Phosphodiesterase 9 localization in cytoplasm and nucleus: the gateway to selective targeting in neuroprotection?

The umbrella term "neurodegenerative disorders" (NDDs) refers to several conditions characterized by a progressive loss of structure and function of cells belonging to the nervous system. Such diseases affect more than 50 million people worldwide. Neurodegenerative disorders are characterized by sundry factors and pathophysiological mechanisms that are challenging to be fully profiled. Many of these rely on cell signaling pathways to preserve homeostasis, involving second messengers such as cyclic adenosine monophosphate (cAMP) and cyclic guanosine 3′,5′-monophosphate (cGMP). Their ability to control the duration and amplitude of the signaling cascade is given by the presence of several common and uncommon effectors. Protein kinases A and G (PKA and PKG), phosphodiesterases (PDEs), and scaffold proteins are among them. The production of cAMP occurs in response to the activation of the Gs subunit of G-protein-coupled receptors and the increase of species like CO2/HCO3–, Ca2+, and ATP. Subsequently, Gs activates transmembrane and soluble adenylate cyclases, leading to an elevation of intracellular cAMP. PKA-I and PKA-II, are stimulated by cAMP and exert their effects by directly activating proteins or modulating other kinase pathways through phosphorylation. For instance, PKA-I inhibits lymphocyte cell proliferation and immune response, while PKA-II is involved in the regulation of neuronal gene expression and motor learning as well as lipolysis and sperm motility. Their functional differences are granted by variations of expression levels in specific cells, by subcellular localization, and by their binding to scaffold proteins. Specifically, their multifunctional and tissue-specific role is related to their spatiotemporal localization which is supported by A- and G-kinase anchoring proteins (AKAPs and GKAPs) that anchor PKA, PDEs, and phosphatases (Corradini et al., 2013). This complex connects PKA to several receptors and subcellular structures such as endoplasmic reticulum (AKAP-100), mitochondria (AKAP-84), and nucleus (AKAP-95), leading to the formation of signalosomes and enabling efficient sensing and functioning of cAMP in various subcellular locations. Importantly, cAMP can regulate gene expression through phosphorylation of the nuclear transcription factor nitric oxide/cGMP/cAMP-response element-binding protein that in neurons is correlated with proliferation, survival, differentiation, neurogenesis, and neuronal plasticity. Another crucial second messenger is cGMP. It is produced by soluble and particulate guanylate cyclases in response to nitric oxide and natriuretic peptides. Also, cGMP is related to PKG-I and PKG-II, and PDEs as effectors. PKG-I and PKG-II are distinguished by cellular localization given that PKG-I is highly expressed in platelets, smooth muscle, cardiomyocytes, and many endothelial and neuronal cells, while PKG-II is mainly expressed in intestine, kidney, and brain cells. PKG is the major intracellular cGMP target in many cell types, but high cGMP concentrations can cross-activate PKA. It is particularly difficult to identify endogenous substrates for PKG, as it is typically expressed at levels that are 10 to 100-fold lower than those of PKA. Again, PKG is reported to be scaffolded by GKAPs involved in compartmentalization, which controls the formation of cGMP-based signalosomes (Ahmad et al., 2015). For instance, PKG-I is mainly localized in the cytoplasm, but it is also involved in nuclear translocation events, thus supporting cGMP-mediated gene expression. Notably, more than 50 genes and transcription factors are directly or indirectly influenced by cGMP signaling (e.g., antiapoptotic effects mediated by B cell lymphoma 2 (Bcl-2) or Bcl-2–associated proteins in neuronal cells and activation/inhibition of MAP kinase pathways) (Pilz and Casteel., 2003). PDEs are known as major effectors of cAMP and cGMP since they hydrolyze their 3′-phosphate bond to generate 5′-AMP and 5′-GMP, respectively. Given the pathophysiological importance of cGMP and cAMP, it is pivotal to fully understand their regulation mechanisms, which contribute to both the central nervous system and peripheral homeostasis. Among them, PDE9A stands out as it is the one with the highest affinity for cGMP (KM = 70–170 nM), over cAMP (KM = 230 mM). This protein is expressed from the PDE9A gene which is located on chromosome 21q22.3, that includes over 20 exons. Currently, more than 15 isoforms generating from alternative splicing are reported. Such species distinguish themselves in different tissues and subcellular distribution, thus suggesting cGMP regulation in a more subtle way. Interestingly, PDE catalytic sites are conserved among families and the main differences can be found in the regulatory N-terminal domain, which might define cellular and subcellular distribution. PDE9A is not an exception since its N-terminal domain is reported to establish its possible localization. In this sense, a different expression pattern for PDE9A6/13 and PDE9A1 regarding its cellular distribution was reported by Wang et al. (2003), which stated that PDE9A mRNA is expressed at the highest rate in the brain, bladder, spleen, kidney, and small intestine. The authors found that PDE9A1 exhibited mostly a nuclear localization, while PDE9A6/13 were preponderantly in the cytoplasm fraction. This difference may be explained by the presence of a nuclear localization signal in PDE9A1, identified as a pat7 motif, which defines the ability to translocate across the nuclear membrane. Pat7 is a well-defined residue pattern that starts with a proline that is followed within 3 residues by a basic segment containing at least 3 lysine/arginine residues. Nuclear localization signal is required for the action of nuclear transport proteins, which recognize macromolecules allowed to translocate in the nucleus with consequences in gene expression. However, there are different studies reporting that these translocation events in PDEs context can be mediated by other scaffold proteins such as β2-arrestin and karyopherins (Laudette et al., 2018). Targeting karyopherins, due to their pivotal role in regulating nuclear events, is an attractive pharmacological strategy: selinexor is a karyopherin (exportin-1; XPO1) inhibitor approved as an antineoplastic agent, while its derivative eltanexor is currently investigated for Parkinson's disease (Liu et al., 2022). Since PDE9A variants have similar enzymatic properties but differential tissue distribution and subcellular localization, and in light of the abovementioned evidence, it is tempting to speculate that PDE9A1 could have a unique role in regulating events through the hydrolysis of compartmentalized cGMP pools which in turns modulate gene expression via cGMP-effector systems. For instance, in central nervous system, it is known that PDE9A-regulated cGMP signaling is involved in synaptic plasticity, even if the mechanism related to compartmentalized cGMP has not been fully elucidated yet. Patel et al. (2018) showed that PDE9A expression and subcellular localization change across the life span in a manner that is isoform, brain-region, and age specific. It must be noted that in recent years, two mechanisms involving the nuclear translocation of other PDEs have been described. In this regard, PDE1A is mostly enriched in the nuclear fraction of vascular smooth muscle cells, while PDE1C is reported to be more present in the cytoplasm. Nagel et al. (2006) found out that PDE1A inhibition led to an alteration of the cell cycle since the reduction of its nuclear expression significantly increased the number of apoptotic cells through the raise of phosphorylated p53 levels. This suggests the pivotal role of PDE1A and PDE1C and their different impacts on the regulation of basal cGMP homeostasis in vascular smooth muscle cells (Nagel et al. 2006). A similar process was reported by Martinez et al. (2023) for another PDE, since PDE4D5, from the PDE4D cAMP-hydrolyzing family, was identified to be enriched in the nuclei of cortical and hippocampal neurons. On the other hand, PDE4D8 and PDE4D9 were mainly detected in the cytoplasm. It must be noted that memory and learning are supported by activation of the expression of immediate early genes promoted by G-protein-coupled receptors. In this work, the authors showed that β2-adrenergic receptor (β2AR) stimulation induced the export of PDE4D5 from the nucleus. Further, they discovered that phosphorylated-β2AR endocytosis mediated by G-protein-coupled receptor kinases is implied in arrestin3-dependent nuclear export of PDE4D5. This mechanism may aim at limiting nuclear cAMP hydrolysis and promoting nuclear signaling and gene expression in hippocampal neurons. Arrestin3-PDE4D5 complex inhibition did not affect β2AR endocytosis, and direct PDE4 inhibition rescued β2AR-induced nuclear cAMP signaling and improved memory deficits in mice (Martinez et al., 2023). These studies further highlighted the translocation events of PDEs as a mechanism to promote second messengers signaling in specific subcellular locations and in the nucleus in particular. The abovementioned PDEs share the ability to cross the nuclear membrane in response to stimuli, and these mechanisms could be comparable with the one of PDE9A1 (Figure 1). In fact, the unique nuclear localization of PDE9A1, regulated by the pat7 motif, differentiates PDE9A1 from other PDE9A splice variants and other cGMP-hydrolyzing PDE families and implies a potential role of PDE9A1 in the regulation of nuclear events.Figure 1: Regulation of nuclear cGMP signaling by preventing PDE9A1 nuclear localization.In the left panel, a case in which PDE9A1 is overactivated is depicted: guanylate cyclase produces cGMP, while PDE9A1 is transferred into the nucleus thanks to the scaffold complex, and this results in increased cGMP hydrolysis in the nucleus. In the right panel, the proposed mechanism for the regulation of nuclear cGMP-related events is represented: an inhibitor of the binding of PDE9A1 to the scaffold complex prevents nuclear translocation of the enzyme, thus resulting in a local increase of cGMP levels. Created with Microsoft PowerPoint Version 2108 (build 14332.20706) using figures from Servier Medical Art, licensed under Creative Commons Attribution 4.0 Unported License. Bcl-2: B-cell lymphoma 2; c-FOS: Finkel–Biskis–Jinkins osteosarcoma proto-oncogene; cGMP: cyclic guanosine 3′,5′-monophosphate; GC: guanylate cyclase; GMP: guanosine 3′,5′-monophosphate; Jun B: transcription factor encoded by the JUNB gene; pat7: residue pattern 7; PDE9A1: phosphodiesterase 9A isoform 1; Scaffold Complex: comprises scaffold proteins.PDE9A plays an important role in the proliferation, differentiation, and apoptosis of cells via intracellular cGMP signaling. These effects are more visible in tissues where PDE9A is more expressed. Indeed, pathophysiological disorders could be addressed also to PDE9A expression alterations. This might be the case of cancers in which PDE9A is considered a potential prognostic marker. In this context, PDE9A is reported to be enriched in prostate and breast cancer cells, in which the inhibition of this enzyme increases the number of apoptotic cells. Currently, it is not clear if these effects are related to nuclear or cytoplasmic events, thus more studies are required to define which PDE9A isoform is involved in the observed effects. Still, more efforts are needed to understand PDE9A subcellular localization, its interaction with other scaffold proteins, and how it influences cGMP signaling at nuclear and cytoplasmic levels. Considering the potential PDE9A interaction network with other proteins, based on the Human Protein Atlas (proteinatlas.org) interactome, proteins of different classes catch the eye. Additionally, it is interesting to note that these are classified as nuclear or cytoplasmic, and are related to different pathways (Cryptochrome Circadian Regulator 2, related to circadian rhythm; CDC Like Kinase 1, a prognostic marker in urothelial cancer; Tripartite Motif Containing 32, involved in liver and head and neck cancers), suggesting possible specific interactions for PDE9A1 or PDE9A6/13 at nuclear or cytoplasmic levels, respectively (Karlsson et al., 2021). Moreover, since PDE9A is also involved in synaptic plasticity due to its expression in central nervous system, several clinical trials enrolling PDE9A inhibitors have been carried out for different brain disorders. As cAMP-response element-binding protein signaling is downregulated in these conditions, EISAI company recently developed E2027, a selective PDE9 inhibitor that is currently involved in a Phase 2 study for patients affected by dementia with Lewy bodies. A single dose of E2027 was already reported to increase cGMP levels in cerebrospinal fluid during Phase 1 (Landry et al., 2022). Additionally, our group recently demonstrated the potential of PDE9 inhibitors in neuroprotection (Ribaudo et al., 2023; Landucci et al., 2023). Thus, this could represent an attractive topic for future drug development and considering that nowadays drug discovery heavily relies on structure-based computer-aided drug design, it is important to point out that 3D structures of the targets are needed. Unfortunately, there are still some caveats. In the case of available structures of PDE9A1, its N-terminal domain is not completely resolved. It is possible that the region containing the pat7 motif, the potential target for neuroprotective drug development, behaves like an intrinsically disordered protein that needs an interactor to reach structural stability. The first question that arises is if PDE9A1 mechanisms could parallel that reported for PDE1A and PDE4D5, and which are the scaffold proteins involved in these processes. Once these events are fully characterized, the possibility of targeting the N-terminal domain, and pat7 in particular, to regulate the intracellular trafficking of PDE9A with drug candidates should be explored. Future drug discovery rationale should rely on these findings to pave the way for a more accurate design of therapeutic candidates, aiming at specifically influencing nuclear or cytoplasmic PDE9A-related events. C-Editors: Zhao M, Liu WJ, Qiu Y; T-Editor: Jia Y