The lysosome and proteostatic stress at the intersection of pediatric neurological disorders and adult neurodegenerative diseases

Abstract

In the last two decades, many gene mutations have been identified that when homozygous, lead to childhood neurological disorders, but when heterozygous, result in adult-onset neurodegenerative disease. A shared feature linking these genes? They encode proteins residing in or impacting the function of the lysosome, a key organelle in macromolecular degradation and recycling whose loss leads to the inability to manage proteostatic stress. Here, we propose that lysosomes connect a subset of genetic neurological and neurodegenerative disorders as they occur in two distinct life epochs—development and aging—that endure high levels of proteostatic and other physiological stresses. In this Perspective, we highlight the differing mechanisms of three genes that exemplify this link: glucocerebrosidase A (GBA: Gaucher’s disease and Parkinson’s disease), progranulin (GRN: neuronal ceroid lipofuscinosis and frontotemporal dementia), and tuberous sclerosis complex 1 (TSC1: tuberous sclerosis complex and frontotemporal dementia). We discuss why neurons seem particularly vulnerable to lysosomal dysfunction and ways in which lysosomes potentially contribute to selective neuronal vulnerability. Finally, as disrupted lysosomal catabolism of macromolecules connects these diseases of the nervous system, we propose that they be jointly conceptualized as “Lysosomal Clearance Disorders.”

Introduction

The declaration of the 1990’s as the “Decade of the Brain” initiated a period of funding that has paid major dividends in neuroscience research. The nascent fields of neurodevelopment and neurodegeneration particularly benefited from this investment. These two disciplines were regarded as distinct entities, separated by a lifetime of human experience. However, as the genomics revolution began to identify new loci related to neurological disease, an unexpected event transpired: neurodevelopment and neurodegeneration became irrefutably linked through common gene mutations. Equally surprisingly, many of the shared genes encoded proteins that directly or indirectly regulate lysosomal function.

Lysosomes are highly specialized, membrane-bound organelles that provide a variety of functions, including nutrient sensing, secretion and, perhaps most importantly, macromolecular breakdown and clearance.¹ As part of the normal process of protein homeostasis (i.e., proteostasis), lysosomes are required for turnover of damaged, misfolded and aggregated proteins.² Lysosomes also contain enzymes crucial for sphingolipid and carbohydrate catabolism.³ To function effectively, lysosomes require a highly acidic pH (~4.5 – 4.7) that is achieved through import of protons from the cytosol into the organellar lumen.⁴ Loss of optimal lysosomal activity profoundly disrupts cellular metabolism, leads to the accumulation of undigested proteins, lipids and carbohydrates, and is linked to aging and age-related diseases such as Alzheimer’s Disease.⁵

While lysosomes have only more recently been implicated in neurodegenerative disease, this organelle has long been associated with neurological disorders categorized as lysosomal storage diseases (LSD).⁶ Historically, the rare autosomal recessive gene mutations associated with neurogenetic disorders, such as Gaucher’s disease (GD), have been thought to exert little to no ill effect upon heterozygous mutation carriers. We now know that possession of a single mutant allele responsible for certain childhood disorders—those involving the lysosome—increases risk for a range of adult neurodegenerative disorders, including Alzheimer’s disease (AD), frontotemporal dementia (FTD) and Parkinson’s disease (PD).^5,7 The intersection of these rare childhood disorders with common adult diseases thus signals the importance of the lysosome in maintaining brain health throughout the lifespan. Despite the recognition that glia make major contributions to neurological disease, neurons nonetheless appear to be particularly susceptible to lysosomal dysfunction. Why? Taking a broader view of these disorders together may provide novel insights into this question and new approaches for treatment.

In this Perspective, we consider the dynamic requirements for lysosomal function across the lifespan. We lay out the clinical, genetic, and pathological overlap of childhood neurogenetic and adult neurodegenerative disorders, followed by a focus on specific molecular mechanisms leading to disease. We conclude with a discussion of how the lysosome may selectively impact neuronal health. As defective lysosomal catabolism of cellular macromolecules is culprit in this class of genetic disorders, we propose that they be jointly conceptualized not as lysosomal storage disorders but as “Lysosomal Clearance Disorders.” Nosologically, as we move toward precision medicine, it is likely that diseases will be further subdivided; however, we propose “lysosomal clearance” as a cohesive framework that more accurately reflects the dynamic contributions of failed lysosomal activity to the pathogenesis of these pediatric neurological and adult neurodegenerative diseases.

The lysosome across the lifespan: one organelle, many functions

In the decades after its discovery in the 1950s, the lysosome was conceptualized as a nonspecific “trash can” whose job was restricted to waste removal.⁸ We now understand that the lysosome has numerous, complex roles. It is a center for cellular signaling and signal integration, has responsibilities in trafficking cargo, and functionally interacts with other organelles.^1,9 Even the canonical role of the lysosome, catabolism, is performed by a highly selective and tightly controlled system of over 60 diverse acid hydrolases, each of which contributes in precise and even unique ways to the degradation of lipids, proteins, carbohydrates and nucleic acids.¹⁰ Thus, lysosomal function is critical for cellular metabolism and homeostasis.

The term proteostasis was coined to describe the ability of the cell to maintain correct levels of appropriately folded and functional proteins.¹¹ Though many cellular components contribute to maintaining proteostasis, the lysosome and its many proteases are vital for preventing diseases of protein misfolding. Both normal and maladaptive processes can enhance proteostatic stress. Thus, it follows that at periods of highest proteostatic stress, cells would be rendered most vulnerable to lysosomal dysfunction. The massive cell division and apoptosis of development and the accumulation of environmental and other stressors of aging represent two life epochs with such high metabolic stress. This may also explain why genetic disorders of lysosomal clearance manifest during these two periods. Importantly, while abnormal proteostasis is a hallmark of aging and misfolded proteins are a key feature in both LSDs and neurodegeneration, other macromolecules are processed by the lysosome, evidenced by the finding that both lipid and carbohydrate metabolism are disrupted in these diseases throughout the lifespan.^12,13

Early life epoch of lysosomal stress: Building and maintaining a healthy CNS

Development is the first life epoch of significant proteostatic and metabolic stress (Figure 1, Epoch 1). This comes from numerous sources, including the metabolic needs of cell growth and cell division, cell migration, programmed cell death, the growth of axons and dendrites, synaptic remodeling and signaling, and myelin production and remodeling. Intact lysosomal function is vital for the brain to successfully emerge from development, and the predominance of neurological phenotypes in individuals with LSDs underscores the metabolic demands of the CNS.¹⁴

Figure 1. Proteostatic stress peaks during two life epochs that require enhanced lysosomal functional capacity.

During an early life epoch that includes gestation, infancy and childhood, rapid cell division, growth, functional maturation and apoptosis are required throughout the body for healthy development. Within the CNS, cells differentiate into neurons, oligodendrocytes, astrocytes and microglia that contribute to neuronal synaptic pruning, synaptic remodeling and axonal myelination/remyelination. Each of these processes contributes to proteostatic stress and thus requires intensive lysosomal function to resolve. During young adulthood and early middle age, injury and infection represent the main sources of proteostatic stress, though these stressors occur throughout the lifespan. With progression into a second life epoch that includes later middle and old age, additional environmental and aging stressors impair proteostasis, particularly in terminally differentiated cells such as neurons. Thus, neurodevelopment and aging represent two major life epochs of increased proteostatic stress in which cells, particularly neurons, become vulnerable to disturbances in lysosomal function.

Starting from the earliest embryonic stage, optimal cell division and development rely on functional lysosomes. During every round of mitosis, lysosomes must degrade approximately 140 proteins, and the disruption of lysosomal function leads to prolonged mitotic timing and mitotic errors.¹⁵ Partial inhibition of mTOR, which mimics nutrient starvation and activates autophagy, can pause mouse blastocyst development, preserving “stemness” for weeks in culture.¹⁶ As the CNS develops, asymmetrical division of neural stem cells (NSCs) is required for the differentiation and migration into the complex cellular subtypes and circuits of the brain. These processes require the activity of Notch proteins, transmembrane receptors vital to embryonic development and self-renewal systems.¹⁷ During division of NSCs, lysosomes are asymmetrically distributed between daughter cells. The daughter cell that receives more lysosomes cleaves and activates more Notch signaling in the lysosome, which contributes to maintaining “stemness.” The other daughter cell with fewer lysosomes and less Notch signaling migrates and differentiates.¹⁸ Lysosome activity more broadly contributes to stemness; if neuronal progenitor cells (NPCs) overexpress transcription factor EB (TFEB), this induces lysosomal biogenesis and differentiation is suppressed. Conversely, knockdown of TFEB and TFE3 causes premature differentiation of NPCs.¹⁹ Thus, even before cells become neurons, lysosomes play an important role in CNS cell fate and development, and this role continues into adulthood, where lysosomes help maintain a set of quiescent neural stem cells for adult neurogenesis.²⁰

Lysosomes are also important in apoptosis, a form of programmed cell death (PCD) that is requisite for the appropriate development of the mammalian nervous system.^21,22. The purpose of apoptosis in development is not entirely clear; however, neurons seem to be born and removed at a particularly high rate. For example, of the 131 cells that undergo apoptosis in C. elegans, the great majority (105) are neurons, thus highlighting the importance of PCD for CNS development.²² Lysosomes are critical for apoptosis in two ways. In the cell undergoing apoptosis, the lysosomal membrane is permeabilized in a controlled fashion, allowing the release of lysosomal hydrolases into the cytosol.²³ To complete the process, the dying cell must be engulfed by neighboring astrocytes and microglia into phagosomes, fuse^24,25with their lysosomes, and then be digested and cleared by lysosomal enzymes. The balance between cell proliferation and programmed cell death is altered in LSDs as well as neurodegenerative disorders.^26,27

Lysosomes are also important for synaptic structure and function. Once neurons have terminally differentiated and mature, they form synapses across which neurotransmission occurs. Dendritic spines are the key site for receiving postsynaptic signaling.²⁸ Initially, too many spines are formed, and then a subset is “pruned” by microglia in early childhood, and impairment of pruning is associated with common neurodevelopmental disorders such as autism.^28,29 Even beyond early embryonic development, the number and morphology of dendritic spines remain dynamic and are ultimately distorted in disease.^30,31 For example, lysosomes contribute to the enlargement of spines through fusion and thus donation to dendritic membrane surface area; concurrently, extrusion of the protease cathepsin B into the extracellular space initiates a downstream cleavage cascade that promotes remodeling of the extracellular matrix (ECM).³² On the axonal side of the synapse, lysosomal fusion is implicated in the development of the growth cone and presynapse.³³ As evidence of its importance in these processes, synaptic dysfunction is a key finding in many lysosomal storage disorders.³⁴

A final component of brain development vulnerable to lysosomal dysfunction is axon myelination by oligodendrocytes. First, myelin is a sphingolipid, and this class of lipids is broken down exclusively in endolysosomes by a series of enzymes, many of which are implicated in the sphingolipidoses, a sub-category of lysosomal storage disorders.³⁵ Further, oligodendrocytes accumulate and store a myelin precursor, proteolipid protein (PLP), in the late endosome/lysosome.³⁶ Upon signaling from neurons, PLP is exocytosed to the plasma membrane to form the myelin sheath.³⁷ There, the myelin is compacted by autophagic removal of the cytosol to form a tight sheath.³⁸ Thus, defective myelination is a feature of many LSDs, as well as neurodegeneration,³⁹ and animal models with defective lysosomal exocytosis (VAMP7 mice) or defective lysosomal protease activity (Cathepsin D mutant mice) have mild to moderate demyelination.^40–42

Overall, the lysosome is necessary for numerous components of both prenatal and postnatal development within a variety of CNS cell types. As a result, it is unsurprising that many LSDs feature neurological symptoms. Inducing lysosomal activation by TFEB has been proposed as a mechanism for treatment of LSDs.⁴³

Mid- and late-life epoch of lysosomal stress: Accumulation of CNS-specific proteostatic stress

The second epoch of enhanced proteostatic and metabolic stress occurs with aging (Figure 1, Epoch 2). This comes from two main sources, the accumulation of environmental stressors (fevers, injuries, infection, hypoxia, etc.) and the steady decline in protein fidelity and proteostasis machinery. Post-differentiation, neurons must survive many decades without dividing or being replaced. Thus, they rely more heavily that cells that are turned over on clearance pathways such as the lysosome-autophagy system.⁴⁴ In contrast to neurodevelopment, the stressors of aging accumulate over decades until ultimately reaching what might be considered a tipping point.

In this latter epoch of proteostatic stress, as organisms encounter repeated bouts of diverse environmental hazards, the lysosome plays an important adaptive (and sometimes maladaptive) role in responding to these stressors. For example, in animal stroke models, neuronal and microglial lysosomal functions are impaired by hypoxia, and induction of autophagy by rapamycin, an mTOR inhibitor, improves stroke outcomes by both preventing neuronal death and improving microglial phagocytosis of damaged cell debris.^45,46 In traumatic brain injury (TBI) models, autophagy is activated following injury, though whether this is protective or detrimental is debated.⁴⁷ Cumulative environmental stressors can, over time, impair the very processes that are used to repair their damage. Examples of such include reactive oxygen species, head injury, heat stress and pollution inhibiting lysosomal function.⁴⁸ Thus, while they are demanding more proteostatic capacity, environmental stressors can impair proteostatic machinery.

Increased protein misfolding occurs with aging due to a variety of processes, including increased DNA repair errors that lead to impaired protein fidelity.⁴⁹ These misfolded proteins are largely cleared through the lysosome and may ultimately overwhelm the proteostatic network. Similarly, damaged organelles are cleared by lysosomes through dedicated machinery (mitophagy, ER-phagy, etc).⁵⁰ Impairment of mitophagy has been implicated in disease through strong genetic links with PD, specifically the genes LRRK2, PARKIN, and PINK1.^51–55

Finally, the machinery of the lysosome itself appears to diminish with age. The lysosome is reliant on maintaining an acidic pH for optimal function. In yeast and C. elegans, lysosomal pH becomes more alkaline with age.^56,57 In primary mouse neuron culture, more days in vitro (DIV) results in de-acidification of lysosomes, particularly in neurites.⁵⁸ Similarly, in transdifferentiated neurons from aging and Alzheimer’s disease patients, there is lysosomal alkanization.⁵⁹ Though it is unknown if lysosomal alkalinization with age occurs in vivo in mammalian models, it would be expected to diminish lysosomal clearance of misfolded proteins, resulting in the pathologic accumulation seen in disease. Moreover, certain disease-associated proteins have been shown to impair lysosomal acidification. In particular, amyloid beta accumulation in neurons in mouse models of AD has been shown to result in faulty lysosomal acidification.⁶⁰ Additionally, chaperone-mediated autophagy (CMA), described in detail below, diminishes with age.⁶¹ Overall, aging appears to represent the confluence of both diminished lysosomal function and accumulating need for lysosomal catabolism.

The two epochs of proteostatic stress explain the genetics of pediatric neurological disorders versus adult neurodegenerative diseases

The early and later-life stress epochs requiring relatively increased lysosomal capacity may explain why lysosomes lie at the nexus of genetic neurodevelopmental and neurodegenerative disorders. In this framework, individuals homozygous for a disease-associated mutant allele cannot meet the lysosomal demand of early development, while heterozygous individuals can tolerate these fundamental early life processes. Then, later in life, heterozygous mutation carriers become susceptible to disease as lysosomes cannot resolve the accumulated stressors of aging (Figure 2). In essence, aging acts as the second mutant allele in neurodevelopmental disorders.

Figure 2. Lysosomal clearance capacity declines gradually in normal aging but is progressively accelerated in those carrying heterozygous and homozygous mutations in lysosomal proteins.

Though lysosomal clearance capacity diminishes with age, by definition, healthy aging (pink line) does not result in neurodegeneration. In our proposed model, individuals who carry a single risk allele for a lysosomal disease gene (green line) can successfully navigate the first epoch of proteostatic stress but become susceptible to impaired proteostasis with the later epoch of aging. Once lysosomal clearance capacity dips below a certain threshold (indicated by the dashed horizontal line), lysosomes are no longer able to adequately support proteostasis, and cell death and neurodegeneration ensue, including in non-CNS cells. Individuals with two risk alleles (blue line) begin life with a much lower lysosomal clearance capacity and thus pass the threshold for cell death and degeneration sooner, during the first epoch of proteostatic stress. Variability in the age of onset and degree of lysosomal dysfunction is denoted by the overlapping colors.

While depicted as separate in Figure 1, in reality, individual genetic variations and differences in environmental exposures can cause these life “epochs” to overlap in regard to age of disease presentation. For example, some “neurodevelopmental” disorders can present later in the teens to early adulthood, while early-onset “neurodegenerative” disorders can be seen in the 4^th or 5^th decade of life. A striking example of this overlap is Kufor-Rakeb syndrome, an autosomal recessive early-onset form of Parkinson’s disease caused by mutations in the lysosomal ATPase protein, ATP13A2. It can present in adulthood, quite late for a neurodevelopmental disorder.^62,63

The clinical and pathologic overlap between neurodevelopmental and neurodegenerative disorders involving the lysosome also provides insight into their pathophysiology and will be discussed below.

The genetic and pathological intersection between neurodevelopmental and neurodegenerative diseases

On the surface, the clinical features of neurodevelopmental and neurodegenerative diseases involving the lysosome would appear to be quite distinct. Besides the highly divergent age of onset, lysosomal storage diseases tend to affect multiple organ systems, while neurodegenerative disorders remain limited to the nervous system, even late in disease (Figure 3). However, upon closer examination, pathological and mechanistic features underpinning these disorders are shared. Most obviously, the nervous system is the earliest and most profoundly affected in these diseases connected by the lysosome.⁶

Figure 3. Comparison of organ systems affected in lysosomal storage disorders compared to neurodegenerative diseases.

There are over 70 LSDs, and they vary widely in their symptomatology, affected organs, and age of onset. Neurodegenerative diseases, conversely, typically involve primarily the nervous system, highlighting the sensitivity of neurons to lysosomal dysfunction.

Lysosomal Storage Disorders (LSDs) are a broad group of conditions, primarily monogenic, resulting from mutations in genes encoding proteins fundamental to lysosomal function. There are over 70 known LSDs which together have an incidence of 1 in 5,000–7,000 people.⁶⁴ LSDs are a set of disorders with heterogenous phenotypes. Neurological symptoms including blindness, deafness, seizures, and cognitive and motor decline, as well as non-CNS involvement with pulmonary, GI, cardiac, hematological, and bone manifestations.⁶⁵ Neurological symptoms, however, are most common, impacting two-thirds of affected individuals.⁶⁶ Many LSDs can feature neurodevelopmental symptoms, such as autism and intellectual disability.^67–69 Importantly, while these are generally disorders of infancy and childhood, the neurological phenotypes of LSDs typically represent both a failure of appropriate development as well as neurodegeneration.¹⁴ In these cases, neurodegeneration is seen phenotypically as developmental milestones are initially gained normally and then lost with disease progression. Thus, neuronal loss is a shared feature of both LSDs and adult neurodegenerative disorders.

Gaucher Disease (GD), the most common LSD, exemplifies the complexity of this group’s presentations, ranging from perinatal lethality to mildly symptomatic.⁷⁰ For half a century, homozygous loss-of-function mutations in glucocerebrosidase A (GBA) have been known to underlie GD.⁷¹ In 2004, a landmark study found that heterozygosity for GBA mutations increased risk for Parkinson’s Disease and was associated with a younger age of onset.⁷² Since then, many additional genes straddling diseases of childhood and aging have been identified (see Table 1). Among these, the majority had been previously defined, based on electron microscopy findings, as lysosomal storage diseases. Most lysosomal storage disorders have classic autosomal recessive inheritance, thus allowing for precise identification of the causative gene in affected families.⁶⁵ With the identification of over 50 different gene mutations that cause LSDs, the term “lysosomal storage disorder” itself has largely fallen out of favor.⁷³ The LSDs can now be classified by the type of macromolecule that accumulates, which is often a direct result of the lysosomal enzyme or protein affected. For example, GD and Krabbe Disease can both be classified as sphingolipidoses, since they are due to deficiencies in enzymes that break down these ceramide-based lipids. The mucopolysaccharidoses are similarly grouped due to the accumulation of heparan-sulfate and related molecules. In contrast, neuronal ceroid lipofuscinoses (NCLs) have a more varied set of affected enzymes.

Table 1. Genes implicated in both neurodevelopmental and neurodegenerative diseases.

Many genes that are causal for lysosomal storage disorders have been found to increase risk for neurodegenerative disorders. This chart compares the LSD, accumulated storage product, and neurodegenerative disorder. PMID for the neurodegenerative risk papers are provided. ALS, amyotropic lateral sclerosis; AD, Alzheimer’s disease; FTD, frontotemporal dementia; MSFD8, major facilitator superfamily domain containing 8; PD, Parkinson’s disease; SAPs, sphingolipid activator protein; SCMAS, subunit C of mitochondrial adenosine triphosphate synthase.

Disease Family	Gene	Protein	Accumulated Substance	Neurodevelopmental Disease	Neurodegenerative Disease	PMID
Sphingolipidoses	GBA	Glucocerebrosidase A	Glucocerebroside	Gaucher’s Disease	PD	15525722,19286695 19846850
	ASAH1	Acid Ceramidase	Ceramide	Farber Lipogranulomatosis	PD	29140481
	ARSA	Aryl Sulfatase	Sulfatide	Metachromatic Leukodystrophy	PD	31312839
	GALC	Galactosylceramidase	Glyoclipids, galactocerbroside	Krabbe Disease	PD	31701892 36370000
	GLB1	Beta-Galactosidase	GM1 gangliosides	GM1 Gangliosidosis	PD	29140481
	NPC1	NPC Intracellular Cholesterol Transporter 1	Cholesterol, glycolipids	Niemann-Pick Type C	PD	29867446 32371106
	NPC2	NPC Intracellular Cholesterol Transporter 2	Cholesterol, glycolipids	Niemann-Pick Type C	Corticobasal Syndrome	27792009
	SCARB2	Lysosomal Integral Membrane Protein Type 2	Glucocerbroside	Action Myoclonus Renal Failure Syndrome/GD	PD	21738487,22223122, 27110593
	SMPD1	Acid Sphingomyelinase	Sphingomyelin	Niemann-Pick Type A/B	PD	23535491,30788890
Neuronal Ceroid Lipofuscinoses	GRN	Progranulin	Unknown lipopigment	NCL11	FTD, PD, AD	16862116,3199628, 23609919
	CLN5	BMP synthase	SC MAS	NCL5	AD	30037983
	CTSD	Cathepsin D	SAPs	NCL10	AD, PD	16543533,29140481
	CTSB	Cathepsin B	N/A	None	AD, PD	35379992,31701892
	CTSF	Cathepsin F	Unknown lipopigment	NCL13	AD	27524508
	MSFD8	MSFD8	SCMAS	NCL8	FTD	30382371
	ATP13A2	ATPase cation transporting 13A2	Unknown lipopigment	NCL12/Kufor Rakeb	Early Onset PD	25900096
Sialic Acid Storage	SLC17A5	Solute Carrier Family 17 Member 5	Sialic acid	Salla Disease	PD	29140481
Mucopolysaccharidosis	NAGLU	N-acetyl-alpha-glucosaminidase	Heparan sulfate	Mucopolysaccharidosis IIIB	Early-onset polyneuropathy	25818867
	IDUA	Alpha-L-iduronidase	Heparan and dermatan sulfate	Mucopolysaccharidosis I	AD, PD	35379992
Other Neurogenetic Disorders (non lysosomal storage disorders)	TSC1	Tuberous Sclerosis Complex 1		Tuberous Sclerosis	FTD, AD, PSP	28828560,34739309
	TMEM106B	TMEM106B		Hypomyelinating Leukodystrophy	FTD, AD	20154673,22855871, 21220649
	SQSTM1	p62		Neurogeneration with Ataxia, Dystonia, and Gaze Palsy	ALS, FTD	22084127

At the other end of the age spectrum, adult-onset neurodegenerative disorders feature a progressive decline in cognitive and/or motor function, with onset typically in the 5^th decade of life or later. Each of the more common adult neurodegenerative disease is characterized by the buildup of specific protein aggregates, with clinical symptoms deriving from affected brain regions. For example, in Alzheimer’s disease (AD), amyloid beta (Aβ)-containing amyloid plaques and neurofibrillary tau tangles accumulate in the hippocampus and limbic system and manifest as memory decline.⁷⁴ In Parkinson’s disease (PD), tremor, bradykinesia and other motor symptoms correlate with alpha-synuclein-containing Lewy body inclusions in the substantia nigra.⁷⁵ In contrast to PD, dementia with Lewy Bodies (DLB) presents with frontal executive dysfunction and cognitive fluctuations associated with Lewy body inclusions in the amygdala and cortex.⁷⁶ Frontotemporal dementia (FTD) can be associated with numerous pathological substrates including tau, TDP-43, and FUS.⁷⁷ The type of inclusion in FTD correlates well with brain regions involved, clinical symptomology, and the overall disease course of FTD.⁷⁸

The presence of a pathognomonic protein aggregate in each of these disorders has allowed for the formation of clinical cohorts and, importantly, the identification of families with autosomal dominant forms of disease. This subsequently led to the identification of some of the earliest disease-associated genes, including APP, PS1 and PS2 in AD, SNCA and LRRK2 in PD and MAPT and GRN in FTD.^79–87 Autosomal dominant familial genetic mutations are responsible for only a small minority of cases, thus, the list of neurodegenerative disease genes remained relatively short until the turn of the century. As more genome-wide association studies (GWAS) and familial cohort sequencing studies have been performed, many additional risk and rare causative alleles have been found.⁸⁸ Because these neurodegenerative disease genes were identified over several decades, only in retrospect did it become clear how many were already associated with a cognate neurodevelopmental disorder.

When the genes are organized based on disease family (Table 1), a correlation, albeit non-perfect, between pediatric disease family and a particular adult neurodegenerative disease becomes apparent. For example, the sphingolipidoses clearly cluster with PD, while AD the FTD are spread across the other disease families. In addition to the LSDs, other lysosomal neurogenetic disorders have been linked to adult neurodegenerative disease. These genes affect autophagy-related proteins (SQSTM1 and TSC1) or lysosomal-resident proteins whose in vivo function has yet to be identified (TMEM106B). The associated childhood disorders highlight the variety of processes in which the lysosome is involved, from myelination (TMEM106B, hypomyelinating leukodystrophy) to cell growth (TSC1, tuberous sclerosis) and bone resorption (SQSTM1).^89–91 Additionally, some genes whose mutations cause childhood disorders (TRPML1, mucolipidosis IV; VAC14, striatonigral degeneration, childhood-onset) encode proteins that are linked to lysosomal dysfunction in neurodegenerative disease, though no genetic risk factors have been identified in these genes for adult diseases.^92,93 We will now dive more deeply into the genetics and mechanisms of three of these genes: GBA1 in sphingolipidosis and PD, GRN in NCLs, AD and FTD and TSC1 in tuberous sclerosis complex (TSC) and FTD.

Genetics and Mechanisms of Lysosomal Dysfunction in Disease

Sphingolipid processing in the lysosome: Sphingolipidoses, glucocerebrosidase and PD

Why do mutations in sphingolipidosis genes affect lysosomal function? Lysosomal membranes are particularly enriched in sphingolipids. With aging, the relative balance of sphingolipids is disrupted, with certain species, such as short-chain ceramide, being relatively increased in aging and sphingomyelin being relatively reduced.⁹⁴ Sphingolipid imbalance has been observed in lipidomic studies of neurodegenerative diseases, including PD.⁹⁵ Thus, reduced clearance of sphingolipids due to loss of sphingolipid progressing enzymes such as GBA may exacerbate derangements of sphingolipid levels with aging, impacting lysosomal function.

Specific to PD pathology, altered sphingolipid metabolism may promote alpha-synuclein accumulation. The GBA substrates glucosylceramide (GlcCer) and its deacetylated derivative glucosylsphingosine (GlcSph) enhance nucleation of alpha-synuclein aggregates. Reducing GlcCer levels by inhibiting GlcCer synthase reduces alpha-synuclein aggregation in mouse models of GD or PD.^96,97 In addition to lipid derangements, loss of function of GBA directly impairs global lysosomal function and protease activity, resulting in reduced alpha-synuclein clearance via macroautophagy.⁹⁸ As part of the cellular response to lysosomal dysfunction, exocytosis of autophagic vesicles containing alpha synuclein are released to the extracellular space, which may promote the spread of alpha synuclein to other cells. In addition to disrupted lipid homeostasis and autophagy, there are several potential mechanisms by which GBA deficiency can impact cell function (Figure 4, left panel), including disrupting CMA (described in a later section), and inducing ER stress through misfolded GBA mutants.⁹⁹

Figure 4. Mechanisms by which glucocerebrosidase deficiency could promote alpha synuclein pathology and Parkinson’s Disease.

Left Panel. Mutations in glucocerebrosidase A (GBA) and other sphingolipidase enzymes predispose to juvenile-onset sphingolipidoses as well as age-associated Parkinson’s Disease with a-synuclein accumulation. The mechanisms by which this occur could be multifold. 1) GBA is translated in the endoplasmic reticulum. Improperly folded mutant GBA can thus cause ER stress. 2) GBA deficiency reduces autophagic flux and lysosomal function, resulting in accumulation of alpha synuclein and other proteins, and exocytosis to the extra cellular space. 3) Mutant GBA binds cytosolic LAMP2A and blocks CMA-mediated degradation of alpha-synuclein. 4) Impaired sphingolipid processing in the lysosome by GBA results in increased GlcCer, which is effluxed to the cytosol and promotes aggregation of alpha-synuclein. Right Panel: Sphingolipids are primarily broken down in the lysosome. Homozygous loss of function mutations in many of these genes lead to sphingolipid accumulation and neurodevelopmental disorders such as Gaucher, Krabbe or Niemann-Pick Disease. For reasons not yet entirely clear, when present in a single copy, these mutations increase risk of familial Parkinson’s Disease. However, the strength of the genetic evidence varies, as described in the text.

Though GBA was the first sphingolipidosis mutation identified to increase risk for PD, it is certainly not the last. SCARB2 has been identified in several PD risk genome-wide association studies (GWAS).^100–103 SCARB2 encodes lysosomal integral membrane protein 2 (LIMP-2), the transmembrane receptor responsible for trafficking GBA to the lysosome, thus its loss would be hypothesized to reduce GBA lysosomal presence and activity. Similarly, GALC encodes galactosylceramidase, and its homozygous loss causes in Krabbe disease. In addition to accumulating glycosphingolipids, infants with Krabbe disease develop alpha-synuclein inclusions that exhibit prion-like spread similar to the alpha-synuclein of PD.¹⁰⁴ GWAS identified the rs979812 locus in GALC as increasing PD risk.^105,106 SMPD1 encodes for acid-sphingomyelinase and has been identified in several PD risk GWAS and confirmatory studies; moreover, reduced acid-sphingomyelinase activity correlates with earlier age of onset of PD.¹⁰⁷

In contrast to the genes above, which have stronger links both genetically and mechanistically to PD, some PD risk genes, such as ASAH1 and GLB1, have only been identified in a single GWAS study of PD risk, with no mechanistic or confirmatory follow up.⁷ Three genes have even more controversial links to PD: ARSA, NPC1 and NPC2. Overall, the odds ratio for developing PD associated with GBA (OR 5.4) is much higher than the other genes described above, with the genes encoding proteins further away from GBA in sphingolipid processing having weaker and weaker evidence (Figure 4, right panel).^72,108

A link between sphingolipidosis and PD may not be just a result of molecular changes caused by reduced lysosomal function, but also an intersection with regional vulnerability. Specifically, microglia in the substantia nigra have a much higher lysosomal content compared to other basal ganglia nuclei.¹⁰⁹ Since sphingolipid-containing membranes are cleared by microglia, it is possible that this region has greater need for sphingolipid clearance and is thus more susceptible to sphingolipidosis-related mutations, though this hypothesis would need much more study.¹¹⁰

Neuronal Ceroid Lipofuscinosis and Intra-lysosomal Proteolysis

In contrast to GBA, which was recognized as a Gaucher’s Disease gene before it was linked to PD, GRN was first discovered as an adult-onset neurodegenerative disease gene.

Heterozygous progranulin (GRN) mutations are associated with FTD with TDP-43 Type A pathology.^79,111 GRN mutations exhibit full penetrance for disease by age 80.^112,113 Subsequently, rare families with biallelic loss of function mutations in GRN were identified, and affected individuals were found to have a juvenile-onset neuronal ceroid lipofuscinosis (NCL).^114–116 It was additionally observed that mono-allelic GRN mutation carriers have lipofuscinosis accumulation similar to the NCLs, suggesting overlapping disease processes.¹¹⁷ The genes associated with NCLs are typically in proteins involved in proteolysis (such as the cathepsin proteases), lipid catabolism, or have unidentified roles in the lysosome (e.g., MSFD8).¹¹⁸ Currently, 14 NCLs have been described, each with their own identified lipo-protein ultrastructural aggregate and causative gene mutation.¹¹⁹ In addition to GRN, a subset of these NCL genes have been associated with neurodegenerative disease, though these links are less strong than GRN (Table 1).¹²⁰

Most NCLs are caused by mutations in genes that encode lysosomal proteases that are collectively referred to as cathepsins. In humans, 15 cathepsins are divided into three families, depending on the active site amino acid: aspartyl (cathepsin D and E), cysteine (B, C, F, H, K, L, O, S, V, X, and W), and serine (A and G) cathepsins.¹²¹ Mutations in several of the cathepsins lead to NCL (D and F), and others have also been identified as risk factors for neurodegenerative disease (B, D, and F).^88,122,123 Even within the three cathepsin families, cathepsins each have preferred cleavage sites which are apparent in how they differentially cleave neurodegenerative disease proteins such as tau, TDP-43, and alpha-synuclein.¹²⁴ Thus, reduced activity of a specific cathepsin may result in the preferential accumulation of some proteins over others. As an example, cathepsin D knockout mice accumulate Ab in addition to lipofuscin, and Alzheimer’s pathology has been observed in postmortem tissues from several NCL patients.^125–127These differences in specificity of NCL-associated proteins may explain why there is a wider variety of neurodegenerative diseases linked to NCL genes (Table 1).

Within the lysosome, cathepsins are modulated by a variety of mechanisms. Most cathepsins operate optimally at the acidic pH of the lysosome. Thus, mutations or disease states that alter lysosomal pH impact the ability of the lysosome to catabolize proteins.¹²⁷ For example, loss of function of PS1, mutations of which cause early-onset AD, alkalinize lysosomes.¹²⁸ Similarly, loss of function of GRN alkalinizes lysosomes, while PD-associated mutations in TMEM175, a lysosomal proton leak channels, hyper-acidifies lysosomes.^129,130 Further, other lysosomal proteins, such as serpins and cystatins, can act as endogenous protease inhibitors.¹²¹ Progranulin, in fact, is one of these cathepsin-modifying proteins. Encoded by the GRN gene, progranulin is composed of 7.5 tandem granulin peptides, each with a non-identical sequence but shared structure of twelve-cysteine motifs. Progranulin is secreted into the extracellular space and then taken up into cells via the sortilin receptor for trafficking to the lysosome, where it is cleaved by lysosomal cathepsins into individual granulins or multi-granulin fragments.¹³¹ In vitro, progranulin promotes the autoactivation of the proprotein of cathepsin D into the mature form, thus enhancing its activity.¹³² In contrast, individual granulins seemingly inhibit CTSD activity.¹³³

In cancer, progranulin and the granulins have opposing effects on inflammation; however, these opposing roles are not as well understood in the brain.¹³⁴ In C. elegans, granulins accumulate with age, and multi-granulin fragments accumulate in diseased brain regions from patients with GRN-FTD and AD.¹³³ Alternatively, individual, specific granulins (F and A) have been shown to promote lysosomal function and rescue disease phenotypes in GRN knockout mice¹³⁵ Thus, progranulin heterozygosity may reflect two mechanisms of disease: both loss of full-length progranulin and relative accumulation of deleterious subsets of granulins and multigranulin fragments. However, more work is needed to determine the contribution of granulins and multi-granulin fragments to neurodegeneration and the mechanisms involved.

In contrast to most of the NCL genes, which appear to disrupt protein homeostasis, NCL5 was very recently identified to encode the bis(monacylglycero)phosphate (BMP) synthase.¹³⁶ BMP is an anionic, late-endosome/lysosome-specific glycerophospholipid that is reduced in several neurodegenerative disorders, including GRN-FTD. BMP stimulates lysosomal lipases and promotes cholesterol egress into the cytosol.¹³⁷ BMP levels are reduced in the brains of GRN KO mice and the CSF of patients with GRN-FTD. Conversely, glucosylsphingosine, the substrate of GBA, accumulates in GRN KO mice, and GBA activity is reduced in these animals as well.¹³⁰ Thus, the clear intersection between GRN and GBA1-mediated pathologies could contribute to the co-pathology observed in these diseases.

Variations on autophagy: Lysosomal import is required for lysosomal degradation, a tau-specific role for TSC1

Some genes not linked to LSDs, but rather to other neurodevelopmental disorders, have been found to have new roles in lysosomal dysfunction in adult neurodegenerative diseases. For example, mutations in the TSC1 gene have for many years been linked to tuberous sclerosis complex (TSC), which is characterized by pathognomonic skin findings, seizures, cognitive disability, and the accumulation of a variety of tumors.¹³⁸ TSC is caused by heterozygous loss of function mutations in TSC1 or TSC2 with a later “loss of heterozygosity” that partially explains its variable phenotype. Tubers isolated from infants with TSC1 mutations demonstrate hyperphosphorylated tau and thus it is recognized as a “juvenile tauopathy.”¹³⁹ In recent years, we showed that TSC1 mutations also increase the risk of FTD with tau accumulation.^90,140 Surprisingly, these individuals accumulate tau inclusions with abnormal acetylation distinct from the tau inclusions observed in AD and FTLD.^90,140 The mechanistic basis for this involves the role of the TSC1/hamartin protein in autophagy.

Autophagy, or the process of “self-digestion” in which cells degrade their own macromolecules, has been sub-divided into macroautophagy, e-microautophagy and chaperone-mediated autophagy (CMA) (for review, see¹⁴¹). In CMA, proteins containing a KFERQ-like degradation signal are recognized by heat shock cognate 71 kDA protein (Hsc70) and directly transported into the lysosome through the transmembrane channel LAMP2A.¹⁴² Almost 50% of the proteins in the mammalian proteome feature a KFERQ-like motif, including all the major aggregating proteins of neurodegenerative diseases (i.e., TDP-43, alpha-synuclein, APP, and tau).^99,143–145

Mechanistically, TSC1 haploinsufficiency results in overactivation of mTORC1, which in turn activates p300 histone acetylase (HAT) and inhibits sirtuin 1 (SIRT1), a histone deacetylase. The resulting imbalance promotes tau acetylation at multiple sites.⁹⁰ Tau acetylation contributes to tau pathology via multiple potential mechanisms: promoting tau fibrillation¹⁴⁶ and impairing tau clearance through CMA both by reducing entry into the lysosome and proteolysis by the lysosome.¹⁴⁷ The resultant failure to clear acetylated tau leads to a novel TSC tauopathy.¹⁴⁸ Reducing tau acetylation restores its degradation through CMA (Figure 5).^90,147 Of note, inhibiting mTORC1 does not fully rescue tau degradation in TSC1 haploinsufficient iNeurons, suggesting that there are also mTORC1-independent mechanisms of tau accumulation in TSC1-mediated tauopathy.⁹⁰

Figure 5. Specialized autophagy mechanisms traffic proteins to the lysosome.

In chaperone-mediated autophagy, proteins containing a KFERQ sequence bind a chaperone complex that includes Hsc70 and other co-chaperones to transport proteins through LAMP2A channels. TSC1 haploinsufficiency results in imbalance between p300 acetylase and SIRT1 deacetylase, resulting in acetylation of tau. Acetylated tau is more prone to aggregation and cannot traffic through chaperone-mediated autophagy (CMA) for degradation.

Failure of CMA has been linked to neurogenetic and neurodegenerative disorders through at least two different mechanisms. First, CMA is diminished with age, thus genetic risk alleles may combine with aging to lead to late-onset neurodegeneration.⁶¹ Second, it was recently shown that mutant glucocerebrosidase A (the protein encoded by GBA1) can be mis-trafficked to the lysosomal surface instead of the lysosomal lumen. There, it binds to and interferes with normal LAMP2A function, impeding processing of CMA client proteins, particularly alpha-synuclein.⁹⁹ At least five other genes that are associated with PD (but not LSDs) also affect CMA: LRRK2, UCHL2, VCP35, DJ-1, and SNCA, which encodes alpha-synuclein.^149–153

Neuronal susceptibility to lysosomal dysfunction

For reasons that are not yet entirely clear, neurons are particularly susceptible to lysosomal dysfunction. Clinically, this is evidenced by the fact that most lysosomal storage disorders feature neurological symptoms such as vision impairment and seizures early in disease progression. At a mechanistic level, the long processes of axons and dendrites mean that neurons rely heavily on local proteostasis that requires robust lysosomal trafficking and positioning.^154,155 Moreover, neurons are terminally differentiated postmitotic cells. Thus, aging neurons 1) lack the diluting power of cell division for neurons, 2) require the maintenance of unusually high lipid content, 3) over time are exposed repeatedly to multiple physiological stressors (e.g. oxidative stress, fever, head injury), 4) must remove damaged organelles such as mitochondria, and 5) experience declining efficiency of all catabolic machinery.

There is a set of diseases that do not fit directly within our paired pediatric and adult framework; however, they highlight the susceptibility of neurons to lysosomal dysfunction and merit dedicated discussion here. Lysosomal acidification is maintained by the vATPase, a proton pump responsible for acidifying the lysosome, and loss of vATPase would be expected to impair processing of numerous substrates. Mutations in ATP6V1A, which encodes a subunit of the vATPase, were identified in patients with encephalopathies as well as moderate intellectual disability and epilepsy.¹⁵⁶ Mutations in ATP6AP2, which endodes another vATPase subunit, cause an X-linked adult-onset Parkinsonism or fulminant postnatal neurodegeneration.^157,158 Broadly, the symptoms of these disease are limited to the neuron, highlighting the reliance of neurons on lysosomal function, while other cell types are able to compensate. Interestingly, ATP6V1A was identified as a key nodal regulator in late-onset Alzheimer’s disease in a multi-omic analysis.¹⁵⁹

Neurons possess a unique structure, with up to 75% of cellular content in long dendrites and axons that can stretch meters from the cell body. To accommodate this structure, autophagy occurs at synaptic terminals and in the distal dendrite, and autolysosomes mature during retrograde trafficking.^{160,161 162} Additionally, in neurons, ribosomes and mRNA hitch a ride along lysosomes for anterograde trafficking.¹⁶³ Lysosomes in neurons versus dividing cells are hypothesized to be different; however, detailed and quantitative comparisons have yet to be performed and are underway.^161,162 Contrary to the idea that neurons are more susceptible to protein stress, it may be that they are, in fact, more resilient. In our studies, we have found that iNeurons are more resistant to the proteostatic stress induced by heavy metal toxicity.¹⁶⁴ Since their identity within neuronal circuits and systems means they can only be turned over at the molecular (rather than cellular) level, neurons may rely more heavily on proteostasis management and, therefore, have more robust systems for dealing with damaged proteins. Newer mechanisms to manage proteostasis, such as exopher formation and extracellular proteostasis, have been discovered in recent years.^165,166 At least for the former, neurons appear to be the prime cell type utilizing this type of stress management. An alternative is that neurons have more lysosomes so as they fail with mutations or age, neurons are particularly burdened, though this has yet to be quantified. Finally, it is not known if neurons in different brain regions or different CNS cell sub-types (e.g., astrocytes, microglia) exhibit variations in lysosome constituency. Future work should address these issues.

Conclusion

The link between neurodevelopmental and neurodegenerative disorders offers an opportunity to use genetic lysosomal diseases of childhood to inform our understanding of neurodegeneration, and vice versa. It also reveals a view of lysosomal dysfunction on a continuum across development and aging, two life epochs of increased proteostatic stress that can lead to lysosomal failure in those with genetic susceptibility. Because the shared disease genes impair lysosomal catabolic function, we propose that they collectively be referred to as lysosomal clearance disorders. This terminology more accurately reflects disease biology since “storage” suggests that cells are actively hoarding abnormal macromolecules rather than the reality that they are struggling (and failing) to clear them. In addition, the term lysosomal clearance disorders would form an umbrella term inclusive of both neurodevelopmental and neurodegenerative diseases, which will promote the recognition that they are genetically and mechanistically related through the lysosome. Our understanding of the lysosome has come a long way since its humble origins as the trash can of the cell. The crucial role of this acidic organelle over the dynamic proteostatic landscape of a lifespan explains the phenotypic variability seen when lysosomal genes are mutated in a dose-dependent manner. The shared genetic and mechanistic pathways truly highlight the importance of the lysosome in the maintenance of brain health. A holistic understanding of these juvenile vs age-associated lysosomal clearance disorders could broadly impact our understanding and ultimate treatment of this class of diseases.