When function follows form: Nuclear compartment structure and the epigenetic landscape of the aging neuron

Abstract

The human brain is heavily affected by cellular aging. Neurons are primarily generated during embryogenesis and early life with a limited capacity for renewal and replacement, making them some of the oldest cells in the human body. Our present understanding of neurodegenerative diseases points towards advanced neuronal age as a prerequisite for the development of these disorders. While significant progress has been made in understanding the relationship between aging and neurological disease, it will be essential to delve further into the molecular mechanisms of neuronal aging in order to develop therapeutic interventions targeting age-related brain dysfunction. In this mini review, we highlight recent findings on the relationship between the aging of nuclear structures and changes in the epigenetic landscape during neuronal aging and disease.

Introduction

Age is the ultimate risk factor for the development of cognitive decline and neurodegenerative diseases including Parkinson’s disease (PD), Alzheimer’s disease (AD) and other dementias (Hou et al., 2019). In the central nervous system, aging is associated with a decrease in synaptic plasticity accompanied by an increase in inflammatory processes (Lynch, 2010; Mosher and Wyss-Coray, 2014). It is well-established that epigenetic factors play a fundamental role in the regulation of synaptic plasticity and cognition (Day and Sweatt, 2011; Dillman et al., 2017), yet the molecular mechanisms that lead to epigenetic and phenotypic changes during aging are still poorly understood. Amongst the myriad of cellular features that characterize the aging process, epigenetic alterations along with genomic instability, telomere attrition, and defective proteostasis have long been considered the primary hallmarks of aging (López-Otín et al., 2013; Rando and Chang, 2012; Sen et al., 2016; Teschendorff et al., 2013). Recent studies have shed light on how epigenetic factors in aged cells interact with cellular compartments such as the nuclear envelope and nuclear pore complex (NPC). Analyses of human postmortem brain specimens revealed that cell populations in the brain accumulate age-related modifications across their epigenetic landscapes—these age-related changes are associated with transcriptional changes, and ultimately, cellular dysfunction (Figure 1). Animal models have further bolstered our understanding of how long-lived protein structures are targeted by age-related damage that could influence many downstream features of aging in the brain (D’Angelo et al., 2009; Mertens et al., 2015). In particular, the nuclear envelope and NPC consist of many extremely long-lived proteins (Toyama et al., 2013), and these structures have emerged as key factors controlling the spatial organization of the genome and transcriptional activity, pointing toward a potential role as a key factor in aging (T. Cremer and C. Cremer, 2001; D’Angelo, 2018; Ibarra et al., 2016; Sood and Brickner, 2014; Toda et al., 2017). Recent advances in the field of cellular reprogramming have allowed for the use of human neural cell models which further narrow the gap between animal models and postmortem studies. Taken together, these studies have confirmed and expanded our understanding of the importance of the nuclear periphery in controlling changes in the aging human epigenome.

Figure 1. Epigenetic alterations and nuclear leakiness contribute to neuronal aging.

Aging in concert with intrinsic, genetic disposition and environmental factors result in changes in the nuclear compartment of neurons over time. Histone loss, increased open chromatin, altered histone modifications, and changes in DNA methylation (DNAm) pattern lead to alterations in the gene expression profile of aged neurons and eventually to neuronal dysfunction. Impairment of nuclear pores (nuclear pore ‘leakiness’) contributes to age-related changes nucleus and cytoplasm configuration.

The basics of nuclear architecture and epigenetic landscape

Nuclear architecture

The nuclear architecture plays an important role in orchestrating the cell’s identity and phenotype by modulating gene expression. Based on its spatial density, nuclear DNA and associated proteins can be subdivided into heterochromatin and euchromatin. Heterochromatin contains condensed DNA and is found predominantly at the periphery of the nucleus, making it inaccessible to the transcriptional machinery and thereby transcriptionally inactive. In contrast, euchromatin is loosely packed and in general permits active transcription. A more detailed insight into the global organization of chromatin has been achieved through the development of chromatin confirmation technologies such as Hi-C (Lieberman-Aiden et al., 2009). Based on Hi-C, genomes can be subdivided into A and B compartments, whereby compartment A delineates predominantly transcriptionally active and compartment B mainly inactive genes (Dekker and Mirny, 2016). Regions of the genome that interact with each other are often referred to as topologically associated domains (TADs) (Dixon et al., 2012; London et al., 1978). TADs are often flanked by CTCF binding sites. CTCF is a transcription factor important for establishing the chromatin architecture, recruiting cell type-specific transcriptional activators, repressors, cohesins, and RNA polymerase II. The interaction between promoters and enhancers can be assessed by methods such as PLAC-seq or HiChIP, which combine Hi-C with chromatin-immunoprecipitation followed by sequencing (ChIP-seq) (Fang et al., 2016; Mumbach et al., 2016). In general, the majority of observed interactions between promoters and enhancers do not cross TAD boundaries. Repressed TADs are found at the periphery of the nucleus and a high overlap of repressed TADs exists with so called lamina-associating domains (LADs). LADs are regions that interact with the nuclear lamina, are flanked by CTCF binding sites at their borders, and contain only a few genes (van Steensel and Belmont, 2017). LADs are characterized by the presence of certain histone modifications such as the repressive marks H3K27me3 and H3K9me3 (see below). Genome-wide sequencing technologies, including multiomics single cell analysis, have advanced greatly over the last decade, allowing for the creation of atlases of histone modifications and changes in DNA methylation across multiple cell types. This advance in technology has vastly increased our knowledge of the epigenetic changes in the aging human brain (Lee et al., 2019). Since it is difficult to address functional consequences of epigenetic aging on both the molecular and behavioral levels in the human brain, animal models have been utilized to uncover fundamental roles of epigenetic regulation in synaptic plasticity and cognition.

Histone modifications

Epigenetic mechanisms governing the expression of genes related to brain function include the posttranslational modification of histone proteins. (Benayoun et al., 2019; Yankner et al., 2008). The N-terminal tails of histones are modified by marks including acetylation and methylation; these changes affect the accessibility of chromatin for transcription factors and other chromatin-associated proteins. Histone acetylation is associated with open chromatin structure, a configuration which allows for increased gene activation. For example, acetylation of lysine residue 27 of the histone protein 3 (H3K27ac) is a mark for active enhancers and promoters of active genes. Acetylation of histone marks such as H3K27ac and H4K16ac are mediated by histone acetyltransferases, whereas histone deacetylases (HDACs) remove acetyl groups. Histone methylation, another category of histone modification, is characterized by the addition of up to three methyl groups to lysine residues; this process is mediated by methyl transferases. Certain types of histone methylation are associated with either gene activation or repression. For example, H3K4me3 and H3K27me3 mark active and silenced gene promoters, respectively, while H3K36me3 marks active gene bodies (Barski et al., 2007). Changes in the pattern of histone modifications have been associated with brain function and aging. SIRT1, a regulatory protein associated with aging, plays a role in age-dependent changes in gene regulation through its deacetylase activity (Oberdoerffer et al., 2008) and reduced histone acetylation is accompanied by impaired memory in aged mice (Benayoun et al., 2019; Peleg et al., 2010). Physiological aging and AD-pathology increase the expression levels of histone deacetylase 2 (HDAC2) and inhibit activity-dependent gene regulation and memory (Gräff et al., 2012; P. Singh and Thakur, 2018). Inhibition of HDAC activity has been reported to reverse age-dependent memory impairment (Gräff et al., 2012; Peleg et al., 2010), but conflicting results have also been reported (Dagnas et al., 2013). Further studies will be needed to follow up on this discrepancy and to elucidate the role of HDAC activity in age-related changes in the brain. Histone methylations that are associated with gene repression (e.g. H3K9me2, H3K9me3, H3K27me3), are also influenced by aging. In general, the loss of histones as well as that of repressive histone modifications were postulated as mechanisms of cellular aging, which compromise the tight orchestrated control of gene regulation. Furthermore, recent studies have shown that AD leads to the loss of repressive histone modifications that cause chromatin relaxation and genome instability (Moerkens and Diesfeldt, 1987; Sun et al., 2018). However, it is not still clear how physiological aging alters the landscape of repressive histone methylation. Of note, an increase of histone methylation associated with gene repression has been observed during aging (Snigdha et al., 2016; Wang et al., 2010). The changes of histone methylations could be sites-dependent or context dependent, and further investigations are required to obtain a comprehensive understanding of age-related histone modification changes and their effects.

DNA methylation

Direct DNA methylation (DNAm) of cytosine residues at the carbon 5 position (5mC) is another mechanism by which gene expression is mediated. This modification is controlled by the DNA methyltransferases DNMT1, DNMT3A, and DNMT3B. Increased methylation of CpG islands in a promoter correlates with and partially mediates the repression of the corresponding gene. DNAm was thought to be a mitotically inherited modification which dictates cell identity at enhancer elements (Song et al., 2019). However, recent studies demonstrated that DNAm can be dramatically modified by environmental stimuli or behavioral experience (Ma et al., 2009; C. A. Miller and Sweatt, 2007). It is now known that DNAm plays a critical role in neural plasticity and memory (Martinowich et al., 2003; C. A. Miller et al., 2010), and is altered during aging, where it is thought to underlie age-dependent memory decline (Liu et al., 2011; Oliveira et al., 2012; Penner et al., 2016; 2011). DNA demethylation is further mediated by the ten-eleven translocation (TET) family via several oxidation or deamination reactions. TET2 oxidizes 5mC, which results in the production of 5-hydroxymethylcytosine (5-hmC) from 5mC. Although the role of 5-hmc is still unclear, one study showed that this pathway is heavily involved in methylation-mediated, age-related changes in murine adult neurogenesis (Gontier et al., 2018). Adult human neurons show high levels of 5-hmC (Kriaucionis and Heintz, 2009), and 5-hmC is enriched around synaptic genes (Khare et al., 2012). Interestingly, non-dividing, postmitotic neurons in the mature human brain show an increased degree of DNAm at non-CpG sites compared to glia (Lister et al., 2013), and the 5-hmC modification is highly dynamic during human brain development (Spiers et al., 2017). Thus, it is likely that 5-hmC mediates gene regulation and aging might influence the degree of 5-hmC.

Epigenetic aging signatures in the human brain

Advances in next-generation sequencing which include RNA-seq, Bisulfite-sequencing and ChIP-seq have enabled genome-wide assessment of gene expression, DNAm, and histone modifications, respectively, during aging and neurodegenerative disease states. Gathering data from postmortem brain samples remains technically difficult due to factors such as postmortem interval, pH, duration of pre-mortem agony, and cause of death; each of these issues affects tissue quality. Furthermore, natural genetic variation, environmental factors, comorbidities, medication and a plethora of other social and clinical features complicate the interpretation of post-mortem studies. Despite these challenges, several studies focusing on human brain tissue were published in recent years. Notably, these studies utilize bulk analysis of brain tissue and do not characterize cell-type-specific changes in the epigenetic landscape.

Performing H3K27ac and H3K4me3 ChIP-seq, which label active chromatin regions and promoters, respectively, we generated a cell type-specific enhancer and promoter of the human brain (Nott et al., 2019). We found a strong enrichment of heritability of variants for psychiatric disorders such as Schizophrenia, major depressive disorders, and autism spectrum disorders. AD DNA variant heritability was mostly enriched in microglia-regulatory elements. We also generated a promoter-enhancer interactome map using PLAC-seq, thereby enabling interpretation of intergenic risk alleles with neurological and psychiatric disorders. However, since tissue from pediatric cases was used, data on cell type-specific chromatin regulatory regions and interactome map derived from aged or individuals with AD is currently lacking.

One study collected postmortem brain tissue from young, old, and AD donors and utilized ChIP-seq to identify genome-wide changes of H4K16ac histone marks (Nativio et al., 2018a). H4K16ac levels increased and redistributed during aging in genes associated with neuroplasticity and immunity. Interestingly, binding sites for REST, a transcriptional repressor, were enriched in genes that lost H4K16ac with age. This resulted in reduced levels of FOXO, which mediates oxidative stress resistance, and the anti-apoptotic protein Bcl-2; PSEN2 and pro-apoptotic proteins like BAX and BID are increased (T. Lu et al., 2014). Variation in genome-wide patterns of H3K27ac was identified in the entorhinal cortex between 24 AD and 23 age-matched control cases (Marzi et al., 2018). In AD, H3K27ac was enriched in genomic regions implicated in amyloid and tau pathology (APP, PSEN1 and 2, MAPT). Another study performing H3K9ac ChIP-seq on 669 aged human prefrontal cortices suggested that changes in acetylation of H3K9, which marks active gene promoters, correlate with tau pathology (Klein et al., 2019).

Recent advances in neuroimaging technology could potentially enable studies of longitudinal alterations in chromatin state. Increased levels of HDACs were consistently detected in postmortem brains of elderly donors and associated to neurodegeneration (Fischer et al., 2010; Stilling and Fischer, 2011). A previous study focusing on postmortem samples of AD patients detected an increase of HDAC2 expression levels (Gräff et al., 2012). Recently, relative HDAC expression in the brain was measured in 41 healthy individuals with different ages using PET neuroimaging (Gilbert et al., 2019). With higher age, relative HDAC expression revealed by PET imaging was found to be increased in white matter. The development of novel neuroimaging tracers to monitor epigenetic changes could shed light on factors shaping the gene expression profile of cell-state specific changes in aging and neurodegeneration.

DNA methylation arrays, whole-genome bisulfite sequencing, and DNAm arrays have identified local and global changes in DNA methylation at various timepoints in brain tissue. Global levels of DNA methylation decrease during aging (Bjornsson et al., 2008; Boks et al., 2009). Fascinatingly, the CpG methylation status of a few hundred sites can accurately measure a person’s chronological age. This subset of genomic sites define the so-called ‘DNAm clock’ (Horvath, 2013). Depending on the selected DNAm sites, alternative ‘clocks’ have been developed, which were optimized to more accurately measure human chronological age, specifically call ages of certain tissues (e.g. blood and skin), or even to predict a person’s time to death (Levine et al., 2018; A. T. Lu et al., 2019; Zhang et al., 2019). Accelerated DNAm ages in the human prefrontal cortex are associated with cognitive decline and neuropathological features of AD (Levine et al., 2015). DNAm was also found to be dysregulated in brain tissues from donors with AD and PD (De Jager et al., 2014; McKinney et al., 2019). Signaling pathways that were implicated in changes in DNA methylation involved synapses and development, such as the TGFb, ErbB, Wnt and Hippo pathways (Sanchez-Mut et al., 2016). More studies will be needed in order to determine whether DNAm clock-associated methylation changes are solely a biomarker for aging, or whether they also represent a major driver of aging (Horvath and Raj, 2018).

One important caveat of studies using postmortem samples is the lack of information about cell-type specific changes of the epigenetic alterations. With the usage of fluorescence activated nuclei sorting (FANS), a method that has harnessed the sorting of nuclei based on cell-of-origin, identification of the epigenetic landscape of specific cell populations is now possible. For example, DNA methylation changes associated with age were observed when neuronal nuclei were sorted from 31 post-mortem brain samples using the neuronal marker NeuN (Gasparoni et al., 2018). Neuronal specific age-associated changes in DNA methylation were identified in genes like CLU, which encodes the protein Clusterin (Apolipoprotein J). Genetic variants in CLU have been identified as a genetic risk factor for AD (Lambert et al., 2013).

Linking the nuclear periphery and epigenetic aging in animal models

Recent years have provided a vast increase in our knowledge of the epigenetic landscape in human brain tissue. However, brain tissue is widely inaccessible and hampered by technical difficulties. Data derived from human samples has yielded limited information on the interface of epigenetic aging and age-dependent deterioration of the nuclear envelope and nuclear pore. An abnormal and irregular distribution of NPCs in the nuclear membrane was identified in post-mortem hippocampus tissue from age-matched controls and AD cases (Sheffield et al., 2006). This pattern was also observed in a recent study in postmortem samples from donors with AD in which mutated forms of Tau seemed to interfere with nucleocytoplasmic transport (Eftekharzadeh et al., 2018).

Regulation of cell-type- and cell-state-specific gene expression profiles depends on the genome’s spatial organization and nuclear architecture. The nuclear envelope and NPCs are crucial in facilitating the compartmentalization of macromolecules between cytoplasm and nucleus; they organize the nuclear architecture through interactions with transcription factors, chromatin regulatory proteins, and with the DNA itself. This precise patterning and organization allows for the establishment of specific gene expression signatures (Peric-Hupkes and van Steensel, 2010). Studying concurrent changes in the epigenetic landscape and nuclear architecture is technically challenging and data on the combination of these features during aging is currently limited. Animal models and the development of human stem cell/reprogramming-based models have provided the majority of our knowledge of nuclear architecture and chromatin landscape during aging (Table 1). Importantly, several features of epigenetic changes in physiological and pathological aging are conserved between mice and humans (Benayoun et al., 2019; Berson et al., 2018; Gjoneska et al., 2015; Pal and Tyler, 2016), suggesting that animal models will continue to provide useful information as the field progresses.

Table 1.

Model systems to study epigenetics and nuclear structure in brain aging.

	Humans	Post-mortem tissue	Animal models	iPSCs	iNs
Human genetics	✓	✓	✕	✓	✓
Human epigenetic age	✓	✓	✕	✕	✓
Human nuclear pore age	✓	✓	✕	✕	✓
Functional causality studies	✕	✕	✓	✓	✓
Genetic manipulation	✕	✕	✓	✓	✓
Epigenetic manipulation	✕	✕	✓	✓	✓
Lifespan experiments	✕	✕	✓	✕	✕
Behavior	✓	✕	✓	✕	✕

Human clinical studies and post-mortem analysis are limited to mostly descriptive studies. Animal models are accessible to manipulations and allow for functional studies in an in-vivo context, but lack human genetics and epigenetics, human nuclear pores, and have comparably low life spans. Human iPSCs and iNs both offer high accessibility, the donor-specific cells can be easily manipulated, and combination of iPSCs and iNs allows for ‘old-versus-rejuvenated’ approaches, while both systems obviously lack the in-vivo context and do not allow for lifespan or behavioral experiments.

Nuclear lamina in aging

The nuclear lamina maintains the integrity of nuclear structure and heterochromatin and also contributes to the regulation of gene expression (Figure 2)(Peric-Hupkes and van Steensel, 2010). Hutchinson-Gilford progeria syndrome (HGPS), a syndrome with accelerated premature aging, has been utilized as a model of cellular aging. Both human cells derived from patients with HGPS and animal models of HGPS have been utilized to study this aging syndrome (Butin-Israeli et al., 2012). The expression of progerin, a mutant form of LaminA and one of the genes responsible for HGPS, disrupts nuclear architecture, induces DNA damage, and promotes loss of heterochromatin (Goldman et al., 2004; Kudlow et al., 2007; Shumaker et al., 2006). These phenomena recapitulate features of physiological cellular aging (Pal and Tyler, 2016); it has been suggested that a disruption of Lamin-dependent nuclear architecture underlies cellular aging (Scaffidi and Misteli, 2008). However, presumably due to the low level of LaminA expression in the central nervous system (Jung et al., 2012; 2013), a progeria knock-in mouse model and many HGPS patients lack several key features of aging in the brain (Yang et al., 2005). Further, ectopic expression of progerin in mice did not change gene expression in the hippocampus despite its pronounced effect on nuclear architecture (Baek et al., 2015). Lamin proteins play a critical role in the segregation of damaged proteins in neural stem cells and may underlie stem cell aging (Moore et al., 2015). In addition, in a model for Alzheimer’s disease pathology in Drosophila melanogaster and in human neurons from AD cases, Lamin-dependent nuclear architecture and gene repression were disrupted (Frost et al., 2016; 2014; Sun et al., 2018). Thus, Lamin-dependent nuclear architecture appears to play vital roles in the aging of the CNS, but the function of these proteins could be cell type-specific and context-dependent during physiological and pathological aging (Nativio et al., 2018b).

Figure 2. Roles of the nuclear pore complex (NPC) and the nuclear lamina in gene expression regulation and cellular aging.

NPCs and the nuclear lamina mesh interact with transcription factors (TFs) and also directly bind to chromatin, thereby regulating cell type-specific gene expression. Accumulation of damage on NPC structures and the nuclear lamina through aging or other molecular insults compromises these cell type-specific gene regulation.

Nuclear pore complex deterioration in aging

The NPC is composed of a scaffold of nucleoporins which make up a cytoplasmic ring, inner pore ring, nuclear ring, and a peripheral compartment containing a nuclear basket and cytoplasmic filaments (Beck and Hurt, 2017). Scaffold nucleoporins have been identified as long-lived, low-turnover proteins with a limited capacity for renewal and repair in postmitotic cells (Daigle et al., 2001; Toyama et al., 2013). In contrast, nucleoporins of the peripheral compartments show a short half-life (Mathieson et al., 2018). However, the implications of these differences are currently unclear. Active nuclear transport of macromolecules to and from the nucleus via the NPC is mediated by the small GTP-binding nuclear protein Ran, which interacts with nucleoporins at the nuclear envelope. Similar to nuclear lamins, NPC interacts with chromatin to regulate global chromatin organization and cell type-specific gene expression (Figure 2)(Breuer and Ohkura, 2015; Ibarra et al., 2016; Jacinto et al., 2015; Toda et al., 2017). Since nuclear pore proteins and lamins are very long-lived proteins and accumulate damage during aging (D’Angelo et al., 2012; DaposAngelo et al., 2009; Toyama et al., 2013; Toyama and Hetzer, 2013), it is likely that age-related defects in the lamina and NPC directly lead to aging-associated changes in the epigenetic landscape (Buchwalter et al., 2019). However, dedicated studies in animal models have yet to be performed to unravel the molecular mechanisms involved.

Nuclear periphery in aging human neurons

Studies of human post-mortem tissue have yielded fundamental insights into cellular mechanisms involved in disease and aging; new technologies such as single-cell, spatial genomic, and protein analysis are beginning to shed light on cell-type specific and mosaic features of aging (Arrojo E Drigo et al., 2019; Mathys et al., 2019; Ximerakis et al., 2019). However, the cellular and molecular insights gleaned from postmortem human tissue are limited, and it is challenging to construct animal models that are predictive due to different physiologies, non-human genetic and molecular compositions, and vastly different life-spans and aging rates (Folgueras et al., 2018). Thus, to establish mechanistic links and causality in the context of human aging, it will be necessary to utilize human neural cell reprogramming models. Human neurons and other cell types can be generated in a patient-specific manner, are amenable to genetic manipulation, and are accessible for most experimental paradigms including functional and live-cell experimentation (Mertens et al., 2016)(Table 1). Human induced pluripotent stem cells (iPSCs) are an attractive model to study the functional implications of nuclear periphery-related and epigenetic ages on neurons, but cellular reprogramming seems to fully rejuvenate old cells, resetting the epigenetic age to pre-natal ages (negative values) (Sardo et al., 2016; Singh et al., 2019). Direct reprogramming of human somatic cells into induced neurons (iNs) appears more appropriate to study age-related phenotypes (Böhnke et al., 2019). Unlike iPSC-derived neurons, iNs show mature neuronal 5-hmC methylation and show transcriptomic aging signatures and preservation of their donors’ DNAm clock ages (Table 1)(Herdy et al., 2019; Huh et al., 2016)., thus underlining their usefulness to study age-related phenotypes in an adult-like human neuronal model (Luo et al., 2016; 2019).

Studies of human fibroblast-derived iNs from newborns, adults, and elderly healthy donors demonstrated age-related protein compartmentalization, nuclear transport dysfunction, and transcriptional changes associated with age. Age-dependent loss of the nucleoporin and importin-β family transport protein RanBP17 appeared to be at least in part causal for the observed ‘nuclear leakiness’, and knock-down of RanBP17 further led to a partial acquisition of an aged transcriptome in young fibroblasts (Figure 2)(T. Lu et al., 2004; Mertens et al., 2015). Further, iNs from old donors showed aging hallmarks that included a loss of H3K9me3, nuclear lamina-associated protein 2α (Lap2α), heterochromatin protein 1γ (HP1γ), as well as disturbed lamina morphologies (Tang et al., 2017). These features of aging were not found in iPSC-derived neurons from the same cohort. No direct causal link between nuclear periphery deterioration and epigenetic marks of aging were established in these cells. INs from familial amyotrophic lateral sclerosis patients revealed compartmentalization defects of the regulator of chromosome condensation 1 (RCC1) and the RNA-binding protein FUS, as well as electrophysiological defects of the neurons (Jovičić et al., 2015; 2016; M.-L. Liu et al., 2016). Since mutated tau also impairs nucleocytoplasmic transport (Eftekharzadeh et al., 2018), it is tempting to speculate that the nuclear pore complex is a common target from physiological aging to pathological phenomena induced by neurodegenerative diseases, and may explain why aging is the most critical risk factor for neurodegenerative diseases. To support this hypothesis, it remains to be seen to what extent the observed neuronal defects were a result of epigenetic and transcriptomic changes guided by a defective nuclear periphery/nuclear pores, or whether these issues were directly caused by the damaged nuclear structures. Indications that the architecture of the nuclear periphery is likely upstream of other disease-related aging features in human neurons come from a study of overexpressed progerin, the truncated HGPS-causing version of LaminA, in otherwise rejuvenated iPSC-derived neurons. The authors observed the formation of premature aging signs in neurons, including structural nuclear envelope impairments, and were able to show that ‘progeria-induced’ neurons from PD patients developed disease-related phenotypes including oxidative stress and neuronal defects (J. D. Miller et al., 2013; Studer et al., 2015).

These studies indicate a central role for nuclear architecture and epigenetic changes in the aging human brain; however, the mechanisms linking the nuclear lamina and NPC aging to epigenetic aging are still poorly understood. While many models of aging do not entirely recapitulate aging human neurons in vivo, iNs appear to be particularly useful to study these links in a human model of postmitotic neurons. These advances in model systems will hopefully complement preexisting post-mortem and animal studies and prove useful in moving the field towards a better understanding of the role of nuclear architecture and chromatin landscape in healthy aging and disease.

Conclusion and outlook

Accumulating evidence from postmortem human brain studies, animal models, and models of reprogrammed human neurons suggests a central role for age-dependent damage at the nuclear periphery in epigenetic regulation. The nuclear periphery, specifically the long-lived components of the nuclear laminar and NPC, are prime targets of cellular aging. Global changes in the epigenetic landscape including histone modifications and DNA methylation are central hallmarks of brain aging, and age-dependent changes in the nuclear periphery can markedly influence the epigenetic landscape. Since it is difficult to address brain aging in human samples due to limited access to patient samples, research utilizing transgenic rodent models has provided substantial information on how aging-related nuclear periphery damage and age-related epigenetic changes influence a myriad of molecular pathways, brain physiology, and behavior. Further, human donor- and patient-derived neural reprogramming-based cell models further underline the importance of nuclear periphery in cellular aging of the brain and suggest a potential role for nuclear lamina and NPCs in orchestrating aging features, including epigenetic changes. The brain is composed of thousands of different neural cell types and the effects of aging are likely to vary among different brain regions and cell types, perhaps even in opposite directions (Ximerakis et al., 2019). Therefore, cell type-specific, single-cell, and high-resolution spatial approaches that account for cell type diversity, and mosaic aging are desired for future studies in human brain samples and animal tissues. Direct evidence for a physiologically aged human nuclear peripheral structure is lacking, and many aspects of this proposed link remain to be explored. Investigation of mechanisms underlying age-dependent epigenetic changes may bring a major breakthrough for developing therapeutic means against age-dependent diseases of the brain.