Developmental Alterations in Adult-Onset Neurodegenerative Disorders: Lessons from Polyglutamine Diseases

A number of neurodegenerative diseases, including Parkinson’s disease, Alzheimer’s disease, and amyotrophic lateral sclerosis, show clinical symptoms late in life. Given the delayed onset, it has long been a question as to when pathogenic events begin. Do cellular derangements begin late in life as do behavioral symptoms, or do they begin much earlier and go unnoticed because they do not cause signs or symptoms?

Addressing this issue is extremely important because without a detailed knowledge of the concatenation of events from gene products to cellular deficits, one has only a partial picture of pathogenesis. One cannot identify how and when to intervene; this clearly has translational implications when it comes to developing optimal treatment strategies.

In this regard, the polyglutamine diseases—those caused by a CAG trinucleotide expansion in the coding region of the disease gene—have been especially illuminating. Huntington’s disease (HD) is the most common polyglutamine disease and results in a movement disorder characterized by chorea stemming from degeneration in the basal ganglia. The other polyglutamine diseases are less prevalent, but they are all neurodegenerative. Most cause incoordination or ataxia from cerebellar degeneration and fall under the broader rubric of spinocerebellar ataxias (SCAs). Spinobulbar muscular atrophy is an exception in that it is primarily a neuromuscular syndrome.

With large multicenter natural history studies, we are learning that subtle pathology does in fact begin before the earliest clinical signs and symptoms of the disease. In HD, at-risk individuals show reductions in brain volume, including specific regions of the basal ganglia, and cerebellar-striatal circuitry show abnormalities much earlier, even in childhood.^1–3 These abnormalities are discerned structurally by magnetic resonance imaging, but magnetic resonance spectroscopy shows metabolic deficits as well.^1,4,5 In the SCAs, such in-depth pediatric studies have yet to be performed, but presymptomatic adults with the mutation exhibit structural and metabolic derangements reminiscent of that seen in HD.^5–7 These changes, it is to be emphasized, are seen in the typical mutations seen in diseases of adult onset, and not from an exaggerated repeat expansion known to cause juvenile versions of the respective polyglutamine disorders.

Parsing out further details about disease onset is currently not possible in patients. Even the most sensitive modalities have their limitations, whereas brain biopsy, the ideal standard for pathological analysis, is clearly not an option. Fortunately, we can turn to animal models for further insights. There are now several well-characterized mouse models for all the polyglutamine disorders.^7–9 These mice have been engineered to express mutant versions of the corresponding polyglutamine protein and remarkably, much like human patients, mice are born healthy only to display neurodegeneration when they reach adulthood. They also are precise in that the same substructures and neuronal subtypes are affected as observed in humans. Mouse models indeed were the first to demonstrate that polyglutamine diseases are disorders of protein homeostasis, with the mutant proteins tending to aggregate most visibly into nuclear inclusions. These aggregates precede the behavioral phenotype, and neuritic and synaptic pathologies can also be seen in the presymptomatic period.⁸ Neurons are also not the only cells of the nervous system that are affected. Astrocytic, microglial, and oligodendrocytic pathologies also occur early, which suggest a role for glial pathology as well.^10,11 In fact, changes in gene expression can be observed in early postnatal life. Some of these events at the levels of proteins and gene expression may well be compensatory,^12,13 but pathogenic processes are definitely at play given that delaying the expression of mutant polyglutamine proteins or rescuing early transcriptional deficits prevents later degeneration.^8,14 In addition, restricting the expression of mutant proteins to early postnatal development still causes significant disease burden in mice, suggesting that these early derangements are indeed crucial for later pathogenesis.¹⁵

Several recent studies suggest, however, that the earliest pathology not only strikes neurons and glia early in life but also neuronal progenitors that give rise to these cells.^16,17 The first from our own group was inspired by the finding that ataxin-1 (ATXN1), the protein mutated in SCA type 1 (SCA1), is expressed in stem cells niches. Curious to test for any derangements in stem cells in the context of the cerebellum, we focused on a relatively understudied progenitor population in the developing cerebellum—those defined by the neuronal stem cell marker Prominin-1 (CD133). We found that mutant ATXN1 causes these stem cells to proliferate in an exaggerated manner with a bias toward a neuronal lineage. The mechanism by which ATXN1 regulates stem cell proliferation is still unknown, but ATXN1 already has well-described properties, suggesting that it regulates gene expression. It binds the promotor region of cyclin D1 and activates its transcription, and it also modulates the RAS-Mitogen-activated protein kinase (RAS-MAPK) pathway critical for cell proliferation.¹⁸ It also forms a complex with its close homologue ATXN1L and a vast array of transcriptional repressors, activators, and additional modulators (eg, Capicua, Transcription Factor 1 (E4F), nuclear receptor corepressor and silencing mediator of retinoic acid and thyroid hormone receptor (NcoR-SMRT), RAR-related orphan receptor α, Tat-interactive protein 60 (Tip60), ANP32, Leucine-rich Acidic Nuclear Protein (LANP), and Yes-associated protein (YAP) deltaC^14,19–23). Mutant ATXN1 interferes or exaggerates the normal function of ATXN1 based on whether it recruits or robs these factors from their normal targets.^22,24 Expanded ATXN1 also misregulates the levels of growth factors such as vascular endothelial growth factor and brain-derived neurotrophic factor.^25–27 It is possible that 1 or more of these functions—or indeed others—are relevant to the stem cell fate. The overall result is an increase in the neurons that derive from these stem cells, GABAergic interneurons that comprise basket cells and stellate cells.¹⁶ This largely occurs by a cell autonomous effect, although one cannot exclude cell–cell interactions.¹⁶ Regardless, the exaggerated number of GABAergic interneurons would be expected to inhibit Purkinje neuron firing, and this indeed was experimentally observed by electrophysiological analysis. This excessive inhibition will likely have long-term implications (Fig. 1A). It could result in a constant inhibitory drag on Purkinje cell output, which in turn would stress the excitatory synapses from other inputs, such as the climbing fibers. Conceivably this could be a cause for excitotoxic damage at the dendrites of Purkinje neurons or metabolic derangements that have been observed.¹⁶

FIG. 1.

Neurodevelopmental abnormalities in polyglutamine disorders. (A) In spinocerebellar ataxia type 1 (SCA1), mutant ATXN1 ataxin-1 (mATXN1) exaggerates the proliferation of postnatal stem cells in the white matter (WM) of the developing cerebellum. These stem cells differentiate in a biased fashion toward γ–aminobutyric acid-ergic (GABAergic) inhibitory interneurons in the molecular layer (ML) and away from a glial lineage (astrocytes). This in turn causes increased inhibition of Purkinje cells (PCs) in the Purkinje cell layer (PCL). (B) In Huntington’s disease (HD), mutant huntingtin (mHTT) mis-localizes to impair the self-renewal capacity of apical progenitors at the ventricular zone of the embryonic cortex. These progenitors are also biased toward developing into a neuronal linage. [Color figure can be viewed at wileyonlinelibrary.com]

Another study by Barnat and colleagues¹⁷ pertains to HD and was prompted by the availability of aborted human fetuses with the HD mutation (13 weeks of gestational age). In these fetuses, huntingtin (HTT) within the apical progenitor cells in the developing ventricular zone was not diffuse as it usually stains, but was largely confined to the apical surface.¹⁷ This mislocalization was confirmed in HD knock-in mice. Moreover, this HTT mislocalization is associated with a disruption of neuroepithelial junctional complexes; this in turn leads to a disruption in the nuclear migration within the apical progenitor cells responsible for their division and differentiation. These events lead to an increased neuronal lineage at the expense of cycling progenitors (Fig. 1B). As with ATXN1, the mechanisms involved have not been entirely solved, but once again there are clues. As with ATXN1, HTT appears to modulate gene expression by its interactions with its binding proteins (eg, transcription activators such as specificity protein 1 (SP1), neurogenic differentiation factor 1 (NEUROD1), and transcriptional repressors such as RE1-silencing transcription factor (REST) and transcription factor II D).^28,29 HTT has other developmental roles as well. It regulates cellular adhesion, polarity, and epithelial organization, and mutant HTT accelerates the process of epithelial-mesenchymal transition where epithelial cells lose their cell polarity and cell–cell adhesion to become progenitor stem cells.^30,31 Mutant HTT also interferes with mitotic spindle orientation through the aberrant recruitment of proteins of the dynein/dynactin complex to the spindle pole.^32–34 It is also intriguing that induced pluripotent stem cells derived from patients with HD show the changes in genes that are involved in neuronal development and maturation.³⁵ By some fascinating convergence of pathological events, in both of these diseases there is a smaller pool of progenitors that are also biased toward a neuronal lineage.^17,36

Unlike SCA1, the consequences of these early changes to the neuronal circuitry in later life in HD have yet to be deciphered, but it is likely that analogous defects to SCA1 will arise in alterations in neuronal number and fate, albeit in the basal ganglia as opposed to the cerebellum. It is also intriguing that in both of these diseases, the aberrant proliferation of neurons is associated with a developmental depletion of the astrocytic pool.^16,36 This would have deleterious consequences given that astrocytes are well known to support neuronal function through their effects on neuronal metabolism or other functions; for instance, astrocytes also alter the pruning of synapses, even influencing the electrophysiological properties of neurons and neurotransmitter release. Recent studies also show deficits in the oligodendroglia lineage as well, which would alter later myelinogenesis.^36,37

It would be important now to decipher which aspects of these early intracellular and circuit events are causal to later pathogenesis because it has implications for treatment. For instance, targeting mutant proteins, which is currently being explored using RNA-reducing strategies, might have to be considered even earlier in the disease course.³⁸ Some developmental pathology might be too late to target, but in the case of SCAs at least it might be possible to rectify some defects given that aspects of cerebellar development occur well into adolescence.³⁹ In the case where developmental processes cannot be reversed, circuit dysfunction could be treated by the judicious manipulation of neurotransmitter pathways. For instance, a downregulation of GABAergic inhibition could be a treatment strategy in the SCAs as it would hamper the inhibition of interneurons on Purkinje cell firing.

Even beyond their translational impact, these findings have fascinating ramifications from evolutionary and developmental standpoints. Neurodegenerative diseases have typically been considered diseases of aging. In fact, scientists in the premolecular medicine era, including such stalwarts as Peter Medawar and J.B.S. Haldane, had speculated that mutations in late-onset diseases such as HD are prevalent because they are not selected against in the course of evolution because they do not impact early life and hence do not preclude presymptomatic individuals from having offspring.⁴⁰ Our findings suggest that these mutations in fact do impact the early organism, except that they do not interfere with fitness until well past the child-bearing years. It is likely that the delay in the expression of the disease has been sculpted by natural selection so as to delay onset. Another exciting implication of these findings is that selective vulnerability of select brain regions might stem from these early developmental changes as they get further stressed by aging pathways. These intersections between developmental and senescence pathways is ripe for further research. Neurodegenerative diseases should not just be seen as diseases of aging, but rather as diseases whose trajectory is presaged early in development.