Progranulin, lysosomal regulation and neurodegenerative disease

Evolution of progranulin

Progranulin dosage and disease

Several lines of evidence indicate that progranulin has gene-dosage-dependent effects on neurodegeneration. As previously indicated, GRN haploinsufficiency causes familial forms of FTD^2,3. GRN mutations that are associated with neurodegenerative disease (for a list of mutations, see http://www.molgen.ua.ac.be/FTDmutations/) largely result in premature stop codons and nonsense mediated decay, although a few familial mutations impair secretion or produce full-length protein with loss of a key cysteine residue^27,36,37. GRN-mutation carriers have serum levels of progranulin that are more than 50% lower than the levels in non-carriers³⁸, and pathologically these patients develop FTLD with neuronal cytoplasmic protein aggregates enriched in hyperphosphorylated, cleaved TDP43 (REF. 39) (BOX 2).

Individuals who carry two mutant GRN alleles, and thus completely lack progranulin and its cleavage products (functional null), develop a lysosomal storage disease that is now known as NCL (or ceroid neuronal lipofuscinosis-11 (CLN11))^10,11,40. These patients have undetectable levels of progranulin in the plasma and develop neurological deficits, such as vision loss, myoclonic seizures and cerebellar ataxia between 20 and 30 years of age. Skin biopsy reveals lipofuscin accumulation in a vacuole-like organelle, confirming the NCL diagnosis. Interestingly, several studies show that cortical neurons and lymphoblasts from patients with FTLD with GRN mutations also contain a marked increase in lysosomal storage materials^41,42. Together, these results further support the gene-dosage effects of progranulin deficiency in neurodegeneration.

The dosage-dependent effects of progranulin deficiency can also be detected in mouse models. For instance, heterozygous Grn^+/− mice develop age-dependent social and behavioural deficits but show no evidence of gliosis or neurodegeneration⁴³. By contrast, homozygous Grn^−/− mice display premature death and age-dependent gliosis and lipofuscin deposits, recapitulating aspects of lysosomal storage disease^24,44–47. Although the function of progranulin in the lysosomes is still unclear, these results have pivoted the field towards this organelle to understand how progranulin exerts its pleiotropic activities.

Trafficking of progranulin

How might progranulin deficiency cause lysosomal defects? During synthesis, progranulin is translated directly into the endoplasmic reticulum lumen where it associates with protein disulfide isomerases⁴⁸, presumably for co-translational folding and, in humans, formation of its up to 44 disulfide bonds. From there, it is likely to be trafficked through the Golgi via transport vesicles, although the characteristics of these vesicles and the circumstances surrounding their secretion are incompletely understood. Once in the extracellular space, potential fates for progranulin include cleavage into granulins by an extracellular proteases^{12,14,49–51} or uptake into target cells via binding of its C terminus to sortilin³¹. Sortilin ultimately delivers progranulin via trafficking from endosomes to the lysosome, where it may be cleaved and/or degraded in a manner that regulates intracellular levels^31,52 (FIG. 2).

Recently, an alternative route to the lysosome involving another potentially secreted glycoprotein, prosaposin, has been described. Prosaposin and progranulin exhibit striking parallels. Like progranulin, prosaposin cleavage results in a set of cysteine-rich peptides, saposins⁵³, which function in the lysosome to promote sphingolipid hydrolysis⁵⁴ and utilize sortilin as a trafficking receptor^55–57. A homozygous loss-of-function mutation in the gene encoding prosaposin leads to a lysosomal storage disease known as sphingolipidoses, which is similar to NCL in that it results from failed lysosomal macromolecular degradation^54,58. Thus, in some ways, it is not surprising that progranulin and prosaposin form heterodimers^59,60 and that prosaposin can regulate both intracellular and secreted progranulin levels⁶¹ (FIG. 2). In addition, prosaposin provides a direct route for progranulin to reach the lysosome, serving as a molecular bridge between progranulin and mannose-6-phosphate receptor (M6PR) or low-density lipoprotein receptor-related protein 1 (LRP1)⁵⁹. Although much remains to be learned about the relationship between progranulin and prosaposin, their similarities and intertwined fates suggest a cellular coordination that is important in lysosome function and regulation⁹.

Lysosome and neurodegeneration

Despite the subcellular localization of progranulin, the exact functions of intracellular progranulin remain obscure. Evidence supporting a role for progranulin in the lysosome includes its delivery to the endolysosomal system by sortilin, its colocalization with lysosome-associated membrane protein 1 (LAMP1), as well as the finding that Grn^−/− mice demonstrate increased reactivity for LAMP1 in certain brain regions^31,62,63. These results support a role for progranulin in regulating the formation and function of lysosomes.

As a major site of macromolecular breakdown and recycling, lysosomes occupy an integral role in cellular proteostasis. Lysosome and autophagy dysfunction have been key players in neurodegenerative disease pathogenesis⁶⁴. For example, mutations in the gene GBA (which encodes β-glucocerebrosidase) increase the risk of Parkinson disease and Lewy body disease⁶⁵. Homozygous loss of GBA has been associated with a form of Gaucher disease, a juvenile onset lysosomal storage disease. Other variants of NCL-associated genes, including cathepsin D (CTSD) and probable cation-transporting ATPase 13A2 (ATP13A2; also known as PARK9) are also linked to AD and Parkinson disease, respectively^66,67. Recently, attention has focused on the lysosome as a nutrient sensor and signal integration centre through the activity of mechanistic target of rapamycin (mTOR)⁶⁸. Progranulin also has links to the master regulator of lysosome biogenesis, the transcription factor TFEB^69,70. Upon translocation into the nucleus, TFEB activates its target genes to induce lysosomal biogenesis, which increases the degradative capacity of the cell (FIG. 2). Both progranulin and prosaposin are targets of TFEB^69,71. In fact, the promoter of GRN contains TFEB binding sites, and progranulin expression is upregulated by TFEB overexpression^69,72. As such, progranulin is highly co-regulated with other lysosomal genes⁷².

A recent study provided evidence that the cleaved granulins may play a part in lysosome function by impairing degradation of TDP43. In C. elegans, when granulins are expressed in the absence of the full-length protein, they increase levels of neuronal TDP43 and exacerbate its toxicity¹³. This first demonstration of a neurotoxic role for granulins supports a role for full-length progranulin and processed granulins converging in the lysosome. Based on these results, it is tempting to propose that a delicate balance of progranulin and granulins in the lysosome is crucial to regulate the biogenesis and function of this important intracellular organelle. Thus, both a complete absence of progranulin and excessive levels of granulins seem to negatively modulate lysosome function. Regardless, further studies exploring in more detail the relationship, regulation and potential biological interactions between the holoprotein and the granulins are warranted.

Progranulin and brain innate immunity

Given the prominent functions of lysosomes in phagocytic cells, including macrophages and microglia, it is not surprising that progranulin has been implicated in the function of the innate immune system. Progranulin has been shown to be a microglial chemoattractant⁷³, whereas lower-molecular-mass granulins bind to, and potentiate signalling from, macrophage toll-like receptor 9 (TLR9), which is involved in innate immunity⁷⁴. Several studies indicate that complete loss of progranulin causes an aberrant increase in phagocytosis and elevated production of pro-inflammatory cytokines in microglia and macrophages^45,46,75. For instance, acute exposure to the neurotoxin MPTP (1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine) induces more robust microglial activation and neurodegeneration in Grn^−/− mice and microglia-specific conditional Grn-knockout (CD11b–Cre;Grn^fl/fl) mice⁴⁵. In this model, many Grn^−/− and CD11b–Cre;Grn^fl/fl microglial cells exhibit prominent reactive features, as if they are actively engulfing injured neurons. These results support the idea that progranulin is normally required to suppress excessive microglial activation in an acute neural injury paradigm. Interestingly, progranulin seems to have similar roles in suppressing microglial activation during ageing. Transcriptomic profiling of several brain regions from a large cohort of ageing Grn^+/+, Grn^+/− and Grn^−/− mice shows a prominent age-dependent upregulation of genes that are related to lysosomal functions and innate immunity in Grn^−/− mouse brain²⁴. Most genes that are dysregulated in the absence of progranulin belong to microglia, suggesting an essential role of progranulin in suppressing microglial activation in the ageing brain. Interestingly, in wild-type microglia, progranulin is localized to late endosomes, recycling endosomes and early lysosomes²⁴, suggesting that progranulin regulates the formation and/or trafficking of these compartments (FIG. 2). By contrast, Grn^−/− microglia show a marked increase in the size and number of lysosomes²⁴. Consistent with a role for progranulin in regulating intracellular trafficking from endosomes to lysosomes, Grn^−/− microglia exhibit increased synaptic pruning activity and preferentially remove inhibitory synapses in the ventral thalamus²⁴.

How might progranulin deficiency promote microglial activation in Grn^−/− mice during ageing? Previous studies showed that loss of progranulin leads to the upregulation of pro-inflammatory cytokines, including TNF and IL-6, in microglia^45,46. In addition, some complement genes, including C1qa, C1qb, C1qc and C3, were upregulated in Grn^−/− mice before the onset of neurodegenerative features. System-level analysis of these dysregulated genes shows extensive functional interactions between the innate immunity genes and lysosomal genes. As the classical complement pathway is a well-recognized arm of the innate immunity surveillance system⁷⁶, these results suggest that aberrant activation of the complement pathway has key mechanistic contributions to the excessive synaptic pruning observed in Grn^−/− mice (FIG. 3). Indeed, genetic deletion of C1qa reduces synaptic pruning activity in Grn^−/− microglia and protects against the excessive synaptic loss that disrupts sensorimotor integration in the thalamocortical neural circuit. Taken together, these results suggest that neurodegeneration due to progranulin deficiency may be partially caused by altered lysosome function and subsequent aberrant microglial activation. This increased activity then results in overly aggressive synapse removal, thus suggesting a crucial role for microglia as potential drivers of neurodegeneration (FIG. 3).

One prominent and unexpected neurodegenerative feature in Grn^−/− mice is the preferential loss of inhibitory synapses in the thalamocortical circuit. Although both excitatory and inhibitory synapses are tagged with complements, only inhibitory synapses show an age-dependent loss²⁴. As a consequence, the thalamocortical circuit in Grn^−/− mice exhibits hyperexcitability that contributes to the excessive grooming behaviour in these mice. Consistent with the prerequisite role of complement activation in synaptic pruning, removal of C1qa protects the loss of inhibitory synapses, mitigates the hyperexcitability in the thalamocortical circuit, reduces excessive grooming and prolongs the survival of Grn^−/−C1qa^−/− mice. The microglial dysfunction and hyper-excitability in the thalamocortical circuit seen in Grn^−/− mice during ageing bring important insights into the pathogenesis of human disease. Indeed, patients with FTLD with GRN mutations show a progressive increase in levels of complements C1qa and C3 in the CSF, as well as intense microglial infiltration in the frontal cortex. Future studies should determine whether GRN carriers also exhibit evidence of hyperexcitability in the thalamocortical circuit.

Therapeutics discovery

In humans and model organisms, age has been observed to increase progranulin levels^25,75. In addition, stressful stimuli, such as hypoxia, acidosis, hyperosmolarity and inhibition of lysosomal function by artificial alkalinization⁵², seem to increase progranulin production and secretion^72,77,78, alter its glycosylation¹⁸ and possibly promote its cleavage^12,14,50,51. Low circulating levels of progranulin in GRN carriers³⁸ suggest that progranulin expression from the wild-type allele might be a potential therapeutic target. Consistent with this, Cenik and colleagues⁷⁹ designed a luciferase-based screen and identified the histone deacetylase inhibitor suberoylanilide hydroxamic acid (SAHA) as an enhancer of progranulin expression. However, dose–response relationship suggests that SAHA reaches a ‘ceiling’ effect in upregulating progranulin mRNA and protein levels. These results imply that progranulin expression is tightly regulated by epigenetic modifications of the GRN locus and by additional post-translational mechanisms.

Indeed, it has been suggested that prosaposin forms heterodimers with progranulin and regulates the intracellular and secreted levels of progranulin^59–61 (FIG. 2). In addition, secretory leukocyte protease inhibitor (SLPI) seems to directly bind to progranulin to prevent its cleavage¹⁴. Similarly, lysosome alkalinization increases expression of progranulin, presumably by altering its transport, degradation or cleavage into granulins⁵². Finally, several microRNAs, including miR-659, miR-29b and miR-107 (REFS 80–83), have been shown to regulate progranulin protein level by altering translation and/or mRNA stability. Individuals who are TT homozygotes at the rs5848 single-nucleotide polymorphism (SNP) locus show tighter miR-659 binding, which leads to a decrease in progranulin levels and to a 3.2-fold increase in risk of FTLD with TDP43 accumulation⁸⁰. Additional understanding of how progranulin expression and cleavage are regulated may aid in future development of therapeutic strategies.

Future perspectives

The discovery that progranulin may affect formation and function of intracellular organelles and activate microglial innate immunity in the CNS provides new and exciting insights into disease mechanisms and potential therapeutic targets for FTLD. First, the profound microglial activation phenotype in the brains of Grn^−/− mice and patients with FTLD with GRN mutations raises the intriguing possibility that progranulin-deficient microglia may disrupt the homeostatic balance of glia–neuron interactions and promote neurodegeneration in the ageing brain (FIG. 3). Although this can be achieved by activating innate immunity pathways, including pathways related to the complement system and pro-inflammatory cytokines in progranulin-deficient microglia, it is possible that progranulin-deficient microglia could recruit and activate astrocytes that may have additional cytotoxic effects on neurons. In support of this idea, a recent study shows that in a lipopolysaccharide-induced neuroinflammation model, microglia can release IL-1α, TNF and C1qa to activate astrocytes⁸⁴. In future studies, it will be important to determine whether Grn^−/− microglia use similar or entirely different mechanism to activate Grn^−/− astrocytes. Alternatively, loss of progranulin may convert astrocytes into a more aggressive phenotype such that Grn^−/− microglia and Grn^−/− astrocytes cooperatively promote neurodegeneration (FIG. 3).

Second, another important question regarding the fundamental role of progranulin in lysosomes is how this function affects cells beyond microglia and macrophages. Given the presence of progranulin in many cell types, it is likely that progranulin may have broader roles in regulating lysosomes in other cell types, including neurons, with consequences on proteostasis and stress response. In support of this idea, neurons in Grn^−/− mice show pathological features similar to those seen in NCL^44–47. Similar lysosomal phenotypes have also been detected in patients with FTLD with GRN mutations^41,42. Future studies are required to determine whether simultaneous losses of progranulin in microglia, astrocytes and neurons have synergistic effects to promote neurodegeneration (FIG. 3).

Third, the highly evolutionarily conserved granulin repeat motif and the diverse downstream effects of progranulin loss, deficiency and excess suggest that progranulin and granulin have broader roles in ageing and proteostasis outside of the CNS. Indeed, a recent study in African turquoise killifish shows that a variant in the killifish grn gene is in a region that is genetically linked to lifespan regulation⁸⁵. Although multiple other potential lifespan genes are also present in that region, the grn gene is a strong candidate to underlie lifespan differences not only because it is in the region that is the most clearly associated with differences in lifespan but also because it harbours coding variants in the granulin repeat domain analogous to FTD variants. Clearly, precise maintenance of progranulin levels is important, as demonstrated by both its disease associations and the multiple layers of regulation around gene expression (promoter and microRNAs), modification (disulfide bonding and glycosylation), trafficking (sortilin, prosaposin and M6PR), processing (into granulins) and degradation. Evidence even exists supporting a role for progranulin in phagocytosis and removal of amyloid plaques⁸⁶.

Last, progranulin is certainly not alone as a neurodegenerative disease gene implicated in microglial functions. In fact, abnormal activation of innate immunity is a common feature of several neurodegenerative diseases. Although several clinical studies have identified rare GRN-mutation carriers with clinical, imaging and neuropathological evidence of AD^87–89, it remains unclear how patients with those mutations or other GRN variants have a higher propensity to develop AD pathology. Given the complex neurodegenerative features in patients with AD, future studies using a larger series of GRN-mutation carriers and quantitative measurements will be required to provide more mechanistic insights into how reduced progranulin levels may affect the development of amyloid and tau pathology. In addition to the potential role of progranulin in AD pathology, it is possible that reduced progranulin level may promote proteostasis defects via related processes such as stress response and protein clearance.

Conclusion

Although regulation of progranulin and its cleavage products affords great therapeutic potential in multiple disease modalities, clarity regarding how they modulate growth, innate immunity and proteostasis is needed. New model systems to explore these questions, as well as the role of progranulin and granulin in ageing and potentially other yet-to-be identified disorders, would also benefit the field. Efforts to boost progranulin levels have somewhat outpaced our knowledge as to its precise responsibilities in the lysosome in phagocytic and other cell types. Further, little is known regarding how cleaved granulins are produced and regulated, and about their function. Such knowledge could shift efforts, for example, away from enhancing progranulin levels in microglia towards increasing neuronal progranulin. Alternatively, an approach such as inhibiting progranulin cleavage into granulins could be explored. Until a more cogent understanding of progranulin and granulin function is gained, efforts to manipulate their levels for clinical application should proceed cautiously. Nonetheless, progranulin and granulin remain complex yet integral factors in human health with pleiotropic downstream effects that aptly illustrate how the precise regulation of a protein and its cleavage products can contribute to cellular homeostasis and prevention of disease.