Our study shows that reduced Foxp1 expression in the Foxp1+/− striatum is accompanied by greatly reduced Pde10a mRNA and protein levels during development and adulthood while the cortex, hippocampus, and thalamus are not affected. We can also demonstrate that continuous administration of the Pde10 antagonist MP-10 from birth can completely reverse early deficits in social behavior, anxiety disorders, and hyperactivity, as well as changes in striatal microglia and synaptic pruning.
It is well known that Pde10a modulates the signaling of the cyclic nucleotides cAMP and cGMP. Highly and almost exclusively expressed in striatal medium spiny neurons, it plays a critical role in regulating striatal activity and basal ganglia circuitry 32, 33. Blocking PDE10A with specific antagonists has previously reduced symptoms of Huntington’s disease (HD) and Parkinsons’s disease (PD) in mice, rats, and monkeys, although PDE10A expression is reduced in both disorders 34–36. We now also show reduced Pde10a expression in Foxp1+/− animals. The reason why inhibition of this phosphodiesterase has a positive effect on symptoms can be explained by the fact that PDE10A antagonists act similarly to D2 receptor blockers by increasing the activity of D2-type medium spiny neurons (D2-MSNs) of the indirect (striatopallidal) pathway 37–40. Due to a lower cAMP threshold in D1-type medium spiny neurons of the direct pathway (D1-MSNs) compared with D2-MSNs, antagonism of Pde10a may result in more pronounced downstream effects in D2-MSNs, ultimately leading to a balance between the direct and indirect pathways 39.
In this context, it is interesting to note that Foxp1+/− mice exhibit hyperexcitability of D2-MSNs, whereas no change was detectable in D1-MSNs 15, 41. Increased intrinsic excitability, which was mainly due to downregulation of two classes of potassium currents (inward rectifying (KIR) and leak currents (KLeak) could also be confirmed in D2-MSNs with homozygous Foxp1 deletion. The downregulation Kcnj2 and Kcnj4 for KIR and Kcnk2 for Kleak detected in these neurons may underlie these currents 41. As Pde10a is known to regulate intracellular signaling in striatal MSNs and exerts strong control over gene expression 42, 43, altered Pde10a expression may be responsible for the downregulation of these potassium channels.
We have demonstrated that Foxp1+/− mice exhibit alterations in striatal microglial morphology and reduced synaptic pruning in pups. Microglia, the resident immune cells of the central nervous system, constantly scan their local microenvironment and sense impairments triggered by endogenous and/or exogenous factors. In pathological conditions, microglia are activated and the resulting dysregulation of certain genes is thought to be an inevitable part of almost all CNS pathologies 44. Reduced complexity and surface area in Foxp1+/− microglia indicate activation. Microglial activation is characterized by a transformation from a branched morphology with small cell bodies and highly branched, filamentous cell processes to a morphology with a large cell body and short, stout, unbranched cell processes, as has been described in early brain development 45. The significantly increased expression of Cd74, Rhoa, Ifi30 and Fcgr2b in the striatum of Foxp1+/− pups supports our findings. Cd74 is strongly upregulated and considered a marker for reactive microglia, as it is more strongly expressed under disease conditions than in the healthy central nervous system 46, 47. Moreover, it is well known that Cd74 mediates binding of the extracellular pro-inflammatory cytokine macrophage migration inhibitory factor, which is released in response to stress or an inflammatory response 48. Rhoa is known to control reactivity and survival of microglia during neuroinflammation and there is new evidence of abnormal RhoA signaling in neurodegenerative diseases 49. It is also assumed that altered Rhoa expression is involved in the pathogenesis of HD. Previous studies have shown increased mRNA levels of RhoA, ROCK and a number of downstream cytoskeleton-associated effector proteins in blood samples and in the frontal cortex of postmortem brain tissue from HD patients and in the striatum of the R6/2 HD mouse model 50. The strongest upregulation in our analysis has been detected for Fcgr2b. Increased expression of Fcgr2b appears to play a crucial role in PD, as it binds aggregated α-synuclein and thus inhibits phagocytosis of microglia through SHP-1 activation 51. It is possible that this mechanism also contributes to the observed reduction of synaptic pruning in Foxp1+/− microglia. Ifi30 is known to play an important role in neurodegenerative processes.52 Altered striatal microglial morphology and reduced synaptic pruning in Foxp1+/− pups could also be normalized by the treatment with MP-10. Our results thus confirm previous studies carried out in mouse models of PD that MP-10, like the Pde10A antagonist papaverine, has anti-inflammatory effects and inhibits both microglial activation and proinflammatory gene expression 18, 19.
There is already good evidence for a contribution of glial cells to the pathophysiology of ASD. Resting microglial cells are known to play an important role in regulating learning and memory, including modulation of memory strength, forgetfulness, and memory quality, through mediation of synaptic pruning. In response to neuroinflammation, microglia are activated and secrete proteins such as cytokines, chemokines, and reactive oxygen species. Through their dynamic morphological and functional properties, they influence synaptic function and plasticity 53. Moreover, glial cell function is associated with an imbalance between excitatory and inhibitory synaptic functions 54, 55. Indeed, brain samples from autistic individuals showed gliosis and increased glial proliferation 30, and animal models of ASD such as Rett syndrome, Fragile X syndrome, and a mouse model of tuberous sclerosis revealed glial abnormalities. Of note, we recently demonstrated mitochondrial dysfunction and increased oxidative stress in the Foxp1+/− striatum and hippocampus 5, 6. Mitochondrial DNA is considered an important activator of inflammation and can lead to inflammasome activation when it escapes from stressed mitochondria 56, 57. For this reason, it is likely that mitochondrial dysfunction in the Foxp1+/− brain contributes to or is even causative for the development of the observed microglial abnormalities. Further studies should clarify whether the improvements in behavior after MP-10 administration are due to an effect of the compound on mitochondrial function and accumulation of ROS in MSNs and possibly microglia.
Alterations in cortical-subcortical circuits have been shown to affect motor function, cognition, and emotional behavior. It is evident today that neurodegenerative diseases such as HD and PD, as well as psychiatric disorders such as schizophrenia (SCZ) and profound developmental disorders such as ASD and anxiety disorders, all have dysfunction of the corticostriatal circuits and imbalances of the direct and indirect basal ganglia pathways as their neurobiological basis 58–60. Moreover, it is well established that movement disorders and altered basal ganglia circuits in HD, PD, and SCZ are associated with changes in PDE10A expression and function 21. In addition, mutations in PDE10A itself cause childhood-onset chorea with striatal lesions 61. Therefore, this molecule is considered a potential therapeutic target for these aforementioned disorders as well as other basal ganglia disorders 21, 62–66. Indeed, autism-like behaviors induced in rats by administration of valproic acid or a serotonin receptor agonist were significantly attenuated by treatment with papaverine, a PDE10A antagonist.65 Furthermore, the treatment had a positive effect on neuroinflammatory processes and oxidative stress. However, compared to MP-10, a potent, orally active and selective PDE10A inhibitor with an IC50 of 0.37 nM, with > 1000-fold selectivity over other PDEs, papaverine has significantly lower potency and selectivity and a very short exposure half-life after systemic administration.27
With regard to HD, it is interesting to note that both HD mouse models and HD patients have reduced levels of PDE10A as well as reduced levels of FOXP1 67–69. Based also on the fact that Pde10a is a Foxp1 target and is decreased in both Nestin-Cre (Foxp1−/−) and Foxp1+/− mice, it is reasonable to assume that reduced Foxp1 expression may also be responsible for the reduced striatal Pde10a levels in HD. Phase II clinical trials with MP-10 (NCT01806896, NCT02197130, NCT02342548) in patients with full-blown symptoms of HD, however, failed to meet the pre-specified endpoints of the trial, despite the promising data in HD mouse models 35, 36. However, this result cannot be extrapolated to FOXP1 syndrome as explained below. First, the timepoint of therapeutic intervention may be crucial for success. It is increasingly understood that therapeutic interventions during sensitive periods in brain development, typically before the onset of symptoms, will affect prospective results. Altered PDE10A expression has already been described well before the symptomatic onset in HD 70, which renders it likely that at the clinical onset, most of the neurons are already affected. This might flag them for destruction, so that dysregulated PDE10A-mediated intracellular signaling may represent an early phenomenon in this late-onset disorder. Second, in cortical neurons expressing mutant htt, elevated Foxp1 expression protected these cells from cell death 67, whereas knockdown of Foxp1 in healthy neurons promoted cell death 67. These findings convincingly show that FOXP1 plays a neuroprotective role in striatal and cortical neurons and that reduced FOXP1 expression causes the selective neurodegeneration of striatal and cortical neurons in HD brains 67. But apart from enlarged lateral ventricles in some individuals with FOXP1 syndrome, there are no obvious signs of striatal neurodegeneration in either humans or mice with loss of a FOXP1 allele 71, 72. Foxp1+/− animals exhibit a decrease in Foxp1 mRNA and protein of ≈ 41% and ≈ 52%, respectively, at P1, P12, and adulthood compared with WT animals 5. Significantly lower FOXP1 levels were also observed in both a HD mouse model and HD patients 70. Therefore, it is reasonable to assume that FOXP1 levels in HD patients gradually decrease with age.
To date, more than 200 individuals with a pathogenic variant in FOXP1 have been documented, but this number is expected to increase due to genetic testing. As the FOXP1 transcription factor and its regulatory networks are highly conserved 73 it is likely that reduced PDE10A expression is a major contributor to the symptomatology not only in mice but also in humans with FOXP1 syndrome. As the 4-week treatment period in newborn mice is equivalent to more than a decade of life in children, further studies can now be carried out to find out whether the start of treatment also at later developmental time points is beneficial and whether inhibition of Pde10a by MP-10 also ameliorates learning and memory deficits during later development.
Most of the individuals affected with FOXP1 syndrome exhibit autistic traits. In addition, there are behavioral problems such as hyperactivity, attention problems, impulsivity, aggression, anxiety, mood instability, obsessions and compulsions. Moreover, many suffer from attention deficit/hyperactivity disorder, often in combination with hyperactivity and inattention. The effect of the antipsychotics administered is largely mediated by blocking the postsynaptic dopamine D2 receptors. Aripiprazole (a partial D2 receptor antagonist) and risperidone (a D2 receptor antagonist) are the most commonly used antipsychotics and the only medications approved by the US Food and Drug Administration for ASD 74, 75. Aripripazole is used for ASD to alleviate irritability, hyperactivity, inappropriate language, and stereotypy 76; risperidone reduces aggression toward others, self-injury, challenging behavior, and rapid mood changes 77. However, both drugs have significant side effects 75, 78 and it is possible that treatment with MP-10 will eliminate the need for both drugs in patients with FOXP1 syndrome.
Overall, our results strongly suggest that individuals with FOXP1 syndrome may benefit from the administration of MP-10 or other potent PDE10A antagonists, especially if treatment is initiated at a young age. Thus, our study may be the first step towards a specific treatment for FOXP1 syndrome that not only alleviates individual symptoms, but also targets one of the overarching causes of the disorder, namely the imbalance between direct and indirect pathway activity due to reduced levels of Pde10a.