Regarding PDE activity, Prune_1 has the ability to hydrolyze the second messengers adenosine and guanosine 3′,5′-cyclic nucleotides [i.e., cyclic adenosine monophosphate (cAMP) and cyclic guanosine monophosphate (cGMP)] in their corresponding 5′-monophosphates (i.e., 5′-AMP and 5′-GMP, respectively). cAMP has major affinity to Prune_1, then cGMP (K _m values: cAMP, 0.9 ± 0.03 M; cGMP, 2.3 ± 0.11 M) (6). Phosphodiesterase enzymes have important roles in cellular homeostasis because of the pivotal functions of cAMP and cGMP (as second messengers) in physiological processes due to the modulation of signaling pathway transduction (21). Furthermore, Prune_1 PDE activity has been reported to enhance cell motility in vitro (6). In this regard, MBA-MB-435 cells that overexpressed mutant Prune_1 proteins lacking PDE activity [through amino acid changes within the region containing motif III (DHRP126-129AAAA, PruneΔ) were found with decreased migration rates (6). Furthermore, dipyridamole, a PDE inhibitor, was also found to affect the motility of breast cancer (BC) cells (i.e., MDA-MB231T) (22, 23). These data, thus, suggest that the PDE enzymatic activity of Prune_1 protein may have a role in the modulation of the migratory properties of tumorigenic cells in which Prune_1 is overexpressed.

The PPX/PPase (polyphosphate–phosphohydrolases) activity of Prune_1 was found to exceed its PDE function (7). Prune_1 has affinity for short-chain over long-chain inorganic polyPs. Indeed, it can hydrolyze linear polyPs by using Mg²⁺ or Co²⁺ as a cofactor. Among the inorganic polyPs with different chain lengths, the highest K _cat values for Prune_1 were observed by using triphosphates (P3), adenosine 5′-tetraphosphate (AP4), and guanosine 5′-tetraphosphate (GP4) as substrates. These values dramatically decreased with polyPs at increased chain lengths. Interestingly, PPi, ATP, and GTP were not significantly hydrolyzed (7). Of importance is that loss of PPX/PPase activity was found in the recombinant mutants p.N24H, p.D28A, p.D179A, and p.R348A Prune_1 proteins when assayed in the presence of P3 and Mg²⁺ as the substrate and cofactor, respectively. The other mutants, p.D106A, p.H107N, and p.H108N Prune_1 proteins, had reduced but measurable PPX/PPase activity. In contrast, the mutant p.R128H Prune_1 protein showed an increased K _cat value in comparison with the wild-type protein (7). These results suggest that variations within the DHH domain (e.g., p.D106A, p.H107N, and p.H108N) did not affect the PPX/PPase activity of Prune_1 (7). In contrast, the same PPX/PPase enzymatic function resulted affected in mutant Prune_1 proteins containing variations in those residues conserved in PPase and PPX enzymes (i.e., p.D28A, p.D179A, and p.R348A) (7).

At this time, questions were raised about the potential role of PPX/PPase activity in cell motility. In this regard, a delayed migratory rate was shown in HaCaT cells overexpressing PPX1 (the Saccharomyces cerevisiae homolog of the human PRUNE_1 gene) by performing would healing assays (24). However, different functions were also postulated due to the role that long-chain polyPs play in the induction of ERK1-2 (i.e., MAP kinase pathway) in human cells (25). Thus, how the PPX/PPAse activity of Prune_1 influences migratory processes and cell motility in both physiological conditions is still not fully understood (in the context of polyP degradation and/or organelle cellular compartment storage). Future research studies and discoveries will address these questions.

Inorganic polyPs are polymers of linear orthophosphate (Pi) units ranging from five and several thousand orthophosphates that are linked through phosphoanhydride bonds (one of the most energy-rich linkages). In human, their amounts are very high in blood plasma, platelets, and osteoblasts, where they modulate blood clotting (26), mineralization processes, and the regulation of ATP level (27). Thus, polyPs are involved in many functions mostly related to additional sources of energy into the cells. Interestingly, polyPs were found to act as cytoprotective agents against human immunodeficiency virus type 1 (HIV-1) (28). In this regard, very recently, polyPs have been found able to exert antiviral actions against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) (29), during early phase of infection and replication (30).

Furthermore, high concentrations of polyPs were also found in mammalian brain, especially in astrocytes, where they were found to induce calcium signaling through P2Y1 purinergic receptors and phospholipase C induction, inositol 3 phosphate (IP3) delivery, and cytosolic Ca²⁺ increase (31, 32). Moreover, PolyPs were also reported to amplify the pro-inflammatory response in endothelial cells via the activation of the same P2Y1 receptor (33), thus overall modulating mitochondrial functions. In this regard, the depletion of mitochondrial polyPs (through the overexpression and the activity of yeast PPX) significantly increased the levels of orthophosphates in the mitochondrial matrix (which is necessary for Ca²⁺ binding), thus resulting in a decreased mitochondrial membrane potential and the inhibition of complex I of the respiratory chain (34, 35). Thus, the crucial role of polyPs in cell metabolism is clear, as also confirmed in transgenic mice widely expressing the scPPX1 exopolyphosphatase that displayed a reduced mitochondrial respiration in muscles (36).

The region of Prune_1 responsible for the interaction with GSK-3β starts from 330 to 394 amino acid residues within the C-terminus domain (8, 40). Because of the co-localization between Prune_1 and the focal adhesion proteins (i.e., F-actin, paxillin, and vinculin) in several cell types, this Prune_1–GSK-3β interaction suggests that Prune_1 can also modulate the microtubular architecture and dynamics (40–42). Importantly, GSK-3β is considered a negative modulator of the canonical WNT/β-catenin pathway through phosphorylation at its amino acid residues S9 and S21 (8, 9). Of importance is that, through GSK-3β interaction, Prune_1 was found to activate β-catenin signaling cascade and to further promote Wnt3a secretion in non-small cell lung cancer (NSCLC) cells (8); thus, a connectome related to the WNT signaling activation was identified (43).

Prune_1 was also shown to physically interact with the C-terminus domain of Nm23-H1 through its D388 and D422 residues (37). Nm23-H1 is described as the first anti-metastatic gene with different functions: a nucleoside diphosphate kinase (NDPK) activity that catalyzes the transfer of a phosphoryl group from a nucleoside triphosphate (NTP) to a nucleotide diphosphate (NDP), a histidine kinase function, and a 3′−5′ exonuclease activity (44). In detail, in vitro studies demonstrated that the complex formation involved the dimer of Prune_1 and the hexameric form of Nm23-H1 (38). Interestingly, Nm23-H1 and GSK-3β could bind Prune_1 at the same time by using non-identical regions of its C-terminus domain for their binding, thus suggesting two independent interaction sites for signaling complex assembly (38). Moreover, other in vitro studies demonstrated that the Prune_1/Nm23-H1 interaction complex requires casein kinase I/II-mediated phosphorylation on three serine residues (S120, S122, and S125) of Nm23-H1. Phosphorylated Nm23-H1 on S120, S122, and S125 was also found to increase its hexameric form (41). Of importance the overexpression of both Prune_1 and Nm23-H1 was found in different tumors, such as MB (11), NBL (37) and BC (45). In this regard, the data indicate that the Prune_1/Nm23-H1 complex could affect the anti-metastatic activity of Nm23-H1 and increase the cell motility properties (45, 46).

Other potential Prune_1 interactors were identified through a pull-down assay coupled to mass spectrometry analysis in BC cells (42). Through this approach, new potential Prune_1 binders were identified. Among these, the interaction between Prune_1 and β-tubulin was proposed because of the crucial functions played by tubulins, MTs, and MAPs in neural development (47, 48). Furthermore, this interaction was also confirmed by co-immunoprecipitation assays (Prune_1 protein and both β- and α-tubulins) in NBL inducible cell clones (SH-SY5Y) (10). Of significance is that Prune_1 protein was also identified as a MAP by performing a cell-based microtubule-binding protein spin-down assay using the same neuronal SH-SY5Y cell clones (10). These data were also confirmed by MT polymerization assays in vitro and “in cells,” thus showing an enhancement of the MT polymerization rate (nucleation, growth, and steady-state phases) in the presence of a recombinant wild-type Prune_1 protein (10). Furthermore, Prune_1 was also found to co-localize with MTs (i.e., β-tubulin) in mitotic spindles during cell division (prometaphase, metaphase, anaphase, and cytokinesis) in Hela cells (10). Taken altogether, these data indicate an important role for Prune_1 protein in cellular division processes.

Prune_1 was found to take part in fundamental processes occurring during neural development. In this regard, Prune-M1 and Nm23-M1 (i.e., the murine homologs of human PRUNE_1 and NME-1, respectively) had similar expression patterns in the developing murine nervous system, from early brain development to adulthood (5). In fact, a significant spatiotemporal co-expression of both Prune-M1 and Nm23-M1 during brain development, within the cortex, hippocampus, and midbrain, was observed. Regarding the telencephalon, both genes were found expressed in the ventricular zone (VZ) of the neopallial cortex, a region involved in proliferation processes. Furthermore, the expressions of Prune-M1 and Nm23-M1 were also shown within the intermediate zone and in the ganglionic eminence. These regions allow neurons to migrate to the pial surface of the forebrain to then undergo differentiation processes. Furthermore, Prune-M1 and Nm23-M1 were also expressed in the cerebellum during development in both granular and Purkinje neuron precursors (5). These findings suggest fundamental functions for Prune_1 and Nm23-H1 during cell proliferation, migration, and differentiation processes occurring in the developing brain and mainly in the cerebellum. Of interest is that the expressions of Prune-M1 and Nm23-M1 were also shown during the murine postnatal phases in cortical layers II and IV and in the basal ganglia of the striatum, thus indicating a potential involvement of the complex in somatosensory processing and in synaptic function and plasticity (5). Interestingly, Nme-X1 and Prune-X1 (i.e., Xenopus homolog proteins) were also reported with a potential function in Muller gliogenesis during retinal development in Xenopus (49). This, thus, indicates a role for the Prune_1 and Nm23-H1 complex also during the development of the retina and in eye morphogenesis.

Recently, homozygous and composite heterozygous mutations in PRUNE_1 locus (1q21.3) were found [through whole-exome sequencing (WES) analyses] in several patients worldwide as causative of the autosomal-recessive NMIHBA (MIM #617481). NMIHBA is a severe disease due to the global developmental delay and the intellectual disability. In fact, most of the affected patients presented MCPH and brain, cerebellar, and optic atrophy. Other cerebral abnormalities were also reported in these patients (through brain imaging), such as cortical atrophy, thickness of the corpus callosum, and hypoplasia of the cerebellum. Some patients were also described with seizure, peripheral spasticity, and slight delay in myelination process. Furthermore, the phenotype is often accompanied by a decreased muscle tone (i.e., “truncal hypotonia”), impossibility to ambulate, or speech and broad dysmorphism.

This complex phenotypic spectrum in NMIHBA patients can be reasoned by the genotypic diversity in PRUNE_1 locus. To date, 64 patients carrying different mutations in the PRUNE_1 gene have been reported worldwide ( Figure 1 and Table 1 ). Regarding the genotypic differences, among the variants identified in patients with PRUNE_1, the most representative was the homozygous c.G316A (p.D106N) variant that was found in 15 subjects: seven from Turkey (12, 16, 18), three from Italy (10, 14), one from Sri Lanka (19), one from Caucasus (17), and three from Lebanon (16). The majority of homozygous mutations were found within the DHH domain of PRUNE_1, including c.G88A (p.D30N) in six patients from Oman (10) and one from Saudi Arabia (12), c.160C>A (p.P54T) in seven Iranian subjects (10), c.383G>A (p.R128Q) in two patients from Saudi Arabia (15), and c.515T>C (p.L172P) in three children from North Africa (16). Two types of homozygous mutations were also found within the DHHA2 domain: the missense variant c.C889T (p.R297W) in two patients from India (10) and the frameshift variant c.874_875insA (p.H292Qfs*3) in one subject from Turkey (52). Interestingly, different compound heterozygous mutations were also identified, comprising the c.[G383A];[G520T] [p.(R128Q); (G174X)] variant in two affected siblings from the USA (12) and two European patients (54) and c.[316G>A];[540T>A] [p.(D106N);(C180*)] in two Japanese subjects (17, 20). Furthermore, two different homozygous splicing variants were also described in several patients. In detail, a mutation in the splicing acceptor site (canonical) within intron 4 of PRUNE_1 (c.521-2A>G [IVS4-2A>G]) was found in an Oji-Cree male (13) and in nine patients belonging to four Manitoba-Cree families (51). A different mutation in the splice donor site within intron 2 of the PRUNE_1 gene (i.e., c.132+2T>C) was also described in five patients from two consanguineous Sudanese families (50). Moreover, homozygous deletion (g.1509-84457–151016-662del) starting from exon 2 to exon 8 including the 3′UTR was also found in two female children from Austria (16). Similarly, a homozygous truncating c.50dup (p.Leu18Serfs*8) variant was found in a Japanese child (17). Recently, a start loss c.3G>A (p.Met1)? variant in the PRUNE_1 gene was also reported in two Iranian patients (53).

A genotype/phenotype correlation study was performed to dissect the function of mutant Prune_1 proteins focusing on the cG88A (p.D30N) and c.C889T (p.R297W) variants (10). Of interest is that the recombinant p.D30N and p.R297W mutant proteins (as synthetized and purified in Escherichia coli) were shown to have an increased PPX/PPase activity in vitro on tetraphosphates (P4) as substrates in comparison to the wild-type protein (K _cat/K _m values: wild-type, 0.014 mM/s; p.D30N, 0.312 mM/s; p.R297W, 0.064 mM/s), thus suggesting a “gain-of-function” activity. The same mutant Prune_1 proteins (i.e., p.D30N and p.R297W) were also shown to affect both the cell proliferation and migration rates in NBL inducible cell clones (SH-SY5Y) (10). Furthermore, the MT polymerization kinetics was found delayed in the presence of the same mutant Prune_1 proteins both in vitro and in cells (10). Altogether, these in vitro results suggest that the cG88A (p.D30N) and c.C889T (p.R297W) variants in the PRUNE_1 gene are gain-of-function mutations responsible for the alteration of cell proliferation and migration via impairment of MT polymerization using a neuronal model (i.e., SH-SY5Y cells). These results are in agreement with the homodimeric form of Prune_1 as reported in vitro (38) and confirmed in cells (41).

In contrast, another study reported that the missense c.G316A (p.D106N) and c.G383A (p.R128Q) variants result in a complete loss of PPX/PPase enzymatic activity on both P3 and P4 as substrates with respect to the wild-type Prune_1 protein, thus suggesting a loss-of-function activity for these mutants (54). Additionally, in the same paper, the authors showed an increased proteasome-dependent intracellular degradation of mutant p.D30N and p.D106N Prune_1 proteins using unsynchronized human embryonic kidney 293 (HEK-293) cells. In contrast, cells expressing the mutant p.G174∗ protein showed absence of Prune_1, while those expressing mutant p.R128Q resulted in an amount of Prune_1 protein comparable to that expressed by wild-type HEK-293 cells (54). Furthermore, homozygous null alleles in Prune_1 knockout mice were embryonic lethal at E9.5 with several vascular anomalies, including those in the yolk sac and in the cephalic vascular system (54). Thus, in contrast to the previous study, these data suggest a reduced function of mutant Prune_1 proteins due to hypomorphic variant alleles. Future research studies should be focused on this topic to functionally dissect these “gain or loss” single amino acid changes in PRUNE_1 locus using in vivo state-of-the-art mouse development technology and/or functional analyses in fibroblast-derived organoids from affected patients.

Due to their prominent role in metastatic tumors, targeting the Prune_1/Nm23-H1 protein complex represents a promising therapeutic target for cancer treatment. The protein complex formation between Prune_1 and Nm23-H1 is dependent on casein kinase 1 (CKI)-mediated phosphorylation in S120, S122, and S125 on the C-terminus domain of Nm23-H1 protein (41). Thus, therapeutic strategies aimed to inhibit Prune_1/Nm23-H1 complex formation via targeting these phosphorylation processes could be applied in tumorigenic cells. In this regard, a CPP was developed to mimic the region of Nm23-H1 (from amino acids 115 to 129) that contains these residues (i.e., S120, S122, and S125) that are phosphorylated by the CKI enzyme. The ability of CPP to impair the Prune_1/Nm23-H1 complex was assayed in different tumorigenic cells in vitro ( Table 2 ) (73). As a consequence, CPP reduced the cell proliferation and cell motility in vitro in BC, prostate cancer (PC), CRC, NBL, and MB cells (11, 73). The delivery of CPP in these tumorigenic cells was performed through adenoviral particles carrying the sequences encoding for CPP and for the transactivating protein of HIV. The absence of cytotoxicity in vitro in non-tumorigenic cells (i.e., HEK-293) indicated a specific action of CPP against cancer cells. The same adenoviral approach was also used in vivo (using xenograft murine models) to show the therapeutic benefits of CPP against PC, NBL, and MB ( Table 2 ) (11, 37, 73). Of importance is that the biosafety of CPP in mice was also reported in terms of hematological parameters. Altogether, these data address the potential future use of CPP for the treatment of PC, NBL, and MB.

The impairment of the PDE or PPX/PPase activities of Prune_1 represents another promising anti-tumorigenic strategy to impair tumor progression. In this regard, the PDE activity of Prune_1 was found to stimulate cellular motility in vitro and, as a consequence, metastatic progression. Furthermore, the anti-tumorigenic action of dipyridamole (an anti-platelet aggregation agent and one of the selective PDE inhibitors) in TNBC was also investigated in vivo using xenograft mice ( Table 2 ) (22). Thus, PDE inhibitors could be used as therapeutic agents for Prune_1-overexpressing tumors. Interestingly, dipyridamole did not affect the PPX/PPase enzymatic function of Prune_1, as measured using P3 as a substrate (7).

The PPX/PPAse activity was shown to be diminished by Nm23-H1 in vitro (7). Indeed, Nm23-H1 was able to inhibit the P3-hydrolyzing activity of Prune_1 at micromolar concentrations (7). These data suggest a potential binding competition between Nm23-H1 and the P3 substrate for Prune_1. Thus, small-molecule activators of Nm23-H1 (e.g., NMac1) (74) represent future strategies to impair cancer progression in tumors characterized by the overexpression of Prune_1 and Nm23-H1.

It is of interest that long-chain polyPs (poor substrates of Prune_1) were also found to reduce the Prune_1-catalyzed hydrolysis of P3 substrate (7). This inhibitory effect may be due to Mg²⁺ chelation or to the competition between P3 and long-chain polyPs for active site binding of Prune_1. To date, other studies have reported on the anti-tumorigenic actions of polyPs. In this regard, the anti-metastatic activity of polyPs (P75) was reported for melanoma in vivo using murine xenograft models (75). Furthermore, treatment with polyPs (P120) was found to reduce the intracellular ATP level in NSCLC and increased their sensitivity to X-irradiation (76). Thus, polyPs can be promising targets for the development of novel anti-tumorigenic therapies in humans.

To date, a pyrimido-pyrimidine derivative (AA7.1) was found with the ability to decrease Prune_1 protein intracellular levels may be via enhancement of its proteasomal-dependent degradation (11). Protein–drug interaction studies (via NMR approaches) showed the amino acid residues of Prune_1 that are mainly responsible for the binding to AA7.1 (i.e., L359 and D364) (11). This molecule was also found able to decrease Prune_1 mRNA and protein levels in different MB group 3 (11) and TNBC (59) cells ( Table 2 ). In vitro and in vivo assays also showed the ability of AA7.1 to decrease the cell proliferative and migratory processes ( Table 2 ). In detail, in MB group 3, AA7.1 impaired the metastatic axis driven by Prune_1, thus leading to impairment in TGF-β, decreased levels of OTX2, upregulation of PTEN, inhibition of EMT, reduction in Nestin, and increases in Tuj1 and GFAP differentiation neuronal markers (11). These results suggest the potential of AA7.1 to inhibit the neural stem/progenitor cell markers and to increase neuronal differentiation processes, thus inhibiting the role of Prune_1 in these very important biological processes during neurogenesis.

Furthermore, the pharmacological inhibition of Prune_1 protein in TNBC (achieved through AA7.1 treatment) decreased the number of metastatic foci in vivo also via inhibiting the switch of TAMs in the TME toward the M2 phenotype (59). In TNBC, AA7.1-mediated tumor inhibition occurred through the impairment of the TGF-β pathway, reduction of inflammatory cytokines (i.e., IL-17F) and the modulation of the protein content of extracellular vesicles (i.e., vimentin) (59). Importantly, this small molecule is not toxic, as suggested by the lack of acute toxicity measured in naive mice (Balb/C) that were intraperitoneally administered escalating doses (15, 30, and 60 mg/kg) of AA7.1 daily for 1 week (11). The results showed no immediate acute toxicity of AA7.1 (in terms of hematological parameters, hepatotoxicity, or nephrotoxicity) in treated mice, as measured via glutamate–pyruvate transaminase 1, glutamic oxaloacetic transaminase, creatinine blood levels, and blood urea nitrogen (11). Thus, altogether, these findings suggest that AA7.1 is a non-toxic potential immunomodulatory molecule with the ability to modulate the inflammatory processes in the TME, thus inhibiting the metastatic spread in Prune_1-overexpressing tumors.

To date, none of these therapeutic approaches (as summarized in Table 2 ) has entered clinical testing. However, the pharmacological approach via Prune_1 downregulation (though AA7.1 treatment) has shown therapeutic benefits in murine models of MB and TNBC with the absence of acute toxicity (11). Regarding the inhibition of Prune_1/Nm23-H1 complex formation, CPP was also tested in vivo in xenograft mice of NBL and prostate cancer, thus also representing a promising strategy against tumors overexpressing both Prune_1 and Nm23-H1. Therefore, future studies should be focused on developing i) small molecules (e.g., AA7.1 derivatives) with a higher affinity to Prune_1 in order to reduce the dose to be used in vivo and ii) synthetic peptides with the ability to impair the interaction between Prune_1 and Nm23-H1, hence overcoming the use of adenoviral infection that might suffer from tissue-specific trophism.