The present invention relates to a novel therapeutic indication for antagonists of dopamine D2 receptors, particularly antipsychotic drugs.

The aetiology of behavioural disorders associated with impaired neurological development, such as schizophrenia, autistic spectrum, and attention deficit hyperactivity disorder (ADHD), is complex and widely unknown, although the scientific literature reveals that there is a high genetic component. Even though many potential risk genes have been iden- tified, the measured effects are weak and the repeatability is not always consistent, probably due to complexity of human polymorphisms, genetic and clinical heterogeneity, and the potential impact of gene-gene and gene-environment interactions.

In this context, the generation of murine models bearing specific genetic variations in risk genes represents an unparalleled tool for studying the neurobiological bases of such highly debilitating diseases and for improving the associated therapies.

In schizophrenia, autism, and ADHD, cognitive alterations are the main long-lasting symptoms that may contribute to exacerbating deficiencies in the social and professional life of patients. These symptoms currently represent an important, and up to now unmet, therapeutic need, since they are still incurable. Therefore, there is a strong need to provide drugs that are effective in the treatment of cognitive alterations linked to schizophrenia, autism, and ADHD. The dysbindin-1 protein encoded by the gene dystrobrevin-binding protein 1 (DTNBP1) is expressed at the synapses both in the human and the murine brain. This protein is involved in the regulation of exocytosis and biogenesis of vesicles both in endocrine cells and in neurons. Genetic variations in DTNBP1 have been suggested to affect cognitive abilities both in humans and in mice (Burdick et al., Human Molecular Genetics, 2006, Vol. 15, No. 10 1563-1568; Karlsgodt et al., Biol Psychiatry, 201 1,69:28-34; Wolf et al., Molecular Psychiatry (2009), 1-11 ; Papaleo et al., Molecular Psychiatry (2010), 1-1). Moreover, several studies have linked genetic variations in DTNBP1 to an increased risk of developing schizophrenia. In patients affected by schizophrenia, decrease in the expression of the dysbindin-1 gene was actually detected in the hippocampus and the prefrontal cortex (PFC).

By investigating the functional genetic variations in the DTNBPl gene and how they can modulate cognitive functions and dysfunctions, the present inventors demonstrated that a decrease in the levels of the dysbindin-1 protein increases the expression of D2, but not Dl, dopamine receptors on the surface of mice cortical neurons. Without wishing to be bound by any theory, this is probably due to an altered reinsertion of the D2 receptors on the neuronal membrane, likely caused by alterations in the lysosomal trafficking. Con- versely, the total levels of Dl and D2 receptors do not vary in the cortex of genetically engineered mice with decreased dysbindin-1 compared to wild type mice. Only the levels of D2 receptors on the cell surface are increased.

On the basis of these observations, the inventors studied the effects of per se known active ingredients capable of interacting with D2 dopaminergic receptors, such as for example conventional antipsychotics and other drugs that act generally on dopaminergic pathways, in order to modulate cognitive functions and dysfunctions.

Particularly, in the experimental part that follows, data are provided which demonstrate modulation of antipsychotic drugs (such as for example risperidone) by genetic modifications which result in decreased levels of the dysbindin-1 protein in genetically modified mice, in healthy human subjects, and in patients affected by schizophrenia. In mice, the decrease of dysbindin-1 has been studied by using heterozygous (dys+/-) or homozygous (-/-) knockout mice. In human beings, the decrease of dysbindin-1 has been studied by using a haplotype consisting of a set of single nucleotide polymorphisms (SNPs) comprising the SNP T allele of Reference SNP cluster ID (rs) 2619538, the A allele of rs3213207, and the A allele of rs 1047631.

Further dysbindin-1 polymorphisms known to be associated with schizophrenia, not con- nected with a reduction in the protein levels, are rsl018381, rs2619522, rslOl 1313, rsl6876759, rs2005976, rs2619539, rs6675281, rs821616, rs909706, rs760761, rs2619528 (Corvin A et al., Neurosci Lett. 2008 Jan 31 ;431(2): 146-9; Trost S et al., Eur Arch Psy- chiatry Clin Neurosci. 2013 Feb;263(l):53-63; Trost S et al., J Psychiatr Res. 2013 Feb;47(2): 188-96; Burdick KE et al., Schizophr Res. 2007 Jan;89(l-3):169-72; Straub RE et al., Am J Hum Genet. 2002 Aug;71(2):337-48). The results obtained by the present inventors show that, specifically in subjects bearing a functional genetic variation in the DTNBP1 gene capable of decreasing the levels of the encoded dysbindin-1 protein, the conventional antipsychotic drugs and more generally substances that act through D2 receptors of the dopaminergic system (antagonists of D2 dopaminergic receptors) are capable of improving cognitive dysfunctions.

This result is surprising as the literature of the art available up to now indicates that the antipsychotic drugs currently used are not effective, and may even be deleterious, in the treatment of cognitive deficiencies in subjects suffering from schizophrenia. See, for instance, Young JW et al. (2009) Pharmacol Ther 122:150-202, in which it is described that cognitive deficiencies in schizophrenia cannot be treated with the conventional antipsychotic therapies; and Michael J. Minzenberg and Cameron S. Carter, Trends in Cognitive Sciences. 2012;16(l):35-42, in which it is described that antipsychotic drugs are ineffective in the treatment of cognitive dysfunctions. According to the Diagnostic and Statistical Manual of Mental Disorders of the American Psychiatric Association, fifth edition, DSM- 5, pages 88, 92 and 100-101, cognitive deficiencies can develop in patients affected by schizophrenia, but it does not fall within the group of the diagnostic features among which negative symptoms, among others, are included, such as diminished emotional expression, avolition, alogia, anhedonia, asociality. The present invention is surprising and unexpected also because from the data obtained by the inventors and detailed hereunder, as well as according to the state of the art, it is inferred that genetic variations in DTNBP1 which decrease dysbindin-1 worsen by themselves the cognitive abilities in humans or mice not treated with D2 receptor antagonists. In this connection, Bray NJ et al. (2005) Hum Mol Genet 14:1947-1954 is mentioned, for ex- ample. On the contrary, these same genetic modifications predict an improvement of the cognitive functions subsequent to use of D2 receptor antagonists, such as for example antipsychotic drugs. The scientific literature describes the correlation between some specific genetic polymorphisms and the effectiveness of antipsychotic treatments in patients suffering from psychiatric diseases, such as for example schizophrenia. In this respect, see for instance Patent Application US 2011/0287417, which however does not mention the DTNBPl gene or the treatment of cognitive deficits and negative symptoms of schizophrenia, and Patent Application EP 2570501 A, which refers specifically to polymorphic markers in ZNF659. On the other hand, United States Patent Application US 2009/0233942 describes the correlation between a few markers, including DTNBPl, and the response to antidepressant drugs. Therefore, one aspect of the present invention is a substance selected from the group consisting of antagonists of D2 dopaminergic receptors for use in the modulation of cognitive dysfunctions in a subject bearing a functional genetic variation in the DTNBPl gene capable of diminishing the levels of the dysbindin-1 protein. The subject preferably suffers from schizophrenia, autistic spectrum, or attention deficit hyperactivity disorder (ADHD).

The decrease in the levels of the dysbindin-1 protein refers to a subject defined as not bearing any of the following SNPs: the T allele of rs2619538, the A allele of rs3213207, and the A allele of rs 1047631.

In an even more preferred embodiment, the genetic background of interest is defined by the haplotype consisting of the following SNPs: rs2619538-rs3213207-rsl047631. Preferred substances suitable for use according to the present invention are antagonists of D2 dopaminergic receptors, such as for example antipsychotic drugs. Non-limiting examples of antipsychotic drugs suitable for use according to the invention are risperidone, ha- loperidol, olanzapine, clozapine, aripiprazole, quetiapine, ziprasidone, sertindole, sulpirideamisulpiride, zotepine, and combinations thereof. These drugs, and pharmaceuti- cal forms and dosages thereof are per se known and thus may be selected and used by those skilled in the art without requiring undue experimentation. The present invention is described in further detail in the following experimental section, which is provided for illustration purposes only and not as a limitation.

In the experimental section, reference is made to the appended drawings and to Table 1, wherein:

Figure 1 shows the results for the attentional set-shifting deficit in dysbindin-1 mutant mice, (a) Total number of trials, (b) time (minutes), and (c) latency (seconds) in giving a response required for reaching the criteria at the various stages of the ID/ED task grouped by wild-type dys +/+ (n = 10) and dys +/- (n = 7). (d) Total number of mice tested in the ID/ED task, (e) Number of days for completing the ID/ED task. Analysis of the performance of dys +/+ mice revealed a significant effect of the stages in relation to the trials (F (8,72) = 6.74, p O.0005) and the time (F (16,72) = 6.85, p O.0005) required for reaching the criterion. Particularly, dys +/+ mice needed more trials (p <0.05; fig. lb) and more time (p <0.05; fig. lc) for the resolution of the EDS compared to CD, IDS, IDS2, IDS2Re, and the EDSRe stage. Analysis of the EDS results showed a significant effect of the genotype on the trials (F (1,15) = 7.69, p <0.05, fig. la) and the time (F (1,15) = 4.75, p <0.05, fig. lb) required for completing the stage. Analysis of the latency of response showed a significant effect of the stages (F (8,240) = 11.40, p O.0005). Particularly, as previously reported (Scheggia D et al. (2014) Biol Psychiatry 75:660-670), the mice improved their response speed in subsequent stages, as demonstrated by a significant decrease in the latency of response to IDS2 compared to the starting discrimination (SD; fig. lc). Furthermore, a significant effect of stage x genotype was present on the latency of response (F (2,60) = 3.43; P = 0.03, fig. lc); * p <0.05, ** p O.005 compared to the performance of +/+ mice in the same stage. The values represent mean ± SEM in Figures 1-3.

Figure 2 shows the results for the modulation of the executive cognitive function of attentional set-shifting by Bray haplotype (rs2619538-rs3213207-rsl047631) in healthy subjects and patients with schizophrenia treated with antipsychotic drugs. Total number of (a) perseverative errors, (b) non-perseverative errors and (c) categories achieved during WCST in healthy volunteers without the Bray haplotype (DysBray-/-) and in subjects bear- ing one or two copies (DysBray +/- and +/+) of the haplotype. Total number of (d) perseverative errors, (e) non-perseverative errors and (f) categories achieved during WCST in patients with schizophrenia without the Bray haplotype (DysBray-/-) and in patients bear- ing one or two copies (DysBray +/- and +/+) of the haplotype. Analysis of the performance of healthy volunteers during WCST revealed a significant effect of the genotype both on the perseverative errors (F (1,204) = 3.57, p = 0.05, fig. 2a) and the non-perseverative ones (F (1,204) - 4.17, p O.05, fig. 2b) made in the WCST. Furthermore, a significant effect of the genotype was also seen in patients with schizophrenia, both on the perseverative errors (F (2,140) = 3.42, p = 0.03, fig. 2d) and the non-perseverative ones (F (2,140) - 3.45, p = 0.03, fig. 2e). Moreover, a significant effect of the genotype was also observed in the analysis of the categories achieved in WCST (F (2,140) = 3.79, p = 0.02, fig. 2f). * p < 0.05 vs DysBray-/-.

Figure 3 shows that risperidone (daily dose of 0.1 mg/kg) improves the attentional set-shifting only in mice with a decrease in dysbindin-1. (a) Total number of trials (b), time (minutes), and (c) latency (seconds) in giving a response required for reaching the criteria during the EDS of the ID/ED task grouped by dys +/+ Veh (n = 6), dys +/+ Ris (n = 7), dys +/- Veh (n = 6) and dys +/- Ris (n = 6). (d) Total number of mice tested after chronic treatment with the Vehicle or Risperidone in the ID/ED task, (e) Number of days for completing the ID/ED task. Analysis of the EDS results showed a significant effect of genotype x treatment on the trials (F (1,21) = 9.05, p = 0.006, fig. 3a) and the time (F (1,21) = 3.88, p = 0.05, fig. 3b) required for completing the stage. Analysis of the latency of response during EDS showed a trend towards a significant effect as regards the interaction treatment x genotype (F (1,44) = 3.09, p = 0.08, fig. 3c). *p O.05 compared to the performance of dys +/+ Veh and dys +/+ Ris in the same stage. # # p <0.005 vs dys +/- Veh.

Table 1 shows the demographic features of the human subjects tested. No significant differences were observed between the three groups of healthy subjects with respect to age (F (2,203) = 0.82, p = 0.44), education (F (2,203) = 0.56, p = 0.57) and Mini Mental State Examination (MMSE) (F (2,203) = 0,20, p = 0.79) which estimates intellectual efficiency disturbances and the presence of cognitive impairment. Likewise, from analysis of demographic data of patients suffering from schizophrenia no differences were detected between the groups in age (F (2,155) = 0.20, p = 0.81), education (F (2,140) = 0.99, p = 0.37) and MMSE (F (2,140) = 0.70, p = 0.49). EXPERIMENTAL SECTION

For the genotyping, genomic DNA was extracted from blood samples by using standard protocols. SNPs present on the DTNBPl gene (rs2619539, rs3213207, rsl047631) were specially selected on the basis of initial studies (Straub RE et al., Am J Hum Genet. 2002 Aug;71(2):337-48) and further analyses that show association with schizophrenia (Williams NM et al. (2004) Arch Gen Psychiatry 61 :336-344; Bray NJ et al. (2005) Hum Mol Genet 14:1947-1954), as listed in the SZGene database (Allen NC et al. (2008) Nat Genet 40:827-834). DNA samples were subjected to genetic analysis for SNPs by using custom- ized Taqman allelic discrimination assays for genotyping (Applied Biosystems) in a realtime PCR AB 17900 (Applied Biosystems). For the genotyping, the protocol recommended by the manufacturer was used. For the amplification and the allelic discrimination, ABI Prism 7900 HT Sequence Detection System and version 2.1 of the SDS software (Applied Biosystems) were used. Genotypes were assigned automatically for allelic discrimination data with SDS, and manually for individual absolute quantification data, both resulting in concordance rates of 100%. Quality values were > 98.8% (genotyping error rates: <1.2%) for all SNPs with the exception of SNP rs 1000117, which therefore was further genotyped via direct sequencing. All polymorphisms were in Hardy- Weinberg equilibrium (P> 0.05; assessed by online resource www.kursus.kvl.dk/shares/vetgen/Popgen/Genetik/applets/kites t.htm, for genotype frequencies, see Table 1).

Protocol:

TaqMan® SNP Genotyping Assays, human: rs2619538 C 3114517_10 rs 1047631 C 7460562_10 rs3213207 C_32386418_10

and TaqMan Universal Master Mix without AmpErase®UNG, real time 96-well plates and the real time ABI7900 instrument were used, all from Applied Biosystem. Each SNP TaqMan® genotyping assay contains two probes labelled at the 5' end for detecting the allele specific for the particular polymorphism of interest. In particular, probe for allele 1 displays the VIC® dye, whereas that for allele 2 displays the FAM® dye. The protocol recommended by the kits user manual was followed:

0.5 microliters of purified genomic DNA (at a starting concentration of 20-50 ng/microliter)

1.25 μΐ 20X Taqman assay

12.5 μΐ Taqman MasterMix

10.75 μΐ water All samples were analysed in duplicate.

A first "standard curve" protocol was carried out:

95°C for 10 minutes,

then 92°C for 15 seconds and 60°C for 1 minute for 40 rounds.

After completion of the PCR reaction, the samples were read for allelic discrimination. The allelic discrimination protocol consists of reading and normalizing the fluorescence of the allele-specific markers (FAM and VIC) with respect to the reference dye in each well. The software provides a graphical representation of the intensity of the two markers for each sample (allelic discrimination plot). In addition to the negative control, the plot groups the samples into three classes, one for each genotypic class: samples homozygous for allele 1 for which only VIC fluorescence was detected, heterozygous samples wherein both VIC and FAM fluorescences were detected, samples homozygous for allele 2 for which only FAM fluorescence was detected.

1. Decreased expression of dysbindin-1 in mouse causes cognitive deficits in the range of executive functions.

In a first set of experiments, the inventors studied how the reduced expression of dys- bindin-1 (that is dys+/-) in mouse can alter the executive cognitive functions that depend on the prefrontal cortex. In order to study the effect of genetic variations in the DTNBP1 gene on the executive cognitive functions, dys +/- mice (with reduced levels of dysbindin- 1) and their control dys +/+ conspecifics (with normal levels of dysbindin-1) were tested in the ID/ED Operon task previously described in Scheggia D et al. (2014) Biol Psychiatry 75:660-670. Five dys+/- mice were excluded as they were not able to complete the test, whereas only one dys +/+ failed to finish the entire test (fig. Id). All the other mice learned to make the whole sequence of presented discriminations without any difficulty. Nevertheless, genetically modified mice with decreased levels of dysbindin-1 (that is dys +/-) took more days to complete the test (p<0.05; fig. le).

In the first place, the performance of dys +/+ mice in the ID/ED test was shown to follow the same pattern displayed by normal mice, as previously reported (Scheggia D et al. (2014) Biol Psychiatry 75:660-670), as they needed more trials (p <0.05; fig. la) and more time (p <0.05; fig. lb) for solving the Extra-dimensional Shift stage (EDS) compared to Compound Discrimination (CD), Intra-dimensional Shift (IDS), Intra-dimensional Shift 2 (IDS2), Intra-dimensional Shift 2 Reversal (IDS2Re) and the Extra-dimensional Shift Reversal stage (EDSRe). It was found that dys +/- mutant mice needed more trials (p <0.05; fig. la) and more time (p <0.05; fig. lb) for solving the EDS stage, compared to their dys +/+ littermate controls, which is the most critical stage for assessing the executive cogni- tive abilities that are dependent on the prefrontal cortex, and is similar to that found from errors made by human subjects tested in the Wisconsin Card Sorting Test (WCST). By contrast, no differences relevant to dysbindin-1 genotype were detected in any other stage of the ID/ED cognitive test. Furthermore, dys +/- mice need more time to make a choice compared to mice with normal levels of dysbindin-1 (p <0.005, fig. lc), as demonstrated by an increase in the latency of response, which is considered as an indication of decision making (Robbins TW (2002) Psychopharmacology (Berl) 163:362-380; Scheggia D et al. (2014) Biol Psychiatry 75:660-670), only during EDS. Altogether, these results demonstrate that genetic modifications with consequent reduction of dysbindin-1 give rise to cognitive deficits in the range of executive functions.

2. Bray haplotype in the DTNBP1 gene modulates the executive cognitive abilities in human beings. On the basis of the results obtained in the dysbindin-1 genetically modified mice (fig. 1), the inventors carried out a number of experiments in order to verify whether functional genetic variants of the DTNBPl gene resulted in the same cognitive effects also in humans in a cognitive test similar to that used in mice and dependent on the prefrontal cortex. For this reason, genetic variations in the DTNBPl gene generated by a haplotype consisting of three SNPs (rs2619538-rs3213207-rsl047631), hereinafter referred to as "Bray haplotype" (Bray NJ et al. (2005) Hum Mol Genet 14:1947-1954), which was previously shown to decrease the expression of dysbindin-1 mRNA in the human brain (in particular, dysbindin-1 levels will be affected as follows: DysBray -/- > DysBray +/-> DysBray +/+), were studied. Thus, the performance of 206 healthy Caucasian volunteers was assessed in WCST, which is largely used for measuring and testing executive cognitive functions in humans, and which is sensitive to PFC functions (Robbins TW, Roberts a C (2007) Cereb Cortex 17 Suppl 1:151-160; Barnett JH et al. (2010) Neurosci Biobehav Rev 34:1161- 1177). The three groups of subjects divided according to the DysBray genotype did not dif- fer in their demographic features (Table 1). Subjects who display Bray haplotype (DysBray +/- and +/+) were shown to make more perseverative errors (p<0.05, fig. 2a) and non- perseverative errors (p <0.05, fig. 2b) compared to subjects of the same age not displaying Bray haplotype (DysBray-/-) in the WCST. Reasonably, no differences were observed in the number of categories completed during the test (fig. 2c, p = 0.45), since the task is rela- tively easy for healthy adult subjects. However, an increase in both the perseverative and non-perseverative errors (or "random") indicates a greater difficulty in solving this cognitive test by bearers of the DysBray haplotype. So, as predicted by the results from mice with decreased dysbindin-1 expression (fig. 1), subjects bearing Bray haplotype (DysBray +/- and +/+), which translates into a relative decrease in dysbindin-1 expression, were shown to have cognitive disadvantage in executive abilities.

3. Bray haplotype produces opposite effects on executive cognitive abilities in patients with schizophrenia. Given the deep impact of cognitive dysfunctions on the functional outcome of schizophrenia, the inventors carried out some experiments in order to verify whether Bray haplotype was able to modulate the executive cognitive abilities in patients with schizophrenia. The performances of 158 subjects who met the diagnosis criteria of the Diagnostic and Statistical Manual of Mental Disorders IV (DSM-IV) for schizophrenia were analysed. 17 patients were under treatment with first-generation antipsychotics (haloperidol), whilst 77 were treated with second-generation antipsychotics (risperidone n = 21 ; olanzapine n = 19; clozapine n = 17; aripiprazole n = 11; quetiapine n = 9); 64 patients instead took a combination of these drugs. The three groups of patients divided by their DTNBPl genotype (DysBray -/-, +/-, +/+) did not differ in their demographic features (Table 1). Surprisingly, it was found that patients bearing the haplotype that reduces the levels of dysbindin-1 (DysBray +/- and +/+) obtained better cognitive results than patients without the Bray hap- lotype in the DTNBPl gene (DysBray-/-). Particularly, DysBray +/+ patients made fewer perseverative (p <0.05, fig. 2d) and non-perseverative (p <0.05, fig. 2e) errors compared to DysBray-/- patients. Consistent with this, DysBray +/+ patients were shown to have completed more categories in WCST (p <0.05, fig. 2f) compared to DysBray-/- patients. Likewise, DysBray +/- patients made fewer perseverative errors (p <0.05) than DysBray-/- pa- tients.

These results demonstrate that genetic variations associated with a decreased DTNBPl expression (in this case bearers of Bray haplotype DysBray +/-, +/+) give cognitive advantages in the range of executive functions in patients treated with antipsychotic drugs com- pared to patients who do not show this haplotype (DysBray -/-). Furthermore, these results point out that genetic variations in the DTNBPl gene modulate the executive control in the opposite way in healthy subjects not pharmacologically treated and in patients with schizophrenia treated with antipsychotic drugs. 4. Chronic treatment with the second-generation antipsychotic Risperidone gives improvement in the case of cognitive disorders in dysbindin-1 mutant mice.

Subsequently, the inventors carried out further experiments to understand why, unlike healthy subjects, patients with schizophrenia bearing Bray haplotype showed better out- comes compared to non-bearing patients. It has been suggested that the effect of antipsychotic drugs taken by these patients could explain the differences in their cognitive performances dependent on genetic variations in the DTNBPl gene. In order to verify this hy- pothesis, the murine dysbindin-1 mutant model of the inventors was used to test the effect of risperidone, one of the second-generation antipsychotics currently most used as a pharmacological therapy in patients with schizophrenia. Particularly, mice with reduced levels of dysbindin-1 (dys+/-) and their control conspecifics (dys+/+) were treated with 0.1 mg/kg risperidone (Ris) or the vehicle (Veh) once daily for two weeks and then were tested in the ID/ED task to assess their executive abilities of attentional set-shifting. One dys +/+ Veh mouse, two dys +/+ Ris mice, four dys +/- Veh mice and four dys +/- Ris mice were excluded because they failed to complete the entire ID / ED procedure (fig. 3d). Dys +/- mutant mice treated with risperidone completed the test, taking fewer days than dys +/- mice treated with the vehicle (p<0.05).

Treatment with risperidone significantly improved the performance of dys +/- mice, in fact they completed the EDS with fewer trials (p <0.005; fig. 3a) and less time (p <0.05; fig. 3b) compared to the dys +/- group treated with the vehicle and to the dys+/+ groups. Thus, risperidone enormously improved the executive cognitive abilities in dys +/- mice, which, otherwise, were again confirmed to be dysfunctional after treatment with the vehicle (p <0.05; fig. 3a). No other stage of the test was influenced by treatment with risperidone (p = 0.58; fig. 2). On the contrary, chronic treatment with risperidone did not affect performances of dys +/+ control mice (p = 0.71).

These results demonstrate that antipsychotic drugs, such as risperidone, improve the executive cognitive abilities measured by the attentional set-shifting only in animals with a genetic modification resulting in partial reduction of dysbindin-1. Moreover, these results are the biological validation that genetic mutations that decrease dysbindin-1 are the cause of the otherwise surprising cognitive improvement observed in patients with schizophrenia under treatment with antipsychotics and bearing genetic mutations that reduce the levels of dysbindin-1. Table 1