Introduction Psychiatric disorders

Psychiatric disorders can affect any aspect of mental health, including cognition, emotionality, and sociability. They are a leading cause of disability worldwide and represent a significant economic burden to the society (Bloom et al., 2012; Vigo et al., 2016). The aim of this research using animal and cellular models is to better understand the neurobiology of psychiatric disorders to develop new treatments for the individuals that experience them. In this review, we discussed models for two common overlapping psychiatric disorders: attention-deficit/hyperactivity disorder (ADHD) and autism spectrum disorder (ASD).

Attention-deficit/hyperactivity disorder and autism spectrum disorder

ADHD and ASD are two of the most widely studied neurodevelopmental disorders. According to the Diagnostic and Statistical Manual of Mental Disorders, 5th ed. (DSM-5), ADHD is a clinically heterogeneous disorder characterized by inattention, hyperactivity, and increased impulsivity (American Psychiatric Association, 2013). Approximately 2–10% of school-age children are affected by ADHD (Hawi et al., 2015), and males are more likely to be affected than females (Polanczyk et al., 2007; Nussbaum, 2012). Despite traditionally being regarded as a childhood disorder, a number of longitudinal-based and population-based epidemiological studies have identified an adult form of ADHD (Franke et al., 2012, 2018). ASD is a group of neurodevelopmental disorders characterized by impairments in learning, sensory filtering, communication, and social interactions together with restricted, repetitive, and stereotypical behaviors and interests. The prevalence of ASD in the population is estimated at ~1%, with three to four times as many males as females affected (Baird et al., 2006). There are currently no objective laboratory-based tests to diagnose ADHD or ASD. The latest version of the DSM acknowledges the high rate of comorbidity for ASD and ADHD, allowing their simultaneous diagnosis for the first time (American Psychiatric Association, 2013). The symptoms of ADHD and ASD often co-exist. Overall, 20–50% of children with ADHD also meet the criteria for ASD (Rommelse et al., 2011). Several studies have shown social deficits, peer relationship deficits, and empathy problems in children with ADHD, and the DSM-5 allows a comorbid diagnosis of both disorders. Although ADHD and ASD comorbidity has mainly been studied in children, there is also evidence that it occurs in adults (Hartman et al., 2016). The drugs most commonly used to treat ADHD are stimulants such as methylphenidate (MPH). Although effective in controlling some symptoms, MPH treatment shows a large degree of variability across patients (Wolraich and Doffing, 2004). There are currently no pharmacological treatments for the core symptoms of ASD. Behavioral therapy can improve the quality of life for patients with ASD, and pharmacological treatments may reduce secondary symptoms such as hyperactivity (LeClerc and Easley, 2015).

Genetic and environmental factors for autism spectrum disorder and attention-deficit/hyperactivity disorder

A combination of both environmental factors and genetic variants are known to influence the development and expression of ADHD and ASD (Sánchez‐Mora et al., 2015; Sjaarda et al., 2017). Meta-analysis of multiple large-scale twin studies estimated the heritability of childhood and adolescent ADHD at between 0.7 and 0.8 (Nikolas and Burt, 2010). The heritability of adult ADHD is reported to be lower, at between 0.3 and 0.4, although this may be an underestimation owing to the self-rating system commonly used in these studies (Franke et al., 2018). The true heritability could be as high as 0.7–0.8 (Brikell et al., 2015; Tick et al., 2015). ASD heritability estimates are similarly high at 0.64–0.91 (Tick et al., 2015). Familial co-aggregation of ADHD and ASD is strongly indicative of an overlap in genetic factors across disorders (Ghirardi et al., 2018). Furthermore, a number of studies have identified genes that are associated with both ADHD and ASD, including contactin-associated protein-like 2 (CNTNAP2) (Strauss et al., 2006; Elia et al., 2010; Poot et al., 2010; Rodenas-Cuadrado et al., 2014).

The genetic basis of ADHD and ASD can be explained by the common disease common variant and common disease rare variant (CDRV) hypotheses (Hawi et al., 2015). However, there is a growing recognition that most individuals with ADHD or ASD carry both common and rare variants. The common disease common variant hypothesis focuses on multiple common polymorphisms, with a frequency greater than 5%. Each polymorphism has a low level of penetrance, but when combined they can increase the risk of developing ADHD or ASD. This hypothesis has formed the basis of molecular genetic research over the past two decades starting with the investigation of preselected candidate genes in relatively small studies (Li et al., 2014). With the advent of high-throughput genetic screening, researchers could conduct hypothesis-free experiments in large samples of patients and controls. These genome-wide association studies have identified a number of polymorphisms, which confer a slightly increased risk of developing a psychiatric disorder. Common variants have been estimated to explain 22% of ADHD heritability and 32% of ASD heritability (Autism Spectrum Disorders Working Group of The Psychiatric Genomics Consortium, 2017; Demontis et al., 2017). Some of the missing heritability can be explained by the CDRV hypothesis. ADHD and ASD might be caused by genetic variants that are rare in the overall population but have high penetrance in individuals. Rare variants such as monogenic or chromosomal abnormalities are a common cause of syndromic ASD, and in some cases, these abnormalities can be identified by genetic testing (Fernandez and Scherer, 2017). Nonsyndromic ASD might also be caused by monogenic mutations in genes such as Neurexin 1 (NXRN1) and SH3 and multiple ankyrin repeat domains (SHANK3) (Caglayan, 2010). Another type of rare variant that has been intensely studied is copy-number variations (CNVs). CNVs are chromosomal deletions or duplications that can span several megabases and which alter normal gene expression (Redon et al., 2006). The relevance of CNVs to a particular phenotype can be linked to their position, length, and heterozygosity. There is increasing evidence that CNVs play a role in the pathogenesis of several neurodevelopmental disorders, and it has been demonstrated that patients with ASD have a higher burden of CNVs compared with healthy controls (Shishido et al., 2013). Testing for CNVs in patients with ADHD has identified a number of genes that had previously been implicated in ASD, demonstrating the shared genetic risk for these disorders (Lionel et al., 2011). Other studies have also reported a greater burden of CNVs in patients with ADHD without reporting comorbidity with ASD (Williams et al., 2010, 2012). These CNVs include the parkin 2 (Park2) and neuropeptide Y loci (Lesch et al., 2011; Jarick et al., 2014) and genes in the metabotropic glutamate receptor family (Elia et al., 2012). There is some evidence that CNVs may contribute to adult ADHD as well (Ramos-Quiroga et al., 2014), although different genes may be affected. The growing consensus among researchers is that ASD and ADHD are caused by a number of synergistic factors including CNVs, rare variants, and common polymorphisms (Kiser et al., 2015).

Environmental factors including epigenetic modifications have also been linked to these disorders. The prenatal and perinatal stages appear to be particularly susceptible to environmental stimuli. Both premature birth and maternal smoking during pregnancy have been linked to an increased risk of developing ADHD (Halmøy et al., 2012; Zhu et al., 2014). However, when genetic and familial confounders are included, the strength of these associations decreases, calling into question any causal relationship (Sciberras et al., 2017). A further meta-analysis of 15 studies failed to find an association between maternal smoking and ASD (Rosen et al., 2014). Even within the prenatal period, the effect of environmental stimuli depends upon the trimester during which exposure occurs. Exposure to antidepressants during the second and third trimesters has been associated with an increased risk of developing ASD (Boukhris et al., 2016). However, a cohort study of more than 1.5 million Swedish children failed to find an association between maternal antidepressant use during the first trimester and risk of either disease (Sujan et al., 2017). In the perinatal period, strong associations between ASD and birth complications leading to trauma, ischemia or hypoxia have been reported (Modabbernia et al., 2017). Both prenatal and postnatal exposures to abnormal levels of metal toxicants and essential elements have been associated with ASD (Arora et al., 2017). After the antenatal period, a number of early life factors have been linked to these disorders. These include an association between maternal anxiety, depression, and ADHD in 3-year olds (Meadows et al., 2007) and increased parental age and ASD risk (Wu et al., 2016). However, whether causality between environmental factors and neurodevelopmental disorders exists can be hard to determine, and further research, including the use of animal models, is required to investigate this issue.

Cross-species models

The use of animal and cellular models to understand the pathogenesis of human disease and develop treatments has greatly contributed to our understanding of several psychiatric disorders. These models are suitable for large-scale trials that can be translated into human clinical research (Crawley, 2012; van der Voet et al., 2014; Falk et al., 2016; Kim et al., 2016; Ardhanareeswaran et al., 2017).

Mouse (Mus musculus)

The mouse is the most commonly used model organism in biomedicine. Their short generation time, small size, and social nature allow a large number of animals to be housed in a small facility. The sequencing of the mouse genome combined with the tools to manipulate it in a targeted manner has revolutionized biomedical research (Hara and Takada, 2017). As mice and humans are very similar genetically (∼99% homology), it is possible to investigate genes identified in human studies in mice. Following genetic manipulation, a combination of highly standardized behavioral and neurobiological techniques can be performed. This can provide insights into how polymorphisms or mutations can lead to the pathology of psychiatric disorders (Arguello and Gogos, 2006). Of particular relevance to psychiatric disorders, mice exhibit a range of complex behaviors including reciprocal social interactions (McCammon and Sive, 2015). Mice are often the first step in testing a new treatment for toxicity and efficacy (Vandamme, 2014).

Zebrafish (Danio rerio)

The zebrafish is an emerging model organism for biomedical research thanks to their short generation time, small size, and genetic tractability. Zebrafish are particularly useful for developmental and live imaging studies, as they develop externally and are transparent at larval stages. Almost 70% of human genes have at least one zebrafish orthologue (Howe et al., 2013). Transient knockdown of genes in zebrafish was initially achieved by injecting morpholino oligonucleotides (Blum et al., 2015). Recently, stable mutant lines have been created using CRISPR/Cas9 genome editing (Schmid and Haass, 2013). However, modelling human disorders in zebrafish is complicated by an additional whole-genome duplication which occurred ~440 million years ago (Amores et al., 2011). Despite this issue, zebrafish are uniquely suited for high-throughput pharmacological studies, as larvae can be generated in large numbers and drugs can be applied by dilution in the embryo medium (Rihel et al., 2010; Norton, 2013; Gutiérrez et al., 2018). Zebrafish can be used to study ADHD and ASD thanks to their stereotypical behavior and social interactions. It is also possible to test cognitive deficits in this species. Spatial working memory can be measured in the habituation to novelty test, and various forms of learning can be assessed using robust behavioral paradigms (Stewart et al., 2014).

Fruit fly (Drosophila melanogaster)

The fruit fly is a widely used model system that has greatly contributed to neuroscience research (Bier, 2005). Fruit flies are small and easy to maintain with a very short generation time, which allows large-scale low-cost experiments. Although the fruit fly brain is relatively small, it is still complex enough to model some aspects of psychiatric disorders (Bellen et al., 2010). Signalling pathways and regulatory molecular networks are also well conserved. Overall, 75% of human disease genes have related sequences in the fly (Bier, 2005). Furthermore, the temporally inducible UAS-GAL4 system can be used to target specific cell types in Drosophila (Brand and Perrimon, 1993). When this system is coupled to the expression of RNA interference molecules, it allows reversible genetic manipulation. Transgenic flies can also be generated using the P-element system (Rubin and Spradling, 1982), and CRISPR/CAS9 technology now permits ubiquitous or tissue-specific genome editing (Port et al., 2014). Drosophila display a behavioral repertoire that varies from simple aggregation to more complex behaviors such as aggression and courtship (O’Kane, 2011). Therefore, even though the nervous system and social behavior of Drosophila are not that similar to vertebrates, the fruit fly presents a valuable model to characterize mechanisms underlying disease-related behaviors (van Alphen and van Swinderen, 2013; van der Voet et al., 2014).

Human induced pluripotent stem cells

The recent development and improvement of human induced pluripotent stem cell (hiPSC) technology permits the conversion of tissue into pluripotent cells that can be redifferentiated into neurons in a noninvasive and ethical manner (Takahashi et al., 2007). This technique offers the possibility to use a range of different samples including skin biopsies, urine samples, and hair follicles. A great advantage of this method is that the genetic makeup of patients is included in the experimental model rather than investigating the influence of single genes. This may be particularly important for multifactorial diseases such as ADHD and ASD. hiPSCs also represent an innovative tool for large-scale drug screening. Moreover, neurons obtained from hiPSCs can be cultured in two dimensions, for single neuron analysis, or in three-dimensional aggregates (brain organoids), allowing evaluation of network-level effects (Falk et al., 2016; Frega et al., 2017; Prytkova and Brennand, 2017). In recent years, hiPSCs have been used as a tool to model psychiatric disorders such as ASD, ADHD, bipolar disorder, Schizophrenia, and the response to lithium at a cellular level (Prilutsky et al., 2014; Ardhanareeswaran et al., 2015; Lim et al., 2015; Soliman et al., 2017; Hoffman et al., 2018; Jansch et al., 2018).

A single model cannot capture all the traits of a complex brain disorder such as ASD and ADHD. As specific behavioral and neurobiological features may be more easily studied in certain organisms, a comparative study across models should be performed whenever possible (Stewart et al., 2015; Wong and Josselyn, 2016).

Contactin-associated protein-like 2, adhesion G protein-coupled receptor L3, and parkin 2

Human genetic studies have identified a large number of genes linked to ADHD and ASD. There are currently 359 genes implicated in ADHD according to the ADHDGene database, whereas 1007 genes are included in the SFARI Gene database for ASD (Zhang et al., 2011; Abrahams et al., 2013). In this review, we have chosen to focus on three candidate genes for ADHD and ASD: CNTNAP2, adhesion G protein-coupled receptor L3 (ADGRL3), and PARK2. As well as being involved in diverse biological processes, the types of variants linking these genes to disease are also different. CNTNAP2 was initially implicated in ASD through a recessive single base pair deletion in the coding region. In contrast, the association between ADGRL3 and ADHD comes from noncoding single nucleotide polymorphisms (SNPs), which may affect gene expression levels (Strauss et al., 2006; Martinez et al., 2016). Larger scale CNVs in PARK2 are associated with ASD (Glessner et al., 2009; Scheuerle and Wilson, 2011). CNTNAP2 and PARK2 are primarily implicated in ASD but have also been linked with ADHD. ADGRL3 is linked with ADHD and substance use disorders (SUD), but there is currently no strong connection to ASD (Gau et al., 2012). Together this review provides an overview of how genes with different biological functions can be studied in model systems to understand the comorbidity of ADHD and ASD.

Contactin-associated protein-like 2 Biological and molecular description

CNTNAP2 is one of the largest human genes. It spans 2.3 Mb and is located in chromosomal region 7q35. CNTNAP2 encodes the neuronal transmembrane protein CASPR2, a member of the neurexin (NRXN) superfamily of single-pass transmembrane proteins. CASPR2 is composed of 1331 amino-acid residues organized into eight extracellular domains, two C-terminal intracellular domains, and one transmembrane domain (Poot, 2015). The intracellular domain of CASPR2 may be involved in protein–protein interactions, whereas the extracellular domains mediate cell–cell interactions and bind ligands, receptors, and extracellular matrix components (Poot, 2015; Baig et al., 2017). CNTNAP2 forms a neuron–glia adhesion complex with contactin 2 that localizes voltage-gated potassium channels to the juxtaparanodal region of myelinated axons (Poliak et al., 1999, 2003). The first link between CNTNAP2 and ASD was a recessive mutation found in an Amish family affected by a syndromic form of ASD called cortical dysplasia-focal epilepsy syndrome (Strauss et al., 2006). Several more studies have provided convergent evidence that rare and common variations in CNTNAP2 confer risk for ASD-related endophenotypes such as developmental language disorders and ASD itself (Peñagarikano and Geschwind, 2012). CNTNAP2 has also been linked to schizophrenia, epilepsy, intellectual disability, learning disability, and ADHD (Rodenas-Cuadrado et al., 2014). However, most studies have focused on the role of CNTNAP2 in ASD.

In this review, we have chosen to focus on these three candidate genes because they are linked to either ADHD (ADGRL3), ASD (CNTNAP2), or both disorders (PARK2). Each gene has been modelled in a number of different species, making these candidates ideal to demonstrate the power of comparative studies to help translate findings to human patients.

Models Mouse

The first Cntnap2 mutant mouse was generated in 2003 (Poliak et al., 2003). After the identification of CNTNAP2 as a candidate gene for ASD, the original strain was backcrossed onto the C57BL/6J background, which is more suitable for behavioral characterization (Crawley et al., 1997; Penagarikano et al., 2011). Cntnap2 KO mice display deficits in core behavioral features of ASD: social deficits, restrictive and repetitive behaviors, and reduced vocal communication (Table 1). When compared with wild-type littermates, these mice appear normal, and there are no differences in growth rate and final size. Interestingly, after six months, Cntnap2−/− mice develop epileptic seizures and their electroencephalogram has an abnormal pattern during slow wave sleep. In addition, knockout mice are hyperactive (Penagarikano et al., 2011). Cntnap2−/− mice show neuronal migration abnormalities that reveal the fundamental role of this gene in cortical projection neuron migration (Penagarikano et al., 2011). Interestingly, it has been found that Cntnap2−/− mice display a reduced number of GABAergic interneurons (Penagarikano et al., 2011). Pharmacological treatment with risperidone rescued the stereotypic behavior of Cntnap2−/− mice, whereas social deficits were unaffected, similar to the pattern of improvements seen in human patients (Penagarikano et al., 2011). Overall, this model recapitulates certain features of patients with CNTNAP2 mutations, and this is one of the best studied genes for this disorder (Poot, 2015).

Table 1:

Mouse models

Zebrafish

A zebrafish cntnap2 mutant model was recently created (Table 2). Cntnap2a−/− and cntnap2b−/− double mutants (cntnap2ab) harbor homozygous loss-of-function mutations in both genes resulting in a complete loss of the protein (Hoffman et al., 2016). Hoffman et al. (2016) analyzed excitatory and inhibitory neuron populations in cntnap2ab during early development, as an imbalance in inhibitory and excitatory signalling could be involved in ASD (Rubenstein and Merzenich, 2003). They observed that mutants have a smaller head than wild type. Moreover, mutants show a decrease in GABAergic cell count; however, there were no prominent deficits in glutamatergic neurons (Hoffman et al., 2016). Loss of inhibitory neurons can increase seizure susceptibility (Cobos et al., 2005). Zebrafish cntnap2ab mutants are more sensitive to drug-induced seizures and show night-time hyperactivity, another phenotype linked to the imbalance between excitatory and inhibitory neurotransmission. The night-time hyperactivity phenotype can be rescued by estrogens (Hoffman et al., 2016). This finding has reinforced the use of whole organism drug screening as an effective system to identify compounds that can be developed into pharmaceutical treatments.

Table 2:

Fish models

Fruit fly

CNTNAP2 and NRXN1 are members of the neurexin superfamily. SNPs and CNVs in both genes have been associated with disorders including ASD. Zweier et al. (2009) have used Drosophila to examine a common synaptic link between these two genes and neuropsychiatric disorders (Table 3). This group studied both Nrx-I and Nrx-IV, the Drosophila orthologs of NRXN1 and CNTNAP2. Knockdown (KO) of Nrx-IV, either ubiquitously or only in neurons, led to late embryonic lethality. Immunostaining of Nrx-IV KO embryos showed an overall reduction in staining for the presynaptic protein bruchpilot that maintains the structural integrity of synaptic active zones (Wagh et al., 2006). They also found a dosage-dependent increase of bruchpilot staining when Nrx-IV was overexpressed in larval neurons. bruchpilot and Nrx-I colocalize at presynaptic active zones (Li et al., 2007). In conclusion, CNTNAP2 and NRXN1 may be involved in a common synaptic mechanism, contributing to the etiology of neurodevelopmental disorders together with bruchpilot (Zweier et al., 2009). Further studies are required to determinate the mechanism by which Nrx-I and Nrx-IV control bruchpilot levels.

Table 3:

Fruit fly

Human induced pluripotent stem cells

hiPSCs carrying a heterozygous deletion in CNTNAP2 have been generated to characterize the associated molecular and cellular phenotypes (Table 4). However, they were derived from a female proband who met the DSM-IV criteria for schizo-affective disorder, not ASD or ADHD. hiPSCs were also generated from both parents as well as five unrelated healthy controls (Lee et al., 2015). Neural migration was quantified using a neurosphere assay. Interestingly, hiPSC neural progenitor cells derived from the heterozygous CNTNAP2 schizo-affective patient showed significantly reduced neural migration relative to controls (Lee et al., 2015). This phenotype is consistent with the neuronal migration abnormalities observed in the Cntnap2−/− mouse model of ASD (Penagarikano et al., 2011). Global gene expression and neuronal activity were analyzed in the same cohort. HiPSC-derived neurons from individuals carrying the heterozygous CNTNAP2 deletion show altered neuronal activity and differential expression of genes involved in synaptic transmission (Flaherty et al., 2017).

Table 4:

Human induced pluripotent stem cells

Adhesion G protein-coupled receptor L3 Biological and molecular description

The latrophilin family of proteins was initially discovered owing their ability to bind α-latrotoxin, a component of black widow spider venom (Davletov et al., 1996). The ADGRL3 (previously known as LPHN3) is of particular interest in relation to neurodevelopmental disorders. Multiple polymorphisms in ADGRL3 have been linked to an increased risk of ADHD and SUD (Arcos-Burgos et al., 2010; Ribasés et al., 2010). ADGRL3 forms a trimeric complex with FLRT3 and UNC5, which supports transcellular adhesion and glutamatergic synapse development (Jackson et al., 2015). The identification of ADGRL3 in human genetic studies, coupled with its putative roles in neuronal migration and synapse development, provides strong support for ADGRL3 polymorphisms in the development of ADHD (O’Sullivan et al., 2014). The lack of naturally occurring coding variants in humans greatly limits our ability to investigate ADGRL3 (Orsini et al., 2016). Coding variants may not affect the genes that they are located in, but may rather have effects elsewhere in the genome. However, there is functional evidence that the noncoding variant rs2271338 affects ADGRL3 expression, further linking this gene to ADHD (Martinez et al., 2016).

Models Mouse

ADGRL3−/− mice displayed hyperactivity in the open-field test, (Table 1) but no significant genotype differences were observed in the amount of time spent in the center of the open field, a common test for anxiety. ADGRL3 null mice also displayed a heightened locomotor response to cocaine. A differential response to cocaine is of interest because of the link between ADGRL3 polymorphisms and SUD (Arcos-Burgos et al., 2012). In a food reward paradigm, ADGRL3−/− mice displayed increased levels of motivation to work for food, whereas motor coordination was unaffected in the rotarod test. In the Porsolt swim test, ADGRL3−/− mice spent more time swimming and had increased latency to immobility, possibly owing to reduced depressive behavior or hyperactivity. Whole brains of ADGRL3−/− and wild-type P0 pups were used to quantify candidate gene expression by qPCR. The genes chosen were involved in neuronal differentiation and survival as well as dopaminergic (DA) and serotonergic neurotransmission. The following genes were differentially expressed in ADGRL3−/−: serotonin transporter (SLC6A4), serotonin 2C receptor (5-Htr2c), dopamine transporter (SLC6A3), dopamine receptor D4 (Drd4), neural cell adhesion molecule 1 (Ncam1), nuclear receptor subfamily 4 group A member 2 (Nr4a2), and tyrosine hydroxylase (Th). High-pressure liquid chromatography also revealed an increased level of dopamine and 5-HT in ADGRL3−/− mice (Wallis et al., 2012). Transcriptomic analysis was performed on the prefrontal cortex, striatum, and hippocampus of ADGRL3−/− and control mice at 4 days, 28 days, and 6 months. Collapsing the RNA-sequencing results across brain regions and time points identified 11 genes that are differentially expressed in ADGRL3−/− mice. Two of these genes (Pcdhgb8, upregulated and Pcdhb9, downregulated) are members of the protocadherin family of calcium-dependent cell–cell adhesion molecules (Orsini et al., 2016), which have been linked to ADHD and ASD (Rivero et al., 2015). Independent analysis of brain regions and time points revealed a large number of differentially expressed genes including overexpression of Htr2c [5-hydroxytryptamine (serotonin) receptor 2c] in the cortex and SLC6A3 (dopamine transporter) in the hippocampus of 6-month old ADGRL3−/− mice. Gene set analysis did not find a statistically significant enrichment of gene sets following adjustment for multiple testing. However, a number of the most enriched pathways were relevant to ADHD and SUD, including synaptic vesicle cycle, DA synapse, glutamatergic synapse, neurotrophin signalling pathway, and both amphetamine and nicotine addiction.

Zebrafish

Zebrafish have two paralogues of human ADGRL3: adgrl3.1 and adgrl3.2. adgrl3.1 shows a more specific expression pattern during embryonic development. adgrl3.1 morphant larvae displayed hyperactivity under both light and dark conditions at 6 days after fertilization (Table 2). The increased distance swum was because of greater acceleration and average speed throughout the trial (Lange et al., 2012). In most patients with ADHD, MPH administration results in a reduction in hyperactivity (Ramos-Quiroga et al., 2009). MPH decreased hyperactivity in adgrl3.1 morphant fish but not in controls. The total distance swam was reduced by decreasing the average speed. Application of another ADHD treatment drug, the selective noradrenaline reuptake inhibitor atomoxetine, reduced the total distance swam by adgrl3.1 morphant zebrafish by increasing the resting time. The number of DA neurons in the posterior tuberculum of adgrl3.1 morphants was reduced at 3 and 6 days after fertilization. Subregions of the posterior tuberculum were also disorganised (Lange et al., 2012). Projectome analysis has indicated a functional correlation between the posterior tuberculum and the mammalian A11 group of DA neurons (Tay et al., 2011). Administration of DA receptor agonists and antagonists showed that adgrl3.1 MO larvae are hyposensitive to DA manipulation compared with adgrl3.1 CO. A possible explanation for this hyposensitivity is that there could be a heightened level of DA in the synaptic cleft of adgrl3.1 morphants, which might desensitize both DA-1 and DA-2 like receptors. A heightened level of DA might also explain the hyperactivity observed in these animals (Lange et al., 2018). An evolutionarily conserved enhancer for adgrl3.1 has been used to express green fluorescent protein in zebrafish. This enhancer was sufficient to produce an expression pattern similar to endogenous adgrl3.1 in the fore, mid, and hindbrain but not in the telencephalon and retina (Martinez et al., 2016).

Fruit fly

The Drosophila orthologue of ADGRL3 is known as latrophilin. Locomotion activity and sleep have been assessed in latrophilin KO flies using the Drosophila Activity Monitor (Table 3). Latrophilin KO flies were found to have heightened locomotor activity and reduced sleep, particularly during the 12-h night period. Testing in constant darkness showed that this hyperactive phenotype was light dependent. As light can buffer the wake-promoting effect of DA (Shang et al., 2011), the presence of a light-sensitive phenotype hints at a disruption of the DAergic system. However, the number and position of DA neurons is normal in latrophilin flies, suggesting that impairment of the DA system may result from changes in downstream signalling. Drosophila with a pan-neuronal KO of the dopamine transporter (DAT) gene displayed a similar light-sensitive hyperactive phenotype. Furthermore, application of MPH normalized the locomotor and sleep phenotype of latrophilin flies, further linking this gene to DA signalling (van der Voet et al., 2015).

Cellular models

Although no hiPSCs-containing ADGRL3 polymorphisms are currently available, a number of conventional cellular models have been used to study this gene. A family-based genetic analysis of 838 individuals identified six evolutionary conserved regions (ECRs) in the ADGRL3 genomic region that contained SNPs or haplotypes linked to ADHD (Martinez et al., 2016). ECR 37 and 47 were the only regions found to drive luciferase expression. The ECR 37 risk variant did not affect luciferase levels whereas the ECR 47 risk haplotype reduced luciferase activity by ~40% in both B35 neuroblastoma and U87 astrocytoma cell lines. Transcription factor binding motif analysis of ECR 47 showed an overrepresentation of transcription factors associated with brain function. Electromobility shift assays showed that the rs2271338 SNP in ECR 47 prevented binding of the Yin Yang 1 (YY1) transcription factor. However, YY1 interfering small RNA molecules did not reduce ADGRL3 expression in SH-SY5Y neuroblastoma cells, perhaps because of an inability to model this complex interaction outside of its native system. Expression quantitative trait loci analysis of post-mortem human brain tissue has linked the rs2271338 AA risk genotype allele to decreased thalamic expression of ADGRL3 (Martinez et al., 2016).

PARK2 Biological and molecular descriptions

Mutations in PARK2 were first discovered in a small percentage of patients with Parkinson’s disease (Lücking et al., 2000). Subsequent whole-genome studies have linked PARK2 CNVs with Han Chinese children with ASD (Yin et al., 2016) and European populations (Glessner et al., 2009; Scheuerle and Wilson, 2011). CNVs within the PARK2 locus have also been found among ADHD patient cohorts (Jarick et al., 2012).

PARK2, also known as PRKN, codes for an E3 ubiquitin ligase previously called Parkin. One of the most studied partners of PARK2 is the serine/threonine protein kinase PINK1. Acting together, PARK2 and PINK1 constitute an internal sensor system for disparate perturbations to cellular homeostasis (Pickrell and Youle, 2015). In fact, different stimuli can lead to PINK1 accumulation on the outer mitochondrial membrane where it can recruit and activate cytosolic PARK2. PARK2 can then target several mitochondrial proteins for degradation (Hang et al., 2015). The interaction between PINK1 and PARK2 plays a pivotal role in mitochondria dynamics and quality control, regulating processes such as mitophagy, fusion and fission, biogenesis, and transport (Scarffe et al., 2014). PARK2 is also linked to apoptotic pathways and could exert a proapoptotic or antiapoptotic effect in a stressor-dependent and environmental-dependent manner (da Costa et al., 2009