Monoamine oxidase (MAO-A and MAO-B)

In brief

What monoamine oxidase actually is

A monoamine is a signaling molecule with one nitrogen-containing side chain hanging off a ring: serotonin, norepinephrine, dopamine, and the trace amines tyramine and beta-phenylethylamine belong to the family. Monoamine oxidase is the enzyme that ends them. It takes the carbon next to the nitrogen, removes two hydrogens from it, and leaves behind an unstable imine that immediately breaks up in water into an aldehyde and ammonia. The two electrons it removed go onto the enzyme's flavin cofactor, and from there onto molecular oxygen, which becomes hydrogen peroxide ^1,2.

That last detail is the one most descriptions skip, and the one that matters for everything in the second half of this page. Every molecule of serotonin or norepinephrine or dopamine this enzyme destroys produces one molecule of hydrogen peroxide inside the cell, immediately outside the mitochondrion. In tissues that handle a lot of amine, that is a standing source of oxidative stress rather than an incidental one ^1,31.

The protein itself is anchored in the outer mitochondrial membrane by a helix at its tail end, so that the working part of the enzyme faces the cytoplasm while the molecule is tethered to the organelle ^1,7. Its cofactor, FAD, is not loosely held the way cofactors usually are: it is covalently bonded to a cysteine inside the sequence Ser-Gly-Gly-Cys-Tyr, which is the same in both enzymes and in every species that has been sequenced ⁵. In human MAO-B, pargyline, a relative of the Parkinson's drug selegiline, was caught in the crystal structure bonded to nitrogen 5 of that flavin, which is exactly how the irreversible inhibitors work: they are substrates that get halfway through the reaction and then stick ⁷.

Human blood platelets carry MAO-B and not MAO-A, which is a biological accident that shaped forty years of psychiatric research, because it made the enzyme measurable from a blood sample in anyone who would hold out an arm ^22,38.

Why there are two of them, and how we know

For forty years after it was found, monoamine oxidase was one enzyme with inconsistent behavior: different tissues and different species gave different answers about which amines it preferred and which drugs blocked it. The split was made in 1968, when J. P. Johnston reported observations on a new inhibitor of monoamine oxidase in brain tissue, the compound now known as clorgyline, and used it to define two activities, which he called A and B ¹³. That paper's PubMed record carries no abstract, so the detail has to be read in the original; what is not in doubt is that the A and B nomenclature starts there and that it was defined pharmacologically, by what a drug did, long before anyone had the proteins.

The proteins came in 1988. Bach and colleagues cloned the complementary DNA for both forms from human liver and found deduced subunit molecular weights of 59,700 for MAO-A and 58,800 for MAO-B, sequences 70 percent identical, the same FAD-binding pentapeptide in both, and messenger RNA patterns indicating that they came from two separate genes rather than one gene spliced two ways ⁵. In 1989 the two genes were mapped to the short arm of the X chromosome, close to each other and to the DXS7 marker, using cell hybrids, in situ hybridization and a deletion in a patient with Norrie disease, an inherited blindness syndrome whose gene lies in the same neighborhood ³. In 1991 the gene structures were published side by side: at least 60 kilobases each, 15 exons apiece, an identical exon-intron organization, and exon 12, which encodes the covalent flavin site, the most conserved of all at 93.9 percent peptide identity ⁴. Two genes, one ancestor, one duplication.

Which amines each prefers was settled in animals. MAO-A preferentially oxidizes serotonin and norepinephrine, MAO-B preferentially oxidizes beta-phenylethylamine, and both handle dopamine. Knocking out MAO-B in mice raised brain phenylethylamine and left serotonin, norepinephrine and dopamine unchanged, which is as clean a demonstration of substrate preference as the field has; those mice were also more reactive to stress and resistant to the parkinsonism-inducing toxin MPTP ⁸.

The crystal structures arrived in the 2000s and explained the drugs. Human MAO-B at 3 angstroms showed a 420 cubic angstrom substrate cavity opening into a 290 cubic angstrom entrance cavity, with two tyrosines forming an aromatic cage around the incoming amino group, and pargyline bonded covalently to the flavin ⁷. Human MAO-A, solved in 2005 as its clorgyline adduct, turned out to have one continuous cavity of about 550 cubic angstroms, to crystallize as a monomer and to behave as a monomer in solution, unlike human MAO-B and rat MAO-A, which are dimers; the authors drew the pointed conclusion that using MAO-A from non-human sources in drug development is questionable ⁶. Later work using pulsed electron paramagnetic resonance on membrane preparations found both human enzymes dimeric when still in their membrane, with distances matching the crystals, so the monomer may be a property of the purified protein ¹.

The whole history, including the reversals

The enzyme was found in liver, in Cambridge, in 1928, by a graduate student in the Biochemical Laboratory whom the record names only as M. L. Hare; she called it tyramine oxidase and published it as a new enzyme system in liver ⁹. The record carries no abstract, so it is cited here for the finding, the name and the date. She is the person a 1999 biographical note calls Mary Bernheim, the married name under which she published the rest of a long career, which is why the discovery is filed sometimes under one name and sometimes the other ¹⁰. Nine years later Blaschko, Richter and Schlossmann published on the inactivation of adrenaline, the work conventionally credited with pulling the separate amine-oxidizing activities into one enzyme; that record has no abstract either, and the unification claim belongs to the historical literature rather than to anything an abstract can be read for ¹¹. By 1952 the enzyme was called monoamine oxidase in print, in a paper in which Zeller and Barsky showed that a new tuberculosis drug inhibited liver and brain monoamine oxidase in a living animal ¹². The name is conventionally credited to Zeller, who drew the line between the enzyme that attacks amines carrying one amino group and the diamine oxidase that attacks those carrying two; the paper in which he is said to have introduced the term in the late 1930s is not indexed in PubMed, so this page records that attribution as conventional rather than checkable and points the reader to a ninety-year historical review for the longer account ⁴⁶. That 1952 paper is the hinge of the whole story.

The tuberculosis drug was iproniazid, a hydrazine derivative of isonicotinic acid marketed as Marsilid, reported in 1952 in both the Journal of the American Medical Association and the American Review of Tuberculosis, the second of those titles naming Rimifon, that is isoniazid, alongside it ^14,15. Neither record carries an abstract, so the trial setting, conventionally given as Sea View Hospital on Staten Island, belongs to the historical literature rather than to anything checkable here. What happened next is the serendipity that the textbooks tell: the patients on the drug were described as unusually cheerful. In 1957 Loomer, Saunders and Kline published a clinical and pharmacodynamic evaluation of iproniazid as a psychic energizer, the paper conventionally credited with the first deliberate psychiatric use of a monoamine oxidase inhibitor; Crane reported iproniazid phosphate as a therapeutic agent for mental disorders and debilitating diseases in the same volume of the same journal, thirteen pages later, so the priority is shared rather than clean ^16,17. Neither of those two records has an abstract or a DOI in PubMed, so neither design can be checked from the database; their dates and their claims are what survive.

The class spread fast, and so did its most famous hazard. In October 1963 two letters appeared in The Lancet three weeks apart: on 5 October one from Asatoor, Levi and Milne titled simply Tranylcypromine and cheese ¹⁹, and on 26 October one from Blackwell reporting hypertensive crisis due to monoamine-oxidase inhibitors, with cheese, phenelzine, tranylcypromine and headache among its indexing terms ¹⁸. Neither carries an abstract. The story told afterwards, that the connection was noticed by a British pharmacist whose wife, taking one of these drugs, had a severe headache every time they ate cheese, is in the literature after all, in a 2009 historical note in the Indian Journal of Psychiatry; that is a secondary account rather than a primary document, and it is cited here as one ⁴⁷. What is in the primary record is what came next: Blackwell and Marley worked the interaction out experimentally in cats and rats in 1966 ²⁰, and in 1967 Blackwell, Marley, Price and Taylor published a full account of hypertensive interactions between monoamine oxidase inhibitors and foodstuffs, indexed under cerebral hemorrhage and subarachnoid hemorrhage ²¹. The reaction could kill, and the class fell out of fashion for a generation.

Meanwhile the biology kept opening. Johnston's 1968 inhibitor split the enzyme into A and B ¹³. The accidental poisoning of four people who injected an illicit drug that turned out to be mainly MPTP, reported in 1983 by Langston and colleagues, who proposed that the chemical selectively damages cells in the substantia nigra ⁴⁵, led to the 1984 demonstration that brain monoamine oxidase converts MPTP into MPP+, the 1-methyl-4-phenylpyridinium ion, which is what actually destroys dopamine neurons, a conversion blocked by deprenyl ⁴¹. The genes were cloned, mapped and sequenced between 1988 and 1991 ^3,4,5. Then the clinical reversals came, one after another, and they are the reason this page exists: a 1989 trial that seemed to show a drug slowing Parkinson's disease and was reinterpreted in 1993 and 1994 as symptomatic ^48,49,50; a 1995 trial reporting higher mortality on the same drug, which a 2004 meta-analysis of 3525 patients did not confirm ^51,52; a gene-by-environment finding in 2002 that replicated, then became a tabloid genetic explanation of violence that the meta-analyses do not support in the form the newspapers used ^53,54,55,56.

The inhibitor class, drug by drug

Monoamine oxidase inhibitors were among the first antidepressants, arriving in 1957 alongside imipramine, the first of the tricyclics, and they have spent seventy years being rediscovered and set aside; which of the two came first is a question of what counts as first, and the iproniazid record in PubMed carries neither an abstract nor a DOI, so this page does not adjudicate it ^16,57. The first generation, built around the hydrazine chemistry of iproniazid, bind the flavin permanently; phenelzine and isocarboxazid are the surviving examples, and both still beat placebo in the modern pooling ⁷³. Tranylcypromine, which is not a hydrazine but an amphetamine-like cyclopropylamine, also binds irreversibly and blocks both forms. The second generation went two ways. Moclobemide blocks MAO-A reversibly, so that a rising concentration of tyramine can displace it, which is why it does not need a diet sheet ³⁴. Selegiline and rasagiline block MAO-B irreversibly and selectively at the doses used in Parkinson's disease, leaving MAO-A in the gut free to do its job. Safinamide, licensed as an add-on in Parkinson's disease, is described in the drug-class reviews as a reversible MAO-B inhibitor ⁴⁴, and it also blocks voltage-dependent sodium and calcium channels and the calcium-mediated release of glutamate ², so its benefit cannot be attributed to the enzyme alone, which the drug's own reviewers concede in writing that its mechanism has not been fully defined ⁷⁷. The clinical standard for using the classic inhibitors today is a 2022 consensus guide, written for prescribers who are not experienced with them and endorsed by more than 70 international experts drawing on six decades of use. Its central recommendations are that phenelzine, tranylcypromine and isocarboxazid should always be considered in treatment-resistant depression, including the melancholic kind, and before electroconvulsive therapy; that in selected patients they can come earlier in the sequence than custom allows and should not be treated as drugs of last resort; and that several drugs long listed as forbidden alongside them, methylphenidate and some tricyclic antidepressants among them, were proscribed on an incorrect reading of the pharmacology ⁷¹. That is expert consensus, which is the weakest form of evidence in a hierarchy and the strongest available where the trials in the relevant patients have not been done, and several of its endorsers have pharmaceutical affiliations, which the author list discloses.

How well do they work for depression? Better than their reputation. A 2024 network meta-analysis of 83 double-blind randomized trials, with 7765 participants on active treatment and 1844 on placebo, found isocarboxazid, phenelzine, tranylcypromine and moclobemide all significantly more effective than placebo, with tolerability and acceptability comparable to other antidepressants; the authors note the gap in the literature is precisely where the drugs are now used, in treatment-resistant and atypical depression ⁷³. A 2020 meta-analysis restricted to tranylcypromine against tricyclic antidepressants pooled eight prospective parallel trials and found a log odds ratio of 0.480 (95 percent confidence interval 0.105 to 0.857, p equals 0.01) favoring tranylcypromine, with no important heterogeneity, but concluded that the two are equal in a mixed sample and that the advantage came from two studies in depression with prominent psychomotor retardation ⁴². That paper's first author was employed by a pharmaceutical company that markets tranylcypromine, which is disclosed in the affiliation and should be weighed.

The skin patch is the cleverest idea in the class. Delivering selegiline through the skin inhibits MAO-A and MAO-B in the brain while largely bypassing the intestinal and hepatic MAO-A that guards against food amines. In the pivotal trial, 289 randomized outpatients took the patch or placebo for up to eight weeks with no tyramine restriction at all; the patch beat placebo on the Montgomery-Asberg scale and the 28-item Hamilton scale but not the 17-item Hamilton scale, blood pressure did not differ between groups, and the commonest problem was a rash where the patch sat, in 31.5 percent against 15.1 percent ³⁵. The authors describe the antidepressant benefit as modest and statistically significant, which is the honest description.

On the Parkinson's side the picture is settled and unglamorous. The 2004 meta-analysis of 17 trials and 3525 patients found that MAO-B inhibitors reduce disability, cut the need for levodopa by nearly half as an odds ratio (0.57, 0.48 to 0.67) and reduce motor fluctuations (0.75, 0.59 to 0.95), with no excess mortality and no substantial excess of side effects, while noting that few trials compared them with other antiparkinsonian drugs ⁵². The largest practical comparison came in 2014. PD MED randomized 1620 newly diagnosed patients, open label, between levodopa and levodopa-sparing therapy, the latter split between dopamine agonists in 632 patients and MAO-B inhibitors in 460. Over a median of three years of follow-up, patient-rated mobility on the 39-item Parkinson's Disease Questionnaire was 1.8 points better with levodopa than with levodopa-sparing therapy (95 percent confidence interval 0.5 to 3.0, p equals 0.005) on a scale where six points is the minimally important difference, and that difference neither grew nor faded across seven years of observation; MAO-B inhibitors were 1.4 points better than dopamine agonists (0.0 to 2.9, p equals 0.05), and 23 percent of patients stopped them for side effects against 28 percent for dopamine agonists and 2 percent for levodopa ⁶³. The honest summary is that they are a reasonable levodopa-sparing option, slightly better tolerated than a dopamine agonist and slightly worse on mobility than levodopa, and that all these differences are smaller than the smallest difference a patient can notice. They are useful symptomatic drugs. Whether any of them changes the course of the disease is the subject of the section below, and the answer, after thirty-seven years, is still no with an asterisk.

The cheese reaction, and the wall it reveals

Tyramine is made when bacteria decarboxylate the amino acid tyrosine, which is why it accumulates in cheese as it ages, in fermented soy, in cured meat and in yeast extracts. Swallowed, it is an indirectly acting sympathomimetic: it enters nerve endings and displaces stored norepinephrine, which then floods out onto blood vessels. In an ordinary person this never happens, because monoamine oxidase in the wall of the intestine and in the liver destroys almost all of it before it reaches the circulation. The cheese reaction is what that wall looks like when it is taken away: a sudden, sometimes catastrophic rise in blood pressure, with a throbbing headache, and in the worst cases cerebral or subarachnoid hemorrhage ^18,21.

The sequence in 1963 is worth getting right, because it is a good example of how a hazard gets found. Two short letters appeared in The Lancet three weeks apart: the first, on 5 October, titled Tranylcypromine and cheese and reporting chromatographic work ¹⁹, the second, on 26 October, reporting hypertensive crisis due to monoamine-oxidase inhibitors and indexing cheese, phenelzine and tranylcypromine among its terms ¹⁸. Neither carries an abstract in PubMed. The popular account is in the literature after all. A 2009 history in the Indian Journal of Psychiatry records that the tyramine connection was first noticed by a British pharmacist whose wife, taking a monoamine oxidase inhibitor, developed a severe headache every time they ate cheese, and it sets out what Blackwell then documented: twelve patients, ten of them women, eleven on tranylcypromine and one on phenelzine, with blood pressures rising from 160/90 to 220/115 mmHg, onset one to two hours after the meal, headache as the main symptom, and complications that included subarachnoid hemorrhage, hemiplegia, intracranial hemorrhage, cardiac arrhythmia, cardiac failure, pulmonary edema and death ⁴⁷. That is a secondary historical account rather than the primary letters, and it is cited as such; the same note records Blackwell's 1965 measurement of 72 to 953 micrograms of tyramine per gram across fourteen cheddar cheeses ⁴⁷. What is documented in the primary record is the follow-through: controlled animal experiments on cheese and its constituents in 1966 ²⁰ and a full clinical series on hypertensive interactions between these drugs and foodstuffs in 1967 ²¹.

Then the diet itself was tested, and most of it turned out to be folklore. High-pressure liquid chromatography of actual foods, with 6 milligrams of tyramine per serving taken as the safety threshold, found no significant tyramine in commercial chain pizzas, including double pepperoni with double cheese, but found clinically significant amounts in tofu stored for a week and in one brand of soy sauce, whose authors concluded that pizzas from large chain outlets are safe with these drugs but that all soybean products should be avoided, especially soy sauce and tofu ⁸⁰. The narrowed modern diet sheet itself, usually given as aged and matured cheese, aged or spoiled meat and fish, fermented soy, tap or home-brewed beer and yeast extracts, is a guideline summary rather than a finding of that paper. Fresh food is still fresh food.

The quantitative difference between the drug generations is the clearest single measurement in this whole subject. Twelve healthy men took part. The mean oral dose of tyramine needed to raise systolic blood pressure by 30 mmHg was 15 mg in the four men challenged on day 28 of phenelzine, and 240 mg after 28 days of moclobemide, and three days after stopping moclobemide no one responded even to hourly doses escalating to 320 mg ⁴⁰. But the margin depends on leaving one enzyme alone: when a reversible MAO-A inhibitor and a reversible MAO-B inhibitor were given together, the pressor effect of oral tyramine was potentiated 13.5-fold plus or minus 6.9, and 51 plus or minus 20 mg was enough to raise pressure by 30 mmHg, which the investigators, who worked for the manufacturer, said means dietary precautions are needed when both forms are blocked ⁷⁸. The same logic explains the selegiline patch, which inhibits the brain enzymes while largely sparing the ones in the gut and liver, and was therefore tested without any dietary restriction at all ³⁵.

Brunner syndrome, and the gene the newspapers renamed

In 1993 a Dutch group described a large family in which several men had borderline intellectual disability and a characteristic pattern of abnormal behavior: impulsive aggression, and in individual cases arson, attempted rape and exhibitionism. Linkage analysis placed the responsible locus on the short arm of the X chromosome with a maximal multipoint lod score of 3.69 at the MAOA locus; a 24-hour urine collection in three affected men showed markedly disturbed monoamine metabolism, and platelet MAO-B activity was normal, which pointed at MAO-A alone ²⁷. Later the same year, in Science, the mutation was found: a single base change in the eighth exon of MAOA turning a glutamine codon into a stop codon, present in each of five affected males, abolishing the enzyme completely and selectively ²⁸.

Two things about that finding are routinely lost. The first is how rare it is: complete MAO-A deficiency has been described in a handful of families worldwide, and it is a loss-of-function mutation in a gene, not a common variant. The second is that it is a metabolic disease with a treatment. Two Australian families reported in 2016, one with a protein-truncating variant and one with a missense change, had mild to borderline intellectual disability, limited friendships, episodic explosive aggression and, in one family, episodes of flushing, headaches and diarrhea; their biochemistry showed high serum serotonin and high urinary metanephrines with low urinary 5-HIAA and VMA. In the second family, a serotonin reuptake inhibitor plus dietary change reversed the biochemical abnormalities, reduced the serotonergic symptoms and improved behavior ²³. The authors recommend screening for the syndrome with serum and urine metanephrine and serotonin measurements in a man with mild intellectual disability and paroxysmal behavioral symptoms.

The picture has widened since. In 2023 an Italian group described a 5-year-old boy with developmental delay, autistic features and myoclonic epilepsy together with his mother, who had mild intellectual disability and recurrent episodes of palpitations, headache, abdominal pain and abdominal bloating; both carried the maternally inherited MAOA variant c.410A>G, p.Glu137Gly, biochemical studies confirmed MAO-A deficiency in both, and in the mother a serotonin reuptake inhibitor with dietary change reversed the biochemical abnormalities and partly reduced the symptoms ⁸¹. A symptomatic carrier mother widens the sex assumptions around an X-linked condition. In 2026 a group in Parma reported a novel splice-site variant that first-line testing had missed, showed from the patient's own messenger RNA that it causes retention of intron 8 and a premature stop codon, and treated him with a drug of the serotonin antagonist and reuptake inhibitor class, with reported improvement in social behavior and sleep; they state that this is the first such use, that it is off-label and that the evidence is preliminary ⁸². Both are single case reports with no comparison group. Together they establish that the condition is treatable in principle, and nothing about how often treatment helps.

The animal work ran alongside. Mice in which a transgene insertion had deleted the MAOA gene had serotonin concentrations up to ninefold higher in pup brains, norepinephrine up to twofold higher, cytoarchitectural changes in the somatosensory cortex, and enhanced aggression in adult males; the pups' trembling and fearfulness were reversed by an inhibitor of serotonin synthesis ³². The developmental half was then nailed down: those mice never form the barrels, the clusters of layer four neurons that map the whiskers, and an inhibitor of serotonin synthesis given early restored them while blocking catecholamine synthesis did not, so the cause is an excess of serotonin during a critical period rather than a lasting lack of enzyme ³³. Transiently inactivating MAOA in normal newborn mice reproduced the barrelless pattern ³³. The lesson is that the enzyme's absence shapes the brain while it is being built, not only while it is running.

Then came the common variant, and the trouble. In 1998 a 30-base-pair repeat 1.2 kilobases upstream of MAOA was described in 3, 3.5, 4 or 5 copies, with the 3.5 and 4 copy versions driving transcription 2 to 10 times more efficiently in cell experiments ⁶⁰. In 2002 Caspi and colleagues reported, in a male birth cohort followed from birth to adulthood, that this variant moderated the effect of childhood maltreatment: maltreated boys with the high-activity genotype were less likely to develop antisocial problems ⁵³. That is a statement about an interaction. It was reported in the press as a warrior gene, which is a statement about a person, and the two are not the same claim.

The replication record deserves to be stated exactly, because it is neither a vindication nor a debunking. In 2006 the original group added 975 seven-year-old boys, replicated the interaction, extended it to childhood closer in time to the maltreatment, and ran a meta-analysis of the published studies showing the association between maltreatment and mental health problems was significantly stronger in males with the low-activity genotype ⁵⁴. In 2014 an independent meta-analysis of 27 peer-reviewed studies through August 2012 found the interaction held across 20 male cohorts (p equals 0.0044), was specific to maltreatment rather than general family hardship (p less than 0.000001), and was robust to sensitivity analyses; across 11 female cohorts there was no interaction with combined adversity, and maltreatment alone predicted antisocial behavior weakly in women with the high-activity genotype, the opposite direction ⁵⁵. The same year a second meta-analysis found the variant significantly associated with antisocial behavior overall, but with substantial heterogeneity across studies that no sample-level or study-level characteristic explained, and summarized the whole area as supporting a modest role for common serotonergic variants ⁵⁶.

Then the method that produced the claim was superseded. Candidate-gene work tests one variant chosen in advance; a genome-wide study tests millions without choosing. The Broad Antisocial Behavior Consortium meta-analyzed 28 discovery samples totaling 85,359 people with five independent replication samples totaling 8058, and the first genome-wide significant association with broad antisocial behavior was in common intronic variants of FOXP2, a gene best known for speech and language, with a lead variant rs12536335; the trait was substantially genetically correlated with depression at 0.63, smoking at 0.54 and insomnia at 0.47, and inversely with intelligence at minus 0.40 and years of schooling at minus 0.46 ⁶⁸. The consortium's earlier paper, which pooled 16,400 discovery and 9381 target participants and set out expressly to reevaluate the candidate-gene era data, found antisocial behavior to be highly polygenic, with three promising loci whose associations differed between the sexes and an inverse genetic correlation with educational attainment of minus 0.52 ⁶⁹. MAOA is not among the loci these studies surface. That does not retract the gene-by-maltreatment interaction, which was never a main-effect claim, but it does mean that no common variant of MAOA emerges as an important determinant of antisocial behavior once the genome is searched without a prior favorite. The genetic correlations are associations between inherited liabilities, not causes.

So: a real, replicated, male-specific interaction with documented maltreatment, of modest size, with heterogeneity nobody has explained and no established effect in women. That is a long way from a gene for violence, and a long way from nothing. The honest sentence is that the variant tells you something about how a population of maltreated boys will distribute, and nothing useful about any individual man. The misreading has had consequences outside the newspapers. The warrior gene label itself was pressed into service in a 2007 editorial in the New Zealand Medical Journal, about the frequency of the low-activity allele in an ethnic population and its supposed bearing on risk-taking, violence and addiction ⁸³; critics answered in the same issue of the same journal that while there is credible evidence for a contribution of this variant to antisocial behavior in people of European descent, there is no direct evidence for such an association in Māori, and that insufficient rigor in interpreting the literature and in generating new data had done both science and Māori a disservice ⁸⁴. And in 2009 an Italian appeal court reduced a convicted murderer's sentence from nine years to eight on the grounds that he carried genetic variants, MAOA among them, said to predispose to aggressiveness; a group of European clinical geneticists replied that the clinical utility of such testing has not been shown and that offering susceptibility testing inside legal proceedings carries risks of its own ⁸⁵. An interaction of modest size, measured across populations of maltreated boys, cannot tell a court anything about one man.

Platelet MAO-B: a marker that measured the wrong thing

Because human platelets carry MAO-B and no MAO-A, a drop of blood gives a number, and in the 1970s that number was asked to carry an enormous amount of psychological weight. The founding study screened 375 college student volunteers, interviewed the highest and lowest tenth, and collected family histories. Students in the low-MAO group reported more psychiatric or psychological counseling and more problems with the law, and families of the low-MAO probands had an eightfold higher rate of suicide or suicide attempts than families of the high-MAO ones; the authors proposed that reduced MAO activity, which had already been reported in affective disorders and chronic schizophrenia, might predict a vulnerability to psychiatric disorder ²². For the next two decades low platelet MAO was linked in print to sensation seeking, impulsivity, and type 2 alcohol dependence.

The claim collapsed from underneath, and the agent of its collapse is in the section after this one. Smoking lowers MAO-B. People with alcohol dependence, psychiatric illness and high sensation seeking smoke more than the comparison groups they were measured against. A 1997 review walked through every candidate confounder in turn, including duration of abstinence, the substrate used in the assay, sex, subtype, other psychopathology and smoking, and concluded that the initial promise of MAO activity as a marker for alcoholism, or at least for a genetically predisposed subgroup, had receded as the negative typology-based studies accumulated, and that the failure to account for smoking in almost all of these studies left the whole hypothesis open to criticism ³⁸.

The confounding was then demonstrated directly in alcohol-dependent patients, although in the type 1 subgroup rather than the type 2 subgroup the trait claim was about. Among type 1 alcohol-dependent subjects, those who smoked cigarettes had significantly lower platelet MAO-B activity, 4.0 plus or minus 1.5 microkatal per kilogram of protein, than those using no tobacco at 5.0 plus or minus 1.9, while those using smokeless tobacco were no different from non-users at 5.1 plus or minus 1.5 ⁸⁶. If the association had been about alcohol or about personality, the form in which the tobacco arrived should not have mattered. It did.

What is left is narrow and real. Platelet MAO-B tracks recent drinking as a state marker rather than a trait: in 98 alcohol-dependent patients it was 6.4 plus or minus 3.1 nanomoles per milligram of protein per hour while they were still under the influence of ethanol, rose by more than 40 percent to 9.3 plus or minus 4.4 after eight days of withdrawal, and then stayed flat at 15 and 22 days, against 9.9 plus or minus 0.9 in 138 healthy people; in that study neither sex nor smoking affected the time course during early withdrawal ⁸⁷. The people who built the claim did not accept that it was finished, and their answer belongs here. Reviewing the same literature, Oreland argued that an association between platelet MAO activity and sensation seeking and impulsiveness does survive once the smoking confounder is handled, that the link to type 2 alcohol dependence has support including in non-human primates, and that the mechanism may not be the one everyone assumed: shared transcription factors of the AP-2 family may regulate both platelet MAO and components of central monoaminergic systems, so that platelet MAO indexes mitochondrial number or outer membrane surface rather than anything about brain MAO ⁸⁸. A later review with Harro reached a similar reading, adding that the findings differ markedly between men and women and that the cross-sectional design of nearly all this work is the likely reason no mechanistic model has been assembled ⁸⁹. Both are narrative reviews by the field's principal advocates, which is how they should be weighed. A measurement that moves with the week is not a measurement of character, and a disputed residue is not the same as a vindicated marker.

• Platelet MAO-B activityHuman blood platelets carry MAO-B only, so a blood sample gives an easily measured number that was treated for two decades as a window on the brain ^22,38.A research measure, not a clinical test; historically used as a proposed marker of personality and of alcohol dependenceReproducible as a measurement: in one Italian series, mean activity was 6.4 plus or minus 3.1 nanomoles per milligram of protein per hour in 98 alcohol-dependent patients still under the influence of ethanol, rising to 9.3 plus or minus 4.4 after eight days of abstinence, against 9.9 plus or minus 0.9 in 138 healthy people ⁸⁷.It is not a window on the brain, and it is heavily confounded. Cigarette smoking lowers it; among type 1 alcohol-dependent subjects, cigarette smokers sat at 4.0 plus or minus 1.5 microkatal per kilogram of protein against 5.0 plus or minus 1.9 for non-tobacco users, while users of smokeless tobacco were unaffected at 5.1 plus or minus 1.5 ⁸⁶. A 1997 review concluded that the failure to account for smoking leaves the whole low-MAO-and-alcoholism hypothesis open to criticism ³⁸.
• Urinary and serum amine metabolitesWhen MAO-A is absent, the products it normally makes disappear from urine and the substrates pile up: high serum serotonin and urinary metanephrines with low 5-hydroxyindoleacetic acid (5-HIAA, the serotonin breakdown product) and low vanillylmandelic acid (VMA, the norepinephrine breakdown product) ²³.Screening for Brunner syndrome, that is complete MAO-A deficiency, in a boy or man with mild intellectual disability and episodic behavioral symptoms, with flushing, headache and diarrhea in some familiesIn the original Dutch kindred a 24-hour urine collection in three affected males showed markedly disturbed monoamine metabolism, and platelet MAO-B was normal, which is what localized the defect to MAO-A ²⁷.It is a test for a rare, complete loss of one enzyme. It says nothing about the common promoter variant, and nothing about how much MAO-A is in anyone's brain.
• PET with carbon-11 harmine (MAO-A)A positron-emitting radioligand, that is a radioactive tracer that sticks to the enzyme and can be counted through the skull, selective for MAO-A ²⁴.Research imaging of MAO-A density in depression, the postpartum period and smokingSensitive enough to find a 34 percent difference between 17 depressed and 17 healthy participants in every brain region examined, with p values from 0.001 to 3 times 10 to the minus 7 ²⁴, and a 43 percent difference in the early postpartum period ²⁹.Small samples, no routine clinical use, and it measures how much enzyme is present rather than how much amine is actually being destroyed.
• PET with carbon-11 deprenyl or fluorine-18 SMBT-1 (MAO-B)MAO-B tracers; the newer fluorine-18 compound was designed as a marker of reactive astrocytes, the support cells that swell and multiply around brain damage ²⁶.Research imaging of smoking, aging and astrogliosis in Alzheimer's diseaseMore than 85 percent of the signal across the brain was blocked by five days of selegiline, confirming the target, and the tracer picked up the known rise in MAO-B with age of about 2.6 percent a year against about 1.9 percent a year measured in tissue ²⁶; in 14 older adults the modeled binding correlated with postmortem MAO-B indices at about r equals 0.8 ⁹⁸. Autopsy work in 2026 showed the binding sits on MAO-B-positive astrocytes clustered around amyloid deposits, more closely than around tau ³⁷.It cannot separate Alzheimer's disease from other tauopathies cleanly, since values in non-Alzheimer tauopathies partly overlapped both controls and Alzheimer's cases ³⁷.
• The tyramine pressor testNot a test of the enzyme but of the consequence: how much tyramine by mouth it takes to raise systolic blood pressure by 30 mmHg ⁴⁰.Establishing how dangerous a given MAO inhibitor makes a given dietQuantitative and reproducible: 15 mg in the four men challenged on day 28 of phenelzine, 240 mg on day 28 of moclobemide, and no response at all up to hourly doses reaching 320 mg three days after stopping moclobemide ⁴⁰; combining a reversible MAO-A and a reversible MAO-B inhibitor still potentiated oral tyramine 13.5-fold plus or minus 6.9 ⁷⁸.It is done in healthy volunteers under supervision with pure tyramine, not in patients eating aged cheese, and the tyramine content of real food varies enormously ⁸⁰.

Smoking, the enzyme, and the Parkinson's puzzle

In 1996 a group at Brookhaven National Laboratory scanned living people and found that the brains of smokers held about 40 percent less MAO-B than the brains of non-smokers or former smokers. The paper is short and its reasoning is explicit: MAO-B breaks down dopamine, so less of it means more dopamine activity and less hydrogen peroxide, and the authors proposed that this reduction may act alongside nicotine to produce the behavioral and epidemiological effects of smoking, including the high rate of smoking in psychiatric illness and the lower rate of Parkinson's disease in smokers ⁵⁸. Later the same year they reported the other enzyme: MAO-A was about 28 percent lower in the brains of 16 smokers than 15 non-smokers, which was roughly half the reduction produced in volunteers by three days of tranylcypromine at 10 mg a day ⁵⁹.

It is not only the brain. In 2003 the same group scanned peripheral organs and found MAO-B significantly reduced in smokers in the heart, lungs and kidneys, by 33 to 46 percent ²⁵. Tobacco smoke is a systemic monoamine oxidase inhibitor, and the organs that handle circulating amines are inhibited along with the brain.

What in the smoke does it? Not nicotine. Working with mainstream smoke collected from commercial cigarettes, investigators showed the smoke itself inhibits both human enzymes, partly reversibly, and identified and quantified two beta-carboline alkaloids in it: norharman, which inhibited MAO-A with an inhibition constant of 1.2 plus or minus 0.18 micromolar and MAO-B at 1.12 plus or minus 0.19 micromolar, and harman, which inhibited MAO-A at 55.54 plus or minus 5.3 nanomolar ⁹⁰. That is laboratory enzymology with isolated enzyme, and it identifies plausible culprits rather than proving that these particular molecules account for the imaging findings in people.

Now the epidemiology, which is where discipline is needed. Smokers really do get Parkinson's disease less often. A meta-analysis of 44 case-control and 4 cohort studies put the relative risk at 0.59 (95 percent confidence interval 0.54 to 0.63) for ever smokers against never smokers, 0.39 (0.32 to 0.47) for current smokers and 0.80 (0.69 to 0.93) for past smokers, with a dose relationship of 0.84 per additional 10 pack-years in case-control studies and 0.78 in cohort studies ⁹¹. The association is large, graded and consistent, and it has an obvious candidate mechanism sitting right there in the PET scans.

But the association is not the same as the mechanism, and it is not even securely causal in the direction people assume. A long-standing alternative explanation is reverse causation: the earliest changes of Parkinson's disease, which begin years before movement symptoms, might make people less likely to take up or continue smoking. Genetic instrumental-variable studies, which use inherited liability to smoking as a stand-in for smoking itself and so are less vulnerable to that problem, have supported a protective direction; a systematic review and meta-analysis of 29 published Mendelian randomization studies plus 123 new analyses across 57 outcomes found genetic liability to smoking associated with increased risk of a long list of diseases and with decreased risk of only two, Parkinson's disease and prostate cancer ⁹², and a 2025 analysis using the same approach reported an odds ratio of 0.288 for Parkinson's disease, p equals 0.008 ⁹³. Even taking that at face value, it tells you smoking is protective, not that MAO-B inhibition is why.

And the test that would settle it has, in effect, been run. If brain MAO-B inhibition were enough to protect dopamine neurons, then selegiline and rasagiline should alter the course of Parkinson's disease. The next section is the story of that question, and the answer is that after thirty-seven years the disease-modification claim still has not been established ^43,50,52. Two further tests have been run, and both cut against the simple story. The reverse-causation worry was given data: among 1808 Danish patients with Parkinson's disease and 1876 matched controls, former smokers who had found quitting harder, or who had needed nicotine substitutes, were less likely to develop the disease than those who found quitting easy, which is what you would expect if reduced responsiveness to nicotine is itself an early feature of the illness rather than smoking being protective; the authors propose that ease of quitting belongs alongside loss of smell, dream-enactment sleep disorder and constipation as a premanifest sign ⁹⁴. The reports of how hard quitting had been were collected after diagnosis, which is that study's main weakness. And the direct experiment failed: in a double-blind placebo-controlled multicenter trial, 163 patients diagnosed within 18 months and not yet needing dopaminergic treatment were randomized to a transdermal nicotine patch or placebo for 52 weeks with 8 weeks of washout, and total scores on the Unified Parkinson's Disease Rating Scale worsened by a mean of 3.5 points on placebo against 6.0 on nicotine, a Hodges-Lehmann difference of minus 3 (95 percent confidence interval minus 6 to 0, p equals 0.06); at 52 weeks the figures were 5.4 against 9.1 points, a difference of minus 4 (minus 7 to minus 1). One year of nicotine did not slow early Parkinson's disease, and if anything the treated group did slightly worse ⁷². The smoking association remains real, the MAO-B reduction in smokers remains real, and the bridge between them remains unbuilt.

Thirty-seven years of asking whether blocking MAO-B slows Parkinson's disease

The question was reasonable and the logic was tight. MAO-B generates hydrogen peroxide when it breaks down dopamine; dopamine neurons die in Parkinson's disease; MAO-B turns MPTP into the ion that kills dopamine neurons, and mice lacking MAO-B are resistant to MPTP ^8,41. If oxidative damage drives the disease, blocking the enzyme should slow it.

DATATOP was designed to find out. Eight hundred patients with early, untreated Parkinson's disease were randomized in a two-by-two factorial design to deprenyl, tocopherol, both or placebo, and the endpoint was the point at which disability forced the clinician to start levodopa. At the interim analysis in 1989, 97 patients on deprenyl had reached that endpoint against 176 not on it, over an average of 12 months, a 57 percent reduction in risk, hazard ratio 0.43 (0.33 to 0.55), with p below 10 to the minus 10 ⁴⁸. The result was read almost everywhere as neuroprotection.

The full report in 1993 is where the reading began to come apart, and it is the authors themselves who did the taking apart. Tocopherol did nothing at all, alone or in combination. Deprenyl still delayed the endpoint, hazard ratio 0.50 (0.41 to 0.62), by about nine months at the median. But the Parkinson's disease ratings improved during the first three months of treatment, and the motor performance of deprenyl-treated patients worsened after the treatment was withdrawn, and the paper's own conclusion reads: the action of deprenyl that accounts for its beneficial effects remains unclear ⁴⁹. A drug with a symptomatic benefit will delay the moment a clinician reaches for levodopa whether or not it changes a single neuron's fate.

In 1994 Ward made the argument formally, modeling the trial on the simple assumption that measured impairment is proportional to the degree of cellular dysfunction, and concluding that the reduced probability of reaching the endpoint was due to a direct treatment effect rather than to neuroprotection ⁵⁰. That is the correction, published in a neurology journal, five years after the headline.

Then the pendulum swung the other way, hard. A United Kingdom trial randomized 520 patients at 93 hospitals to levodopa or levodopa plus selegiline and, after an average of five to six years, reported a mortality ratio of 1.57 (1.09 to 2.30) against levodopa alone, significant on the log rank test at p equals 0.015, with no clinical benefit, more disabling peak-dose dyskinesia and more on-off fluctuation ⁵¹. Prescribing collapsed. But a 2004 meta-analysis of 17 randomized trials covering 3525 patients found no significant difference in mortality (odds ratio 1.13, 0.94 to 1.34, p equals 0.2), alongside better rating-scale scores at three months, substantially less need for additional levodopa (0.57, 0.48 to 0.67) and fewer motor fluctuations (0.75, 0.59 to 0.95), with no excess of side effects or withdrawals ⁵². One trial's alarming result did not survive the company of the others. Both reversals belong in the record.

The disease-modification question was then re-asked with a cleverer design. ADAGIO randomized 1176 untreated patients to rasagiline for 72 weeks, at 1 or 2 mg a day, or to placebo for 36 weeks followed by the same dose, and required the early-start group to clear three hierarchical endpoints. At 1 mg daily it cleared all three: a slower rate of worsening between weeks 12 and 36 (0.09 plus or minus 0.02 against 0.14 plus or minus 0.01 UPDRS points per week, p equals 0.01), less total worsening at 72 weeks (2.82 plus or minus 0.53 against 4.52 plus or minus 0.56 points, p equals 0.02) and non-inferiority of the rate of change between weeks 48 and 72. At 2 mg daily it cleared none, with 3.47 against 3.11 points at week 72, p equals 0.60. The authors' own conclusion is the one to quote: because the two doses were associated with different outcomes, the study results must be interpreted with caution ⁴³. A dose-response that runs backwards is not what a disease-modifying effect looks like. An earlier cohort followed for 6.5 years had reported a persistent 2.5-point advantage for early starters, which the authors said might reflect neuroprotection or might reflect effects on compensatory mechanisms ⁷⁹.

Today MAO-B inhibitors are prescribed for what they reliably do, which is to make people with Parkinson's disease better while they take them. Safinamide, described in the drug-class reviews as a reversible MAO-B inhibitor ⁴⁴ and one that also blocks voltage-dependent sodium and calcium channels and the calcium-mediated release of glutamate ², added 1.42 hours a day of good on time against 0.57 for placebo in 549 patients over 24 weeks, a difference of 0.96 hours (0.56 to 1.37), p below 0.001, at the cost of dyskinesia in 14.6 percent against 5.5 percent ⁶⁴. Reviewers of that drug note frankly that its mechanism has not been fully defined, which means its benefit cannot be credited to the enzyme alone ⁷⁷. A 2026 systematic review pooled three randomized trials of safinamide 100 mg daily added to levodopa, 1127 participants in all, with its protocol registered in advance: daily off time fell by 1.21 hours against placebo (minus 1.60 to minus 0.82, moderate certainty), motor scores on part III of the rating scale improved by 3.13 points (minus 4.89 to minus 1.37, rated low certainty because the trials disagreed with each other, with an I-squared of 72 percent), and the risk of dyskinesia tripled, relative risk 3.01 (1.92 to 4.70, moderate certainty) ⁷⁶. Three trials is a thin evidence base, and the authors say so, but the size of both the benefit and the harm is now pooled rather than taken from one study.

The same question was asked in the other disease, and answered more cleanly. A Cochrane review pooled 17 double-blind randomized trials of selegiline in Alzheimer's disease. It found a few scattered effects on combined cognitive tests at 4 to 6 weeks (standardized mean difference 0.39, 95 percent confidence interval 0.07 to 0.72) and at 8 to 17 weeks (0.44, 0.04 to 0.84) and on activities of daily living at 4 to 6 weeks, no effect on emotional state or on global rating, no excess of adverse effects, and it concluded that despite its initial promise selegiline for Alzheimer's disease had proved disappointing, with no evidence of clinically meaningful benefit on any outcome measure and no justification for using it or for studying it further in that disease ⁶¹. Two decades later a purpose-built drug was tried. Sembragiline, a potent, selective, long-acting reversible MAO-B inhibitor, was given to 542 patients with moderate Alzheimer's disease, Mini-Mental State Examination scores of 13 to 20, at 1 or 5 mg daily against placebo for 52 weeks on top of standard treatment. It was well tolerated and it missed its primary endpoint: the difference from placebo on the 11-item Alzheimer's Disease Assessment Scale cognitive subscale was minus 0.15 points (p equals 0.865) at 1 mg and 0.90 points (p equals 0.312) at 5 mg. The secondary endpoints were missed too, the behavioral differences that appeared were present at week 52 only, the subgroup finding was post hoc, and most of the authors were employed by the company that developed the drug ⁶⁵. Two diseases, four decades, and no demonstration that blocking this enzyme changes either course.

Too much MAO-A: depression, the postpartum weeks, and what the imaging shows

The monoamine hypothesis of depression said that monoamines are low. It never said why. In 2006, a Toronto group gave the most direct answer anyone has produced: they imaged MAO-A density with carbon-11 harmine in 17 people in an untreated major depressive episode, medication-free for at least five months, and 17 healthy people, all of them non-smokers because smoking lowers the enzyme. MAO-A density was significantly elevated in every one of the ten brain regions examined, with p values running from 0.001 to 3 times 10 to the minus 7, and the average elevation across the brain was 34 percent, about two standard deviations ²⁴. The authors concluded that elevated MAO-A density is the primary monoamine-lowering process in major depression. That conclusion is an interpretation; the 34 percent is a measurement.

Four years later the same group looked at the one time in life when mood changes on a known biological schedule. In 15 healthy women 4 to 6 days after delivery and 15 who had not recently given birth, all non-smoking and medication-free, MAO-A binding was elevated by a mean of 43 percent throughout every region analyzed ²⁹. The reasoning is mechanistic and testable: estrogen falls 100- to 1000-fold in the first three to four days after birth, estrogen levels move inversely with MAO-A density, postpartum blues is common, and major depression follows delivery in about 13 percent of women. The proposal is a biological chain rather than a purely psychosocial one, and the elevated enzyme is the measured link in it. Fifteen women per group is a small study and it has not been followed by a large one.

Imaging the other enzyme in depression has produced a 2026 result that is interesting and must be handled carefully, because it is a preprint and has not been peer reviewed. Using a MAO-B tracer in 95 adults, the same center reported that MAO-B distribution volume was about 10 percent higher in gray matter in women than in men, and about 23 percent higher in the prefrontal cortex in people with treatment-resistant depressive episodes than in healthy controls and 6 percent higher than in non-treatment-resistant episodes; in a subgroup scanned before and after treatment, rasagiline and phenelzine each occupied about 94 percent of the enzyme at clinical doses, while duloxetine had minimal effect ⁷⁵. If it survives review, it suggests the enzyme worth blocking may differ between patients, which is not how the drugs are currently chosen.

What this body of work does not show is that MAO-A elevation causes depression, but it is no longer all cross-sectional. The same Toronto group scanned 16 people during a depressive episode and again after six weeks of a selective serotonin reuptake inhibitor, 18 people with major depressive disorder in recovery, and 28 healthy people. MAO-A binding was significantly elevated in every region during the episode and stayed elevated after the drug, which means the antidepressant does not normalize the enzyme; it was still elevated in recovery; and among those in recovery, the people who went on to a new episode within six months had significantly higher MAO-A in the prefrontal and anterior cingulate cortex than those who did not ³⁰. The authors read that as a monoamine-lowering process which outlasts the episode, and as a reason to continue treatment during early remission. It is still a case-control design with small groups, from the laboratory that produced the original finding, and an enzyme that rose in response to low mood and stayed up would look much the same; nobody has scanned a healthy cohort forward into a first illness. The clinical fact that sits alongside it, and fits it, is that blocking the enzyme treats the illness: a 2024 network meta-analysis of 83 double-blind randomized trials found isocarboxazid, phenelzine, tranylcypromine and moclobemide all significantly better than placebo, with tolerability and acceptability comparable to other antidepressants, and specifically noted the shortage of trials in the treatment-resistant and atypical depression where these drugs are actually reached for ⁷³.

The peroxide problem: heart, vessels and the reactive astrocyte

Everything above treats monoamine oxidase as a disposal system. The rest of the body has to treat it as a source of hydrogen peroxide, and the evidence for that runs from mouse to man in a way worth separating carefully.

The animal evidence is strong and mechanistic. In mice, MAO-A activity in heart muscle cells was triggered by norepinephrine, followed by enlargement of the cells, reactive oxygen species and the molecular signature of maladaptive hypertrophy, partly independently of the usual adrenergic receptors and blocked by the specific MAO-A inhibitor clorgyline. In mice with left ventricular dilation and pump failure from pressure overload, norepinephrine breakdown by MAO-A was increased with worse oxidative stress; inhibiting MAO-A prevented that, reversed the fetal gene reprogramming and reduced myocardial cell death, and mice expressing a dominant-negative MAO-A were protected relative to their wild-type littermates ⁹⁵. That is as complete an animal causal chain as cardiology produces. It is still mice.

The human evidence is observational, and it is about tissue rather than treatment. In right atrial appendage taken from 244 patients during cardiac surgery, monoamine oxidase and NADPH oxidase, an enzyme complex whose job is to make reactive oxygen species on purpose, together generated them at roughly ten times the rate of intact coupled mitochondria, and the risk of postoperative atrial fibrillation rose across quartiles of MAO activity, from an adjusted relative risk of 1.0 in the lowest quartile to 3.8 (1.9 to 7.5) in the highest, with a trend p of 0.009; total glutathione, the cell's main antioxidant reserve, ran in the opposite direction ³¹. The authors are explicit that MAO needs validating as a predictive marker and that its role in generating the arrhythmia is not established. In atrial tissue from patients with and without type 2 diabetes, MAO-A and MAO-B content and maximal activity were both higher in diabetes and tracked body mass index; catecholamine levels in the tissue were lower, consistent with increased flux through the enzymes; protein adducts formed by the aldehydes that monoamine oxidase produces were more abundant and were MAO-dependent; and norepinephrine suppressed the mitochondrial production of ATP, the molecule cells spend as energy, in permeabilized muscle fibers from those patients in an MAO-dependent manner ⁹⁶. A 2026 study extended the observation to veins: in 20 patients having varicose veins stripped, both enzymes were expressed in the vein wall and higher in the varicose than the adjacent perforator veins, MAO-A was significantly higher in the obese patients with raised C-reactive protein, a blood protein that rises with inflammation, than in the non-obese (p less than 0.01), angiotensin II raised both isoforms and hydrogen peroxide production in tissue outside the body, and 10 micromolar clorgyline or selegiline reduced the oxidative stress ⁹⁷. All three studies are associations in human tissue, and none of them is a trial of an inhibitor as a cardiovascular treatment.

In the brain the same chemistry has a different and surprising consequence. MAO-B in reactive astrocytes, the support cells that enlarge and multiply around damage, was shown in 2014 to synthesize GABA, the brain's principal inhibitory transmitter, which the astrocytes then release through the bestrophin 1 channel; in the dentate gyrus of mouse models of Alzheimer's disease this released GABA reduced the firing probability of granule cells, and suppressing its production or release fully restored firing, synaptic plasticity and learning and memory in the mice, while astrocytic GABA and MAO-B were both significantly increased in postmortem brain tissue from people with Alzheimer's disease ³⁶. This is why a MAO-B tracer has become an astrocyte tracer: the 2022 first-in-human characterization of fluorine-18 SMBT-1 found more than 85 percent of its brain signal blocked by five days of selegiline, a distribution matching known MAO-B and an age-related rise of about 2.6 percent a year ²⁶, its kinetics were formally modeled in 2024 with binding correlating at about r equals 0.8 against postmortem MAO-B indices ⁹⁸, and in 2026 autopsy work showed the binding sitting on MAO-B-positive astrocytes clustered around amyloid plaques and vessels, more closely than around tau pathology, with non-Alzheimer tauopathies partly overlapping both controls and Alzheimer's cases ³⁷.

The therapeutic follow-through is just beginning, and it is a genuinely new development rather than a rediscovery. The old MAO-B inhibitors bond permanently, which in astrocytes produces a compensatory problem: prolonged selegiline failed to lower astrocytic GABA or to rescue memory in APP/PS1 mice, a mouse model of Alzheimer's disease, because activity rose in a compensating GABA-synthesizing enzyme, diamine oxidase. That is why a reversible compound, KDS2010, with a half-maximal inhibitory concentration of 7.6 nanomolar and 12,500-fold selectivity over MAO-A, was developed; long treatment with it induced no such compensation and did rescue learning and memory in the same mice ⁹⁹. All of that is mouse work. In 2025 a study that combined cross-sectional and longitudinal human imaging, postmortem tissue and mouse models implicated MAO-B-dependent astrocytic GABA in post-traumatic stress disorder, with elevated prefrontal GABA in patients that normalized with recovery, and reported that KDS2010 restored astrocytic GABA, cerebral blood flow and fear extinction retrieval in the mouse models, and had completed phase 1 trials with a favorable safety profile, with phase 2 planned ⁷⁴. Phase 1 means it was safe in volunteers. It does not mean it works.

The interactions that still kill people

Two drug interactions matter with these medicines. The first is the cheese reaction, dealt with above: an irreversible inhibitor plus dietary tyramine, or plus any indirectly acting sympathomimetic such as the decongestants in cold remedies, equals a hypertensive crisis ^18,21. The second is serotonin toxicity, which is the opposite problem: too much serotonin acting on its receptors, usually when a drug that blocks serotonin reuptake meets a drug that stops serotonin being destroyed ¹⁰⁰.

Serotonin toxicity is a clinical diagnosis made at the bedside, and it has a rule. The Hunter criteria were derived from 2222 consecutive admissions after overdose of a serotonergic drug over 15 years, with a learning set of 473 overdoses of a selective serotonin reuptake inhibitor alone, and they reduce to seven features: clonus, whether spontaneous, inducible or ocular, plus agitation, sweating, tremor and hyperreflexia, with muscle rigidity and a temperature above 38 degrees Celsius added because those were universal in the life-threatening cases. Against a clinical toxicologist's diagnosis, the criteria were 84 percent sensitive and 97 percent specific, better on both counts than the older Sternbach criteria at 75 and 96 percent ¹⁰¹. Clonus, a rhythmic beating of a muscle when it is stretched, is the finding that separates this from most other causes of agitation and fever.

The reason this belongs on a page about an enzyme is that two widely used drugs inhibit monoamine oxidase without anyone thinking of them as monoamine oxidase inhibitors. Linezolid is an antibiotic, reached for in resistant Gram-positive infection, and it inhibits both MAO-A and MAO-B. A review of post-marketing reports to the United States Food and Drug Administration identified 29 cases meeting a definition of serotonin toxicity in which linezolid was the primary suspect drug given with at least one other drug raising central serotonin, most often a selective serotonin reuptake inhibitor in 26 of 43 patients; thirteen patients required intervention to prevent permanent impairment or were hospitalized for the event ⁶². Spontaneous reports cannot give a rate, but they establish the hazard. Sixteen years later someone counted. In a population-based retrospective cohort of 1134 outpatients aged 66 or older in Ontario prescribed oral linezolid between October 2014 and January 2021, of whom 215, that is 19.0 percent, were also taking an antidepressant, clinically significant serotonin syndrome within 30 days occurred in fewer than 6 patients, under 0.5 percent; in a propensity-score-matched comparison the adjusted risk difference between those on antidepressants and those not was minus 1.2 percent (95 percent confidence interval minus 2.9 to 0.5), with similar rates of altered mental status, hospitalization and death ⁷⁰. The authors conclude that linezolid is likely safe in patients taking antidepressants, while asking prescribers to stay alert. Both things are true at once: the interaction is real and documented in the spontaneous reports, and it is rare enough that a cohort this size could not measure an increase, which bounds the risk rather than excluding it.

Methylene blue is the other, and the story of its recognition is a small triumph of prediction. It had been reported to precipitate serotonin toxicity in patients on serotonin reuptake inhibitors, and the prediction was that it must be inhibiting MAO-A. Kinetic experiments on purified human enzyme confirmed it: methylene blue is a potent, tight-binding and reversible inhibitor of MAO-A, with much weaker action on MAO-B, interacting with the MAO-A active site both as an oxidizing substrate and as a one-electron reductant; at the concentrations reported in the literature after intravenous administration, MAO-B would be partly inhibited and MAO-A completely inhibited ³⁹. A dye used in surgery for parathyroid localization, and as the antidote to methemoglobinemia, is a full-strength MAO-A inhibitor by intravenous injection.

The practical conclusion is dull, which is what makes it good advice. When any new drug is added for a patient on a monoamine oxidase inhibitor, including an antibiotic, a dye, a cough remedy or an analgesic, the question is not whether the new drug is an antidepressant but whether it raises serotonin or releases norepinephrine, and whether anyone has checked ^39,62,100.

The enzyme outside the nervous system: tumors and T cells

Two bodies of work have grown since 2014 that would have surprised anyone who thought of this as a brain enzyme. The immunology strand has not reached patients. The prostate cancer strand has, once: an open-label phase 2 trial of phenelzine in 20 men with biochemically recurrent, castration-sensitive disease ⁶⁶. Both are worth stating at their true strength.

The first is prostate cancer. Tumors from men with high-grade aggressive disease express more MAO-A, and a 2014 study worked out a route from the enzyme to the behavior of the cells: MAO-A drove the epithelial-to-mesenchymal transition, the change by which a tumor cell loosens from its neighbors and becomes migratory, and stabilized the oxygen-sensing transcription factor HIF1-alpha by raising reactive oxygen species, with a chain through vascular endothelial growth factor and its co-receptor neuropilin-1 into AKT, FOXO1 and TWIST1. Knocking MAO-A down reduced or even eliminated tumor growth and metastasis in mouse xenograft models, pharmacological inhibition reduced xenograft growth, and high expression in human tumor tissue correlated with worse clinical outcomes ¹⁰². Everything causal in that paper is cell culture and mice; the human part is an association between expression and outcome. The one human test of the idea is small and equivocal. Twenty men with biochemically recurrent, castrate-sensitive prostate cancer, of mean age 66.9 plus or minus 4.8 years and with a mean prostate-specific antigen of 4.7 plus or minus 5.8 nanograms per deciliter as the paper gives the unit, took phenelzine 30 mg twice daily in an open-label single-arm phase 2 study. The maximal fall in prostate-specific antigen was at least 30 percent in 5 of 20 and at least 50 percent in 2 of 20; dizziness occurred in 80 percent and hypertension of grade 2 or worse in 30 percent, and there was one episode of grade 4 hypertension and two of grade 3 syncope that forced treatment to stop ⁶⁶. Twenty patients, no control group and a surrogate endpoint: that is a signal worth a randomized trial, not a treatment, and the senior author is the senior author of the 2014 mechanism paper, which is a reason to weigh it carefully.

The second is immunology, and it inverts the usual logic. In 2021 a group reported that the Maoa gene is switched on in immune cells that infiltrate tumors, that knockout mice mount a stronger antitumor T cell response and grow tumors more slowly, that MAO inhibitors suppressed tumor growth in mouse syngeneic and human xenograft models in a way that depended on T cells, and that combining an inhibitor with an anti-PD-1 antibody produced synergistic suppression; the mechanism they identified is that MAO-A restrains T cells by controlling serotonin signaling that the T cells produce for themselves. Intratumoral expression of the gene correlated with T cell dysfunction and shorter survival across a range of cancers ⁶⁷. The proposal is to repurpose antidepressants as immunotherapy, and that proposal currently rests on mice, xenografts and correlation.

The medicinal chemistry has not slowed either. A 2026 review traced quinoxaline-based inhibitors from 1996 to 2026 and the structural rules that make a compound selective for one isoform ¹⁰³; new isatin-derived compounds reported in 2026 reached half-maximal inhibitory concentrations of 0.040 micromolar against MAO-A and 0.018 micromolar against MAO-B with a selectivity index above 15,000, binding competitively and reversibly ¹⁰⁴; and a hydrogen-sulfide-releasing MAO-B inhibitor reported the same year reached a half-maximal inhibitory concentration of 0.35 nanomolar and relieved motor deficits in MPTP-treated mice given the compound by mouth ¹⁰⁵. These are all test-tube and rodent results. The gap between a selectivity index of 15,000 in an assay and a benefit in a person is the whole of drug development, and most compounds do not cross it.

What was believed, and is not

Monoamine oxidase is one enzyme. It is two, with different genes, different substrate preferences, different active-site shapes and different drugs, and the division was made with a chemical before anyone had the proteins ^5,6,7,8,13.

Platelet MAO activity is a window on the brain and a marker of personality. It is a measure of one isoform in one cell type, it falls with cigarette smoking, and the smoking confound was not handled in most of the studies that built the claim; the review that worked through the confounders concluded that the promise had receded ^22,38,86. What survives is its behavior as a state marker of recent drinking ⁸⁷. The investigators who built the claim dispute that reading and argue that a residual association with sensation seeking and impulsiveness outlives the smoking confounder, through shared transcription factors rather than through brain MAO at all; theirs are narrative reviews by the claim's own advocates ^88,89.

Deprenyl slows Parkinson's disease. The 1989 interim result was read that way, but the same trial's full report showed the benefit appearing within three months and motor performance worsening after withdrawal, and said the mechanism of benefit was unclear; a 1994 re-evaluation concluded the endpoint effect was a direct treatment effect, not neuroprotection ^48,49,50. The delayed-start trial run twenty years later cleared its endpoints at 1 mg and failed them at 2 mg, and its authors asked for caution ⁴³.

Selegiline added to levodopa increases mortality. One trial in 520 patients reported exactly that, with a mortality ratio of 1.57 ⁵¹, but pooling 17 randomized trials and 3525 patients found no significant difference, odds ratio 1.13 (0.94 to 1.34) ⁵². The scare was a reversal, and the pooled analysis that failed to confirm it was a second one. Neither paper was withdrawn: the 1995 trial stands in the record as published, and what changed was the weight it carries once the other sixteen trials are set beside it.

Patients on these drugs must avoid a long list of foods. Chromatographic analysis of real meals found no significant tyramine in commercial chain pizzas including double pepperoni and double cheese, while finding it in week-old tofu and in one soy sauce, whose authors concluded that chain pizzas are safe with these drugs but that all soybean products should be avoided ⁸⁰. The narrowed modern diet sheet itself, usually given as aged and matured cheese, aged or spoiled meat and fish, fermented soy, tap or home-brewed beer and yeast extracts, is a guideline summary rather than a finding of that paper. With a reversible MAO-A inhibitor the margin is sixteen times wider than with an irreversible one ⁴⁰, and with a skin patch that spares the gut enzyme the pivotal trial was run with no dietary restriction at all ³⁵.

MAOA is a warrior gene. The common promoter repeat does alter transcription by 2 to 10 times in cell experiments ⁶⁰, and its interaction with documented childhood maltreatment in males has replicated and survived two meta-analyses ^53,54,55,56. But the effect is an interaction, not a main effect; it is modest; the heterogeneity between studies is substantial and unexplained; and in eleven female cohorts the interaction with combined adversity was absent, with the weak maltreatment effect running toward the high-activity genotype instead ^55,56. And when the whole genome is searched in 85,359 people rather than one gene tested in a few hundred, the signal that emerges is FOXP2, not MAOA ⁶⁸. Complete loss of the enzyme, which is a different thing entirely, causes a rare recognizable syndrome that comes with a biochemical fingerprint and appears to respond to treatment ^{23,27,28,81,82}.

Smokers get less Parkinson's disease because tobacco smoke inhibits MAO-B. The first half is as solid as observational epidemiology gets, with a relative risk of 0.59 overall and 0.39 in current smokers, dose-graded, and supported in the protective direction by genetic instrumental variable studies ^91,92,93. The enzyme reduction is also real, 40 percent in the brain and 33 to 46 percent in heart, lungs and kidneys ^25,58, and beta-carbolines in the smoke inhibit the enzymes in a test tube ⁹⁰. The causal bridge between the two halves is not built, the drugs that make that bridge testable have not shown disease modification ^43,50, and two further results push the other way: ease of quitting looks like an early feature of the disease itself ⁹⁴, and a year of transdermal nicotine did not slow early Parkinson's disease in a randomized trial of 163 patients ⁷².

What is still unknown

Whether elevated MAO-A in depression is a cause, a consequence or a parallel. The original comparisons were small, 17 against 17 and 15 against 15 ^24,29. The one prospective element in the literature is that the elevation persists after a selective serotonin reuptake inhibitor and into recovery, and is larger in those who then relapse ³⁰, which is compatible with a cause and equally compatible with a trait that outlasts the episode. Nobody has followed healthy people from health into a first episode with repeat scans, which is the study that would settle the direction.

Why the sexes differ. A 2026 preprint reports MAO-B distribution volume 10 percent higher in gray matter in women than men ⁷⁵, and the best meta-analysis of the MAOA promoter variant found the gene-by-maltreatment interaction in 20 male cohorts and not in 11 female cohorts ⁵⁵. X inactivation, in which one of a woman's two X chromosomes is silenced cell by cell, is an obvious candidate explanation and has not been shown to be the answer.

Whether anything that blocks MAO-B can change the course of a neurodegenerative disease. Thirty-seven years, two large trials and one reinterpretation later, the honest answer is that it has not been demonstrated ^43,48,50. The Alzheimer's attempts, the old drug and the purpose-built one, both failed ^61,65. The new reversible compounds aimed at astrocytes are a different pharmacological idea rather than a repeat of the old one, and the first of them has only completed phase 1 ⁷⁴.

Whether inhibiting monoamine oxidase would help the heart. The mouse evidence for MAO-A in pressure-overload heart failure is strong and causal ⁹⁵, and three independent human observations point the same way in atrial tissue and vein wall ^31,96,97. No randomized trial of a monoamine oxidase inhibitor for a cardiac endpoint has been done, and the gap between a mechanism in mice and a benefit in people is exactly the gap this page keeps describing.

Whether MAO-A inhibition will do anything useful in cancer. A mouse checkpoint story with xenograft support and human expression correlation is an interesting hypothesis ^67,102, and the drugs are cheap, old and available. One trial has been run: phenelzine in 20 men with biochemically recurrent prostate cancer, open label and single arm, with a prostate-specific antigen fall of at least 50 percent in 2 of them and real cardiovascular toxicity ⁶⁶. The immunotherapy trial that the mouse work implies has not been run, and neither has the randomized trial that the prostate signal deserves.

And a very old question that has never had a satisfying answer: what MAO-B is actually for. Its preferred substrate, beta-phenylethylamine, is a trace amine whose physiological job is unclear, mice without the enzyme are healthy enough to breed, and the clearest thing the enzyme does in a living person may be to make hydrogen peroxide and to activate the occasional poison ^8,41.

Where it connects

Sources

105 sources, numbered as they are cited. Every one was checked against PubMed or its publisher before it was cited here; the note under each says what it shows and what it does not.

1. 1
   Edmondson DE, Binda C, Wang J, Upadhyay AK, Mattevi A. Molecular and mechanistic properties of the membrane-bound mitochondrial monoamine oxidases.Biochemistry · 2009 · 48(20):4220-30doi:10.1021/bi900413g · PMID 19371079

   Review by the structural groups themselves: C-terminal helices anchor both enzymes to the outer mitochondrial membrane, and pulsed electron paramagnetic resonance shows both are dimeric in membrane preparations, with distances matching the crystals. A review, not new primary data.

2. 2
   Binda C, Milczek EM, Bonivento D, Wang J, Mattevi A, Edmondson DE. Lights and shadows on monoamine oxidase inhibition in neuroprotective pharmacological therapies.Curr Top Med Chem · 2011 · 11(22):2788-96doi:10.2174/156802611798184355 · PMID 22039878

   Review of the reaction chemistry and of the argument that part of the benefit of these drugs may come from preventing the oxidative stress the reaction produces rather than from raising neurotransmitters. Cited for the mechanism, not for any clinical claim.

3. 3
   Lan NC, Heinzmann C, Gal A, Klisak I, Orth U, Lai E, Grimsby J, Sparkes RS, Mohandas T, Shih JC. Human monoamine oxidase A and B genes map to Xp 11.23 and are deleted in a patient with Norrie disease.Genomics · 1989 · 4(4):552-9doi:10.1016/0888-7543(89)90279-6 · PMID 2744764

   The mapping paper. Its title says Xp11.23 and its abstract places the two genes close to each other and to the DXS7 locus at Xp11.3; both designations are the paper's own. Mapping used somatic cell hybrids, in situ hybridization, field-inversion gel electrophoresis and a deletion in one patient.

4. 4
   Grimsby J, Chen K, Wang LJ, Lan NC, Shih JC. Human monoamine oxidase A and B genes exhibit identical exon-intron organization.Proc Natl Acad Sci U S A · 1991 · 88(9):3637-41doi:10.1073/pnas.88.9.3637 · PMID 2023912

   Source for at least 60 kilobases, 15 exons, identical exon-intron organization and 93.9 percent peptide identity at exon 12, which codes the covalent FAD site, supporting duplication from one ancestral gene.

5. 5
   Bach AW, Lan NC, Johnson DL, Abell CW, Bembenek ME, Kwan SW, Seeburg PH, Shih JC. cDNA cloning of human liver monoamine oxidase A and B: molecular basis of differences in enzymatic properties.Proc Natl Acad Sci U S A · 1988 · 85(13):4934-8doi:10.1073/pnas.85.13.4934 · PMID 3387449

   Source for the subunit molecular weights of 59,700 and 58,800, the 70 percent sequence identity, the shared FAD-binding pentapeptide Ser-Gly-Gly-Cys-Tyr and the conclusion that the two forms come from separate genes.

6. 6
   De Colibus L, Li M, Binda C, Lustig A, Edmondson DE, Mattevi A. 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 U S A · 2005 · 102(36):12684-9doi:10.1073/pnas.0505975102 · PMID 16129825

   The MAO-A structure as its clorgyline adduct: one cavity of about 550 cubic angstroms against about 700 for deprenyl-inhibited human MAO-B and about 450 for rat MAO-A, monomeric in crystal and in solution, with an explicit warning against developing human drugs against non-human MAO-A.

7. 7
   Binda C, Newton-Vinson P, Hubalek F, Edmondson DE, Mattevi A. Structure of human monoamine oxidase B, a drug target for the treatment of neurological disorders.Nat Struct Biol · 2002 · 9(1):22-6doi:10.1038/nsb732 · PMID 11753429

   The MAO-B crystal structure at 3 angstroms: C-terminal transmembrane helix, 420 cubic angstrom substrate cavity with a 290 cubic angstrom entrance cavity, aromatic cage of Tyr 398 and Tyr 435, and pargyline bonded covalently to flavin N5.

8. 8
   Grimsby J, Toth M, Chen K, Kumazawa T, Klaidman L, Adams JD, Karoum F, Gal J, Shih JC. Increased stress response and beta-phenylethylamine in MAOB-deficient mice.Nat Genet · 1997 · 17(2):206-10doi:10.1038/ng1097-206 · PMID 9326944

   Source for the substrate preferences and for the MAO-B knockout phenotype: phenylethylamine rose, serotonin, norepinephrine and dopamine did not, stress reactivity increased, and the mice resisted MPTP. Mouse data.

9. 9
   Hare ML. Tyramine oxidase: A new enzyme system in liver.Biochem J · 1928 · 22(4):968-79doi:10.1042/bj0220968 · PMID 16744124

   The first description of the enzyme, under the name tyramine oxidase, from the Biochemical Laboratory in Cambridge. The PubMed record carries no abstract, so it is cited for the finding, the name and the date, not for any number. The author is the person later known as Mary Bernheim; see source 99.

10. 10
    Slotkin TA. Mary Bernheim and the discovery of monoamine oxidase.Brain Res Bull · 1999 · 50(5-6):373doi:10.1016/s0361-9230(99)00110-0 · PMID 10643441

    A one-page biographical note indexed as a Biography and Historical Article; no abstract in the PubMed record. Cited for the identity of the discoverer, that is for the link between the M. L. Hare of the 1928 paper and the Mary Bernheim of the later career, and for nothing else.

11. 11
    Blaschko H, Richter D, Schlossmann H. The inactivation of adrenaline.J Physiol · 1937 · 90(1):1-17doi:10.1113/jphysiol.1937.sp003497 · PMID 16994875

    The work conventionally credited with bringing the separate amine-oxidizing activities under one enzyme. No abstract in the record; the unification claim rests on the historical literature, not on anything verifiable from the database entry.

12. 12
    Zeller EA, Barsky J. In vivo inhibition of liver and brain monoamine oxidase by 1-Isonicotinyl-2-isopropyl hydrazine.Proc Soc Exp Biol Med · 1952 · 81(2):459-61doi:10.3181/00379727-81-19910 · PMID 13027339

    The paper that connects the tuberculosis drug iproniazid to the enzyme in a living animal, and in which the name monoamine oxidase already appears in the title. No abstract in the record.

13. 13
    Johnston JP. Some observations upon a new inhibitor of monoamine oxidase in brain tissue.Biochem Pharmacol · 1968 · 17(7):1285-97doi:10.1016/0006-2952(68)90066-x · PMID 5659776

    The paper that defined the A and B forms with the inhibitor clorgyline. No abstract in the record; cited for the split and its date. The indexing confirms work across rat, rabbit, dog and human tissue.

14. 14
    Selikoff IJ, Robitzek EH, Ornstein GG. Treatment of pulmonary tuberculosis with hydrazide derivatives of isonicotinic acid.J Am Med Assoc · 1952 · 150(10):973-80doi:10.1001/jama.1952.03680100015006 · PMID 12990328

    One of the 1952 tuberculosis reports on the hydrazide class that included iproniazid. No abstract in the record; cited for the date and the clinical setting in which the mood effect was first noticed.

15. 15
    Robitzek EH, Selikoff IJ. Hydrazine derivatives of isonicotinic acid (rimifon marsilid) in the treatment of active progressive caseous-pneumonic tuberculosis; a preliminary report.Am Rev Tuberc · 1952 · 65(4):402-28doi:10.1164/art.1952.65.4.402 · PMID 14903507

    The companion report; the title names both trade products, Rimifon (isoniazid) and Marsilid (iproniazid). No abstract in the record.

16. 16
    Loomer HP, Saunders JC, Kline NS. A clinical and pharmacodynamic evaluation of iproniazid as a psychic energizer.Psychiatr Res Rep Am Psychiatr Assoc · 1957 · 8:129-41PMID 13542681

    The paper conventionally credited with the first deliberate psychiatric use of a monoamine oxidase inhibitor. The priority is shared rather than clean: Crane published on iproniazid for mental disorders in the same volume of the same journal (source 80). Neither abstract nor DOI is in PubMed, so the trial design cannot be checked from the database; it is cited for the date and the claim.

17. 17
    Crane GE. Iproniazid (marsilid) phosphate, a therapeutic agent for mental disorders and debilitating diseases.Psychiatr Res Rep Am Psychiatr Assoc · 1957 · 8:142-52PMID 13542682

    Published in the same 1957 volume of the same journal as Loomer, Saunders and Kline, thirteen pages later. No abstract in PubMed; cited only for its date and its claim, which is why this page treats the priority as shared rather than clean.

18. 18
    Blackwell B. Hypertensive crisis due to monoamine-oxidase inhibitors.Lancet · 1963 · 2(7313):849-50doi:10.1016/s0140-6736(63)92743-0 · PMID 14056007

    The letter that put cheese and these drugs together. No abstract; the indexing confirms cheese, dietary proteins, headache, hypertension, phenelzine and tranylcypromine as its subject matter.

19. 19
    Asatoor AM, Levi AJ, Milne MD. Tranylcypromine and cheese.Lancet · 1963 · 2(7310):733-4doi:10.1016/s0140-6736(63)90368-4 · PMID 14057166

    The companion Lancet letter, three weeks earlier, indexed for chromatography. No abstract in the record.

20. 20
    Blackwell B, Marley E. Interactions of cheese and of its constituents with monoamine oxidase inhibitors.Br J Pharmacol Chemother · 1966 · 26(1):120-41doi:10.1111/j.1476-5381.1966.tb01816.x · PMID 5950679

    The experimental follow-up in animals, including cats and rats. No abstract in the record; cited for the existence of the controlled work rather than for a number.

21. 21
    Blackwell B, Marley E, Price J, Taylor D. Hypertensive interactions between monoamine oxidase inhibitors and foodstuffs.Br J Psychiatry · 1967 · 113(497):349-65doi:10.1192/bjp.113.497.349 · PMID 6034391

    The full clinical account. No abstract, but the indexing includes cerebral hemorrhage and subarachnoid hemorrhage alongside cheese, yeast and tyramine, which is the severity of the reaction stated in the record itself.

22. 22
    Buchsbaum MS, Coursey RD, Murphy DL. The biochemical high-risk paradigm: behavioral and familial correlates of low platelet monoamine oxidase activity.Science · 1976 · 194(4262):339-41doi:10.1126/science.968488 · PMID 968488

    The founding platelet MAO study: 375 students screened, top and bottom tenth interviewed, more counseling and legal trouble in the low group and an eightfold higher rate of suicide or attempts in their families. A selected-extremes design with no control for smoking.

23. 23
    Palmer EE, Leffler M, Rogers C, Shaw M, Carroll R, Earl J, Cheung NW, Champion B, Hu H, Haas SA, Kalscheuer VM, Gecz J, Field M. New insights into Brunner syndrome and potential for targeted therapy.Clin Genet · 2016 · 89(1):120-7doi:10.1111/cge.12589 · PMID 25807999

    Two further families with a protein-truncating and a missense MAOA variant, the biochemical fingerprint of high serum serotonin and urinary metanephrines with low 5-HIAA and VMA, and improvement of symptoms and biochemistry on a serotonin reuptake inhibitor with dietary change. Case reports, not a trial.

24. 24
    Meyer JH, Ginovart N, Boovariwala A, Sagrati S, Hussey D, Garcia A, Young T, Praschak-Rieder N, Wilson AA, Houle S. Elevated monoamine oxidase A levels in the brain: an explanation for the monoamine imbalance of major depression.Arch Gen Psychiatry · 2006 · 63(11):1209-16doi:10.1001/archpsyc.63.11.1209 · PMID 17088501

    17 depressed and 17 healthy non-smoking adults, medication free for at least five months: MAO-A binding elevated in all ten regions (p = .001 to 3 x 10 to the minus 7), averaging 34 percent, about two standard deviations. Cross-sectional with 17 per group; the conclusion that this is the primary monoamine-lowering process is the authors' interpretation.

25. 25
    Fowler JS, Logan J, Wang GJ, Volkow ND, Telang F, Zhu W, Franceschi D, Pappas N, Ferrieri R, Shea C, Garza V, Xu Y, Schlyer D, Gatley SJ, Ding YS, Alexoff D, Warner D, Netusil N, Carter P, Jayne M, King P, Vaska P. Low monoamine oxidase B in peripheral organs in smokers.Proc Natl Acad Sci U S A · 2003 · 100(20):11600-5doi:10.1073/pnas.1833106100 · PMID 12972641

    Extends the finding outside the brain: MAO-B significantly reduced in smokers in heart, lungs and kidneys, by 33 to 46 percent, using deuterium substitution to confirm binding specificity. The exposure is systemic, not cerebral.

26. 26
    Villemagne VL, Harada R, Dore V, Furumoto S, Mulligan R, Kudo Y, Burnham S, Krishnadas N, Bozinovski S, Huang K, Lopresti BJ, Yanai K, Rowe CC, Okamura N. First-in-Humans Evaluation of 18F-SMBT-1, a Novel 18F-Labeled Monoamine Oxidase-B PET Tracer for Imaging Reactive Astrogliosis.J Nucl Med · 2022 · 63(10):1551-1559doi:10.2967/jnumed.121.263254 · PMID 35086898

    77 participants: more than 85 percent of the signal blocked by selegiline 5 mg twice daily for 5 days, regional distribution matching known MAO-B (r = 0.84), and an age-related rise of about 2.6 percent a year against about 1.9 percent a year in vitro. A tracer validation study; the authors say larger series in mild cognitive impairment and Alzheimer's disease are needed.

27. 27
    Brunner HG, Nelen MR, van Zandvoort P, Abeling NG, van Gennip AH, Wolters EC, Kuiper MA, Ropers HH, van Oost BA. X-linked borderline mental retardation with prominent behavioral disturbance: phenotype, genetic localization, and evidence for disturbed monoamine metabolism.Am J Hum Genet · 1993 · 52(6):1032-9PMID 8503438

    The clinical and linkage paper on the Dutch kindred: maximal multipoint lod score 3.69 at the MAOA locus, disturbed monoamine metabolism on 24-hour urine in three affected males, and normal platelet MAO-B, which localized the defect to MAO-A. No DOI in the PubMed record.

28. 28
    Brunner HG, Nelen M, Breakefield XO, Ropers HH, van Oost BA. Abnormal behavior associated with a point mutation in the structural gene for monoamine oxidase A.Science · 1993 · 262(5133):578-80doi:10.1126/science.8211186 · PMID 8211186

    The mutation itself: a point mutation in the eighth exon changing a glutamine codon to a termination codon in each of five affected males, with complete and selective loss of MAO-A activity. The behavioral description is of this one family and is not a general claim about the gene.

29. 29
    Sacher J, Wilson AA, Houle S, Rusjan P, Hassan S, Bloomfield PM, Stewart DE, Meyer JH. Elevated brain monoamine oxidase A binding in the early postpartum period.Arch Gen Psychiatry · 2010 · 67(5):468-74doi:10.1001/archgenpsychiatry.2010.32 · PMID 20439828

    15 healthy women 4 to 6 days postpartum against 15 who had not recently given birth: MAO-A binding elevated by a mean of 43 percent in every region analyzed. The paper also states that estrogen falls 100- to 1000-fold in the first 3 to 4 days and that postpartum depression affects about 13 percent of women. A case-control imaging study of 30 women.

30. 30
    Meyer JH, Wilson AA, Sagrati S, Miler L, Rusjan P, Bloomfield PM, Clark M, Sacher J, Voineskos AN, Houle S. Brain monoamine oxidase A binding in major depressive disorder: relationship to selective serotonin reuptake inhibitor treatment, recovery, and recurrence.Arch Gen Psychiatry · 2009 · 66(12):1304-12doi:10.1001/archgenpsychiatry.2009.156 · PMID 19996035

    28 healthy participants, 16 scanned during a depressive episode and again after six weeks of a selective serotonin reuptake inhibitor, and 18 in recovery followed for six months. MAO-A binding stayed elevated in every region after treatment and in recovery, and prefrontal and anterior cingulate elevation in recovery was associated with subsequent recurrence. It carries the only prospective element in this literature, but the groups are small and it comes from the laboratory that produced the original finding.

31. 31
    Anderson EJ, Efird JT, Davies SW, O'Neal WT, Darden TM, Thayne KA, Katunga LA, Kindell LC, Ferguson TB, Anderson CA, Chitwood WR, Koutlas TC, Williams JM, Rodriguez E, Kypson AP. Monoamine oxidase is a major determinant of redox balance in human atrial myocardium and is associated with postoperative atrial fibrillation.J Am Heart Assoc · 2014 · 3(1):e000713doi:10.1161/JAHA.113.000713 · PMID 24572256

    Right atrial appendage from 244 cardiac surgery patients: monoamine oxidase and NADPH oxidase generated reactive oxygen species at roughly ten times the rate of intact coupled mitochondria, and postoperative atrial fibrillation risk rose across quartiles of MAO activity to an adjusted relative risk of 3.8 (1.9 to 7.5), trend p = 0.009. An association in human tissue; the authors say the enzyme still needs validating as a predictive marker.

32. 32
    Cases O, Seif I, Grimsby J, Gaspar P, Chen K, Pournin S, Muller U, Aguet M, Babinet C, Shih JC. Aggressive behavior and altered amounts of brain serotonin and norepinephrine in mice lacking MAOA.Science · 1995 · 268(5218):1763-6doi:10.1126/science.7792602 · PMID 7792602

    The MAOA-deficient mouse: serotonin up to ninefold higher in pup brains, norepinephrine up to twofold, cytoarchitectural change in somatosensory cortex, enhanced aggression in adult males, pup behavior reversed by a serotonin synthesis inhibitor. Mouse data, published two years after the human syndrome.

33. 33
    Cases O, Vitalis T, Seif I, De Maeyer E, Sotelo C, Gaspar P. Lack of barrels in the somatosensory cortex of monoamine oxidase A-deficient mice: role of a serotonin excess during the critical period.Neuron · 1996 · 16(2):297-307doi:10.1016/s0896-6273(00)80048-3 · PMID 8789945

    Shows the cortical abnormality is developmental and serotonin-driven: barrels absent, restored by early parachlorophenylalanine but not by blocking catecholamine synthesis, and reproduced by transiently inactivating MAOA in normal newborns.

34. 34
    Fulton B, Benfield P. Moclobemide. An update of its pharmacological properties and therapeutic use.Drugs · 1996 · 52(3):450-74doi:10.2165/00003495-199652030-00013 · PMID 8875133

    Drug review: reversible selective MAO-A inhibition, only weak potentiation of the tyramine pressor response, efficacy similar to tricyclics, selective serotonin reuptake inhibitors and irreversible MAO inhibitors. A commercially produced drug profile, which is a form of review with its own incentives.

35. 35
    Amsterdam JD. A double-blind, placebo-controlled trial of the safety and efficacy of selegiline transdermal system without dietary restrictions in patients with major depressive disorder.J Clin Psychiatry · 2003 · 64(2):208-14doi:10.4088/jcp.v64n0216 · PMID 12633131

    365 outpatients enrolled at 16 sites, 289 randomized: superiority to placebo on the Montgomery-Asberg scale (p = .001) and the 28-item Hamilton scale (p = .039) but not the 17-item Hamilton scale (p = .069), application-site reactions in 31.5 versus 15.1 percent, no blood pressure difference, no tyramine restriction. The author calls the effect modest.

36. 36
    Jo S, Yarishkin O, Hwang YJ, Chun YE, Park M, Woo DH, Bae JY, Kim T, Lee J, Chun H, Park HJ, Lee DY, Hong J, Kim HY, Oh SJ, Park SJ, Lee H, Yoon BE, Kim Y, Jeong Y, Shim I, Bae YC, Cho J, Kowall NW, Ryu H, Hwang E, Kim D, Lee CJ. GABA from reactive astrocytes impairs memory in mouse models of Alzheimer's disease.Nat Med · 2014 · 20(8):886-96doi:10.1038/nm.3639 · PMID 24973918

    Reactive astrocytes make GABA through MAO-B and release it through the bestrophin 1 channel, reducing granule cell firing in mouse models; suppressing production or release restored firing, plasticity and memory. Astrocytic GABA and MAO-B were raised in postmortem human Alzheimer's brain, which is an association, while the causal work is in mice.

37. 37
    Abrahamson EE, Kofler JK, Lopez OL, Cohen AD, Gogola A, Aizenstein HJ, Karikari TK, Mufson EJ, Villemagne VL, Ikonomovic MD. Characterization of [18F]SMBT-1 binding substrates in Alzheimer's disease and related disorders: A postmortem biochemical and immunohistochemical study.Alzheimers Dement · 2026 · 22(8):e71685doi:10.1002/alz.71685 · PMID 42609050

    Autopsy tissue: tracer binding correlated with MAO-B activity and with amyloid binding, and corresponded to astrocytes positive for MAO-B and for GFAP, glial fibrillary acidic protein, the structural marker used to identify reactive astrocytes, associated with amyloid in plaques and vessels more closely than with tau. In non-Alzheimer tauopathies, binding and activity partially overlapped both controls and Alzheimer's cases, which limits its use as a disease-specific marker.

38. 38
    Farren CK. Platelet monoamine oxidase (MAO) activity and alcoholism: is there a genuine association?Addict Biol · 1997 · 2(2):171-80doi:10.1080/13556219772714 · PMID 26735634

    The review that works through every confounder of the platelet MAO literature and concludes the initial promise had receded, with the failure to account for smoking leaving the hypothesis open to criticism. A narrative review, not a pooled analysis.

39. 39
    Ramsay RR, Dunford C, Gillman PK. Methylene blue and serotonin toxicity: inhibition of monoamine oxidase A (MAO A) confirms a theoretical prediction.Br J Pharmacol · 2007 · 152(6):946-51doi:10.1038/sj.bjp.0707430 · PMID 17721552

    Kinetic work on purified human enzyme: methylene blue is a potent tight-binding reversible MAO-A inhibitor, far weaker against MAO-B, and at concentrations reported after intravenous dosing MAO-A would be completely inhibited. Enzymology interpreted against published case reports, not a clinical study.

40. 40
    Simpson GM, Gratz SS. Comparison of the pressor effect of tyramine after treatment with phenelzine and moclobemide in healthy male volunteers.Clin Pharmacol Ther · 1992 · 52(3):286-91doi:10.1038/clpt.1992.143 · PMID 1526086

    Twelve healthy men, 28 days of open-label treatment: the mean oral tyramine dose raising systolic pressure by 30 mmHg was 15 mg on phenelzine and 240 mg on moclobemide, with no response three days after stopping moclobemide up to hourly doses of 320 mg. The phenelzine day-28 tyramine challenge was performed in four of the twelve volunteers, and the moclobemide figure is the day-28 mean in the moclobemide arm. Small, open-label, healthy volunteers, not patients eating food.

41. 41
    Chiba K, Trevor A, Castagnoli N. Metabolism of the neurotoxic tertiary amine, MPTP, by brain monoamine oxidase.Biochem Biophys Res Commun · 1984 · 120(2):574-8doi:10.1016/0006-291x(84)91293-2 · PMID 6428396

    Rat brain mitochondrial fractions converted MPTP to the 1-methyl-4-phenylpyridinium ion at 0.91 plus or minus 0.02 nanomoles per milligram of protein per minute, blocked by 100 nanomolar deprenyl or pargyline. The bioactivation that made MAO-B inhibition look neuroprotective; rat tissue.

42. 42
    Ulrich S, Ricken R, Buspavanich P, Schlattmann P, Adli M. Efficacy and Adverse Effects of Tranylcypromine and Tricyclic Antidepressants in the Treatment of Depression: A Systematic Review and Comprehensive Meta-analysis.J Clin Psychopharmacol · 2020 · 40(1):63-74doi:10.1097/JCP.0000000000001153 · PMID 31834088

    Pooled log odds ratio 0.480 (0.105 to 0.857, p = 0.01) favoring tranylcypromine over tricyclics across eight trials, without important heterogeneity; the authors conclude the two are equal in a mixed sample and that the advantage was driven by two trials in psychomotor-retarded depression. The first author's affiliation is a pharmaceutical company that markets tranylcypromine.

43. 43
    Olanow CW, Rascol O, Hauser R, Feigin PD, Jankovic J, Lang A, Langston W, Melamed E, Poewe W, Stocchi F, Tolosa E; ADAGIO Investigators. A double-blind, delayed-start trial of rasagiline in Parkinson's disease.N Engl J Med · 2009 · 361(13):1268-78doi:10.1056/NEJMoa0809335 · PMID 19776408

    1176 untreated patients: 1 mg daily met all three hierarchical endpoints (0.09 versus 0.14 UPDRS points per week between weeks 12 and 36, p = 0.01; 2.82 versus 4.52 points at week 72, p = 0.02; non-inferiority weeks 48 to 72), while 2 mg met none (3.47 versus 3.11 points, p = 0.60). The authors state the discordance between doses demands caution.

44. 44
    Alborghetti M, Nicoletti F. Different Generations of Type-B Monoamine Oxidase Inhibitors in Parkinson's Disease: From Bench to Bedside.Curr Neuropharmacol · 2019 · 17(9):861-873doi:10.2174/1570159X16666180830100754 · PMID 30160213

    Review of the three licensed MAO-B inhibitors, and the source used here for the statement that selegiline and rasagiline are irreversible inhibitors forming a covalent bond in the active site while safinamide is a reversible one that also inhibits voltage-sensitive sodium channels and glutamate release. A review, with no new data and with its own expectation that these drugs may yet prove neuroprotective.

45. 45
    Langston JW, Ballard P, Tetrud JW, Irwin I. Chronic Parkinsonism in humans due to a product of meperidine-analog synthesis.Science · 1983 · 219(4587):979-80doi:10.1126/science.6823561 · PMID 6823561

    Four people developed marked parkinsonism after intravenous use of an illicit drug that on analysis was mainly MPTP; the authors proposed that the chemical selectively damages cells in the substantia nigra. A four-person case series. The record does not say where they lived or how old they were, so the familiar description of them as young Californians is not taken from here.

46. 46
    Tipton KF. 90 years of monoamine oxidase: some progress and some confusion.J Neural Transm (Vienna) · 2018 · 125(11):1519-1551doi:10.1007/s00702-018-1881-5 · PMID 29637260

    A long historical and mechanistic review by one of the field's principal authorities, indexed as a Historical Article, covering the reaction, the inhibitor classes and the areas of continuing uncertainty. A review: cited for history and framing, not for any primary number, and the abstract does not itself carry the naming attribution this page describes as conventional.

47. 47
    Sathyanarayana Rao TS, Yeragani VK. Hypertensive crisis and cheese.Indian J Psychiatry · 2009 · 51(1):65-6doi:10.4103/0019-5545.44910 · PMID 19742203

    A short historical account with free full text, not primary data: it records the pharmacist-and-wife origin story, summarizes Blackwell's series of 12 patients with the blood pressure range of 160/90 to 220/115 mmHg and the list of complications, and notes Blackwell's 1965 measurement of 72 to 953 micrograms of tyramine per gram across 14 cheddar cheeses. Cited for the history, with its secondary status stated.

48. 48
    Parkinson Study Group. Effect of deprenyl on the progression of disability in early Parkinson's disease.N Engl J Med · 1989 · 321(20):1364-71doi:10.1056/NEJM198911163212004 · PMID 2509910

    DATATOP interim report: 800 patients in a two-by-two design, 97 on deprenyl versus 176 not on it reaching the need for levodopa over an average of 12 months, hazard ratio 0.43 (0.33 to 0.55). An interim analysis of a surrogate endpoint, read at the time as neuroprotection.

49. 49
    Parkinson Study Group. Effects of tocopherol and deprenyl on the progression of disability in early Parkinson's disease.N Engl J Med · 1993 · 328(3):176-83doi:10.1056/NEJM199301213280305 · PMID 8417384

    The full report after a mean 14 plus or minus 6 months: no effect of tocopherol, hazard ratio 0.50 (0.41 to 0.62) for deprenyl with about nine months' delay, but ratings improved in the first three months and motor performance worsened after withdrawal, and the authors state the action accounting for the benefit remains unclear. This is the paper that undermines the neuroprotection reading of its own predecessor.

50. 50
    Ward CD. Does selegiline delay progression of Parkinson's disease? A critical re-evaluation of the DATATOP study.J Neurol Neurosurg Psychiatry · 1994 · 57(2):217-20doi:10.1136/jnnp.57.2.217 · PMID 8126510

    Models the trial on the assumption that measured impairment is proportional to cellular dysfunction and concludes the reduced probability of reaching the endpoint was a direct treatment effect rather than neuroprotection. An analysis of published data, not new data.

51. 51
    Lees AJ; Parkinson's Disease Research Group of the United Kingdom. Comparison of therapeutic effects and mortality data of levodopa and levodopa combined with selegiline in patients with early, mild Parkinson's disease.BMJ · 1995 · 311(7020):1602-7doi:10.1136/bmj.311.7020.1602 · PMID 8555803

    520 patients at 93 UK hospitals, open label, average 5 to 6 years: mortality ratio 1.57 (1.09 to 2.30) for the combination, log rank p = 0.015, adjusted hazard ratios 1.49 and 1.57, with no significant disability benefit and more dyskinesia and on-off fluctuation. Open-label design; the mortality signal was not confirmed by later pooling.

52. 52
    Ives NJ, Stowe RL, Marro J, Counsell C, Macleod A, Clarke CE, Gray R, Wheatley K. Monoamine oxidase type B inhibitors in early Parkinson's disease: meta-analysis of 17 randomised trials involving 3525 patients.BMJ · 2004 · 329(7466):593doi:10.1136/bmj.38184.606169.AE · PMID 15310558

    No significant mortality difference (odds ratio 1.13, 0.94 to 1.34, p = 0.2), better rating scale scores at three months, less need for levodopa (0.57, 0.48 to 0.67) and fewer motor fluctuations (0.75, 0.59 to 0.95), with no excess side effects. The authors note few trials compared these drugs with other antiparkinsonian agents.

53. 53
    Caspi A, McClay J, Moffitt TE, Mill J, Martin J, Craig IW, Taylor A, Poulton R. Role of genotype in the cycle of violence in maltreated children.Science · 2002 · 297(5582):851-4doi:10.1126/science.1072290 · PMID 12161658

    The original gene-by-environment finding in a male birth cohort: maltreated children with the high-activity genotype were less likely to develop antisocial problems. The claim is about an interaction with maltreatment, not about the gene causing violence on its own.

54. 54
    Kim-Cohen J, Caspi A, Taylor A, Williams B, Newcombe R, Craig IW, Moffitt TE. MAOA, maltreatment, and gene-environment interaction predicting children's mental health: new evidence and a meta-analysis.Mol Psychiatry · 2006 · 11(10):903-13doi:10.1038/sj.mp.4001851 · PMID 16801953

    Replication in 975 seven-year-old boys plus a meta-analysis showing the maltreatment-outcome association is significantly stronger in males with the low-activity genotype. Conducted by the original authors, which is a reason to weigh the independent meta-analyses alongside it.

55. 55
    Byrd AL, Manuck SB. MAOA, childhood maltreatment, and antisocial behavior: meta-analysis of a gene-environment interaction.Biol Psychiatry · 2014 · 75(1):9-17doi:10.1016/j.biopsych.2013.05.004 · PMID 23786983

    Independent meta-analysis of 27 studies through August 2012: interaction confirmed across 20 male cohorts (p = .0044), specific to maltreatment (p = .00000082); across 11 female cohorts no interaction with combined adversity, and maltreatment alone predicted antisocial behavior weakly in women with the high-activity genotype, the opposite direction.

56. 56
    Ficks CA, Waldman ID. Candidate genes for aggression and antisocial behavior: a meta-analysis of association studies of the 5HTTLPR and MAOA-uVNTR.Behav Genet · 2014 · 44(5):427-44doi:10.1007/s10519-014-9661-y · PMID 24902785

    Finds the MAOA repeat significantly associated with antisocial behavior across studies, with substantial heterogeneity that no sample-level or study-level characteristic explained and no evidence of publication bias for this marker, and describes the role of common serotonergic variants as modest.

57. 57
    López-Muñoz F, Alamo C. Monoaminergic neurotransmission: the history of the discovery of antidepressants from 1950s until today.Curr Pharm Des · 2009 · 15(14):1563-86doi:10.2174/138161209788168001 · PMID 19442174

    Historical review: the 1950s saw the clinical introduction of the first two specifically antidepressant drugs, iproniazid and imipramine. Cited here for the fact that the two arrived together, which is why this page says the inhibitors were among the first antidepressants rather than the first. A review, not primary data.

58. 58
    Fowler JS, Volkow ND, Wang GJ, Pappas N, Logan J, MacGregor R, Alexoff D, Shea C, Schlyer D, Wolf AP, Warner D, Zezulkova I, Cilento R. Inhibition of monoamine oxidase B in the brains of smokers.Nature · 1996 · 379(6567):733-6doi:10.1038/379733a0 · PMID 8602220

    The 40 percent lower brain MAO-B in living smokers compared with non-smokers and former smokers, with the authors' proposal that this may act alongside nicotine to produce the behavioral and epidemiological features of smoking. An imaging comparison, not an experiment.

59. 59
    Fowler JS, Volkow ND, Wang GJ, Pappas N, Logan J, Shea C, Alexoff D, MacGregor RR, Schlyer DJ, Zezulkova I, Wolf AP. Brain monoamine oxidase A inhibition in cigarette smokers.Proc Natl Acad Sci U S A · 1996 · 93(24):14065-9doi:10.1073/pnas.93.24.14065 · PMID 8943061

    15 non-smokers against 16 smokers with carbon-11 clorgyline: MAO-A about 28 percent lower in every region, roughly half the reduction produced in four volunteers by tranylcypromine 10 mg daily for three days.

60. 60
    Sabol SZ, Hu S, Hamer D. A functional polymorphism in the monoamine oxidase A gene promoter.Hum Genet · 1998 · 103(3):273-9doi:10.1007/s004390050816 · PMID 9799080

    Describes the 30-base-pair repeat 1.2 kilobases upstream of MAOA in 3, 3.5, 4 or 5 copies, with 3.5 and 4 copies transcribed 2 to 10 times more efficiently in gene fusion and transfection experiments in three cell types. Cell experiments; it does not measure enzyme in anyone's brain.

61. 61
    Birks J, Flicker L. Selegiline for Alzheimer's disease.Cochrane Database Syst Rev · 2003 · (1):CD000442doi:10.1002/14651858.CD000442 · PMID 12535396

    17 included double-blind randomized trials, mostly completers' data, with individual patient data from seven. A few scattered effects on combined cognitive tests and on activities of daily living, none on emotional state or global rating, no excess of adverse effects, and the authors' conclusion that despite its initial promise selegiline for Alzheimer's disease proved disappointing, with no evidence of clinically meaningful benefit and no justification for using it or studying it further there. A systematic review, not new data.

62. 62
    Lawrence KR, Adra M, Gillman PK. Serotonin toxicity associated with the use of linezolid: a review of postmarketing data.Clin Infect Dis · 2006 · 42(11):1578-83doi:10.1086/503839 · PMID 16652315

    29 cases from the US Food and Drug Administration's adverse event reporting system in which linezolid was the primary suspect with a serotonergic co-prescription, most often a selective serotonin reuptake inhibitor (26 of 43 patients); 13 required hospitalization or intervention to prevent permanent impairment. Spontaneous reports establish that the interaction occurs but cannot give an incidence.

63. 63
    PD MED Collaborative Group; Gray R, Ives N, Rick C, Patel S, Gray A, Jenkinson C, McIntosh E, Wheatley K, Williams A, Clarke CE. Long-term effectiveness of dopamine agonists and monoamine oxidase B inhibitors compared with levodopa as initial treatment for Parkinson's disease (PD MED): a large, open-label, pragmatic randomised trial.Lancet · 2014 · 384(9949):1196-205doi:10.1016/S0140-6736(14)60683-8 · PMID 24928805

    1620 patients randomized between November 2000 and December 2009, ISRCTN69812316: 528 to levodopa, 632 to a dopamine agonist and 460 to an MAO-B inhibitor, median follow-up three years with up to seven years of observation. Open-label and patient-rated, which is its main weakness; its strengths are size, duration and a quality-of-life primary outcome rather than a surrogate. Rates of dementia, institutionalization and death did not differ significantly.

64. 64
    Schapira AHV, Fox SH, Hauser RA, Jankovic J, Jost WH, Kenney C, Kulisevsky J, Pahwa R, Poewe W, Anand R. Assessment of Safety and Efficacy of Safinamide as a Levodopa Adjunct in Patients With Parkinson Disease and Motor Fluctuations: A Randomized Clinical Trial.JAMA Neurol · 2017 · 74(2):216-224doi:10.1001/jamaneurol.2016.4467 · PMID 27942720

    549 patients at 119 centers over 24 weeks: on time without troublesome dyskinesia rose 1.42 hours against 0.57 for placebo, a least-squares mean difference of 0.96 hours (0.56 to 1.37), p < .001; dyskinesia was reported in 14.6 versus 5.5 percent. One listed author's affiliation is the company developing the drug.

65. 65
    Nave S, Doody RS, Boada M, Grimmer T, Savola JM, Delmar P, Pauly-Evers M, Nikolcheva T, Czech C, Borroni E, Ricci B, Dukart J, Mannino M, Carey T, Moran E, Gilaberte I, Muelhardt NM, Gerlach I, Santarelli L, Ostrowitzki S, Fontoura P. Sembragiline in Moderate Alzheimer's Disease: Results of a Randomized, Double-Blind, Placebo-Controlled Phase II Trial (MAyflOwer RoAD).J Alzheimers Dis · 2017 · 58(4):1217-1228doi:10.3233/JAD-161309 · PMID 28550255

    542 patients with Mini-Mental State Examination scores of 13 to 20 randomized 1:1:1 to 1 mg, 5 mg or placebo for 52 weeks on top of standard treatment (NCT01677754). Well tolerated; primary and secondary endpoints missed. The behavioral signals were present at week 52 only and the subgroup finding was post hoc. Most authors were employed by F. Hoffmann-La Roche or Genentech, which developed the drug.

66. 66
    Gross ME, Agus DB, Dorff TB, Pinski JK, Quinn DI, Castellanos O, Gilmore P, Shih JC. Phase 2 trial of monoamine oxidase inhibitor phenelzine in biochemical recurrent prostate cancer.Prostate Cancer Prostatic Dis · 2020 · 24(1):61-68doi:10.1038/s41391-020-0211-9 · PMID 32123315

    Open-label single-arm phase 2 in 20 evaluable men, mean age 66.9 plus or minus 4.8 years and prostate-specific antigen 4.7 plus or minus 5.8 ng/dl: maximal declines of at least 30 percent in 25 percent and at least 50 percent in 10 percent, dizziness in 80 percent, hypertension of grade 2 or worse in 30 percent, one grade 4 hypertension and two grade 3 syncopes forcing discontinuation. No control arm, a surrogate endpoint and real cardiovascular toxicity; the senior author is the author of the 2014 mechanism paper, which is a reason to weigh it carefully.

67. 67
    Wang X, Li B, Kim YJ, Wang YC, Li Z, Yu J, Zeng S, Ma X, Choi IY, Di Biase S, Smith DJ, Zhou Y, Li YR, Ma F, Huang J, Clarke N, To A, Gong L, Pham AT, Moon H, Pellegrini M, Yang L. Targeting monoamine oxidase A for T cell-based cancer immunotherapy.Sci Immunol · 2021 · 6(59):eabh2383doi:10.1126/sciimmunol.abh2383 · PMID 33990379

    Maoa induced in tumor-infiltrating immune cells; knockout mice showed stronger antitumor T cell immunity, MAO inhibitors suppressed tumor growth in mouse syngeneic and human xenograft models in a T-cell-dependent way and synergized with anti-PD-1, acting through autocrine serotonin signaling in T cells, with intratumoral expression correlating with T cell dysfunction and shorter survival in patients. Mice, xenografts and correlation, not a trial.

68. 68
    Tielbeek JJ, Uffelmann E, Williams BS, Colodro-Conde L, Gagnon É, Mallard TT, et al. Uncovering the genetic architecture of broad antisocial behavior through a genome-wide association study meta-analysis.Mol Psychiatry · 2022 · 27(11):4453-4463doi:10.1038/s41380-022-01793-3 · PMID 36284158

    28 discovery samples totaling 85,359 people and five replication samples totaling 8058. The first genome-wide significant association with broad antisocial behavior was in common intronic variants of FOXP2, lead variant rs12536335, with supporting mouse data and polygenic score associations with conduct disorder and criminal convictions. MAOA is not among the loci it surfaces. The genetic correlations it reports are associations between inherited liabilities, not causes.

69. 69
    Tielbeek JJ, Johansson A, Polderman TJC, Rautiainen MR, Jansen P, Taylor M, et al. Genome-Wide Association Studies of a Broad Spectrum of Antisocial Behavior.JAMA Psychiatry · 2017 · 74(12):1242-1250doi:10.1001/jamapsychiatry.2017.3069 · PMID 28979981

    16,400 discovery and 9381 target participants, all of European descent, with an explicit reevaluation of candidate-gene-era data. Three promising loci with sex-discordant associations, a highly polygenic architecture, and an inverse genetic correlation with educational attainment of minus 0.52.

70. 70
    Bai AD, McKenna S, Wise H, Loeb M, Gill SS. Association of Linezolid With Risk of Serotonin Syndrome in Patients Receiving Antidepressants.JAMA Netw Open · 2022 · 5(12):e2247426doi:10.1001/jamanetworkopen.2022.47426 · PMID 36534400

    Population-based retrospective cohort of 1134 outpatients aged 66 or older in Ontario prescribed oral linezolid between October 2014 and January 2021, 215 of them also on an antidepressant; outcomes by physician diagnosis, Sternbach criteria or Hunter criteria within 30 days. Older outpatients and oral linezolid only, with so few events that the confidence interval is wide: it bounds the risk rather than excluding it.

71. 71
    Van den Eynde V, Abdelmoemin WR, Abraham MM, Amsterdam JD, Anderson IM, Andrade C, et al.; Gillman PK. The prescriber's guide to classic MAO inhibitors (phenelzine, tranylcypromine, isocarboxazid) for treatment-resistant depression.CNS Spectr · 2022 · 28(4):427-440doi:10.1017/S1092852922000906 · PMID 35837681

    A specialist-consensus standard with more than 70 international expert endorsers, explicitly based on empirical evidence plus expert opinion drawn from six decades of use rather than on new data; the authors present it as correcting misconceptions in the existing literature. Several endorsers have pharmaceutical affiliations, including Neurawell Therapeutics and Aristo Pharma, and the lead and senior authors run a private research organization, all of which the author list discloses.

72. 72
    Oertel WH, Müller HH, Unger MM, Schade-Brittinger C, Balthasar K, Articus K, Brinkman M, Venuto CS, Tracik F, Eberling J, Eggert KM, Kamp C, Kieburtz K, Boyd JT. Transdermal Nicotine Treatment and Progression of Early Parkinson's Disease.NEJM Evid · 2023 · 2(9):EVIDoa2200311doi:10.1056/EVIDoa2200311 · PMID 38320207

    Double-blind placebo-controlled multicenter trial, 163 randomized and 101 assessed for the primary endpoint over 52 weeks of treatment and 8 of washout. Negative: total rating-scale worsening was 3.5 points on placebo against 6.0 on nicotine. Patients could stay on stable MAO-B inhibition, 34.6 percent started dopaminergic therapy during the trial, and dropout for skin reactions and adverse events was substantial. Funded in part by the Michael J. Fox Foundation; several authors were employed by Novartis Pharma.

73. 73
    Gimenez-Palomo A, Chamdal AK, Gottlieb N, Lotfaliany M, Jokinen T, Bastawy EM, Adlington K, Benachar N, Dodd S, Pacchiarotti I, Vieta E, Berk M, Stokes PRA. Efficacy and tolerability of monoamine oxidase inhibitors for the treatment of depressive episodes in mood disorders: A systematic review and network meta-analysis.Acta Psychiatr Scand · 2024 · 150(6):500-515doi:10.1111/acps.13728 · PMID 39001570

    83 double-blind randomized trials, 7765 participants on active treatment and 1844 on placebo: isocarboxazid, phenelzine, tranylcypromine and moclobemide significantly better than placebo, tolerability and acceptability comparable to other antidepressants. The authors note the trials are not concentrated where the drugs are actually used, in treatment-resistant and atypical depression.

74. 74
    Yoon S, Won W, Lee S, Han K, Ha E, Lee J, Hyeon SJ, Joo Y, Hong H, Lee H, Song Y, Park KD, Huber BR, Lee J, Edden RAE, Suh M, Ryu H, Lee CJ, Lyoo IK. Astrocytic gamma-aminobutyric acid dysregulation as a therapeutic target for posttraumatic stress disorder.Signal Transduct Target Ther · 2025 · 10(1):240doi:10.1038/s41392-025-02317-5 · PMID 40717119

    Human cross-sectional and longitudinal imaging showing elevated prefrontal GABA in post-traumatic stress disorder that normalized with recovery, postmortem and mouse work implicating MAO-B-dependent astrocytic GABA, and a report that the reversible MAO-B inhibitor KDS2010 completed phase 1 with a favorable safety profile, with phase 2 planned. Phase 1 is a safety study in volunteers; one author's institute developed the compound.

75. 75
    Kuik EJ, Baena-Tan C, Moriguchi S, Braga J, Miler L, Rusjan P, Bagby M, Husain MI, Vasdev N, Tomoda T, Banasr M, Kloiber S, Meyer JH. Effects of sex, treatment resistance, and antidepressant treatment on monoamine oxidase B total distribution volume.medRxiv · 2026 · preprint, posted 10 September 2026doi:10.64898/2026.09.08.26362462 · PMID 42818885

    A preprint, not peer reviewed. In 95 adults, MAO-B distribution volume was 10 percent higher in gray matter in women, 23 percent higher in prefrontal cortex in treatment-resistant depression than in controls and 6 percent higher than in non-treatment-resistant episodes; in 13 participants scanned before and after treatment, rasagiline and phenelzine occupied about 94 percent of the enzyme while duloxetine had minimal effect.

76. 76
    Ahmed AAO, Ahmed OAO, Dishean AS, Gilbert C, Elkheir MMAS, Sota HA, Abdulaziz MAM, Mahmoud EYM, Ahmed RAO. Efficacy and Safety of Safinamide as an Adjunct to Levodopa for Motor Fluctuations in Parkinson's Disease: A Systematic Review and Meta-Analysis of Randomized Controlled Trials.Parkinsons Dis · 2026 · 2026:9352886doi:10.1155/padi/9352886 · PMID 42846716

    Three randomized controlled trials and 1127 participants, PRISMA-compliant with Cochrane RoB 2 risk-of-bias assessment and GRADE certainty, PROSPERO CRD420261298124. Only three trials, with substantial heterogeneity for the motor outcome (I-squared 72 percent), which is why that estimate is rated low certainty.

77. 77
    deSouza RM, Schapira A. Safinamide for the treatment of Parkinson's disease.Expert Opin Pharmacother · 2017 · 18(9):937-943doi:10.1080/14656566.2017.1329819 · PMID 28504022

    Review stating that safinamide, marketed as Xadago, began as a first-generation anticonvulsant, that its mechanism has not been fully defined, and that it improves motor function and quality of life as an adjunct in advanced disease. A review, and the basis for not crediting its benefit to MAO-B inhibition alone.

78. 78
    Dingemanse J, Hussain Y, Korn A. Tyramine pharmacodynamics during combined administration of lazabemide and moclobemide.Int J Clin Pharmacol Ther · 1996 · 34(4):172-7PMID 8861736

    Combining reversible MAO-A and MAO-B inhibitors potentiated oral tyramine 13.5-fold plus or minus 6.9, with 51 plus or minus 20 mg needed for a 30 mmHg rise, leading the authors to conclude dietary precautions are needed when both isoforms are inhibited. Thirteen volunteers in two panels; the authors were employed by the manufacturer. No DOI in the record.

79. 79
    Hauser RA, Lew MF, Hurtig HI, Ondo WG, Wojcieszek J, Fitzer-Attas CJ; TEMPO Open-label Study Group. Long-term outcome of early versus delayed rasagiline treatment in early Parkinson's disease.Mov Disord · 2009 · 24(4):564-73doi:10.1002/mds.22402 · PMID 19086083

    404 randomized with 306 entering open-label extension and a mean 3.6 years in study: an adjusted mean difference of 2.5 UPDRS units (SE 1.1, p = 0.021) favoring early start over 6.5 years. The authors say this might reflect neuroprotection or effects on compensatory mechanisms; an open-label extension cannot distinguish them.

80. 80
    Shulman KI, Walker SE. Refining the MAOI diet: tyramine content of pizzas and soy products.J Clin Psychiatry · 1999 · 60(3):191-3PMID 10192596

    Chromatographic analysis with 6 mg of tyramine or less taken as safe: no significant tyramine in chain-outlet pizzas including double pepperoni with double cheese, clinically significant amounts in week-old tofu and in one soy sauce. No DOI in the PubMed record. Cited for the chromatographic food measurements and the 6 mg threshold only; its own recommendation was that all soybean products be avoided, especially soy sauce and tofu, while chain-outlet pizzas are safe. Food analysis, not a clinical study.

81. 81
    Minniti ML, Kalantari S, Pasca L, Bruno S, Arceri S, Novello E, Giorgio E, Rizzo V, Borgatti R, Valente EM, Pisani A, Orcesi S, Sirchia F. Expanding the phenotype of Brunner syndrome from childhood to adulthood: Description of the second pediatric patient and his mother.Am J Med Genet A · 2023 · 194(1):82-87doi:10.1002/ajmg.a.63413 · PMID 37750385

    Two-person case report with biochemical confirmation of MAO-A deficiency in a 5-year-old boy and in his carrier mother, both with the maternally inherited variant c.410A>G, p.Glu137Gly. It widens the sex assumptions around an X-linked condition. A case report, with no control for the natural course of symptoms.

82. 82
    Gravagno E, Bellini M, Ambrosini E, Luberto A, Busciglio S, Vitetta G, Cannizzaro IR, Taiani A, Barili V, Percesepe A, Uliana V, Martorana D. Genotype-Driven Diagnosis Enables Targeted Pharmacological Treatment in Brunner Syndrome: A Novel Splice-Site Variant and Case-Based Review.Int J Mol Sci · 2026 · 27(14):6223doi:10.3390/ijms27146223 · PMID 42511567

    One patient, with functional messenger RNA work confirming intron 8 retention and a premature stop codon, plus a review of reported variants. The treatment claim is a single off-label observation with no comparison group, which the authors describe as preliminary evidence of a potential benefit.

83. 83
    Lea R, Chambers G. Monoamine oxidase, addiction, and the "warrior" gene hypothesis.N Z Med J · 2007 · 120(1250):U2441PMID 17339897

    An editorial, indexed under aggression, gene frequency, racial groups, risk-taking and violence; no abstract and no DOI in the PubMed record. Cited as the published origin of the warrior gene framing, not as evidence for it.

84. 84
    Merriman T, Cameron V. Risk-taking: behind the warrior gene story.N Z Med J · 2007 · 120(1250):U2440PMID 17339896

    The reply in the same issue of the same journal: credible evidence exists for a contribution of an MAOA variant to antisocial behavior in people of European descent, but no direct evidence supports such an association in Māori, and insufficient rigor in interpreting and generating the data did science and Māori a disservice. A commentary, not new data.

85. 85
    Forzano F, Borry P, Cambon-Thomsen A, Hodgson SV, Tibben A, de Vries P, van El C, Cornel M. Italian appeal court: a genetic predisposition to commit murder?Eur J Hum Genet · 2010 · 18(5):519-21doi:10.1038/ejhg.2010.31 · PMID 20216573

    Commentary by clinical geneticists on the 2009 Italian sentencing decision that cut a murder sentence from nine years to eight, arguing the lack of evidence for clinical utility and the risks of susceptibility testing inside legal proceedings. Opinion, not data.

86. 86
    Berggren U, Eriksson M, Fahlke C, Blennow K, Balldin J. Different effects of smoking or use of smokeless tobacco on platelet MAO-B activity in type 1 alcohol-dependent subjects.Alcohol Alcohol · 2007 · 42(3):267-71doi:10.1093/alcalc/agm031 · PMID 17526638

    Shows the confound directly: cigarette smokers 4.0 plus or minus 1.5 microkatal per kilogram of protein against 5.0 plus or minus 1.9 in non-tobacco users, while smokeless tobacco users were unaffected at 5.1 plus or minus 1.5. The subjects were type 1 alcohol-dependent patients, not the type 2 subgroup the trait claim was about. Small groups: 33 cigarette smokers, 46 non-tobacco users and 14 users of smokeless tobacco, indexed under male subjects only.

87. 87
    Coccini T, Castoldi AF, Gandini C, Randine G, Vittadini G, Baiardi P, Manzo L. Platelet monoamine oxidase B activity as a state marker for alcoholism: trend over time during withdrawal and influence of smoking and gender.Alcohol Alcohol · 2002 · 37(6):566-72doi:10.1093/alcalc/37.6.566 · PMID 12414548

    Source for 6.4 plus or minus 3.1 nanomoles per milligram per hour under the influence of ethanol in 98 patients, rising to 9.3 plus or minus 4.4 after eight days and staying flat thereafter, against 9.9 plus or minus 0.9 in 138 healthy people. Supports a state marker, not a trait marker.

88. 88
    Oreland L. Platelet monoamine oxidase, personality and alcoholism: the rise, fall and resurrection.Neurotoxicology · 2004 · 25(1-2):79-89doi:10.1016/S0161-813X(03)00115-3 · PMID 14697883

    A narrative review by the principal investigator of much of the platelet MAO literature, arguing for a residual true association with personality after the smoking confounder is handled, citing support for the type 2 alcohol dependence link including in non-human primates, and proposing that shared AP-2 family transcription factors may index mitochondrial number or outer membrane surface rather than brain MAO. A review, not pooled data, and not independent of the claim it defends.

89. 89
    Harro J, Oreland L. The role of MAO in personality and drug use.Prog Neuropsychopharmacol Biol Psychiatry · 2016 · 69:101-11doi:10.1016/j.pnpbp.2016.02.013 · PMID 26964906

    Review covering genotype, blood and brain enzyme activity and gene methylation, concluding that the low degree of generalization across studies is likely caused by their cross-sectional design and that male and female findings differ substantially. A review by the same group, with the same caveat.

90. 90
    Herraiz T, Chaparro C. Human monoamine oxidase is inhibited by tobacco smoke: beta-carboline alkaloids act as potent and reversible inhibitors.Biochem Biophys Res Commun · 2005 · 326(2):378-86doi:10.1016/j.bbrc.2004.11.033 · PMID 15582589

    Mainstream smoke inhibited both human isoforms; norharman inhibited MAO-A with a Ki of 1.2 plus or minus 0.18 micromolar and MAO-B at 1.12 plus or minus 0.19 micromolar, harman inhibited MAO-A at 55.54 plus or minus 5.3 nanomolar. In vitro enzymology identifying candidate constituents; it does not prove these account for the imaging findings.

91. 91
    Hernan MA, Takkouche B, Caamano-Isorna F, Gestal-Otero JJ. A meta-analysis of coffee drinking, cigarette smoking, and the risk of Parkinson's disease.Ann Neurol · 2002 · 52(3):276-84doi:10.1002/ana.10277 · PMID 12205639

    44 case-control and 4 cohort studies for smoking: relative risk 0.59 (0.54 to 0.63) for ever smokers, 0.39 (0.32 to 0.47) for current and 0.80 (0.69 to 0.93) for past, with a dose gradient. Observational; the authors call for research on the biological mechanism, which was still open.

92. 92
    Larsson SC, Burgess S. Appraising the causal role of smoking in multiple diseases: A systematic review and meta-analysis of Mendelian randomization studies.EBioMedicine · 2022 · 82:104154doi:10.1016/j.ebiom.2022.104154 · PMID 35816897

    Across 29 published Mendelian randomization studies and 123 new analyses of 57 outcomes, genetic liability to smoking was associated with increased risk of a long list of diseases and decreased risk of only two, Parkinson's disease and prostate cancer. Useful for keeping the protective association in proportion to the harm.

93. 93
    Zhang X, Zhu Z, Zhu L, Guan Y, Zhu Z, Liu B, Ren H, Yang X. Integrating Mendelian Randomization With Single-Cell Sequencing Data Reveals the Causal Effect and Related Mechanisms of Smoking on Parkinson's Disease.Nicotine Tob Res · 2025 · 27(5):884-892doi:10.1093/ntr/ntae180 · PMID 39030896

    Genetic instrumental variable analysis reporting smoking as a protective factor for Parkinson's disease with an odds ratio of 0.288 (p = .008), together with expression work on candidate genes. A single analysis of summary statistics; it addresses the direction of causation, not the mechanism.

94. 94
    Ritz B, Lee PC, Lassen CF, Arah OA. Parkinson disease and smoking revisited: ease of quitting is an early sign of the disease.Neurology · 2014 · 83(16):1396-402doi:10.1212/WNL.0000000000000879 · PMID 25217056

    Case-control study of 1808 patients and 1876 matched controls from Danish registries with lifestyle interviews. Risk was lowest among former smokers who described quitting as extremely difficult or who had used nicotine substitutes, which supports reverse causation. Quitting difficulty was self-reported after diagnosis, which is the study's main weakness.

95. 95
    Kaludercic N, Takimoto E, Nagayama T, Feng N, Lai EW, Bedja D, Chen K, Gabrielson KL, Blakely RD, Shih JC, Pacak K, Kass DA, Di Lisa F, Paolocci N. Monoamine oxidase A-mediated enhanced catabolism of norepinephrine contributes to adverse remodeling and pump failure in hearts with pressure overload.Circ Res · 2010 · 106(1):193-202doi:10.1161/CIRCRESAHA.109.198366 · PMID 19910579

    Mouse and isolated myocyte work: norepinephrine triggered MAO-A activity, reactive oxygen species and maladaptive hypertrophy partly independently of adrenergic receptors; clorgyline prevented the changes in pressure-overloaded mice, and dominant-negative MAO-A mice were protected. Entirely animal and cell data.

96. 96
    Nelson MM, Efird JT, Kew KA, Katunga LA, Monroe TB, Doorn JA, Beatty CN, Shi Q, Akhter SA, Alwair H, Robidoux J, Anderson EJ. Enhanced Catecholamine Flux and Impaired Carbonyl Metabolism Disrupt Cardiac Mitochondrial Oxidative Phosphorylation in Diabetes Patients.Antioxid Redox Signal · 2021 · 35(4):235-251doi:10.1089/ars.2020.8122 · PMID 33066717

    Human atrial myocardium with and without type 2 diabetes: higher MAO-A and MAO-B content and activity associated with body mass index, lower tissue catecholamines consistent with greater flux, more MAO-dependent catecholaldehyde protein adducts, and norepinephrine suppressing mitochondrial ATP production in an MAO-dependent way. Human tissue, observational plus ex vivo manipulation.

97. 97
    Ratiu S, Sturza A, Muntean PS, Borza C, Bratu T, Muntean DM. Monoamine oxidases are mediators of oxidative stress in human varicose veins: interactions with obesity, inflammation, and angiotensin II.Mol Cell Biochem · 2026 · 481(1):297-306doi:10.1007/s11010-025-05398-6 · PMID 41021120

    Twenty patients having varicose veins stripped, randomized by body mass index into obese and non-obese groups: both isoforms expressed in the vein wall and higher in varicose than adjacent perforator veins, MAO-A significantly higher in obese patients with raised C-reactive protein (p < 0.01), angiotensin II raising both isoforms and hydrogen peroxide ex vivo, and 10 micromolar clorgyline or selegiline reducing oxidative stress. The authors call it exploratory.

98. 98
    Lopresti BJ, Stehouwer J, Reese AC, Mason NS, Royse SK, Narendran R, Laymon CM, Lopez OL, Cohen AD, Mathis CA, Villemagne VL. Kinetic modeling of the monoamine oxidase-B radioligand 18F-SMBT-1 in human brain with positron emission tomography.J Cereb Blood Flow Metab · 2024 · 44(11):1262-1276doi:10.1177/0271678X241254679 · PMID 38735059

    Formal kinetic modeling in 14 older adults with arterial sampling: total volume of distribution well identified and correlated at about r = 0.8 with postmortem MAO-B indices, with a 70 to 90 minute ratio to cerebellar gray matter correlating at r = 0.93 with the full analysis. Methodological; a small sample.

99. 99
    Park JH, Ju YH, Choi JW, Song HJ, Jang BK, Woo J, Chun H, Kim HJ, Shin SJ, Yarishkin O, Jo S, Park M, Yeon SK, Kim S, Kim J, Nam MH, Londhe AM, Kim J, Cho SJ, Cho S, Lee C, Hwang SY, Kim SW, Oh SJ, Cho J, Pae AN, Lee CJ, Park KD. Newly developed reversible MAO-B inhibitor circumvents the shortcomings of irreversible inhibitors in Alzheimer's disease.Sci Adv · 2019 · 5(3):eaav0316doi:10.1126/sciadv.aav0316 · PMID 30906861

    Mouse work in the APP/PS1 model: prolonged selegiline failed to lower astrocytic GABA or rescue memory because activity of a compensating GABA-synthesizing enzyme, diamine oxidase, rose, while KDS2010, a reversible inhibitor with a half-maximal inhibitory concentration of 7.6 nanomolar and 12,500-fold selectivity over MAO-A, induced no such compensation. Animal data only.

100. 100
     Boyer EW, Shannon M. The serotonin syndrome.N Engl J Med · 2005 · 352(11):1112-20doi:10.1056/NEJMra041867 · PMID 15784664

     The standard clinical review of the syndrome and of the drug combinations that cause it. No abstract in the PubMed record; it is a review, so it carries no new data.

101. 101
     Dunkley EJ, Isbister GK, Sibbritt D, Dawson AH, Whyte IM. The Hunter Serotonin Toxicity Criteria: simple and accurate diagnostic decision rules for serotonin toxicity.QJM · 2003 · 96(9):635-42doi:10.1093/qjmed/hcg109 · PMID 12925718

     Derived from 2222 serotonergic overdose admissions over 15 years with a learning set of 473: clonus, agitation, diaphoresis, tremor and hyperreflexia, plus hypertonicity and temperature above 38 degrees, giving 84 percent sensitivity and 97 percent specificity against a clinical toxicologist's diagnosis, against 75 and 96 percent for the older Sternbach criteria. Derived in overdose patients, which is not identical to therapeutic-dose interactions.

102. 102
     Wu JB, Shao C, Li X, Li Q, Hu P, Shi C, Li Y, Chen YT, Yin F, Liao CP, Stiles BL, Zhau HE, Shih JC, Chung LWK. Monoamine oxidase A mediates prostate tumorigenesis and cancer metastasis.J Clin Invest · 2014 · 124(7):2891-908doi:10.1172/JCI70982 · PMID 24865426

     MAO-A driving epithelial-to-mesenchymal transition and stabilizing HIF1-alpha through reactive oxygen species, with knockdown reducing or eliminating tumor growth and metastasis in mouse xenografts and pharmacological inhibition reducing xenograft growth; high expression in human tumors correlated with worse outcomes. Cell and mouse causal work with human association.

103. 103
     Bindra S, Pandian R, Mathew B, Anal JMH. Quinoxaline-Based Monoamine Oxidase Inhibitors: Design Strategies, Synthesis, Structure-Activity Relationships, and Therapeutic Potential in Neurological Disorders: A Review From 1996 to 2026.ChemMedChem · 2026 · 21(13):e70330doi:10.1002/cmdc.70330 · PMID 42410707

     A 2026 medicinal chemistry review of one inhibitor scaffold and the structural features that set MAO-A versus MAO-B selectivity, metabolic stability and blood-brain barrier penetration. Cited as evidence that the chemistry is still active; it contains no clinical data.

104. 104
     Jooste MJ, Cloete SJ, Petzer A, Petzer JP, Cloete TT. Evaluation of Novel Isatin-Schiff Base Derivatives: Monoamine Oxidase Inhibition and Antimicrobial Assessment.ChemMedChem · 2026 · 21(16):e70437doi:10.1002/cmdc.70437 · PMID 42603777

     New compounds reaching half-maximal inhibitory concentrations of 0.040 plus or minus 0.002 micromolar against MAO-A and 0.018 plus or minus 0.003 micromolar against MAO-B with a selectivity index of 15,389, inhibiting competitively and reversibly. Test-tube enzymology against recombinant human enzymes; no animal or human data.

105. 105
     Cong S, Tang K, Li J, Gu R, Chen C, Luo J, Wen J, Xu Y, Ke Y, Liu X, Huang Y, Deng Y. Design and synthesis of 6-benzyloxybenzo[c]thiophen-1(3H)-ones as selective monoamine oxidase-B inhibitors with antioxidant and anti-neuroinflammatory activities against Parkinson's disease.Eur J Med Chem · 2026 · 319:119223doi:10.1016/j.ejmech.2026.119223 · PMID 42586021

     A hydrogen-sulfide-releasing MAO-B inhibitor with a half-maximal inhibitory concentration of 0.35 nanomolar and a selectivity index of 6342.8, which by mouth relieved motor deficits and preserved tyrosine hydroxylase positive neurons in MPTP-treated mice. Rodent model work; MPTP toxicity is exactly the situation in which MAO-B inhibition is most expected to help, so it does not predict benefit in human Parkinson's disease.