The response of the hippocampus to the unilateral removal of the entorhinal cortex provides an illustration of the general principles of reactive synaptogenesis

One of the best illustrations of the general principles and mechanisms of reactive synaptogenesis is represented by the response of the hippocampus to the unilateral removal of the entorhinal cortex input [4]. Fibers of the entorhinal cortex represent a major input to the dentate gyrus and terminate in the outer two-thirds of the dentate gyrus molecular layer. This circuitry is of particular interest because of its critical role in higher cognitive functions, such as learning and memory, and its vulnerability to degeneration in Alzheimer's disease and, to a lesser degree, during the course of aging (see Chaps. 30, 46 and 50). Thus, this system is an excellent model not only for minor cell loss but also for changes triggered by significant injury to the brain. These studies also provide insight into the mechanisms underlying more subtle events, such as learning, that occur over much longer time periods.

Following unilateral ablation of the entorhinal cortex, over 80% of the synapses degenerate and are subsequently replaced by sprouting and new synapse formation from remaining fibers, including cholinergic projections from the medial septum, the commissural-associational pathway of CA4 neurons and projections from the contralateral entorhinal cortex. The projections from the ipsilateral and contralateral CA4 neurons normally terminate in the inner one-third of the dentate gyrus molecular layer. Following entorhinal lesions, these fibers sprout into the denervated zone, eventually occupying the inner half of the molecular layer (Fig. 29-2).

Figure 29-2

Changes in the dentate gyrus molecular layer following a unilateral entorhinal lesion. A: Normal distribution of entorhinal inputs to the outer two-thirds of the molecular layer and commissural/associational (Comm/Assoc) inputs and kainic acid (KA) receptors (more...)

Fibers from the contralateral entorhinal cortex, normally sparse in the molecular layer, sprout extensively in the denervated zone after a unilateral lesion. The behavior of single fibers can be traced using anterograde labeling methods with the lectin tracer Phaseolus vulgaris leucoagglutinin (PHAL), a marker that is transported via axoplasmic flow throughout the axon and its branches. Single neurons in the entorhinal cortex are injected with the tracer, and after a few days the tissues are processed histochemically to trace the fibers. In control animals, these fibers exhibit sparse branching and form primarily en passant, bouton-type synapses. In lesioned animals, the fiber density increases and single axons exhibit more branch points (Fig. 29-3). Some axons exhibit high-density, localized sprouting, and in some, but not all animals, tangle-like structures are observed in the denervated outer molecular layer. The number of synapses increases approximately 100-fold.

Figure 29-3

After a unilateral entorhinal lesion, the entorhinal—dentate pathway originating from the opposite entorhinal cortex shows extensive axon sprouting. A: Control; B, C: two examples of the behavior of single fibers after lesion. D: Quantitative (more...)

Because contralateral entorhinal input is essentially homologous to the lost entorhinal input, the sprouted fibers likely participate in the recovery of function following unilateral entorhinal ablation. Consistent with this idea, studies show that a unilateral entorhinal lesion causes temporary deficits in spontaneous or reinforced alternation tasks and that the rate of behavioral recovery corresponds to the rate of sprouting of the homologous contralateral fibers [5–7]. This response of the brain may maintain function in the same manner as the early recovery after partial transection of the sciatic nerve in the PNS. Remaining fibers are known to sprout and restore function prior to regeneration.

To some extent, remodeling in the normal, healthy brain may involve processes that parallel those involved in remodeling that occurs after injury [3,8]. It appears that the brain has some intrinsic capabilities for plasticity and that these are enhanced when the homeostasis of the brain is challenged. It is interesting, in this respect, that, following injury, regions not primarily associated with the lesion also exhibit synaptic density changes and subsequent recovery of control levels over a long period of time. The synaptic changes occur despite the absence of degenerating terminals within these zones. Thus, pronounced transneuronal changes may occur after major trauma to the CNS, suggesting that reactive synaptogenesis may adjust the functional integrity of complex circuitry in areas with and without a primary lesion [9].

Sprouting in the adult brain results in an increase in the inputs already present without new pathway formation

The capacity for extensive remodeling and growth must be restrained when such remodeling is not required. Because the anatomy of the hippocampal formation is very well defined, examination of synaptic plasticity in the entorhinal/dentate circuit allows the establishment of a set of principles regulating axon sprouting. All new synapses appear to originate from fiber systems normally present in the circuit. The spatial arrangements of inputs, termed the lamination pattern, reorganizes so that CA4 synapses terminate further distally on the granule cell dendritic tree and septal cholinergic synapses reduce their domain to the reduced entorhinal zone. Relamination of inputs in the dentate gyrus may be due to (i) an invasion of CA4 fibers into regions previously occupied by entorhinal inputs or (ii) an outward growth of the dendritic tree, which carries CA4 synapses and fibers more distal from the cell body. If the dendrites grow, then the first-order branch points would be located farther from the soma. Indeed, the first-order stem appears to lengthen by 35 μm, which is exactly the increase in the width of the CA4 termination zone. Apparently, the outward growth of the dendritic tree causes the migration of CA4 fibers toward the outer molecular layer and a new lamination pattern [10,11].

Thus, although no new circuits are formed, fiber connections assume new spatial relationships on their normal target cells. Clearly, restraints are placed on the extent of axon growth and synapse formation in the mature brain. This is not the case in the developing brain, in which entire new pathways may develop (Chap. 27). The restraints operational in the mature brain may involve myelin-based or other types of growth inhibitors in neuronal pathways, such as has been defined in the spinal cord (see below).

Collateral sprouting and reactive synaptogenesis occur in discrete stages

During development, a specific sequence of events occurs that results in the formation of specific connections and neuronal circuits. When an injury occurs in the mature brain, the growth process must be executed in the context of a damaged system. The old system must be cleared and coordinated with the initiation of growth and the formation of new synapses.

The interaction of a growing axon and its environment involves several key factors and stages. Over the past several years, studies in cell culture have shown that neuronal growth requires a minimum of two extrinsic conditions: a supply of select neurotrophic factors (Chap. 19) and a proper substrate or set of cell-adhesion molecules (Chap. 7). These are also critical conditions for axon sprouting and regeneration in vivo. Thus, as illustrated in Figure 29-4, the key cells and molecular events required in vivo are (i) glial involvement in clearing degeneration, (ii) neurite outgrowth-promoting factors, (iii) composition of the extracellular matrix and expression of cell-adhesion molecules, (iv) events specifying targeting and (v) synapse formation and the expression of molecular systems regulating neurotransmitter release and proper postsynaptic receptors.

Figure 29-4

Stages in the mechanisms of axon sprouting and reactive synaptogenesis. The CNS has to face the complex problem of clearing the damage as the circuitry is being rebuilt and remodeled. Microglia and astrocytes clear the products of degeneration. Neurite (more...)

Glial cells set the pace for reactive synaptogenesis

Microglia are reactive and increase in number within the denervated zone by the first day, peak by postlesion day 3 and return to normal conditions by about 8 days. Astrocyte reactivity appears to follow the microglial response, peaking at approximately postlesion day 4 and remaining reactive for approximately 2 weeks. Microglia, in particular, engage in a robust phagocytotic reaction and engulf degenerating terminals, myelin and debris. Astrocytes also appear to be involved primarily in the removal of degenerating terminals and the stripping of them from postsynaptic densities.

The rate of synapse formation appears to be inversely related to the rate at which degenerating terminals are cleared. Thus, the clearance reaction appears to be critical to subsequent synapse formation. To determine if the glial response and the removal of degenerating material paces the re-innervation process, it is necessary to delay degeneration and to determine if there is a corresponding delay in sprouting. The “Ola” mouse exhibits delayed axonal degeneration, and in this strain, the glial reaction and sprouting response are similarly delayed [12].

Cellular lipids also appear to follow a process of degradation and recycling involving the induction of apolipoprotein E (ApoE) within astrocytes in the outer molecular layer. Lipids released from degenerating axons and dendrites are salvaged by astrocytes, released in the form of an ApoE—cholesterol lipoprotein complex and accumulated in neurons via the low-density lipoprotein receptor pathway [13]. These lipids are thus reused in the processes of dendritic remodeling and synapse formation.

Cytokines and neurotrophic factors are induced after lesions of the entorhinal cortex

In response to entorhinal cell loss, increased syntheses of nerve growth factor (NGF), brain-derived neurotrophic factor (BDNF), ciliary neurotrophic factor (CNTF), interleukin 1B (IL-1B), basic fibroblast growth factor (FGF-2) and transforming growth factor (TGF-β1) have been described. The list will probably continue to grow as additional factors are recognized and analyzed in this model. Some of these factors regulate glial responses and others are directed at neurons and their growth responses (see Chap. 19).

It is generally believed that most of these changes are involved in the evolution of the growth process. Detailed studies, however, have suggested that another mechanism may complicate the analysis. After entorhinal lesions, there is a surge of glutamate release, which can and does induce a transient neurotrophin increase; for example, NGF increases briefly after the lesion. If this increase is blocked by an NMDA antagonist, the growth response is not suppressed, indicating that the initial induction is not an essential part of the growth response. NGF, however, does appear to be necessary for cholinergic sprouting.

Over a period of several days, NGF protein concentrations appear to increase after entorhinal lesions, reaching a maximum 8 days postlesion and returning to control values by 30 days. To determine if this increase is essential for cholinergic sprouting, an antibody to NGF was given for 8 days prior to an entorhinal lesion. The antiserum blocked the collateral sprouting of cholinergic fibers, suggesting that a supply of NGF and its action on these cells are important parts of the growth response [14,15].

Cell adhesion molecules influence neuronal growth

There are several cell adhesion molecules that can support and stimulate neuronal growth. Isoforms of the neural cell adhesion molecule (NCAM) have been shown to mediate axonal outgrowth after peripheral nerve injury and during reinnervation through interactions with other cell-surface molecules (see Chap. 7). After entorhinal cortex lesions, the embryonic form of NCAM is expressed on dendrites and axons within the denervated outer molecular layer by 24 hr. Astrocytes express NCAM at a later time, between 2 and 4 days, which may support and guide the sprouting reaction [16].

The adhesion molecule L1 can also enhance neuronal growth and may participate in the reinnervation process. L1 has been reported to mediate neuritic outgrowth, axon fasciculation and regeneration. Two to twelve days after entorhinal lesions, L1 immunoreactivity decreases but then increases again as synapses reform. The time course and reappearance of L1 parallel the maturation phase of synapse formation, which suggests that L1 is involved in the maturation and stabilization of synapses after they have made contact [17,18]. Thus, after entorhinal lesions, neurotrophic factors and cell adhesion molecules appear to participate in the process of reactive synaptogenesis.

Cytoskeletal protein concentrations increase after lesions

Sprouting in the CNS may involve a reactivation of mechanisms that operate primarily during development. There are many examples of developmentally regulated genes, including those encoding cytoskeletal proteins normally expressed during development. For example, the fetal form of α-tubulin, referred to as Tα1 in the rat, is expressed at high levels in the fetal brain during development of neuronal processes, and its expression is normally greatly reduced in the mature brain. Other examples are tau protein and microtubule-associated protein 2 (MAP2), two of the major microtubule-associated proteins that form the cytoskeleton of neurons in the vertebrate nervous system (see Chap. 8). MAP2 and tau proteins promote the assembly and integrity of microtubules in axons and dendrites, suggesting a major role in the determination of neuronal morphology.

In mature neurons, altered gene expression and increased synthesis of both microtubules and associated proteins are observed during periods of neurite outgrowth. For example, following axonal injury to the mature nervous system, Tα1 is rapidly reinduced and maintained at high concentrations during axon outgrowth [19]. Other mRNAs induced include tau, α-tubulin, β-tubulin and β-actin. These results support the idea that the neuronal cytoskeleton is dynamically modified in response to lesions, which may involve the re-expression of mechanisms normally activated during development. Furthermore, following entorhinal cortex lesions in adult rats, MAP2 protein, which is expressed primarily in the inner molecular layer, is redistributed from the inner layer to the outer layer in the processes of sprouting neurons [20].

Synaptic proteins are produced in response to lesions

Synaptic proteins and growth-associated proteins increase in response to entorhinal lesions. The neuronal growth-associated phosphoprotein GAP-43 is expressed on growth cone membranes during periods of axonal growth. GAP-43 is decreased in the outer molecular layer for the first few days after entorhinal cortex lesions and then returns to normal by 2 to 3 weeks with an increase of approximately 300% in the number of labeled terminals. Many of these also exhibit an increase in synaptophysin, indicating that new terminals are being formed [21]. Because the new synapses are functional, it can be assumed that the essential molecules are regenerated. The exact mechanistic sequence for reinnervation and assembly of components is as yet undefined.

Molecular cascades involving cytokines appear to regulate the growth response

Increasing evidence indicates that the biological role of growth factors exceeds that of simply promoting cell growth, and indeed, they have major roles as physiological regulators. Even though growth factors generally promote growth, depending on interactions with other molecules, they can also inhibit growth. Recent studies in vitro and in vivo demonstrate that there are physiological interactions between various growth factors, consistent with the idea that multiple growth factors coordinate their actions in molecular cascades.

Following brain injury, several cellular and molecular events occur near the injury site that determine the physiological response of the remaining cells and appear to be associated with the action of growth factors. After injury, one of the earliest cellular responses appears to involve microglia. This is followed by increases in astrocytic reactivity. This cellular sequence has led to various investigations examining possible cascades of growth factors involving these cells. Indeed, glial cells such as microglia and astrocytes are common sources of growth factors and have a determining role in the injury response. One such cascade is illustrated in Figure 29-5. Microglia release cytokines, such as IL-1, that induce reactive astrocytes. In this process, IL-1 promotes the release of other growth factors, such as IL-6 and NGF, by astrocytes. Reactive astrocytes also secrete CNTF and FGF-2, which may have an autocrine role for astrocytes [22]. In the normal brain, CNTF is low, but it and its receptor (CNTFRα) are induced in astrocytes within 3 days. The receptor is also induced on neurons between 7 and 10 days, which correlates with the remodeling process, suggesting that CNTF plays a role in the overall response. Microglia also release TGFβ1, which regulates the action of FGFs on cells and protects and strengthens the extracellular matrix. Extracellular matrix components such as proteoglycans, in turn, potentiate the action of growth factors such as FGF-2 (see above). As growth factors are released into the extracellular fluid, they become available to other cells, such as neurons, which also produce growth factors, such as NGF. Not only do these growth factors work in concert to regulate growth factor production, but some also increase the responsiveness of cells to other growth factors. For example, studies have shown that cells in culture cannot respond to NGF until they are primed with FGF-2 [23]. In this case, FGF-2 induces receptors for NGF in cells, preparing them for the action of NGF. Thus, it is highly likely that growth factors are organized in molecular cascades to act as major regulators of cell development and plasticity.

Figure 29-5

Simplified mechanism by which growth factors may regulate neuronal plasticity following injury to the CNS. Primed by the original insult, microglia, astroglia and neurons interact with each other by releasing growth factors to the extracellular space. (more...)

The hippocampus in Alzheimer's disease shows plasticity similar to that observed in the rodent brain after entorhinal lesions

One of the goals of modern neurochemistry is to employ the findings from basic research and animal models to predict and evaluate mechanisms in human disease. Alzheimer's disease causes extensive neuronal degeneration in select brain areas, including the entorhinal cortex, the origin of the major excitatory projection to the hippocampus. The course of degeneration is such that the neurons of the entorhinal cortex in layers II and III projecting into the hippocampus are among the first affected. However, while degeneration is a prominent feature of Alzheimer's disease, reactive growth in both neurons and glial cells is exhibited in this disease as well (Chap. 46).

In the dentate gyrus of the normal brain, there is a light cholinergic input, whereas in the Alzheimer brain, the cholinergic input is increased in the denervated zone, as predicted from the animal models discussed earlier [24]. The fetal forms of several cytoskeletal proteins are also re-expressed in the Alzheimer brain. As predicted from entorhinal cortex lesion models, the message for Tα1 is present in the Alzheimer brain at high levels [19]. These results suggest that as neurons are lost in the early period of the disease, the remaining cells sprout and form new synapses to compensate for lost connections and to maintain neuronal circuitry.

In Alzheimer's disease, plasticity may become pathological and result in plaque biogenesis

One of the neuropathological hallmarks of Alzheimer's disease is the presence of senile plaques in select brain areas. Plaques are extracellular deposits that consist primarily of a polypeptide product called amyloid. Plaques also have certain neuritic and cellular involvement, such as with astrocytes and microglia. In the dentate gyrus, plaques appear to form along the areas of interface between degeneration and neuronal sprouting and are a locus of concentrated sprouted axons, such as has been observed in some animals after entorhinal lesions (see Fig. 29-3). It has been suggested that there is an abortive turning of the sprouting reaction into plaque formation [25]. In fact, Ramon y Cajal in 1928 [26] had suggested that sprouting fibers were attracted to plaques by some neurotrophic factor, which was a remarkable insight considering that at the time there was little knowledge regarding the role of neurotrophic factors in the brain.

Plaques located along the sprouting zone in the dentate gyrus accumulate a variety of neurotrophic factors and cell adhesion molecules, which probably attract neuronal growth responses [27]. Once in the area, the processes become dystrophic, degenerate and then regenerate again because of the favorable growth-stimulating environment. Further, the process of degeneration causes an inflammatory response, which continues to drive the evolution of the cycle. In this proposed mechanism, β-amyloid binds neurotrophic factors and cell adhesion molecules and stimulates inflammatory responses and degeneration. Because β-amyloid persists in the tissue, it may convert a reversible acute phase response that would usually increase and decrease, as described above for an acute injury, into a chronic response [28].

It is ironic that the initial and/or subsequent mechanisms that promote growth and slow degeneration may in certain circumstances contribute to the disorganization of the environment and the evolution of disease. This mechanism may abort the compensation process of axon sprouting. That is, as entorhinal neurons degenerate, other entorhinal neurons form connections to replace those lost and to delay functional decline; however, some of the growth response is misdirected into plaques and drives decline.