3.1 GRN mutations in neurodegenerative diseases

3.2 Animal models with PGRN insufficiency

As GRN mutations appear to act through a loss-of-function mechanism, PGRN-deficient mice (Grn^± and Grn^-/-) are the suitable disease models to simulate the disease process of FTLD.

4.1 PGRN acts as a growth factor

PGRN is a growth factor that promotes neurogenesis and brain development. PGRN stimulates adult neurogenesis in the DG of the hippocampus (Asakura et al., 2011; Chiba et al., 2007; Suzuki et al., 2009), the proliferation of PC12 cells (Daniel et al., 2000), and the male-specific differentiation of the neonatal hypothalamus and contributes to normal brain development (Suzuki et al., 1998, 2000). Moreover, it regulates neurite outgrowth and enhances neuronal survival (Van Damme et al., 2008; Gao et al., 2010; Ryan et al., 2009). In a previous study, neurons with Grn deletion showed significantly decreased neurite outgrowth and branching, and PGRN overexpression and exogenous PGRN was able to rescue this phenotype (Gass, Lee, et al., 2012). The neurite outgrowth promoting activity of PGRN has been shown to depend on GRN(E) (De Muynck et al., 2013; Suzuki et al., 2000). Both PGRN and GRN(E) stimulate the proteolytic activity of CTSD in vitro to promote axonal outgrowth and neuronal survival (Beel et al., 2017; Van Damme et al., 2008).

The expression of PGRN in the developing cerebral microvasculature (Daniel et al., 2003) and a profound disruption of the blood-brain barrier (BBB) in Grn knockout (KO) mice after brain ischemia-reperfusion (Jackman et al., 2013) raised the possibility that PGRN is involved in maintaining the structure and function of cerebral microvessels. Under the same injury, the lack of PGRN in the brain increases the severity of the injury. Mice with Grn deletion displayed exaggerated axonal injury in the brain following traumatic brain injury (TBI) (Suzuki et al., 1998). Loss of PGRN increases the susceptibility of the brain to BBB breakdown (Jackman et al., 2013).

PGRN is required for the expression of several receptors in the brain. Estrogen receptor α (ERα) is a classic well-studied ER subtype and is expressed in both neurons and astrocytes (Fang et al., 2018). In Grn-KO mice, PGRN deficiency resulted in a lack of expression of ERα in astrocytes, but did not affect ERα expression in neurons (Doke et al., 2016). The serotonin 1A receptor (5-HT1A) is involved in the inhibition of aggression and anxiety. Hippocampal 5-HT1A mRNA levels were found to be significantly reduced in PGRN-deficient mice after aggressive encounters, suggesting that GRN plays a role in establishing sexual dimorphic behaviors at least partially, by modulating the brain serotonergic system (Kayasuga et al., 2007).

The mechanism of action of PGRN as a growth factor and neurotropic factor has not been fully elucidated. Although SORT1 is regarded as a neuronal receptor for PGRN, PGRN has been reported to regulate neuronal outgrowth independent of sortilin (Gass, Lee, et al., 2012). Furthermore, the cleavage of full-length PGRN into GRNs is required for increased neuronal outgrowth, suggesting that certain GRNs are involved in the neurotrophic functions of PGRN (Gass, Lee, et al., 2012). Other related receptors and signals involved in this regard need to be investigated further.

4.2 PGRN/GRNs in neuroinflammation

Inflammation is a physiological protective response to pathogen exposure, cell injury, and stress. However, dysregulation of the degree and duration of the inflammatory response results in tissue damage. Alterations in PGRN levels are associated with many chronic inflammatory conditions such as rheumatoid arthritis (RA) (Cerezo et al., 2015; Liu, 2011), osteoarthritis (Wei et al., 2017), and inflammatory bowel disease (Thurner et al., 2014), as well as conditions involving acute inflammation such as acute lung injury (Luo et al., 2020; Xie et al., 2018), septic shock (Yu et al., 2016), and acute brain injury (Kanazawa et al., 2015).

PGRN deficiency may promote or amplify the inflammatory response. Elevated brain inflammation in Grn-KO mice was confirmed by increased transcription of pro-inflammatory cytokines and decreased transcription of anti-inflammatory cytokines (Menzel et al., 2017). When exposed to bacterial lipopolysaccharide, Grn-deficient macrophages produce more proinflammatory cytokines and chemokines including CCL2, CXCL1, IL-6, IL-12p40, and TNF-α and less of the anti-inflammatory cytokine IL-10 (Yin, Banerjee, et al., 2010). Moreover, microglia in PGRN-deficient mice showed greater cytotoxicity than normal microglia under the stimulation of inflammatory factors (Yin, Banerjee, et al., 2010). In models of brain injury, Grn-KO aggravated the neuroinflammatory response, axonal injury, astrogliosis, and neuron loss (Martens et al., 2012; Menzel et al., 2017). Most recently, microglial toxicity has been proven to be a key driving factor that promotes neurodegeneration in PGRN deficiency (Zhang et al., 2020). PGRN deficiency promotes microglial toxicity, resulting in endolysosomal dysfunction and neurodegeneration. Deleting C1qa and C3 mitigated this microglial toxicity and rescued TDP-43 proteinopathy and neurodegeneration (Zhang et al., 2020).

PGRN has been shown to inhibit inflammatory responses in brain (Jian et al., 2018; Mendsaikhan, Tooyama, & Walker, 2019). Among the multiple pro-inflammatory cytokines, TNF-α is at the peak of the inflammatory cascade, and the increase of TNF-α correlates with the manifestation of an inflammatory reaction (Bradley, 2008). PGRN is found to directly bind to Types 1 and 2 receptors of TNF (TNFR1 and TNFR2) and offset TNF-α-mediated inflammatory signaling (Tang et al., 2011; Tian et al., 2014).

PGRN and GRNs have been found to possess opposing inflammatory functions. GRNs are pro-inflammatory and may neutralize the anti-inflammatory activities of full-length PGRN (Zhu et al., 2002). Microglia, the resident immune cells in the brain (especially those that have become reactive), produce and secrete high levels of PGRN. In a rat model of cerebral ischemia, the amount of microglial PGRN and the activity of NE increased at an early stage after cerebral ischemia, resulting in the production of GRNs. The inhibition of NE was reported to suppress GRNs production as well as the increase in pro-inflammatory cytokines, suggesting that an increase in PGRN cleavage by NE (to produce GRNs) is involved in the inflammatory response (Horinokita et al., 2019).

4.3 PGRN/GRNs in lysosomal functions

The presence of intracellular aggregates in the brains of dementia patients implies a failure in the protein clearance mechanism and notably, a profound disruption of lysosome function (Chung et al., 2019; Götzl et al., 2016; Hwang et al., 2019; Lie & Nixon, 2019).

PGRN localizes to the lysosomes (Lee et al., 2017) and promotes lysosomal biogenesis and function (Paushter et al., 2018). SORT1 is a high-affinity binding partner of PGRN, and SORT1-mediated sorting is likely to occur both at TGN in the biosynthetic pathway and at the plasma membrane in the process of endocytic trafficking to the lysosomes (Paushter et al., 2018). Like SORT1, the soluble lysosomal protein PSAP is able to transport PGRN from the secretory pathway as well as from the extracellular space. In Sort1-KO mice, PGRN is still successfully transported to neuronal lysosomes (Hu et al., 2010; Zhou et al., 2015), which might be mediated indirectly by the combination of PGRN and PSAP. The transport of PGRN by PSAP depends on CI-M6PR or LRP1, both of which are trafficking receptors of PSAP (Paushter et al., 2018). When transported into lysosomes, PGRN co-localizes with the active CTSL, which is an intracellular PGRN protease within the lysosome (Lee et al., 2017). It has been reported that GRNs, rather than PGRN, are the predominant stable species present in lysosomes and that the GRNs levels are regulated by SORT1 and TMEM106B (Holler et al., 2017).

PGRN facilitates lysosomal clearance (Kao et al., 2017), possibly by controlling lysosomal acidification (Tanaka et al., 2017) and by acting as a chaperone of degradation enzymes (Beel et al., 2017; Evers et al., 2017; Valdez et al., 2017). In neurons, PGRN co-localizes with cathepsin D (CTSD), an aspartyl protease degrading protein present in the lysosomes (Beel et al., 2017), and increases the activity of CTSD but not that of cathepsin B (CTSB) or CTSL (Valdez et al., 2017). PGRN also stimulates the maturation of the CTSD precursor (proCTSD) at an acidic pH (Butler et al., 2019). PGRN-insufficient neurons exhibit impaired lysosomal proteolysis and decreased CTSD activity (Valdez et al., 2017). Elevated CTSD protein levels were observed in the brains of Grn-KO mice (Tanaka et al., 2014) and in FTLD patients with GRN mutations (FTLD-GRN) (Götzl et al., 2014). However, CTSD activity was significantly decreased in a range of tissue lysates from Grn deletion mice (Zhou, Paushter, et al., 2017). Notably, the addition of recombinant PGRN rescued the decreased activity of CTSD in the brain lysates of geriatric Grn-KO mice (Beel et al., 2017). GRNs are localized in the endolysosomal compartment; they cause impairment of lysosomal functions and induce a compensatory HLH-30/TFEB transcriptional response (Butler et al., 2019c). However, the recombinant peptides GRN(BAC) and GRN (CDE) were reported to induce a significant destabilizing effect on proCTSD. This destabilization has been correlated with enhanced CTSD maturation and activity at an acidic pH (Butler et al., 2019). GRN(E) was found to increase the proteolytic activity of recombinant CTSD (Beel et al., 2017; Valdez et al., 2017).

Another lysosomal enzyme regulated by PGRN or GRNs is glucocerebrosidase (GBA), a β-glucosidase that cleaves glucocerebroside into glucose and ceramide. As a co-chaperone of heat shock protein 70 (HSP70), PGRN recruits the GBA trafficking receptor, lysosome integral membrane protein 2 (LIMP-2), and GBA (Jian et al., 2016). Thus, PGRN, mainly through the C-terminal GRN(E) region, is required for the lysosomal localization of GBA, and its deficiency leads to cytoplasmic accumulation and activity reduction of GBA (Jian, Tian, et al., 2016; Jian, Zhao, et al., 2016; Zhou et al., 2019). GBA deficiency has been reported in the brains of patients with Gaucher disease (GD) (Stirnemann et al., 2017), lysosomal storage disease (LSD), and sporadic PD (Gegg & Schapira, 2018). Significantly decreased serum PGRN levels have been reported in GD patients (Jian, Chen, et al., 2018; Jian, Zhao, et al., 2016). Impaired GBA was observed in the brains of FTLD-GRN patients; this was associated with lower levels of mature GBA protein and the accumulation of insoluble, incompletely glycosylated GBA (Arrant et al., 2019). GBA activity was significantly reduced in tissue lysates from Grn-KO mice (Zhou et al., 2019). In addition, GRN-mutant neurons showed lipid accumulation and increased insoluble α-synuclein (Valdez et al., 2020). A recent study using FTLD-GRN patient-derived cortical neurons differentiated from induced pluripotent stem cells (iPSCs) as well as post-mortem tissue from patients with FTLD-GRN showed that PGRN haploinsufficiency resulted in the impaired processing of PSAP to saposin C, a critical activator of GBA (Valdez et al., 2020).

TMEM106B, a highly expressed neuronal protein that is mainly located in the late endosome/lysosome compartments, participates in the regulation of lysosomal morphology and lysosomal trafficking (Schwenk et al., 2014; Stagi et al., 2014). TMEM106B deficiency results in defects in lysosome size, mobility, and lysosomal trafficking (Schwenk et al., 2014; Werner et al., 2020). In mice, Tmem106b deletion leads to accumulation of lysosome vacuoles at the distal end of the axon initial segment in motor neurons and the development of FTLD-related pathology during aging (Feng et al., 2020). However, overexpressing TMEM106B resulted in the accumulation of enlarged lysosomes, delayed the degradation of endocytic cargoes, and exacerbated lysosomal abnormalities caused by the loss of PGRN (Brady et al., 2013; Zhou et al., 2017). PGRN and TMEM106B may be linked synergistically. PGRN deficiency leads to increased TMEM106B protein levels in the mouse cortex with aging and is accompanied by exaggerated lysosomal abnormalities and increased lipofuscin accumulation (Zhou, Sun, Brady, et al., 2017). Thus, PGRN is thought to promote TMEM106B degradation to maintain the proper level of TMEM106B on lysosomal membranes in the aged brain. In mice, TMEM106B deficiency exacerbates PGRN deficit-related FTLD pathologies (Feng et al., 2020; Werner et al., 2020; Zhou et al., 2020). Paradoxically, in another study, loss of TMEM106B ameliorated lysosomal and FLTD phenotypes in Grn-KO mice (Klein et al., 2017).

4.4 PGRN/GRNs in tau aggregation

Tau is a microtubule-associated protein. Tauopathies are a heterogeneous group of neurodegenerative dementias (including AD and FTLD) characterized by the accumulation of abnormal tau, leading to the formation of aggresomes such as NFTs (Almansoub et al., 2019; Mroczko et al., 2019; Papanikolopoulou & Skoulakis, 2020). Independent of the aggregation and destabilization of microtubules, phosphorylated tau increased Aβ toxicity and led to the impairment of synaptic function and memory formation in a mouse model of AD (Ittner et al., 2010). How the abnormal tau protein evades the monitoring of intracellular protein control and metabolic system and finally forms these aggresomes has not been fully clarified (Papanikolopoulou & Skoulakis, 2020). To a large extent, tauopathies may result from mutations in the tau gene (MAPT), dysregulation of alternative splicing, and post-translational modifications or truncation.

In humans, GRN is located on chromosome 17q21, 1.7 Mb centromeric of MAPT (Cruts et al., 2006; Gass et al., 2006). PGRN reduction exacerbates tau pathology. The total level of tau protein was found to be reduced in the brain (Papegaey et al., 2016) but increased in the plasma (Foiani et al., 2018) of FTLD patients. In P301L tau transgenic mice, PGRN protein reduction might contribute to tau hyperphosphorylation and accumulation by upregulating the activity of cyclin-dependent kinases (Hosokawa et al., 2015; Takahashi et al., 2017). In a study of a mouse with the R504X mutation of GRN KI, synapse pathology was specifically associated with tau phosphorylation at Ser203 (corresponding to human Ser214) in the early-stage pathology of FTLD. Specifically, PGRN blocked the activation of Tyro3 and suppressed the activation of its downstream MAPK signaling; PGRN deficiency de-repressed Tyro3 resulting in the activation of PKCα via PLCγ, inducing tau phosphorylation at Ser203, the mislocalization of tau to the dendritic spines, and spine loss (Fujita et al., 2018). However, in the middle temporal gyrus samples of AD patients, increased PGRN protein levels correlating with tau phosphorylation at serine 205 were detected, though PGRN-positive NFTs were not observed (Mendsaikhan, Tooyama, Bellier, et al., 2019).

4.5 PGRN/GRNs in Aβ aggregation

In addition to NFTs, another major feature of AD pathology is the deposition of extracellular SPs composed of aggregated Aβ; Aβ is formed by the amyloidogenic cleavage of the amyloid precursor protein (APP) (Fagiani et al., 2019; Lloret et al., 2019; Wang et al., 2020). Several pathophysiological roles of APP have been reported in neurodevelopmental diseases (autism, fragile X syndrome, and Lesch-Nyhan disease), neurodegenerative disorders [AD, ALS], and multiple sclerosis) and metabolic disorders (diabetes) (Nguyen, 2019). Aβ is toxic to neurons in various ways including promoting apoptosis, causing synaptic loss, disrupting the cytoskeleton, disrupting cellular calcium balance, and loss of membrane potential. There is abundant evidence showing that the soluble oligomeric Aβ generated extracellularly and intracellularly is the primary noxious form (Fagiani et al., 2019; Wang et al., 2020).

In the middle temporal gyrus of the human brain, Aβ plaques with PGRN deposits were identified in non-demented cases with low/smaller plaques, suggesting that this is an early event in plaque formation (Mendsaikhan, Tooyama, Bellier, et al., 2019). In the brains of AD mice, PGRN protein levels, but not mRNA levels, were significantly reduced when there were no or fewer SPs (Minami et al., 2014). However, GRNs have been found to colocalize with Aβ plaques in the brains of AD patients (Pickford et al., 2011) and transgenic AD mice (Pereson et al., 2009). GRN(B) binds to monomeric and oligomeric Aβ₄₂ to promote rapid fibril formation. As Aβ oligomers are well-established neurotoxins, the rapid promotion of fibrils by GRN(B) mitigates Aβ₄₂-induced cellular apoptosis (Bhopatkar et al., 2019).

Selective reduction of microglial PGRN in AD mice impaired phagocytosis, increased plaque load threefold, and exacerbated cognitive deficits (Minami et al., 2014). In contrast, PGRN deficiency significantly reduced Aβ deposition in APP mice and APP/PS1 mice (Hosokawa et al., 2018; Takahashi et al., 2017). It has been reported that global PGRN reduction induces the expression of microglial TYROBP network genes (TNG) and increases AD risk by exacerbating neuronal injury and tau pathology rather than by accelerating Aβ pathology (Takahashi et al., 2017). All these reports suggest that neuronal PGRN and microglia PGRN might have different roles in Aβ production and toxicity.

4.6 PGRN/GRNs in TDP-43 aggregation

TDP-43 is a DNA/RNA binding protein that controls the expression of thousands of genes and is a major component of neuronal and glial aggregates in the brains of tau-negative FTLD, AD, DLB, HS, ALS-D, and TBI patients with dementia (Amador-Ortiz et al., 2007; Arai et al., 2006; Flanagan et al., 2018; Heyburn et al., 2019; Ling et al., 2013; Neumann et al., 2006). It is also associated with Huntington's disease (HD) and stroke (Hatsuta et al., 2019; Heyburn et al., 2019).

Pathological modifications of TDP-43 involve both the loss of its normal function in the nucleus and the toxicity of its cytoplasmic components. Pathological TDP-43 is hyperphosphorylated, ubiquitinated, cleaved into fragments comprising of TDP-43 CTD, and is mislocalized to cytoplasmic protein inclusions and the mitochondria (Bayram et al., 2019; Cascella et al., 2019; Huang et al., 2018). The degree of TDP-43 pathology correlates with the degree of neuronal loss (Mackenzie et al., 2013). Aberrant cleavage of TDP-43 enhances aggregation and cellular toxicity (Zhang et al., 2009). Two C-terminal fragments of TDP-43, a 25-kDa fragment (TDP-25) and a 35-kDa (TDP-35) fragment, are generated by caspases and exist in the neurons of patients with ALS and FTLD (Neumann et al., 2006) or other neurodegenerative diseases (Bayram et al., 2019; Gao et al., 2018). To date, the toxicities of TDP-25 and TDP-35 remain unclear. TDP-35 retains both the RRM1 and RRM2 sequences responsible for the interaction with RNA, thus maintaining its ability to regulate RNA maturation (Kitamura et al., 2016). However, this truncated form has a defective nuclear localization signal and accumulates in the cytoplasm, where it is highly prone to aggregate formation (Bozzo et al., 2016). TDP-25, which is defective in RNA-binding motifs, is not able to bind mRNAs (Kitamura et al., 2016). Both full-length TDP-43 and fragmented TDP-43 accumulate in the cytoplasm and form insoluble inclusions by coacervation with RNA in the cytoplasm and are further toxic to mammalian neurons (Weskamp et al., 2020). TDP-43 also induces mitochondrial damage by decreasing the mitochondrial membrane potential, inhibiting mitochondrial complex I activity, reducing mitochondrial ATP synthesis, and elevating the production of reactive oxygen species (ROS) (Wang et al., 2019). Recently, it was reported that TDP-43 pathology contributes to dementia status and progression in a variety of ways in different phosphorylation states involving both neurons and glia, independent of age and classic AD-related pathologies (Buciuc et al., 2020).

TDP-43 pathology is associated with GRN mutations (Baker et al., 2006; Cruts et al., 2006; Ghidoni et al., 2008; Mackenzie, 2007; Sleegers et al., 2009). PGRN mediates the proteolytic cleavage of TDP-43 to generate TDP-25 and TDP-35 and reduces insoluble TDP-43 levels in mutant TDP-43 mice (Beel et al., 2018). Suppression of PGRN expression leads to caspase-dependent accumulation of TDP-43 fragments (Zhang et al., 2007). PGRN-deficient mice are reported to display greater activation of microglia and astrocytes and more neuronal cytosolic phosphorylated TDP-43 accumulation than wild-type mice (Yin, Banerjee, et al., 2010). Furthermore, mutations or genetic variants of PGRN underlie the abnormal deposition of TDP-43 and the co-occurrence of abnormal deposition of TDP-43 and tau in neurodegeneration pertaining to dementia (Arai et al., 2009; Baker et al., 2006; Cruts et al., 2006). It is still unknown as to how PGRN deficiency causes TDP-43 aggregation. The derailment of lysosomal degradation pathways with reduced clearance of TDP-43 may be involved (Götzl et al., 2014). Cells lacking PGRN show reduced autophagic flux (Chang et al., 2017) and lysosomal dysfunction (Kao et al., 2017; Paushter et al., 2018). Consequently, Grn deletion neurons more readily accumulate pathological forms of TDP-43 (Caccamo et al., 2009; Chang et al., 2017; Paushter et al., 2018). Additionally, PGRN deficiency also affects autophagy and lysosome-independent pathways to induce TDP-43 pathology and neurodegeneration such as impairing Ran-mediated nuclear import of cytoplasmic TDP-43 (Ward et al., 2014). It has been confirmed that PGRN overexpression reduces insoluble TDP-43 levels in TDP-43 (A315T) mice (Beel et al., 2018).

Different from PGRN, specific GRNs may exacerbate TDP-43 cytotoxicity and increase TDP-43 levels (Salazar et al., 2015). Recently, it was shown that GRN(B) and GRN(C) interact with TDP-43 CTD and differentially modulated TDP-43 CTD aggregation in vitro; further, it has been reported that GRN(B) promotes the formation of insoluble aggregates of the TDP-43 CTD, while GRN(C) mediates liquid-liquid phase separation (Bhopatkar et al., 2020).

4.7 PGRN/GRNs in neuronal loss and synaptic deficit

PGRN has multiple neurotrophic effects including stimulating neurite growth, enhancing synaptic connectivity, and promoting neuron survival (Van Damme et al., 2008). PGRN deficiency reduces neural connectivity and leads to increased neuronal loss following injury (Gass, Lee, et al., 2012; Martens et al., 2012). In rat primary hippocampal cultures, downregulation of PGRN decreased neuronal arborization and length, as well as synapse densities. However, the number of synaptic vesicles per synapse and the frequency of mEPSCs increased, suggesting an increase in release at the remaining synapses. The number of vesicles per synapse was also found to be increased in post-mortem brain sections from FTLD-GRN patients with PGRN haploinsufficiency (Tapia et al., 2011). In 10- to 12-month-old Grn-KO male mice, altered synaptic connectivity and impaired synaptic plasticity were observed. In the hippocampal CA1 region of Grn-KO male mice, apical dendrites in pyramidal cells displayed an altered morphology, and decreased spine density was observed (Petkau et al., 2012). GluN2B-containing NMDAR (NR2B) and tau phosphorylation induced by NR2B are important for the regulation of the structural plasticity of neurons (Laurier-Laurin et al., 2014; Tan et al., 2017). PGRN deficiency increased tau AT8 and AT180 pathologies in P301L tau mice (Takahashi et al., 2017). In contrast, PGRN decline diminished the dendritic arborization of primary cortical neurons by reducing the levels and density of NR2B as well as tau phosphorylation (at serine 396/404 and serine 262); this effect was prevented by acute NR2B stimulation (Longhena et al., 2017). In APP/PS1 mice, PGRN-deficient mice exhibited reduced diffuse Aβ plaque growth and less severe axonal dystrophy but enhanced C1q complement deposition at the synapses (Takahashi et al., 2017). PGRN deficiency led to an age-dependent, progressive increase in complement production and enhanced synaptic pruning in microglia (Lui et al., 2016). During aging, Grn-KO mice showed profound microglial infiltration and preferential elimination of inhibitory synapses in the ventral thalamus, which lead to hyperexcitability in the thalamocortical circuits and obsessive-compulsive disorder (OCD)-like grooming behaviors. Remarkably, deletion of the C1qa gene significantly reduced synaptic pruning by Grn-KO microglia and mitigated neurodegeneration, behavioral phenotypes, and premature mortality in Grn-KO mice (Lui et al., 2016). Thus, the upregulation of microglial synaptic pruning might be a key factor that promotes synaptic deficit in PGRN deficiency.

In different pathological models, PGRN overexpression has different effects on neuronal morphology and connections. In TDP-43 mutant mice, PGRN reduced insoluble TDP-43 levels, slowed axonal degeneration, and prolonged survival (Beel et al., 2018). In AD mice, the overexpression of PGRN in hippocampus significantly reduced Aβ plaque burden and revised synaptic atrophy (Van Kampen & Kay, 2017). Fragile X syndrome is an inheritable form of intellectual disability and autistic behavior caused by the loss of the Fragile X mental retardation protein (FMRP, encoded by FMR1) (Salcedo-Arellano et al., 2020), which is an RNA-binding protein that regulates the local translation of numerous mRNAs at the synapses (Prieto et al., 2020). The absence of FMRP causes the overexpression of PGRN. PGRN mRNA and protein were upregulated in the medial prefrontal cortex of Fmr1-KO mice, causing insufficient dendritic spine pruning and late-phase long-term potentiation (LTP). In Fmr1-KO mice, the partial downregulation of PGRN restored spine morphology and reversed behavioral deficits including impaired fear memory, hyperactivity, and motor inflexibility (Zhang et al., 2017). It is noteworthy that FMRP and TDP-43 are associated with ribonuclear protein particles and share mRNA targets in neurons (Ferro et al., 2018; Majumder et al., 2016).

4.8 PGRN/GRNs in the extracellular fluid of patients with dementia

The levels of PGRN and GRNs in the extracellular fluid (CSF or serum/plasma) have diagnostic and predictive value for AD and FTLD. However, they are not helpful in discriminating diseases because a variety of diseases with inflammation also present changes in the levels of PGRN/GRNs.

Mutations in GRN have been considered to be among the most common causes of familial FTLD (Mackenzie et al., 2011; Schymick et al., 2007). Both PGRN and multiple GRNs are haploinsufficient in the brains of FTLD-GRN carriers. It was reported that GRNs were significantly reduced (by ~60%) compared to controls, while PGRN was reduced by ∼40% in soluble lysates of the frontal cortex (Brodmann Area 9) brain tissue from FTLD-GRN cases (Holler et al., 2017). In primary human fibroblasts from FTLD-GRN patients, PGRN and GRNs were significantly decreased by ~50% (Holler et al., 2017). The PGRN levels of FTLD-GRN patients decreased by ~50% in the plasma and CSF (Finch et al., 2009; Ghidoni et al., 2012; Goossens et al., 2018; Meeter et al., 2016). The associations between the decreased levels of CSF PGRN and impairments in categoric and letter fluency, naming, and overall cognition, as well as olfactory dysfunction were observed in FTLD cases (Körtvélyessy et al., 2015). Recently, it was confirmed that plasma PGRN levels predict GRN mutations even in pre-symptomatic carriers more than four decades before disease onset. Although plasma PGRN does not correlate with age at onset and FTLD phenotypes, stable plasma PGRN levels constitute a reliable and cost-effective marker that is suitable as a screening tool in patients with FTLD and more broadly in patients without a family history or with atypical presentations of the disease (Sellami et al., 2020).

A common rs5848 allele in the 3′UTR of GRN is associated with decreased serum and brain PGRN levels and an increased risk of developing AD (Hsiung et al., 2011; Perry et al., 2013; Sheng et al., 2014). PGRN localizes around SPs in the post-mortem brain tissue of AD patients (Gliebus et al., 2009; Mendsaikhan, Tooyama, Bellier, et al., 2019; Pereson et al., 2009). CSF PGRN increases in the course of AD and is associated with soluble triggering receptors expressed on myeloid cells 2 (sTREM2), neurodegeneration, and cognitive decline (Suárez-Calvet et al., 2018). PGRN mRNA is elevated in the peripheral blood of patients with AD and mild cognitive impairment (MCI) (Cooper et al., 2018). In a study involving 107 AD patients and 107 healthy controls, there were no differences in plasma PGRN levels between patients with AD and healthy controls. However, a positive correlation between plasma PGRN levels and age was observed in female AD patients who had a higher PGRN level than male AD patients (Piscopo et al., 2013). Collectively, these data suggest that PGRN has limited diagnostic utility for AD (Morenas-Rodríguez et al., 2016; Piscopo et al., 2013).