If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Cyclic adenosine monophosphate (cAMP) and cyclic guanosine monophosphate (cGMP) are both cyclic nucleotides. This nomenclature originates from the fact that their phosphate group is attached to two sugar hydroxyl groups, forming a ring or “cyclic” structure. Cyclic nucleotides function as second messengers, which serve to relay and strongly amplify incoming signals at receptors on the cell surface. Cyclic nucleotides are thus important elements in signal transduction cascades (1). cAMP and cGMP are hydrophilic and therefore transmit signals within the cytosol, activating mainly protein kinases and ion channels. The study of cyclic nucleotide signaling in the last decades has revealed a stunning complexity (2). The effects of cAMP and cGMP are namely not only dependent on changes in their concentration, since temporal and spatial components are also of major importance. Signals can vary in timeframe, from milliseconds to hours, and often show a very strict compartmentalization in a specific part of the cell. This enables a wide variety of outcomes of cAMP and cGMP signaling, enabling the cell to react appropriately and nuanced on stimulation. These characteristics expose the requirement for a meticulous regulation of cyclic nucleotides. cAMP is synthesized from adenosine 5′-triphosphate (ATP) by membrane-bound adenylyl cyclase (AC), which is mainly regulated in neurons by G-proteins and additionally stimulated by Ca2+ and calmodulin. Synthesis of cGMP is regulated by guanylyl cyclase (GC), which converts guanosine 5′-triphosphate (GTP) into cGMP. The key activator of the most common GC in neurons, that is, soluble GC, is the gaseous signaling molecule nitric oxide (NO). However, the most important regulation of cyclic nucleotides is seemingly not achieved by its synthesis, but by the breakdown of cAMP and cGMP in their inactive forms, 5′AMP and 5′GMP, respectively. The enzymes responsible for this process are phosphodiesterases (PDEs).
The PDE superfamily consists of 11 subtypes (PDE1–PDE11) based largely on their sequence homology and are coded by 21 identified genes. In addition, each gene product may have multiple splice variants (e.g., PDE4D1–PDE4D9), which add up to a total of more than 100 different PDE proteins (3). Subtypes are differentiated based on several characteristics, such as localization, subcellular distribution, regulatory mechanisms, and enzymatic and kinetic properties. Most of these subtypes have more than one gene product (e.g., PDE4A, PDE4B, PDE4C, and PDE4D). One fundamental distinction between subfamilies is made on the basis of the difference in affinity for the two distinct cyclic nucleotides. A differentiation is possible between cAMP-specific enzymes (PDE4, 7, and 8), cGMP-specific enzymes (PDE5, 6, and 9), and the so-called dual-substrate PDEs, which have affinity for both cyclic nucleotides (PDE1, 2, 3, 10, and 11; for an overview, see Fig. 1).
CYCLIC NUCLEOTIDES AND PDEs IN NEUROPLASTICITY
As mentioned before, cAMP and cGMP, and concomitantly also adequate PDE functioning, are essential in cellular signaling and a variety of cellular functions. Furthermore, there are indications that they also affect neuronal cell survival and, when functioning incorrectly, may be involved in neurodegenerative processes (4).
Multiple downstream signaling targets of cyclic nucleotide signaling may account for these neuroprotective effects. An important target of both cyclic nucleotides, cAMP and cGMP, in neuronal signaling is cAMP-responsive element binding protein (CREB) (5, 6). CREB is an activity-inducible transcription factor. Crucial to the activation of CREB is its Serine 133 (Ser133) region, where multiple kinases can bind, including protein kinase A (PKA) and mitogen-activated protein kinase (MAPK). On phosphorylation of Ser133, transcription coactivators CREB-binding protein and p300 bind to CREB, which offsets CRE-mediated transcription. Studies investigating CRE-regulated gene expression have associated CREB with upregulation of neurotransmitters, growth factors, and other signaling molecules with important functions in neuroplasticity and neuronal survival (6). The changes in PDE expression and subsequent cyclic nucleotide signaling will therefore affect the level of neuroprotection via CREB (7). Furthermore, cAMP and cGMP trigger the signal translocation signal translocation into the nucleus via activation of MAPK, resulting in downstream activation of antiapoptotic factors such as bcl-2, and conversely the inactivation of proapoptotic Bad (8–12). In vitro, cAMP elevation can rapidly recruit tyrosine-kinase B (TrkB) receptors at the membrane surface by translocation from intracellular stores, which are the main receptors of brain-derived neurotrophic factor (BDNF), thereby enhancing responsiveness to these neurotrophic factors that are essential in neuronal growth and survival (13). BDNF, which activates the MAPK pathway, is also one of the major gene products of CREB-mediated transcription that will be upregulated on cyclic nucleotide level elevation. Furthermore, BDNF/TrkB signaling is also able to activate phosphatidylinositol-3-kinase/Akt cascades, which are generally renowned for their beneficial effects on neuronal survival neuronal survival via bcl-2 activation and Bad inactivation (12).
After acute damage to the central nervous system (CNS), elevation of cyclic nucleotide levels via inhibition of PDEs are able to ameliorate recovery processes. In various in vitro neurotoxicity models, including hypoxia/hypoglycemia-induced and glutamate-induced neurotoxicity, inhibition of PDEs shows a neuroprotective profile, possibly via the suppression of proapoptotic caspase-3 activity (14). Stimulation of cGMP signaling via cGMP analogs and selective inhibition of cGMP-specific PDE5 protects motor and nonmotor neurons to acute reactive oxygen species-induced neurotoxicity in vitro (15, 16). In that same study, no beneficial effects of other dual-substrate or cAMP-specific PDE inhibitors were reported. However, others have found that after spinal cord injury in preclinical settings, elevation of cAMP levels, for example, via inhibition of PDE4, can enhance myelination, survival, and growth of axons as well as functional recovery (17–19). Furthermore, it was demonstrated that in preclinical models of ischemia, PDE3 and PDE5 inhibition increase neurogenesis (20, 21) and also initiate structural changes in preexisting neurons, such as synaptic sprouting and axonal remodeling, to facilitate functional recovery (21–23).
Because of the importance of adequate cyclic nucleotide signaling in neuroplasticity and the distinct characteristics of the different PDE isoforms, PDE inhibition is being evaluated as a target for treatment of a broad spectrum of neurodegenerative disorders. However, in this review, we chose to focus on genetic and protein expression studies on PDEs in neurodegeneration. Literature involving possible therapeutic effects of inhibition of PDEs is beyond the scope of this review (for review on PDE inhibition and CNS disorders, see ref. 24).
The most studied age-related neurodegenerative disorder is without any doubt Alzheimer's disease (AD). The most prominent symptoms of the disease are the progressive decline in cognitive functions, in particular memory. Underlying these symptoms, the disease is characterized by synaptic and neuronal loss, leading to atrophy in mainly temporal and parietal lobe areas. Although the precise causes of the disease are still to beunraveled, the presence of amyloid β-plaques and neurofibrillary tangles in the brain are thought to be critical in these processes.
Both cAMP and cGMP have been suggested to be affected in AD. cAMP and especially its main target protein kinase, PKA, are thought to play a role in the etiology of neurofibrillary tangles via phosphorylation of tau (25, 26). A downregulation of the AC/cAMP/PKA signaling pathway has been reported in human patients with AD (27–29). As a main activator of CREB, a decrease in AC/cAMP/PKA signaling can also account for loss in synaptic plasticity and memory decline in AD (30).
A number of studies have showed changes in expression of cAMP-specific PDEs mRNAs in AD brains. Specifically, an increase in the expression of PDE4A, PDE4B, and PDE7A are observed in early stages of AD, whereas the severest clinical stages are associated with an increase in PDE8B expression in the brain regions associated with memory such as the enthorhinal cortex (31). Furthermore, a decrease of cAMP-specific PDE7A mRNA was reported in white matter tracts, probably representing loss of oligodendrocytes (32). In addition, McLachlan et al. examined the expression of isoforms of PDE4D, which in preclinical research has been suggested to be of particular importance for cognition, in the hippocampus of a patient with AD (33, 34). Although they reported a reduction in the expression of most isoforms of PDE4D (PDE4D 3, 5, 6, 7, 8 and 9), the predominant short form of PDE4D in the human brain, PDE4D1, was strongly increased (33). These findings are in line with the expectations of decreased cAMP signaling in AD. Because the activity of some subtypes of PDEs, such as PDE4, are regulated among other pathways by cAMP/PKA signaling itself (3, 35), the activity and expression of PDE are disturbed due to dysfunctional cell signaling. This implies that changes are secondary to the neuropathology rather than causative.
Similar to the cAMP/PKA pathway, the NO/cGMP pathway is also thought to play an important role in memory processes and is known to be altered in aged brains (36, 37). Therefore, in literature, cGMP is suggested to be linked to AD (36). As for cAMP, the eventual downstream activation of CREB provides an interesting link to cognitive dysfunction and decreased synaptic plasticity in AD (5, 38). Along this line, antiapoptotic effects of NO/cGMP have also been attributed to the upregulation of Bcl-2, an antiapoptotic protein of which the expression is decreased in AD (39). However, in AD postmortem brains, no changes in expression of cGMP-specific PDE9 or dual-substrate PDE2 mRNA was found in comparison with control brains (40).
Major depressive disorder (MDD) is generally considered a neurodegenerative disease, as it is thought to be associated with a decrease in neurogenesis in hippocampal dentate gyrus (41) and neuronal atrophy in the hippocampal CA3 area (42–45).
There is ample evidence from postmortem human brain material that multiple components of the cAMP signaling cascade including AC, PKA, and CREB are downregulated in MDD (46–49). One subtype that has been of particular interest in this regard is the cAMP-specific PDE4. A very recent study quantified in vivo the binding of 11C-(R)-rolipram, a specific PDE4 inhibitor, using PET scans in unmedicated patients with MDD (50). The patients with MDD were reported to show a significant reduction in PDE4 levels, representing a compensatory mechanism in response to decreased cAMP signaling caused by decreased neurotransmission (51).
Wong et al. (52) genotyped single-nucleotide polymorphism (SNPs) in all 21 genes coding for the PDE superfamily in a patient and control population. They found strong evidence for an association between certain polymorphisms in PDE9A and PDE11A and the diagnosis of MDD. In addition, their data are suggestive of an involvement of other PDE genes (PDE2A, PDE5A, PDE6C, and PDE10A) in MDD (52). Furthermore, the PDE11A global haplotype was not only associated with the diagnosis of MDD but also with response to antidepressant treatment (53), with some polymorphisms of PDE11A showing an significant correlation with remission on antidepressants.
Multiple sclerosis (MS) is an autoinflammatory disease of the CNS characterized by white matter lesions. Although the main cause of dysfunctions is due to inflammatory processes in CNS, disease progression and especially irreversible neurological disability are associated with axonal loss (54). cAMP has been identified as an important player in regulatory T-cell-mediated suppression (55). It is also known that increasing cAMP levels reduces inflammatory cellular responses (56–58). A recent study reported that patients with MS show increased mRNA expression in peripheral blood lymphocytes of several cAMP-targeting PDE subfamilies, that is, PDE2, 3, 4, and 7 (59). The levels of expression in the brain itself are however lacking to our knowledge. Taken together, this suggests an important involvement of cAMP signaling in neuroinflammatory processes, possibly originating from changes in PDE levels.
Striatal Motor Disorders
PDEs are also thought to play a role in the pathological mechanisms leading to diseases characterized by striatal degeneration, such as Huntington's disease (HD) and Parkinson's disease (PD). In these disorders, specific interneuron populations in the striatum show disturbed dopaminergic innervations (60). This causes severe motor dysfunction, with HD being characterized by aberrant uncontrollable movement, whereas in PD, the hallmark symptoms include slowness of movement and tremor. With regard to striatal neurodegeneration, the most studied PDE in this respect is dual-substrate enzyme PDE10A, as PDE10A mRNA is being expressed at highest levels in the striatum (61). More specifically, PDE10A seems to be localized at the membrane of dendrites and dendritic spines in the GABAergic spiny projection interneurons (62). In postmortem brains of patients with severe HD, decreased levels of PDE10A in caudate nucleus, putamen, and nucleus accumbens were reported when compared with control subjects (63).
Yet another dual-substrate PDE, Ca2+/calmodulin-dependent PDE1B, is expressed in a similar localization pattern as PDE10A, with highest levels in caudate putamen and nucleus accumbens (61, 64). In general, localization of PDE1B coincides with brain areas rich in D1 dopamine receptors and dopaminergic innervation. In the striatum, PDE1B is localized in spiny interneurons, and this strongly suggests that PDE1B is critically involved in its dopaminergic signaling. The fact that PDE1B and PDE10A have opposite effects on motor behavior points toward a differential involvement of the PDE subtypes in striatal intracellular signaling cascades. Thus, PDE10A is predominantly expressed in medial spinal neurons in the striatopallidal D2 indirect pathway. D2 receptors are negatively coupled to AC and their stimulation decreases cAMP levels leading to behavioral inhibition (65). In contrast, PDE1B is mainly expressed in the striatonigral D1 direct pathway. D1 receptors enhance AC activation, leading to an increase of cAMP levels and behavioral activation (65).
Beyond motor function, HD is characterized by cognitive problems that occur even before onset of motor symptoms (66–68). A recent study has reported a hippocampal hyperactivity of PKA signaling in human patients with HD (69). This was associated with a downregulation in the expression of a number of PDE4 isoforms, including PDE4D1, in hippocampal areas, which suggests that a hyperactivity of cAMP signaling could underlie early cognitive problems in HD.
A disease that is related to HD and PD is the rare autosomal-dominant striatal degeneration (ADSD), which is characterized by dysfunction and morphological changes of the striatum (70). Clinical features resemble PD; however, tremor is absent and response to L-dopa treatment is poor. Recently, it was shown using genetic linkage analysis of an ADSD family that ADSD is caused by a complex frameshift mutation in the gene coding for PDE8B (71). This cAMP-specific PDE8B is highly expressed throughout the brain, with highest levels present in the putamen (71). In ADSD, PDE8B proteins are severely truncated with five of six isoforms having lost all their functional domains. In this light, it might be relevant to take PDE8B into account for other striatal motor disorders, such as PD, although it might be reasoned that, considering the unresponsiveness to l-dopa treatment, ADSD is caused by mechanisms downstream of dopaminergic striatal signaling, whereas the cause of PD is upstream.
Cyclic nucleotides are vital in intracellular signaling and neuroplasticity throughout the brain. Therefore, dysfunctional signaling can be expected to lead to important changes in the functioning of cells and thus might lead to pathological outcomes. Indeed, many neurodegenerative diseases are associated with aberrant cyclic nucleotide signaling related to PDE expression. In some cases, such as in ADSD, evidence even points toward a direct causative link between PDE dysfunction and the disorder. In most other disorders, however, such a straightforward association cannot be made, and the relationship seems rather a compensatory mechanism in response to dysfunctional signal transduction rather than the cause of neuropathology. Nevertheless, abnormal PDE expression and function can be linked to abnormal cyclic nucleotide signaling and thus cellular function possibly leading to neurodegenerative processes. Normalizing cyclic nucleotide levels therefore represents at least a symptomatic approach in the treatment of neurodegenerative disorders. This explains the current large interest in PDE inhibitors as a pharmaceutical drug target. However, given the limited amount of studies that examined PDE expression in human patients, more research should be conducted to obtain a better understanding of the relationship between PDE functioning and neuropathology.