MAOA uVNTR Genetic Variant and Major Depressive Disorder: A Systematic Review

Major Depressive Disorder (MDD) is a highly prevalent multifactorial psychopathology affected by neurotransmitter levels. Monoamine Oxidase A (MAOA) influences several neural pathways by modulating these levels. This systematic review (per PRISMA protocol and PECOS strategy) endeavors to understand the MAOA uVNTR polymorphism influence on MDD and evaluate its 3R/3R and 3R* genotypic frequencies fluctuation in MDD patients from different populations. We searched the Web of Science, PubMed, Virtual Health Library, and EMBASE databases for eligible original articles that brought data on genotypic frequencies related to the MAOA uVNTR variant in patients with MDD. We excluded studies with incomplete data (including statistical data), reviews, meta-analyses, and abstracts. Initially, we found 43 articles. After removing duplicates and applying the inclusion/exclusion criteria, seven articles remained. The population samples studied were predominantly Asians, with high 3R and 4R allele frequencies. Notably, we observed higher 3R/3R (female) and 3R* (male) genotype frequencies in the healthy control groups and higher 4R/4R (female) and 4R* (male) genotype frequencies in the MDD groups in the majority of different populations. Despite some similarities in the articles analyzed, there is still no consensus on the MAOA uVNTR variant’s role in MDD pathogenesis.


Introduction
Major Depressive Disorder (MDD) is a psychiatric condition marked by various mood changes (e.g., sad, empty, or irritable) and reduced interest and pleasure for at least two weeks, among other varying symptoms. Considered one of the most prevalent mood disorders [1,2], the intensity of symptoms can be disabling with clinically significant suffering [3]. The high rates of first-line treatment remission (30-40%) have patients often undergoing multiple subsequent courses of antidepressants or augmentation strategies, making MDD a formidable public health challenge for front-line clinicians and researchers [4].
Worldwide more than 350 million people were estimated to suffer from MDD in 2017 [3,5], and by the year 2030, MDD should become the highest global burden of disease, according to the World Health Organization (WHO) [1,6,7]. This burden would represent not only a critical compromise of the affected population's quality of life but also a significant economic impact-even without not considering the impact of the COVID-19 pandemic on the disorder. The COVID-19 pandemic triggered a 25% increase in depression
For this, we included observational or interventional studies that presented data on the MAOA uVNTR's variant genotypic and allelic frequencies in human research participants with MDD and described their laboratory methods according to our eligibility criteria. However, studies with incomplete data (including statistical data), reviews, meta-analyses, abstracts, and studies not in English, Spanish, or Portuguese were excluded.
We searched, on the 11th of May 2022, employing the databases Web of Science, PubMed, Virtual Health Library (BVS), and EMBASE with no adopted filters, including the year of publication of the articles. The indexed terms (descriptors) researched were "MAOA uVNTR," "Depressive Disorder, Major," and "Polymorphism, Genetic," the last two as described by the Medical Subject Headings (MeSH) vocabulary thesaurus, combined by the boolean operator "AND."

Study Selection and Data Extraction
Two reviewers (CF and AG) collaborated on the article selection in two phases. Each reviewer in the first phase independently analyzed each article's title and abstract, verifying their eligibility according to the PECOS strategy. The Rayyan tool, developed by the Qatar Computing Research Institute (QCRI), was used to assist this initial analysis, while Mendeley Desktop version 1.19.4 software helped remove duplicates. In the second phase, the same two reviewers (CF and AG) independently performed the full-text analysis of the pre-selected articles, always observing the pre-established eligibility criteria. For this, Mendeley Desktop software version 1.19.4 was also used.
In both phases, disagreements or doubts were discussed between the two reviewers, and a third reviewer (CS) was consulted if a disagreement was unresolved. The two reviewers (CF and AG) independently extracted pre-defined data into an electronic spreadsheet in Microsoft Office Excel; these were: author, study title, objective, year of publication, the country in which the study was carried out, variants studied, 3R/3R and 3R* genotypic frequencies, sample size, laboratory methodology, main result, and p-value. Relevant or other up-to-date original publications in the field have also been included in the introduction and discussion section.

Bias Risk in Each Study
Risk models are usually based on either examining genetic variants or analyzing genetic and environmental risk factors. We examined the selected articles' bias risk using the Genetic Risk Prediction Studies (GRIPS) Guideline [28,29], comprised of 25 items, to rate the articles' quality. This review only considered 20 items (verifying the item's presence or absence) when evaluating the studies' quality. We considered that an article was of good quality if it presented at least 75% of the items.
This step was performed independently by two reviewers (CF and AG). Any disagreements were resolved after discussing with the third reviewer (IS).

Articles' Search, Selection, and Quality Assessment
Initially, we identified 43 articles by searching four databases. After removing duplicates, 20 studies were selected for the title and abstract analysis, observing the aspects delimited by PECOS. Among these, 11 were selected for full-text analysis. After considering previously established inclusion and exclusion criteria, seven articles were eligible for inclusion in this systematic review ( Figure 1, Table 1).
Initially, we identified 43 articles by searching four databases. After removing duplicates, 20 studies were selected for the title and abstract analysis, observing the aspects delimited by PECOS. Among these, 11 were selected for full-text analysis. After considering previously established inclusion and exclusion criteria, seven articles were eligible for inclusion in this systematic review ( Figure 1, Table 1).

Figure 1.
Flowchart outlining the steps adopted in the articles' selection.   Articles that did not meet the eligibility criteria (the PECOS strategy) were excluded, and their reason for exclusion is described in Table S1. Table S2 presents the results of the selected articles' bias risk analysis and quality determination using the Genetic Risk Prediction Studies (GRIPS) Guideline (20 items and two sub-items of 25 items). Of articles analyzed in this study, 57.1% (4) had at least 75% (16.5) or more items and were considered of good quality [26,27,30,33]. The lowest score was 68.2% (15 items) [36] Table S3 summarizes the selected articles' selection criteria for MDD and control groups, the participants' ethnicity/race, and the articles' employed statistical analysis.

Selected Studies' General Characteristics
As shown in Table 2, most studies were conducted on the Asian continent, mainly in Taiwan [27,30], followed by the American and European continents. Regarding demographic data, most studies were conducted with adults over 18 years old, and females were more frequent within the MDD patient groups.

MAOA uVNTR Genotypic Frequency
In general, the MAOA uVNTR variant's high-activity 4R allele had the highest frequency in the populations studied, followed by the low-activity 3R allele. While the 2R, 3.5R, 5R, and 6R alleles, in most studies, were considered rare, being found in few patients or not found at all [23,26,27,30,33,34]. This finding confirms Sabol et al.'s [17] report that the 3R and 4R alleles are the most common among distinct ethnic populations. Taking into account the 3R/3R (female homozygote) and 3R* (male hemizygote) genotype frequencies reported by the articles included in this systematic review, we noted that in all studies, their individual and combined frequencies were lower than 60% in populations diagnosed with MDD [23,26,27,30,33,34] (Tables 1 and 2).
Du et al. [23] found a combined 3R/3R and 3R* genotypic frequency of 37.2%, similar to Yu et al. [33] who estimated 35.1%, while Lung et al. [27], in turn, found a 41.2% frequency and Sanabrais-Jiménez et al. [36] a 26.6% frequency. The highest frequency estimated was 47.3% by Huang et al. [30], in contrast, the lowest frequency was 13.1%, estimated by Rivera et al. [26]. Won et al. [34] found a 41.93% 3R/3R frequency in the female MDD patients they analyzed (Tables 1 and 2). The lower percentage of the 3R/3R and 3R* genotypes in the majority of MDD groups might indirectly indicate a protective role against MDD, depending on the population (Tables 1 and 2).

MAOA uVNTR and Its Genotypic Frequency in Major Depressive Disorder (MDD)
Different factors can influence genotypic frequency in different populations and study groups. Among them, we highlight population stratification and ethnicity, sample size, different symptom detection and diagnosis methodologies applied to case and control groups and also between different studies, phenotype definition (case group), unscreened control groups, multiallelic gene organization, and variations related to sex, e.g., the MAOA gene is located on the X chromosome (X-inactivation process also called lyonization and males only have one allele), or even hormonal effects [26,30,33,34]. Some might explain the significant variation in MAOA uVNTR polymorphism genotype frequencies between the analyzed populations. In the studies conducted in an Asian population, female patients diagnosed with Major Depressive Disorder (MDD) and a 3R/3R (female homozygote) genotype had frequencies that varied between 25 and 42%, while in male MDD patients with a 3R* (male hemizygote) genotype, the frequency varied from 48 to 58% (Tables 1 and 2). In contrast, in Caucasians/Hispanics/Latino, these frequency values ranged between 11 and 35% and from 26 to 44%, respectively (Tables 1 and 2).
In Rivera et al.'s [26] study, the combined 3R/3R and 3R* genotype frequency in a Spanish population with MDD was estimated at 13.1%, similar to another Spanish study with MDD patients (16.8%) [25]. Notably, the Rivera et al. [26] study used two instruments for diagnosis, the Diagnostic and Statistical Manual of Mental Disorders-Fourth Edition (DSM-IV) and the Composite International Diagnostic Interview (CIDI) per the International Classification of Diseases-10 th edition (ICD-10), with separate analyses for the different phenotypes (ICD-10 depressive episode; ICD-10 severe depressive episode and DSM-IV major depression, MDD). Furthermore, the authors grouped the MAOA uVNTR variants' high-activity 3.5R and 4R alleles, the 5R allele, and their combinations (homozygous and heterozygous genotypes) as high-activity alleles and genotypes, respectively [26]. Du et al. [23] also grouped these alleles into a single allele group for analysis against the low-activity allele 3R. Although 5R is rare, it is controversial regarding its transcriptional efficiency (see the Introduction), so subdivisions, such as these might contribute to a sample reduction for each analysis, influencing the frequency estimates [23][24][25][26]40]. The 3R/3R and 3R* genotypes frequency obtained by Du et al. [23] in a Canadian MDD patient sample (Hamilton Depression Rating Scale, HAM-D: males, 22.9 ± 3.5; females, 23.6 ± 3.2) was 37.2%, close to the 40% reported by Schulze et al. [40] for recurrent depression (DSM-IV) in a German population.
Most control groups had the 3R/3R and 3R* genotype frequencies higher than those of the MDD groups (Tables 1 and 2), a fact that may represent an intriguing finding as the 3R allele presence may represent a protective factor for MDD. This protection might be due to the lower MAOA transcriptional efficiency [17,41] and enzymatic activity [18] that the 3R allele confers, contributing to the maintenance of central serotonin levels [33].
Du et al. [23] investigated MAOA (EcoRV and uVNTR) polymorphisms' connection with MDD and depressive symptoms and found that MAOA uVNTR did not associate with MDD, although it trended towards significance in male patients (p = 0.055). Du et al. [23] did find a highly significant linkage disequilibrium (a nonrandom association of alleles of different loci) between MAOA's uVNTR and EcoRV variants (D' controls = −0.79; p < 0.0001; D' patients = 0.84; p < 0.0001). They also found that the EcoRV allele with the EcoRV site present and MAOA uVNTR's 3R allele significantly correlated with depression in males better than any of the polymorphic alleles alone (p = 0.008; OR = 2.5, OR CI = 1.3-4.8) and with higher insomnia scores compared to other haplotype carriers (HAM-D clusters compared by unpaired Student t-test, t = 2.7, p = 0.008), even after correction for multiple testing (p = 0.048 in both cases) [23]. Huang et al. [30] also found a strong linkage disequilibrium between these MAOA variants but found no significant differences in the haplotype frequencies in total MDD patients (DSM-IV major depression; HAM-D ≥ 18) or MDD clinical subgroups (18 ≤ HAM-D ≤ 24: moderate MDD; HAM-D > 24: severe MDD; MDD with and without a first-degree family member with a history of MDD or bipolar disorder) versus controls (p > 0.05, data not shown).
Rivera et al. [26] found an association between high-activity alleles (3.5R; 4R; and the supposedly 5R; see the Introduction) and MDD in female Spanish patients (p = 0.048), similarly to Huang et al. [30]; who found a weak association between severe MDD group and its genotypic frequencies among female Taiwanese participants (p = 0.041). However, like with Lung et al. [27]; the association was not maintained after multiple logistic regression analyses; suggesting that the MAOA uVNTR variant may not play a central role in the risk of developing MDD [30]. Yu et al. [33] also assessed the association between the MAOA uVNTR polymorphism and MDD (DSM-IV major depression; HAM-D-1967 ≥ 18) and found that the 4R allele was more common in MDD women (X 2 = 6.93; df = 1; p = 0.009) and MDD men (X 2 = 6.27; df = 1; p = 0.015) compared to the control group (mostly composed of medical staff) [33]. These findings suggest a non-gender-specific effect on MDD given the increased 4R allele presence in the MDD group (X 2 = 12.48; df = 1; p < 0.001), even considering X-inactivation at the MAOA locus (X 2 = 10.7,2; df = 1; p = 0.001) [33].
Like with the Brazilian population; there is a lack of studies analyzing the MAOA uVNTR variant effect on patients diagnosed with MDD. Other studies trying to relate this variant to other monoamine-metabolism-affected neurological/psychopathological disorders: such as aggression and antisocial behavior; mood disorders; schizophrenia; autism spectrum disorders; substance use disorders; and even Alzheimer's disease; were found in the literature [43][44][45]. For instance: the high-activity alleles' presence correlated with clinical improvement of opposing symptoms in male children and adolescents with Attention Deficit Hyperactivity Disorder using methylphenidate (p < 0.001) [46]; the occurrence of bilateral tonic-clonic seizures in patients with epilepsy (p = 0.032) [47]; and increased consumption of lipid-dense foods in preschool children (p = 0.009) [48]. At the same time: the low-activity allele correlated with an early-age onset of alcohol dependence (p = 0.01) and a more considerable amount of antisocial personality symptoms after age 15 (p = 0.02) [49].
Such associations and those with MDD seem to indicate this polymorphism's influence on monoaminergic metabolic pathways. The WHO estimates a 5.8% prevalence of depressive disorders in Brazil [1], a country composed of a population with a heterogeneous genetic makeup-Native Americans, Europeans, and Africans [43]. This heterogeneity highlights the need for further studies evaluating the behavior of the MAOA uVNTR variant's genotypic distribution and allelic frequencies in multiracial populations, together with more single ethnicity populations studies, to fill the knowledge gaps and build better risk prediction models.

MAOA uVNTR Genetic Variant and Cortical Thickness
Neuroimaging studies are essential when analyzing mood disorders as they help identify neuroanatomical alterations in brain regions brought on by the disorder, its treatment, and other factors [4,34]. Suh et al.'s [4] systematic review and meta-analysis analyzed neurobiological differences between MDD (medicated and medication-naïve) patients (n = 1073) and healthy controls (n = 936). They found that MDD patients have a significant cortical thinning in the bilateral medial orbitofrontal cortex (OFC), left pars opercularis, and left calcarine fissure/lingual gyrus, as well as a significant thickening in the left supramarginal gyrus compared to healthy controls. However, it is still unclear what factors may contribute to these alterations in MDD [4,34].
The OFC seems to regulate mood, reward-guided behavior, and impulse control; thus, damage to this structure might create deficits in an individual's emotional and social regulation and mood lability [34]. With the majority of imaging genetics studies conducted on MAOA uVNTR polymorphism reporting its genotype relevance to OFC structure in healthy controls, Won et al. [34]  Although these results suggest that OFC structural alterations are associated more with MDD pathophysiology than MAOA uVNTR polymorphism, this lack of correlation might be due to the samples' composition, as MAOA uVNTR genotype-dependent OFC structural changes have usually been observed in males but not females [43,[50][51][52][53][54].
Tzeng et al. [56], in turn, investigated the association between MAOA uVNTR polymorphism's genotypes and their response to the Mirtazapine, an atypical antidepressant belonging to the tetracyclic piperazine-azepine class, in a Taiwanese sample and found an association between medication use and treatment variations according to the genotype. In their study, 3R allele carriers had a better therapeutic response and higher remission rates after a 7-week Mirtazapine treatment. Noticeably, 3R allele carriers also had a lower dropout rate; however, the sample size was small (n = 58), which might generate a bias [56].

Quality Assessment and Limitations of Selected Articles
Genomic-wide association studies investigate the association of genotypic variations and characteristics expressed by individuals in different populations [57], including susceptibility/resistance to certain diseases and disorders. These studies have great applicability; among them is the implementation of these possible connections to construct analytical designs/models to predict the genetic risk of several diseases [58]. Studies to better understand genetic and environmental factors are increasingly important in scientific areas, specifically health. In this sense, the Genetic Risk Prediction Studies (GRIPS) guideline ensures a higher methodological quality of a study, enabling the minimization of biases that can compromise the interpretation of results [28,29].
We employed the GRIPS guideline, composed of 25 items, to assess the quality of the articles selected to integrate this systematic review (Table S2). Criteria chosen to analyze the quality of methods section were ten items (two sub-items), results: seven items, and discussion: three items, totaling 20 items (and two sub-items). We considered the study quality adequate if it had 75% (15) of the items, and all articles included met this criterion. We considered the study quality adequate if it had at least 75% (15) of the items, and all articles included in this review met this criterion. All articles [23,26,27,30,33,34,36] described the participants' eligibility criteria and sources and methods of selecting participants. They also made a generalized discussion and, when pertinent, revealed the importance of their study. Nevertheless, for instance, 71.4% [23,27,33,34,36] examined their limitations, one (14,3%) did not report their population's demographic and clinical characteristics [30], and none specified how they handled missing data in the analysis.
High-quality studies must be prepared transparently to facilitate the interpretation of the results' validity and assess their relevance. Observations in different populations are essential to generalize the research and help minimize the existing heterogeneity factor [59]. In this way, these studies allow the information to be extrapolated to different contexts and help construct an integrative data analysis model for eventual practical application [28].

Conclusions
Although some evidence suggests that the MAOA uVNTR variant's alleles are related to MDD manifestation, there is still no concrete accord on this association. In general, the MAOA uVNTR variant's 3R and 4R alleles had the highest frequency in the populations studied, whereas the 2R, 3.5R, 5R, and 6R alleles, in most studies, were considered rare or not found at all. Interestingly, the low-activity allele 3R presented a higher frequency in most control groups, and the reverse was true for the 4R allele in the studies when not grouped with other alleles. Nevertheless, the number of retrieved studies that met our inclusion criteria was small, representing a limiting factor in this systematic review. The different allele groupings used in some of the articles' analyses also increased the difficulty in interpretation as some analyzed 3R/3R and 3R* genotype and allelic frequencies against 4R/4R and 4R*, while others grouped low-activity genotypes and alleles (3R) against high-activity ones (3.5R and 4R, that sometimes included 5R). Another limitation was the few represented populations.
Depression is a heterogeneous and multifactorial disease with both genetic and environmental factors influencing its development. Sex-specific effects and possible interactions with other polymorphisms and genes (e.g., linkage disequilibrium), as well as epigenetic and environmental factors, must be taken into account to understand this polymorphism influence on MDD better, thus contributing to the elucidation of the mechanisms related to this disorder-further studies on a larger scale and in other populations will help fill these gaps.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/cells11203267/s1, Table S1: Articles excluded according to criteria related to the defined PECOS strategy; Table S2: Quality Evaluation of articles according to the adapted GRIPS guideline; Table S3: The selected articles' selection criteria (MDD and control groups), ethnicity/race, and employed statistical analysis per their original authors, with few modifications.

Data Availability Statement:
No new data were created in this study, and the data presented in this review are available in its tables (Table 1, Tables S1 and S2) and the reviewed original articles.