THE brain, the centre of the nervous system in all vertebrates, is perhaps the most complex of biological structures. It has multiple components that together regulate, on the one hand, basic life processes such as breathing, maintenance of body temperature, and motor functions and, on the other, higher functions such as cognition, learning and memory, creative abilities and emotions.
The seat of these higher functions is not confined to a single part of the brain. Many parts of the brain are responsible for them, but together they can be broadly identified as a single subsystem called the limbic system, which includes the hippocampus and the amygdala (Figure 1). The amygdala, which is not much bigger than an almond, is located deep within the brain below the hippocampus. While the hippocampus is associated with complex cognitive processing, learning and memory, the amygdala is involved in complex emotional responses, such as fear, anxiety and aggression, and the long-term memories linked to them. Although major strides have been made in the understanding of how the brain functions, much of the detailed mechanisms involved in a given function are not very well understood and thus constitute a major discipline in ongoing biological research.
There are certain cognitive and emotional disorders that are caused by inherited (genetic) defects. A class of such disorders is known as fragile X syndrome (FXS), which includes a range of physical and cognitive disabilities as well as emotional and behavioural features many of which have an overlap with autism or have autism-like characteristics. FXS is said to be the most common cause of inherited mental retardation. This is an X chromosome-linked disorder and hence affects males more severely than females because the second normal X chromosome present in females lessens the effect of the chromosome with the disorder.
People with FXS may display heightened anxiety, fear, mood instability, aggression and hyperactive behaviour such as fidgeting, excessive physical movements or impulsive actions. They have mental retardation and learning (especially mathematical) disabilities. They may also have attention deficit disorder and problems of social interaction, including shyness, limited eye contact, an inability to encode faces, memory problems and cluttered or nervous speech. Non-neurological features include a long face, large ears, hyperextensible joints, flat feet, soft skin, low muscle tone and enlarged testes in men.
FXS is caused by a specific mutation of the gene called fragile X mental retardation 1 (FMR1), which is located in the long arm of the X chromosome. The FMR1 carries instructions to make a protein called fragile X mental retardation protein (FMRP) but, in the presence of this mutation, makes very little or no protein. In this mutation of FMR1, a DNA (deoxyribonucleic acid) segment, which has a repeat of the nucleotide triplet CGG (C and G are two of the four nucleotide bases A, T, C and G that make up DNA), is expanded a great deal more than the normal. Normally, this CGG repeat occurs five to 50 times. In people with FXS, it is repeated 200 to 4,000 times. This abnormally expanded CGG repeat segment results in methylation of that portion, which effectively silences the expression of FMR1. The methylation is believed to cause a constriction of the X chromosome at that point, which appears fragile under the microscope, and hence the name for the syndrome. While the mutation itself is found in one out of every 2,000 males and one out of every 250 females, the incidence of FXS is about one in every 3,000 males and one in every 4,000-6,000 females.
FMRP plays a critical role in the development of the connections between nerve cells (neurons), or the synapses, which are important for transmitting nerve signals. Neurons are the cells that pass signals to individual target cells and synapses are the junctions between the nerves that enable them to do so (Figure 2). At the synapse, where the axon the long and slender projection that conducts electrical impulses away from the body of the nerve cell, or soma from the signal-emitting neuron contacts the dendrite of the signal-receiving neuron, neurotransmitters are released when the pre-synaptic side is activated either spontaneously or by external electrical stimulation. Glutamate, for example, is an important neurotransmitter in the brain. The released neurotransmitter drifts across the synaptic gap and reaches the neuron on the post-synaptic side, where it binds to receptors on the cell membrane. Glutamate can bind with three classes of receptors: metabotropic glutamate receptors (mGluRs), AMPA receptors and NMDA receptors. This binding enables the signal to be transmitted to the post-synaptic side. The loss or shortage of FMRP, therefore, leads to a disruption in the normal communication between neurons, resulting in the spectrum of disorders that characterise FXS.
At present there is no effective treatment for FXS or autism. Recent research on FXS has led to a useful framework to understand its cellular and molecular basis. This is the so-called mGluR theory, which was put forward by M.F. Bear and others in 2004. According to them, the lack of FMRP synthesis and the regulation of signal transmission by it result in exaggerated signalling in certain communication pathways that involve the mGluRs, and this is responsible for many of the symptoms of FXS and probably autism too. In particular, a subclass of these mGluRs, called mGluR5, is implicated in the abnormally high signalling involved in FXS. This suggested that mGluR5 blockers could be potential therapeutic agents for FXS. One such potential drug is the chemical compound called MPEP.
The underpinnings of the mGluR theory came from studies that found neurotransmission defects in the hippocampus. Recently, however, a team of neuroscientists, led by Sumantra Chattarji of the National Centre for Biological Sciences (NCBS) of the Tata Institute of Fundamental Research (TIFR) in Bangalore, identified novel synaptic defects in the amygdala. The study was carried out using mice that were genetically engineered to model FXS by knocking out the FMR1 gene in them FMR1-KO mice. The findings are also of therapeutic significance because investigations showed that a brief pharmacological intervention of the knockout mice with MPEP could reverse some of the synaptic defects of the amygdala. Chattarji carried out this research in association with his graduate student Aparna Suvrathan and scientists from New York University (NYU). The work was published recently in the American journal Proceedings of the National Academy of Sciences (PNAS).
Why did Chattarji suspect that the amygdala may have a significant role in FXS? According to him, from his interactions with clinicians he had learnt that FXS patients often displayed severe emotional problems, such as abnormally high fear and high social anxiety, which overwhelmed whatever cognitive skills they possessed. Also, Chattarji's group at the NCBS has been involved in detailed studies of how the underlying cellular and synaptic mechanisms in the amygdala get affected as a result of emotionally significant experiences that leave a long-term imprint or memory, leading to affective disorders, such as severe stress-induced chronic anxiety, in an otherwise normal nervous system. So, it was natural for his group to extend the investigations to FXS, where there are genetically inherited emotional dysfunctions.
In his work on affective disorders, Chattarji had found that signalling and cellular changes in the amygdala were not quite the same as what had been observed in the hippocampus and the cortex. For instance, while in the hippocampus a shrinkage of nerve cells and a loss in the total neural volume was seen, in the amygdala nerve cells, dendrites, and so on, actually grew bigger, leading to chronic and stronger anxiety and fear. In the FXS-induced synaptic changes, too, he and his student found that the changes were in contrast to what had been seen in the hippocampus.
Earlier studies in the hippocampus showed that the glutamate-mediated transmission had deficits only in the post-synaptic side. But the present experiment in the amygdala of knockout mice revealed something very interesting and strikingly different from the observations in the hippocampus there were defects in both sides of the synaptic gap. While the post-synaptic defects were similar, Chattarji and Aparna Suvrathan also found hitherto unknown pre-synaptic deficits relating to glutamate release. Such pre-synaptic defects have not been seen in the hippocampus. Since glutamate is the essential trigger for the signal to be transmitted, this implied that, in the amygdala, the neurons were not communicating properly. Eric Klann and Charles Hoeffer of NYU identified the molecular correlates of these defects, which led to a firm understanding of the exact nature of the defects.
All past work has focussed primarily on post-synaptic defects. The entire field has been obsessed with the other side. So, our finding of a pre-synaptic dysfunction is a breakthrough and opens up new directions for therapeutic targets, said Chattarji. But this finding obviously raised doubts in their minds about the utility of the drugs that have been identified on the basis of the mGluR theory, which is based on the findings mainly related to the post-synaptic defects in the hippocampus. In fact, the drug MPEP, which had been found to reverse the post-synaptic defects in the hippocampus, did not produce the same effects in the post-synaptic defects in the amygdala. Surprisingly, however, MPEP seemed to act positively on the pre-synaptic defect that seemed unique to the amygdala.
One of the characteristic features of signal transmission between two neurons is what is known as synaptic plasticity, the ability of the neural connection, or the synapse, to change in strength. There are several underlying mechanisms that cooperate to achieve synaptic plasticity, including changes in the quantity of neurotransmitters released into a synapse, changes in the response of receiving cells to the neurotransmitters because of post-synaptic receptor activity or receptor density. Synaptic plasticity is supposed to be an important neurochemical basis for learning and memory. Long-term depression (LTD) and long-term potentiation (LTP) are two of the several phenomena underlying synaptic plasticity. While the former is a weakening of signal transmission between two neurons, the latter is a strengthening, or enhancement, of the synapses. In some sense, LTP is the opposite of LTD. These are called long term because these changes last for hours or longer.
In the studies on the hippocampus with FMR1-KO mice, LTD was found to be enhanced. Given the basic fact that FMRP is silenced in FXS, the mGluR theory was proposed as a natural way of explaining this increased synaptic weakening in LTD. The basic premise of the mGluR theory is as follows: Activation of mGluR, in particular mGluR5, normally stimulates synthesis of proteins involved in the stabilisation of LTD and, in addition, FMRP. The FMRP, in turn, inhibits further synthesis of the proteins and thus puts a brake on LTD. If FMRP is absent, mGluR-mediated protein synthesis goes on unregulated and there is a continual weakening of LTD.
Now, it is known that mGluR5 also mediates LTP, or synaptic strengthening, in the amygdala. LTP in the lateral amygdala, for example, is responsible for classical fear conditioning. So, if we extended this naive mGluR model to LTP in the amygdala, and if FMRP is lacking, excessive and unchecked mGluR signalling should actually enhance LTP. When Chattarji and co. began investigating, nothing was known about the impact of FXS on synaptic function in the amygdala and, in particular, whether synaptic abnormalities in the amygdala were consistent with the mGluR theory.
To monitor both pre-synaptic and post-synaptic aspects of transmission, Chattarji and Aparna Suvrathan took electrophysiological recordings on the signal-receiving neurons in the slices of lateral amygdala of both wild-type and mutant, or FMR1-KO, mice and compared them. Since input signals to the transmitting neuron came from the thalamus and it was difficult to reach the thalamus with measurement probes, they chose to monitor the effects from the receiving end, Chattarji explained. These single-cell physiological recordings provided measures of the following: pre-synaptic release of glutamate, post-synaptic changes (including biochemical quantification of AMPA receptors) and synaptic plasticity (specifically LTP). What they found was that LTP in the amygdala, instead of being enhanced, was impaired; in fact, no LTP could be detected. So, as Chattarji pointed out, not only was there contrasting manifestation of mGluR-dependent synaptic plasticity between the amygdala and the hippocampus (LTP vs LTD), but the defects in these forms of synaptic plasticity also diverged.
This prompted them to ask the question: Does this contradict the mGluR theory? While there is an apparent contradiction in the naive picture, some work done in 2007 showed that protein synthesis triggered by excessive mGluR signalling also resulted in the internalisation of the other class of receptors, the AMPA receptors, on the cell surface. That is, there was a reduction in the surface expression of the AMPA receptors. In the hippocampus, this receptor suppression would, therefore, reinforce the enhancement of mGluR-mediated LTD, or synaptic weakening. In the experiment with the amygdala of the knockout mice, too, Chattarji's group found that there was indeed a reduced surface expression of AMPA receptors (Figure 3). We provide the first direct evidence for a reduction in surface AMPA receptors in adult amygdalar slices [of knockout mice] using both electrophysiological and molecular assays, said Chattarji.
From this perspective, therefore, it becomes easy to understand the impairment of LTP in the amygdala. In this picture, the dominant effect of FXS on the synaptic process is the reduction of receptors on the post-synaptic surface, which implies reduced signal transmission. The picture also now offers a consistent explanation for both enhanced LTD in the hippocampus and impaired LTP in the amygdala. One could, of course, imagine that, in the case of amygdalar LTP, excessive mGluR-signalling-induced protein synthesis and repressed expression of AMPA receptors would be competing with each other. However, according to Chattarji, protein synthesis actually gets triggered a little after LTP sets in. But if LTP itself is totally impaired, protein synthesis does not begin at all. He, however, added that this was an area where researchers had only scratched the surface and that there were a lot of open questions.
The clear indication of impaired synaptic plasticity does not, however, provide any information on the basic characteristics of synaptic transmission in the amygdala of the knockout mice. In a bid to gain an insight into that, the research group measured some parameters that also would reflect pre-synaptic processes, such as glutamate release. This included the frequency and amplitude of what is called the spontaneous miniature excitatory post-synaptic currents (mEPSCs). While the frequency gives a measure of the pre-synaptic glutamate release probability, the amplitude is indicative of post-synaptic deficits. In particular, this would reflect the reduction in surface AMPA receptors, which would be an electrophysiological corroboration of what was observed through biochemical assays.
As expected, there was a decrease in mEPSC amplitude. But there was also a decrease in mEPSC frequency, a deficit not seen earlier in any of the studies on the hippocampus or the cortex. This suggested a decrease in the pre-synaptic glutamate release probability. This possibility was corroborated by measuring another parameter called paired-pulse ratios (PPRs) using external electrical stimuli at different inter-stimulus intervals as thalamic inputs to the amygdala. [This] warrants a more rigorous analysis of the decrease in probability of release especially because no such pre-synaptic effect has been reported in the hippocampus, the PNAS paper reports.
From a therapeutic point of view, the most important prediction of the mGluR theory is that suppressing mGluR activity (with mGluR5 blockers) should reverse the synaptic defect in the amygdala of the knockout mice as it does in the hippocampus. Strikingly, the drug MPEP failed to reverse both the impaired LTP and the post-synaptic deficit in the surface AMPA receptors at the amygdalar synapses. One possible explanation for this, according to the paper, is the restoration of AMPA receptors to the surface could also be mediated by other non-mGluR plasticity mechanisms. Thus, our results, gathered from the amygdala of knockout mice, contradict some key aspects of the mGluR theory as envisioned in the hippocampal framework [with regard to therapeutic intervention], Chattarji said.
But when Chattarji and associates investigated whether MPEP could reverse the pre-synaptic deficit in glutamate release probability, they, surprisingly, found that MPEP was able to return mEPSC frequency in knockout neurons to levels not significantly different from wild-type neurons. Strikingly, said Chattarji, MPEP treatment for the same duration [as for post-synaptic defects] was able to reverse pre-synaptic deficits in the amygdala. So in this regard, the mGluR theory does hold up in the amygdala but through a novel pre-synaptic route. In the hippocampus, on the other hand, MPEP treatment has been shown to rescue the post-synaptic deficit in surface AMPA receptors, but no pre-synaptic changes in transmitter release have been reported in knockout mice (emphasis added).
As mentioned earlier, emotional dysfunctions are a major problem in FXS and related autism spectrum disorders. Yet, virtually nothing was known about how the amygdala, which plays a critical role in emotional responses, was affected before Chattarji and co. entered the scene. Therapeutically, without any insight into this, it would be difficult to address both the cognitive and emotional defects effectively. Our findings, said Chattarji, for the first time, provide a rigorous platform to bridge this gap in knowledge by showing how some of the cellular and synaptic deficits underlying emotional symptoms may be unique, and give hope that treatment may be possible even later in life.
Although the findings of the Chattarji-led research are consistent with certain aspects of the mGluR theory, the paper has emphasised the need to modify the existing framework to better explain the synaptic dysfunctions in the FXS amygdala. Notwithstanding the important finding that deficits in the synaptic transmission and plasticity in the hippocampus and the amygdala of knockout mice diverge (Figure 4), the work has demonstrated that the latter are still amenable to pharmacological interventions that target mGluR5, but in a manner different from what the original mGluR theory suggested.
One could, of course, argue that since the amygdalar post-synaptic defects in the knockout neurons are not reversed by MPEP-like mGluR5 blockers, these cannot yet be considered effective therapeutic interventions. The key issue, Chattarji said in an e-mail exchange, is the following. Even with a one-hour in vitro treatment, we were able to reverse the pre-synaptic deficit, that too in a fully adult brain where FXS had lots of time to do damage. So, if we treat with drugs in vivo for days, weeks or months, there is every reason to be optimistic about reversing the post-synaptic defects, [and] maybe also the plasticity deficits. Now that we have such a detailed map of what exactly goes wrong in terms of information encoding, we are all set to use better drugs in vivo and see what is corrected and to what extent.
