Characteristics of the tissue used. Dx, diagnosis; PMI, post-mortem interval.
Autism is a severe neuropsychiatric disorder characterized by impaired communication, significant reduction in social interaction, and repetitive and stereotyped behaviour. It is highly hereditable (Hoekstra et al., 2007); however, genomic alterations associated to autism have been found only in less than a fifth of the total number of cases. How those alterations ultimately cause the autistic phenotype is still very poorly understood. Besides genomic abnormalities, environmental and epigenetic factors may also increase the risk of developing autism or autistic traits. Prenatal exposure to rubella virus, cytomegalovirus, or to the chemical substances thalidomide and valproate are among the non-genetic causes linked to autism (Persico & Bourgeron, 2006), however causal relationships are not established. Regardless of the origins of autism, neuropathological observations are consistently found in several areas of autistic brains (Bauman & Kemper, 1985; Ritvo et al., 1986), and abnormal patterns of synaptic connectivity are thought to be at the core of the autistic disorder (Belmonte et al., 2004). Indeed, many of the genes associated with high risk for autism and those increasing susceptibility are directly, or indirectly, involved in axon guidance, neuronal signalling, metabolism, cell differentiation and synaptic homeostasis (Weiss et al., 2009; Autism genome project consortium, 2007; Tabuchi et al., 2007; Toro et al., 2010). Therefore, along with an early diagnosis (Limon 2007), the prevention and correction of the abnormal connectivity, and the modulation of the synaptic function are the main goals of current and future treatments of the pathological characteristics of the autistic disorders.
Glutamate and GABA are the main excitatory and inhibitory neurotransmitters in the human brain and both have important roles during early development of the nervous system, an ontological stage when the evidence indicates that autism begins. Therefore, it is important to analyse the functional status of glutamatergic and GABAergic neurotransmission in the autistic brain. Cumulative evidence indicates that dysfunctional excitatory and inhibitory synaptic activities underlie several of the characteristics of autism and are, consequently, important targets of pharmacological intervention. In this chapter we describe how glutamate and GABA receptors may participate in the aetiology of autistic disorders and will discuss some of the methods that we have developed to study functional and pharmacological properties of human membrane receptors. We include information demonstrating that functional studies of GABA and glutamate receptors from autistic tissue are feasible and important for the development of new drugs aimed at the manipulation of the GABAergic or glutamatergic systems in the autistic brain.
2. Evidence of glutamatergic dysfunction in autism
Glutamate is the main excitatory neurotransmitter in the vertebrate brain and its effects are exerted mainly through metabotropic (mGlu) and ionotropic (iGlu) glutamate receptors localized in the cellular membranes of neurons and glia. The iGlu receptors are tetrameric proteins that mediate fast synaptic transmission and are grouped into three classes, according to their differential affinity for the agonists N-methyl-D-aspartate (NMDA), kainate (KA), and α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid (AMPA) (Dingledine et al., 1999). The metabotropic receptors belong to the G-protein-coupled receptor superfamily, and can be divided in the group I (mGluR1 and mGluR5), group II (mGluR2 and mGluR3) and group III (mGluR4 and mGluR6-8) according to their agonist pharmacology, primary sequence and G-protein effector coupling (Fagni et al., 2004). The increased probability of epilepsy in the autistic population (Tuchman & Rapin, 2002) suggests that abnormally enhanced glutamatergic signalling may contribute to some of the autistic characteristics. Indeed, there is a slight positive correlation between plasma levels of glutamate and the severity of autism that supports this hypothesis (Shinohe et al., 2006). Moreover, the cerebellum of autistic patients had increased expression of mRNAs encoding the excitatory amino acid transporter 1 (EAAT 1) and the AMPA 1 receptor. Increments in those proteins would elevate the extracellular concentration of glutamate and enhance the postsynaptic activity of glutamatergic synapses (Purcell et al., 2001). Genome studies have found associations between autism and
3. Evidence of GABAergic dysfunction in autism
GABA is the most abundant and versatile neurotransmitter in the Central Nervous System (CNS). GABA is excitatory in the immature brain and inhibitory in the mature one (Ben-Ari, 2008). At early stages of neural development GABA has a paracrine action on immature neurons. It modulates neuronal migration, stimulating developing networks and exerting a wide range of trophic actions that lead to the correct establishment of neural circuits (Ben-Ari, 2008). In the adult brain, GABA participates in the generation of synchronous rhythms of cortical assemblies. This allows local time-precise communication among neurons and coherent communication with other cerebral centres, thus creating behavioural relevant processes (Somogyi et al., 2005). GABA actions on the membrane potential are mediated mostly by ionotropic receptors. Ionotropic GABA receptors in the CNS are pentameric channels made up by the combination of α (1-6), β (2-3), γ (1-3) and δ subunits in a γ−β−α−β−α arrangement, with δ substituting γ in some extrasynaptic receptors (Olsen & Sieghart, 2008). They are permeable to chloride ions and whether they depolarize or hyperpolarize the cell membrane potential depends on the developmentally regulated expression of the Na+/K+/Cl−-cotransporter, NKCC1, and the K+/Cl−-cotransporter, KCC2. NKCC1 is highly expressed in immature neurons and transports chloride into the neuron producing a high intracellular concentration of chloride, thus making GABA depolarizing. In mature neurons the increased expression of KCC2 and reduction of NKCC1 lowers the concentration of intracellular chloride and makes GABA hyperpolarizing (Mathews, 2007).
Due to the concerted and intertwined activity of GABAergic and glutamatergic neurotransmission, even small net deviances of GABAergic activity could affect the excitation-inhibition balance in the autistic brain. Such an imbalance would reduce the ratio signal to noise of the sensory and procedural information in mild cases (Casanova et al., 2006) and, in the extreme ones, it could lead to epilepsy. Actually, the incidence of epilepsy in one third of people with autism (Tuchman & Rapin, 2002) and the presence of paroxystic activity in the electroencephalogram (EEG) of approximately 68% of autistic people (Kim et al., 2006) is consistent with this hypothesis. These gradual alterations of the EEG traces suggest that aberrant electrical activity in the autistic brain is expressed as a continuum and it may be present even in cases of autism with normal EEG recordings. Although abnormal electrical activity could have scores of causes, genetic association studies have implicated genes coding for the subunits γ1, α2, α4, β1 and β3 of GABAA receptors as likely contributors for autism (Blatt et al., 2001; Hussman, 2001; Cook et al., 1998; Ma et al., 2005; Vincent et al., 2006; Kakinuma & Sato, 2008). Indeed, evidence from disorders that share overlapping autistic characteristics like Prader-Willi/Angelman syndrome (AS) and Rett syndrome, supports the idea that GABA receptors are convergent nodes in autistic phenotypes of different genetic origins. Angelman syndrome is an imprinted disorder caused by a maternal deficiency of chromosome 15q11-q13 (Magenis et al., 1987; Lalande, 1996) that includes autistic characteristics and developmental delay, seizures and stereotyped behaviours (Samaco et al., 2005). Because the 15q11-q13 region contains genes for the β3, α5 and γ3 subunits of GABAA receptors, the chromosomal deficiencies in Angelman syndrome may produce alterations in the expression of these GABA subunits. Knockout mice with deletions in the GABAA α5 and γ3 subunits did not show a drastic phenotype; but a deletion in the gene for the GABAA β3 subunit produced a neonatal mortality of 90-95% and the survivors displayed a phenotype resembling severe forms of AS (Sinkkonen et al., 2003). These alterations were associated to a decreased number of GABA receptors in areas like the hippocampus (Sinkkonen et al., 2003) which is commonly affected in autism. Rett syndrome is a pervasive disorder classified within the autism spectrum category. Patients diagnosed with Rett syndrome, in addition to the triad of autistic characteristics, also show cognitive deficits, apraxia, ataxia, seizures and respiratory abnormalities (Chahrour & Zoghbi, 2007). Rett syndrome is the result of mutations of GABA is produced by the enzymes Gad65 and Gad67 which are respectively coded by the genes Gad1 and Gad2. In MECP2 deficient mice the expression of Gad1 and Gad2 and the immunoreactivity to GABA in interneurons were importantly reduced. These reductions were associated with smaller miniature postsynaptic events (mIPSC) without a change in their frequency (Chao et al., 2010).
GABA is produced by the enzymes Gad65 and Gad67 which are respectively coded by the genes Gad1 and Gad2. In MECP2 deficient mice the expression of Gad1 and Gad2 and the immunoreactivity to GABA in interneurons were importantly reduced. These reductions were associated with smaller miniature postsynaptic events (mIPSC) without a change in their frequency (Chao et al., 2010).
The wide number of factors with effects on GABA receptors and the different degrees of changes in GABAergic signalling may explain, at least in part, the high heterogeneity found in the clinical phenotypes within the autism spectrum. However whether changes in animal models apply to humans and what other qualitative changes are present in the human brain is a matter of current research. Blatt et al. (2001) reported a reduced binding of benzodiazepines and muscimol in the hippocampus of autistic brains, suggesting a decrease in the number of GABAA receptors. Posterior studies using different concentrations of [H3]flunitrazepan showed that the decrement of benzodiazepine binding was due to reductions of binding sites with no changes in the binding affinity of the receptors (Guptill et al., 2007). Western blot analyses of four GABAA subunits (α1-3 and β3) showed reductions of all subunits in parietal cortex, of α1 in frontal cortex and of α1 and β3 in cerebellum of post-mortem autistic brains (Fatemi et al., 2009). And recent studies showed reductions in the binding sites to muscimol and benzodiazepines in cingulate cortex and fusiform gyrus of autistic brains (Oblak et al., 2009; Oblak et al., 2011). These authors also found a reduction in the affinity of the binding sites to muscimol, suggesting pharmacological changes in the receptors due to changes in the properties of the same receptors or switching of GABAA subunits. Indeed, an increment in the expression of α5 has been reported in the autistic brain (Purcell et al., 2001), and a potential remodelling of GABA subunits may explain the several reports of paradoxical benzodiazepine-based sedatives on severe autistic individuals with mental retardation (Marrosu et al., 1987; Sandman & Barron; 1992; Aman & Langworthy, 2000).
Undoubtedly binding experiments are, and will continue, providing important information about the status of GABA receptors in the autistic brain, particularly about density and tissue localization. However, their resolution is limited and the absence of functional information on the receptors is a great drawback compared with electrophysiological studies, where even the ion current through an activated single channel can be detected. Therefore, important differences between GABAA receptors might be overlooked. Another concern is that the benzodiazepine binding site is not present in all GABAA receptor isoforms and alterations in GABA receptors containing δ, α4 and α6 subunits, which do not bind classical benzodiazepines but are highly expressed in cerebellum (Sieghart, 2006), have not yet been appropriately addressed. It is worth noting that because of the duplication of the 4p12 chromosome (Ma et al., 2005), a gain of function of the α4 subunits is expected, but not yet explored. Also important is the fact that normal binding does not necessarily mean normal activation; therefore, even though the agonists and antagonists can bind to the receptor, determining whether the receptors are functional or not requires a multidisciplinary approach, using binding, biochemical and electrophysiological methodologies. Because GABA and glutamate are main targets of pharmacological intervention, a detailed analyses of their kinetic and biophysical properties will help to evaluate new drugs and therapeutic treatments.
4. Pharmacology of human receptors beyond binding studies
We have developed two methods that allow in-depth studies of neurotransmitter receptors and ion channels of the human brain. The first, now widely used, involves the heterologous expression of human receptors in
4.1. Heterologous expression of human receptors
The evidence that single mutations of glutamate- and calcium channels can lead to autism indicates that abnormal network excitability and intracellular signalling are critical factors in the generation of autistic phenotypes; and highlights the importance of understanding the electrophysiological properties of the receptors and channels, even in cases where no mutations have been reported. The expression of mRNA or cloned human receptors in
4.1.1. Methodological insight into heterologous expression in
In order to study ion channels and other membrane proteins via the expression approach, the gene must be available in a plasmid. Nuclear injections of plasmidic DNAs can be done, provided a relative high purity sample is used. We have found that plasmidic DNA isolated with spin columns is good enough to express receptors. The human cytomegalovirus (CMV) promoter is our promoter of choice. We routinely inject 14 nL at a concentration of about 200 ng/mL, aiming at the animal pole and injecting deep to increase the likelihood of reaching the nucleus. Notwithstanding, the rate of successful nuclear injection is not very high, ending always with some oocytes that do not express the protein and with increased oocyte “mortality”. We also use the T7 promoter, but the level of expression is considerably lower (see also Geib et al., 2001). Another factor to consider is our observation that the receptor induced membrane currents achieved in the oocytes successfully injected with DNA are lower than those achieved after injecting cRNA. Therefore, our preferred choice for expression of channels is the cytoplasmic injection of synthetic cRNA. If the gene of interest is placed after any RNA Polymerase promoter such as T7, T3, SP6, synthetic cRNA can be produced in large quantities and with a desirable level of purity. One can use the template DNA and add the Polymerase, rNTPs and a reaction of a couple of hours will yield microgram quantities of RNA (Krieg et al., 1984). The ready to use kits are very convenient and even incorporate a 5’ cap analog and have the option of adding an enzyme that will incorporate a Poly(A)+ tail at the 3’ end to resemble more closely a natural RNA.
Of all the kits we have tested, the mMessage mMachine kit from Ambion is the one that has given us the best results. That kit uses ARCA (Anti Reverse Cap Analog) 5’ cap analog (Stepinski et al., 2001), which prevents incorporation in the reverse orientation and maximizes translation efficiency. In such a way, we obtain RNAs that have strong expressional potency after injecting 50 nL (at ca. 1 mg/mL) into the equator of the oocyte (Limon et al., 2007; Reyes-Ruiz et al., 2010; Limon et al., 2010). Besides its use for the expression of recombinant proteins,
4.2. Microtransplantation of human receptors
Cell membranes, mostly in the form of small vesicles, are adjusted to a protein concentration of 1-2 mg/mL and injected into an oocyte. Within a few hours the membranes, carrying their original neurotransmitter receptors and ion channels, begin to fuse with the oocyte plasma membrane. Voltage-clamp recording is then used to study the functional characteristics of the transplanted receptors. Oocytes with transplanted receptors can be studied up to several days post-injection (Miledi et al., 2004; Limon et al., 2008).
We have assessed whether autistic brains with long post-mortem intervals still contain functional neurotransmitter receptors and voltage-operated ion channels that could be microtransplanted into
It will be interesting to microtransplant receptors from other areas of the brain. The hippocampus is important in memory and learning processes; and together with the amygdala and prefrontal cortex participates in the negative feedback of the hypothalamic-pituitary-adrenal (HPA) axis which in turn controls the body’s response to stress (Morris, 2007). The hippocampus and amygdala have clear neuroanatomical alterations in autistic brains (Bauman & Kemper, 2005) and there is a hypothesis that prenatal stress may affect the HPA axis during development and increase the risk of developing autism (O’Donnell et al., 2009). Accordingly we evaluated if neurotransmitter receptors from hippocampi with long post-mortem intervals can be transplanted to
4.2.1. Microtransplantation of native receptors from the hippocampus of autistic brains
For these experiments we used the anterior hippocampus from two autistic brains and two matching control brains (Table 1). Cell membranes were prepared as previously reported (Limon et al., 2008) and then injected into the equator of
|Case||Dx||Age (years)||Gender||PMI (h)|
An important advantage of the microtransplantation method is that the biophysical studies are done directly on native receptors that were once in the human brain and are still embedded in their original lipids and with their own cohort of associated proteins. Another advantage is the rapid and high yield of functional and pharmacological information, obtained from minimal amounts of protein. For comparative purposes, consider for example that to prepare a 2-dimensional gel ranges of 2-2000 µg of protein are needed (Adams & Gallagher, 2005); for a Western Blot about 40 µg is recommended. Even for mass spectrometry-based proteomics few µg of total protein is used. In contrast, with the microtransplantation method 50 ng of protein are injected into a single oocyte and a full dose response plus several pharmacological experiments can be done with such a small amount of protein.
Strong evidence indicates that autism is a developmental synaptic disorder that affects the processing of behavioural relevant information. Although the causes of autism are still not known, GABAergic and glutamatergic synapses appear to be convergent nodes of genetic, epigenetic, and probably environmental factors causing the autistic phenotype. Even small alterations in GABAergic or glutamatergic signalling produce autistic characteristics in animal models, and it is highly probable that a similar phenomenon is present in the human brain. GABA and glutamate receptors are also important targets of pharmacological interventions and a detailed knowledge of their function in the human brain will improve the use and design of molecules with therapeutic activity. The microtransplantation of the original receptors from postmortem brains coupled to the expression of receptors by post-mortem mRNAs will help to determine in great detail the type, number, and functional properties of autistic neurotransmitter receptors and channels. These procedures will help to decipher the functional impact of genetic and epigenetic factors in autism. Furthermore, the microtransplantation method will help to determine the mode of action of the medicines presently used to treat autism, and help to develop new medicines and evaluate their pharmacological activity.
We are extremely grateful to all the persons that donated their brains for scientific research. Human brain tissue samples were generously provided by the Autism Tissue Program via the Harvard Brain Tissue Resource Center (Belmont, MA). This work was supported by the King Abdul Aziz City for Science and Technology, (Riyadh, SA; Grant KACST-46749).
- GABA is produced by the enzymes Gad65 and Gad67 which are respectively coded by the genes Gad1 and Gad2. In MECP2 deficient mice the expression of Gad1 and Gad2 and the immunoreactivity to GABA in interneurons were importantly reduced. These reductions were associated with smaller miniature postsynaptic events (mIPSC) without a change in their frequency (Chao et al., 2010).