To what extent can external factors, including nutritional influences, modify the way the brain is wired and behaves? This is an issue of intensive research that has advanced greatly in recent years due to the development and application of new technologies. Multiphoton laser imaging now allows researchers to look into the living brain at the microscopic level and investigate synapses and glial cells. The mobility of neurotransmitter receptors can be studied by another new technique, i.e., single-molecule tracking. These two methodologies both permit analyses of processes and structure in real time, which is an enormous advantage. Another area of investigation considers the extracellular matrix molecules and how these molecules, interposed between the cells of the central nervous system, affect function and restrain structural plasticity. The importance of the extracellular matrix molecules has been markedly underrated, but this is about to be rectified by new research.
The number of synapses in the brain may be in the range of 1014. Synapses provide contact points between neurons and are the computational units that help humans think and act. Malleability operates at the level of these synapses. It is necessary to recognize that the traditional approaches involving microscopic analyses have deceptively led to the perception of synapses as rigid physical connections. In the electron microscope, a typical synapse is seen to consist of a presynaptic element, a synaptic cleft, and a spine. The spine receives information from the presynaptic element, and the synaptic cleft appears as a tiny part of the extracellular space separating the presynaptic element from the spine. With the immunogold technique,1 it is now possible to identify the neurotransmitter receptors, which are impacted by transmitting signals from the presynaptic element. Most of the synapses discussed here use glutamate as the transmitter. Glutamate is by far the most prevalent transmitter in the central nervous system.2
The classical nonvital microscopic appearance of a glutamate synapse shows complex structures, with glutamate transporters evident in both the neuronal and the glial membranes. An important type of glutamate receptor is the α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor. The AMPA receptors are concentrated in the postsynaptic density, i.e., the membrane domain that receives the signaling at a synapse. In principle, a glutamate synapse is designed to transmit impulses in a very precise manner, with the receptors located very close to the transmitter release sites. Transmitter is released from a vesicle in the presynaptic membrane and diffuses across the synapse to the receptors in the postsynaptic element. How dynamic is this arrangement and how fixed are the numbers of synapses for this particular transmitter? The discussion of these issues will draw on recent data that include results obtained by the Nordic Centre of Excellence in Molecular Medicine, a new center devoted to in vivo imaging by multiphoton microscopy, and two European Union Framework Programme projects known by their acronyms, KAR-TRAP (Kainate and AMPA Receptor Trafficking at Brain Synapses) and GRIPANNT (Glutamate Receptor Interacting Proteins as Novel Neuroprotective Targets). These projects focus on the mobility of glutamate receptors and the plasticity of glutamate synapses.
In some of the first attempts to follow the movement of individual receptors, a bead was coupled to the receptor in the membrane by way of a receptor-specific antibody.3 Using video microscopy, the movement of the bead could be monitored, reflecting the mobility of a single receptor or, perhaps, of a couple of receptors. More recently, the technique has evolved immensely, and the state-of-the art technique combines physics with biology. Specifically, using in vitro systems, receptor movement is analyzed by means of quantum dots, which emit pulses of light that allow the mobility of individual receptors to be traced. These studies have shown that the receptors, previously thought to be stable, are in fact moving rapidly. A receptor can move into a synapse and then along the cell's surface and from one synapse to another. Thus, AMPA glutamate receptors can alternate within seconds between a state of rapid diffusion and stationary behavior, with the latter being characteristic of the receptors when they are inside the synapse. So, receptors are not faithful to any given synapse in the brain but show a pronounced promiscuity. Whether this extraordinary mobility applies to the in vivo state is unknown, but the assumption is that an element of receptor transfer between synapses exists in the living brain as well.
Choquet,4 one of the members of the European Union consortium GRIPANNT, has produced a dynamic picture of AMPA receptors in which some of them are located on the spines and are partly faithful in the sense that they stay there for a while. Other receptors, however, are moving rapidly from one synapse to another. AMPA glutamate receptors can be seen to alternate within seconds between rapid diffusive and stationary behavior. A critic might ask whether the labeled receptors actually enter the synapses. Dahan et al.5 have addressed this question by combining the single-molecule tracking technique with electron microscopy and the use of a silver intensification technique. Using this combination of techniques, glycine receptors6 can be observed moving from one synapse to another, and the labeled receptors clearly do reside within the synapse.
Why should we care about the mobility and number of receptors and whether they relate to learning and memory? Long-term potentiation of synapses is a term used to denote synaptic strengthening and is associated with an increased number of AMPA receptors in the synapse. The GRIPPANT consortium was long preoccupied with the mechanisms that allowed receptors to move into and out of synapses and thereby regulate the number of receptors in synapses. It seems now there are two different mechanisms at play in receptor removal from the synapse. One is tangential diffusion of receptors along the plasma membrane, and the other is an endocytotic/exocytotic process.7 The latter implies the receptors are brought into the cell and then recycled to the membrane through a vesicular carrier. Many of the molecules involved in this recycling process have now been identified. In electron micrographs, spines with small pits covered with a protein called clathrin can be found.8 These clathrin-coated pits are thought to represent the sites where the receptors are brought into the cell by endocytosis.
Receptor mobility also seems to be critically involved in a process called paired-pulse depression. Paired-pulse depression refers to a mechanism in which a synapse that is stimulated twice shows a reduced response the second time compared with the first. Heine et al.9 have shown the degree of paired-pulse depression to be greater if the receptors in the synapse are immobilized, i.e., prevented from moving (also see Mikasova et al.10). Normally, there is a diffusion of receptors in and out of the synapse so that new, fresh receptors are recruited, which allows the signal to continue acting. If this is prevented from happening, there is an attenuation of the signal when the synapse is stimulated the next time.10 This can be shown in a very elaborate way by observing the transition between closed receptors, open receptors, and desensitized receptors.
Since the receptors are mobile,10,11 the properties of the membrane itself can be of importance when it comes to signal transfer, and this can affect the function of the central nervous system at large. This raises the question as to what governs the fluidity of the membrane and the ease by which the receptors move. Fatty acids in the diet may affect membrane fluidity and, hence, the mobility of the receptors (discussed in Chytrova et al.12).
Turnover of brain synapses
The next issue to consider is the structural stability of the synapses themselves. How do synapses change in terms of number and structure? Multiphoton imaging technology reveals dendrites with protruding spines, representing sites of synapses, and makes it possible to monitor these synapses and dendrites over periods of days and even weeks by repeatedly revisiting exactly the same spot in the mouse brain. It is not necessary to make a hole in the cranium to perform the analysis, as the laser beam easily penetrates the thinned cranium. Multiphoton imaging also allows analysis of the astrocyte cell bodies and the thin processes that connect them to the perivascular endfeet that abut the small vessels in the central nervous system. Thus, it is now possible at the microscopic level to analyze structure, physiology, and also the pathophysiology in the living brain.
In vivo optical imaging technology is based on the use of light in the infrared spectrum, a possibility already predicted by the physicist Maria Göppert-Mayer in 1931 in her doctoral thesis, long before she won the Nobel Prize. She predicted optical electronic transitions in molecules, i.e., the excitation of molecules, can be induced even in response to low-energy photons, i.e., by light in the infrared spectrum, if there are enough photons present. If the photon density is high enough, a given molecule can absorb more than one photon at a time, so that the added energy induces excitation. However, at that time, there were no means available to produce the high density of photons required, and the realization of her proposition had to wait until the laser was invented in 1958.
The laboratory of K Svoboda was one of the first to take full advantage of multiphoton analysis in investigating synaptic plasticity and showed that synapses are not necessarily stable over time. Some synapses turn over quite rapidly, with some sites showing particularly high synapse turnover from day to day, as shown by Trachtenberg et al.13 Sensory input also altered synaptic turnover, and electron microscopy showed these new structures are indeed new synapses with a mature nerve terminal in contact with a newly formed spine.
Many laboratories worldwide, including our own, now use in vivo optical imaging techniques to assess how sensory experience and nutritional factors affect synapse turnover. Again, using the Trachtenberg et al.13 study as an example, the removal of vibrissae has been shown to profoundly alter synapse turnover in the barrel cortex (the cortical representation area of the vibrissae). Current studies are now attempting to demonstrate, by use of the same approach, how different learning paradigms affect synapse turnover and stability.
The multiphoton imaging technique is also suitable for analyzing the molecular mechanisms of disease, including the generation of plaques in Alzheimer's disease. Recently, Meyer-Luehmann et al.14 used the multiphoton technique to study the establishment of amyloid plaques in a mouse model of Alzheimer's disease. Contrary to the widely held view that plaques develop slowly, at least some of the plaques seem to build up extremely rapidly, from one day to another in some cases. These plaques were termed “popcorn” plaques because they pop up so quickly. These data illustrate how modern imaging techniques have changed our perception of the timescale of events in the central nervous system. Thus, several physiological and pathological features appear to develop much more rapidly than previously thought.
Nuntagij et al.15 have recently shown amyloid plaques can be reconstructed at the electron microscopic level by using a series of more than 200 consecutive ultrathin sections. The analysis shows these plaques develop in close contiguity to specialized membrane domains. When considered with the data by Meyer-Luehmann et al.,16,17 it is proposed there are factors in these specialized membrane domains that precipitate the formation of amyloid fibrils in the extracellular space.
Multiphoton imaging can also be used to investigate the molecular mechanisms of disease in real time. In studies of the pathophysiology of brain edema, a condition in which water accumulates in the brain, e.g., after a stroke or head injury, Nase et al.18 induced brain edema in the mouse by intraperitoneal injection of distilled water. They used multiphoton imaging to identify cells in the cerebellar cortex at different intervals after water injection and found astrocytes, particularly those close to brain capillaries, were the first cells to swell when edema was starting. Astrocytes are equipped with a high density of AQP4 water channels at their site of contact with the brain microcapillaries, corresponding to the perivascular endfeet. These AQP4 water channels, shown by immunogold techniques, probably account for the initial swelling of the astrocytes.19,20
Importance of a malleable extracellular matrix
The malleability of synapses and the mobility of membrane molecules appear to be restrained by extracellular matrix molecules. For example, the above-mentioned AQP4 molecule is tethered by dystrophin and dystroglycan21,22 to extracellular matrix molecules, in this case laminin and agrin.22 However, there are enzymes that can break down molecules in the extracellular matrix and limit the constraints imposed by the extracellular matrix on cellular and molecular plasticity.23 One important class of enzymes is the metalloproteinases, which include a metalloproteinase called MMP9. These proteinases can act on molecules in the extracellular space. With Wilczynski et al.,24 our group has shown, using immunogold cytochemistry, that the MMP9 metalloproteinase is found in the extracellular space in conjunction with the endfeet of astrocytes and around synapses, and some of the MMP9 molecules seem to be in close apposition to synaptic spines.25 Thus, there are molecules close to the synapses with the ability, at least in theory, to modify the extracellular matrix in order to make it easier for the spines to assume new shapes or to shuttle molecules into or out of the membrane.
The ability of MMP9 to facilitate structural plasticity was tested in a model of epilepsy, which is normally associated with a significant pruning or loss of dendrites.24 When the metalloproteinase MMP9 is knocked out in mice, the marked pruning of spines is not observed, implying the complement of synapses is largely retained. Therefore, by removing the action of the proteinase acting on the extracellular matrix, the plasticity normally seen in this particular pathological condition is reduced. Knocking out the MMP9 also lessened the severity of the epileptic condition, with seizures being less intense than those observed in the normal wild-type animals. Mice with an overexpression of MMP9 show an increased rate of seizures. It was also recently reported that metalloproteinases, which affect the composition of the extracellular matrix, alter the movement of receptors, as predicted.