RICERCA ITALIANA     

Biochemical and genetic studies of autistic disorder

Abstract

Autism is a severe neuropsychiatric disorder, whose incidence has apparently risen during the last decade from 2-5 to 15-20/10,000 children. This disease is today recognized as the consequence of biochemical alterations, deranging prenatal neurodevelopmental processes. It represents the neuropsychiatric disorder displaying the heaviest genetic component, with heritability estimates greater than 90% according to twin studies.

To study this complex disease, we have set up and coordinate since 1997 a long-standing network of nine clinical groups and four laboratories, which have collected DNA, plasma and urines from 273 simplex and 27 multiplex families, including 328 primary autistic patients, and our recruitment is continuing to reach its final target of 400 families. Since specific subgroups of autistic patients are characterized by “biochemical markers of disease” or “endophenotypes”, namely hyperserotoninemia, oligopeptiduria, and macrocephaly, we already have assessed serotonin blood levels, and urinary peptide excretion rates in 152, and 180 autistic patients, and in 325 and 400 first-degree relatives, respectively. Fronto-occipital cranial circumference has been recorded in 247 patients and 54 unaffected siblings. These data are particularly important because “markers” are more closely related to underlying genetic variants and biochemical pathogenetic mechanisms than the complex, polymorphic and variable clinical symptomatology of autism.

This project is aimed at performing genetic linkage/association studies in our sample of families using polymorphisms located in strong candidate genes likely involved in the biological processes underlying elevated serotonin blood levels (ITGB3 and SLC6A4), enlarged head circumference (OMGP and HOXB1), and urinary loss of small peptides (LRP2 and SLC25A12). In particular, we shall focus upon the SLC25A gene and on its protein product, the mitochondrial aspartate/glutamate carrier (AGC1), studying in parallel the biochemical and genetic levels in postmortem brain tissue specimens from the superior temporal gyrus (Brodmann areas 41/42, 52 or 22) of autistic patients and age-, sex-, and post-mortem interval-matched controls. AGC1 in fact transports aspartate from mitochondria into the cytosol in exchange for glutamate and plays a role in the transport of NADH reducing equivalents from the cytosol into the mitochondria, as a member of malate-aspartate shuttle. A genetic linkage/association study has identified the SLC25A12 gene encoding AGC1, as a susceptibility locus for autism. Finally, we shall produce a genomewide assessment of gene expression levels in these same brain tissue specimens, using the most up-to-date microarray technologies, in order to evidence transcriptional alterations pointing toward the biochemical pathogenesis underlying the disease, and to search for novel genes potentially involved in the disease.

The identification and biochemical characterization of alterations at the DNA and/or expression level either causing autism, conferring vulnerability, or explaining biochemical “markers” associated with this disease, will undoubtedly enhance our understanding of the neurobiological bases of autism, and likely pave the path to earlier and more reliable diagnoses. <<<

Principal Investigator

Antonio Maria Persico Università "Campus Bio-Medico" ROMA

Research Objectives

The objective of the present Research Program is to study the biochemical and neurobiological alterations underlying autistic disorder. In order to reduce the degree of clinical and etiological heterogeneity, while maximizing the probability to identify biochemical and genetic factors involved in the pathogenesis of the disease, we shall employ three different and complementary strategies: [a] the characterization and study of biological markers of disease or “endophenotypes”, i.e. biochemical or morphological parameters characterizing specific subgroups of patients, yielding phenotypes closer to the biochemical and genetic levels than the complex and variable behavioral symptomatology; [b] parallel biochemical and genetic studies of molecules potentially involved in autism pathogenesis and/or in the abovementioned endophenotypes; [c] an hypothesis-free search for autism susceptibility genes, based upon a genome-wide transcriptional assessment performed using microarray technologies.

In reference to the first strategy, the best-characterized endophenotypes in autism research consist of hyperserotoninemia, oligopeptiduria and macrocephaly (i.e., fronto-occipital cranial circumference &gt;97th percentile). Each endophenotype characterizes approximately 20-40% of the patients (see sect. 2.2). We shall thus employ a large sample of patients and first-degree relatives already largely characterized in reference to these three parameters, and we shall study six candidate genes, namely ITGB3 and SLC6A4 for hyperserotoninemia, HOXB1 and OMGP for macrocephaly, LRP2 and SLC25A12 for oligopeptiduria. These genes have been selected either due to strong preliminary evidence, or to a high probability of involvement in the biochemical pathogenesis underlying these endophenotypes and possibly the disease.

The second strategy will be applied to the SLC25A12 gene, recently identified as an autism susceptibility gene (see sect. 2.2), as well as to its protein product, the mitochondrial aspartate/glutamate transporter (AGC1). We shall proceed with functional assessments of transport rates following reconstitution in liposomes, DNA sequencing of both cDNA and genomic promoter regions, and quantification of gene expression, all three assessed in parallel in the same neocortical tissue specimens belonging to autistic patients and controls matched by age, sex, and post-mortem interval. The functional influence of mutations or polymorphisms on mitochondrial aspartate/glutamate transport rates will subsequently be characterized in vitro. Furthermore, we shall employ cellular models to characterize AGC1 roles in neuronal development and differentiation.

Finally, the third strategy will consist in a genome-wide gene expression post-mortem study, performed using microarray technology on total RNA extracted from the same neocortical tissue specimens of patients and controls described above. Therefore this approach represents a broad-based, non hypothesis-driven gene search, aimed at defining the biochemical pathophysiology of autism starting at the transcriptional level.

We expect all parts of the project to be completed in 24 months. Our understanding of the neurobiological bases of autism can significantly benefit from the identification of additional polymorphisms or mutations either conferring vulnerability to autism or directly causing the disease, from the definition of new biochemical markers and from a better understanding of the significance and mechanisms underlying known endophenotypes in autism. These studies hold the potential to yield preventive measures, as well as earlier and more reliable diagnoses, in a disease which, to this date, is still recognized on an exclusively-behavioral basis, leading to frequent misdiagnosing resulting in even more severe impairment despite intensive rehabilitative efforts. <<<

Timescale

24 months

National and international background

Autistic disorder is a severe Pervasive Developmental Disorder of childhood characterized by impaired language, communication and social skills, as well as by repetitive and stereotypic patterns of behavior (1). The central nervous system of individuals with autism processes information by activating neural networks distinct from those employed by non-autistic individuals, particularly for socially-relevant stimuli (2). Altered neurodevelopment, occurring as early as during the first trimester of pregnancy, is widely recognized as the underlying neuropathological cause of the disease. Postmortem studies of autistic brains have described reduced programmed cell death and/or increased cell proliferation, altered cell migration, abnormal cell differentiation with reduced neuronal size, and altered synaptogenesis (3). The cause of these neurodevelopmental alterations is more complex than initially anticipated. On one hand, in no other neuropsychiatric disorder do genetic factors provide as significant a contribution as in autism, with concordance rates of 82-92% in monozygotic twins vs 1-10% in dizygotic twins, sibling recurrence risk at 2-3% and heritability estimates above 90% (4). On the other hand, the most recent studies have conclusively proven the existence of broader genetic heterogeneity, complex epistasis involving 15-20 genes and possibly gene-environment interactions (5). These interactions spur interest particularly in light of the steady increase in autism prevalence recorded starting during the 1980s, with current estimates averaging 15-20/10,000 children vs 2-5/10,000 until the 1970s (6). Finally, additional complexity is added by the phenotypic expression of autism-predisposing genes ranging from minimal autistic traits to full-blown autism, identifying a broad clinical entity referred to as “autism-spectrum disease” (7).

The study of complex and heterogeneous behavioral disorders, such as autism, requires several converging and complementary experimental strategies, as discussed above (see sect. 2.1). For each one of the three strategies employed in this Research Program, we havee briefly summarized below the relevant Literature, especially in reference to autistic disorder:

[a] Identification of “biological markers of disease” or “endophenotypes”, i.e., biochemical or morphological parameters characterizing specific subgroups of patients.

The biochemical and morphological endophenotypes most studied and consistently proven reliable in autism research include serotonin (5-HT) blood levels, cranial circumference and urinary peptide excretion rates. Their main features, and the candidate genes most likely involved in these endophenotypes, are discussed below:

Serotonin blood levels: consistently elevated in at least 25% of autistic patients, they represent a marker for familial forms of the disease (9,10). Elevated 5-HT blood levels derive from increased density of functionally-active 5-HT transporter molecules on platelet membranes from autistic patients, with no changes in affinity for 5-HT and no elevation in free 5-HT plasma level (11,12).
One of the primary candidate genes for this endophenotype and for autism is thus represented by the SLC6A4 gene (i.e., 5-HT transporter or 5-HTT), also considering the neurotrophic roles exerted by 5-HT during development (13). The 5-HT transporter responsible for platelet 5-HT uptake is identical in its primary sequence to the 5-HTT expressed in serotoninergic neurons: both are indeed produced by a single gene, located on chromosome 17q12. A variable-number-tandem-repeats (VNTR) polymorphism located in the promoter region and known as 5-HTTPLPR profoundly affects 5-HTT gene expression (14). We have previously shown that the 5-HTTPLPR is not associated with autism and provides little, if any, contributions to determine serotonin blood levels in autism (15,16). Subsequently, however, Kim et al (17) have described significant linkage/association between autism and a VNTR located in the second intron.
Another candidate gene of interest has been recently identified by Weiss et al (18), who have found a significant association between the integrin beta-3 (ITGB3) Leu33Pro polymorphism and serotonin blood levels using quantitative trait locus (QTL) analysis on a large sample of Hutterites, a founder population.

Cranial Circumference: macrocephaly (i.e., head circumference &gt; 97th percentile) is present in approximately 20% of autistic patients (19,20). We have shown that the HOXA1 A218G polymorphism affects cranial circumference in autistic patients (21). Similarly to HOXA1, also the HOXB1 gene plays an important role in brainstem and cranial morphogenesis, and has been found associated with autism in at least one initial study (22). The potential relationship between HOXB1 gene variants and macrocephaly has not been addressed.
Brain imaging studies suggest the probable existence of alterations in neurite pruning and myelination in autism (23,24). The oligodendrocyte and myelin glycoprotein (OMGP) is heavily involved in myelination of the central nervous system, a process that is critical to the determination of head circumference and that may be altered in the brains of some autistic patients. The OMGP gene, located at (D17S250) linked to autism in several independent genome scans, was previously shown by Vourc’h et al (25) to be associated with autism in a sample of French autistic patients.

Urinary peptide excretion rates: Oligopeptides, 3-10 aminoacids in length, are physiologically filtered by the renal glomeruli and undergo tubular reabsorption through an LDL receptor family member, the LRP2 protein, also known as gp330 or megalin (26). Oligopeptide binding to LRP2 is followed by endocytosis of the oligopeptide-receptor complex and digestion by peptidases. Up to 60% of autistic children display significantly enhanced urinary oligopeptide excretion rates compared to age- and sex-matched controls (27). This finding, consistently present in indipendent samples of autistic patients although to a different extent, has been confirmed in our sample where approximately 38% of autistic patients display urinary peptide amounts significantly higher than those of age- and sex-matched controls. Reduced LRP2 amounts and/or altered LRP2 function could explain the enhanced urinary excretion rates present in this subgroup of autistic patients. In addition, LRP2 is expressed in the trophoblast and plays an important role in placental uptake of maternal LDL cholesterol, which is essential to normal neurodevelopment particularly during the first trimester of pregnancy, as suggested by studies of LRP2 knock-out mice (28).
Following LRP2 binding to oligopeptides, tubular uptake is an energy-dependent endocytotic process (26). Mitochondrial function is therefore essential to this second phase of the process. Interestingly, mitochondrial disorders can be associated with seizures (and 30% of autistic patients have a history of seizures), repetitive and stereotypic behaviors, and autism (29). The SLC25A12 gene, encoding the mitochondrial aspartate/glutamate carrier (AGC1), has been found strongly linked and associated with autism (30). This is particularly interesting, given the influence exerted by AGC1 on the energy balance of the cell through the aspartate/malate NADH shuttle (31). Finally, both the LRP2 and the SLC25A12 loci are located on human chromosome 2q31-q32.2, within a region yielding very promising lod scores in several genome scans performed to date on affected sib-pair families (30,32).
For these reasons, the SLC25A12 gene will be studies more thoroughly both at the biochemical and genetic levels.


[b] Parallel biochemical and genetic characterization of molecules potentially involved in the above-mentioned endophenotypes:

The SLC25A12 gene, recently identified as an autism susceptibility gene (30), encodes the mitochodrial aspartate/glutamate transporter (AGC1), which catalyzes the transport of aspartate from mitochondria into cytosol in exchange for glutamate and plays a role in the transport of NADH reducing equivalents from cytosol into mitochondria, as a member of the malate-aspartate shuttle (MAS) (33). AGC belongs to the mitochondrial carrier family whose members catalyze the transport of various metabolites and cofactors essential for the coordination of mitochondrial metabolism to cell energy status in various physiopathological conditions (34-36). Their molecular characterization has revealed that, in spite of the wide variety of substrates and kinetic mechanisms, they all belong to the same protein family (34-36). Members of the mitochondrial carrier family have a tripartite structure, made up of three related sequences about 100 amino acids in length. Each repetitive domain contains two hydrophobic stretches and a characteristic sequence motif P-h-D/E-X-X-K/R-h-R/K-(20-30 amino acids)-D/E-G-(4 amino acids)-a-K/R-G (where h = hydrophobic, a = aromatic and X = variable). A breakthrough was recently achieved when Pebay-Peyroula et al. (37) succeeded in solving the crystal structure of the bovine ADP/ATP carrier whose monomer appears as a six-helix bundle with a pseudo three-fold symmetry consistent with the available structural information and the topological models previously built for various members of the mitochondrial carrier family.
The structure of the two paralogues AGC1 (encoded by SLC25A12 on ch. 2q31) and AGC2 (encoded by SLC25A13 on ch. 7q21) share 77.8% identical amino acids and are different from those of most mitochondrial carriers, due to a long N-terminal extension harbouring EF-hand Ca2+-binding motifs. We have previously shown that AGC is activated by calcium binding in the mitochondrial intermembrane space and thereof it is involved in calcium signalling (31).
AGC1 is expressed in almost every tissue with the highest expression in brain, skeletal muscle and heart (38). In the brain, AGC1 is mainly expressed in neurons (39). AGC2 is predominantly expressed in liver, kidney and heart. Its deficiency causes adult-onset type II citrullinemia (CTLN2) (40), which is characterized biochemically by a liver-specific deficiency of argininosuccinate synthetase (ASS) and clinically by psychotic symptoms (41). During embryogenesis, AGC1 and AGC2 have a widely overlapping expression pattern, and the full expression of AGC1 is only attained postnatally (39). AGC1 levels increased during neuronal differentiation and are correlated with an increase in the malate/ aspartate NADH shuttle activity. Accordingly, AGC1-/- mouse embryos are produced in normal numbers (42); however, soon after birth many AGC1-/- mice develop severe growth defects and neuromuscular deficiencies (42). Along with that, AGC1-/- mice have a reduced N-acetyl aspartate (NAA) content and myelination in the central nervous system, but not in peripheral nerves (42). Interestingly, NAA content was shown to be reduced in certain brain regions of autistic patients (43).On the other hand, the presence of overexpression of AGC1 can lead to increased mitochondrial ATP production (44), so genetic variations that change the expression of AGC1 would be predicted to impact ATP production. Altered expression could result in an altered transduction of calcium signalling. Mitochondria are in fact important regulators of cellular calcium homeostasis and calcium concentrations regulate the production of ATP via the Krebs cycle and oxidative phosphorylation. Recent work has shown that AGC1 plays an essential role in the transduction of small calcium signals to mitochondria in neurons (45). Further evidence suggesting that abnormal calcium signalling may contribute to autism also comes from a recent study showing that Timothy syndrome (TS), a complex multisystem disorder affecting multiple organ systems including heart and brain, is caused by de novo missense mutations in Cav1.2, an L-type calcium channel that is ubiquitously expressed in man (46,47). Notably, children suffering from TS have developmental delays including language, motor and generalized cognitive impairment and a significant proportion of patients suffering meet the criteria of autism spectrum disorders. Functional expression of mutant channels in heterologous systems demonstrated that the mutations lead to nearly complete absence of voltage-dependent inactivation that is likely to induce intracellular calcium overload (46,47).



[c] Broad-spectrum, non hypothesis-driven search for autism susceptibility genes, performed using functional transcriptional approaches:

Another avenue of investigation toward the identification of candidate genes involved in complex disorders, such as autism, is represented by genome-wide expression studies performed using microarray technology on total RNA extracted post-mortem from brain tissue of patients and matched controls (48). Despite some limitations particularly inherent to the quality of the available tissue samples, studies of this sort have already provided significant contributions to our understanding of other complex behavioral disorders, such as schizophrenia (49). Furthermore, the advantage of using total RNA-based strategies rather than proteomic analyses is that the total RNA is significantly more labile than proteins, yet its quality is much easier to control, providing greater reliability when the starting tissue comes from postmortem specimens (49). To this date, whole-genome expression studies of autistic brains implementing microarray technologies are still to be performed, with the single exception of one microarray study reporting gene expression patterns in the cerebellar cortex of autistic patients and controls (50). This study identified alterations in gene expression particularly affecting genes encoding molecules involved in glutamatergic neurotransmission (50). Interestingly, these results led to two subsequent genetic studies both reporting positive associations of glutamate receptor subunit-encoding genes with autism (51,52), confirming the usefulness of this approach. Finally, since the publication of this initial microarray study, several other brain regions in addition to the cerebellum have been shown to be structurally and functionally altered in autism, particularly the superior temporal sulcus (2,53) which will be the object of a set of experiments proposed in this application. <<<