from genes to behavior
.
Annu. Rev. Neurosci.
22
,
197
-
217
.
Son
,
S. Y.
,
Ma
,
J.
,
Kondou
,
Y.
,
Yoshimura
,
M.
,
Yamashita
,
E.
and
Tsukihara
,
T.
(
2008
).
Structure of human monoamine oxidase A at 2.2-Å resolution: the control of opening the entry for substrates/inhibitors
.
Proc. Natl. Acad. Sci. USA
105
,
5739
-
5744
.
Stachowiak
,
M. K.
,
Stricker
,
E. M.
,
Jacoby
,
J. H.
and
Zigmond
,
M. J.
(
1986
).
Increased tryptophan hydroxylase activity in serotonergic nerve terminals spared by 5,7-dihydroxytryptamine
.
Biochem. Pharmacol.
35
,
1241
-
1248
.
Starke
,
K.
,
Gothert
,
M.
and
Kilbinger
,
H.
(
1989
).
Modulation of neurotransmitter release by presynaptic autoreceptors
.
Physiol. Rev.
69
,
864
-
989
.
Tao
,
R.
and
Hjorth
,
S.
(
1992
).
a2-Adrenoceptor modulation of rat ventral hippocampal 5-hydroxytryptamine release in vivo
.
Naunyn. Schmiedebergs. Arch. Pharmacol.
345
,
137
-
143
.
Trendelenburg
,
A. U.
,
Trendelenburg
,
M.
,
Starke
,
K.
and
Limberger
,
N.
(
1994
).
Release-inhibiting α2-adrenoceptors at serotonergic axons in rat and rabbit brain cortex: evidence for pharmacological identity with α2-autoreceptors
.
Naunyn. Schmiedebergs. Arch. Pharmacol.
349
,
25
-
33
.
Veselovsky
,
A. V.
,
Ivanov
,
A. S.
and
Medvedev
,
A. E
. (
2004
).
Computer modelling and visualization of active site of monoamine oxidases
.
Neurotoxicology
25
,
37
-
46
.
Vrana
,
S. L.
,
Azzaro
,
A. J.
and
Vrana
,
K. E.
(
1992
).
Chronic selegiline administration transiently decreases tyrosine hydroxylase activity and mRNA in the rat nigrostriatal pathway
.
Mol. Pharmacol.
41
,
839
-
844
.
Yang
,
J.
,
Yan
,
R.
,
Roy
,
A.
,
Xu
,
D.
,
Poisson
,
J.
and
Zhang
,
Y.
(
2015
).
The I-TASSER suite: protein structure and function prediction
.
Nat. Methods
12
,
7
-
8
.
Yang
,
J.
,
Wang
,
Y.
and
Zhang
,
Y.
(
2016
).
ResQ: an approach to unified estimation of B-factor and residue-specific error in protein structure prediction
.
J. Mol. Biol.
428
,
693
-
701
.
Youdim
,
M. B. H.
,
Edmondson
,
D.
and
Tipton
,
K. F.
(
2006
).
The therapeutic potential of monoamine oxidase inhibitors
.
Nat. Rev. Neurosci.
7
,
295
-
309
.
Ziegler
,
C.
and
Domschke
,
K.
(
2018
).
Epigenetic signature of MAOA and MAOB genes in mental disorders
.
J. Neural Transm.
125
,
1581
-
1588
.

Competing interests

The authors declare no competing or financial interests.

Supplementary information