Grossman et al.,
A Brain Adaptation View of Plasticity:
Is Synaptic Plasticity An Overly Limited Concept?
Grossman?, Churchill, Kleim, Bates, Greenough?
1Beckman Institute, 2Neuroscience Program, 3Medical Scholars Program, 4Departments of Psychology, 5Psychiatry, and 6Cell and Structural Biology, University of Illinois at Urbana-Champaign, Illinois, USA. 7Canadian Centre for Behavioural Neuroscience, University of Lethbridge, Alberta, Canada.
Address correspondence and reprint requests to:
William T. Greenough, Ph.D.
405 N. Mathews
University of Illinois at Urbana-Champaign
Urbana, IL 61801
There is a long tradition, traceable to the early speculations of Ramon y Cajal, of focusing on the neuron as the only plastic cell type of any importance within the brain, and synaptic plasticity as the only important process for modulating interactions between neurons. While neuronal and synaptic plasticity are without question important aspects of brain function, it has become increasingly clear that other cellular elements of brain are malleable and that their plasticity can contribute to brain function. For example, oligodendrocytes, astrocytes, vasculature, and perhaps other neuropil components also exhibit plasticity in the developing and mature brain, and each of these cellular elements appears to be differentially influenced by distinct components of an experience. It is also becoming clear that various forms of brain plasticity likely have different functional purposes. Exposure to a complex environment, for example, causes synaptogenesis in animals genetically rendered incapable of potentiation of some post-synaptic responses, suggesting that these forms of plasticity serve different purposes in the brain. In short, while research has focused largely on naturally- and artificially-induced changes in synaptic connectivity, the brains of animals (and presumably people) in real-world situations are in a dynamic state in which synaptic adjustment may in some cases be a relatively small part of the mix. Here we review some of the data regarding multiple forms of plasticity, and briefly discuss how such changes might differentially affect functional brain organization.
Since the early speculations of Tanzi and Ramon y Cajal, the synapse has been the principal proposed site of plasticity underlying learning and memory in the brain. Tanzi (1893) initially emphasized the possibility of strength changes in pre-existing connections while Ramon y Cajal (1893) stressed both the formation and loss of connections. The ability to appropriately investigate these possibilities was limited by the availability of suitable tools, but, by the early 1970s, electrophysiological and anatomical evidence of the nervous system’s ability to alter its functional connectivity in accord with its experience was becoming well established.
Physiologically, activity-dependent modification of post-synaptic responses were described in invertebrates (Castellucci, et al., 1978), and in the mammalian visual system (Hubel, et al., 1977), and long-term potentiation was proposed as a model of vertebrate neural learning (Bliss and Lomo, 1973). Anatomically, there was evidence that dendrites and synapses could both form and change in size in response to altered patterns of synaptic activity, not only in the developing nervous system, but also in adult animals. Functional reinnervation of neurons by surviving axons, for example, was shown to occur spontaneously subsequent to denervation (Lynch et al., 1973; Raisman and Field, 1973). Likewise, morphological changes have been described in well-characterized models of invertebrate learning (e.g., Bailey and Chen, 1988), enabling researchers to better understand the relationship between structure and function that underlies plastic neural change. Here we summarize progress in understanding the various types of brain plasticity thought to be associated with learning and memory, focusing heavily on our own work, since those relatively early beginnings.
The complex environment housing paradigm, pioneered by Hebb (1949) and his students (e.g., Hymovitch, 1952; [Forgays, 1952 #207] was first used as a tool for exploring brain plasticity by Bennett, Diamond, Krech, Rosenzweig, and colleagues (Bennett, et al., 1964), who reported some of the earliest evidence for morphological brain plasticity in response to experience. Subsequently, dendritic field dimensions (Volkmar and Greenough, 1972) and synaptic size (West and Greenough, 1972) were reported to increase in the visual cortex of rats exposed to a complex environment (EC) from weaning through adolescence. A later report by Turner and Greenough (1985) specifically demonstrated that there were more synapses per neuron in upper layers of the visual cortex in rats reared in a complex environment than in control animals reared in standard laboratory housing conditions. Taken together, these observations demonstrate that behavioral experiences could be used as a tool to study very specific, measurable aspects of neuronal plasticity.
It should be noted that there is a reason that we use the term “complex environment” instead of the oft-used term “enriched environment.” We contend that laboratory environments simply are not enriched relative to the norm of wild or feral animals. We would argue that no one has published studies of the brains of rats that have been exposed to levels of environmental complexity and challenges beyond the level provided by the natural environment, and studies of wild animals have for years confirmed that feral animal brains are larger than those of domestically reared animals (old german and other literature). Nevertheless, studying different degrees of environmental complexity can provide information about brain responses that are likely to generalize to higher levels of stimulation, as suggested also by studies of the differential effects of environment on human development (e.g., Hart and Risley, 1995).
In addition to the effects of complex environment on the visual cortex, multiple brain areas suspected to be involved in the processing and/or responding to environmental stimuli have also been shown to exhibit plasticity. For example, morphological changes have been reported in the auditory cortex (Greenough, et al., 1973) and physiological modifications of forepaw representation in primary somatosensory cortex (Coq and Xerri, 1998) in response to exposure to a complex environment. Beyond cortical sites, areas such as hippocampus [Moser, 1997 #80;Rampon, 2000 #87] basal ganglia (Comery et al., 1995; 1996), and cerebellar cortex (Greenough, et al., 1986) are also responsive to experience. These observations should not be interpreted to suggest that exposure to a complex environment induces plasticity in an ubiquitous manner throughout the brain. Indeed, certain forms of activity-induced plasticity have been shown not to occur in specific brain regions (e.g. Kleim, et al., 1998; see also below). Instead, these data suggest that different brain systems are involved in the processing of specific components of an animal’s experience. Following a description of the multiple types of brain plasticity, experimental methods for dissociating the specific components of experience will be discussed.
1 Neuronal and Synaptic Plasticity
Among the most exciting recent developments in the field of neuronal plasticity are the data suggesting that the brain responds to experience by adding new neurons. Using a thymidine analog (BrdU) that incorporates into replicating DNA, it has been demonstrated that neurogenesis occurs in the hippocampal formation in response to complex environment housing (Kempermann, et al., 1998; Kempermann, et al., 1998; Nilsson, et al., 1999). More specifically, using a learning paradigm in which the underlying neural pathways necessary to perform the task have been very well-characterized, it was reported that neurogenesis is dramatically increased in the hippocampus when this structure is critically involved in learning the task, yet when the contingency does not demand involvement of this structure, the rate of neurogenesis is unaffected (Gould, et al., 1999). Although these increases in neuron number are small relative to the contingent of neurons already present in the brain, theories of brain plasticity that have largely focused on changes in the number and strength of synapses in neural networks must now consider the profound effects that integration of new neurons could have on both the composition and function of neural networks. While neurogenesis represents an exciting areas, it is a principal focus of other chapters in this volume. Thus, we highlight here specific aspects of plasticity of existing neurons as they relate to elements of non-neuronal plasticity and to the functional implications of the existence of multiple forms of brain plasticity.
2 Plasticity of synapse number
On an anatomical level, the malleability of neuronal systems and individual neurons can be quantified using a number of parameters. Dendritic arborization is an indirect measure of available postsynaptic space (suggestive of synapse number), and can be quantified according to the methods described by Sholl ([Sholl, 1956 #227]). A more specific measure of synapse number is the ratio of synapses per neuron [Cragg, 1975 #226], which should be estimated using stereologically unbiased techniques (West, 1999). Exposing weanling rats to a complex environment leads to increases in dendritic branching and in synapses per neuron in the visual cortex (Greenough, et al., 1973; Turner and Greenough, 1985). The magnitude of these two experience effects was in the same range, a 20-25% increase, suggesting that the synaptogenesis associated with visual experience may reflect an increase in dendritic length, upon which new synapses form, more than it reflects an increase in the density of synapses per unit length of dendrite. Experience-induced increases in dendritic spine density, however, have been described (Globus, et al., 1973; Comery, et al., 1995)[Rampon, 2000 #87].
Although the effects of experience on measures of synapse number are most profound in young animals, synaptogenesis appears to occur across the age spectrum and is not restricted to particular critical or sensitive periods of development. Exposure to a complex environment has been shown to affect dendritic field dimensions in the visual cortex of young adult, middle-aged and even elderly rats (e.g., Greenough, et al., 1979; Juraska et al., 1980; Green et al., 1983; Black et al., 1986; Connor et al., 1981). Synapse number per neuron is also increased in rats placed in complex environments as adults (Briones, et al., In preparation). Behavioral experience appears to induce anatomical change throughout the lifespan in the cerebellum as well as the cerebral cortex .
2 Plasticity of synapse morphology
In addition to changes in neuronal morphology, the morphology of individual synapses is affected by experience, as well. A number of these morphological effects are typically viewed as modifications of preexisting synapses, although they could reflect intermediate stages of synapse remodeling or the formation and/or loss of synapses with different properties. The size of both pre- and post-synaptic components, the shape of dendritic spines, and the , or the size of components such as the postsynaptic density (PSD) length, vesicle aggregates and the pre and postsynaptic processes themselves (Sirevaag and Greenough, 1985; West and Greenough, 1972).
In response to physiological activity (
While changes in spine shape appear to be sensitive to experience, Harris and Stevens (Harris, 1988 #110) have reported that the diameter of the spine neck is stable. These observations have been interpreted to suggest that the spine neck serves to isolate intracellular events to activated synapses without altering the transfer of synaptic charge to the dendrite. THE WESA STUFF SEEMS OUT OF PLACE, CERTAINLY OUT OF FLOW As yet another demonstration of plasticity at the synaptic complex, Wesa et al. (1982) reported an increase in concavity of the presynaptic element in visual cortex associated with rearing rats in a complex environment. Taken together, these findings support the contention that neuronal plasticity can be expressed in a number of manners…..
It has been theorized (by harris KM) that as the synapse grows, and the PSD gets larger, it finally develops a peforation. (or some such transition)
2 PSD Perforations and Multiple Synaptic Boutons
Although it is widely appreciated that the changes in synapse size and number have functional relevance, other types of synaptic plasticity exist whose functional relevance is less well understood. The curious discontinuities in the postsynaptic density characteristic of “perforated” synapses have been regularly associated with plastic synaptic changes since they were first reported to increase during development and with complexity of experience in rat visual cortex (Greenough et al., 1978; see also Jones and Calverley, 1991). The frequency of perforated synapses also increases after motor skill training (Jones et al., 1999; Kleim et al., in preparation). Similarly, Geinisman and others reported that some types of these synapses increased in number following kindling or induction of LTP (Geinisman et al., 1990; 1991).
Another form of synaptic morphology with a proposed relevance to plasticity is the multiple synaptic bouton (MSB). Multiple post-synaptic spine and dendritic shaft synapes occur on vesicle-filled presynaptic processes in a number of brain regions including cerebellar cortex, cerebral cortex and hippocampal formation. While there is debate about the processes whereby terminal boutons or en passant varicosities become contacted by more than one postsynaptic process, there are consistent reports that the frequency of these MSBs is increased in behavioral and electrical plasticity paradigms.
Early evidence that their occurrence could be influenced by experience was the Friedlander et al. (1991) report of a greater number of postsynaptic contacts on presynaptic terminals associated with the open eye in monocularly-deprived kittens. Subsequently Jones et al. (1997) reported a higher number of MSBs in visual cortex of EC (vs. SC or IC) rats. Jones et al. (1999) have also described increases in these multiple synapses in the intact cortex in the course of compensatory changes following unilateral cortical lesions. Geinisman et al., (2001) reported an elevated number of MSBs in hippocampal subfield CA1 following associative eyeblink conditioning using the trace paradigm, which requires an intact hippocampus in order to be learned. Similarly, Federmeier et al (submitted) found a dramatic increase in cerebellar parallel fiber varicosity MSBs with motor skill training. Thus the formation of new postsynaptic contacts on previously-innervated varicosities appears to be a common form of plastic synaptic change.
Several theories have proposed that perforated synapses and multiple synaptic boutons, in addition to branched dendritic spines, might represent intermediate stages of a single synapse “splitting” into separate dendritic spines (e.g., Carlin and Siekevitz, 1983). In addition to the work described above, studies on synapse morphology following the induction of synaptic activity (Geinisman et al., 1989; Comery et al., 1996; Toni et al., 1999) have suggested the likely involvement of these intermediates in the establishment and/or refinement of synaptic contacts. Alternatively, evidence suggests that the phenomenon of spine splitting is not supported by the existence of perforations, branched dendritic spines and multiple synaptic boutons (Sorra, et al., 1998; Fiala, et al., 2002). Transmission electron microscopy precludes investigation of morphological dynamics, but recently developed techniques hold promise for real-time visualization of dendritic spine motility and will surely shed more light on this subject (e.g., Engert and Bonhoeffer, 1999; Maletic-Savatic, et al., 1999). It is clear that the final words have yet to be written regarding modification of synaptic morphology in response to altered patterns of activation.
1 Persistence of neuronal changes
For the most part, neuronal changes induced by differential experience (e.g., Volkmar and Greenough, 1972; Turner and Greenough, 1985) appear to persist after experience has been discontinued. In rats that had been exposed to 30 days of complex environment (EC) and then returned to standard housing conditions (IC), the dendritic arborization (Camel et al., 1986) and number of synapses per neuron (Figure X; Briones, et al., in preparation) were not different from animals that were exposed to EC for 60 days. Both groups, however, differed significantly from animals that were individually caged for 60 days. The increases in number of synapses per Purkinje cell in response to motor skill training (Black, et al., 1990) appear to persist for a minimum of four weeks in the absence of continued training (Kleim, et al., 1997).
1 Non-neuronal Plasticity
The complex environment paradigm has been used extensively to examine plasticity of neurons and synapses. This paradigm has also been used to study a range of plasticity of non-neural elements, yet these data have largely failed to become incorporated into theories of brain plasticity. Early in the history of this field, Diamond et al. (1966) reported that exposure to a complex environment induced changes in the number of glial cells in cerebral cortex. We have subsequently examined plasticity in glia and other non-neuronal elements of the cerebral cortex in response to complex environment exposure and find that, in general, the changes in these elements parallel those observed in neuronal dendrites and synapses. These observations suggest that most, if not all, types of brain tissue exhibit plasticity, at least in regions such as the cerebral and cerebellar cortices where this phenomenon has been systematically examined.
. ALTHOUGH GLIAL-GENESIS COULD BE A VERY INTERESTING ASPECT OF PLASTICITY AS WELL, YET WE DON’T HAVE TIME TO DISCUSS SUCH.
2 Plasticity in astrocytes
Although early reports on experience-induced plasticity of astrocytes (Diamond et al. 1966; Szeligo and Leblond, 1977) were suggestive of this phenomenon, the techniques for quantifying morphology that were available at that time limited the conclusions that could be drawn from the data. Subsequent reports used unbiased stereological techniques to quantify morphological changes and have described increases in both the surface density of glial fibrillary acidic protein-immunoreactive processes and the percent of tissue that is taken up by astrocyte cell nuclei (volume fraction) in the visual cortex in response to EC exposure (Sirevaag and Greenough, 1987; Jones et al., 1996; Sirevaag and Greenough, 1991). Additional complex experience appears to cause hypertrophy of astrocytic processes and cell number, although the effects of environment on astrocytes depend both on the duration of EC exposure and on the cortical layer examined (reviewed in Jones and Greenough, in press).
Astrocytic changes in response to EC seem, for the most part, to follow a similar time course, paralleling measures of neural plasticity (e.g. Wallace et al., 1992; Jones et al., 1996). Moreover, at least in cerebellar cortex, the volume fraction of astrocytes parallels plasticity of synapse number on an individual animal basis [Sirevaag, 1988 #40; Anderson, 1994 #11]. Ultrastructurally, it appears that proliferation of astrocytic processes is closely associated with experience-induced synapse addition. Jones and Greenough (1996) found that in rats exposed to a complex environment synapses are more completely ensheathed by astrocytic processes than synapses in control animals, possibly reflecting a role for astrocytes in optimizing the synaptic microenvironment in response to increased neural activity.
There are numerous reasons to expect astrocytic plasticity to have functional consequences. Glia influence synaptic function in ways that range from modulating synaptic efficacy (Araque et al., 1998; Smit et al., 2001) to dissociating presynaptic from postsynaptic processes during synaptic remodeling (Hatton, 1997; Meshul et al., 1987; Salm, 2000). Astrocytes can conduct excitation via propagated Ca2+ waves (Araque et al., 1999; Dani et al., 1992) which can directly influence neuronal activity (Rouach et al., 2000). These data suggest that astrocytes may constitute a separate, relatively local (and slower) system for activating brain regions.
Glial cells are involved in the re-uptake and metabolism of GABA and glutamate (Schousboe et al., 1992; Bezzi et al., 1999), suggesting that they could play a role in synaptic modulation. Glial cells also have receptors for many neurotransmitters including norepinephrine (Shao and Sutin, 1992) and glutamate (Muller et al., 1993; Shelton and McCarthy, 1999), two neurotransmitter systems that have been implicated in the formation and modulation of memory processes. Considering the apparent close interaction between astrocytes and events typically thought to be “synaptic,” it is likely that plasticity of astrocytes as well as neurons is critical to the process of learning and memory.