Analysis of DA and its metabolites in rat mPFC by high throughput liquid chromatography (HPLC) showed that DA concentrations were significantly lower in juvenile and adolescent rats than in adults. DA concentration rose steadily between the juvenile (P25) and late adolescent stages (P45) and increased particularly sharply between the end of adolescence and adulthood. In parallel, a decrease in DA turnover ratios was observed with increasing age, an effect that could contribute to the overall increase in DA availability in the mPFC (Naneix et al., 2012). Analysis of DA tissue concentrations in rhesus monkey PFC showed that DA levels fluctuated between 2, 5, 8 and 15-18 months old animals and significantly increased in 2–3 years old animals (Goldman-Rakic and Brown, 1982). These data suggest that both in rats and rhesus monkey, the overall DA concentration coincides with the increase in DA fiber innervation of the PFC. However, whether this increase in concentration correlates with active DA release has not been investigated in the developing PFC. The recent development of genetically encoded DA sensors that allow the monitoring of DA release in the behaving animal, offer the opportunity to correlate behavior, PFC function and DA release in real-time in adolescent and adult animals (Labouesse et al., 2020).

Once released from the axonal varicosities of DA axons, DA binds to DA receptors (DRDs) of the D1-like or D2-like subfamily of G-protein coupled receptors. DRD1 and DRD5 belong to the D1-like subfamily, while DRD2, DRD3, and DRD4 are subtypes of the D2-like subfamily. Unlike Drd1 and Drd5, the D2-like subfamily receptor genes contain introns that allow differential splicing of the transcripts, generating additional isoforms. Drd2 comes in two alternatively spliced variants, Drd2s (short form) and Drd2l (long form), and isoforms of Drd3 and Drd4 have also been identified (Missale et al., 1998). D1-like receptors signal by coupling to G proteins G_as and G_aolf, which stimulate adenylyl cyclase and lead to activation of protein kinase A (PKA). D2-like receptors stimulate G_ai and G_ao proteins, blocking adenylyl cyclase and consequently inhibiting PKA activity (Missale et al., 1998; Tritsch and Sabatini, 2012). Furthermore, DRDs can activate a signaling cascade by interacting with ß-arrestin (Beaulieu et al., 2005) or induce phospholipase C-mediated increase of intracellular calcium levels (Lee et al., 2004), although the signal transduction pathway of this modulation remains to be resolved (Chun et al., 2013). The striatum and the nucleus accumbens receive dense projections from mDA neurons and have high expression levels of DRDs. In the PFC, the expression levels of the DRDs are considerably lower, correlating with relatively sparse innervation by DA fibers.

The distribution and expression of DRDs and their transcripts in rodent PFC have been studied using multiple histological methods, real-time quantitative PCR and in recent years, genetic tools and single-cell transcriptome analysis (Table 1). Early studies include autoradiographic experiments employing radiolabeled agonist or antagonist of DRDs (Boyson et al., 1986; Noisin and Thomas, 1988), immunohistochemical and immunoblotting approach targeting the receptor protein (Levey et al., 1993; Sesack et al., 1994) and in-situ hybridization technique detecting Drd transcripts (Gaspar et al., 1995). Some of the radioligands used in binding assays were later found to lack selectivity for specific subtypes of DRD (Landwehrmeyer et al., 1993) and similar doubts have been expressed for commercially available antibodies for the receptors (Bodei et al., 2009). RNA in situ hybridization methods have characterized the distribution of certain Drd mRNAs within the subregions of the PFC (Santana and Artigas, 2017) and RT-qPCR approaches were used to quantify the relative gene expression of the Drd subtypes in the PFC (Araki et al., 2007). Whether the transcript levels reliably correspond to the expression levels of DRD protein is not known.

Taking into account these methodological limitations, studies on DRD proteins and their transcripts indicate that of the five DRD subtypes, DRD1 and its mRNA are most highly expressed in the adult rodent PFC, followed by DRD2/Drd2. In comparison, DRD3, 4 and 5 show limited expression (Tarazi and Baldessarini, 2000; Lidow et al., 2003; Araki et al., 2007; Rajput et al., 2009; Santana et al., 2009). DRD1 and DRD2 are expressed in both pyramidal neurons and interneurons of rodent PFC but are rarely colocalized (Santana et al., 2009; Zhang et al., 2010). RNA in situ hybridization studies in adult rats show that cells expressing Drd1 mRNA are most prominent in layer VI, extending into layer V, with an additional thin band of positive cells in layer II. Drd2-expressing cells are mainly localized in layer V and VI, with few positive cells in layer II and III (Gaspar et al., 1995; Santana and Artigas, 2017). This laminar distribution pattern of DRD1 and DRD2 in rat mPFC was also observed in an earlier receptor binding study using fluorescently coupled receptor antagonists (Vincent et al., 1993). More recently, genetic labeling has emerged as an additional tool to monitor Drd1- and Drd2-expressing neurons in rodents. Genetic labeling studies involve transgenic mice that accommodate a BAC (bacterial artificial chromosome) construct containing Drd1 or Drd2 regulatory regions directing expression of Cre recombinase (Drd1-Cre or Drd2-Cre mice) (Gong et al., 2007). These Cre mice are crossed with reporter mice that express fluorescent proteins upon Cre-mediated recombination (such as Ai14 or Ai6 mice) allowing the identification of cells that express Drd1 or Drd2 (Madisen et al., 2010; Wei et al., 2018). In rats, Drd2-Cre knock-in animals have been generated and crossed with a fluorescent rat reporter line (Ai9) (Madisen et al., 2010; Yu et al., 2019). An important aspect to keep in mind with these Cre reporter systems is that recombination of the reporter allele is permanent, meaning that if the Drd1 or Drd2 promoter is transiently active in certain cell populations during embryonic or postnatal development, these cells will be recombined and continue to express the fluorescent protein in the adult brain even when these neuronal populations may no longer express Drd1 or Drd2 in the adult. Furthermore, in this system, the expression level of the fluorescent protein does not correspond to the level of endogenous gene or protein expression. Despite these caveats, in Drd2-Cre, Ai9 reporter rats, the distribution of recombined cells (expressing fluorescent reporter protein) is largely in agreement with previous findings on Drd2 expression in the mPFC (Santana and Artigas, 2017). Analysis of recombined cells in the anterior cingulate cortex show them mostly to be putative pyramidal neurons of upper and deep layers. Only a small number of inhibitory interneurons exhibit fluorescent labeling in this region (Yu et al., 2019). Similarly, in Drd1-Cre, Ai6 or Drd1-Cre, Ai14 reporter mice, fluorescently labeled cells show a laminar distribution comparable to what has been reported for Drd1 transcript expression in mPFC, with a higher overall density of Drd1 expression in deep layers. In Drd2-Cre Ai6/Ai14 mice, however, distribution of fluorescently labeled cells in mPFC is strikingly distinct from the one reported in Drd2-Cre, Ai9 reporter rats or the expression patterns observed in RNA in situ hybridization studies, showing high expression of Drd2 in superficial layers rather than in deep layers (Wei et al., 2018). Whether this is due to the different approaches used to generate the Cre-lines (BAC transgenic mice versus knock-in rats) or reflects a transient expression of Drd2 in superficial layers of the mPFC during development in the mouse is unclear (Beil et al., 2012; Yu et al., 2019). BAC transgenic mice expressing enhanced green fluorescent protein (EGFP) under the transcriptional regulation of Drd3 (Drd3-Egfp mice) or Drd4 (Drd4-Egfp mice) locus have also been used to study the expression of Drd3 and Drd4 in different regions of the brain (Gong et al., 2003). In the Drd3-Egfp mouse model, the fluorescent cells in the caudal cingulate cortex are mainly located in layer VI (Li and Kuzhikandathil, 2012). Analysis of Drd4-Egfp mice showed strongly labeled EGFP-expressing neurons in layer V and VI of prelimbic and cingulate cortices (Noaín et al., 2006). DRD5 immunoreactivity has been detected in layer II to layer VI of prelimbic and cingulate cortices, with more labeled cells in layer II and III. In mice, DRD5 is more uniformly distributed across the cortical layers of the mPFC (Ciliax et al., 2000; Lidow et al., 2003; Table 1).

The developmental time course of DRD expression in rodent PFC is not well characterized and appears to vary considerably between rats and mice. RT-qPCR analysis in the murine cingulate cortex (both at rostral and caudal levels) at P0, P21, and P60 reveals that other than Drd4, which has the highest expression at birth followed by a rapid postnatal decrease in expression, transcript levels of the Drd subtypes do not show any significant developmental change between P0 and P60 (Araki et al., 2007). In the frontal cortex of rats, in situ hybridization signals for Drd1 or Drd2 transcripts have been detected around E14 or E18, respectively (Schambra et al., 1994). According to the same study, expression levels for both Drd1 and Drd2 appear to reach maximal levels between P14 and P30, although the change in signal intensity has not been quantified. Another study, however, shows that Drd1, Drd5, Drd4, Drd2l (but not Drd2s) expression in the mPFC of rats reaches peak expression only at P45 and then decreases between P45 and P70 (Naneix et al., 2012). At the protein level, there is a marked decline of DRD1 and DRD2 density in PFC of rats between adolescence (P40) and adulthood (P120) (Andersen et al., 2000). An earlier study using quantitative autoradiography in rats has described a similar pattern for DRD1 in mPFC, but with peak receptor binding density at P14 and P21, and a decrease in binding between P21-P42 (Leslie et al., 1991). A certain population of mPFC pyramidal neurons projecting to the nucleus accumbens also shows differential expression of DRD1 across postnatal development. In retrogradely traced prelimbic pyramidal neurons projecting to the nucleus accumbens core, the number of DRD1 immunoreactive cells was significantly higher in adolescents (P44) than in juveniles (P27) or adults (P105) (Brenhouse et al., 2008). Tarazi and Baldessarini, however, report a different temporal expression pattern in frontal cortex of rats. In their investigation, binding of radioligands to DRD1, DRD2 and DRD4 receptors gradually rises from P7 to maximal levels at P60 (Tarazi and Baldessarini, 2000). Overall, the data from various published studies do not deliver a conclusive picture on the time course and distribution of DRD/Drd expression in the developing mPFC (Table 2).

An additional potent tool to investigate the distribution of Drd transcripts is single-cell mRNA sequencing (scRNA seq). DropViz is an extensive collection of scRNA seq data, assembled from analysis of RNA expression of thousands of individual cells across different regions of mouse brain (P60–P70) (Macosko et al., 2015; Saunders et al., 2018). Based on gene expression profiles in the frontal cortex (including the mPFC, orbital cortices, frontal association cortex, anterior parts of primary and secondary motor cortices, insular cortex and somatosensory cortex), Drd1 and Drd5 expression is highest in deep layer pyramidal neurons and Drd4 is mostly expressed in pyramidal cells of layer II/III. Drd2 transcript levels are notably low and are predominantly found in interneurons rather than in projection neurons. Additionally, a rather remarkable observation is that the highest level of Drd1 and Drd2 expression is found in microglia. Drd3 expression is not included in this transcriptional analysis of the frontal cortex, possibly because of low expression levels. The transcriptional dynamics of the Drds in the frontal cortex during development has not yet been investigated. However, dynamic regulation of Drd1 has been demonstrated in the context of mouse models of drug abuse. Bhattacherjee and colleagues have shown that chronic cocaine addiction induces cell type-specific transcriptional changes in the murine mPFC. The effect of cocaine addiction on gene expression changes was particularly striking during the withdrawal period, with excitatory neurons in the deeper layers being more affected. While the most significantly affected excitatory clusters expressed Drd1, the analysis also detected Drd1 expressing excitatory clusters that did not respond robustly to cocaine. Although the functional role of each subtype remains to be investigated, this suggests that certain Drd1-expressing neuronal subtypes in the PFC may be more involved in the process of cocaine addiction than others (Bhattacherjee et al., 2019). Further analysis on dataset of cocaine-addicted mice revealed that Drd1 and Drd2 genes are both upregulated in cocaine addiction and are almost solely expressed in excitatory neurons, with Drd1 also being found at lower levels in inhibitory neurons, oligodendrocyte and endothelial cells in the mPFC (Bhattacherjee et al., 2019; Navandar et al., 2021).

In addition to DRD expression patterns, maturation of receptor function could also contribute to changing impact of the mesoprefrontal system over time. Investigations into DRD function have shown that DRD1-mediated modulation of NMDA receptor transmission prompt recurrent depolarizing plateaus in pyramidal neurons of mPFC slices, an effect that develops only after P45 (Tseng and O’Donnell, 2005). Furthermore, DRD2-mediated increase in excitability of fast-spiking interneurons in PFC slices appears only after P50 (Tseng and O’Donnell, 2007). Thus, while the changes in postnatal expression levels of DRD/Drd are still unclear, there is indeed a change in activity of DRDs in the post-pubertal stage, hinting towards the role of DA in the remodeling of PFC microcircuits during the transition from adolescence to adulthood.

Another gap in our understanding of DRD receptor expression and function in the developing and adult mPFC is that we know little about the subcellular localization of receptors in DRD-expressing neurons. Because existing antibodies against DRDs have limited utility for detecting DRDs in brain tissue (Bodei et al., 2009), alternative approaches should be considered for investigating this question. Vincent and colleagues analyzed cellular localization of D1- and D2-like family of receptors in the mPFC using receptor antagonists coupled to fluoroprobes and observed that around 25% of all fluoroprobe-labeled cells displayed both D1 and D2-like subfamily receptor binding fluorescence along the outer edge of the soma. Further analysis on cell size distribution suggested that the cells in which colocalization could be detected were non-pyramidal (Vincent et al., 1995). A recent promising technique to examine subcellular localization of DRDs may be the application of CRISPR/Cas9 based genome editing tools to introduce fluorescent tags to endogenous receptor proteins. A recent study used a modified CRISPR/Cas9 knock-in strategy with two guide RNAs to knock-in a fluorescent protein to α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) and N-methyl D-aspartic acid (NMDA) receptor subunits in primary mouse cortical cultures (Fang et al., 2021). The application of these epitope tags in vivo is also possible. Using the so-called ORANGE (Open Resource for the Application of Neuronal Genome Editing) toolbox, adeno-associated virus plasmids containing fluorescent tag knock-in constructs for PSD95 and AMPA receptor subunit (GLUA1) were injected into the hippocampus of Cas9-P2A-GFP transgenic mice resulting in robust labeling of both proteins (Willems et al., 2020). Applying these methods for the fluorescent tagging of DRDs has the potential to aid in determining the subcellular localization of DRDs in fixed tissue as well as monitoring dynamics of receptor localization in dissociated cell cultures in vitro or in acute slices.

The distribution of DRDs in the rodent mPFC correlates largely with the innervation pattern of DA fibers, suggesting that DRD expression might be influenced by DA release in the mPFC. In this context, DA might play a role during the phase when projections are established (as in a critical developmental period) and/or influence DRD expression levels in the adult brain. Indeed, there is evidence from the striatum that ablating DA innervation during early postnatal development (using 6-OHDA-mediated lesion of the nigrostriatal and mesolimbic pathway at neonatal stages) results in reduced binding of radioligand to DRD1 in the caudate putamen and nucleus accumbens of the adult (P90) rat. Radioligand binding to DRD2 is not affected (Thomas et al., 1998). In the adult brain, the loss of striatal DA input in Parkinson’s disease patients or in animal models of the disease leads to compensatory upregulation of DRDs, while drug-induced DA increase in the nucleus accumbens leads to reduced expression of DRDs to adjust for elevated DA in the system (Hisahara and Shimohama, 2011; Volkow and Morales, 2015). In the mPFC, the influence of DA on DRD expression has not been studied in detail. One study has examined the effect of depletion of DA projections in the postnatal rat by intracisternal injection of 6-OHDA 5 days after birth and found that DRD1 receptor binding remains unaltered (Leslie et al., 1991). Mouse models interfering with the development of mesoprefrontal projections, such as the Dcc and Netrin-1 haploinsufficient mice that elevate DA transmission in the mPFC (Vosberg et al., 2020) or mouse models that lack mesoprefrontal innervation (Kabanova et al., 2015) may offer a suitable approach to determine the role of DA innervation in the developmental trajectory of DRD expression in the mPFC.

In the adult PFC of rhesus monkeys, autoradiographic receptor binding assays showed that DRD1 is most densely present in layers I, II, IIIa, V, and VI, while DRD2 shows the highest expression density in layer V of adult PFC (Lidow and Rakic, 1992) (Table 1). Immunohistochemistry for DRD1 and DRD5 in rhesus monkey PFC (area 9) demonstrated that these receptors widely colocalize on spines of pyramidal neurons and axon terminals (Bordelon-Glausier et al., 2008). In the adult human PFC, DRD1, DRD2 and DRD4 are highly expressed in deeper layers (layer V and VI) and layer II (Weickert et al., 2007; Table 1). These studies did not report on the expression of DRD3.

DRDs appear to be dynamically expressed in the developing PFC in primates. In the adult rhesus monkey PFC (5–6 years old), DRD1 and DRD2 density (examined by autoradiographic receptor binding assays) was found to be significantly lower compared to 2 months of age (Lidow and Rakic, 1992). Another study, using [¹¹C] FLB 457 (high-affinity radioligand for DRD2/3) in positron emission tomography (PET) on human subjects (age range 19–74 years), detected a significant decline in DRD2/3 expression with age in the frontal cortex area (Kaasinen et al., 2000). An immunohistochemistry study was not performed for these receptors. At the transcriptional level, a cohort study of human post-mortem PFC tissue revealed that DRD1 mRNA is expressed at neonatal stages. Expression levels decline within the first year of life, are highest during adolescence and young adults, and gradually decline again in adult and aged cohorts (Weickert et al., 2007). However, in a similar cohort study of the human dorsolateral PFC, DRD1 mRNA expression was reported to increase steadily until adolescence but to decrease slightly thereafter. Western blot analysis of DRD1 expression indicated that protein levels also increase gradually with age, but the highest expression was found in the young adult and adult groups (Rothmond et al., 2012). Moreover, a similar layer-specific pattern was observed across all studied ages: DRD1 transcript levels were not detected in layer I of the human dorsolateral PFC, were present at an intermediate level in layers III and IV and highest expression was found in layers II, V, and VI (Weickert et al., 2007). Unlike DRD1, DRD2 expression levels peak at neonatal age, followed by a significant decrease in infants. At all later developmental time points examined, expression levels remain below neonatal levels. Similarly, mRNA levels of the short (DRD2S) and long (DRD2L) DRD2 isoform are highest at the neonatal stage and decrease with age in the dorsolateral PFC. A layer-specific pattern was observed also for DRD2 with highest expressions in layers II, V, and VI. DRD1 and DRD2 mRNA was found in both pyramidal and non-pyramidal neurons in adult brain (Weickert et al., 2007). DRD4 mRNA expression was detected in presumed non-pyramidal neurons and glia but was barely present in pyramidal cells (Weickert et al., 2007). Generally, DRD4 did not show any age-specific changes in expression and highest signal intensity was detected in layer V (Weickert et al., 2007; Rothmond et al., 2012). DRD5 expression levels did not show any significant differences between age groups (Rothmond et al., 2012). To the best of our knowledge, the distribution of DRD3 expression in the developing primate PFC has not yet been reported (Table 2).

Similar to what we have highlighted above for the investigation of Drd expression in the rodent brain, high-throughput techniques for transcriptome analysis, such as scRNAseq, give now the opportunity to explore the cell-type specific expression of DRD transcripts in the developing and adult human PFC in further detail (Allen Institute for Brain Science, 2010; Fan et al., 2018, 2020; Zhong et al., 2018; Polioudakis et al., 2019; Tanaka et al., 2020; Maynard et al., 2021). This will be instrumental in defining temporal dynamics and cell type-specific responsiveness to DA.

Finally, neither DA release nor receptor expression may offer a full reflection of how DA impacts on cortical neurons in the PFC. As discussed above, DRDs act on DA-receiving cells by modulating PKA activity. Thus, monitoring PKA activity may offer additional insight into the effects of DA on cortical neuronal function. A recent study used a PKA activity sensor to monitor the effect of DA release on Drd1- versus Drd2-expressing medium spiny neurons in the nucleus accumbens during learning in real-time (Lee et al., 2021). However, given that DA innervation and release is much sparser in the mPFC than in the nucleus accumbens and other modulatory neurotransmitters released in the mPFC (e.g., NA, Serotonin) act also via G-protein coupled receptors and modulation of PKA activity, further studies would be needed to determine whether a similar approach could be applied in the PFC.

In summary, a better understanding of the developmental time course of DA release; the laminar distribution, neuronal subtype expression, and subcellular localization of DRD receptors as well as downstream signaling events would greatly contribute to our knowledge of the functional role of DA in the developing and adult PFC.

Schizophrenia is a neuropsychiatric disorder with severe symptoms that usually become manifest in full during adolescence or early adulthood. These include the so-called positive symptoms (psychosis), negative symptoms (deficits in emotional responses and thought processes), and cognitive dysfunction (e.g., deficits in working memory, long-term memory, semantic processing, learning) (Marder and Cannon, 2019). According to the so-called DA hypothesis of schizophrenia, alterations in DA signaling are a major factor in these disease symptoms: DA hyperactivity in the striatum promotes psychosis, while DA hypoactivity in other brain areas, including the PFC, contributes to the negative symptoms and cognitive dysfunction. There is ample evidence from human studies to support this hypothesis. To name a few: (1) DA agonists and stimulants such as cocaine or amphetamine can induce psychosis in healthy individuals and exacerbate psychosis in patients with schizophrenia; (2) antipsychotic drugs act on the DA system via DRD2 receptors (e.g., haloperidol); (3) postmortem studies have demonstrated increased levels of DRDs, DA, and DA metabolites in the striatum of patients with schizophrenia; (4) imaging studies in patients with schizophrenia show that stimulant-induced presynaptic DA release is decreased in most brain regions, except for the striatum, where it is increased. For further details, we refer the interested reader to a collection of reviews on the DA hypothesis of schizophrenia (Biol Psychiat, 2017).

With respect to the mesoprefrontal DA system, its hypoactivity is most likely associated with the cognitive dysfunctions in schizophrenia. The cause of the overall DA imbalance may be caused by deficits in local cortical or hippocampal networks that in turn lead to changes in the inputs to the VTA from these regions and ultimately to DA hypoactivity in VTA targets. Alternatively, or in addition, defects in the regulation of DA release in target regions (including the PFC) or in the developmental of the mesoprefrontal system could contribute to the DA hypoactivity (Rice et al., 2016; Abi-Dargham, 2017; Chuhma et al., 2017; Grace, 2017; Walker et al., 2017; Sonnenschein et al., 2020; Braun et al., 2021). Whether the development of the mesoprefrontal DA system (or other parts of the DA system) is altered in patients with schizophrenia has not yet been studied in detail.

Autism spectrum disorder (ASD) encompasses a group of severe neurodevelopmental disorders that exhibit core symptoms of social and communication deficits and stereotyped, repetitive behaviors (Association, 2013; Fein et al., 2021). Many studies highlight similar behavioral and cognitive impairments between ASD and schizophrenia such as social and language deficits and there is a high co-occurrence of both neurodevelopmental disorders (Spek and Wouters, 2010; King and Lord, 2011; Chisholm et al., 2015; Crescenzo et al., 2019). Based on this, it has been speculated that dysfunction in the DA system may also contribute to the cognitive disorders in ASD and, similar to schizophrenia, a DA hypothesis has been proposed for ASD. According to this hypothesis, aberrant mesocorticolimbic and nigrostriatal DA circuitry may contribute to reward deficits and goal-directed motor impairments manifested in ASD children (Pavăl, 2017; Pavăl and Micluţia, 2021). Initial evidence for impairments in the DA system in ASD came from a study that found elevated levels of DA metabolites, such as homovanillic acid, in the cerebrospinal fluid of autistic children (age 1- 16 years old) (Gillberg and Svennerholm, 1987). Further evidence supporting this hypothesis comes from (1) the discovery that de novo genetic variants of the gene encoding the dopamine transporter (DAT) (Neale et al., 2012; Hamilton et al., 2013; Bowton et al., 2014; Cartier et al., 2015) and gene polymorphisms in DRD3 and DRD4 (Gadow et al., 2010; Staal, 2014; Staal et al., 2015) are associated with ASD; (2) the therapeutic efficacy of DRD blockers (risperidone and aripiprazole) in alleviating stereotypic and/or abnormal social behaviors in children with autism (McCracken et al., 2002; McDougle et al., 2005; Ghaeli et al., 2014) and (3) studies showing that the reward circuitry is hypoactivated in autistic patients in response to social and monetary rewards (Zeeland et al., 2010; Dichter et al., 2012; Kohls et al., 2012). According to the DA hypothesis in ASD, this diminished ability to register rewards for social cues could lead to the decreased pursuit of social interaction and ultimately to the deficits in social and communication skills observed in ASD patients (Pavăl, 2017). Regarding the mesoprefrontal system, an early PET scanning study for fluorine-18-labeled fluorodopa (F-DOPA) revealed significantly decreased F-DOPA ratio in the anterior mPFC of autistic children compared to healthy subjects, indicating decreased DA activity in the mPFC in autistic patients (Ernst et al., 1997). ASD patients underperform in working memory tasks involving planning, cognitive flexibility, and high working memory load compared to control subjects, which could be due to, or at least influenced by, a dysfunctional mesoprefrontal DA system (Kercood et al., 2014). Moreover, computational models predict that decreasing DA modulation in the PFC could lead to executive dysfunctions such as decreased cognitive flexibility, as occurs in ASD (Kriete and Noelle, 2015). Nevertheless, it remains largely unclear whether impairments of the mesoprefrontal DA system contribute to cognitive deficits in ASD patients and whether the development of mesoprefrontal mDA neurons is altered in ASD. The phenotypic heterogeneity of ASD and largely unknown disease mechanisms complicate the investigations of these potential deficits.

To uncover the potential role of altered development of the DA system and in particular the mesoprefrontal DA neurons in ASD etiology and associated social and executive dysfunctions, further DA system-focused studies in patients and ASD mouse models are needed. Evidence from mouse models for the involvement of the DA system in ASD is discussed in detail in a recent review (Kosillo and Bateup, 2021), thus we will only discuss two examples here. Mutations in the gene encoding SH3 and multiple ankyrin repeat domains 3 (SHANK3), a postsynaptic scaffolding protein, have been discovered in ASD patients, making it a prominent autism gene candidate (Gauthier et al., 2009; Phelan and McDermid, 2012; Boccuto et al., 2013). Studies on the Shank3 haploinsufficient mouse model show that impaired preference for social interactions is due to decreased DA activity in the VTA (Bariselli et al., 2016, 2018). Whether this hypoactivity results in decreased DA release in the nucleus accumbens and/or the mPFC has not yet been addressed. A potential link between autistic-like phenotypes and aberrant development of the DA system emerges from animal models for Mucopolysaccharidosis (MPS). MPS are hereditary lysosomal storage diseases, in which dysfunctions in lysosomal hydrolases lead to the accumulation of undegraded glycosaminoglycans in lysosomes and eventually to disturbances in cellular metabolism. In MPS IIIa, in which the gene coding for the lysosomal hydrolase sulfamidase is mutated, the metabolic cellular deficits result in neurodegeneration and dementia in children. Dementia is preceded by severe autistic-like behaviors (Valstar et al., 2010; Rumsey et al., 2014). In a mouse model of MPS IIIa, inactivation of the gene coding for sulfamidase, results in severely impaired behavior that that can be considered autism-like. These behavioral deficits are associated with increased DA release in the dorsal and ventral striatum and can be ameliorated with a DRD1 antagonist. This hyperdopaminergic state in MPS IIIa mice appears to be caused by developmental changes in the DA system: increased proliferation of mDA progenitors results in an increased number of mDA neurons in the SNpc and the VTA in the adult brain. Moreover, the same study shows that autistic-like behaviors and increased DA cell number are also present in a mouse model for a different type of MPS (MPS-II) (Risi et al., 2021). While this study suggests that altered development of the mDA system may be one of the causes of autism-like behaviors, it has not been investigated whether the increase in VTA neurons in these animal models leads also to alterations in the mesoprefrontal DA system. Further investigation of existing and potentially novel ASD candidate genes in animal models will be necessary to uncover developmental, structural, and/or functional impairments of the mesoprefrontal DA system in association with ASD.

Attention deficit/hyperactivity disorder (ADHD) is a highly heritable, early-onset neurodevelopmental disorder, characterized by symptoms of hyperactivity, short attention span, and impulsivity. The PFC is a key region afflicted in this disorder. Studies report thinning of PFC areas, reduced density of the dorsolateral PFC, and decreased PFC activity in ADHD patients compared to controls (Arnsten and Pliszka, 2011; Cortese, 2012; Klein et al., 2019). Shaw and colleagues reported that the PFC in children with ADHD takes significantly longer to reach peak cortical thickness compared to the PFC in typically developing individuals, suggesting a delay in PFC maturation (Shaw et al., 2007a). The typical ADHD symptoms also reflect impaired executive functioning of PFC, which in turn is related to dysregulated NA and DA signaling in the PFC (Arnsten and Pliszka, 2011). There are several points of evidence that suggest that alterations in the DA system may contribute to ADHD symptoms. An F-DOPA PET study showed low DOPA-decarboxylase activity in the PFC of adult ADHD patients compared to healthy controls, an effect that could however not be replicated in adolescents with ADHD (Del Campo et al., 2011). Methylphenidate and amphetamine, which are used in the treatment of ADHD, act by inhibiting DA and NA reuptake and consequently by increasing DA and NA transmission in the PFC. Low doses of methylphenidate have been shown to improve PFC function in rats and monkeys, which can be counteracted by blocking DRD1 receptor. Moreover, mice heterozygous for the gene encoding dopamine transporter (DAT hypofunction mice), show behavior typical for ADHD such as hyperactivity, inattention, and impulsivity. Inattentive and impulsive behavior in these mice can be rescued by amphetamine. In humans, using radiolabeled altropane, a high-affinity selective probe for DAT, neuroimaging studies point towards evidence of increased DAT activity in striatum of children and adults with ADHD. However, due to its limited expression, it has been challenging to analyze DAT levels in the cortex using PET imaging techniques and it is still poorly characterized in the PFC of ADHD patients (Spencer et al., 2005; Prince, 2008). In addition, there is a significant association between ADHD and polymorphism in the genes that encode DRD4, DRD5, and DAT. DRD4 has a high number of polymorphisms in its nucleotide sequence. Comprehensive meta-analyses showed that the so-called DRD4 7-repeat allele (DRD4 7R; a 7-repeat form of the 48–base pair (bp) variable number tandem repeat) elevates the risk of ADHD (Wu et al., 2012). Shaw and colleagues showed that presence of DRD4 7R was linked to cortical thinning in orbitofrontal and inferior prefrontal cortex that was augmented in ADHD patients (Shaw et al., 2007b). Another study suggests a considerable reduction in gyrification of inferior frontal gyrus in children with ADHD, who were DRD4 7R allele carrier. The authors hypothesize that this DRD4 polymorphism could affect early stages of cortical development in children who later develop ADHD (Palaniyappan et al., 2019). Additionally, a 148-bp and a 136-bp dinucleotide repeat allele from the DRD5 gene have also received considerable attention while the most extensively studied DAT polymorphism involves the 40 bp 9-repeat and 10-repeat alleles (Gizer et al., 2009; Wu et al., 2012).

An evolutionary perspective on ADHD argues for an adaptive role of the mesoprefrontal system in the disorder. Symptoms associated with ADHD, such as hyperactivity or limited sustained attention, could help animals to detect threats more rapidly and hence serve as beneficial features in endangered situations (Jensen et al., 1997; Lee and Goto, 2015). When delayed PFC maturation puts animals at a disadvantage in an adverse environment, ADHD symptoms arising from reduced mesoprefrontal DA could emerge as a compensative mechanism to make animals less vulnerable to the environmental threats. While such an adaptive response may have aided ancestral humans in stressful conditions, it does not translate well to modern social settings (Lee and Goto, 2015).

In summary, these data indicate that changes in DA signaling, in particular in the PFC may play a critical role in the pathophysiology of ADHD. However, it remains challenging to separate the impact of altered DA versus NA signaling on PFC dysfunction in ADHD. It also should be taken into consideration, that similar to schizophrenia and ASD, alterations in DA signaling could be secondary to functional changes in cortical areas (Arnsten and Dudley, 2005; Gamo et al., 2010; Arnsten and Pliszka, 2011; Mereu et al., 2017). The etiology of ADHD is multifaceted, having a strong genetic background but also contributions from environmental risk factors. Beside PFC, other brain regions having reciprocal connection to PFC, such as caudate and cerebellum are affected and there is an intricate interplay of neurotransmitters distinctive to each region (Arnsten and Pliszka, 2011; Cortese, 2012). Our understanding of the role of reduced mesoprefrontal signaling among these complex interactions is still evolving (Stanford and Heal, 2019) and requires further studies to better understand both, its specific function, and its complementary role along with DA signaling in the subcortical brain regions, in the pathophysiology of ADHD.