Localization of 63-kDa calmodulin-stimulated phosphodiesterase mRNA in the rat brain by in situ hybridization histochemistry

Cyclic nucleotide phosphodiesterases (PDEs) and endothelial function in ischaemic stroke. A review

BACKGROUND

Endothelial dysfunction is a hallmark of cerebrovascular disease, including ischemic stroke. Modulating endothelial signalling by cyclic nucleotides, cAMP and cGMP, is a potential therapeutic target in stroke. Inhibitors of the cyclic nucleotide degrading phosphodiesterase (PDE) enzymes may restore cerebral endothelial function. Current knowledge on PDE distribution and function in cerebral endothelial cells is sparse. This review explores data on PDE distribution and effects of PDEi in cerebral endothelial cells and identifies which PDEs are potential treatment targets in stroke.

METHOD

We performed a systematic search of electronic databases (Medline and Embase). Our search terms were cerebral ischaemia, cerebral endothelial cells, cyclic nucleotide, phosphodiesterase and phosphodiesterase inhibitors.

RESULTS

We found 23 publications which described effects of selective inhibitors of only three PDE families on endothelial function in ischemic stroke. PDE3 inhibitors (PDE3i) (11 publications) and PDE4 inhibitors (PDE4i) (3 publications) showed anti-inflammatory, anti-apoptotic or pro-angiogenic effects. PDE3i also reduced leucocyte infiltration and MMP-9 expression. Both PDE3i and PDE4i increased expression of tight junction proteins and protected the blood-brain barrier. PDE5 inhibitors (PDE5i) (6 publications) reduced inflammation and apoptosis. In preclinical models, PDE5i enhanced cGMP/NO signalling associated with microvascular angiogenesis, increased cerebral blood flow and improved functional recovery. Non-specific PDEi (3 publications) had mainly anti-inflammatory effects.

CONCLUSION

This review demonstrates that non-selective and selective PDEi of PDE3, PDE4 and PDE5 modulated endothelial function in cerebral ischemic stroke by regulating processes involved in vascular repair and neuroprotection and thus reduced cell death and inflammation. Of note, they promoted angiogenesis, microcirculation and improved functional recovery; all are important in stroke prevention and recovery, and effects should be further evaluated in humans.

Phosphodiesterases in the central nervous system

Phosphodiesterases (PDEs) represent important cornerstones of cGMP signaling in various tissues. Since the discovery of PDE activity in 1962, it has become clear that the functional characteristics of PDEs and their role in cyclic nucleotide signaling are fairly complex. On the one hand, members of the PDE family responsible for the hydrolysis of cGMP affect cellular responses by shaping cGMP signals derived from the activation of soluble cytosolic and/or membrane bound particulate guanylyl cyclases. Conversely, PDEs may function as downstream effectors in the cGMP signaling cascade. To make things even more sophisticated, cGMP modulates the activity of several PDEs either directly, by binding to a regulatory domain, or indirectly, through phosphorylation, and the result can be either inhibition or stimulation of the enzyme, depending on the subtype. Furthermore, cross-talk between cGMP and cAMP signaling is achieved by cGMP-dependent modulation of PDEs hydrolyzing cAMP and vice versa. Mammals possess at least 21 PDE genes and often express a set of PDEs in a tissue- and differentiation-dependent manner. Given these premises, it is still a challenging task to elucidate the physiological function(s) of individual PDE genes. The present chapter focuses on the role of PDEs as regulators of neuronal functions. Useful information regarding this topic has been gained by studying (1) the expression pattern of PDEs in the CNS, (2) the association of PDEs with specific macromolecular signaling complexes and (3) the phenotypes associated with mutations or ablation of PDE genes in man, mice and fruit flies, respectively. PDEs degrading cGMP and/or being regulated by cGMP have been implicated in cognition and learning, Parkinson's disease, attention deficit hyperactivity disorder, psychosis and depression. Correspondingly, modulators of PDEs have become attractive tools for treatment of these disorders of CNS function.

Behavioral and neurochemical characterization of mice deficient in the phosphodiesterase-1B (PDE1B) enzyme

PDE1B is a calcium-dependent cyclic nucleotide phosphodiesterase that is highly expressed in the striatum. In order to investigate the physiological role of PDE1B in the central nervous system, PDE1B knockout mice (C57BL/6N background) were assessed in behavioral tests and their brains were assayed for monoamine content. In a variety of well-characterized behavioral tasks, including the elevated plus maze (anxiety-like behavior), forced swim test (depression-like behavior), hot plate (nociception) and two cognition models (passive avoidance and acquisition of conditioned avoidance responding), PDE1B knockout mice performed similarly to wild-type mice. PDE1B knockout mice showed increased baseline exploratory activity when compared to wild-type mice. When challenged with amphetamine (AMPH) and methamphetamine (METH), male and female PDE1B knockout mice showed an exaggerated locomotor response. Male PDE1B knockout mice also showed increased locomotor responses to higher doses of phencyclidine (PCP) and MK-801; however, this effect was not consistently observed in female knockout mice. In the striatum, increased dopamine turnover (DOPAC/DA and HVA/DA ratios) was found in both male and female PDE1B knockout mice. Striatal serotonin (5-HT) levels were also decreased in PDE1B knockout mice, although levels of the metabolite, 5HIAA, were unchanged. The present studies demonstrate increased striatal dopamine turnover in PDE1B knockout mice associated with increased baseline motor activity and an exaggerated locomotor response to dopaminergic stimulants such as methamphetamine and amphetamine. These data further support a role for PDE1B in striatal function.

Phosphodiesterase 1B differentially modulates the effects of methamphetamine on locomotor activity and spatial learning through DARPP32-dependent pathways: evidence from PDE1B-DARPP32 double-knockout mice

Mice lacking phosphodiesterase 1B (PDE1B) exhibit an exaggerated locomotor response to D-methamphetamine and increased in vitro phosphorylation of DARPP32 (dopamine- and cAMP-regulated phosphoprotein, M r 32 kDa) at Thr34 in striatal brain slices treated with the D1 receptor agonist, SKF81297. These results indicated a possible regulatory role for PDE1B in pathways involving DARPP32. Here, we generated PDE1B x DARPP32 double-knockout (double-KO) mice to test the role of PDE1B in DARPP32-dependent pathways in vivo. Analysis of the response to d-methamphetamine on locomotor activity showed that the hyperactivity experienced by PDE1B mutant mice was blocked in PDE1B-/- x DARPP32-/- double-KO mice, consistent with participation of PDE1B and DARPP32 in the same pathway. Further behavioral testing in the elevated zero-maze revealed that DARPP32-/- mice showed a less anxious phenotype that was nullified in double-mutant mice. In contrast, in the Morris water maze, double-KO mice showed deficits in spatial reversal learning not observed in either single mutant compared with wild-type mice. The data suggest a role for PDE1B in locomotor responses to psychostimulants through modulation of DARPP32-dependent pathways; however, this modulation does not necessarily impact other behaviors, such as anxiety or learning. Instead, the phenotype of double-KOs observed in these latter tasks may be mediated through independent pathways.

HIV-1 coat protein gp120 decreases NO-dependent cyclic GMP accumulation in rat brain astroglia by increasing cyclic GMP phosphodiesterase activity

The human immunodeficiency virus type-1 (HIV-1) coat glycoprotein gp120 has been proposed as a likely etiologic agent of HIV-associated dementia (HAD). The pathogenic mechanisms underlying HAD have not yet been fully elucidated, but different evidences indicate that glial cells play an essential role in the development and amplification of the disease. The NO/cyclic GMP (cGMP) system is a widespread signal transduction pathway in the CNS involved in numerous physiological and pathological functions. Increased expression of NO synthase has been reported in the brain of AIDS patients and in cultured rodent glial cells exposed to gp120. The aim of this study was to investigate if gp120 could cause alterations in the metabolism of the NO physiological messenger cGMP that could contribute to the pathogenesis of HAD. Here, we show that long-term treatment (more than 24 h) of rat cerebellar astrocyte-enriched cultures with gp120 (10 nM) induces changes in the cultured cells--astrocyte stellation and proliferation of ameboid microglia--compatible with the acquisition of a reactive phenotype and reduces the capacity of the astrocytes to accumulate cGMP in response to NO in a time-dependent manner (maximal after 72 h). Measurements in cell extracts show that gp120 enhances Ca2+-independent cGMP phosphodiesterase activity by 80-100% without significantly affecting soluble guanylyl cyclase (sGC). Experiments in whole cells using specific phosphodiesterase inhibitors indicate that the viral protein increases the activity of cGMP specific phosphodiesterase 5.

Species differences in the localization of cGMP-producing and NO-responsive elements in the mouse and rat hippocampus using cGMP immunocytochemistry

The aim of the study was to compare the localization of the nitric oxide (NO)-cGMP pathway in hippocampus of mice and rats using cGMP- and soluble guanylyl cyclase (GC) immunocytochemistry and in situ hybridization of the cGMP-hydrolysing phosphodiesterase types 2, 5 and 9. In vitro incubation of hippocampus slices in the absence of a guanylyl cyclase stimulator or a phosphodiesterase inhibitor resulted in cGMP-positive astrocytes mainly in the CA1 area in mouse slices. In contrast, no cGMP immunoreactivity was observed under these conditions in the rat hippocampus. Treatment with an NO synthase inhibitor or inhibitors of soluble or particulate GC did not abolish cGMP immunoreactivity in astrocytes. Incubation with the NO donors sodium nitroprusside or diethylamino NONOate, or with the NO-independent activators of soluble GC, YC-1 and BAY 41-2272, in combination with phosphodiesterase inhibitors, resulted in an increase in cGMP immunoreactivity in numerous astrocytes throughout the mouse hippocampus. In contrast, under these conditions cGMP immunoreactivity was primarily observed in varicose fibers in rat hippocampus. Comparison of the cellular localization of the beta1 subunit of soluble GC and the mRNAs of PDE2, PDE5 and PDE9 revealed that in both species the beta1 subunit was observed in pyramidal and granule cells, which also expressed the mRNAs of the three phosphodiesterase families. Although the beta1 subunit was observed in astrocytes, none of the phosphodiesterases were detected in these cells. We conclude that, although the expression profiles of the soluble GC beta1 subunit and cGMP-hydrolysing phosphodiesterase mRNAs were identical, the cellular patterns of cGMP immunoreactivity differ between rat and mouse hippocampus.

Molecular comparison of rat cyclic nucleotide phosphodiesterase 8 family: unique expression of PDE8B in rat brain

Cyclic nucleotide phosphodiestease (PDE) type 8 is categorized into a family of 3-isobutyl-1-methylxanthine-insensitive PDE hydrolyzing cAMP with high affinity. We have isolated cDNAs encoding rat PDE8A and PDE8B from brain and testis, respectively. The sequence analysis demonstrated that rat PDE8A was a protein of 823 amino acid residues. Rat PDE8B protein was predicted as an N-terminal truncated form of 760 amino acid residues. Both of rat PDE8 proteins include REC, PAS and catalytic PDE domains. Tissue-specific expression patterns of rat PDE8A and PDE8B transcripts were demonstrated by Northern blot analysis. Rat PDE8A transcripts were rich in the liver and testis, and those of rat PDE8B were particularly abundant in the brain and were not expressed in the thyroid gland, while human thyroid gland contains PDE8B transcripts at a high level. Rat PDE8B transcripts were found in all brain regions other than cerebellum and shown to exist in the neuronal cells in in situ hybridization. Mouse PDE8B1 sequence was also identified by a database search and sequence alignment, revealing a protein of 885 amino acid residues, which is 99% and 96% identical to rat and human PDE8B1, respectively. As well as rat PDE8B, expression of mouse PDE8B transcripts was not confined to the thyroid gland. Species-dependent tissue expression pattern was quite unique features of PDE8B.

mRNA expression patterns of the cGMP-hydrolyzing phosphodiesterases types 2, 5, and 9 during development of the rat brain

Recent evidence indicates that cGMP plays an important role in neural development and neurotransmission. Since cGMP levels depend critically on the activities of phosphodiesterase (PDE) enzymes, mRNA expression patterns were examined for several key cGMP-hydrolyzing PDEs (type 2 [PDE2], 5 [PDE5], and 9 [PDE9]) in rat brain at defined developmental stages. Riboprobes were used for nonradioactive in situ hybridization on sections derived from embryonic animals at 15 days gestation (E15) and several postnatal stages (P0, P5, P10, P21) until adulthood (3 months). At all stages PDE9 mRNA was present throughout the whole central nervous system, with highest levels observed in cerebellar Purkinje cells, whereas PDE2 and PDE5 mRNA expression was more restricted. Like PDE9, PDE5 mRNA was abundant in cerebellar Purkinje cells, although it was observed only on and after postnatal day 10 in these cells. In other brain regions, PDE5 mRNA expression was minimal, detected in olfactory bulb, cortical layers, and in hippocampus. PDE2 mRNA was distributed more widely, with highest levels in medial habenula, and abundant expression in olfactory bulb, olfactory tubercle, cortex, amygdala, striatum, and hippocampus. Double immunostaining of PDE2, PDE5, or PDE9 mRNAs with the neuronal marker NeuN and the glial cell marker glial fibrillary acidic protein revealed that these mRNAs were predominantly expressed in neuronal cell bodies. Our data indicate that three cGMP-hydrolyzing PDE families have distinct expression patterns, although specific cell types coexpress mRNAs for all three enzymes. Thus, it appears that differential expression of PDE isoforms may provide a mechanism to match cGMP hydrolysis to the functional demands of individual brain regions.

Striatal phosphodiesterase mRNA and protein levels are reduced in Huntington′s disease transgenic mice prior to the onset of motor symptoms

Inheritance of a single copy of the gene encoding huntingtin (HD) with an expanded polyglutamine-encoding CAG repeat leads to neuronal dysfunction, neurodegeneration and the development of the symptoms of Huntington's disease (HD). We have found that the steady-state mRNA levels of two members of the phosphodiesterase (PDE) multi-gene family decrease over time in the striatum of R6 transgenic HD mice relative to age-matched wild-type littermates. Phosphodiesterase 10A (PDE10A) mRNA and protein levels decline in the striatum of R6/1 and R6/2 HD mice prior to motor symptom development. The rate of reduction in PDE10A protein correlates with the rate of decline of the message and the decrease in PDE10A mRNA and protein is more rapid in R6/2 compared with R6/1 mice. Both PDE10A protein and mRNA, therefore, decline to minimum levels prior to the onset of overt physical symptoms in both strains of transgenic mice. Moreover, protein levels of PDE10A are decreased in the caudate-putamen of grade 3 HD patients compared with age-matched neuropathologically normal controls. Striatal PDE1B mRNA levels also decline in R6/1 and R6/2 HD mice; however, the decrease in striatal PDE10A levels (>60%) was greater than that observed for PDE1B and immediately preceded the onset of motor symptoms. In contrast, PDE4A mRNA levels are relatively low in the striatum and do not differ between age-matched wild-type and transgenic HD mice. This suggests that the regulation of PDE10A and PDE1B, but not PDE4A, mRNA levels is dependent on the relative expression of or number of CAG repeats within the human HD transgene. The loss of phosphodiesterase activity may lead to dysregulation of cyclic adenosine monophosphate (cAMP) and cyclic guanosine monophosphate (cGMP) levels in the striatum, a region of the brain that contributes to the control of movement and cognition.

Phosphodiesterase 1B knock-out mice exhibit exaggerated locomotor hyperactivity and DARPP-32 phosphorylation in response to dopamine agonists and display impaired spatial learning

Using homologous recombination, we generated mice lacking phosphodiesterase-mediated (PDE1B) cyclic nucleotide-hydrolyzing activity. PDE1B(-/-) mice showed exaggerated hyperactivity after acute D-methamphetamine administration. Striatal slices from PDE1B(-/-) mice exhibited increased levels of phospho-Thr34 DARPP-32 and phospho-Ser845 GluR1 after dopamine D1 receptor agonist or forskolin stimulation. PDE1B(-/-) and PDE1B(+/-) mice demonstrated Morris maze spatial-learning deficits. These results indicate that enhancement of cyclic nucleotide signaling by inactivation of PDE1B-mediated cyclic nucleotide hydrolysis plays a significant role in dopaminergic function through the DARPP-32 and related transduction pathways.

Expression of cGMP-specific phosphodiesterase 9A mRNA in the rat brain

cGMP has been implicated in the regulation of many essential functions in the brain, such as synaptic plasticity, phototransduction, olfaction, and behavioral state. Cyclic nucleotide phosphodiesterase (PDE) hydrolysis of cGMP is the major mechanism underlying the clearance of cGMP and is likely to be important in any process that depends on intracellular cGMP. PDE9A has the highest affinity for cGMP of any PDE, and here we studied the localization of this enzyme in the rat brain using in situ hybridization. PDE9A mRNA is widely distributed throughout the brain with varying regional expression. The pattern of PDE9A mRNA expression closely resembles that of soluble guanylyl cyclase (sGC) in the rat brain, suggesting a possible functional association or coupling of these two enzymes in the regulation of cGMP levels. Most of the brain areas expressing PDE9A mRNA also contain neuronal nitric oxide synthase (NOS), the enzymatic source of NO and the principal activator of sGC. PDE9A is the only cGMP-specific PDE with significant expression in the forebrain, and as such is likely to play an important role in NO-cGMP signaling.

The effects of phosphodiesterase inhibition on cyclic GMP and cyclic AMP accumulation in the hippocampus of the rat

The effects of selective and non-selective 3',5'-cyclic nucleotide phosphodiesterase (PDE) inhibitors on cGMP and cAMP accumulation were studied in rat hippocampal slices incubated in vitro. The following PDE inhibitors were used: vinpocetine and calmidazolium (PDE1 selective), erythro-9-(2-hydroxy-3-nonyl)adenine (EHNA, PDE2 selective), SK&F 95654 (PDE3 selective), rolipram (PDE4 selective), SK&F 96231 (PDE5 selective), the mixed type inhibitors zaprinast and dipyridamole, and the non-selective inhibitors 3-isobutyl-1-metylxanthine (IBMX) and caffeine. cGMP levels were increased in the presence of different concentrations of IBMX, EHNA, dipyridamole, vinpocetine and rolipram. cGMP immunocytochemistry showed that incubation with different inhibitors in the presence and/or absence of sodium nitroprusside resulted in pronounced differences in the extent and regional localization of the cGMP response and indicate that PDE activity in the hippocampus is high and diverse in nature. The results suggest an interaction between cGMP and cAMP signalling pathways in astrocytes of the rat hippocampus.

Identification and characterisation of a human calmodulin-stimulated phosphodiesterase PDE1B1

A cDNA encoding a calmodulin-stimulated 3',5'-cyclic nucleotide phosphodiesterase (PDE) was isolated from a human brain cDNA library. The cDNA, designated HSPDE1B1, encoded a protein of 536 amino acids that shared 96% sequence identity with the bovine "63 kDa" calmodulin-stimulated PDE. The recombinant protein had cyclic nucleotide phosphodiesterase activity that was stimulated approximately 2-fold by Ca2+/calmodulin and preferred cGMP as substrate. In addition, the enzymatic activity of HSPDE1B1 was inhibited by phosphodiesterase inhibitors with potencies similar to that displayed toward the bovine PDE1 enzymes: IBMX approximately equal to 8-methoxymethyl-IBMX > vinpocetine approximately equal to zaprinast > cilostamide > rolipram. HSPDE1B1 mRNA was found predominantly in the brain. Lower mRNA levels were found in heart and skeletal muscle. In situ hybridisation of brain revealed expression of HSPDE1B1 predominately in neuronal cells of the cerebellum, hippocampus and caudate. The HSPDE1B1 gene was mapped to human chromosome 12. A partial genomic sequence of HSPDE1B1 was isolated and shown to contain two splice junctions that are conserved in the rat PDE4 and the Drosophila dunce genes.

Effects of the 3',5'-phosphodiesterase inhibitors isobutylmethylxanthine and zaprinast on NO-mediated cGMP accumulation in the hippocampus slice preparation: an immunocytochemical study

The effect of inhibition of 3',5'-phosphodiesterase (PDE) activity on the cGMP accumulation was studied in control and nitric oxide (NO) stimulated hippocampal slices incubated in vitro using immunohistochemical visualisation of cGMP. Isobutylmethylxanthine (IBMX) was used as a non-selective PDE inhibitor and zaprinast was used as a selective inhibitor of cGMP-specific PDE activity. In the absence of PDE inhibitors cGMP-immunoreactivity (cGMP-IR) was found in blood vessel walls only. After incubation with the NO-donor sodium nitroprusside (SNP) cGMP-IR was found in a few isolated varicose fibres which were distributed throughout the slice. Incubation in the presence of either 1 mM IBMX or 10 microM zaprinast resulted in cGMP-IR in small numbers of varicose fibres distributed throughout the hippocampal slice. SNP in combination with IBMX resulted in cGMP-IR in small numbers multitude of varicose fibres throughout the slice; occasionally cell somata were observed. After incubation with SNP and zaprinast cGMP-IR was found in varicose fibres, although with a more restricted distribution and less numerous than in the presence of IBMX. In the latter combination, varicose fibres were observed predominantly in the CA2/CA3 region and in the stratum lacunosum molecular of the hippocampus, and cell somata were occasionally observed throughout the hippocampus. The differential distribution of cGMP-IR in the presence of different PDE inhibitors is consistent with the notion that there are regional differences in the localization of cGMP hydrolyzing enzymes in the hippocampus.

Differential distribution of mRNA encoding cAMP-specific phosphodiesterase isoforms in the rat brain

We studied the distributions of four different cyclic AMP-specific phosphodiesterase isoform mRNAs (APDE1-4) and compared them with that of 63 kDa calmodulin-stimulated phosphodiesterase (CPDE) in the rat brain by in situ hybridization histochemistry using specific radiolabeled oligonucleotides. The distribution patterns were unique for all the APDE isoforms examined here. Although no significant signals for APDE1 could be detected anywhere in the rat brain, all other isoforms were expressed ubiquitously but unevenly and showed overlapping distribution patterns. Among all the APDE isoforms studied here, APDE3 showed the strongest and the most extensive expression. Its distribution pattern implies that it may modulate different cellular processes associated with learning and memory. Compared to APDE3, the levels of expression of APDE2 and APDE4 were weaker, the latter showing the weakest expression. Our study suggests that different isoforms of APDE are expressed together in the same class of neurons implying complex interactions among different signaling pathways, thereby mediating distinct and specific functions.